computerized decision-support systems for chronic pain management in primary care
TRANSCRIPT
PAIN MEDICINE
Volume 8
bull
Number S3
bull
2007
copy American Academy of Pain Medicine 1526-237507$1500S155 S155ndashS166 doi101111j1526-4637200700278x
Blackwell Publishing IncMalden USAPMEPain Medicine1526-2375American Academy of Pain Medicine 20078S3S155S166Review Articles
Computerized Decision-Support System for PainSmith et al
Reprint requests to
Meredith Y Smith MPA PhD PurduePharma LP One Stamford Forum Stamford CT 06901USA Tel 203-588-8248 Fax 203-588-6242 E-mailmeredithsmithpharmacom
Computerized Decision-Support Systems for Chronic Pain Management in Primary Care
Meredith Y Smith MPA PhD Judith D DePue EdD MPH
dagger
and Christine Rini PhD
Mount Sinai School of Medicine New York New York
dagger
Miriam HospitalBrown Medical School Providence Rhode Island
A B S T R A C T
USA
ABSTRACT
Objective
Computerized decision-support systems (CDSSs) can offer clinical guidance as well aspromote doctorndashpatient collaboration and patient self-care As such they have great potential forimproving chronic pain management particularly in the primary care setting where physiciansoften lack sufficient pain-specific clinical expertise and communication skills The objective of thisstudy was to examine the use of CDSSs in chronic pain management and to review the evidencefor their feasibility and effectiveness
Design
A review of the available literature using search terms associated with computerized deci-sion-support and chronic pain management Major databases searched included MEDLINECINAHL PsychINFO HealthSTAR EMBASE Cochrane Library Computer and InformationSystems Abstracts and Electronics and Communications Abstracts Descriptive and evaluativestudies were included
Results
Nine studies describing eight CDSSs met study inclusion criteria With but two exceptionsCDSSs were specific to a pain-related condition(s) All were designed to assist clinicians to managepain medically Aside from pain status input specifications differed markedly Evaluative studieswere exclusively feasibility studies and varied widely in design and level of description All werenonexperimental most were methodologically weak Two primary care studies were reportedPatient and clinician acceptability ratings of CDSSs ranged from moderate to high Due to insuf-ficient data definitive conclusions concerning the impact of CDSSs on provider performance andpatient outcomes were not possible
Conclusion
Research on CDSSs in chronic pain management is limited The effects of CDSSs onprovider and patient outcomes remain understudied and their potential to improve doctorndashpatientcollaboration and self-care largely untested
Key Words
Computerized Decision-Support Systems Expert System Primary Care Chronic
Pain DoctorndashPatient Communication Disease Management for Chronic Pain
Introduction
n estimated 9 of adults in America sufferfrom chronic pain and its sequelae and over
half of these individuals seek treatment for thiscondition from a primary care physician (PCP) [1]In recognition of this fact recent American Pain
A
Society guidelines advocate that PCPs should ldquo participate in the process of screening diagnosisand long-term follow-up treatment of patientswho suffer from chronic pain [j]ust as PCPs diag-nose and maintain patients with other chronicdiseases rdquo [2] In order to accomplish this goalhowever PCPs must be equipped not only withthe necessary clinical tools and expertise but alsowith the communication and related interpersonalskills to build and sustain a strong alliance withtheir patient Collaborative management thatstrengthens and supports self-care is recognized as
S156
Smith et al
the most appropriate and cost-effective way totreat chronic pain and is moreover the approachmost often preferred by both patients and healthcare providers [2ndash6]
In reality however deficits in physiciansrsquotraining and knowledge regarding pain manage-ment coupled with time constraints during theprimary care visit frequently prevent PCPs frommeeting these dual clinical and psychosocialexpectations [57ndash15] PCPs commonly focusmore on technical aspects of care when treatingchronic pain patients and less on promotingpatient self-management behaviors [7] In addi-tion doctorndashpatient communication regardingpain is frequently inadequate [16ndash19] Conse-quently physiciansrsquo perceptions of patient painare often incongruent with patientsrsquo self-ratingstreatment goals are frequently developed withoutpatient input and patient adherence to treatmentplans is less than optimal [20ndash25]
Faced with declining reimbursement rates fromboth public and private payers physicians need analternative that enables them to reconcile the twinimperatives of providing high-quality pain carewhile maximizing efficiency in their clinicalpractice Computerized decision-support sys-tems (CDSSs) offer one potential solution tothis dilemma CDSSs are information systemsdesigned to enhance the quality of clinical decisionmaking and to minimize deviations in clinical per-formance from accepted professional standards[26] Individual patient data are entered into thesystem whereupon predetermined algorithmsguided by a resident logic library of expert-basedclinical data generate patient-specific recommen-dations [26] Data can be collected via a personalcomputer or a handheld device such as a palmpilot Typically output is available in real time foruse during the medical visit
The process involved in using a CDSS based onpersonal computer technology is illustrated as fol-lows Upon arrival at a medical appointment apatient is directed to a computer workstationequipped with a laptop stylus and printer Fol-lowing instructions on the computer screen thepatient completes an electronic questionnaire con-taining the following types of items current painstatus (eg intensity location duration) past andpresent pain treatment regimens degree of adher-ence to prescribed pain medication side effects ofcurrent medication risk factors for opioid abuseimpact of pain on respondentsrsquo quality of life(QOL) psychological status pain-related copingand self-management strategies type and degree
of social support personal goals for daily livingand (for initial visit only) demographic informa-tion The data collected are automatically storedelectronically and then run through a series ofalgorithms In the final step a report is generated(electronically in hard-copy format or both) thatcontains a synopsis of the patientrsquos pain statusrecommendations as to possible improvements inthe current pain treatment regimen key issues todiscuss with the patient (eg problems with med-ication adherence or side effects signs of depres-sion) and a list of patient-targeted suggestionsdesigned to encourage pain self-management (egrecommendations to pursue a specific exerciseregimen)
Both the type and application of CDSSs havebeen expanding rapidly in recent years To dateCDSSs have been used to diagnose a variety ofmedical conditions to enhance the provision ofpreventive care (eg cancer screening vaccina-tion) to facilitate disease management and toassist with drug prescribing [27ndash32] Evidencesuggests that CDSSs can improve health carepractitioner performance and patient outcomesparticularly in the areas of vaccinations tetanusimmunizations breast cancer screening colorectalcancer screening cardiovascular risk reductionand smoking cessation [2633ndash36]
CDSSs can be designed not only to provideclinical guidance but to capture and integratepatientsrsquo perspective on their illness and to pro-mote positive patient health-related behaviorsSometimes referred to as ldquoexpert systemsrdquo thisvariant of a CDSS is particularly promising foruse in the clinical management of chronic painfor two reasons (i) pain is a highly subjectiveexperience hence patient-reported data areessential to obtain and (ii) successful outcomesof pain management are at least equally if notmore dependent on appropriate self-care (egself-monitoring of pain adherence to prescribedmedications regular exercise weight control)than on the quality of the clinical diagnosis orrecommended therapeutic regimen [5] Based onthe data entered the expert system producestailored recommendations often derived frombehavioral change models that are specific to theneeds of the individual patient [37] Recommen-dations are directed at either the clinician or thepatient or separate output can be generated foreach party To date expert systems have beenused widely in the arena of health behavior mod-ification such as smoking cessation and weightreduction [38ndash43]
Computerized Decision-Support System for Pain
S157
The purpose of this study was to systematicallyreview the research evidence for CDSSs to addressthe following questions (i) To what extent haveCDSSs been utilized in the context of chronic painmanagement (ii) What are the characteristics ofthese systems and (iii) To what degree have theybeen evaluated and in what types of clinicalsettings
Methods
Data Sources
We conducted an automated literature searchusing the Ovid search engine With the assistanceof a research librarian we searched the followingdatabases MEDLINE (1966 to April 2006)CINAHL (1982 to April 2006) PsychINFO (1967to April 2006) HealthSTAR (1981 to April 2006)EMBASE Cochrane Library Computer andInformation Systems Abstracts Electronics andCommunications Abstracts Proust Digital Disser-tations Computer Retrieval of Informationon Scientific Projects (CRISP) LISA ERICComputer and Information Systems Abstractsand Dissertation Abstracts Key search wordsemployed included the following computer-generated decision support systems and expert sys-tems Additional terms included chronic painprimary care tailored reports personalized com-puter-based information disease management forchronic pain patient goals pain diagnosis andmanagement decision support systems neuralnetworks and fuzzy logic We also conducted amanual search to supplement the automatedsearch The manual search was not limited in timeperiod and included articles that had been refer-enced in other articles
Study Selection Criteria
Eligibility for inclusion in the final set includedany studies describing the development andorapplication of a CDSS or expert system in thecontext of chronic pain management We defineda CDSS as any electronic system designed to assistin clinical decision making regarding chronic painmanagement and in which patient-specific assess-ments and recommendations were generated foruse by a clinician andor patient [44] Consistentwith this definition we excluded any CDSSs thatexamined pain as only one component of an over-all assessment of QOL as well as any thataddressed acute pain management only We alsoexcluded studies that were written in languagesother than English
Results
The cross-database search yielded 70 publishedarticles and three federally funded research studiesdescribing ongoing investigations No additionalstudies were identified via the manual search pro-cess Full-text articles were retrieved for all titlesconsidered to be potentially relevant by theauthors Nine studies describing eight discreteCDSSs were identified as meeting our inclusioncriteria Lack of homogeneity among the final setof studies precluded a quantitative meta-analysisof the data Due to these analytic constraints weconducted a descriptive literature review only
As shown in Table 1 all eight CDSSs weredesigned to assist in the diagnosis andor manage-ment of chronic pain Two of these systems thePain Management Advisor (PMA) and the Diag-nostic Headache Diary were also designed to offereducation to the health care provider One thePMA had an interactive capability that permittedusers to query the system for explanations thera-peutic rationale and therapy guidelines Twoother systems the SymptomReport and thePAIN
ReportIt
featured adjunctive software pro-grams (SymptomConsult PAIN
ConsultN
) thatwere expressly designed as interactive educationaltools for the patient
All but two of the CDSSs were designed totarget a specific type of chronic pain or pain-related condition These included headache (2)low back (1) and cancer-related (3) Input specifi-cations also varied widely both in terms of thetype and amount of data required and in terms ofthe party responsible for data entry (ie physicianother health care provider or patient) All of theCDSSs required detailed input regarding painsymptomology Half of the systems reviewed alsoelicited data on pain medications currently usedand three requested QOL information Otherdata such as psychological status history of priortherapies (both pharmaceutical and nondrug-related therapies) patient goals for pain care andbarriers to pain management were specified asrequired inputs in only two of the CDSSsreviewed Three of the CDSSs used standard psy-chometrically validated instruments (eg McGillPain Questionnaire Medical Outcomes StudyShort Form 36 Oswestry Low Back Pain Disabil-ity Index) to collect input data
The majority of the CDSSs (5 of 8) producedoutput that was intended for clinician use(eg physician or nurse) only Three targetedboth the clinician and the patient Output varied
S158
Smith et al
Tab
le 1
Key
fea
ture
s of
clin
ical
dec
isio
n-su
ppor
t sy
stem
s (C
DS
Ss)
dev
elop
ed f
or c
hron
ic p
ain
man
agem
ent
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
RH
INO
S [
45]
To a
ssis
t ph
ysic
ians
in t
he d
iagn
osis
of
chro
nic
head
ache
sor
fac
ial p
ain
Bas
ed o
n ex
pert
kno
wle
dge
of h
eada
che
and
faci
alpa
in s
peci
alis
ts U
ses
4 se
ts o
f co
nditi
onal
pr
obab
ility
-bas
ed
rule
s 1
) ex
clus
ive
rule
s (i
e
if pa
tient
has
dis
ease
D
s
he m
ust
have
sym
ptom
s S
1
S
2
S
n
) 2)
incl
usiv
e ru
les
3)
asso
ciat
e ru
les
and
4)
dise
ase
imag
e ru
les
Pro
toty
pe o
nly
no t
estin
gS
yste
m d
evel
oped
usi
ngth
e P
rolo
g-K
AB
Apr
ogra
mm
ing
lang
uage
R
uns
on p
erso
nal
com
pute
rs w
ith C
PU
808
6m
emor
y 38
2 K
byte
s
Phy
sici
an in
put
base
d on
patie
nt in
terv
iew
ndash
Pat
ient
dem
ogra
phic
sndash
Ons
et o
f he
adac
hendash
His
tory
sin
ce o
nset
ndash P
ain
char
acte
ristic
sndash
Neu
rolo
gica
l sig
nsas
soci
ated
with
pai
nndash
Sle
ep s
tatu
sndash
Per
sona
l and
fam
ilial
hist
ory
of h
eada
che
ndash Jo
lt he
adac
he (
Yes
No)
ndash S
cler
osis
of
retin
al a
rter
y
Initi
al o
utpu
t af
ter
first
set
of
scre
enin
g qu
estio
ns
ndash Li
st o
f di
seas
es n
ot r
ejec
ted
Out
put
afte
r ad
ditio
nal
ques
tion(
s)
ndash Li
st o
f po
ssib
le d
isea
ses
ndash D
iagn
ostic
con
clus
ions
ndash E
xpla
natio
n of
the
dis
ease
requ
ired
exam
inat
ions
and
sugg
este
d th
erap
yH
ospi
tal c
hart
can
be
prin
ted
from
the
inpu
t da
ta in
the
form
of
a n
atur
al la
ngua
ge
repr
esen
tatio
n
Phy
sici
ans
IVA
N [
46]
To p
rovi
dere
com
men
datio
nsfo
r co
ntro
lling
pai
nan
d pr
ovid
ing
sym
ptom
rel
ief
inca
ncer
os
teo-
and
rheu
mat
oid
arth
ritis
Cas
e-ba
sed
reas
onin
gst
rate
gy t
o re
cord
and
retr
ieve
info
rmat
ion
stor
edin
an
inte
rnal
kno
wle
dge
base
Pro
toty
pe o
nly
no t
estin
gIV
AN
sof
twar
e w
ritte
n in
LPA
Pro
log
prog
ram
min
gla
ngua
ge fo
r W
indo
ws
(ldquoW
inP
rolo
grdquo)
runs
on
PC
Win
dow
s 95
NT
Win
Pro
log
ndash P
ain
sym
ptom
che
cklis
tndash
Cur
rent
pai
n di
agno
sis
ndash C
urre
nt p
ain
trea
tmen
tre
gim
en
Com
pute
r sc
reen
dis
play
ndash
Dia
gnos
tic c
onfir
mat
ion
ndash D
escr
iptio
n of
sym
ptom
s th
atm
ay a
ppea
r la
ter
ndash Tr
eatm
ents
pro
ven
succ
essf
ulin
sim
ilar
or r
elat
ed c
ases
ndash P
ossi
ble
alte
rnat
ive
caus
esof
the
pai
n
Phy
sici
ans
and
patie
nts
The
Dia
gnos
ticH
eada
che
Dia
ry[4
7]
To e
duca
te a
ndpr
ovid
e di
agno
stic
supp
ort
to p
rimar
yca
re p
rovi
ders
inor
der
to im
prov
em
anag
emen
t of
head
ache
s
Rul
e-ba
sed
expe
rt s
yste
mus
ing
Boo
lean
logi
c A
set
of d
iagn
ostic
rul
es u
sed
tode
term
ine
a di
agno
sis
base
dup
on t
he d
ata
ente
red
inpa
tient
dia
ry D
iary
dat
a ar
etr
ansf
orm
ed in
to a
dia
gnos
isfo
llow
ing
the
Inte
rnat
iona
lH
eada
che
Soc
iety
rsquoscl
assi
ficat
ion
Pro
toty
pede
velo
ped
syst
em-g
ener
ated
diag
nose
s w
ere
valid
ated
aga
inst
phys
icia
n ex
pert
-ge
nera
ted
diag
nose
s
Sta
nd-a
lone
Win
dow
s 95
prog
ram
writ
ten
in D
elph
ipr
ogra
mm
ing
lang
uage
ndash P
atie
nt d
ata
ndash H
eada
che
diar
y en
trie
sndash
Med
icat
ions
use
d to
alle
viat
e he
adac
he
ndash D
iagn
osis
of
head
ache
typ
eP
CP
s
Pai
n M
anag
emen
tA
dvis
or (
PM
A)
(Nov
aInt
ellig
ence
Inc
S
an D
iego
C
A)
[48]
To e
nhan
ce p
rimar
yca
re p
rovi
ders
rsquo(P
CP
s) m
anag
emen
tof
chr
onic
pai
n
ndash R
elie
s on
rul
e-ba
sed
algo
rithm
s de
rived
fro
mex
pert
kno
wle
dge
ofpa
in s
peci
alis
tsndash
Use
r as
ked
a se
ries
ofqu
estio
ns t
o re
fine
the
diag
nosi
s an
d de
term
ine
appr
opria
te t
hera
pyndash
Inte
ract
ive
capa
bilit
y(e
g
for
expl
anat
ions
th
erap
eutic
rat
iona
les
ther
apy
guid
elin
es)
Wor
king
ver
sion
deve
lope
d s
ome
field
tes
ting
cond
ucte
d
ndash P
entiu
m-b
ased
PC
sndash
Win
dow
s 95
ndash P
MA
writ
ten
inM
icro
Sof
t Vis
ual B
asic
v
50
ru
n as
an
expe
rtap
plic
atio
n in
Xpe
rtR
ule
ndash A
lgor
ithm
s st
ored
inM
icro
Sof
t A
cces
s da
taba
sendash
Mic
roS
oft
Hel
p U
tility
used
for
expl
anat
ions
and
quer
ies
ndash P
atie
nt d
emog
raph
ics
ndash D
iagn
osis
ndash P
ain
char
acte
ristic
sndash
Labo
rato
ry t
ests
ampim
agin
g st
udie
sndash
Cur
rent
med
icat
ions
ndash P
rior
ther
apie
sndash
Con
curr
ent
dise
ase
cond
ition
sndash
Alle
rgie
sndash
Psy
chol
ogic
al s
tatu
s
ndash A
prio
ritiz
ed li
st o
fre
com
men
datio
ns 1
) m
edic
alm
anag
emen
t (p
harm
acol
ogic
and
nonp
harm
acol
ogic
alm
anag
emen
t ph
ysic
al
psyc
hoso
cial
mod
aliti
es)
2)in
vasi
ve p
roce
dure
s 3
) re
ferr
als
PC
Ps
Computerized Decision-Support System for Pain
S159
Sym
ptom
Rep
ort
and
Sym
ptom
Con
sult
[49]
To a
ssis
t cl
inic
ians
in a
sses
sing
can
cer-
rela
ted
chro
nic
pain
and
fatig
ue
and
clar
ify p
atie
ntsrsquo
mis
belie
fs a
bout
pain
ass
essm
ent
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
Pat
ient
sel
f-re
port
ed in
put
ndash P
atie
nt d
emog
raph
ics
ndash 19
70 v
ersi
on o
f M
cGill
Pai
n Q
uest
ionn
aire
(M
PQ
)ndash
Bar
riers
Que
stio
nnai
re(B
Q)
ndash S
chw
artz
can
cer
fatig
uesc
ale
(SC
FS
-6)
1) H
ard
copi
es o
f ex
pert
sys
tem
repo
rt g
iven
to
patie
nt a
nd t
ocl
inic
ian
2) P
atie
nt r
ecei
ves
educ
atio
nal
mat
eria
ls o
n ho
w t
o re
port
pa
in
use
pain
med
icat
ions
sa
fely
an
d m
anag
e fa
tigue
M
ater
ials
are
cus
tom
ized
to
the
patie
ntrsquos
nee
ds a
nd
pres
ente
d in
an
inte
ract
ive
m
ultim
edia
form
at P
atie
nts
have
opt
ion
to re
ad o
r lis
ten
to
info
rmat
ion
on t
he c
ompu
ter
prin
t an
y or
all
of t
he
mat
eria
ls
or d
o bo
th
Onc
olog
ynu
rses
ot
her
clin
icia
ns
Dec
isio
n-su
ppor
tco
mpu
ter
prog
ram
for
canc
er p
ain
man
agem
ent
[50]
To im
prov
e th
eon
colo
gy n
urse
srsquode
cisi
on m
akin
gre
late
d to
can
cer
pain
man
agem
ent
amon
g cu
ltura
llydi
vers
e fe
mal
eon
colo
gy p
atie
nts
ndash S
urve
y da
ta o
n m
ultic
ultu
ral
canc
er p
ain
char
acte
ristic
sw
ere
anal
yzed
usi
ng f
uzzy
infe
renc
e lo
gic
to d
evel
op 4
mod
ules
1)
a ge
neric
know
ledg
e ba
se 2
) a
cultu
re-s
peci
fic k
now
ledg
e ba
se 3
) de
cisi
on-m
akin
g
and
4) s
elf-
adap
tatio
nndash
Dec
isio
n-m
akin
g m
odul
eco
nsis
ts o
f 2
sets
of
fuzz
yin
fere
nce
logi
c de
velo
ped
via
a ge
netic
alg
orith
m
Har
dwar
e no
t de
scrib
edndash
Ada
ptiv
e fu
zzy
logi
cso
ftwar
e us
ed t
o de
velo
pan
d ru
n th
e kn
owle
dge
base
gen
erat
ion
and
the
deci
sion
-mak
ing
and
self-
adap
tatio
n m
odul
es
Nur
se-e
nter
ed d
ata
base
don
pat
ient
inte
rvie
w
ndash P
atie
nt d
emog
raph
ics
ndash P
ain
char
acte
ristic
s
Com
pute
r sc
reen
dis
play
of
anal
gesi
c tr
eatm
ent
reco
mm
enda
tions
bas
ed o
n th
eW
orld
Hea
lth O
rgan
izat
ion
(WH
O)rsquos
ana
lges
ic la
dder
Onc
olog
ynu
rses
PAIN
Rep
or
tIt a
ndPA
IN
Con
sultN
[51
52]
To a
ssis
t cl
inic
ians
in a
sses
sing
chr
onic
pain
and
to
educ
ate
patie
nts
rega
rdin
gpa
in m
onito
ring
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
data
sto
red
in A
cces
s 97
data
base
Pat
ient
sel
f-re
port
ed in
put
1)
Pat
ient
dem
ogra
phic
sndash
McG
ill P
ain
Que
stio
nnai
rendash
Pai
n st
atus
ndash P
atie
nt g
oals
for
and
expe
ctat
ions
abo
ut p
ain
ndash Ty
pe a
nd e
ffect
iven
ess
of p
revi
ous
pain
trea
tmen
ts
Har
d co
pies
of
expe
rt s
yste
mre
port
giv
en t
o pa
tient
and
to
clin
icia
n o
n-sc
reen
vie
win
g of
repo
rt is
als
o po
ssib
le
Onc
olog
ynu
rses
and
othe
rcl
inic
ians
Touc
h-sc
reen
Com
pute
rA
sses
smen
t of
Chr
onic
Low
Bac
k P
ain
[53]
To c
olle
ct p
ain
sym
ptom
sta
tus
and
othe
r he
alth
info
rmat
ion
from
patie
nts
with
low
back
pai
n
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
lim
ited
field
tes
ting
Web
-bas
ed s
yste
m u
sing
Del
l Ins
piro
n 11
00 la
ptop
with
Mic
roso
ft X
P o
pera
ting
syst
em 1
4
prime
tou
ch s
cree
n(M
agic
Tou
ch
Key
tec)
ndash P
atie
nt d
emog
raph
ics
ndash O
swes
try
Low
Bac
k P
ain
Dis
abili
ty I
ndex
(V
ersi
on 2
)ndash
Bec
k D
epre
ssio
n In
vent
ory
ndash M
OS
Sho
rt F
orm
-36
(MO
S S
F-3
6)
Not
des
crib
edP
hysi
cian
s
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
S160
Smith et al
considerably in terms of content format anddelivery (eg electronic paper or both) SeveralCDSSs scored and summated patient responseson standard pain and QOL-related assessmentmeasures Based on the published descriptions atleast five of the CDSSs were designed to gener-ate output in real time at the patientrsquos medicalvisit
In terms of systems architecture all CDSSsreviewed were stand-alone personal computer-based systems None interfaced with existingelectronic medical records systems pharmacyappointment scheduling or laboratory resultsreporting Three either were Web-based or hadthe capacity to use a Web-based platform
Table 2 summarizes the types of studies con-ducted to date to evaluate chronic pain CDSSs Ofthe eight CDSSs identified five had publishedevaluation results With one exception all werefeasibility studies exclusively Studies were con-ducted in both inpatient and outpatient settingsAmong the outpatient studies two had been con-ducted with PCPs the remainder involved special-ists in tertiary care settings Study designs used forevaluating these CDSSs varied two were cross-sectional two involved immediate pre- and post-assessments of CDSS use one was longitudinal(12-month follow-up) and one was a focus groupStudy sample sizes ranged from 213 to 4 with themajority having 50 or fewer subjects
Patient acceptability of the CDSS was the sin-gle most commonly assessed variable Evaluationsof both the SymptomReport and the PAIN
Report-It
employed a common tool to assess patientacceptability Results across four pilot testsinvolving a total of 254 subjects consistentlyshowed high acceptability of these two CDSSsand 100 completion rates in terms of data input[495152] The average amount of time requiredranged from a high of 38 minutes to a low of14 minutes
Two studies addressed the issue of medicalaccuracy of the system-generated recommenda-tion Both studies examined this issue by compar-ing system-generated diagnoses andor treatmentrecommendations with those generated by physi-cian experts based on a select sample of patientcases Results showed moderate to high agreementbetween the system- and expert-generated recom-mendations [4748]
Clinician perceptions concerning ease of useand value of a CDSS for chronic pain managementwere examined in two studies Overall physiciansfound the system to be moderately easy to use and
of some clinical worth [4852] Knab and col-leagues [48] reported that the average cliniciantime spent per case on the PMA to obtain outputwas approximately 5 minutes (
plusmn
34 minutes) [48]In addition 85 of physicians adopted the PMA-generated recommendations and 25 of thestudy patients seen were referred to a pain special-ist with the average time to referral being37 months (
plusmn
06 months) [48]Analyses concerning the impact of CDSSs on
patient outcomes were limited Huang and col-leagues [51] assessed changes in pain intensity pre-and post-CDSS use in a sample of radiationoncology clinic patients Although there was adownward trend in pain levels over time resultswere not statistically significant possibly dueto the small size of the sample (N
=
15) Wilkieand colleagues [4952] reported qualitative dataregarding the impact of the SymptomReport andthe PAIN
ReportIt
on patient behavior Resultswere contradictory Of the 41 outpatients whoused the SymptomReport approximately 68stated that it had not affected their pain-relatedcommunication Some however felt that theytalked more precisely and explicitly about theirpain as a result of using it Other comments asso-ciated with use of the SymptomReport includedan increased awareness of pain symptoms andgreater compliance with pain symptom manage-ment In contrast 86 of users of the PAIN-
ReportIt
cited it as beneficial for patientndashdoctorpain-related communication and that it ldquofreedthem to describe their painrdquo
Discussion
Over the past two decades there has been a gradualbut steady growth in research on the use of CDSSsfor chronic pain management To date the numberof such systems is small but expanding Advanceshave also been evident in terms of both the quantityand quality of evaluation studies conducted Whilethe earliest versions were presented in the litera-ture as prototypes only [4546] CDSSs developedsince 2001 have all undergone some form of fieldtesting The majority of these studies howeverhave been nonexperimental in design and focusedexclusively on process measures such as patientor clinician ratings of system acceptability andusability Other salient process measures such asthe degree to which the clinician andor patientactually reviewed and utilized system output orhad confidence in its accuracy have not been con-sistently assessed Poor usability and practitioner
Computerized Decision-Support System for Pain
S161
Tab
le 2
Eva
luat
ion
of c
ompu
teri
zed
deci
sion
-sup
port
sys
tem
s (C
DS
Ss)
in c
hron
ic p
ain
man
agem
ent
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
Nie
lsen
et a
l
[47
]D
iagn
ostic
Hea
dach
eD
iary
PC
Ps
Not
des
crib
edndash
a
gree
men
t of c
ompu
ter-
gene
rate
d vs
exp
ert
phys
icia
n di
agno
ses
ndash 10
0 a
gree
men
t of c
ompu
ter-
gene
rate
d vs
ex
pert
phy
sici
an d
iagn
oses
Kna
b et
al
[48
]P
ain
Man
agem
ent
Adv
isor
N
=
50
PC
Ps
N
=
50
chro
nic
pain
pat
ient
sLo
ngitu
dina
l w
ith 1
2-m
onth
PC
P fo
llow
-up
ndash E
ase
of u
sendash
Med
ical
app
ropr
iate
ness
of
reco
mm
enda
tions
ndash
phy
sici
ansrsquo
ado
ptin
g P
ain
Man
agem
ent
Adv
isor
tre
atm
ent
reco
mm
enda
tions
ndash
pat
ient
s re
ferr
ed t
o pa
insp
ecia
lty c
linic
Phy
sici
an a
dopt
ion
of s
yste
m-g
ener
ated
re
com
men
datio
ns 8
5 o
f ca
ses
Ave
rage
phy
sici
an t
ime
spen
t pe
r ca
se
49
min
utes
(S
D
plusmn
34)
Eas
e of
use
as
rate
d by
phy
sici
ans
42
(
plusmn
28
cm)
on s
cale
of
1ndash10
25
of
patie
nts
refe
rred
to
pain
spe
cial
ty
clin
icA
vera
ge t
ime
to p
ain
spec
ialty
ref
erra
l 3
7 m
onth
s (S
D
plusmn
06)
ndash 70
o
f no
nref
erre
d pa
tient
s st
ill r
ecei
ving
co
mpu
ter-
reco
mm
ende
d tr
eatm
ent
1 ye
ar
post
Wilk
ie e
t al
[49
]S
ympt
omR
epor
tN
=
41
outp
atie
nts
with
can
cer
Cro
ss-s
ectio
nal
tele
phon
e in
terv
iew
sndash
13-it
em p
atie
nt a
ccep
tabi
lity
scal
e as
sess
ing
ease
of
use
ofS
ympt
omR
epor
tndash
Inpu
t co
mpl
etio
n tim
endash
Qua
litat
ive
asse
ssm
ent
of d
egre
eof
com
mun
icat
ion
with
hea
lth c
are
prov
ider
s re
gard
ing
pain
and
oth
ersy
mpt
oms
ndash M
ean
time
to c
ompl
ete
Sym
ptom
Rep
ort
382
min
utes
(S
D
plusmn
202
)ndash
Mea
n tim
e to
com
plet
e S
ympt
omC
onsu
lt
209
min
utes
(S
D
plusmn
18
6)ndash
71
of
part
icip
ants
rat
ed S
ympt
omR
epor
t as
eas
y e
njoy
able
an
d in
form
ativ
endash
68
rep
orte
d th
at th
e am
ount
and
con
tent
of
the
ir pa
in-r
elat
ed c
omm
unic
atio
n w
ith
thei
r do
ctor
had
not
cha
nged
muc
hndash
Qua
litat
ive
patie
nt c
omm
ents
1)
help
ed
them
tal
k m
ore
expl
icitl
y ab
out
pain
2)
gave
the
m g
reat
er a
war
enes
s of
pai
n sy
mpt
oms
3)
incr
ease
d un
ders
tand
ing
of
and
enha
nced
com
plia
nce
rega
rdin
g sy
mpt
om m
anag
emen
t ndash
Pat
ient
com
men
ts r
e S
ympt
omC
onsu
lt 1
) co
ntai
ned
too
muc
h in
form
atio
n 2
) no
t ta
rget
ed t
o in
divi
dual
nee
ds (
3) p
rovi
ded
no n
ew in
form
atio
n
Wilk
ie e
t al
[52
]PA
IN
Rep
ortIt
N
=
213
of
who
m 1
06w
ere
canc
er in
patie
nts
10 m
etas
tatic
can
cer
outp
atie
nts
and
97
wer
e in
divi
dual
sex
perie
ncin
g ac
ute
orch
roni
c pa
in r
ecru
ited
from
non
heal
th c
are
setti
ngs
Des
crip
tive
cro
ss-
sect
iona
l stu
dy in
3se
tting
s 1
) te
rtia
ryca
re
2) r
adia
tion
onco
logy
clin
ic
3)m
obile
clin
ic
ndash 13
-item
sca
le m
easu
ring
acce
ptab
ility
of
PAIN
Rep
ortIt
(ie
tim
e to
com
plet
e e
ase
of u
se
unde
rsta
ndab
ility
of
dire
ctio
ns
ergo
nom
ic e
lem
ents
of
syst
em
and
com
plet
enes
s of
res
pons
es)
ndash 86
o
f res
pond
ents
rate
d th
e PA
IN
Rep
ortIt
as
ben
efici
al f
or p
ain
com
mun
icat
ion
ndash 10
0 p
atie
nt c
ompl
etio
n ra
te o
f PA
IN
Rep
ortIt
ndash M
ean
time
to p
atie
nt
com
plet
ion
15
8 m
inut
es (
SD
plusmn
67
)ndash
Mea
n pa
tient
acc
epta
bilit
y sc
ore
11
7 (S
D
plusmn
16
) sc
ores
ran
ged
from
6 t
o 13
80
o
f pa
tient
s ra
ted
acce
ptab
ility
as
grea
ter
than
min
imum
crit
erio
n of
10
ndash U
ser
com
men
ts 1
) so
me
mec
hani
cal
diffi
culti
es 2
) re
activ
ity t
o us
e of
sys
tem
(e
g
vom
iting
) 3)
pre
fere
nce
to r
elay
info
rmat
ion
dire
ctly
to
prov
ider
PC
P p
rimar
y ca
re p
hysi
cian
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
S156
Smith et al
the most appropriate and cost-effective way totreat chronic pain and is moreover the approachmost often preferred by both patients and healthcare providers [2ndash6]
In reality however deficits in physiciansrsquotraining and knowledge regarding pain manage-ment coupled with time constraints during theprimary care visit frequently prevent PCPs frommeeting these dual clinical and psychosocialexpectations [57ndash15] PCPs commonly focusmore on technical aspects of care when treatingchronic pain patients and less on promotingpatient self-management behaviors [7] In addi-tion doctorndashpatient communication regardingpain is frequently inadequate [16ndash19] Conse-quently physiciansrsquo perceptions of patient painare often incongruent with patientsrsquo self-ratingstreatment goals are frequently developed withoutpatient input and patient adherence to treatmentplans is less than optimal [20ndash25]
Faced with declining reimbursement rates fromboth public and private payers physicians need analternative that enables them to reconcile the twinimperatives of providing high-quality pain carewhile maximizing efficiency in their clinicalpractice Computerized decision-support sys-tems (CDSSs) offer one potential solution tothis dilemma CDSSs are information systemsdesigned to enhance the quality of clinical decisionmaking and to minimize deviations in clinical per-formance from accepted professional standards[26] Individual patient data are entered into thesystem whereupon predetermined algorithmsguided by a resident logic library of expert-basedclinical data generate patient-specific recommen-dations [26] Data can be collected via a personalcomputer or a handheld device such as a palmpilot Typically output is available in real time foruse during the medical visit
The process involved in using a CDSS based onpersonal computer technology is illustrated as fol-lows Upon arrival at a medical appointment apatient is directed to a computer workstationequipped with a laptop stylus and printer Fol-lowing instructions on the computer screen thepatient completes an electronic questionnaire con-taining the following types of items current painstatus (eg intensity location duration) past andpresent pain treatment regimens degree of adher-ence to prescribed pain medication side effects ofcurrent medication risk factors for opioid abuseimpact of pain on respondentsrsquo quality of life(QOL) psychological status pain-related copingand self-management strategies type and degree
of social support personal goals for daily livingand (for initial visit only) demographic informa-tion The data collected are automatically storedelectronically and then run through a series ofalgorithms In the final step a report is generated(electronically in hard-copy format or both) thatcontains a synopsis of the patientrsquos pain statusrecommendations as to possible improvements inthe current pain treatment regimen key issues todiscuss with the patient (eg problems with med-ication adherence or side effects signs of depres-sion) and a list of patient-targeted suggestionsdesigned to encourage pain self-management (egrecommendations to pursue a specific exerciseregimen)
Both the type and application of CDSSs havebeen expanding rapidly in recent years To dateCDSSs have been used to diagnose a variety ofmedical conditions to enhance the provision ofpreventive care (eg cancer screening vaccina-tion) to facilitate disease management and toassist with drug prescribing [27ndash32] Evidencesuggests that CDSSs can improve health carepractitioner performance and patient outcomesparticularly in the areas of vaccinations tetanusimmunizations breast cancer screening colorectalcancer screening cardiovascular risk reductionand smoking cessation [2633ndash36]
CDSSs can be designed not only to provideclinical guidance but to capture and integratepatientsrsquo perspective on their illness and to pro-mote positive patient health-related behaviorsSometimes referred to as ldquoexpert systemsrdquo thisvariant of a CDSS is particularly promising foruse in the clinical management of chronic painfor two reasons (i) pain is a highly subjectiveexperience hence patient-reported data areessential to obtain and (ii) successful outcomesof pain management are at least equally if notmore dependent on appropriate self-care (egself-monitoring of pain adherence to prescribedmedications regular exercise weight control)than on the quality of the clinical diagnosis orrecommended therapeutic regimen [5] Based onthe data entered the expert system producestailored recommendations often derived frombehavioral change models that are specific to theneeds of the individual patient [37] Recommen-dations are directed at either the clinician or thepatient or separate output can be generated foreach party To date expert systems have beenused widely in the arena of health behavior mod-ification such as smoking cessation and weightreduction [38ndash43]
Computerized Decision-Support System for Pain
S157
The purpose of this study was to systematicallyreview the research evidence for CDSSs to addressthe following questions (i) To what extent haveCDSSs been utilized in the context of chronic painmanagement (ii) What are the characteristics ofthese systems and (iii) To what degree have theybeen evaluated and in what types of clinicalsettings
Methods
Data Sources
We conducted an automated literature searchusing the Ovid search engine With the assistanceof a research librarian we searched the followingdatabases MEDLINE (1966 to April 2006)CINAHL (1982 to April 2006) PsychINFO (1967to April 2006) HealthSTAR (1981 to April 2006)EMBASE Cochrane Library Computer andInformation Systems Abstracts Electronics andCommunications Abstracts Proust Digital Disser-tations Computer Retrieval of Informationon Scientific Projects (CRISP) LISA ERICComputer and Information Systems Abstractsand Dissertation Abstracts Key search wordsemployed included the following computer-generated decision support systems and expert sys-tems Additional terms included chronic painprimary care tailored reports personalized com-puter-based information disease management forchronic pain patient goals pain diagnosis andmanagement decision support systems neuralnetworks and fuzzy logic We also conducted amanual search to supplement the automatedsearch The manual search was not limited in timeperiod and included articles that had been refer-enced in other articles
Study Selection Criteria
Eligibility for inclusion in the final set includedany studies describing the development andorapplication of a CDSS or expert system in thecontext of chronic pain management We defineda CDSS as any electronic system designed to assistin clinical decision making regarding chronic painmanagement and in which patient-specific assess-ments and recommendations were generated foruse by a clinician andor patient [44] Consistentwith this definition we excluded any CDSSs thatexamined pain as only one component of an over-all assessment of QOL as well as any thataddressed acute pain management only We alsoexcluded studies that were written in languagesother than English
Results
The cross-database search yielded 70 publishedarticles and three federally funded research studiesdescribing ongoing investigations No additionalstudies were identified via the manual search pro-cess Full-text articles were retrieved for all titlesconsidered to be potentially relevant by theauthors Nine studies describing eight discreteCDSSs were identified as meeting our inclusioncriteria Lack of homogeneity among the final setof studies precluded a quantitative meta-analysisof the data Due to these analytic constraints weconducted a descriptive literature review only
As shown in Table 1 all eight CDSSs weredesigned to assist in the diagnosis andor manage-ment of chronic pain Two of these systems thePain Management Advisor (PMA) and the Diag-nostic Headache Diary were also designed to offereducation to the health care provider One thePMA had an interactive capability that permittedusers to query the system for explanations thera-peutic rationale and therapy guidelines Twoother systems the SymptomReport and thePAIN
ReportIt
featured adjunctive software pro-grams (SymptomConsult PAIN
ConsultN
) thatwere expressly designed as interactive educationaltools for the patient
All but two of the CDSSs were designed totarget a specific type of chronic pain or pain-related condition These included headache (2)low back (1) and cancer-related (3) Input specifi-cations also varied widely both in terms of thetype and amount of data required and in terms ofthe party responsible for data entry (ie physicianother health care provider or patient) All of theCDSSs required detailed input regarding painsymptomology Half of the systems reviewed alsoelicited data on pain medications currently usedand three requested QOL information Otherdata such as psychological status history of priortherapies (both pharmaceutical and nondrug-related therapies) patient goals for pain care andbarriers to pain management were specified asrequired inputs in only two of the CDSSsreviewed Three of the CDSSs used standard psy-chometrically validated instruments (eg McGillPain Questionnaire Medical Outcomes StudyShort Form 36 Oswestry Low Back Pain Disabil-ity Index) to collect input data
The majority of the CDSSs (5 of 8) producedoutput that was intended for clinician use(eg physician or nurse) only Three targetedboth the clinician and the patient Output varied
S158
Smith et al
Tab
le 1
Key
fea
ture
s of
clin
ical
dec
isio
n-su
ppor
t sy
stem
s (C
DS
Ss)
dev
elop
ed f
or c
hron
ic p
ain
man
agem
ent
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
RH
INO
S [
45]
To a
ssis
t ph
ysic
ians
in t
he d
iagn
osis
of
chro
nic
head
ache
sor
fac
ial p
ain
Bas
ed o
n ex
pert
kno
wle
dge
of h
eada
che
and
faci
alpa
in s
peci
alis
ts U
ses
4 se
ts o
f co
nditi
onal
pr
obab
ility
-bas
ed
rule
s 1
) ex
clus
ive
rule
s (i
e
if pa
tient
has
dis
ease
D
s
he m
ust
have
sym
ptom
s S
1
S
2
S
n
) 2)
incl
usiv
e ru
les
3)
asso
ciat
e ru
les
and
4)
dise
ase
imag
e ru
les
Pro
toty
pe o
nly
no t
estin
gS
yste
m d
evel
oped
usi
ngth
e P
rolo
g-K
AB
Apr
ogra
mm
ing
lang
uage
R
uns
on p
erso
nal
com
pute
rs w
ith C
PU
808
6m
emor
y 38
2 K
byte
s
Phy
sici
an in
put
base
d on
patie
nt in
terv
iew
ndash
Pat
ient
dem
ogra
phic
sndash
Ons
et o
f he
adac
hendash
His
tory
sin
ce o
nset
ndash P
ain
char
acte
ristic
sndash
Neu
rolo
gica
l sig
nsas
soci
ated
with
pai
nndash
Sle
ep s
tatu
sndash
Per
sona
l and
fam
ilial
hist
ory
of h
eada
che
ndash Jo
lt he
adac
he (
Yes
No)
ndash S
cler
osis
of
retin
al a
rter
y
Initi
al o
utpu
t af
ter
first
set
of
scre
enin
g qu
estio
ns
ndash Li
st o
f di
seas
es n
ot r
ejec
ted
Out
put
afte
r ad
ditio
nal
ques
tion(
s)
ndash Li
st o
f po
ssib
le d
isea
ses
ndash D
iagn
ostic
con
clus
ions
ndash E
xpla
natio
n of
the
dis
ease
requ
ired
exam
inat
ions
and
sugg
este
d th
erap
yH
ospi
tal c
hart
can
be
prin
ted
from
the
inpu
t da
ta in
the
form
of
a n
atur
al la
ngua
ge
repr
esen
tatio
n
Phy
sici
ans
IVA
N [
46]
To p
rovi
dere
com
men
datio
nsfo
r co
ntro
lling
pai
nan
d pr
ovid
ing
sym
ptom
rel
ief
inca
ncer
os
teo-
and
rheu
mat
oid
arth
ritis
Cas
e-ba
sed
reas
onin
gst
rate
gy t
o re
cord
and
retr
ieve
info
rmat
ion
stor
edin
an
inte
rnal
kno
wle
dge
base
Pro
toty
pe o
nly
no t
estin
gIV
AN
sof
twar
e w
ritte
n in
LPA
Pro
log
prog
ram
min
gla
ngua
ge fo
r W
indo
ws
(ldquoW
inP
rolo
grdquo)
runs
on
PC
Win
dow
s 95
NT
Win
Pro
log
ndash P
ain
sym
ptom
che
cklis
tndash
Cur
rent
pai
n di
agno
sis
ndash C
urre
nt p
ain
trea
tmen
tre
gim
en
Com
pute
r sc
reen
dis
play
ndash
Dia
gnos
tic c
onfir
mat
ion
ndash D
escr
iptio
n of
sym
ptom
s th
atm
ay a
ppea
r la
ter
ndash Tr
eatm
ents
pro
ven
succ
essf
ulin
sim
ilar
or r
elat
ed c
ases
ndash P
ossi
ble
alte
rnat
ive
caus
esof
the
pai
n
Phy
sici
ans
and
patie
nts
The
Dia
gnos
ticH
eada
che
Dia
ry[4
7]
To e
duca
te a
ndpr
ovid
e di
agno
stic
supp
ort
to p
rimar
yca
re p
rovi
ders
inor
der
to im
prov
em
anag
emen
t of
head
ache
s
Rul
e-ba
sed
expe
rt s
yste
mus
ing
Boo
lean
logi
c A
set
of d
iagn
ostic
rul
es u
sed
tode
term
ine
a di
agno
sis
base
dup
on t
he d
ata
ente
red
inpa
tient
dia
ry D
iary
dat
a ar
etr
ansf
orm
ed in
to a
dia
gnos
isfo
llow
ing
the
Inte
rnat
iona
lH
eada
che
Soc
iety
rsquoscl
assi
ficat
ion
Pro
toty
pede
velo
ped
syst
em-g
ener
ated
diag
nose
s w
ere
valid
ated
aga
inst
phys
icia
n ex
pert
-ge
nera
ted
diag
nose
s
Sta
nd-a
lone
Win
dow
s 95
prog
ram
writ
ten
in D
elph
ipr
ogra
mm
ing
lang
uage
ndash P
atie
nt d
ata
ndash H
eada
che
diar
y en
trie
sndash
Med
icat
ions
use
d to
alle
viat
e he
adac
he
ndash D
iagn
osis
of
head
ache
typ
eP
CP
s
Pai
n M
anag
emen
tA
dvis
or (
PM
A)
(Nov
aInt
ellig
ence
Inc
S
an D
iego
C
A)
[48]
To e
nhan
ce p
rimar
yca
re p
rovi
ders
rsquo(P
CP
s) m
anag
emen
tof
chr
onic
pai
n
ndash R
elie
s on
rul
e-ba
sed
algo
rithm
s de
rived
fro
mex
pert
kno
wle
dge
ofpa
in s
peci
alis
tsndash
Use
r as
ked
a se
ries
ofqu
estio
ns t
o re
fine
the
diag
nosi
s an
d de
term
ine
appr
opria
te t
hera
pyndash
Inte
ract
ive
capa
bilit
y(e
g
for
expl
anat
ions
th
erap
eutic
rat
iona
les
ther
apy
guid
elin
es)
Wor
king
ver
sion
deve
lope
d s
ome
field
tes
ting
cond
ucte
d
ndash P
entiu
m-b
ased
PC
sndash
Win
dow
s 95
ndash P
MA
writ
ten
inM
icro
Sof
t Vis
ual B
asic
v
50
ru
n as
an
expe
rtap
plic
atio
n in
Xpe
rtR
ule
ndash A
lgor
ithm
s st
ored
inM
icro
Sof
t A
cces
s da
taba
sendash
Mic
roS
oft
Hel
p U
tility
used
for
expl
anat
ions
and
quer
ies
ndash P
atie
nt d
emog
raph
ics
ndash D
iagn
osis
ndash P
ain
char
acte
ristic
sndash
Labo
rato
ry t
ests
ampim
agin
g st
udie
sndash
Cur
rent
med
icat
ions
ndash P
rior
ther
apie
sndash
Con
curr
ent
dise
ase
cond
ition
sndash
Alle
rgie
sndash
Psy
chol
ogic
al s
tatu
s
ndash A
prio
ritiz
ed li
st o
fre
com
men
datio
ns 1
) m
edic
alm
anag
emen
t (p
harm
acol
ogic
and
nonp
harm
acol
ogic
alm
anag
emen
t ph
ysic
al
psyc
hoso
cial
mod
aliti
es)
2)in
vasi
ve p
roce
dure
s 3
) re
ferr
als
PC
Ps
Computerized Decision-Support System for Pain
S159
Sym
ptom
Rep
ort
and
Sym
ptom
Con
sult
[49]
To a
ssis
t cl
inic
ians
in a
sses
sing
can
cer-
rela
ted
chro
nic
pain
and
fatig
ue
and
clar
ify p
atie
ntsrsquo
mis
belie
fs a
bout
pain
ass
essm
ent
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
Pat
ient
sel
f-re
port
ed in
put
ndash P
atie
nt d
emog
raph
ics
ndash 19
70 v
ersi
on o
f M
cGill
Pai
n Q
uest
ionn
aire
(M
PQ
)ndash
Bar
riers
Que
stio
nnai
re(B
Q)
ndash S
chw
artz
can
cer
fatig
uesc
ale
(SC
FS
-6)
1) H
ard
copi
es o
f ex
pert
sys
tem
repo
rt g
iven
to
patie
nt a
nd t
ocl
inic
ian
2) P
atie
nt r
ecei
ves
educ
atio
nal
mat
eria
ls o
n ho
w t
o re
port
pa
in
use
pain
med
icat
ions
sa
fely
an
d m
anag
e fa
tigue
M
ater
ials
are
cus
tom
ized
to
the
patie
ntrsquos
nee
ds a
nd
pres
ente
d in
an
inte
ract
ive
m
ultim
edia
form
at P
atie
nts
have
opt
ion
to re
ad o
r lis
ten
to
info
rmat
ion
on t
he c
ompu
ter
prin
t an
y or
all
of t
he
mat
eria
ls
or d
o bo
th
Onc
olog
ynu
rses
ot
her
clin
icia
ns
Dec
isio
n-su
ppor
tco
mpu
ter
prog
ram
for
canc
er p
ain
man
agem
ent
[50]
To im
prov
e th
eon
colo
gy n
urse
srsquode
cisi
on m
akin
gre
late
d to
can
cer
pain
man
agem
ent
amon
g cu
ltura
llydi
vers
e fe
mal
eon
colo
gy p
atie
nts
ndash S
urve
y da
ta o
n m
ultic
ultu
ral
canc
er p
ain
char
acte
ristic
sw
ere
anal
yzed
usi
ng f
uzzy
infe
renc
e lo
gic
to d
evel
op 4
mod
ules
1)
a ge
neric
know
ledg
e ba
se 2
) a
cultu
re-s
peci
fic k
now
ledg
e ba
se 3
) de
cisi
on-m
akin
g
and
4) s
elf-
adap
tatio
nndash
Dec
isio
n-m
akin
g m
odul
eco
nsis
ts o
f 2
sets
of
fuzz
yin
fere
nce
logi
c de
velo
ped
via
a ge
netic
alg
orith
m
Har
dwar
e no
t de
scrib
edndash
Ada
ptiv
e fu
zzy
logi
cso
ftwar
e us
ed t
o de
velo
pan
d ru
n th
e kn
owle
dge
base
gen
erat
ion
and
the
deci
sion
-mak
ing
and
self-
adap
tatio
n m
odul
es
Nur
se-e
nter
ed d
ata
base
don
pat
ient
inte
rvie
w
ndash P
atie
nt d
emog
raph
ics
ndash P
ain
char
acte
ristic
s
Com
pute
r sc
reen
dis
play
of
anal
gesi
c tr
eatm
ent
reco
mm
enda
tions
bas
ed o
n th
eW
orld
Hea
lth O
rgan
izat
ion
(WH
O)rsquos
ana
lges
ic la
dder
Onc
olog
ynu
rses
PAIN
Rep
or
tIt a
ndPA
IN
Con
sultN
[51
52]
To a
ssis
t cl
inic
ians
in a
sses
sing
chr
onic
pain
and
to
educ
ate
patie
nts
rega
rdin
gpa
in m
onito
ring
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
data
sto
red
in A
cces
s 97
data
base
Pat
ient
sel
f-re
port
ed in
put
1)
Pat
ient
dem
ogra
phic
sndash
McG
ill P
ain
Que
stio
nnai
rendash
Pai
n st
atus
ndash P
atie
nt g
oals
for
and
expe
ctat
ions
abo
ut p
ain
ndash Ty
pe a
nd e
ffect
iven
ess
of p
revi
ous
pain
trea
tmen
ts
Har
d co
pies
of
expe
rt s
yste
mre
port
giv
en t
o pa
tient
and
to
clin
icia
n o
n-sc
reen
vie
win
g of
repo
rt is
als
o po
ssib
le
Onc
olog
ynu
rses
and
othe
rcl
inic
ians
Touc
h-sc
reen
Com
pute
rA
sses
smen
t of
Chr
onic
Low
Bac
k P
ain
[53]
To c
olle
ct p
ain
sym
ptom
sta
tus
and
othe
r he
alth
info
rmat
ion
from
patie
nts
with
low
back
pai
n
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
lim
ited
field
tes
ting
Web
-bas
ed s
yste
m u
sing
Del
l Ins
piro
n 11
00 la
ptop
with
Mic
roso
ft X
P o
pera
ting
syst
em 1
4
prime
tou
ch s
cree
n(M
agic
Tou
ch
Key
tec)
ndash P
atie
nt d
emog
raph
ics
ndash O
swes
try
Low
Bac
k P
ain
Dis
abili
ty I
ndex
(V
ersi
on 2
)ndash
Bec
k D
epre
ssio
n In
vent
ory
ndash M
OS
Sho
rt F
orm
-36
(MO
S S
F-3
6)
Not
des
crib
edP
hysi
cian
s
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
S160
Smith et al
considerably in terms of content format anddelivery (eg electronic paper or both) SeveralCDSSs scored and summated patient responseson standard pain and QOL-related assessmentmeasures Based on the published descriptions atleast five of the CDSSs were designed to gener-ate output in real time at the patientrsquos medicalvisit
In terms of systems architecture all CDSSsreviewed were stand-alone personal computer-based systems None interfaced with existingelectronic medical records systems pharmacyappointment scheduling or laboratory resultsreporting Three either were Web-based or hadthe capacity to use a Web-based platform
Table 2 summarizes the types of studies con-ducted to date to evaluate chronic pain CDSSs Ofthe eight CDSSs identified five had publishedevaluation results With one exception all werefeasibility studies exclusively Studies were con-ducted in both inpatient and outpatient settingsAmong the outpatient studies two had been con-ducted with PCPs the remainder involved special-ists in tertiary care settings Study designs used forevaluating these CDSSs varied two were cross-sectional two involved immediate pre- and post-assessments of CDSS use one was longitudinal(12-month follow-up) and one was a focus groupStudy sample sizes ranged from 213 to 4 with themajority having 50 or fewer subjects
Patient acceptability of the CDSS was the sin-gle most commonly assessed variable Evaluationsof both the SymptomReport and the PAIN
Report-It
employed a common tool to assess patientacceptability Results across four pilot testsinvolving a total of 254 subjects consistentlyshowed high acceptability of these two CDSSsand 100 completion rates in terms of data input[495152] The average amount of time requiredranged from a high of 38 minutes to a low of14 minutes
Two studies addressed the issue of medicalaccuracy of the system-generated recommenda-tion Both studies examined this issue by compar-ing system-generated diagnoses andor treatmentrecommendations with those generated by physi-cian experts based on a select sample of patientcases Results showed moderate to high agreementbetween the system- and expert-generated recom-mendations [4748]
Clinician perceptions concerning ease of useand value of a CDSS for chronic pain managementwere examined in two studies Overall physiciansfound the system to be moderately easy to use and
of some clinical worth [4852] Knab and col-leagues [48] reported that the average cliniciantime spent per case on the PMA to obtain outputwas approximately 5 minutes (
plusmn
34 minutes) [48]In addition 85 of physicians adopted the PMA-generated recommendations and 25 of thestudy patients seen were referred to a pain special-ist with the average time to referral being37 months (
plusmn
06 months) [48]Analyses concerning the impact of CDSSs on
patient outcomes were limited Huang and col-leagues [51] assessed changes in pain intensity pre-and post-CDSS use in a sample of radiationoncology clinic patients Although there was adownward trend in pain levels over time resultswere not statistically significant possibly dueto the small size of the sample (N
=
15) Wilkieand colleagues [4952] reported qualitative dataregarding the impact of the SymptomReport andthe PAIN
ReportIt
on patient behavior Resultswere contradictory Of the 41 outpatients whoused the SymptomReport approximately 68stated that it had not affected their pain-relatedcommunication Some however felt that theytalked more precisely and explicitly about theirpain as a result of using it Other comments asso-ciated with use of the SymptomReport includedan increased awareness of pain symptoms andgreater compliance with pain symptom manage-ment In contrast 86 of users of the PAIN-
ReportIt
cited it as beneficial for patientndashdoctorpain-related communication and that it ldquofreedthem to describe their painrdquo
Discussion
Over the past two decades there has been a gradualbut steady growth in research on the use of CDSSsfor chronic pain management To date the numberof such systems is small but expanding Advanceshave also been evident in terms of both the quantityand quality of evaluation studies conducted Whilethe earliest versions were presented in the litera-ture as prototypes only [4546] CDSSs developedsince 2001 have all undergone some form of fieldtesting The majority of these studies howeverhave been nonexperimental in design and focusedexclusively on process measures such as patientor clinician ratings of system acceptability andusability Other salient process measures such asthe degree to which the clinician andor patientactually reviewed and utilized system output orhad confidence in its accuracy have not been con-sistently assessed Poor usability and practitioner
Computerized Decision-Support System for Pain
S161
Tab
le 2
Eva
luat
ion
of c
ompu
teri
zed
deci
sion
-sup
port
sys
tem
s (C
DS
Ss)
in c
hron
ic p
ain
man
agem
ent
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
Nie
lsen
et a
l
[47
]D
iagn
ostic
Hea
dach
eD
iary
PC
Ps
Not
des
crib
edndash
a
gree
men
t of c
ompu
ter-
gene
rate
d vs
exp
ert
phys
icia
n di
agno
ses
ndash 10
0 a
gree
men
t of c
ompu
ter-
gene
rate
d vs
ex
pert
phy
sici
an d
iagn
oses
Kna
b et
al
[48
]P
ain
Man
agem
ent
Adv
isor
N
=
50
PC
Ps
N
=
50
chro
nic
pain
pat
ient
sLo
ngitu
dina
l w
ith 1
2-m
onth
PC
P fo
llow
-up
ndash E
ase
of u
sendash
Med
ical
app
ropr
iate
ness
of
reco
mm
enda
tions
ndash
phy
sici
ansrsquo
ado
ptin
g P
ain
Man
agem
ent
Adv
isor
tre
atm
ent
reco
mm
enda
tions
ndash
pat
ient
s re
ferr
ed t
o pa
insp
ecia
lty c
linic
Phy
sici
an a
dopt
ion
of s
yste
m-g
ener
ated
re
com
men
datio
ns 8
5 o
f ca
ses
Ave
rage
phy
sici
an t
ime
spen
t pe
r ca
se
49
min
utes
(S
D
plusmn
34)
Eas
e of
use
as
rate
d by
phy
sici
ans
42
(
plusmn
28
cm)
on s
cale
of
1ndash10
25
of
patie
nts
refe
rred
to
pain
spe
cial
ty
clin
icA
vera
ge t
ime
to p
ain
spec
ialty
ref
erra
l 3
7 m
onth
s (S
D
plusmn
06)
ndash 70
o
f no
nref
erre
d pa
tient
s st
ill r
ecei
ving
co
mpu
ter-
reco
mm
ende
d tr
eatm
ent
1 ye
ar
post
Wilk
ie e
t al
[49
]S
ympt
omR
epor
tN
=
41
outp
atie
nts
with
can
cer
Cro
ss-s
ectio
nal
tele
phon
e in
terv
iew
sndash
13-it
em p
atie
nt a
ccep
tabi
lity
scal
e as
sess
ing
ease
of
use
ofS
ympt
omR
epor
tndash
Inpu
t co
mpl
etio
n tim
endash
Qua
litat
ive
asse
ssm
ent
of d
egre
eof
com
mun
icat
ion
with
hea
lth c
are
prov
ider
s re
gard
ing
pain
and
oth
ersy
mpt
oms
ndash M
ean
time
to c
ompl
ete
Sym
ptom
Rep
ort
382
min
utes
(S
D
plusmn
202
)ndash
Mea
n tim
e to
com
plet
e S
ympt
omC
onsu
lt
209
min
utes
(S
D
plusmn
18
6)ndash
71
of
part
icip
ants
rat
ed S
ympt
omR
epor
t as
eas
y e
njoy
able
an
d in
form
ativ
endash
68
rep
orte
d th
at th
e am
ount
and
con
tent
of
the
ir pa
in-r
elat
ed c
omm
unic
atio
n w
ith
thei
r do
ctor
had
not
cha
nged
muc
hndash
Qua
litat
ive
patie
nt c
omm
ents
1)
help
ed
them
tal
k m
ore
expl
icitl
y ab
out
pain
2)
gave
the
m g
reat
er a
war
enes
s of
pai
n sy
mpt
oms
3)
incr
ease
d un
ders
tand
ing
of
and
enha
nced
com
plia
nce
rega
rdin
g sy
mpt
om m
anag
emen
t ndash
Pat
ient
com
men
ts r
e S
ympt
omC
onsu
lt 1
) co
ntai
ned
too
muc
h in
form
atio
n 2
) no
t ta
rget
ed t
o in
divi
dual
nee
ds (
3) p
rovi
ded
no n
ew in
form
atio
n
Wilk
ie e
t al
[52
]PA
IN
Rep
ortIt
N
=
213
of
who
m 1
06w
ere
canc
er in
patie
nts
10 m
etas
tatic
can
cer
outp
atie
nts
and
97
wer
e in
divi
dual
sex
perie
ncin
g ac
ute
orch
roni
c pa
in r
ecru
ited
from
non
heal
th c
are
setti
ngs
Des
crip
tive
cro
ss-
sect
iona
l stu
dy in
3se
tting
s 1
) te
rtia
ryca
re
2) r
adia
tion
onco
logy
clin
ic
3)m
obile
clin
ic
ndash 13
-item
sca
le m
easu
ring
acce
ptab
ility
of
PAIN
Rep
ortIt
(ie
tim
e to
com
plet
e e
ase
of u
se
unde
rsta
ndab
ility
of
dire
ctio
ns
ergo
nom
ic e
lem
ents
of
syst
em
and
com
plet
enes
s of
res
pons
es)
ndash 86
o
f res
pond
ents
rate
d th
e PA
IN
Rep
ortIt
as
ben
efici
al f
or p
ain
com
mun
icat
ion
ndash 10
0 p
atie
nt c
ompl
etio
n ra
te o
f PA
IN
Rep
ortIt
ndash M
ean
time
to p
atie
nt
com
plet
ion
15
8 m
inut
es (
SD
plusmn
67
)ndash
Mea
n pa
tient
acc
epta
bilit
y sc
ore
11
7 (S
D
plusmn
16
) sc
ores
ran
ged
from
6 t
o 13
80
o
f pa
tient
s ra
ted
acce
ptab
ility
as
grea
ter
than
min
imum
crit
erio
n of
10
ndash U
ser
com
men
ts 1
) so
me
mec
hani
cal
diffi
culti
es 2
) re
activ
ity t
o us
e of
sys
tem
(e
g
vom
iting
) 3)
pre
fere
nce
to r
elay
info
rmat
ion
dire
ctly
to
prov
ider
PC
P p
rimar
y ca
re p
hysi
cian
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
Computerized Decision-Support System for Pain
S157
The purpose of this study was to systematicallyreview the research evidence for CDSSs to addressthe following questions (i) To what extent haveCDSSs been utilized in the context of chronic painmanagement (ii) What are the characteristics ofthese systems and (iii) To what degree have theybeen evaluated and in what types of clinicalsettings
Methods
Data Sources
We conducted an automated literature searchusing the Ovid search engine With the assistanceof a research librarian we searched the followingdatabases MEDLINE (1966 to April 2006)CINAHL (1982 to April 2006) PsychINFO (1967to April 2006) HealthSTAR (1981 to April 2006)EMBASE Cochrane Library Computer andInformation Systems Abstracts Electronics andCommunications Abstracts Proust Digital Disser-tations Computer Retrieval of Informationon Scientific Projects (CRISP) LISA ERICComputer and Information Systems Abstractsand Dissertation Abstracts Key search wordsemployed included the following computer-generated decision support systems and expert sys-tems Additional terms included chronic painprimary care tailored reports personalized com-puter-based information disease management forchronic pain patient goals pain diagnosis andmanagement decision support systems neuralnetworks and fuzzy logic We also conducted amanual search to supplement the automatedsearch The manual search was not limited in timeperiod and included articles that had been refer-enced in other articles
Study Selection Criteria
Eligibility for inclusion in the final set includedany studies describing the development andorapplication of a CDSS or expert system in thecontext of chronic pain management We defineda CDSS as any electronic system designed to assistin clinical decision making regarding chronic painmanagement and in which patient-specific assess-ments and recommendations were generated foruse by a clinician andor patient [44] Consistentwith this definition we excluded any CDSSs thatexamined pain as only one component of an over-all assessment of QOL as well as any thataddressed acute pain management only We alsoexcluded studies that were written in languagesother than English
Results
The cross-database search yielded 70 publishedarticles and three federally funded research studiesdescribing ongoing investigations No additionalstudies were identified via the manual search pro-cess Full-text articles were retrieved for all titlesconsidered to be potentially relevant by theauthors Nine studies describing eight discreteCDSSs were identified as meeting our inclusioncriteria Lack of homogeneity among the final setof studies precluded a quantitative meta-analysisof the data Due to these analytic constraints weconducted a descriptive literature review only
As shown in Table 1 all eight CDSSs weredesigned to assist in the diagnosis andor manage-ment of chronic pain Two of these systems thePain Management Advisor (PMA) and the Diag-nostic Headache Diary were also designed to offereducation to the health care provider One thePMA had an interactive capability that permittedusers to query the system for explanations thera-peutic rationale and therapy guidelines Twoother systems the SymptomReport and thePAIN
ReportIt
featured adjunctive software pro-grams (SymptomConsult PAIN
ConsultN
) thatwere expressly designed as interactive educationaltools for the patient
All but two of the CDSSs were designed totarget a specific type of chronic pain or pain-related condition These included headache (2)low back (1) and cancer-related (3) Input specifi-cations also varied widely both in terms of thetype and amount of data required and in terms ofthe party responsible for data entry (ie physicianother health care provider or patient) All of theCDSSs required detailed input regarding painsymptomology Half of the systems reviewed alsoelicited data on pain medications currently usedand three requested QOL information Otherdata such as psychological status history of priortherapies (both pharmaceutical and nondrug-related therapies) patient goals for pain care andbarriers to pain management were specified asrequired inputs in only two of the CDSSsreviewed Three of the CDSSs used standard psy-chometrically validated instruments (eg McGillPain Questionnaire Medical Outcomes StudyShort Form 36 Oswestry Low Back Pain Disabil-ity Index) to collect input data
The majority of the CDSSs (5 of 8) producedoutput that was intended for clinician use(eg physician or nurse) only Three targetedboth the clinician and the patient Output varied
S158
Smith et al
Tab
le 1
Key
fea
ture
s of
clin
ical
dec
isio
n-su
ppor
t sy
stem
s (C
DS
Ss)
dev
elop
ed f
or c
hron
ic p
ain
man
agem
ent
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
RH
INO
S [
45]
To a
ssis
t ph
ysic
ians
in t
he d
iagn
osis
of
chro
nic
head
ache
sor
fac
ial p
ain
Bas
ed o
n ex
pert
kno
wle
dge
of h
eada
che
and
faci
alpa
in s
peci
alis
ts U
ses
4 se
ts o
f co
nditi
onal
pr
obab
ility
-bas
ed
rule
s 1
) ex
clus
ive
rule
s (i
e
if pa
tient
has
dis
ease
D
s
he m
ust
have
sym
ptom
s S
1
S
2
S
n
) 2)
incl
usiv
e ru
les
3)
asso
ciat
e ru
les
and
4)
dise
ase
imag
e ru
les
Pro
toty
pe o
nly
no t
estin
gS
yste
m d
evel
oped
usi
ngth
e P
rolo
g-K
AB
Apr
ogra
mm
ing
lang
uage
R
uns
on p
erso
nal
com
pute
rs w
ith C
PU
808
6m
emor
y 38
2 K
byte
s
Phy
sici
an in
put
base
d on
patie
nt in
terv
iew
ndash
Pat
ient
dem
ogra
phic
sndash
Ons
et o
f he
adac
hendash
His
tory
sin
ce o
nset
ndash P
ain
char
acte
ristic
sndash
Neu
rolo
gica
l sig
nsas
soci
ated
with
pai
nndash
Sle
ep s
tatu
sndash
Per
sona
l and
fam
ilial
hist
ory
of h
eada
che
ndash Jo
lt he
adac
he (
Yes
No)
ndash S
cler
osis
of
retin
al a
rter
y
Initi
al o
utpu
t af
ter
first
set
of
scre
enin
g qu
estio
ns
ndash Li
st o
f di
seas
es n
ot r
ejec
ted
Out
put
afte
r ad
ditio
nal
ques
tion(
s)
ndash Li
st o
f po
ssib
le d
isea
ses
ndash D
iagn
ostic
con
clus
ions
ndash E
xpla
natio
n of
the
dis
ease
requ
ired
exam
inat
ions
and
sugg
este
d th
erap
yH
ospi
tal c
hart
can
be
prin
ted
from
the
inpu
t da
ta in
the
form
of
a n
atur
al la
ngua
ge
repr
esen
tatio
n
Phy
sici
ans
IVA
N [
46]
To p
rovi
dere
com
men
datio
nsfo
r co
ntro
lling
pai
nan
d pr
ovid
ing
sym
ptom
rel
ief
inca
ncer
os
teo-
and
rheu
mat
oid
arth
ritis
Cas
e-ba
sed
reas
onin
gst
rate
gy t
o re
cord
and
retr
ieve
info
rmat
ion
stor
edin
an
inte
rnal
kno
wle
dge
base
Pro
toty
pe o
nly
no t
estin
gIV
AN
sof
twar
e w
ritte
n in
LPA
Pro
log
prog
ram
min
gla
ngua
ge fo
r W
indo
ws
(ldquoW
inP
rolo
grdquo)
runs
on
PC
Win
dow
s 95
NT
Win
Pro
log
ndash P
ain
sym
ptom
che
cklis
tndash
Cur
rent
pai
n di
agno
sis
ndash C
urre
nt p
ain
trea
tmen
tre
gim
en
Com
pute
r sc
reen
dis
play
ndash
Dia
gnos
tic c
onfir
mat
ion
ndash D
escr
iptio
n of
sym
ptom
s th
atm
ay a
ppea
r la
ter
ndash Tr
eatm
ents
pro
ven
succ
essf
ulin
sim
ilar
or r
elat
ed c
ases
ndash P
ossi
ble
alte
rnat
ive
caus
esof
the
pai
n
Phy
sici
ans
and
patie
nts
The
Dia
gnos
ticH
eada
che
Dia
ry[4
7]
To e
duca
te a
ndpr
ovid
e di
agno
stic
supp
ort
to p
rimar
yca
re p
rovi
ders
inor
der
to im
prov
em
anag
emen
t of
head
ache
s
Rul
e-ba
sed
expe
rt s
yste
mus
ing
Boo
lean
logi
c A
set
of d
iagn
ostic
rul
es u
sed
tode
term
ine
a di
agno
sis
base
dup
on t
he d
ata
ente
red
inpa
tient
dia
ry D
iary
dat
a ar
etr
ansf
orm
ed in
to a
dia
gnos
isfo
llow
ing
the
Inte
rnat
iona
lH
eada
che
Soc
iety
rsquoscl
assi
ficat
ion
Pro
toty
pede
velo
ped
syst
em-g
ener
ated
diag
nose
s w
ere
valid
ated
aga
inst
phys
icia
n ex
pert
-ge
nera
ted
diag
nose
s
Sta
nd-a
lone
Win
dow
s 95
prog
ram
writ
ten
in D
elph
ipr
ogra
mm
ing
lang
uage
ndash P
atie
nt d
ata
ndash H
eada
che
diar
y en
trie
sndash
Med
icat
ions
use
d to
alle
viat
e he
adac
he
ndash D
iagn
osis
of
head
ache
typ
eP
CP
s
Pai
n M
anag
emen
tA
dvis
or (
PM
A)
(Nov
aInt
ellig
ence
Inc
S
an D
iego
C
A)
[48]
To e
nhan
ce p
rimar
yca
re p
rovi
ders
rsquo(P
CP
s) m
anag
emen
tof
chr
onic
pai
n
ndash R
elie
s on
rul
e-ba
sed
algo
rithm
s de
rived
fro
mex
pert
kno
wle
dge
ofpa
in s
peci
alis
tsndash
Use
r as
ked
a se
ries
ofqu
estio
ns t
o re
fine
the
diag
nosi
s an
d de
term
ine
appr
opria
te t
hera
pyndash
Inte
ract
ive
capa
bilit
y(e
g
for
expl
anat
ions
th
erap
eutic
rat
iona
les
ther
apy
guid
elin
es)
Wor
king
ver
sion
deve
lope
d s
ome
field
tes
ting
cond
ucte
d
ndash P
entiu
m-b
ased
PC
sndash
Win
dow
s 95
ndash P
MA
writ
ten
inM
icro
Sof
t Vis
ual B
asic
v
50
ru
n as
an
expe
rtap
plic
atio
n in
Xpe
rtR
ule
ndash A
lgor
ithm
s st
ored
inM
icro
Sof
t A
cces
s da
taba
sendash
Mic
roS
oft
Hel
p U
tility
used
for
expl
anat
ions
and
quer
ies
ndash P
atie
nt d
emog
raph
ics
ndash D
iagn
osis
ndash P
ain
char
acte
ristic
sndash
Labo
rato
ry t
ests
ampim
agin
g st
udie
sndash
Cur
rent
med
icat
ions
ndash P
rior
ther
apie
sndash
Con
curr
ent
dise
ase
cond
ition
sndash
Alle
rgie
sndash
Psy
chol
ogic
al s
tatu
s
ndash A
prio
ritiz
ed li
st o
fre
com
men
datio
ns 1
) m
edic
alm
anag
emen
t (p
harm
acol
ogic
and
nonp
harm
acol
ogic
alm
anag
emen
t ph
ysic
al
psyc
hoso
cial
mod
aliti
es)
2)in
vasi
ve p
roce
dure
s 3
) re
ferr
als
PC
Ps
Computerized Decision-Support System for Pain
S159
Sym
ptom
Rep
ort
and
Sym
ptom
Con
sult
[49]
To a
ssis
t cl
inic
ians
in a
sses
sing
can
cer-
rela
ted
chro
nic
pain
and
fatig
ue
and
clar
ify p
atie
ntsrsquo
mis
belie
fs a
bout
pain
ass
essm
ent
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
Pat
ient
sel
f-re
port
ed in
put
ndash P
atie
nt d
emog
raph
ics
ndash 19
70 v
ersi
on o
f M
cGill
Pai
n Q
uest
ionn
aire
(M
PQ
)ndash
Bar
riers
Que
stio
nnai
re(B
Q)
ndash S
chw
artz
can
cer
fatig
uesc
ale
(SC
FS
-6)
1) H
ard
copi
es o
f ex
pert
sys
tem
repo
rt g
iven
to
patie
nt a
nd t
ocl
inic
ian
2) P
atie
nt r
ecei
ves
educ
atio
nal
mat
eria
ls o
n ho
w t
o re
port
pa
in
use
pain
med
icat
ions
sa
fely
an
d m
anag
e fa
tigue
M
ater
ials
are
cus
tom
ized
to
the
patie
ntrsquos
nee
ds a
nd
pres
ente
d in
an
inte
ract
ive
m
ultim
edia
form
at P
atie
nts
have
opt
ion
to re
ad o
r lis
ten
to
info
rmat
ion
on t
he c
ompu
ter
prin
t an
y or
all
of t
he
mat
eria
ls
or d
o bo
th
Onc
olog
ynu
rses
ot
her
clin
icia
ns
Dec
isio
n-su
ppor
tco
mpu
ter
prog
ram
for
canc
er p
ain
man
agem
ent
[50]
To im
prov
e th
eon
colo
gy n
urse
srsquode
cisi
on m
akin
gre
late
d to
can
cer
pain
man
agem
ent
amon
g cu
ltura
llydi
vers
e fe
mal
eon
colo
gy p
atie
nts
ndash S
urve
y da
ta o
n m
ultic
ultu
ral
canc
er p
ain
char
acte
ristic
sw
ere
anal
yzed
usi
ng f
uzzy
infe
renc
e lo
gic
to d
evel
op 4
mod
ules
1)
a ge
neric
know
ledg
e ba
se 2
) a
cultu
re-s
peci
fic k
now
ledg
e ba
se 3
) de
cisi
on-m
akin
g
and
4) s
elf-
adap
tatio
nndash
Dec
isio
n-m
akin
g m
odul
eco
nsis
ts o
f 2
sets
of
fuzz
yin
fere
nce
logi
c de
velo
ped
via
a ge
netic
alg
orith
m
Har
dwar
e no
t de
scrib
edndash
Ada
ptiv
e fu
zzy
logi
cso
ftwar
e us
ed t
o de
velo
pan
d ru
n th
e kn
owle
dge
base
gen
erat
ion
and
the
deci
sion
-mak
ing
and
self-
adap
tatio
n m
odul
es
Nur
se-e
nter
ed d
ata
base
don
pat
ient
inte
rvie
w
ndash P
atie
nt d
emog
raph
ics
ndash P
ain
char
acte
ristic
s
Com
pute
r sc
reen
dis
play
of
anal
gesi
c tr
eatm
ent
reco
mm
enda
tions
bas
ed o
n th
eW
orld
Hea
lth O
rgan
izat
ion
(WH
O)rsquos
ana
lges
ic la
dder
Onc
olog
ynu
rses
PAIN
Rep
or
tIt a
ndPA
IN
Con
sultN
[51
52]
To a
ssis
t cl
inic
ians
in a
sses
sing
chr
onic
pain
and
to
educ
ate
patie
nts
rega
rdin
gpa
in m
onito
ring
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
data
sto
red
in A
cces
s 97
data
base
Pat
ient
sel
f-re
port
ed in
put
1)
Pat
ient
dem
ogra
phic
sndash
McG
ill P
ain
Que
stio
nnai
rendash
Pai
n st
atus
ndash P
atie
nt g
oals
for
and
expe
ctat
ions
abo
ut p
ain
ndash Ty
pe a
nd e
ffect
iven
ess
of p
revi
ous
pain
trea
tmen
ts
Har
d co
pies
of
expe
rt s
yste
mre
port
giv
en t
o pa
tient
and
to
clin
icia
n o
n-sc
reen
vie
win
g of
repo
rt is
als
o po
ssib
le
Onc
olog
ynu
rses
and
othe
rcl
inic
ians
Touc
h-sc
reen
Com
pute
rA
sses
smen
t of
Chr
onic
Low
Bac
k P
ain
[53]
To c
olle
ct p
ain
sym
ptom
sta
tus
and
othe
r he
alth
info
rmat
ion
from
patie
nts
with
low
back
pai
n
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
lim
ited
field
tes
ting
Web
-bas
ed s
yste
m u
sing
Del
l Ins
piro
n 11
00 la
ptop
with
Mic
roso
ft X
P o
pera
ting
syst
em 1
4
prime
tou
ch s
cree
n(M
agic
Tou
ch
Key
tec)
ndash P
atie
nt d
emog
raph
ics
ndash O
swes
try
Low
Bac
k P
ain
Dis
abili
ty I
ndex
(V
ersi
on 2
)ndash
Bec
k D
epre
ssio
n In
vent
ory
ndash M
OS
Sho
rt F
orm
-36
(MO
S S
F-3
6)
Not
des
crib
edP
hysi
cian
s
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
S160
Smith et al
considerably in terms of content format anddelivery (eg electronic paper or both) SeveralCDSSs scored and summated patient responseson standard pain and QOL-related assessmentmeasures Based on the published descriptions atleast five of the CDSSs were designed to gener-ate output in real time at the patientrsquos medicalvisit
In terms of systems architecture all CDSSsreviewed were stand-alone personal computer-based systems None interfaced with existingelectronic medical records systems pharmacyappointment scheduling or laboratory resultsreporting Three either were Web-based or hadthe capacity to use a Web-based platform
Table 2 summarizes the types of studies con-ducted to date to evaluate chronic pain CDSSs Ofthe eight CDSSs identified five had publishedevaluation results With one exception all werefeasibility studies exclusively Studies were con-ducted in both inpatient and outpatient settingsAmong the outpatient studies two had been con-ducted with PCPs the remainder involved special-ists in tertiary care settings Study designs used forevaluating these CDSSs varied two were cross-sectional two involved immediate pre- and post-assessments of CDSS use one was longitudinal(12-month follow-up) and one was a focus groupStudy sample sizes ranged from 213 to 4 with themajority having 50 or fewer subjects
Patient acceptability of the CDSS was the sin-gle most commonly assessed variable Evaluationsof both the SymptomReport and the PAIN
Report-It
employed a common tool to assess patientacceptability Results across four pilot testsinvolving a total of 254 subjects consistentlyshowed high acceptability of these two CDSSsand 100 completion rates in terms of data input[495152] The average amount of time requiredranged from a high of 38 minutes to a low of14 minutes
Two studies addressed the issue of medicalaccuracy of the system-generated recommenda-tion Both studies examined this issue by compar-ing system-generated diagnoses andor treatmentrecommendations with those generated by physi-cian experts based on a select sample of patientcases Results showed moderate to high agreementbetween the system- and expert-generated recom-mendations [4748]
Clinician perceptions concerning ease of useand value of a CDSS for chronic pain managementwere examined in two studies Overall physiciansfound the system to be moderately easy to use and
of some clinical worth [4852] Knab and col-leagues [48] reported that the average cliniciantime spent per case on the PMA to obtain outputwas approximately 5 minutes (
plusmn
34 minutes) [48]In addition 85 of physicians adopted the PMA-generated recommendations and 25 of thestudy patients seen were referred to a pain special-ist with the average time to referral being37 months (
plusmn
06 months) [48]Analyses concerning the impact of CDSSs on
patient outcomes were limited Huang and col-leagues [51] assessed changes in pain intensity pre-and post-CDSS use in a sample of radiationoncology clinic patients Although there was adownward trend in pain levels over time resultswere not statistically significant possibly dueto the small size of the sample (N
=
15) Wilkieand colleagues [4952] reported qualitative dataregarding the impact of the SymptomReport andthe PAIN
ReportIt
on patient behavior Resultswere contradictory Of the 41 outpatients whoused the SymptomReport approximately 68stated that it had not affected their pain-relatedcommunication Some however felt that theytalked more precisely and explicitly about theirpain as a result of using it Other comments asso-ciated with use of the SymptomReport includedan increased awareness of pain symptoms andgreater compliance with pain symptom manage-ment In contrast 86 of users of the PAIN-
ReportIt
cited it as beneficial for patientndashdoctorpain-related communication and that it ldquofreedthem to describe their painrdquo
Discussion
Over the past two decades there has been a gradualbut steady growth in research on the use of CDSSsfor chronic pain management To date the numberof such systems is small but expanding Advanceshave also been evident in terms of both the quantityand quality of evaluation studies conducted Whilethe earliest versions were presented in the litera-ture as prototypes only [4546] CDSSs developedsince 2001 have all undergone some form of fieldtesting The majority of these studies howeverhave been nonexperimental in design and focusedexclusively on process measures such as patientor clinician ratings of system acceptability andusability Other salient process measures such asthe degree to which the clinician andor patientactually reviewed and utilized system output orhad confidence in its accuracy have not been con-sistently assessed Poor usability and practitioner
Computerized Decision-Support System for Pain
S161
Tab
le 2
Eva
luat
ion
of c
ompu
teri
zed
deci
sion
-sup
port
sys
tem
s (C
DS
Ss)
in c
hron
ic p
ain
man
agem
ent
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
Nie
lsen
et a
l
[47
]D
iagn
ostic
Hea
dach
eD
iary
PC
Ps
Not
des
crib
edndash
a
gree
men
t of c
ompu
ter-
gene
rate
d vs
exp
ert
phys
icia
n di
agno
ses
ndash 10
0 a
gree
men
t of c
ompu
ter-
gene
rate
d vs
ex
pert
phy
sici
an d
iagn
oses
Kna
b et
al
[48
]P
ain
Man
agem
ent
Adv
isor
N
=
50
PC
Ps
N
=
50
chro
nic
pain
pat
ient
sLo
ngitu
dina
l w
ith 1
2-m
onth
PC
P fo
llow
-up
ndash E
ase
of u
sendash
Med
ical
app
ropr
iate
ness
of
reco
mm
enda
tions
ndash
phy
sici
ansrsquo
ado
ptin
g P
ain
Man
agem
ent
Adv
isor
tre
atm
ent
reco
mm
enda
tions
ndash
pat
ient
s re
ferr
ed t
o pa
insp
ecia
lty c
linic
Phy
sici
an a
dopt
ion
of s
yste
m-g
ener
ated
re
com
men
datio
ns 8
5 o
f ca
ses
Ave
rage
phy
sici
an t
ime
spen
t pe
r ca
se
49
min
utes
(S
D
plusmn
34)
Eas
e of
use
as
rate
d by
phy
sici
ans
42
(
plusmn
28
cm)
on s
cale
of
1ndash10
25
of
patie
nts
refe
rred
to
pain
spe
cial
ty
clin
icA
vera
ge t
ime
to p
ain
spec
ialty
ref
erra
l 3
7 m
onth
s (S
D
plusmn
06)
ndash 70
o
f no
nref
erre
d pa
tient
s st
ill r
ecei
ving
co
mpu
ter-
reco
mm
ende
d tr
eatm
ent
1 ye
ar
post
Wilk
ie e
t al
[49
]S
ympt
omR
epor
tN
=
41
outp
atie
nts
with
can
cer
Cro
ss-s
ectio
nal
tele
phon
e in
terv
iew
sndash
13-it
em p
atie
nt a
ccep
tabi
lity
scal
e as
sess
ing
ease
of
use
ofS
ympt
omR
epor
tndash
Inpu
t co
mpl
etio
n tim
endash
Qua
litat
ive
asse
ssm
ent
of d
egre
eof
com
mun
icat
ion
with
hea
lth c
are
prov
ider
s re
gard
ing
pain
and
oth
ersy
mpt
oms
ndash M
ean
time
to c
ompl
ete
Sym
ptom
Rep
ort
382
min
utes
(S
D
plusmn
202
)ndash
Mea
n tim
e to
com
plet
e S
ympt
omC
onsu
lt
209
min
utes
(S
D
plusmn
18
6)ndash
71
of
part
icip
ants
rat
ed S
ympt
omR
epor
t as
eas
y e
njoy
able
an
d in
form
ativ
endash
68
rep
orte
d th
at th
e am
ount
and
con
tent
of
the
ir pa
in-r
elat
ed c
omm
unic
atio
n w
ith
thei
r do
ctor
had
not
cha
nged
muc
hndash
Qua
litat
ive
patie
nt c
omm
ents
1)
help
ed
them
tal
k m
ore
expl
icitl
y ab
out
pain
2)
gave
the
m g
reat
er a
war
enes
s of
pai
n sy
mpt
oms
3)
incr
ease
d un
ders
tand
ing
of
and
enha
nced
com
plia
nce
rega
rdin
g sy
mpt
om m
anag
emen
t ndash
Pat
ient
com
men
ts r
e S
ympt
omC
onsu
lt 1
) co
ntai
ned
too
muc
h in
form
atio
n 2
) no
t ta
rget
ed t
o in
divi
dual
nee
ds (
3) p
rovi
ded
no n
ew in
form
atio
n
Wilk
ie e
t al
[52
]PA
IN
Rep
ortIt
N
=
213
of
who
m 1
06w
ere
canc
er in
patie
nts
10 m
etas
tatic
can
cer
outp
atie
nts
and
97
wer
e in
divi
dual
sex
perie
ncin
g ac
ute
orch
roni
c pa
in r
ecru
ited
from
non
heal
th c
are
setti
ngs
Des
crip
tive
cro
ss-
sect
iona
l stu
dy in
3se
tting
s 1
) te
rtia
ryca
re
2) r
adia
tion
onco
logy
clin
ic
3)m
obile
clin
ic
ndash 13
-item
sca
le m
easu
ring
acce
ptab
ility
of
PAIN
Rep
ortIt
(ie
tim
e to
com
plet
e e
ase
of u
se
unde
rsta
ndab
ility
of
dire
ctio
ns
ergo
nom
ic e
lem
ents
of
syst
em
and
com
plet
enes
s of
res
pons
es)
ndash 86
o
f res
pond
ents
rate
d th
e PA
IN
Rep
ortIt
as
ben
efici
al f
or p
ain
com
mun
icat
ion
ndash 10
0 p
atie
nt c
ompl
etio
n ra
te o
f PA
IN
Rep
ortIt
ndash M
ean
time
to p
atie
nt
com
plet
ion
15
8 m
inut
es (
SD
plusmn
67
)ndash
Mea
n pa
tient
acc
epta
bilit
y sc
ore
11
7 (S
D
plusmn
16
) sc
ores
ran
ged
from
6 t
o 13
80
o
f pa
tient
s ra
ted
acce
ptab
ility
as
grea
ter
than
min
imum
crit
erio
n of
10
ndash U
ser
com
men
ts 1
) so
me
mec
hani
cal
diffi
culti
es 2
) re
activ
ity t
o us
e of
sys
tem
(e
g
vom
iting
) 3)
pre
fere
nce
to r
elay
info
rmat
ion
dire
ctly
to
prov
ider
PC
P p
rimar
y ca
re p
hysi
cian
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
S158
Smith et al
Tab
le 1
Key
fea
ture
s of
clin
ical
dec
isio
n-su
ppor
t sy
stem
s (C
DS
Ss)
dev
elop
ed f
or c
hron
ic p
ain
man
agem
ent
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
RH
INO
S [
45]
To a
ssis
t ph
ysic
ians
in t
he d
iagn
osis
of
chro
nic
head
ache
sor
fac
ial p
ain
Bas
ed o
n ex
pert
kno
wle
dge
of h
eada
che
and
faci
alpa
in s
peci
alis
ts U
ses
4 se
ts o
f co
nditi
onal
pr
obab
ility
-bas
ed
rule
s 1
) ex
clus
ive
rule
s (i
e
if pa
tient
has
dis
ease
D
s
he m
ust
have
sym
ptom
s S
1
S
2
S
n
) 2)
incl
usiv
e ru
les
3)
asso
ciat
e ru
les
and
4)
dise
ase
imag
e ru
les
Pro
toty
pe o
nly
no t
estin
gS
yste
m d
evel
oped
usi
ngth
e P
rolo
g-K
AB
Apr
ogra
mm
ing
lang
uage
R
uns
on p
erso
nal
com
pute
rs w
ith C
PU
808
6m
emor
y 38
2 K
byte
s
Phy
sici
an in
put
base
d on
patie
nt in
terv
iew
ndash
Pat
ient
dem
ogra
phic
sndash
Ons
et o
f he
adac
hendash
His
tory
sin
ce o
nset
ndash P
ain
char
acte
ristic
sndash
Neu
rolo
gica
l sig
nsas
soci
ated
with
pai
nndash
Sle
ep s
tatu
sndash
Per
sona
l and
fam
ilial
hist
ory
of h
eada
che
ndash Jo
lt he
adac
he (
Yes
No)
ndash S
cler
osis
of
retin
al a
rter
y
Initi
al o
utpu
t af
ter
first
set
of
scre
enin
g qu
estio
ns
ndash Li
st o
f di
seas
es n
ot r
ejec
ted
Out
put
afte
r ad
ditio
nal
ques
tion(
s)
ndash Li
st o
f po
ssib
le d
isea
ses
ndash D
iagn
ostic
con
clus
ions
ndash E
xpla
natio
n of
the
dis
ease
requ
ired
exam
inat
ions
and
sugg
este
d th
erap
yH
ospi
tal c
hart
can
be
prin
ted
from
the
inpu
t da
ta in
the
form
of
a n
atur
al la
ngua
ge
repr
esen
tatio
n
Phy
sici
ans
IVA
N [
46]
To p
rovi
dere
com
men
datio
nsfo
r co
ntro
lling
pai
nan
d pr
ovid
ing
sym
ptom
rel
ief
inca
ncer
os
teo-
and
rheu
mat
oid
arth
ritis
Cas
e-ba
sed
reas
onin
gst
rate
gy t
o re
cord
and
retr
ieve
info
rmat
ion
stor
edin
an
inte
rnal
kno
wle
dge
base
Pro
toty
pe o
nly
no t
estin
gIV
AN
sof
twar
e w
ritte
n in
LPA
Pro
log
prog
ram
min
gla
ngua
ge fo
r W
indo
ws
(ldquoW
inP
rolo
grdquo)
runs
on
PC
Win
dow
s 95
NT
Win
Pro
log
ndash P
ain
sym
ptom
che
cklis
tndash
Cur
rent
pai
n di
agno
sis
ndash C
urre
nt p
ain
trea
tmen
tre
gim
en
Com
pute
r sc
reen
dis
play
ndash
Dia
gnos
tic c
onfir
mat
ion
ndash D
escr
iptio
n of
sym
ptom
s th
atm
ay a
ppea
r la
ter
ndash Tr
eatm
ents
pro
ven
succ
essf
ulin
sim
ilar
or r
elat
ed c
ases
ndash P
ossi
ble
alte
rnat
ive
caus
esof
the
pai
n
Phy
sici
ans
and
patie
nts
The
Dia
gnos
ticH
eada
che
Dia
ry[4
7]
To e
duca
te a
ndpr
ovid
e di
agno
stic
supp
ort
to p
rimar
yca
re p
rovi
ders
inor
der
to im
prov
em
anag
emen
t of
head
ache
s
Rul
e-ba
sed
expe
rt s
yste
mus
ing
Boo
lean
logi
c A
set
of d
iagn
ostic
rul
es u
sed
tode
term
ine
a di
agno
sis
base
dup
on t
he d
ata
ente
red
inpa
tient
dia
ry D
iary
dat
a ar
etr
ansf
orm
ed in
to a
dia
gnos
isfo
llow
ing
the
Inte
rnat
iona
lH
eada
che
Soc
iety
rsquoscl
assi
ficat
ion
Pro
toty
pede
velo
ped
syst
em-g
ener
ated
diag
nose
s w
ere
valid
ated
aga
inst
phys
icia
n ex
pert
-ge
nera
ted
diag
nose
s
Sta
nd-a
lone
Win
dow
s 95
prog
ram
writ
ten
in D
elph
ipr
ogra
mm
ing
lang
uage
ndash P
atie
nt d
ata
ndash H
eada
che
diar
y en
trie
sndash
Med
icat
ions
use
d to
alle
viat
e he
adac
he
ndash D
iagn
osis
of
head
ache
typ
eP
CP
s
Pai
n M
anag
emen
tA
dvis
or (
PM
A)
(Nov
aInt
ellig
ence
Inc
S
an D
iego
C
A)
[48]
To e
nhan
ce p
rimar
yca
re p
rovi
ders
rsquo(P
CP
s) m
anag
emen
tof
chr
onic
pai
n
ndash R
elie
s on
rul
e-ba
sed
algo
rithm
s de
rived
fro
mex
pert
kno
wle
dge
ofpa
in s
peci
alis
tsndash
Use
r as
ked
a se
ries
ofqu
estio
ns t
o re
fine
the
diag
nosi
s an
d de
term
ine
appr
opria
te t
hera
pyndash
Inte
ract
ive
capa
bilit
y(e
g
for
expl
anat
ions
th
erap
eutic
rat
iona
les
ther
apy
guid
elin
es)
Wor
king
ver
sion
deve
lope
d s
ome
field
tes
ting
cond
ucte
d
ndash P
entiu
m-b
ased
PC
sndash
Win
dow
s 95
ndash P
MA
writ
ten
inM
icro
Sof
t Vis
ual B
asic
v
50
ru
n as
an
expe
rtap
plic
atio
n in
Xpe
rtR
ule
ndash A
lgor
ithm
s st
ored
inM
icro
Sof
t A
cces
s da
taba
sendash
Mic
roS
oft
Hel
p U
tility
used
for
expl
anat
ions
and
quer
ies
ndash P
atie
nt d
emog
raph
ics
ndash D
iagn
osis
ndash P
ain
char
acte
ristic
sndash
Labo
rato
ry t
ests
ampim
agin
g st
udie
sndash
Cur
rent
med
icat
ions
ndash P
rior
ther
apie
sndash
Con
curr
ent
dise
ase
cond
ition
sndash
Alle
rgie
sndash
Psy
chol
ogic
al s
tatu
s
ndash A
prio
ritiz
ed li
st o
fre
com
men
datio
ns 1
) m
edic
alm
anag
emen
t (p
harm
acol
ogic
and
nonp
harm
acol
ogic
alm
anag
emen
t ph
ysic
al
psyc
hoso
cial
mod
aliti
es)
2)in
vasi
ve p
roce
dure
s 3
) re
ferr
als
PC
Ps
Computerized Decision-Support System for Pain
S159
Sym
ptom
Rep
ort
and
Sym
ptom
Con
sult
[49]
To a
ssis
t cl
inic
ians
in a
sses
sing
can
cer-
rela
ted
chro
nic
pain
and
fatig
ue
and
clar
ify p
atie
ntsrsquo
mis
belie
fs a
bout
pain
ass
essm
ent
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
Pat
ient
sel
f-re
port
ed in
put
ndash P
atie
nt d
emog
raph
ics
ndash 19
70 v
ersi
on o
f M
cGill
Pai
n Q
uest
ionn
aire
(M
PQ
)ndash
Bar
riers
Que
stio
nnai
re(B
Q)
ndash S
chw
artz
can
cer
fatig
uesc
ale
(SC
FS
-6)
1) H
ard
copi
es o
f ex
pert
sys
tem
repo
rt g
iven
to
patie
nt a
nd t
ocl
inic
ian
2) P
atie
nt r
ecei
ves
educ
atio
nal
mat
eria
ls o
n ho
w t
o re
port
pa
in
use
pain
med
icat
ions
sa
fely
an
d m
anag
e fa
tigue
M
ater
ials
are
cus
tom
ized
to
the
patie
ntrsquos
nee
ds a
nd
pres
ente
d in
an
inte
ract
ive
m
ultim
edia
form
at P
atie
nts
have
opt
ion
to re
ad o
r lis
ten
to
info
rmat
ion
on t
he c
ompu
ter
prin
t an
y or
all
of t
he
mat
eria
ls
or d
o bo
th
Onc
olog
ynu
rses
ot
her
clin
icia
ns
Dec
isio
n-su
ppor
tco
mpu
ter
prog
ram
for
canc
er p
ain
man
agem
ent
[50]
To im
prov
e th
eon
colo
gy n
urse
srsquode
cisi
on m
akin
gre
late
d to
can
cer
pain
man
agem
ent
amon
g cu
ltura
llydi
vers
e fe
mal
eon
colo
gy p
atie
nts
ndash S
urve
y da
ta o
n m
ultic
ultu
ral
canc
er p
ain
char
acte
ristic
sw
ere
anal
yzed
usi
ng f
uzzy
infe
renc
e lo
gic
to d
evel
op 4
mod
ules
1)
a ge
neric
know
ledg
e ba
se 2
) a
cultu
re-s
peci
fic k
now
ledg
e ba
se 3
) de
cisi
on-m
akin
g
and
4) s
elf-
adap
tatio
nndash
Dec
isio
n-m
akin
g m
odul
eco
nsis
ts o
f 2
sets
of
fuzz
yin
fere
nce
logi
c de
velo
ped
via
a ge
netic
alg
orith
m
Har
dwar
e no
t de
scrib
edndash
Ada
ptiv
e fu
zzy
logi
cso
ftwar
e us
ed t
o de
velo
pan
d ru
n th
e kn
owle
dge
base
gen
erat
ion
and
the
deci
sion
-mak
ing
and
self-
adap
tatio
n m
odul
es
Nur
se-e
nter
ed d
ata
base
don
pat
ient
inte
rvie
w
ndash P
atie
nt d
emog
raph
ics
ndash P
ain
char
acte
ristic
s
Com
pute
r sc
reen
dis
play
of
anal
gesi
c tr
eatm
ent
reco
mm
enda
tions
bas
ed o
n th
eW
orld
Hea
lth O
rgan
izat
ion
(WH
O)rsquos
ana
lges
ic la
dder
Onc
olog
ynu
rses
PAIN
Rep
or
tIt a
ndPA
IN
Con
sultN
[51
52]
To a
ssis
t cl
inic
ians
in a
sses
sing
chr
onic
pain
and
to
educ
ate
patie
nts
rega
rdin
gpa
in m
onito
ring
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
data
sto
red
in A
cces
s 97
data
base
Pat
ient
sel
f-re
port
ed in
put
1)
Pat
ient
dem
ogra
phic
sndash
McG
ill P
ain
Que
stio
nnai
rendash
Pai
n st
atus
ndash P
atie
nt g
oals
for
and
expe
ctat
ions
abo
ut p
ain
ndash Ty
pe a
nd e
ffect
iven
ess
of p
revi
ous
pain
trea
tmen
ts
Har
d co
pies
of
expe
rt s
yste
mre
port
giv
en t
o pa
tient
and
to
clin
icia
n o
n-sc
reen
vie
win
g of
repo
rt is
als
o po
ssib
le
Onc
olog
ynu
rses
and
othe
rcl
inic
ians
Touc
h-sc
reen
Com
pute
rA
sses
smen
t of
Chr
onic
Low
Bac
k P
ain
[53]
To c
olle
ct p
ain
sym
ptom
sta
tus
and
othe
r he
alth
info
rmat
ion
from
patie
nts
with
low
back
pai
n
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
lim
ited
field
tes
ting
Web
-bas
ed s
yste
m u
sing
Del
l Ins
piro
n 11
00 la
ptop
with
Mic
roso
ft X
P o
pera
ting
syst
em 1
4
prime
tou
ch s
cree
n(M
agic
Tou
ch
Key
tec)
ndash P
atie
nt d
emog
raph
ics
ndash O
swes
try
Low
Bac
k P
ain
Dis
abili
ty I
ndex
(V
ersi
on 2
)ndash
Bec
k D
epre
ssio
n In
vent
ory
ndash M
OS
Sho
rt F
orm
-36
(MO
S S
F-3
6)
Not
des
crib
edP
hysi
cian
s
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
S160
Smith et al
considerably in terms of content format anddelivery (eg electronic paper or both) SeveralCDSSs scored and summated patient responseson standard pain and QOL-related assessmentmeasures Based on the published descriptions atleast five of the CDSSs were designed to gener-ate output in real time at the patientrsquos medicalvisit
In terms of systems architecture all CDSSsreviewed were stand-alone personal computer-based systems None interfaced with existingelectronic medical records systems pharmacyappointment scheduling or laboratory resultsreporting Three either were Web-based or hadthe capacity to use a Web-based platform
Table 2 summarizes the types of studies con-ducted to date to evaluate chronic pain CDSSs Ofthe eight CDSSs identified five had publishedevaluation results With one exception all werefeasibility studies exclusively Studies were con-ducted in both inpatient and outpatient settingsAmong the outpatient studies two had been con-ducted with PCPs the remainder involved special-ists in tertiary care settings Study designs used forevaluating these CDSSs varied two were cross-sectional two involved immediate pre- and post-assessments of CDSS use one was longitudinal(12-month follow-up) and one was a focus groupStudy sample sizes ranged from 213 to 4 with themajority having 50 or fewer subjects
Patient acceptability of the CDSS was the sin-gle most commonly assessed variable Evaluationsof both the SymptomReport and the PAIN
Report-It
employed a common tool to assess patientacceptability Results across four pilot testsinvolving a total of 254 subjects consistentlyshowed high acceptability of these two CDSSsand 100 completion rates in terms of data input[495152] The average amount of time requiredranged from a high of 38 minutes to a low of14 minutes
Two studies addressed the issue of medicalaccuracy of the system-generated recommenda-tion Both studies examined this issue by compar-ing system-generated diagnoses andor treatmentrecommendations with those generated by physi-cian experts based on a select sample of patientcases Results showed moderate to high agreementbetween the system- and expert-generated recom-mendations [4748]
Clinician perceptions concerning ease of useand value of a CDSS for chronic pain managementwere examined in two studies Overall physiciansfound the system to be moderately easy to use and
of some clinical worth [4852] Knab and col-leagues [48] reported that the average cliniciantime spent per case on the PMA to obtain outputwas approximately 5 minutes (
plusmn
34 minutes) [48]In addition 85 of physicians adopted the PMA-generated recommendations and 25 of thestudy patients seen were referred to a pain special-ist with the average time to referral being37 months (
plusmn
06 months) [48]Analyses concerning the impact of CDSSs on
patient outcomes were limited Huang and col-leagues [51] assessed changes in pain intensity pre-and post-CDSS use in a sample of radiationoncology clinic patients Although there was adownward trend in pain levels over time resultswere not statistically significant possibly dueto the small size of the sample (N
=
15) Wilkieand colleagues [4952] reported qualitative dataregarding the impact of the SymptomReport andthe PAIN
ReportIt
on patient behavior Resultswere contradictory Of the 41 outpatients whoused the SymptomReport approximately 68stated that it had not affected their pain-relatedcommunication Some however felt that theytalked more precisely and explicitly about theirpain as a result of using it Other comments asso-ciated with use of the SymptomReport includedan increased awareness of pain symptoms andgreater compliance with pain symptom manage-ment In contrast 86 of users of the PAIN-
ReportIt
cited it as beneficial for patientndashdoctorpain-related communication and that it ldquofreedthem to describe their painrdquo
Discussion
Over the past two decades there has been a gradualbut steady growth in research on the use of CDSSsfor chronic pain management To date the numberof such systems is small but expanding Advanceshave also been evident in terms of both the quantityand quality of evaluation studies conducted Whilethe earliest versions were presented in the litera-ture as prototypes only [4546] CDSSs developedsince 2001 have all undergone some form of fieldtesting The majority of these studies howeverhave been nonexperimental in design and focusedexclusively on process measures such as patientor clinician ratings of system acceptability andusability Other salient process measures such asthe degree to which the clinician andor patientactually reviewed and utilized system output orhad confidence in its accuracy have not been con-sistently assessed Poor usability and practitioner
Computerized Decision-Support System for Pain
S161
Tab
le 2
Eva
luat
ion
of c
ompu
teri
zed
deci
sion
-sup
port
sys
tem
s (C
DS
Ss)
in c
hron
ic p
ain
man
agem
ent
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
Nie
lsen
et a
l
[47
]D
iagn
ostic
Hea
dach
eD
iary
PC
Ps
Not
des
crib
edndash
a
gree
men
t of c
ompu
ter-
gene
rate
d vs
exp
ert
phys
icia
n di
agno
ses
ndash 10
0 a
gree
men
t of c
ompu
ter-
gene
rate
d vs
ex
pert
phy
sici
an d
iagn
oses
Kna
b et
al
[48
]P
ain
Man
agem
ent
Adv
isor
N
=
50
PC
Ps
N
=
50
chro
nic
pain
pat
ient
sLo
ngitu
dina
l w
ith 1
2-m
onth
PC
P fo
llow
-up
ndash E
ase
of u
sendash
Med
ical
app
ropr
iate
ness
of
reco
mm
enda
tions
ndash
phy
sici
ansrsquo
ado
ptin
g P
ain
Man
agem
ent
Adv
isor
tre
atm
ent
reco
mm
enda
tions
ndash
pat
ient
s re
ferr
ed t
o pa
insp
ecia
lty c
linic
Phy
sici
an a
dopt
ion
of s
yste
m-g
ener
ated
re
com
men
datio
ns 8
5 o
f ca
ses
Ave
rage
phy
sici
an t
ime
spen
t pe
r ca
se
49
min
utes
(S
D
plusmn
34)
Eas
e of
use
as
rate
d by
phy
sici
ans
42
(
plusmn
28
cm)
on s
cale
of
1ndash10
25
of
patie
nts
refe
rred
to
pain
spe
cial
ty
clin
icA
vera
ge t
ime
to p
ain
spec
ialty
ref
erra
l 3
7 m
onth
s (S
D
plusmn
06)
ndash 70
o
f no
nref
erre
d pa
tient
s st
ill r
ecei
ving
co
mpu
ter-
reco
mm
ende
d tr
eatm
ent
1 ye
ar
post
Wilk
ie e
t al
[49
]S
ympt
omR
epor
tN
=
41
outp
atie
nts
with
can
cer
Cro
ss-s
ectio
nal
tele
phon
e in
terv
iew
sndash
13-it
em p
atie
nt a
ccep
tabi
lity
scal
e as
sess
ing
ease
of
use
ofS
ympt
omR
epor
tndash
Inpu
t co
mpl
etio
n tim
endash
Qua
litat
ive
asse
ssm
ent
of d
egre
eof
com
mun
icat
ion
with
hea
lth c
are
prov
ider
s re
gard
ing
pain
and
oth
ersy
mpt
oms
ndash M
ean
time
to c
ompl
ete
Sym
ptom
Rep
ort
382
min
utes
(S
D
plusmn
202
)ndash
Mea
n tim
e to
com
plet
e S
ympt
omC
onsu
lt
209
min
utes
(S
D
plusmn
18
6)ndash
71
of
part
icip
ants
rat
ed S
ympt
omR
epor
t as
eas
y e
njoy
able
an
d in
form
ativ
endash
68
rep
orte
d th
at th
e am
ount
and
con
tent
of
the
ir pa
in-r
elat
ed c
omm
unic
atio
n w
ith
thei
r do
ctor
had
not
cha
nged
muc
hndash
Qua
litat
ive
patie
nt c
omm
ents
1)
help
ed
them
tal
k m
ore
expl
icitl
y ab
out
pain
2)
gave
the
m g
reat
er a
war
enes
s of
pai
n sy
mpt
oms
3)
incr
ease
d un
ders
tand
ing
of
and
enha
nced
com
plia
nce
rega
rdin
g sy
mpt
om m
anag
emen
t ndash
Pat
ient
com
men
ts r
e S
ympt
omC
onsu
lt 1
) co
ntai
ned
too
muc
h in
form
atio
n 2
) no
t ta
rget
ed t
o in
divi
dual
nee
ds (
3) p
rovi
ded
no n
ew in
form
atio
n
Wilk
ie e
t al
[52
]PA
IN
Rep
ortIt
N
=
213
of
who
m 1
06w
ere
canc
er in
patie
nts
10 m
etas
tatic
can
cer
outp
atie
nts
and
97
wer
e in
divi
dual
sex
perie
ncin
g ac
ute
orch
roni
c pa
in r
ecru
ited
from
non
heal
th c
are
setti
ngs
Des
crip
tive
cro
ss-
sect
iona
l stu
dy in
3se
tting
s 1
) te
rtia
ryca
re
2) r
adia
tion
onco
logy
clin
ic
3)m
obile
clin
ic
ndash 13
-item
sca
le m
easu
ring
acce
ptab
ility
of
PAIN
Rep
ortIt
(ie
tim
e to
com
plet
e e
ase
of u
se
unde
rsta
ndab
ility
of
dire
ctio
ns
ergo
nom
ic e
lem
ents
of
syst
em
and
com
plet
enes
s of
res
pons
es)
ndash 86
o
f res
pond
ents
rate
d th
e PA
IN
Rep
ortIt
as
ben
efici
al f
or p
ain
com
mun
icat
ion
ndash 10
0 p
atie
nt c
ompl
etio
n ra
te o
f PA
IN
Rep
ortIt
ndash M
ean
time
to p
atie
nt
com
plet
ion
15
8 m
inut
es (
SD
plusmn
67
)ndash
Mea
n pa
tient
acc
epta
bilit
y sc
ore
11
7 (S
D
plusmn
16
) sc
ores
ran
ged
from
6 t
o 13
80
o
f pa
tient
s ra
ted
acce
ptab
ility
as
grea
ter
than
min
imum
crit
erio
n of
10
ndash U
ser
com
men
ts 1
) so
me
mec
hani
cal
diffi
culti
es 2
) re
activ
ity t
o us
e of
sys
tem
(e
g
vom
iting
) 3)
pre
fere
nce
to r
elay
info
rmat
ion
dire
ctly
to
prov
ider
PC
P p
rimar
y ca
re p
hysi
cian
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
Computerized Decision-Support System for Pain
S159
Sym
ptom
Rep
ort
and
Sym
ptom
Con
sult
[49]
To a
ssis
t cl
inic
ians
in a
sses
sing
can
cer-
rela
ted
chro
nic
pain
and
fatig
ue
and
clar
ify p
atie
ntsrsquo
mis
belie
fs a
bout
pain
ass
essm
ent
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
Pat
ient
sel
f-re
port
ed in
put
ndash P
atie
nt d
emog
raph
ics
ndash 19
70 v
ersi
on o
f M
cGill
Pai
n Q
uest
ionn
aire
(M
PQ
)ndash
Bar
riers
Que
stio
nnai
re(B
Q)
ndash S
chw
artz
can
cer
fatig
uesc
ale
(SC
FS
-6)
1) H
ard
copi
es o
f ex
pert
sys
tem
repo
rt g
iven
to
patie
nt a
nd t
ocl
inic
ian
2) P
atie
nt r
ecei
ves
educ
atio
nal
mat
eria
ls o
n ho
w t
o re
port
pa
in
use
pain
med
icat
ions
sa
fely
an
d m
anag
e fa
tigue
M
ater
ials
are
cus
tom
ized
to
the
patie
ntrsquos
nee
ds a
nd
pres
ente
d in
an
inte
ract
ive
m
ultim
edia
form
at P
atie
nts
have
opt
ion
to re
ad o
r lis
ten
to
info
rmat
ion
on t
he c
ompu
ter
prin
t an
y or
all
of t
he
mat
eria
ls
or d
o bo
th
Onc
olog
ynu
rses
ot
her
clin
icia
ns
Dec
isio
n-su
ppor
tco
mpu
ter
prog
ram
for
canc
er p
ain
man
agem
ent
[50]
To im
prov
e th
eon
colo
gy n
urse
srsquode
cisi
on m
akin
gre
late
d to
can
cer
pain
man
agem
ent
amon
g cu
ltura
llydi
vers
e fe
mal
eon
colo
gy p
atie
nts
ndash S
urve
y da
ta o
n m
ultic
ultu
ral
canc
er p
ain
char
acte
ristic
sw
ere
anal
yzed
usi
ng f
uzzy
infe
renc
e lo
gic
to d
evel
op 4
mod
ules
1)
a ge
neric
know
ledg
e ba
se 2
) a
cultu
re-s
peci
fic k
now
ledg
e ba
se 3
) de
cisi
on-m
akin
g
and
4) s
elf-
adap
tatio
nndash
Dec
isio
n-m
akin
g m
odul
eco
nsis
ts o
f 2
sets
of
fuzz
yin
fere
nce
logi
c de
velo
ped
via
a ge
netic
alg
orith
m
Har
dwar
e no
t de
scrib
edndash
Ada
ptiv
e fu
zzy
logi
cso
ftwar
e us
ed t
o de
velo
pan
d ru
n th
e kn
owle
dge
base
gen
erat
ion
and
the
deci
sion
-mak
ing
and
self-
adap
tatio
n m
odul
es
Nur
se-e
nter
ed d
ata
base
don
pat
ient
inte
rvie
w
ndash P
atie
nt d
emog
raph
ics
ndash P
ain
char
acte
ristic
s
Com
pute
r sc
reen
dis
play
of
anal
gesi
c tr
eatm
ent
reco
mm
enda
tions
bas
ed o
n th
eW
orld
Hea
lth O
rgan
izat
ion
(WH
O)rsquos
ana
lges
ic la
dder
Onc
olog
ynu
rses
PAIN
Rep
or
tIt a
ndPA
IN
Con
sultN
[51
52]
To a
ssis
t cl
inic
ians
in a
sses
sing
chr
onic
pain
and
to
educ
ate
patie
nts
rega
rdin
gpa
in m
onito
ring
and
man
agem
ent
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
fiel
dte
stin
g co
nduc
ted
Mic
roS
oft W
indo
ws
959
8pe
rson
al c
ompu
ters
with
touc
h sc
reen
(P
en-T
ab)
data
sto
red
in A
cces
s 97
data
base
Pat
ient
sel
f-re
port
ed in
put
1)
Pat
ient
dem
ogra
phic
sndash
McG
ill P
ain
Que
stio
nnai
rendash
Pai
n st
atus
ndash P
atie
nt g
oals
for
and
expe
ctat
ions
abo
ut p
ain
ndash Ty
pe a
nd e
ffect
iven
ess
of p
revi
ous
pain
trea
tmen
ts
Har
d co
pies
of
expe
rt s
yste
mre
port
giv
en t
o pa
tient
and
to
clin
icia
n o
n-sc
reen
vie
win
g of
repo
rt is
als
o po
ssib
le
Onc
olog
ynu
rses
and
othe
rcl
inic
ians
Touc
h-sc
reen
Com
pute
rA
sses
smen
t of
Chr
onic
Low
Bac
k P
ain
[53]
To c
olle
ct p
ain
sym
ptom
sta
tus
and
othe
r he
alth
info
rmat
ion
from
patie
nts
with
low
back
pai
n
Not
des
crib
edW
orki
ng v
ersi
onde
velo
ped
lim
ited
field
tes
ting
Web
-bas
ed s
yste
m u
sing
Del
l Ins
piro
n 11
00 la
ptop
with
Mic
roso
ft X
P o
pera
ting
syst
em 1
4
prime
tou
ch s
cree
n(M
agic
Tou
ch
Key
tec)
ndash P
atie
nt d
emog
raph
ics
ndash O
swes
try
Low
Bac
k P
ain
Dis
abili
ty I
ndex
(V
ersi
on 2
)ndash
Bec
k D
epre
ssio
n In
vent
ory
ndash M
OS
Sho
rt F
orm
-36
(MO
S S
F-3
6)
Not
des
crib
edP
hysi
cian
s
Nam
e of
CD
SS
Pur
pose
Des
crip
tion
Sta
ge o
fD
evel
opm
ent
Har
dwar
e amp
Sof
twar
eR
equi
rem
ents
Dat
a In
put(
s)O
utpu
t(s)
Targ
et
Rec
ipie
nt(s
)of
Out
put
S160
Smith et al
considerably in terms of content format anddelivery (eg electronic paper or both) SeveralCDSSs scored and summated patient responseson standard pain and QOL-related assessmentmeasures Based on the published descriptions atleast five of the CDSSs were designed to gener-ate output in real time at the patientrsquos medicalvisit
In terms of systems architecture all CDSSsreviewed were stand-alone personal computer-based systems None interfaced with existingelectronic medical records systems pharmacyappointment scheduling or laboratory resultsreporting Three either were Web-based or hadthe capacity to use a Web-based platform
Table 2 summarizes the types of studies con-ducted to date to evaluate chronic pain CDSSs Ofthe eight CDSSs identified five had publishedevaluation results With one exception all werefeasibility studies exclusively Studies were con-ducted in both inpatient and outpatient settingsAmong the outpatient studies two had been con-ducted with PCPs the remainder involved special-ists in tertiary care settings Study designs used forevaluating these CDSSs varied two were cross-sectional two involved immediate pre- and post-assessments of CDSS use one was longitudinal(12-month follow-up) and one was a focus groupStudy sample sizes ranged from 213 to 4 with themajority having 50 or fewer subjects
Patient acceptability of the CDSS was the sin-gle most commonly assessed variable Evaluationsof both the SymptomReport and the PAIN
Report-It
employed a common tool to assess patientacceptability Results across four pilot testsinvolving a total of 254 subjects consistentlyshowed high acceptability of these two CDSSsand 100 completion rates in terms of data input[495152] The average amount of time requiredranged from a high of 38 minutes to a low of14 minutes
Two studies addressed the issue of medicalaccuracy of the system-generated recommenda-tion Both studies examined this issue by compar-ing system-generated diagnoses andor treatmentrecommendations with those generated by physi-cian experts based on a select sample of patientcases Results showed moderate to high agreementbetween the system- and expert-generated recom-mendations [4748]
Clinician perceptions concerning ease of useand value of a CDSS for chronic pain managementwere examined in two studies Overall physiciansfound the system to be moderately easy to use and
of some clinical worth [4852] Knab and col-leagues [48] reported that the average cliniciantime spent per case on the PMA to obtain outputwas approximately 5 minutes (
plusmn
34 minutes) [48]In addition 85 of physicians adopted the PMA-generated recommendations and 25 of thestudy patients seen were referred to a pain special-ist with the average time to referral being37 months (
plusmn
06 months) [48]Analyses concerning the impact of CDSSs on
patient outcomes were limited Huang and col-leagues [51] assessed changes in pain intensity pre-and post-CDSS use in a sample of radiationoncology clinic patients Although there was adownward trend in pain levels over time resultswere not statistically significant possibly dueto the small size of the sample (N
=
15) Wilkieand colleagues [4952] reported qualitative dataregarding the impact of the SymptomReport andthe PAIN
ReportIt
on patient behavior Resultswere contradictory Of the 41 outpatients whoused the SymptomReport approximately 68stated that it had not affected their pain-relatedcommunication Some however felt that theytalked more precisely and explicitly about theirpain as a result of using it Other comments asso-ciated with use of the SymptomReport includedan increased awareness of pain symptoms andgreater compliance with pain symptom manage-ment In contrast 86 of users of the PAIN-
ReportIt
cited it as beneficial for patientndashdoctorpain-related communication and that it ldquofreedthem to describe their painrdquo
Discussion
Over the past two decades there has been a gradualbut steady growth in research on the use of CDSSsfor chronic pain management To date the numberof such systems is small but expanding Advanceshave also been evident in terms of both the quantityand quality of evaluation studies conducted Whilethe earliest versions were presented in the litera-ture as prototypes only [4546] CDSSs developedsince 2001 have all undergone some form of fieldtesting The majority of these studies howeverhave been nonexperimental in design and focusedexclusively on process measures such as patientor clinician ratings of system acceptability andusability Other salient process measures such asthe degree to which the clinician andor patientactually reviewed and utilized system output orhad confidence in its accuracy have not been con-sistently assessed Poor usability and practitioner
Computerized Decision-Support System for Pain
S161
Tab
le 2
Eva
luat
ion
of c
ompu
teri
zed
deci
sion
-sup
port
sys
tem
s (C
DS
Ss)
in c
hron
ic p
ain
man
agem
ent
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
Nie
lsen
et a
l
[47
]D
iagn
ostic
Hea
dach
eD
iary
PC
Ps
Not
des
crib
edndash
a
gree
men
t of c
ompu
ter-
gene
rate
d vs
exp
ert
phys
icia
n di
agno
ses
ndash 10
0 a
gree
men
t of c
ompu
ter-
gene
rate
d vs
ex
pert
phy
sici
an d
iagn
oses
Kna
b et
al
[48
]P
ain
Man
agem
ent
Adv
isor
N
=
50
PC
Ps
N
=
50
chro
nic
pain
pat
ient
sLo
ngitu
dina
l w
ith 1
2-m
onth
PC
P fo
llow
-up
ndash E
ase
of u
sendash
Med
ical
app
ropr
iate
ness
of
reco
mm
enda
tions
ndash
phy
sici
ansrsquo
ado
ptin
g P
ain
Man
agem
ent
Adv
isor
tre
atm
ent
reco
mm
enda
tions
ndash
pat
ient
s re
ferr
ed t
o pa
insp
ecia
lty c
linic
Phy
sici
an a
dopt
ion
of s
yste
m-g
ener
ated
re
com
men
datio
ns 8
5 o
f ca
ses
Ave
rage
phy
sici
an t
ime
spen
t pe
r ca
se
49
min
utes
(S
D
plusmn
34)
Eas
e of
use
as
rate
d by
phy
sici
ans
42
(
plusmn
28
cm)
on s
cale
of
1ndash10
25
of
patie
nts
refe
rred
to
pain
spe
cial
ty
clin
icA
vera
ge t
ime
to p
ain
spec
ialty
ref
erra
l 3
7 m
onth
s (S
D
plusmn
06)
ndash 70
o
f no
nref
erre
d pa
tient
s st
ill r
ecei
ving
co
mpu
ter-
reco
mm
ende
d tr
eatm
ent
1 ye
ar
post
Wilk
ie e
t al
[49
]S
ympt
omR
epor
tN
=
41
outp
atie
nts
with
can
cer
Cro
ss-s
ectio
nal
tele
phon
e in
terv
iew
sndash
13-it
em p
atie
nt a
ccep
tabi
lity
scal
e as
sess
ing
ease
of
use
ofS
ympt
omR
epor
tndash
Inpu
t co
mpl
etio
n tim
endash
Qua
litat
ive
asse
ssm
ent
of d
egre
eof
com
mun
icat
ion
with
hea
lth c
are
prov
ider
s re
gard
ing
pain
and
oth
ersy
mpt
oms
ndash M
ean
time
to c
ompl
ete
Sym
ptom
Rep
ort
382
min
utes
(S
D
plusmn
202
)ndash
Mea
n tim
e to
com
plet
e S
ympt
omC
onsu
lt
209
min
utes
(S
D
plusmn
18
6)ndash
71
of
part
icip
ants
rat
ed S
ympt
omR
epor
t as
eas
y e
njoy
able
an
d in
form
ativ
endash
68
rep
orte
d th
at th
e am
ount
and
con
tent
of
the
ir pa
in-r
elat
ed c
omm
unic
atio
n w
ith
thei
r do
ctor
had
not
cha
nged
muc
hndash
Qua
litat
ive
patie
nt c
omm
ents
1)
help
ed
them
tal
k m
ore
expl
icitl
y ab
out
pain
2)
gave
the
m g
reat
er a
war
enes
s of
pai
n sy
mpt
oms
3)
incr
ease
d un
ders
tand
ing
of
and
enha
nced
com
plia
nce
rega
rdin
g sy
mpt
om m
anag
emen
t ndash
Pat
ient
com
men
ts r
e S
ympt
omC
onsu
lt 1
) co
ntai
ned
too
muc
h in
form
atio
n 2
) no
t ta
rget
ed t
o in
divi
dual
nee
ds (
3) p
rovi
ded
no n
ew in
form
atio
n
Wilk
ie e
t al
[52
]PA
IN
Rep
ortIt
N
=
213
of
who
m 1
06w
ere
canc
er in
patie
nts
10 m
etas
tatic
can
cer
outp
atie
nts
and
97
wer
e in
divi
dual
sex
perie
ncin
g ac
ute
orch
roni
c pa
in r
ecru
ited
from
non
heal
th c
are
setti
ngs
Des
crip
tive
cro
ss-
sect
iona
l stu
dy in
3se
tting
s 1
) te
rtia
ryca
re
2) r
adia
tion
onco
logy
clin
ic
3)m
obile
clin
ic
ndash 13
-item
sca
le m
easu
ring
acce
ptab
ility
of
PAIN
Rep
ortIt
(ie
tim
e to
com
plet
e e
ase
of u
se
unde
rsta
ndab
ility
of
dire
ctio
ns
ergo
nom
ic e
lem
ents
of
syst
em
and
com
plet
enes
s of
res
pons
es)
ndash 86
o
f res
pond
ents
rate
d th
e PA
IN
Rep
ortIt
as
ben
efici
al f
or p
ain
com
mun
icat
ion
ndash 10
0 p
atie
nt c
ompl
etio
n ra
te o
f PA
IN
Rep
ortIt
ndash M
ean
time
to p
atie
nt
com
plet
ion
15
8 m
inut
es (
SD
plusmn
67
)ndash
Mea
n pa
tient
acc
epta
bilit
y sc
ore
11
7 (S
D
plusmn
16
) sc
ores
ran
ged
from
6 t
o 13
80
o
f pa
tient
s ra
ted
acce
ptab
ility
as
grea
ter
than
min
imum
crit
erio
n of
10
ndash U
ser
com
men
ts 1
) so
me
mec
hani
cal
diffi
culti
es 2
) re
activ
ity t
o us
e of
sys
tem
(e
g
vom
iting
) 3)
pre
fere
nce
to r
elay
info
rmat
ion
dire
ctly
to
prov
ider
PC
P p
rimar
y ca
re p
hysi
cian
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
S160
Smith et al
considerably in terms of content format anddelivery (eg electronic paper or both) SeveralCDSSs scored and summated patient responseson standard pain and QOL-related assessmentmeasures Based on the published descriptions atleast five of the CDSSs were designed to gener-ate output in real time at the patientrsquos medicalvisit
In terms of systems architecture all CDSSsreviewed were stand-alone personal computer-based systems None interfaced with existingelectronic medical records systems pharmacyappointment scheduling or laboratory resultsreporting Three either were Web-based or hadthe capacity to use a Web-based platform
Table 2 summarizes the types of studies con-ducted to date to evaluate chronic pain CDSSs Ofthe eight CDSSs identified five had publishedevaluation results With one exception all werefeasibility studies exclusively Studies were con-ducted in both inpatient and outpatient settingsAmong the outpatient studies two had been con-ducted with PCPs the remainder involved special-ists in tertiary care settings Study designs used forevaluating these CDSSs varied two were cross-sectional two involved immediate pre- and post-assessments of CDSS use one was longitudinal(12-month follow-up) and one was a focus groupStudy sample sizes ranged from 213 to 4 with themajority having 50 or fewer subjects
Patient acceptability of the CDSS was the sin-gle most commonly assessed variable Evaluationsof both the SymptomReport and the PAIN
Report-It
employed a common tool to assess patientacceptability Results across four pilot testsinvolving a total of 254 subjects consistentlyshowed high acceptability of these two CDSSsand 100 completion rates in terms of data input[495152] The average amount of time requiredranged from a high of 38 minutes to a low of14 minutes
Two studies addressed the issue of medicalaccuracy of the system-generated recommenda-tion Both studies examined this issue by compar-ing system-generated diagnoses andor treatmentrecommendations with those generated by physi-cian experts based on a select sample of patientcases Results showed moderate to high agreementbetween the system- and expert-generated recom-mendations [4748]
Clinician perceptions concerning ease of useand value of a CDSS for chronic pain managementwere examined in two studies Overall physiciansfound the system to be moderately easy to use and
of some clinical worth [4852] Knab and col-leagues [48] reported that the average cliniciantime spent per case on the PMA to obtain outputwas approximately 5 minutes (
plusmn
34 minutes) [48]In addition 85 of physicians adopted the PMA-generated recommendations and 25 of thestudy patients seen were referred to a pain special-ist with the average time to referral being37 months (
plusmn
06 months) [48]Analyses concerning the impact of CDSSs on
patient outcomes were limited Huang and col-leagues [51] assessed changes in pain intensity pre-and post-CDSS use in a sample of radiationoncology clinic patients Although there was adownward trend in pain levels over time resultswere not statistically significant possibly dueto the small size of the sample (N
=
15) Wilkieand colleagues [4952] reported qualitative dataregarding the impact of the SymptomReport andthe PAIN
ReportIt
on patient behavior Resultswere contradictory Of the 41 outpatients whoused the SymptomReport approximately 68stated that it had not affected their pain-relatedcommunication Some however felt that theytalked more precisely and explicitly about theirpain as a result of using it Other comments asso-ciated with use of the SymptomReport includedan increased awareness of pain symptoms andgreater compliance with pain symptom manage-ment In contrast 86 of users of the PAIN-
ReportIt
cited it as beneficial for patientndashdoctorpain-related communication and that it ldquofreedthem to describe their painrdquo
Discussion
Over the past two decades there has been a gradualbut steady growth in research on the use of CDSSsfor chronic pain management To date the numberof such systems is small but expanding Advanceshave also been evident in terms of both the quantityand quality of evaluation studies conducted Whilethe earliest versions were presented in the litera-ture as prototypes only [4546] CDSSs developedsince 2001 have all undergone some form of fieldtesting The majority of these studies howeverhave been nonexperimental in design and focusedexclusively on process measures such as patientor clinician ratings of system acceptability andusability Other salient process measures such asthe degree to which the clinician andor patientactually reviewed and utilized system output orhad confidence in its accuracy have not been con-sistently assessed Poor usability and practitioner
Computerized Decision-Support System for Pain
S161
Tab
le 2
Eva
luat
ion
of c
ompu
teri
zed
deci
sion
-sup
port
sys
tem
s (C
DS
Ss)
in c
hron
ic p
ain
man
agem
ent
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
Nie
lsen
et a
l
[47
]D
iagn
ostic
Hea
dach
eD
iary
PC
Ps
Not
des
crib
edndash
a
gree
men
t of c
ompu
ter-
gene
rate
d vs
exp
ert
phys
icia
n di
agno
ses
ndash 10
0 a
gree
men
t of c
ompu
ter-
gene
rate
d vs
ex
pert
phy
sici
an d
iagn
oses
Kna
b et
al
[48
]P
ain
Man
agem
ent
Adv
isor
N
=
50
PC
Ps
N
=
50
chro
nic
pain
pat
ient
sLo
ngitu
dina
l w
ith 1
2-m
onth
PC
P fo
llow
-up
ndash E
ase
of u
sendash
Med
ical
app
ropr
iate
ness
of
reco
mm
enda
tions
ndash
phy
sici
ansrsquo
ado
ptin
g P
ain
Man
agem
ent
Adv
isor
tre
atm
ent
reco
mm
enda
tions
ndash
pat
ient
s re
ferr
ed t
o pa
insp
ecia
lty c
linic
Phy
sici
an a
dopt
ion
of s
yste
m-g
ener
ated
re
com
men
datio
ns 8
5 o
f ca
ses
Ave
rage
phy
sici
an t
ime
spen
t pe
r ca
se
49
min
utes
(S
D
plusmn
34)
Eas
e of
use
as
rate
d by
phy
sici
ans
42
(
plusmn
28
cm)
on s
cale
of
1ndash10
25
of
patie
nts
refe
rred
to
pain
spe
cial
ty
clin
icA
vera
ge t
ime
to p
ain
spec
ialty
ref
erra
l 3
7 m
onth
s (S
D
plusmn
06)
ndash 70
o
f no
nref
erre
d pa
tient
s st
ill r
ecei
ving
co
mpu
ter-
reco
mm
ende
d tr
eatm
ent
1 ye
ar
post
Wilk
ie e
t al
[49
]S
ympt
omR
epor
tN
=
41
outp
atie
nts
with
can
cer
Cro
ss-s
ectio
nal
tele
phon
e in
terv
iew
sndash
13-it
em p
atie
nt a
ccep
tabi
lity
scal
e as
sess
ing
ease
of
use
ofS
ympt
omR
epor
tndash
Inpu
t co
mpl
etio
n tim
endash
Qua
litat
ive
asse
ssm
ent
of d
egre
eof
com
mun
icat
ion
with
hea
lth c
are
prov
ider
s re
gard
ing
pain
and
oth
ersy
mpt
oms
ndash M
ean
time
to c
ompl
ete
Sym
ptom
Rep
ort
382
min
utes
(S
D
plusmn
202
)ndash
Mea
n tim
e to
com
plet
e S
ympt
omC
onsu
lt
209
min
utes
(S
D
plusmn
18
6)ndash
71
of
part
icip
ants
rat
ed S
ympt
omR
epor
t as
eas
y e
njoy
able
an
d in
form
ativ
endash
68
rep
orte
d th
at th
e am
ount
and
con
tent
of
the
ir pa
in-r
elat
ed c
omm
unic
atio
n w
ith
thei
r do
ctor
had
not
cha
nged
muc
hndash
Qua
litat
ive
patie
nt c
omm
ents
1)
help
ed
them
tal
k m
ore
expl
icitl
y ab
out
pain
2)
gave
the
m g
reat
er a
war
enes
s of
pai
n sy
mpt
oms
3)
incr
ease
d un
ders
tand
ing
of
and
enha
nced
com
plia
nce
rega
rdin
g sy
mpt
om m
anag
emen
t ndash
Pat
ient
com
men
ts r
e S
ympt
omC
onsu
lt 1
) co
ntai
ned
too
muc
h in
form
atio
n 2
) no
t ta
rget
ed t
o in
divi
dual
nee
ds (
3) p
rovi
ded
no n
ew in
form
atio
n
Wilk
ie e
t al
[52
]PA
IN
Rep
ortIt
N
=
213
of
who
m 1
06w
ere
canc
er in
patie
nts
10 m
etas
tatic
can
cer
outp
atie
nts
and
97
wer
e in
divi
dual
sex
perie
ncin
g ac
ute
orch
roni
c pa
in r
ecru
ited
from
non
heal
th c
are
setti
ngs
Des
crip
tive
cro
ss-
sect
iona
l stu
dy in
3se
tting
s 1
) te
rtia
ryca
re
2) r
adia
tion
onco
logy
clin
ic
3)m
obile
clin
ic
ndash 13
-item
sca
le m
easu
ring
acce
ptab
ility
of
PAIN
Rep
ortIt
(ie
tim
e to
com
plet
e e
ase
of u
se
unde
rsta
ndab
ility
of
dire
ctio
ns
ergo
nom
ic e
lem
ents
of
syst
em
and
com
plet
enes
s of
res
pons
es)
ndash 86
o
f res
pond
ents
rate
d th
e PA
IN
Rep
ortIt
as
ben
efici
al f
or p
ain
com
mun
icat
ion
ndash 10
0 p
atie
nt c
ompl
etio
n ra
te o
f PA
IN
Rep
ortIt
ndash M
ean
time
to p
atie
nt
com
plet
ion
15
8 m
inut
es (
SD
plusmn
67
)ndash
Mea
n pa
tient
acc
epta
bilit
y sc
ore
11
7 (S
D
plusmn
16
) sc
ores
ran
ged
from
6 t
o 13
80
o
f pa
tient
s ra
ted
acce
ptab
ility
as
grea
ter
than
min
imum
crit
erio
n of
10
ndash U
ser
com
men
ts 1
) so
me
mec
hani
cal
diffi
culti
es 2
) re
activ
ity t
o us
e of
sys
tem
(e
g
vom
iting
) 3)
pre
fere
nce
to r
elay
info
rmat
ion
dire
ctly
to
prov
ider
PC
P p
rimar
y ca
re p
hysi
cian
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
Computerized Decision-Support System for Pain
S161
Tab
le 2
Eva
luat
ion
of c
ompu
teri
zed
deci
sion
-sup
port
sys
tem
s (C
DS
Ss)
in c
hron
ic p
ain
man
agem
ent
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
Nie
lsen
et a
l
[47
]D
iagn
ostic
Hea
dach
eD
iary
PC
Ps
Not
des
crib
edndash
a
gree
men
t of c
ompu
ter-
gene
rate
d vs
exp
ert
phys
icia
n di
agno
ses
ndash 10
0 a
gree
men
t of c
ompu
ter-
gene
rate
d vs
ex
pert
phy
sici
an d
iagn
oses
Kna
b et
al
[48
]P
ain
Man
agem
ent
Adv
isor
N
=
50
PC
Ps
N
=
50
chro
nic
pain
pat
ient
sLo
ngitu
dina
l w
ith 1
2-m
onth
PC
P fo
llow
-up
ndash E
ase
of u
sendash
Med
ical
app
ropr
iate
ness
of
reco
mm
enda
tions
ndash
phy
sici
ansrsquo
ado
ptin
g P
ain
Man
agem
ent
Adv
isor
tre
atm
ent
reco
mm
enda
tions
ndash
pat
ient
s re
ferr
ed t
o pa
insp
ecia
lty c
linic
Phy
sici
an a
dopt
ion
of s
yste
m-g
ener
ated
re
com
men
datio
ns 8
5 o
f ca
ses
Ave
rage
phy
sici
an t
ime
spen
t pe
r ca
se
49
min
utes
(S
D
plusmn
34)
Eas
e of
use
as
rate
d by
phy
sici
ans
42
(
plusmn
28
cm)
on s
cale
of
1ndash10
25
of
patie
nts
refe
rred
to
pain
spe
cial
ty
clin
icA
vera
ge t
ime
to p
ain
spec
ialty
ref
erra
l 3
7 m
onth
s (S
D
plusmn
06)
ndash 70
o
f no
nref
erre
d pa
tient
s st
ill r
ecei
ving
co
mpu
ter-
reco
mm
ende
d tr
eatm
ent
1 ye
ar
post
Wilk
ie e
t al
[49
]S
ympt
omR
epor
tN
=
41
outp
atie
nts
with
can
cer
Cro
ss-s
ectio
nal
tele
phon
e in
terv
iew
sndash
13-it
em p
atie
nt a
ccep
tabi
lity
scal
e as
sess
ing
ease
of
use
ofS
ympt
omR
epor
tndash
Inpu
t co
mpl
etio
n tim
endash
Qua
litat
ive
asse
ssm
ent
of d
egre
eof
com
mun
icat
ion
with
hea
lth c
are
prov
ider
s re
gard
ing
pain
and
oth
ersy
mpt
oms
ndash M
ean
time
to c
ompl
ete
Sym
ptom
Rep
ort
382
min
utes
(S
D
plusmn
202
)ndash
Mea
n tim
e to
com
plet
e S
ympt
omC
onsu
lt
209
min
utes
(S
D
plusmn
18
6)ndash
71
of
part
icip
ants
rat
ed S
ympt
omR
epor
t as
eas
y e
njoy
able
an
d in
form
ativ
endash
68
rep
orte
d th
at th
e am
ount
and
con
tent
of
the
ir pa
in-r
elat
ed c
omm
unic
atio
n w
ith
thei
r do
ctor
had
not
cha
nged
muc
hndash
Qua
litat
ive
patie
nt c
omm
ents
1)
help
ed
them
tal
k m
ore
expl
icitl
y ab
out
pain
2)
gave
the
m g
reat
er a
war
enes
s of
pai
n sy
mpt
oms
3)
incr
ease
d un
ders
tand
ing
of
and
enha
nced
com
plia
nce
rega
rdin
g sy
mpt
om m
anag
emen
t ndash
Pat
ient
com
men
ts r
e S
ympt
omC
onsu
lt 1
) co
ntai
ned
too
muc
h in
form
atio
n 2
) no
t ta
rget
ed t
o in
divi
dual
nee
ds (
3) p
rovi
ded
no n
ew in
form
atio
n
Wilk
ie e
t al
[52
]PA
IN
Rep
ortIt
N
=
213
of
who
m 1
06w
ere
canc
er in
patie
nts
10 m
etas
tatic
can
cer
outp
atie
nts
and
97
wer
e in
divi
dual
sex
perie
ncin
g ac
ute
orch
roni
c pa
in r
ecru
ited
from
non
heal
th c
are
setti
ngs
Des
crip
tive
cro
ss-
sect
iona
l stu
dy in
3se
tting
s 1
) te
rtia
ryca
re
2) r
adia
tion
onco
logy
clin
ic
3)m
obile
clin
ic
ndash 13
-item
sca
le m
easu
ring
acce
ptab
ility
of
PAIN
Rep
ortIt
(ie
tim
e to
com
plet
e e
ase
of u
se
unde
rsta
ndab
ility
of
dire
ctio
ns
ergo
nom
ic e
lem
ents
of
syst
em
and
com
plet
enes
s of
res
pons
es)
ndash 86
o
f res
pond
ents
rate
d th
e PA
IN
Rep
ortIt
as
ben
efici
al f
or p
ain
com
mun
icat
ion
ndash 10
0 p
atie
nt c
ompl
etio
n ra
te o
f PA
IN
Rep
ortIt
ndash M
ean
time
to p
atie
nt
com
plet
ion
15
8 m
inut
es (
SD
plusmn
67
)ndash
Mea
n pa
tient
acc
epta
bilit
y sc
ore
11
7 (S
D
plusmn
16
) sc
ores
ran
ged
from
6 t
o 13
80
o
f pa
tient
s ra
ted
acce
ptab
ility
as
grea
ter
than
min
imum
crit
erio
n of
10
ndash U
ser
com
men
ts 1
) so
me
mec
hani
cal
diffi
culti
es 2
) re
activ
ity t
o us
e of
sys
tem
(e
g
vom
iting
) 3)
pre
fere
nce
to r
elay
info
rmat
ion
dire
ctly
to
prov
ider
PC
P p
rimar
y ca
re p
hysi
cian
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
S162
Smith et al
Hua
ng e
t al
[51
]PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1 N
=
9pa
tient
s w
ith b
one
met
asta
sis-
rela
ted
pain
Pilo
t st
udy
2 N
=
15
patie
nts
with
can
cer
and
bone
met
asta
sis
Phy
sici
an fo
cus
grou
pN
=
4 r
adia
tion
onco
logi
sts
1) P
ilot
test
1
Fea
sibi
lity
stud
y us
ing
a te
stndashr
etes
t w
ithin
-su
bjec
t de
sign
2) P
ilot
test
2
Fea
sibi
lity
stud
y us
ing
an 1
1-da
y te
stndashr
etes
tw
ithin
-sub
ject
des
ign
3) P
hysi
cian
focu
sgr
oup
Out
com
e m
easu
res
used
for
both
pilo
t st
udie
sndash
Acc
epta
bilit
yndash
Com
plet
enes
sndash
Tim
e to
com
plet
endash
Val
idity
Phy
sici
an fo
cus
grou
pndash
Rec
eptiv
ity t
o PA
IN
Rep
ortIt
and
PAIN
Con
sultN
Pilo
t st
udy
1
ndashM
ean
time
to c
ompl
ete
PAIN
Rep
ortIt
at
pret
est
12 m
inut
es (
SD
plusmn
4)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at
post
-te
st 1
ndash7 m
inut
es p
er p
atie
ntndash
Mea
n ac
cept
abili
ty s
core
11
2 (S
D
plusmn
18
)ndash
100
com
plet
ion
rate
Pilo
t st
udy
2ndash
Mea
n tim
e to
com
plet
e PA
IN
Rep
ortIt
at
pret
est
17 m
inut
es (
SD
plusmn
6)
ndashT
ime
to c
ompl
ete
PAIN
Rep
ortIt
at p
ostte
st
14 m
inut
es (
SD
plusmn
8)
ndashM
ean
acce
ptab
ility
sco
re 1
22
(SD
plusmn
13
)ndash
100
com
plet
ion
rate
ndashPA
IN
Con
sultN
rec
omm
ende
d a
med
ian
of
4 dr
ugs
phy
sici
ans
pres
crib
ed a
med
ian
of
3 dr
ugs
post
use
ndashP
atie
nt p
ain
inte
nsity
ave
rage
4 a
t bas
elin
e an
d 2
7 at
pos
ttest
(no
t si
gnifi
cant
)
Foc
us g
roup
ndashP
hysi
cian
s sa
w v
alue
of
PAIN
Rep
ortIt
1)
incr
ease
d ef
ficie
ncy
durin
g cl
inic
vis
it 2
) su
pple
men
ted
pain
ser
vice
con
sulta
tion
3)
prov
ided
out
com
e da
tandash
PAIN
Con
sultN
was
vie
wed
as
clin
ical
ad
junc
t bu
t fo
rmat
ting
need
ed
impr
ovem
ent
Koe
stle
r et
al
[53
]To
uch-
scre
en C
ompu
ter
Ass
essm
ent
of C
hron
icLo
w B
ack
Pai
n
N
=
30
low
bac
kpa
in p
atie
nts
Cro
ss-s
ectio
nal d
esig
nin
ter
tiary
car
e cl
inic
ndashP
atie
nt-r
atin
gs o
f er
gono
mic
desi
gn d
egre
e of
tec
hnic
aldi
fficu
lties
ac
cept
abili
ty
ease
of u
se d
ata
secu
rity
ndashM
ean
time
to c
ompl
ete
the
67-it
em t
ouch
sc
reen
27
4 m
inut
es (
SD
plusmn
138
)
Stu
dyN
ame
of C
DS
SS
ampl
eD
esig
nO
utco
mes
Ass
esse
dR
esul
ts
PC
P p
rimar
y ca
re p
hysi
cian
Tab
le 2
Con
tinue
d
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
Computerized Decision-Support System for Pain
S163
nonacceptance of computer recommendations canserve as significant barriers to system adoption inroutine clinical practice [365455]
User preferences regarding the presentation ofcomputer output including content formatting(eg color graphics) and length have not beensolicited in most instances either Similarly thereare few published data concerning technical diffi-culties (eg type and number of system crashes ortouch-screen calibration problems) encounteredby CDSS users Both issues have important rami-fications for future system refinements [52] Addi-tionally there is a paucity of information oncontextual circumstances (eg presence of a localldquochampionrdquo of the system) or the processes usedto integrate the CDSS into the existing clinicalworkflow key considerations for successful systemimplementation Not least testing has been con-fined almost exclusively to either inpatient ortertiary care settings with only two studies con-ducted in the primary care context to date
The effects of these systems on patient out-comes remain understudied Two studies reportedqualitative data concerning CDSS impact onpatientsrsquo perceived pain-related communicationwith their physician however sample sizes weresmall and results inconsistent [4952] One studyreported system impact on patient pain intensitylevel over time but the study lacked adequate sta-tistical power to detect clinically important differ-ences [52] Other major patient outcomes such ashealth care utilization health care costs painrelief pain medication usage communication withhealth care provider about pain functional statusand QOL have not been examined One studyreported evidence that CDSS use may invokepatient reactivity (eg vomiting intensified painsymptoms) Potential adverse effects on patientsshould be measured in future investigations [52]Similarly there is need for more extensive andconsistent examination of system impact on clini-cian pain management performance
While we sought to be as comprehensive aspossible in our literature search our review wasrestricted to only English-language studies Inaddition it is possible that there are other CDSSsunder development that we failed to identify Thelimited size of the available literature as well asthe methods used in these primary studies pre-vented us from conducting a meta-analysis ofresearch findings and from reaching more defin-itive conclusions about the impact of these systemson physician performance and patient pain func-tioning and other aspects of QOL Notably in all
the studies we examined study investigators andCDSS developers were one and the same a factthat may have resulted in more positive findings[36] Lastly we did not conduct a separate evalu-ation of the clinical appropriateness of either theCDSS algorithms or treatment recommendationsnor of the underlying logic employed to generatesuch algorithms
The potential for these computer-based sys-tems to improve the quality of chronic pain man-agement in the primary care context is substantialTo manage chronic pain effectively PCPs firstneed to conduct a comprehensive patient assess-ment [56] Information on the patientrsquos pain ex-perience history of and preferences for paintreatment psychological status approach to self-management and personal goalspriorities are keyvariables to collect during assessment as they arecritical for making an accurate diagnosis and fordeveloping an appropriate treatment plan to whichthe patient will adhere [56] An expert system-typeCDSS provides a way to elicit and integrate suchpatient-specific information in a manner that isconvenient and timely for both physicians andpatients Moreover the ensuing system-generatedrecommendations are individualized to the needsand circumstances of the specific pain patient perbest clinical practices [56]
CDSSs developed for chronic pain manage-ment have as yet however to fulfill this promiseAs our review indicates systems developed thus farhave been predominantly biomedical in focus anddesigned to assist physicians and other health careproviders in the medical management of painsymptoms (including invasive procedures andreferrals) exclusively Only a few of these systemshave reached a sufficiently advanced stage ofdevelopment to warrant more rigorous testingin large-scale randomized controlled trials[264952] Such trials are imperative for under-standing system effect on provider performanceand patient outcomes
Significantly none of the systems reviewedwere integrated with existing electronic recordssystems nor did they include reminder or docu-mentation functionalities features which have allbeen shown to increase the likelihood of physicianadoption [5758] This lack of integration mayreflect the fact that widespread adoption of elec-tronic records systems by health care institutionshas been a relatively recent occurrence Potentiallythis trend coupled with pressures from majoraccrediting agencies to document the provision ofpain screening and treatment along with the
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
S164
Smith et al
recent publication of primary care pain manage-ment guidelines may serve to spur additionalmore rigorous research on the use of CDSSs forchronic pain management in primary care[565960] Demonstrating the clinical value ofthese systems is a critical step in convincing healthcare organizations and clinicians that the benefitsof investing in a CDSS for pain management out-weigh potential risks In particular physiciansneed to be assured that this type of system canenhance rather than erode their decision-makingabilities and that time spent learning how to usea CDSS yields measurable improvement in patienthealth and well-being
References
1 American Pain Society Chronic pain in AmericaRoadblock to relief 1999 American Pain SocietyAPS News and Announcements Available at httpwwwampainsocorgwhatsnewtoc_roadhtmtoc
2 Lande SD The problem of pain in managed careIn Lande SD Kulich RJ eds Managed Care andPain Glenview IL American Pain Society 200019
3 Turner JA Leresche L Von Korff M Ehrlich LPrimary care back pain patient characteristics visitcontent and short term outcomes Spine199823463ndash9
4 Von Korff M Gruman J Schaefer J Curry SJ Wag-ner EH Collaborative management of chronic ill-ness Ann Intern Med 1997127(12)1097ndash102
5 Von Korff M Pain management in primary care Anindividualized stepped-care approach In GatchelRJ Turk DL eds Psychosocial Factors in PainCritical Perspectives New York The GuilfordPress 1999
6 Von Korff M Katon W Bush T et al Treatmentcosts cost offset and cost-effectiveness of collabo-rative management of depression Psychosom Med199860(2)143ndash9
7 Bertakis KD Azari R Callahan EJ Patient pain Itsinfluence on primary care physicianndashpatient inter-action Fam Med 200335(2)119ndash23
8 Green CR Wheeler JRC Laporte F Marchant BGuerrero E How well is chronic pain managedWho does it well Pain Med 20023(1)56ndash62
9 Von Roenn JH Cleeland CS Gonin R HatfieldAK Pandya KJ Physician attitudes and practice incancer pain management A survey from the EasternCooperative Oncology Group Ann Intern Med1993119121ndash6
10 Cleeland CS Cleeland LM Dar R Rinehardt LCFactors influencing physician management of can-cer pain Cancer 198658796ndash800
11 Fife BL Irick N Painter JD A comparative studyof the attitudes of physicians and nurses towards themanagement of cancer pain J Pain Symptom Man-age 19938132ndash9
12 Wilson JF Brockoop GW Kryst S Steger H WittWO Medical studentsrsquo attitudes towards painbefore and after a brief course on pain Pain199250251ndash6
13 Weinstein SM Laux LF Thornby JI et al Physi-ciansrsquo attitudes towards pain and the use of opioidanalgesics Results of a survey from the Texas Can-cer Pain Initiative South Med J 200093(5)479ndash87
14 Weinstein SM Laux LF Thornby JI et al Medicalstudentsrsquo attitudes towards pain and the use of opi-oid analgesics Implications for changing medicalschool curriculum South Med J 200093(5)472ndash8
15 Whedon M Ferrell BR Professional and ethicalconsiderations in the use of high-tech pain manage-ment Oncol Nurs Forum 1991181135ndash43
16 Ward S Goldberg N Miller-Mccauley V et alPatient-related barriers to management of cancerpain Pain 199352319ndash24
17 Pargeon KL Hailey BJ Barriers to effective cancerpain management A review of the literature J PainSymptom Manage 199918358ndash68
18 Breitbart W Passik S Mcdonald MV et al Patient-related barriers to pain management in ambulatoryAIDS patients Pain 199352319ndash24
19 Gunnarsdottir S Donovan HS Serlin RC Voge CWard S Patient-related barriers to pain manage-ment The barriers questionnaire II (BQ-II) Pain200299385ndash96
20 Arora NK Interacting with cancer patients Thesignificance of physiciansrsquo communication behaviorSoc Sci Med 200357791ndash806
21 Mccaffery M Thorpe DM Differences in percep-tion of pain and the development of adversarial rela-tionships among health care providers In Hill CSFields W eds Advances in Pain Research and Ther-apy Drug Treatment of Cancer Pain in a Drug-Oriented Society Vol 11 New York Raven Press1989
22 Ong LML de Haes JCJM Hoos AM Lammes FBDoctorndashpatient communication A review of the lit-erature Soc Sci Med 199540903ndash18
23 Jones WL Rimer BK Levy MH Kinman JL Can-cer patientsrsquo knowledge beliefs and behaviorregarding pain control regimens Implications foreducation programs Patient Educ Couns19845(4)159ndash64
24 Lukoschek P Fazzari M Marantz P Patient andphysician factors predict patientsrsquo comprehensionof health information Patient Educ Couns200350201ndash10
25 Donovan JL Blake DR Patient non-complianceDeviance or reasoned decision-making Soc SciMed 199234377ndash94
26 Garg AX Adhikari NKJ Mcdonald H et al Effectsof computerized clinical decision support systemson practitioner performance and patient outcomesa systematic review JAMA 2005293(10)1223ndash38
27 Nilasena DS Lincoln MJ A computer-generatedreminder system improves physician compliance
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
Computerized Decision-Support System for Pain
S165
with diabetes preventive care guidelines Proc AnnSymp Comput Appl Med Care 1995640ndash5
28 Chambers CV Balaban DJ Charlson BL Grass-berger DM The effect of microcomputer-gener-ated reminders on influenza vaccination rates in auniversity-based family practice center J Am BoardFam Pract 1991419ndash26
29 Flanagan JR Doebbeling BN Dawson J BeekmannS Randomized study of online vaccine reminders inadult primary care Proc AMIA Symp 1999755ndash9
30 Burack RC Gimotty PA Promoting screeningmammography in inner-city settings The sustainedeffectiveness of computerized reminders in a ran-domized controlled trial Med Care 199735921ndash31
31 Rossi RA Every NR A computerized interventionto decrease the use of calcium channel blockers inhypertension J Gen Intern Med 199712672ndash8
32 Montgomery AA Fahey T Peters TJ MacintoshC Sharp DJ Evaluation of computer based clinicaldecision support system and risk chart for manage-ment of hypertension in primary care Randomizedcontrolled trial BMJ 2000320686ndash90
33 Shea S Dumouchel W Bahamonde L A meta-analysis of 16 randomized controlled trials toevaluate computer-based reminder systems for pre-ventive care in the ambulatory setting J Am MedInform Assoc 19963399ndash409
34 Unrod M Smith MY DePue J Spring B WinkelG Randomized controlled trial of a computer-based tailored intervention to increase smokingcessation counseling by primary care physicians JGen Intern Med 200722478ndash84
35 Balas EA Austin SM Mitchell JA et al The clinicalvalue of computerized information services Areview of 98 randomized clinical trails Arch FamMed 19965271ndash8
36 Hunt DL Haynes RB Hanna SE Smith K Effectsof computer-based clinical decision support systemson physician performance and patient outcome Asystematic review JAMA 1998280(15)1339ndash46
37 Revere D Dunbar PJ Review of computer-gener-ated outpatient health behavior interventions Clin-ical encounters ldquoin absentiardquo J Am Med InformAssoc 20018(1)62ndash79
38 Prochaska JO Velicer WF Redding C et al Stage-based expert systems to guide a population of pri-mary care patients to quit smoking eat healthierprevent skin cancer and receive regular mammo-grams Prev Med 200541(2)406ndash16
39 Prochaska JO Velicer WF Fava JL Rossi JS TsohJY Evaluating a population-based recruitmentapproach and a stage-based expert system inter-vention for smoking cessation Addict Behav200126(4)583ndash602
40 Strecher VJ Kreuter M Den Boer DJ et al Theeffects of computer-tailored smoking cessation mes-sages in family practice settings J Fam Pract199439(3)262ndash70
41 Strecher VJ Shiffman S West R Randomized con-trolled trial of a web-based computer-tailored smok-ing cessation program as a supplement to nicotinepatch therapy Addiction 2005100(5)682ndash8
42 Dijkstra A De Vries H Roijackers J Long-termeffectiveness of computer-generated tailored feed-back in smoking cessation Health Educ Res199813(2)207ndash14
43 Dijkstra A De Vries H Roijackers J van BreukelenG Tailored interventions to communicate stage-matched information to smokers in differentmotivational stages J Consult Clin Psychol199866(3)549ndash57
44 Kawamoto K Houlihan CA Balas EA Lobach DFImproving clinical practice using clinical decisionsupport systems A systematic review of trials toidentify features critical to success BMJ2005330765ndash73
45 Matsumura Y RHINOS A consultation system fordiagnosis of headache and facial pain ComputMethods Programs Biomed 19862365ndash71
46 Thomas J IVAN An expert system for pain controland symptom relief in advance cancer PC AI199913(4)28ndash30
47 Nielsen KD Rasmussen C Russell MB The Diag-nostic Headache Diary A headache expert systemIn Paiva T Penzel T eds European NeurologicalNetwork Amsterdam IOS Press 2000
48 Knab JH Wallace MS Wagner RL Tsoukatos JWeinger MB The use of a computer-based decisionsupport system facilitates primary care physiciansrsquomanagement of chronic pain Anesth Analg200193712ndash20
49 Wilkie DJ Huang H Berry DL et al Cancersymptom control Feasibility of a tailored interac-tive computerized program for patients Fam Com-munity Health 200124(3)48ndash62
50 Im E Chee W Decision support computer pro-gram for cancer pain management Comput InformNurs 200321(1)12ndash21
51 Huang H Wilkie DJ Zong S et al Developing acomputerized data collection and decision supportsystem for cancer pain management ComputInform Nurs 200321(4)206ndash17
52 Wilkie DJ Judge MK Berry DL Dell J Zong SGilespie R Usability of a computerized PAIN
Repor-tIt
in the general public with pain and peoplewith cancer pain J Pain Symptom Manag200325(3)213ndash24
53 Koestler ME Libby E Schofferman J Redmon TWeb-based screen computer assessment of chroniclow back pain A pilot study Comput Inform Nurs200523(5)275ndash84
54 Wyatt JC Spiegelhalter DJ Evaluating medicalexpert systems What to test and how Med Inform(Lond) 199015(3)205ndash17
55 Reisman Y Computer-based clinical decision-aidsA review of methods and assessment of systemsMed Inform (Lond) 199621179ndash97
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8
S166
Smith et al
56 Gruener D Lande SD eds Pain Control in thePrimary Care Setting Glenview IL American PainSociety 2006
57 Shiffman RN Liaw Y Brandt CA Corb GJComputer-based guideline implementation sys-tems A systematic review of functionality andeffectiveness J Am Med Inform Assoc 19996104ndash14
58 Muller ML Ganslandt T Eich HP Lang K Ohm-ann C Prokosch HU Towards integration of
clinical decision support in commercial hospitalinformation systems using distributed reusablesoftware and knowledge components Int J MedInform 200163369ndash77
59 Phillips D for the Joint Commission on Accredita-tion of Healthcare Organizations JCAHO painmanagement standards are unveiled JAMA2000284428ndash9
60 Frankenstein RS Letters to the editor reply JAMA20002842317ndash8