comprehensive care programs for patients with multiple chronic conditions: a systematic literature...

R

Cc

SCa

b

Tc

a

ARRA

KCCEFMI

(l((c

0h

Health Policy 107 (2012) 108– 145

Contents lists available at SciVerse ScienceDirect

Health Policy

j ourna l ho me p ag e: www.elsev ier .com/ locate /hea l thpol

eview

omprehensive care programs for patients with multiple chroniconditions: A systematic literature review

imone R. de Bruina,∗, Nathalie Versnelb,c, Lidwien C. Lemmensa,laudia C.M. Molemaa, Franc ois G. Schellevisb,c, Giel Nijpelsb, Caroline A. Baana

National Institute for Public Health and the Environment, Centre for Prevention and Health Services Research, PO Box 1, 3720 BA Bilthoven, The NetherlandsDepartment of General Practice/EMGO+ Institute for Health and Care Research, VU University Medical Center, v.d. Boechorststraat 7, 1081 BT Amsterdam,he NetherlandsNIVEL, Netherlands Institute for Health Services Research, PO Box 1568, 3500 BN Utrecht, The Netherlands

r t i c l e i n f o

rticle history:eceived 16 November 2011eceived in revised form 19 June 2012ccepted 21 June 2012

eywords:hronic careomprehensive careffectivenessrailtyultimorbidity

ntegrated care

a b s t r a c t

Objective: To provide insight into the characteristics of comprehensive care programs forpatients with multiple chronic conditions and their impact on patients, informal caregivers,and professional caregivers.Methods: Systematic literature search in multiple electronic databases for English languagepapers published between January 1995 and January 2011, supplemented by referencetracking and a manual search on the internet. Wagner’s chronic care model (CCM) wasused to define comprehensive care. After inclusion, the methodological quality of eachstudy was assessed. A best-evidence synthesis was applied to draw conclusions.Results: Forty-two publications were selected describing thirty-three studies evaluatingtwenty-eight comprehensive care programs for multimorbid patients. Programs varied inthe target patient groups, implementation settings, number of included interventions, andnumber of CCM components to which these interventions related. Moderate evidence wasfound for a beneficial effect of comprehensive care on inpatient healthcare utilization andhealthcare costs, health behavior of patients, perceived quality of care, and satisfaction ofpatients and caregivers. Insufficient evidence was found for a beneficial effect of compre-hensive care on health-related quality of life in terms of mental functioning, medicationuse, and outpatient healthcare utilization and healthcare costs. No evidence was found fora beneficial effect of comprehensive care on cognitive functioning, depressive symptoms,functional status, mortality, quality of life in terms of physical functioning, and caregiver
burden.Conclusion: Because of the heterogeneity of comprehensive care programs, it is as yet tooearly to draw firm conclusions regarding their effectiveness. More rigorous evaluation stud-ies are necessary to determine what constitutes best care for the increasing number of
le chro
people with multip
∗ Corresponding author. Tel.: +31 0 30 2744352.E-mail addresses: [email protected], [email protected]

S.R. de Bruin), [email protected], [email protected] (N. Versnel),[email protected] (L.C. Lemmens), [email protected]. Molema), [email protected], [email protected]. Schellevis), [email protected] (G. Nijpels),[email protected] (C.A. Baan).

168-8510/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.healthpol.2012.06.006

nic conditions.© 2012 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Chronic diseases are the leading cause of disabilityand death in the western part of the world [1]. Over
the coming years, the prevalence of chronic diseases ispredicted to increase because of the rapid ageing of theworld population and the greater longevity of people withchronic conditions [2–4]. An increasing proportion of the
dx.doi.org/10.1016/j.healthpol.2012.06.006

http://www.sciencedirect.com/science/journal/01688510

http://www.elsevier.com/locate/healthpol

mailto:[email protected]










dx.doi.org/10.1016/j.healthpol.2012.06.006

lth Polic
S.R. de Bruin et al. / Hea
chronically ill is multimorbid [5,6]. The term “multimor-bidity” refers to any co-occurrence of multiple chronicconditions within one person, and is to be distinguishedfrom the term “co-morbidity” which refers to the combina-tion of additional diseases beyond an index disorder [6–8].The prevalence of multiple chronic conditions among indi-viduals increases with age and is substantial among olderadults [5]. In people over age 65 about 65% is multimorbid;in people over age 85 this proportion is about 85% [6,8,9].

Having multiple chronic conditions is associated withpoor quality of life, disability, psychological distress, andan increased mortality risk [6,10]. The complex healthcareneeds of most patients with multiple chronic conditionsusually require the involvement of a large number ofhealthcare providers in the care process. Coordination ofcare is therefore difficult and healthcare providers oftenfail to address the integral healthcare demand, includ-ing psychosocial issues, of patients with multiple chronicconditions [8,11–13]. As a result, multimorbid patientsare prone to receive fragmented, incomplete, inefficient,and ineffective care, which in turn may lead to unneces-sary hospitalizations, increased use of emergency facilities,polypharmacy, adverse drug events, duplicative tests, andconflicting medical advices [5,8,14,15].

Increasingly, comprehensive care programs are imple-mented in healthcare systems worldwide to addressthe integral healthcare demand of multimorbid patients[5,16,17]. Models are considered to be comprehensiveif they address several healthcare needs of people withmultiple chronic conditions, functional disabilities, and/orhigh healthcare utilization and if healthcare services areprovided by multiple healthcare providers [17]. Compre-hensive care programs are patient-centered and aim tostructure and coordinate delivery of healthcare services[5,16,17]. Such programs are referred to as e.g. integratedcare programs, disease management programs, guidedcare, transitional care or shared care [13,17]. By improv-ing quality and continuity of care, these programs aim toimprove patient health outcomes, whilst making efficientuse of healthcare resources [13,18,19].

Reviews that provide an overview of implementedcomprehensive care programs for patients with multiplechronic conditions indicate that evidence for their impacton patients and caregivers is inconsistent [17,20–25]. Inmost of these reviews criteria for the inclusion of studieswere rather narrow. Some reviews predominantly includedrandomized controlled trials [17,21,23], whereas othersjust evaluated the impact of programs on patient out-comes [17,23–25] or included studies suggesting positiveeffects of comprehensive care programs [17]. Furthermore,most reviews provide limited information about the con-tents of the comprehensive care programs included. As aresult, relevant information from non-randomized trialsand information about the impact of programs on caregiveroutcomes have hardly been summarized yet. Moreover, lit-tle is known about the program characteristics that maybe related to positive outcomes of care and about the
patient groups that may benefit most from comprehen-sive care. This information is, however, of importance sinceinterest in what constitutes best care for patients with mul-timorbidity is growing. We therefore performed a thorough
y 107 (2012) 108– 145 109

review of the literature using less narrow inclusion crite-ria than earlier review studies. We aimed to provide moreinsight into the characteristics of comprehensive care pro-grams for patients with multiple chronic conditions andtheir impact on patients, informal caregivers, and profes-sional caregivers.

2. Methods

2.1. Study design and search strategy

We conducted a systematic literature search in theelectronic databases Cinahl, Cochrane, Embase, Medline,PsycInfo, and SciSearch. The databases were searched forEnglish language papers focusing on comprehensive careprograms for people aged 18 years and older and publishedbetween January 1995 and January 2011. January 1995 wasused as a starting point since from that year on compre-hensive care has become an increasingly important focusof attention in the medical literature [26].

A comprehensive search strategy was developed by alibrarian and two members of the project team (SdB andLL) to identify studies matching the following search terms(Medical Subject Headings): case management, compre-hensive healthcare, critical pathways, disease management,continuity of patient care, patient care management, planningor team, patient-centered care, delivery of integrated health-care, guided care, integrated care, managed care (programs),shared care, transmural care and variations of the key-words chronic disease, chronic illness, co-morbidity, frailty,multimorbidity, multiple chronic conditions, and specificchronic conditions. These search terms were combinedwith variations of the following search terms: benefits,effects, effectiveness, efficacy, impact, outcomes, and specific(health) outcomes. In addition to the search in the elec-tronic databases, relevant papers were identified throughreference tracking and through a manual literature searchon the internet. The internet was searched if the electronicdatabase search yielded papers that reported on the designof an intervention study or a pilot intervention study. Wethen searched the internet to determine the status of thesestudies and to find out whether their results had alreadybeen published in scientific papers.

2.2. Study selection

Four reviewers (SdB, NV, LL, CM) worked in pairs andindependently reviewed the papers yielded by the searchfor their relevance by screening their title and abstract.When considered relevant by both reviewers, the full-text paper was retrieved. Any disagreement between thereviewers was resolved by consensus. In line with ear-lier studies [e.g. 27–31], the chronic care model (CCM)of Wagner et al. [18,32] was used to define comprehen-sive care. The model suggests that comprehensive careprograms ideally comprise six interrelated components.Four components refer to the actual delivery of care by
healthcare providers: 1. self-management support that helpspatients and their families to obtain skills and confidenceto manage their chronic condition (e.g. regular health edu-cation and coaching); 2. delivery system design that ensures

1 lth Polic

tipmlcspittcvpnoaritfsastdt

ttcpdis

2a

eeo45caq

pimcfccrhccd

papers were selected by the reviewers to be retrieved full-text for in-depth screening. This screening process resulted

10 S.R. de Bruin et al. / Hea

he delivery of effective, efficient patient care through e.g.nvolvement of all members of the multidisciplinary team,lanned patient interactions, regular follow-up, and caseanagement; 3. decision support by evidence-based guide-

ines providing clinical standards for high-quality chronicare, and 4. development of clinical information systems thatupply care teams with feedback, remind them to com-ly with practice guidelines, provide registries for planning

ndividual and population-based care, and enable proac-ive interactions between caregivers and patients. Thewo remaining components refer to the context in whichhronic care is provided: 5. the healthcare system that pro-ides the organizational context in which chronic care isrovided and encompasses the aforementioned compo-ents. A healthcare system that endorses improvementf the quality of care must be well-organized, motivated,nd prepared to change; and 6. links towards communityesources and policies. The healthcare system is embeddedn a community that includes organizations and programshat may support or expand a healthcare system’s careor chronically ill patients (e.g. senior centers that addressocial issues of patients or fitness centers that offer physicalctivity programs) [32,33]. In our study, only comprehen-ive care programs that included interventions related towo or more components of the CCM were included. Weecided on a minimum of two components since, accordingo the CCM, components must be interrelated.

For the present study, papers were eligible if they methe following inclusion criteria: 1. the program described inhe paper met our operational definition of comprehensiveare; 2. the target population of the program consisted ofeople with multiple chronic conditions; and 3. the studyescribed in the paper was an intervention study evaluat-

ng the impact of a comprehensive care program. Duplicatetudies were identified and excluded.

.3. Data extraction, quality assessment, and datanalysis

Four authors (SdB, NV, LL, CM) worked in pairs toxtract relevant data from the studies included. Dataxtracted from the papers were: 1. study design; 2. lengthf follow-up of study; 3. target population of the program;. setting in which the program had been implemented;. content of the program described in terms of the CCMomponents; 6. characteristics of the usual care condition;nd 7. study outcomes. Our analyses were descriptive andualitative in nature.

Of the studies that compared healthcare costs of com-rehensive care with those of usual care, we reported the

ncremental healthcare costs of comprehensive care. Incre-ental costs can be defined as the difference in healthcare

osts for patients receiving comprehensive care and thoseor patients receiving care as usual. Negative incrementalosts imply that healthcare costs for patients receivingomprehensive care are lower than those for patientseceiving usual care. Positive incremental costs imply that
ealthcare costs for patients receiving comprehensiveare are higher than those for patients receiving usualare. For the present paper, we only report incrementalirect healthcare costs, such as costs of primary care visits,
y 107 (2012) 108– 145

emergency room (ER) visits, and hospitalization. Costestimates were adjusted for cross-country purchasingpower differences (PPP), using 2008 US$ PPP, and forinflation, using Gross Domestic Product prices in order tomake meaningful comparisons across studies [34]. Fourstudies did not report the year of data collection. For thesestudies the year of publication was used as reference year.

The methodological quality of the selected studies wasscored independently by three reviewers (SdB, NV, CM) onthe basis of six items adapted from two quality criterialists [35,36] (Table 2). Our list included criteria that in ouropinion were most relevant for studies on comprehensivecare. The reviewers rated each criterion as ‘+’ (i.e. crite-rion fulfilled), ‘−’ (i.e. criterion not fulfilled), ‘?’ (i.e. lackof information) or ‘N.A.’ (i.e. not applicable). Since we didnot use a complete standardized set of assessment crite-ria, we decided not to assign an overall quality rating perstudy. Hence, we only provide a total quality sum-score(ranging from 0 to 6) per study that was determined bycounting the number of criteria scored positively. To drawconclusions regarding the effectiveness of comprehensivecare programs, in line with earlier studies a best-evidenceanalysis was applied [37–39]. We distinguished four lev-els to rate the strength of evidence for each of theoutcomes:

1. Strong evidence: consistent evidence for a beneficialeffect of comprehensive care across multiple studieswith high quality sum-scores;

2. Moderate evidence: consistent evidence for a beneficialeffect of comprehensive care across multiple studies,including at least one study with a high quality sum-score and studies with lower quality sum-scores;

3. Insufficient evidence: inconsistent evidence for a ben-eficial effect of comprehensive care across multiplestudies;

4. No evidence: only few studies available and/or consis-tent evidence for no effect of comprehensive care acrossmultiple studies.

Similar to previous reviews that applied a best-evidencesynthesis [37,38] results were considered consistent whenat least 75% of the studies showed results in the samedirection. Findings of studies with relatively higher qual-ity sum-scores were valued more than findings of studieswith relatively lower quality sum-scores.

3. Results

3.1. Study retrieval

Our literature search yielded 3544 potentially relevantpublications. On the basis of their title and abstract, 184

in 32 publications for inclusion in our study. The addition of10 papers from our manual search resulted in a total of 42publications that were classified as eligible for our review.Reasons for exclusion are shown in Fig. 1.

S.R. de Bruin et al. / Health Policy 107 (2012) 108– 145 111

Potentiall y rele vant pa per s identified by electronic

datab ase sea rch: N= 3544

Cinahl: n=536; Co chra ne: n=494; Embas e/

PsycInfo/SciSea rch: n=182 4; Medline: n=690

Papers exclude d: n= 3360. Rea sons for

exclusion:

1. No co mpr ehensive ca re (n=2620)

2. No multi morbi dity (n=513)

3. No inte rventi on study (n= 103)

4. Other re asons * (n=70)

5. Duplicates (n=54 )

Full-text pa pers retrie ved fo r in- depth scree nin g:

n=184

Cinahl: n=31; Cochr ane: n=79 ; Embas e/Ps ycInfo/

SciSearch: n=30; Medline: n=4 4

Papers exclude d: n= 152. Reas ons for

exclusion:

1. No co mpr ehensive ca re (n=31)

2. No multi morbi dity (n=80)

3. No inte rventi on study (n= 22)

4. Other re asons * (n=19)

Total nu mbe r of paper s eligible for ou r study: n= 32

Cinahl: n=6; Coch rane: n=18; Em base/PsycInf o/

SciSearch: n=4; Medline: n= 4

Papers id entified in

manual searc h:

N=10

Total nu mbe r of pape rs in clud ed in ou r sy ste matic

literature review : n=42

not be re

n study o

f literat

*e.g. papers were not written in Engli sh, papers co uld

papers, papers described t he design of an interventio

Fig. 1. Flow diagram o

3.2. Study characteristics

3.2.1. Study designs and length of follow-upThe 42 included publications described 33 different

studies i.e. fourteen randomized controlled trials, eightpretest-post-test studies, five controlled clinical trials,three cluster randomized controlled trials, two post-testonly studies, and one case control study (Table 1). Acrossall studies, sample size varied from 25 to 65,132 subjects. Of
the studies with a longitudinal design, twenty studies hada follow-up of ≤12 months and seven studies had a follow-up of >12 months. In three studies the length of follow-upwas variable (e.g. depending on time to discharge).
trieve d, identified docum ents were no n-sc ientific

nly.

ure screening process.

3.2.2. Methodological quality of studiesNone of the studies fulfilled all quality criteria (i.e.

sum-score of 6) based on what could be retrieved fromthe information provided in the papers (Table 2). Theobserved minimum sum-score was 0 (n = 6) and the maxi-mum sum-score was 5 (n = 1). The quality criteria “similarat baseline” and “drop-out rate” were most frequentlyscored positively. Only three papers reported enough infor-mation to assess the quality criterion regarding compliance
[40–42]. The quality criteria “randomization”, “similarity atbaseline”, “application of intention-to-treat-analysis”, and“adjustment for confounders” were not applicable for stud-ies with a non-experimental design (n = 11).

112S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Table 1Characteristics and outcomes of studies evaluating the impact of comprehensive care programs for multimorbid patients.

Author (year) Study design(N)a

Length offollow-up(months)b

Targetpopulation

Setting Contents of comprehensive care programc Usual cared Outcomese

Anonymous [63] Pretest-posttestdesign (pilot)(N = 60)

12 Highly complexpatients withabout 7diagnoses andhigh healthcareconsumption.

Hospital care,communitycare, and homecare in the USA

Enhanced Care Initiatives (ECI)CR: coordination of transportation, meals, and othersupport services in the community.DS: ECI for physicians in case of questions or concerns.DSD: home visit by nurse for initial geriatricassessment, then weekly or biweekly contact andsometimes daily contact; physicians are regularlyinformed about patients; nurse accompanies patientto physician.HS: fee for physicians for their time and cooperationwith the program.SMS: group sessions at a community center foreducation, exercise, and entertainment; nurse giveseducation, reviews and upgrades dietary choices,suggests home safety improvements, and implementsan exercise program tailored to the patients needs.

N.A. PatientsInpatient and outpatient hospital

expenditures decreased considerably (i.e.$265.085) after implementation of theprogram (change over time was nottested for significance).

Béland et al. [43] RCT (N = 1309) 22 Frail olderpeople aged ≥65years withfunctionaldisabilities andcomplex serviceneeds and theirinformalcaregivers.

Community,primary, andhospital care byIntegratedServices for FrailElders (SIPA)sites (locatedwithin localcommunityservice centers)in Canada

Integrated Services for Frail Elders (SIPA)CIS: development IT system to monitor performanceof team with regard to administrative procedures andservices received by patients.CR: partnership with local community service centersto rapidly mobilize community based services.DS: design of interdisciplinary intervention protocolsand service delivery processes based on the latestknowledge and adapted to local population; SIPAtraining for all personnel.DSD: responsibility of SIPA site for the delivery of allservices, including health and social, acute andlong-term, and community and institutional care;supervision of SIPA by program director beingresponsible for clinical staff; integrated health andsocial services offered under responsibility of a casemanager and a multidisciplinary team; regularadaptation of intervention plans including services tobe provided to patients’ circumstances and needs;24/7 on-call services.HS: SIPA sites clinically and financially independent;system based on community primary care services(health and social, acute and long-term, andcommunity and institutional, including acute carehospitals and nursing homes); system created toassess quality of services and management onongoing basis; implementation of changes in clinical,organizational, and management practices;development of inter-organizational agreements (e.g.shared clinical responsibility); installation ofcommittees to support team and patients (regional,site monitoring, inter-site coordination, and workingcommittee).SMS: involvement of patients and caregivers in careplanning.

Usual servicesavailable to frailolder people.

PatientsAverage total healthcare costs per

patient during the study did not differsignificantly between groups; 1. costs forcommunity services were significantlyhigher for intervention than for controlgroup; 2. costs for institutional serviceswere significantly lower for interventionthan for control group. Incrementalhealthcare costs were −$204 per patientper year.

Utilization of home care and GPservices, and number of hours nursinghome care during the study weresignificantly higher in intervention thanin control group. Utilization of specialistcare, prescribed drugs, number of days inacute care, number of days in long-termacute care, number of days in nursinghomes, number of hours of socialservices, and number of hours at ERduring the study did not differsignificantly between groups.

% of patients with long-term stay wassignificantly lower during the study inintervention than in control group. % ofpatients accessing acute and ER care andwho were institutionalized did not differsignificantly between groups.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145113

Table 1 (Continued )



Targetpopulation


Boult et al. [45];Boyd et al.[44]; Leff et al.[46], andWolff et al.[47]

cRCT (N = 904patients; N = 308informalcaregivers;N = 33professionalcaregivers)

8 (healthcarecosts andutilization), 12(professionalcaregivers), and18 (otherpatientoutcomes andinformalcaregivers)

Multimorbidpeople aged ≥65years with highprobability ofusing healthservicesintensively/athigh risk forincurring highhealthcare costsin the comingyear.

Primary care by8 primary carepractices in theUSA

Guided Care (GC): integrating registered nurse (GCnurse) into primary care practice to collaborate with2-5 physicians in providing comprehensive care to50-60 patients.CIS: clinical processes provided by GC nurse areguided by web-accessible electronic health records.CR: establishing access to community resources forpatients by GC nurse.DS: education for GC nurses to provide comprehensivecare; clinical processes provided by GC nurse areguided by scientific evidence.DSD: comprehensive assessment at home;development care plan; monthly monitoring; supporttransitions into and out of hospitals; coordination ofall providers of care by GC nurse.SMS: involvement of patient in development self-careplan; referral to self-management course; coachingfor self-management; education and support of familycaregivers by GC nurse.

N.R. PatientsMean annual healthcare costs and

healthcare utilization (i.e. hospitaladmissions, hospital days, skilled nursingfacility admissions, skilled nursing facilitydays, ER visits, GP visits, specialist visits,home healthcare) during the study didnot differ significantly between groups.Incremental healthcare costs were−$1394 per patient per year.

Perceived quality of care at follow-upwas significantly higher in interventionthan in control group; 1. mean overallquality of care score, and mean goalsetting, coordination of care, problemsolving, and patient activation sub scoreswere significantly higher in interventionthan in control group; 2. interventiongroup was significantly (1.5–2 times)more likely than control group to ratetheir overall quality of care, decisionsupport, and coordination of care as“high-quality”.

Mean decision support sub score andrating of goal setting, problem solving,and patient activation at follow-up didnot differ significantly between groups.

Patients in intervention group thatrated baseline quality of care as “low”were significantly (twice) more likely torate their overall quality of care atfollow-up as “medium or high”.Informal caregivers

Caregiver depression and caregiverstrain at follow-up did not differsignificantly between groups.

Perceived quality of care at follow-upwas significantly higher in caregivers ofpatients receiving GC than of thosereceiving usual care. Caregivers reportedsignificantly higher quality for 4 out of 5subscales: goal setting, coordination ofcare, decision support, and patientactivation. Perceived quality of care forproblem solving did not differsignificantly between groups.

Productivity loss in terms of regularactivity and work productivity atfollow-up did not differ significantlybetween caregivers from both groups.

114S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Professional caregiversOver time, physicians providing guided

care significantly more improved theirsatisfaction with 5 out of 11 aspects of care(i.e. communication with patients,communication with family caregivers,education of family caregivers, motivationof patients, and referral to communityresources) and with 1 out of 6 aspects ofknowledge (i.e. knowledge of medicationspatients are taking) than physiciansproviding care as usual.

Change in rating of provider team’sproblem-solving performance, patientactivation, other aspects of knowledge, timespent on patient care, and care coordinationover time did not differ significantlybetween physicians from both groups.

Boyd et al. [48]and Sylviaet al. [49]

CCT (pilot)(N = 150)

6 Multimorbidpeople aged ≥65years and withhigh probabilityof using healthservicesintensively inthe coming year.

Primary care by1 primary carepractice in theUSA

Guided Care (see Boult et al. [45]) N.R. PatientsHealthcare costs and healthcare

utilization (i.e. hospital admissions, hospitaldays, ER visits, primary care visits did notdiffer significantly between groups duringthe study. Incremental healthcare costswere −$3089 per patient per year.

Change in quality of patient-physiciancommunication over time differedsignificantly between intervention groupand control group: quality ofpatient-physician communicationsignificantly improved in interventiongroup.

Change in physician’s comprehensiveknowledge of patient, quality of integrationof care, interpersonal treatment, and trust inphysician over time did not differsignificantly between groups.

Brand et al. [65] CCT (N = 166) 3 and 6 General medicalpatients aged≥65 years withan inpatient stayof >24 hours andwith either ahistory ofreadmissions toacute care,multiplemedicalco-morbidities,or admissiondiagnosis ofCHF.

Hospital care by1 tertiaryteachinghospital inAustralia

Transitional care service model.CR: within 2 weeks after discharge (if necessary)coordination of community services by chronicdisease nurse consultant (CDNC).DSD: consult patient with CDNC within 24 hoursbefore discharge from hospital (pre-discharge data,screening risk factors for readmission, liaison withdischarge planners, nursing staff and allied healthstaff, inform GP about patient’s discharge); consultpatient with CDNC in chronic care clinic within 2weeks after discharge (evaluate patient’s healthstatus, medication, inform patient’s GP, coordinationof liaison between CDNC and GP, organization ofhome visiting).SMS: development of follow-up and action plan (atdischarge) and review of action plan (within 2 weeksafter discharge) by CDNC and patient; addressingself-management and social issues within 2 weeksafter discharge by CDCN.

Medical, nursingand allied healthinterventionconsistent with thepatient’s diagnosisand resourcesavailable on thegeneral medicalwards. Alsoincludes dischargeplanning by themultidisciplinaryteam and couldincludeout-patientfollow-up.

PatientsGP visits and quality of life at 3-month

follow-up did not differ significantlybetween groups.

ER visits and readmission rate at 6-monthfollow-up did not differ significantlybetween groups.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145115




Targetpopulation


Burns et al. [66] RCT (N = 130) 12 Hospitalizedveterans aged≥65 yearsdiagnosed with≥1 ADL deficit,≥2 chronicmedicalconditions, ≥2acutehospitalizationsin previous year,or taking ≥6scheduledprescriptiondrugs.

Hospital care byVeterans Affairsmedical centerin the USA.

Outpatient Geriatric Evaluation andManagement (GEM) clinicDS: participation of interdisciplinary team inextensive team training and development.DSD: interdisciplinary primary care team(geriatric physicians, nurse practitioner,social worker, psychologists, clinicalpharmacists) that focused on evaluation andlong-term management or primary care(particular focus on special elderlyproblems); review by team of patients’ andcaregivers’ resources and needs; initial2-hour assessment was performed by entireteam; team developed goals, interventions,and treatment and provided continuousindividualized follow-up and aftercare afterdischarge (e.g. short-term hospital admission,routine care in clinic by relevant healthcareproviders with telephone follow-up).

After hospitaldischargeinpatienttreatment teamreferred patientsback to previousprovider(s) or tonew provider(s)if necessary.Outpatient carewas provided inambulatory careclinics of the VAmedical centeror by localphysicians inthe community.

PatientsOverall, ADL deteriorated

significantly over time but did notdiffer significantly between groups.

Overall, cognition and IADLimproved significantly over time butdid not differ significantly betweengroups.

Overall, depressive symptoms andhealth perception were significantlyless and higher, respectively inintervention than in control group,and change over time differedsignificantly between the groups infavor of the intervention group.

Overall, the number ofhospitalizations increasedsignificantly over time, but did notdiffer significantly between groups.

Overall, life satisfaction did notdiffer significantly between groupsand did not change significantly overtime. However, change over timediffered significantly between groups,in favor of the intervention group.

Overall, the number of medicationsduring the study was significantlylower in intervention than in controlgroup, but change in the number ofmedications over time did not differsignificantly between groups.

Overall, social activity wassignificantly higher in interventionthan in control group, and changeover time differed significantlybetween groups in favor of theintervention group.

116S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Chumbler et al.[79]

Case controldesign in whichsubjectsreceivingcomprehensivecare werecompared withsubjects fromanotherlongitudinalstudy (N = 226)

12 Veterans with,hypertension,congestive heartfailure,respiratorydisease, ordiabetes, whohad ≥2 ED visits,hospitalizations,or unscheduledwalk-in visits inthe previous 12months, and/orobtained carecosting≥$25,000.

Home care byVeterans HealthSystem in theUSA

Rural Home Care Project (RHCP)CIS: each patient was given one of three types ofhome-telehealth with communication opportunities:1. hand-held in-home messaging device (calledHealth Buddy) with disease management dialogues;monitoring of health status of patients by carecoordinators on daily basis with interactive sessions(± 30 minutes) to determine risk of health-care crises;2. telemonitor with two-way audio video connectivityallowing biometric monitoring; care coordinatorsassess patients in weekly sessions (15-20 minutes);3. videophone with two-way audio video connectivitywithout biometric monitoring; assessment of patientsby care coordinators patients in weekly sessions(15-20 minutes).appropriate system was selected by RHCP team tomeet specific patient needs and to enhanceeffectiveness of care coordinator.DS: decision tree for selecting most appropriatesystem.DSD: provision of care by RHCP team (i.e. geriatricnurse practitioners, social worker, program supportassistant); pro-active management of patients by carecoordinators to prevent ER-visits and hospitaladmissions.SMS: education of all patients by care coordinators onchronic disease; encouragement of self-managementbehavior (e.g. to prevent health crises); enhancementof communication between patient and health-caresystem.

Disabled elders≥60 years, with ≥1(I)ADL limitationrecruited fromanotherlongitudinal study.

PatientsWith the exception of change in

cognitive functioning measured with theMMSE change over time in all outcomesdiffered significantly between the groups,in favor of the intervention group.

IADL and cognitive functioningmeasured on FIM cognitive subscaleimproved significantly over time inintervention group and significantlydeteriorated in the control group. ADL,cognitive functioning measured with theMMSE, and independent functioningsignificantly improved in the interventiongroup and remained stable in the controlgroup.

Coleman et al.[50]

RCT (N = 169) 24 Frail older adultsaged ≥65 yearsat high risk forhospitalizationand functionaldecline in thesubsequent 4years.

Primary care by9 primary carephysicianpractices in theUSA

Chronic care clinicsDS: training to physician and nurses inpopulation-based medicine and managementstrategies (geriatric syndromes); on-the job coachingof nurses by study staff.DSD: 3–4 monthly visits to primary care practice fordevelopment of shared treatment plan by physicianand nurse; session with pharmacist aboutpolypharmacy and medications associated withfunctional decline; health status assessment.SMS: 3–4 monthly visits to primary care practice forself-management group session led by nurse or socialworker.

N.R. PatientsNo significant differences between

intervention and control group indepressive symptoms, proportion ofpatients reporting falling in the past year,health-related quality of life (physicalfunctioning), urinary incontinence, andrate of use of high-risk medications atfollow-up.

No significant difference betweenintervention and control group inproportion of patients rating their carecoordination and overall medical care as“excellent” at follow-up.

No significant differences in healthcareutilization (i.e. primary care visits, ERvisits, hospitalization, hospital days) andhealthcare costs during the studybetween intervention and control group.Incremental healthcare costs were −$726per patient per year.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145117




Targetpopulation


Professional caregiversPhysicians expressed a high level of

support for chronic care levels andbelieved that they were better able tomeet the needs of frail older patients.

Counsell et al.[51,52]

cRCT (N = 951) 24 Older adultsaged ≥65 yearswith an incomeless than 200%of the federalpoverty level.


Geriatric Resources for Assessment and Care of Elders(GRACE)CIS: use of electronic medical record to supportphysician practices, facilitate monitoring of clinicalparameters, and coordination of care across multiplegeriatric syndromes, providers, and sites of care.CR: community-based services liaison from primarycare practice and affiliated health system is memberof interdisciplinary team.DS: training of support team in implementing GRACEprotocols and working as interdisciplinary team; useof institutionally endorsed clinical practice guidelines.DSD: GRACE support team (including an advancedpractice nurse and social worker) that cares for lowincome seniors in collaboration with the patient’sprimary care physician and a geriatricsinterdisciplinary team led by a geriatrician;comprehensive geriatric assessment by support teamat patient’s home; development individualized careplan by GRACE interdisciplinary team; support teamdiscusses individualized care plan with patient’sprimary care physician; implementation of care planthrough contacts with patient, family members orcaregivers by support team in collaboration withphysician and support by GRACE interdisciplinaryteam; at least 1 follow-up home visit to review careplan; 1 monthly contact; face-to-face contact after ERvisit or hospitalization; coordination of care betweenproviders and settings by support team.SMS: encouragement support team for goal settingand self-care; education by support team inproblem-solving skills and in usinglow-health-literacy materials.

Access to allprimary andspecialty careservices available(i.e. outpatientgeriatricassessment andmultispecialtycentre, inpatientacute care forelders unit andconsult centre,skilled nursingfacility, physicianhouse callsprogram,psychiatric care).

PatientsAll patients

No significant differences in changeover time in ADLs, IADLs, days in bed, and4 aspects of quality of life (physicalfunctioning, role-physical, bodily pain,role-emotional), physical functioning(physical component summary score ofSF-36) between intervention and controlgroup.

Significant difference in change overtime in the 4 other aspects of quality oflife (general health, vitality, socialfunctioning, and mental health) and inmental functioning (mental componentsummary score of SF-36) betweenintervention and control group, all infavor of intervention group that improvedsignificantly on these aspects.

Total healthcare costs during the studydid not differ significantly betweenintervention and control group: 1. totalcosts for acute care (i.e. ER andhospitalization) did not differsignificantly between intervention andcontrol group; 2. total costs for chronicand preventive care (i.e. GRACE, primaryand specialty care, procedures,rehabilitation, mental health) weresignificantly higher in intervention thanin control group. Incremental healthcarecosts were $1321 per patient per year.

No significant difference betweengroups at follow-up in proportion ofpatients rating their overall satisfactionwith care as very good or excellent,mortality, and time to death.

ER visit rate was significantly lowerduring the study in intervention than incontrol group. No significant differencebetween groups during the study inhospitalization rate, number of hospitaldays, and readmission within 30 daysafter the first hospitalization.

118S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Patients at high risk of hospitalizationER visit rate and hospital admission

rate were significantly lower during thestudy in intervention than in controlgroup. Number of hospital days duringthe study did not differ significantlybetween groups.

Courtney et al.[67]

RCT (N = 128) 6 Older adultsaged ≥65 yearswith acutemedicaladmission to thehospital and atrisk forreadmission dueto multipleco-morbidities,impairedfunctionality,age ≥75 years,recent multiplereadmissions,poor socialsupport orhistory ofdepression.

Hospital care(partly providedat home) by 1tertiary hospitalin Australia

Older Hospitalized Patients’ Discharge Planning andIn-home Follow-up Protocol (OHP-DP), in addition tousual hospital care.CR: assessment by registered nurse (RN) of need forassistance with activities of daily living andcommunity services.DS: evidence-based exercise intervention; writtenguidelines on discharge management for the patient(health promotion activities, management oftreatment regimens and the exercise plan).DSD: comprehensive patient assessment within 72 hof admission by RN and physiotherapist; design ofgoal-directed individualized care plan incollaboration with patient, health professionals,family and caregivers. Care plan included:1. Individualized exercise intervention designed byphysiotherapist and facilitated by RN and that startedin hospital and continued for 24 weeks.2. Nursing intervention while patient is in hospitalincluding daily visits of patients by RN in hospital;discharge planning by RN; design transitional careplan by RN including functional ability, need forassistance with activities of daily living,post-discharge treatments and follow-up care, socialsupport, chronic disease management plans andinformation, medication information, communityservices, and assistance with exercise program; RNand physiotherapist combined their visits regardingexercise program to ensure continuity.3. Intervention after discharge including home visit ofRN within 48 hours after discharge to assessavailability of support, transitional concern, provideadvice and support, and ensure safe environment tofollow exercise program; extra home visits ifnecessary; weekly telephone monitoring (advice,information, support) by RN for 4 weeks, thereaftermonthly for 5 months.SMS: involvement of patient in designing care plan;diary kept by patient (describing steps and distancerecorded with a pedometer).

Usual hospital care(routine care,dischargeplanning,rehabilitationadvice, in-homefollow up (ifnecessary)normallyprovided).

PatientsNumber of readmissions, emergency

GP visits, and emergency allied healthservice visits during the study weresignificantly lower in intervention than incontrol group. % of patients requiringemergency hospital readmissions andemergency GP visits during the study wassignificantly lower in intervention than incontrol group. % of patients requiring ERvisits did not differ between groups.

Significant difference in change overtime in physical component score andmental component score, general healthsubscale, physical function subscale, rolephysical subscale, and bodily painsubscale scores between intervention andcontrol group. All in favor of theintervention group that significantlyimproved on these aspects.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145119


Author (year) Study design (N)a Length offollow-up(months)b

Target population Setting Contents of comprehensive care programc Usual cared Outcomese

Daly et al. [72]and Douglaset al. [73]

RCT (N = 334) 2 ICU patients with>72 h mechanicalventilation

Hospital care,nursing homecare, andrehabilitationcare provided bylong-term acutecare facility,nursing home,and rehabilitationcentre,respectively inthe USA.

In addition to usual care, Disease Management Programto reduce hospital readmissionsCIS: documentation of all contacts of advanced practicenurse (APN) with patient, family, and healthcareproviders using data collection forms.DS: access of APN to pulmonologist and geriatrician forguidance and collaboration by phone (once or twice amonth), consultation of hospital care team about patientsneeds by APN.DSD: case management of APN who meets with patientand family prior to discharge and makes a care plan forthe patient, this plan was send to all relevant out-of thehospital healthcare providers; weekly visits of APN topatients in first 4 weeks and biweekly in next 4 weeksand whenever necessary (in total at least 8 visits);coordination of care by APN.SMS: patients and family are informed about the care planby the APN; preparation of family for return of patients totheir homes; provision of emotional support to family.

When patients andfamily membersasked for advicethey were referredback to theirprimary careprovider, extendedcare facility staff orhome care agency.

Patients(I)ADL at follow-up did not differ significantly

between groups.Total healthcare costs after discharge did not

differ significantly between groups. Only costsfor readmission during the study differedsignificantly between groups. Cost savingsassociated with intervention were $6198 perreadmitted patient.

Significantly less total hospital days duringreadmission during the study in interventionthan in control group. Relative risk for hospitalreadmission, % of patients having >1 hospitalreadmission, and mean number of days fromhospital discharge to readmission during thestudy did not differ significantly between groups.

Change over time in health-related quality oflife at follow-up did not differ significantlybetween groups.

Mortality during readmission and survivalrates during the study did not differ significantlybetween groups.

Dorr et al. [53,54] CCT (N = 3432, ofwhom 1671 hadcomplex illnessincludingdiabetes)

24 Older adults aged≥65 years withcomplex illness.

Primary care by13 primary careclinics in the USA

Care Management Plus, multidisease care managementprogramCIS: specialized information technology tools (e.g.structured protocols, guidelines, patient worksheet, caremanagement tracking database) for care managers tohelp manage their patients and to track and react topatient-related data by access to care plans, best practicesand facilitating communication through messagingsystem and documenting all contacts with patients,families, and healthcare providers.CR: care manager and care team focus on linking patientsand their caregivers to community resources.DS: training of care managers every other month (e.g.care for seniors, specific chronic illness assessment, carestandards); yearly physician training to encouragereferral for any complex patient (multiple co-morbidities,frailty) and patients with certain illnesses; guidelines andprotocols were regularly revised and updated by localexperts.DSD: employment of care managers (registered nurses) inprimary care clinic who assesses patients for enrollmentin care management services using validatedinstruments; identification and referral of patients to caremanager by physician.HS: administrative and clinical management teamssupport process and quality improvement at all levels ofthe organization, and these leaders encourage opendiscussion of key safety and quality concerns;infrastructure for innovations in treatment of chronicdisease that revolves around primary care teams andinvolves multiple stakeholders; quality indicators tomonitor and improve quality of care.SMS: self-management support by care managers; caremanager and patient develop care plan together.

N.R. PatientsAll patients with complex illness

(PQI) hospitalizations and mortality duringthe study did not differ significantly betweengroups. Significantly more ER visits during thestudy in intervention than in control group.Patients with complex illness including diabetes

No significant difference in ER visits and (PQI)hospitalizations during the study betweenintervention and control group.

Significantly lower mortality during the studyin intervention than in control group.

120S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Duke [55] Pretest-posttest design(N = 107)

36 Frailcommunity-dwellingelderly aged≥65 years.

Primary care by1 geriatricclinic in theUSA

Frail Elderly Community-Based Case ManagementProjectCIS: telemedicine assessment for medicallycomprised patients; documentation of all relevantdata into electronic database which was used tocommunicate patients’ overall status and to keepphysician informed.DS: weekly patient care conferences at geriatricclinic (for case managers, geriatricians, geriatricfellows, support staff, pharmacy and medicalstudents); monthly staff education on patients’needs.DSD: collaboration between primary carephysician, geriatric clinic, assisted livingcommunity staff, nurse and social-work casemanager; case management of medical and socialconditions, utilization of hospice and end-of-lifedecision making; adaptation of interventions topatients’ needs.HS: ongoing program development and evaluationby a steering committee.SMS: monthly education by case manager onspecific care needs/concerns for patients, familycaregivers and assisted living staff.

N.A. Patients(Changes were not tested forsignificance)

There are indications for decline incognitive functioning andimprovement in perception of qualityof life over time.

54% reduction in ER visits over time,resulting in 36% reduction in ER costs.

69% reduction in hospitaladmissions and 64% reduction inhospital length of stay over time,resulting in 60% reduction in hospitalcosts (i.e. $580.189).

Elzen et al. [77] RCT (N = 144) 6 People aged≥59 years,diagnosed withangina pectorisor heart failure,COPD orasthma, orarthritis, ordiabetes, and inaddition other(minor)diseases.

Communitycare providedby hospi-tal/psychologistof theuniversityhospital in theNetherlands

Chronic Disease Self-Management Program(CDSMP)DS: 20 h CDSMP-training of session leaders (pair ofpeer leader or lay leader and health professional);use of standardized leader manual during CDSMPsessions.DSD: session leaders were a trained psychologist(CDSMP Master Trainer) and a second psychologistor peer leader.SMS: 6-session workshop of 2-1/2 h with 10-13participants with mixed diagnoses;sessions focused on problem solving, appropriateexercise and medication, communicationtechniques, nutrition, and how to evaluate newtreatments; activities to enhance self-efficacyincluded weekly action planning, modeling ofbehaviors, group problem solving, and individualdecision-making.

Notspecified.

PatientsHealth behavior (i.e. exercise,

cognitive symptom management),health-related quality of life, andperceived self-efficacy at follow-up didnot differ significantly between groups.

Intervention group was enthusiasticand rated the course with an averageof 8,5 (range 0-10). Majority of thegroup enjoyed the course, indicated itas useful, and was satisfied with coursematerials, preparations, teaching,length of meetings, group size, andmeeting rooms. 25% of the patientsindicated the course as strenuous.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145121




Targetpopulation


Kobb et al. [40] CCT (N = 1401) Ongoing,periodic datacollection at 6month intervals.In this study 12monthsfollow-up.

Veterans withhigh-costmedical careneeds(>$25.000) andhigh use (two ormore hospitaladmissions,frequentemergencyroom visits, andunscheduledwalk-in visits,and 10 or moreprescriptions) inthe yearprecedingenrollment.

Home careprovided byVeterans HealthAdministrationin the USA

Rural Home Care Project (RHCP) (see Chumbler et al.[79])

N.R. Patients(Changes were not tested for significance)

Healthcare utilization (i.e. hospitaladmissions, bed days of care, nursinghome admissions, nursing home bed daysof care, ER visits, clinic visits, pharmacyvisits) at follow-up was lower forintervention group and higher for controlgroup. Health utilization decreased in theintervention groups whereas it increasedin the control group.

Overall, both groups improved in 3aspects of health-related quality of lifeover time (i.e. physical health, mentalhealth, and social functioning) andremained stable in the other aspects(physical role, bodily pain, general health,vitality, emotional role).

Intervention group using Health Buddydevice improved from being notcompliant with immunization usage(hypertensive medication, influenza andpneumococcal vaccines) to 90-92%compliant.

Medication compliance by patients inthe intervention group increased from68% to 93%.

Intervention group reported highlevels of satisfaction with project teammembers and technology and found thetechnology easy to use and helpful inmanaging their chronic conditions.Professional caregivers

Providers reported high satisfactionreflected by the majority of providersvaluing the communication betweentheir selves and care coordinatortimely/appropriate; finding RHCP ofbenefit to their patients; and willingnessto refer patients to RHCP.

122S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Krause et al. [64] Pretest-posttestdesign (N = 39)

12 Patients withmultiple chronicconditions.

Hospital careprovided bydifferenthospitals in theUSA.

Integrated multidisciplinary advocacy programDSD: multidisciplinary team of health advocates (notspecified) to review physical and psychological medicalrecords and performing an extensive multidisciplinary,whole-person intake assessment of patient.SMS: advocate team collaborates with patient to designintegrated plan with goals for improvement inself-efficacy, education and experience in appropriate useof the health care system as well as whole-person, family,financial, and social well-being.

N.A. PatientsRating of effectiveness of healthcare

services under the program in helping themto deal with their conditions and satisfactionwith healthcare services was significantlyhigher at follow-up than at baseline.

Patients reported to exercise significantlymore and to eat significantly healthier atfollow-up than at baseline.

Healthcare costs were significantly lowerthan projected costs (using HCUP-3 data).

Health-related quality of life wassignificantly higher at follow-up than atbaseline.

Participants perceived their physicalwell-being significantly higher at follow-upthan at baseline.

Life satisfaction was significantly higherat follow-up than at baseline.

Locus of control at follow-up did not differsignificantly from locus of control atbaseline.

Self-efficacy regarding patient’s ability tomanage their healthcare needs wassignificantly higher at follow-up than atbaseline.

Leveille et al. [56] RCT (N = 201) 12 Frail older peopleaged ≥70 yearswith ≥1 chronicillness

Primary/community careby a senior centerin the USA.

Health Enhancement Project (HEP)CR: partnership of community senior center with primarycare; referral of patients to senior centre or communityprograms in case of smoking and excessive alcohol use.DSD: collection of information at GP by Geriatric NursePractitioner (GNP) on current health problems of patientsand GP’s goals; reports were sent to GPs to inform themabout participation of patient in program; sessions withpatients with GNP at senior centre; monitoring ofprogress towards health goals by follow-up visits andphone contact; changes in medication were discussedwith GP; referral to specialized programs of caregivers ifnecessary (smoking, alcohol abuse, depression).SMS: design of targeted health management plan by GNPand patient addressing risk factors for disability andself-management issues; individual counseling by GNP inchronic illness self-management; encouragement ofpatients to attend physical activities at senior center or toperform activities at home and to attend chronic illnessself-management course (7 week, 2 h) conducted bytrained lay leaders (combining peer support with healthpromotion information and disease managementconcepts) using a self-management workbook; provisionof nutrition tip sheets developed in consultation withregistered dieticians.

Patients were givena tour of the seniorcenter and a list ofscheduled activities.They did not meetwith GNP but hadaccess to all seniorcentre activities thatwere available tothe interventiongroup.

PatientsChange in ADL significantly differed

between groups, in favor of interventiongroup.

Attitude and behavior related to physicalactivity at follow-up were significantlybetter in intervention than in control group.Alcohol use, smoking status, and nutritionalrisk at follow-up did not differ significantlybetween groups.

Change in depressive symptoms andhealth-related quality of life over time didnot differ significantly between groups.

Significantly fewer bed disability days atfollow-up in intervention than in controlgroup. Restricted activity days did not differsignificantly between groups.

Change in total number of hospital daysdiffered significantly between groups, infavor of intervention group. Change innumber of hospitalized patients, ER visits,and outpatient visits over time did not differsignificantly between groups.

Senior centre participation wassignificantly higher at follow-up inintervention than in control group.

Use of psychoactive drugs at follow-upwas significantly lower in interventiongroup than in control group.

Change in physical performance over timedid not differ significantly between groups.Physical performance measured with PASEat follow-up was significantly higher inintervention than in control group.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145123




Targetpopulation


Liddy et al. [41] Post-test onlystudy (N = 22;purposefulsample of alarger RCT)

Time ofmeasurementwas variableand dependedon periodtelehome careunits were used(9-339 days)

People aged ≥50years at risk forfunctionaldecline orphysicaldeterioration.

Primary care by1 primary carepractice (i.e.family healthnetwork) inCanada

TelehomecareCIS: tele-homecare unit installed at patient’s home;use of remote home-based monitoring equipmentlinked to offices of primary care providers.DSD: individualized care plan collaborativelydeveloped by primary care physician, nursepractitioner, and pharmacist.SMS: training to enter clinical data on vital signs andhealth information into the system, either manuallyor through supplied peripherals.

N.A. PatientsPatients found the technology

user-friendly and useful.Informal caregivers

Informal caregivers found thetechnology user-friendly and useful.Professional caregivers

Healthcare providers were satisfiedwith the technology and foundequipment useful; they thought it mightreduce number of office visits and tracklong-term trends.

Littleford andKralik [57]

Pretest-posttestdesign (pilotstudy) (N = 220)

Variable,depending onadmission toICCOP in the 24month pilotperiod.

Older peopleaged ≥65 yearsor Aboriginalsaged ≥45 yearswith multiplecomplex healthproblems and atrisk of or with ahistory offrequentunplannedhospitaladmissions orER visits.

Integratedcommunity careby a partnershipbetween acommunitynursingorganization, GPpractice, andhospital inAustralia

Integrated Community Care for Older People (ICCOP)CIS: assessments and care plans are hosted on secureICCOP website in individualized client e-healthrecords which are accessible for GP, geriatrician, EDmedical staff, and ICCOP nurse.CR: referral to appropriate services of communityservice system to meet complex needs of patients.DS: collaboration and sharing of knowledge betweeninvolved disciplines to manage patient.DSD: referral of patients to ICCOP by geriatrician orGP; transfer of referral to ICCOP nurse who contactspatient for geriatric assessment; design care plan(including health goals, actions to be taken) by ICCOPnurse based on geriatric assessment; regular updateof care plan based on events and contacts; admissionand care plan discussed with patient, family, GP,practice nurse and geriatrician; cooperation betweenhealth and care professionals, patients and services;coordination of provision of care and funding for careby ICCOP nurse; involvement of ICCOP nurse indischarge planning.SMS: cooperation of ICCOP nurse with patients toestablish desired health goals.

N.A. PatientsER visits reduced with 51% during the

study compared with the number of visitsbefore enrollment. Hospital length of stayreduced with 37% during the studycompared with the total number ofhospital days before enrollment.

124S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Markle-Reidet al. [58]

RCT (N = 288) 6 Frail olderpeople aged ≥75years eligible forhome careservices.

Home care byhome careorganization inCanada

In addition to usual home care, proactive nursinghealth promotion interventionCR: referral to and coordination of communityservices.DSD: health assessment to support personalresources; identifying and managing risk factors forfunctional decline; regular home visits or telephonecontacts by registered nurse from acommunity-nursing agency; providing caregiversupport.SMS: health education by registered nurse aboutmanagement of illness and healthy lifestyles usingempowerment strategies.

Usual home careservices consistingof casemanagement,personal care,home support(home making),nursing,occupationaltherapy,physiotherapy,social work, andspeech languagetherapy throughcommunity-basedagencies.

PatientsChange over time in coping styles did

not differ significantly between groups.Change over time in depressive

symptoms differed significantly betweengroups, in favor of the intervention groupthat had a significantly greater reductionin depressive symptoms.

No significant difference betweengroups during the study in mean costs ofall types of health and social services, andthe total annual per person direct costs ofhealth services. Significantly lower perperson cost of prescription medicationsin intervention than in control group.

Change over time in physical function,role functioning related to physicalhealth, bodily pain, general healthperception, vitality, social functioning,and physical health component summaryscore did not differ significantly betweengroups.

Change over time in role functioningrelated to emotional health, mentalhealth, and mental health componentsummary score differed significantlybetween groups, in favor of theintervention group that significantlyimproved on these aspects.

Change over time in perceived socialsupport differed significantly betweengroups, in favor of the intervention groupthat experienced a significant increase ofsocial support.

Marsteller et al.[59]

cRCT (N = 49) 12 Primary carephysiciansproviding careto multimorbidpeople aged ≥65years and athigh risk forincurring highhealthcare costsin the comingyear.


Guided Care (see Boult et al. [45]) Not specified Professional caregiversKnowledge of the clinical

characteristics of patients and satisfactionwith patient/family communication atfollow-up were significantly higher inphysicians providing guided care than inphysicians providing usual care.

No significant differences betweenphysicians from both groups at follow-upwith regard to satisfaction with caremanagement, time spent on chronic care,knowledge of patients’ personalcircumstances, and care coordination.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145125




Targetpopulation


Mitton et al. [60] Pretest-posttestdesign (N = 37)

12 Older patientswith highhealthcareneeds andchronic orcomplexmedicalproblems.

Primary care bycollaborativepartnershipbetween homecareorganizationand primarycare practice inCanada.

Nurse Physician Collaborative PartnershipCIS: electronic patient health record system in whichnurses could access patients’ records andcommunicate with physicians by e-mail.CR: linkage to community supports (no further detailsavailable).DS: shared care guidelines; professional courses forproject nurses with knowledge and skill gaps;consultation with partners and other healthcareproviders (e.g. pharmacist); discussion between nurseand physician at monthly care planning meeting.DSD: collaborative partnership between home carenurses and family physicians, services includedreferral and admissions, development, andmodification of shared care plan for each patient;comprehensive biopsychological assessment; activepatient management; standard and advanced nursinginterventions.SMS: health education and self management.

N.A. PatientsADL, cognitive functioning, and

health-related quality of life at follow-updid not differ significantly from ADL,cognitive functioning, and health-relatedquality of life at baseline.

Total healthcare costs significantlydecreased during the study, mainly due tosignificantly less inpatient costs than inthe year preceding the study. Otherhealthcare costs did not changesignificantly during the study. However,when program costs were included, totalhealthcare costs did not differsignificantly with the year preceding thestudy, although they were $2463 lowerper patient.

Healthcare utilization decreased duringthe study: 25% reduction in hospitaladmissions, 50% reduction in days spentin hospital, 32% reduction in ER visits, 28%reduction in visits for diagnostic tests,15% reduction in ambulatory care visits.

Patients reported periods of improvedhealth and felt well cared for. Theyparticularly appreciated nurse-physicianpartnership, the holistic approach, theearly intervention, technologicalcapabilities, and connections to otherhealthcare providers.Informal caregivers (N = not reported)

Caregivers appreciated the improvedaccess and quality of services. Theyreported decreased anxiety and improvedability to cope (not tested for statisticalsignificance).Professional caregivers (N = not reported)(Changes not tested for significance)

Small changes in job satisfaction ofphysicians and nurses at follow-upcompared to baseline. Partnershipimproved accessibility and availability,enabled continuity of care, and enabledeffective case management.Pharmaceutical knowledge andinterpretation of laboratory tests werethe 2 major areas of additional learningfor nurses.

126S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Naylor et al. [74] RCT (N = 239) 12 Patients aged 65or older andhospitalizedwith heartfailure.

Hospital care by6 hospitals inthe USA

Transitional care interventionDS: standardized orientation and training programguided by a multidisciplinary team of heart failureexperts to prepare the advanced practice nurse (APN)to address the unique needs of older adults and theircaregivers throughout an acute episode of heartfailure; implementation of an evidence-basedprotocol designed for patients with an acute episodeof heart failure complicated by multiple co-morbidconditions.DSD: 3-month APN-directed discharge planning andhome follow-up protocol; use of care managementstrategies including identification of patient’s andcaregiver’s goals, individualized care plans designedand implemented by APN in collaboration withpatient’s physicians; use of expert nurses to deliverand manage clinical services to high risk patients;continuity of care and care coordination acrosssettings; initial APN visit within 24 hours of indexhospitalization; daily visits during hospitalization andat least 8 visits at home after discharge (first weeklythen biweekly during second and third month) formonitoring and in home assessments; 24/7availability of APN by phone.SMS: APN performs educational and behavioralstrategies to address patients’ and caregivers’ learningneeds

Routine careprovided by theadmitting hospital,includingsite-specific heartfailure patientmanagement anddischarge planningcritical paths, andif referred,standard homecare agency care.The dischargeplanning wassimilar in allhospitals, alsohome care wassimilar. Home careincluded use ofliaison nurses,availability ofskilled home careservices, and nurse24 a day hours oncall.

PatientsADL, health-related quality of life, and

mortality at follow-up did not differsignificantly between groups.

Healthcare costs during the study weresignificantly lower in intervention than incontrol group. Incremental healthcarecosts were −$5708 per patient per year.

Number of rehospitalizations duringthe study was significantly lower inintervention than in control group. Timeto first readmission was significantlylonger in intervention than in controlgroup. Number of patients requiringreadmission and number of hospital daysduring the study did not differsignificantly between groups.

Patient satisfaction at 6 weeksfollow-up was significantly higher inintervention group than control group.

Procter andSingle [42]

Pretest-posttestdesign (N = 25)

4 Older peopleaged ≥65 years,diagnosed withheart failure orCOPD and earlystage dementia(combinationswith otherchronicconditions wereallowed), livingat home, andexperiencingrepeatedhospitaladmissions.

Primary care inthe UK, settingunknown.

Home Telehealth careCIS: installation tele-homecare unit at patient’s homefor biometric monitoring; observations required byclinician were discussed with patient, tele-homecareunit was programmed to prompt patient to undertakethese observations on daily basis; results weredownloaded to a central web-based server which wasaccessed daily by a nurse; tele-monitoring equipmentwas programmed to ask patients a series of questionsabout their health; equipment could be programmedto remind patients to take their medications and toattend appointments; system could be adapted fornew observations.CR: liaison with other providers.DSD: project nurse was appointed and was responsiblefor assessing patient, developing care plan, installingtele-monitoring equipment, and daily monitoring; incase daily monitoring identified problems, nursecontacted patient (telephone or home visit) orarranged GP visit; nurse was first contact for anytechnical difficulties experienced by patients; nursenegotiated with patient and healthcare providers andwas a liaison for other healthcare providers.SMS: patients enter and upload measurement results(of vital signs) daily; reminders on medication andappointments (hospital and GP).

N.A. Patients(Changes were not tested for significance)

Patients reported increased confidencein managing diseases at follow-up.

Number of hospitalizations decreasedwith 38% during the study.Informal caregivers

Informal caregivers reported increasedconfidence in managing diseases atfollow-up.Professional caregivers

In 14 out of 25 patients early detectionand treatment of problems occurredduring the study.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145127




Targetpopulation


Rose et al. [78] Pretest-posttestdesign (N = 175)

6 Low incomeolder urbanAfricanAmerican adultswith ≥1 chroniccondition.

Community careprovided bycommunity sites(i.e. seniorcitizen centers,senior housingfor low incomeolder people,churches) in theUSA.

Chronic Disease Self-Management Program (CDSMP)(see Elzen et al. [77]DSD: in addition to session leaders reported by Elzenet al. [77] co-leaders were lay African American olderadults, who were part of the community.SMS: workshops consisted of 10-20 participantsinstead of 10-13 participants as reported by Elzenet al. [77].

N.A. PatientsHealth behavior regarding stretching

or strengthening exercise andcommunication with GP was significantlyimproved at follow-up. Health behaviorregarding cognitive symptommanagement and aerobic exercise wasnot changed significantly at follow-up.

Health-related quality of life, impact ofdisease on role and social activities,perceived self-efficacy, and role functionwere not changed significantly atfollow-up, with the exception of energythat was significantly improved.

Healthcare utilization (i.e. GP visits, ERvisits, hospital stays, nights in hospital)was not changed significantly atfollow-up.

Rubenstein et al.[61]

RCT (N = 792) 36 Community-dwellingveterans aged≥65 years withat least 1 clinicvisit in the last18 months.

Primary careprovided byDepartment ofVeterans affairsambulatory carecentre in theUSA.

Managed care plus screening, case finding, andreferral system for older veterans in primary careDS: supervision of case manager by studygeriatricians; structured format to validate and clarifypositive findings from structured geriatric telephoneassessment of patient.DSD: structured geriatric telephone assessment ofspecific risks and unmet needs by physician assistantcase manager; specific referrals andrecommendations by case manager to further assess,treat, and manage unmet needs; coordinatedoutpatient follow up by case manager; referral ofpatients to either geriatric assessment clinic,home-based primary care program or primary careprovider or other specific services; provision ofoverview of recommendations, referrals, andscheduled appointments to patient by case manager;on-going telephone case management (every 3months) to encourage patients to adhere to referralsand recommendations and to monitor changes inhealth.SMS: health promotion recommendations and healtheducation by case manager to patient.

Managed carewithout screening,case finding andreferral system.

PatientsChange over time in (I)ADL, depressive

symptoms, fall incidence, health-relatedquality of life, and urinary incontinencedid not differ significantly betweengroups.

Healthcare utilization (i.e. number ofhospital days, number of patientsadmitted to hospital) during the study didnot differ significantly between groups.

128S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Saltvedt et al.[68,69]

RCT (N = 254) 12 Acutely sick,frail olderpatients aged≥75 years whowere acutelyadmittedhospital.

Hospital care by1 hospital inNorway.

Geriatric Evaluation Management Unit (GEMU)DS: Meetings of all disciplines (twice a week) toreport assessments, discuss problems, set goals, andplan discharge. Primary health care nurses weretelephoned to discuss arrangements of discharge,when considered appropriate by the staffDSD: interdisciplinary approach with closecollaboration between all disciplines; designindividual rehabilitation plans for each patient.

Usual treatment ofDepartment ofInternal Medicine

Patients(I)ADL, general well-being,

depression, and cognitive functioningat follow-up did not differsignificantly between groups.

% of patients deceased at follow-updid not differ significantly betweengroups. Heart disease was the majorcause of death in both groups. Duringthe study infections were morefrequently the cause of death in theintervention than in the controlgroup.

Schmader et al.[70]

RCT (N = 864) Inpatientperiod: variableOutpatientperiod: 12months minusinpatient period.

Hospitalizedfrail veteransaged ≥65 yearadmitted to amedical orsurgical wardfor ≥3 days.

Hospital care by11 VeteransAffairs medicalcenters in theUSA.

Geriatric Evaluation and Management (GEM)In addition to GEM described by Burns et al. (1995):DS: evaluation and management by team membersaccording to published guidelines and VeteransAffairs’ standards for patients.DSD: team consisting of geriatrician, social worker,and nurse; implementation of evaluation andmanagement protocols for GEM patients by team;regular assessments and recommendations regardingmedications by pharmacists.

Usual inpatientcare: customarymedical or surgicaltreatment byphysicians (orothers under theirdirection).Usual outpatientcare: customarycare delivered byambulatory carephysicians (orothers under theirdirection).

PatientsInpatient period

Total number of detected adversedrug reactions during the study wassignificantly higher in interventionthan in control group.

Change in number of unnecessarydrugs, in medication appropriateness,in number of inappropriate drugs,and in number of conditions withomitted drugs during the study wassignificantly larger in interventionthan in control group, all in favor ofthe intervention group.Outpatient period

Total number of detected adversedrug reactions during the study didnot differ significantly betweengroups. Number of detected seriousadverse drug reactions wassignificantly higher in interventionthan in control group.

Change in number of unnecessarydrugs, medication appropriateness,and number of inappropriate drugsduring the study did not differsignificantly between groups.

Change in number of conditionswith omitted drugs during the studywas significantly smaller inintervention than in control group, infavor of the intervention group.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145129


Author (year) Study design (N)a Length offollow-up(months)b

Target population Setting Contents of comprehensive care programc Usual cared Outcomese

Sorrento et al.[62]

Pretest-posttestdesign (N = 273)

N.R. Community-dwelling olderpeople aged ≥75years with ≥2chronic illnessesand with ≥1 ofthe followingproblems:memory deficits,history of falls,incontinence,depression,anxiety, and/ormalnutrition.

Primary careprovided bypartnership ofphysicians andhospitals in theUSA.

Geriatric Assessment Service (GAS)DS: weekly team meeting discussing all enrolled patientsand determining the disciplines initiating the evaluation;use of standardized instrument to evaluate aspects ofpatients’ functioning, impairments and social support.DSD: multidimensional evaluations of patients byinterdisciplinary team (consultant pharmacist, socialworker, registered nurse); home evaluations by nurseand/or social worker (required assessments variedaccording to needs of patients); involvement ofconsultant pharmacist in screening patients identified asat-risk by primary care provider and in evaluatingappropriateness of medication regimens; results of allassessments, goals, recommendations and plans werewritten in a summary letter that was sent to GP and whodecided which actions should be undertaken.SMS: family/caregiver education and recommendationsto improve medication management by pharmacist.

N.A. Patients(Changes were not tested for significance).

On average 62% of patients reported animprovement in medication management atfollow-up.

On average 90% of the patients was satisfiedwith the program.Informal caregivers

Informal caregivers (n = 13) reported onaverage a 28% improvement in caregiverburden at follow-up.

Taylor et al. [76] RCT (N = 169) 12 Patients withlongstandingdiabetes with oneor more majormedicalco-morbidconditions(hypertension,dyslipidemia orCVD) and HbA1c>10%.

Primary careprovided bymanaged careorganization inthe USA.

Nurse Care Management SystemDS: treatment algorithms based on national guidelinesused by nurse care-managers use; education fornurse-care managers on diabetes, cholesterol,hypertension, and depression.DSD: initial individual meeting with a registered nurse toassess patients health status and needs; telephonefollow-up calls by program nurse manager; primary carephysician was called if a new medication was indicated orto report any unusual findings.SMS: development initial self-management plan; groupclasses once a week for 4 weeks, with focus on groupdiscussion, participation, and problem-solving; about 8telephone follow-up calls to review patients’ goals andmedical outcomes.

Usual medical care,patients wereinstructed andencouraged toremain under careof their physician,got a folder withinformation ondiabetes and aMedical Alertpamphlet, and wereencouraged toattend generaldiabetes education.

PatientsDepression improved significantly over time

in both groups. Change over time in depressionand health-related quality of life did not differsignificantly between groups.

Healthcare utilization (i.e. GP visits, ER visits,hospitalizations) during the study did notchange significantly within both groups anddid not differ significantly between groups.

Changes over time in HbA1c, total, and LDLcholesterol differed significantly betweengroups, in favor of intervention group thatimproved significantly on these outcomeswhereas control group did not changesignificantly over time. Changes in fastingglucose, systolic and diastolic blood pressure,and BMI were not significant within bothgroups and did not differ significantly betweengroups.

Significantly larger % of patients in theintervention group than in the control groupmet the goals for HbA1c (i.e. HbA1c <7.5%) atfollow-up. % of patients meeting the otheroutcome goals did not differ between groups.

90% of patients indicated at follow-up thatthey found the program moderately toextremely helpful. 92% of patients indicated atfollow-up that the program was moderately toextremely helpful in preparing them toself-manage their conditions.

Patients reported high levels of confidence toengage in the behaviors necessary to managetheir diabetes. Confidence increased in bothgroups over time.

130S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Professional caregivers (N = 13)69% of caregivers indicated that they

strongly recommended adoption of theprogram by their healthcare system, thatthey felt that the program decreased timespent with patients. 31% indicated thatthey felt that the program increased thetime spent with patients.

Tibaldi et al. [71] RCT (N = 109) Variable(dependingon time todischarge)

Frail olderpatients withdementiarequiringadmission to ahospitalemergencydepartment foracute illness.

Hospital care(provided athome) by 1hospital in Italy

Geriatric Home Hospitalization Service (GHHS)DSD: GHHS team (5 geriatricians, 14 nurses, 4physiotherapists, 1 dietician, 1 social worker, 1counselor) providing physician and nursing care,medicines, and appropriate diagnostic and therapeuticinterventions at home; service operates 12 h a day 7days a week; GHHS doctors and nurses are availablefor patients at all times.SMS: caregivers are informed about the emergencyplan and encouraged to telephone if problems orquestions arise.

Admission togeneral medicalward.

PatientsSignificantly lower % of patients with

behavioral disturbances (i.e. sleepingdisorders, agitation, feeding disorders)and using antipsychotic drugs atdischarge in intervention than in controlgroup.

No significant difference in mortalityduring the study between groups.Informal caregivers

Change in caregiver stress over timewas significantly larger in interventionthan in control group, in favor of theintervention group where caregiver stressreduced significantly.

Wright et al.[11]

Post-test onlystudy (withoutcontrol) (pilot)(N = 118)

Time ofmeasure-ment was12 monthsafter startof AD-LIFE

Low-incomeolder patientswith chronicconditions andfunctionalimpairment athigh risk for re-hospitalizationor nursing homeplacement.

Primary care bynon-profithealthcaredeliverynetwork in theUSA.

After Discharge Care Management of Low-Income FrailElderly (AD-LIFE) (integrated medical and social care)DS: use of evidence-based protocols for designing careplan; education of primary care physicians regardingcare management process.DSD: advanced practice nurse assesses eligiblepatients for program at admission and assist dischargeplanning; assessment findings are shared withhospital-based interdisciplinary team that includespatient, geriatrician, registered nurse (RN) caremanager, pharmacist, social worker, and possibleother healthcare providers; team generates care plan;RN care manager implements care plan afterdischarge in collaboration with patient’s primary carephysician and provides ongoing follow-up evaluation;each primary care practice gets assigned own RN caremanager; reimbursement for primary care physiciansto review care plans and receive updates; follow-upby RN care manager to check whether patientimplements recommendations.

N.A. PatientsNumber of hospital admissions per

1000 patients decreased resulting indecreased healthcare costs of $12260 perpatient per year during the study (nottested for significance).

About 70% of patients indicated atfollow-up that the program had improvedtheir health, had made it easier to gethealthcare services, had them providedwith a better understanding of theirdisease, and had not experienced asubjective decline in health during theyear. 93% of patients rated theirexperience as “good” or “excellent” and92% of patients would recommendprogram to their friends.Professional caregivers

Physicians indicated to be highlysatisfied at follow-up.

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145131




Targetpopulation


Zhang et al. [75] CCT (N = 65132) 24 Medicaidpatients havingone of five of thefollowingdiseases:diabetes,hyperten-sion/congestiveheart failure,depression,gastro-esophagal refluxdisease/pepticulcer diseaseandAsthma/COPDand/orco-morbidities.

Primary and/orpharmaceuticalcare by primarycare practicesand/orpharmacies inthe USA

Disease State Management (DSM) programI1: physicians intervenedI2: both physicians and pharmacists intervenedCIS: Mailing with claims-data based routine feedbacksheet and clinical summary with latest results onoutcome measures.DS: Mailing of educational materials on the fivediseases and their co-morbidities on quarterly basisfor 3 years to participating physicians andpharmacists.SMS: physicians and pharmacists were asked toconsult with their patients and/or patientrepresentatives about their life styles, treatment, anddrug use.

Nointerventions

PatientsPatients with co-morbidity

Change in ER visits andhospitalizations over time differedsignificantly between intervention groupsand control group, in favor of the controlgroup. Change in GP visits did not differsignificantly between intervention group1 and control group and differedsignificantly between intervention group2 and control group, in favor of controlgroup and between intervention group 1and 2, in favor of intervention group 1.All patients

Overall, health-related quality of lifedeteriorated over time. Health-relatedquality of life reduced significantly slowerover time in the intervention groups thanin the control group for 9 of 11 subscales(mental, social, general, and perceivedhealth, self-esteem, anxiety, depression,anxiety/depression, and disability).

Average cost savings perhospitalization for intervention groupwere $50. It was estimated that theprogram would save approximately $1.8million/year during the study, and wouldachieve a $2.0 return for each dollarinvested.

a CCT = controlled clinical trial; cRCT = cluster-randomized controlled trial; RCT = randomized controlled trial; N = the number of patients/caregivers allocated at study entry.b N.A. = not applicable.c CIS = clinical information system; CR = community resources; DS = decision support; DSD = delivery system design; HS = health system; SMS = self management support.d N.A. = not applicable; N.R. = not reported.e ADL = activities of daily living; ER = emergency room; FIM = Functional Independence Measure; GP = general practitioner; HbA1c = glycated hemoglobin; HCUP = Healthcare Cost and Utilization Project; IADL = instrumental

activities of daily living; LDL = low-density lipoprotein; MMSE = Mini Mental State Examination; PASE = Physical Activity Scale for the Elderly; PQI hospitalization = hospitalization for one of the prevention quality indicators;SF-36 = Medical Outcomes 36-item short-form.

1 lth Polic

3

c(u(tmi

3

3

hwAdtpwdhmaaocdff

3

watA(apsttiilctonDAI

3

pdto


.2.3. Usual care conditionsIn most studies (n = 23) the effects of comprehensive

are programs were compared with those of care as usuali.e. no comprehensive care). In twelve of these studies, thesual care conditions were not or only poorly describedTable 1). Usual care was mostly described as normal accesso services available to frail older people, routine home, pri-

ary and hospital care, and/or encouragement to remainn care of their current physician.

.3. Program characteristics

.3.1. Target populationsThe 33 different studies evaluated 28 different compre-

ensive care programs. The comprehensive care programsere implemented in the USA (n = 17), Canada (n = 4),ustralia (n = 3), Italy (n = 1), Norway, (n = 1), United King-om (n = 1), and one program was implemented in bothhe Netherlands and the USA. Fifteen comprehensive carerograms focused on frail people aged 65 years and olderith multiple, not specified chronic conditions, functionalisabilities, complex healthcare needs, and/or high use ofealthcare services [11,43–64]. Seven programs focused onultimorbid older people who were (acutely) hospitalized

nd who sometimes had a history of readmissions or weret risk for readmissions [65–74]. Five programs focusedn older people with (a combination of) specific chroniconditions such as diabetes mellitus, heart failure, COPD,ementia and/or arthritis [40,42,75–79], and one programocused on people aged 50 years or older who were at risk ofunctional decline or physical deterioration [41] (Table 1).

.3.2. SettingsThe settings in which the comprehensive care programs

ere implemented varied from home care organizationsnd community centers to primary care practices, hospi-als, specialized clinics (e.g. geriatric clinics and Veteransffairs medical centers), and managed care organizations

Table 1). Consequently, the type of care that was providedlso differed widely between the programs. Various com-rehensive care programs incorporated a mix of care types,uch as the Integrated Services for Frail Elders (SIPA) andhe Nurse Physician Collaborative Partnership in Canada,he Integrated Community Care for Older People (ICCOP)n Australia, and the Health Enhancement Project (HEP)n the USA. These programs mostly included newly estab-ished partnerships between community centers, homeare organizations, primary care practices, and/or hospi-als [43,56,57,60]. Other programs enabled the provisionf regular care services in settings in which they were notormally provided, such as the Older Hospitalized Patients’ischarge Planning and In-home Follow-up Protocol inustralia and the Geriatric Home Hospitalization Service in

taly both providing hospital care services at home [67,71].

.3.3. ContentsTable 1 presents the contents of the comprehensive care

rograms included in our review. The table illustrates theiversity in the comprehensive care programs with regardo the number of included interventions and the numberf related CCM components. We observed comprehensive

y 107 (2012) 108– 145

care programs that included interventions related to twoCCM components (n = 5), three CCM components (n = 12),four CCM components (n = 3), five CCM components (n = 6),and six CCM components (n = 2).

Except for one program [75], all comprehensive careprograms included interventions related to the CCM com-ponent delivery system design. Examples are employing acase manager, working in multidisciplinary teams, design-ing individualized care plans, and making home visitsto patients. The majority of comprehensive care pro-grams (n = 25) additionally included interventions relatedto the CCM component self-management support, suchas involving patients in the design of their care plan,educating patients to improve self-management, and stim-ulating patients to participate in self-management groupsessions. Twenty comprehensive care programs includedinterventions related to the CCM component decisionsupport. Interventions to support healthcare providers’decision-making included implementing evidence-basedguidelines, training teams to implement new protocols,and supervising newly appointed case managers. Interven-tions related to the CCM components clinical informationsystems (e.g. feedback sheets, electronic patient records,and tele-homecare units) and community resources (e.g.establishing access to community resources and part-nerships with local community service centers) wereless frequently observed (n = 12 for both components).A small minority of the comprehensive care programs(n = 4) included interventions related to the CCM com-ponent health system. Examples of such interventionsare installing committees to support new partnerships,employing management teams to support process andquality improvement, and enabling infrastructure for inno-vations in chronic care.

3.4. Impact of comprehensive care programs

Table 1 presents all outcomes reported in the studiesincluded. Only outcomes that were reported in at least fivestudies are described in the text. More detailed informationabout the actual differences between comprehensive caregroups and usual care groups or about actual changes overtime for clinical outcomes are provided in Tables 3a and 3b,respectively. Quality of reporting on baseline and follow-up scores on the clinical patient outcomes widely variedacross studies.

3.4.1. Patient related outcomes—clinical outcomes3.4.1.1. Cognitive functioning. Five studies evaluated theeffect of comprehensive care on cognitive functioning[55,60,66,68,79] (Table 3a), of which almost all had rela-tively low quality sum-scores (Table 2). All studies used theMini Mental State Examination [80] to measure cognitivefunctioning. Three studies compared cognitive function-ing between patients receiving comprehensive care andpatients receiving usual care [66,68,79]. None reported asignificant difference between both patient groups. One
study, however, measured cognitive functioning also withthe Functional Independence Measure [81] and showedthat cognitive functioning significantly improved over timein the comprehensive care group, whereas it significantly

S.R. de Bruin et al. / Health Policy 107 (2012) 108– 145 133

Table 2Results of methodological quality assessment of included studies.

Study Randomizationa Similar atbaselineb

Compliancec Drop-outrated

ITT-analysise Adjustments forconfounding variablesin analysisf

Totalscore

Anonymous [63] N.A. N.A. ? N.A. N.A. N.A. 0Béland et al. [43] + + ? + + + 5Boult et al., Boyd et al.,

Leff et al., and Wolffet al. [44–47]

+ − ? + + + 4

Boyd et al. and Sylviaet al. [48,49]

N.A. − ? + + ? 2

Brand et al. [65] N.A. − ? + N.A. + 2Burns et al. [66] + + ? ? − N.A. 2Chumbler et al. [79] N.A. − ? + N.A. + 2Coleman et al. [50] ? + ? + + + 4Counsell et al. [51,52] + + ? + + N.A. 4Courtney et al. [67] + + ? + + N.A. 4Daly et al. and Douglas

et al. [72,73]+ + ? ? − N.A. 2

Dorr et al. [53,54] N.A. + ? + N.A. N.A. 2Duke et al. [55] N.A. N.A. ? + N.A. N.A. 1Elzen et al. [77] + + ? + − N.A. 3Kobb et al. [40] N.A. ? + + N.A. ? 2Krause et al. [64] N.A. N.A. ? + N.A. N.A. 1Leveille et al. [56] + − ? + + + 4Liddy et al. [41] N.A. N.A. + N.A. N.A. N.A. 1Littleford and Kralik [57] N.A. N.A. ? + N.A. N.A. 1Markle – Reid et al. [58] + + ? + + N.A. 4Marsteller et al. [59] ? + ? + − ? 2Mitton et al. [60] N.A. N.A. ? + N.A. N.A. 1Naylor et al. [74] + + ? + + N.A. 4Procter and Single [42] N.A. N.A. − ? N.A. N.A. 0Rose et al. [78] N.A. N.A. ? − N.A. N.A. 0Rubenstein et al. [61] ? + ? + + ? 3Saltvedt et al. [68,69] + + ? + − N.A. 3Schmader et al. [70] + + ? ? − N.A. 2Sorrento et al. [62] N.A. N.A. ? ? N.A. N.A. 0Taylor et al. [76] ? + ? + − N.A. 2Tibaldi et al. [71] ? ? ? ? + ? 1Wright et al. [11] N.A. N.A. ? N.A. N.A. N.A. 0Zhang et al. [75] N.A. ? ? ? N.A. N.A. 0

N.A. = not applicable; + = criterion fulfilled; − = criterion not fulfilled; ? = lack of information.a Was the method of randomization adequate?b Were the groups similar at baseline regarding the most important prognostic indicators?c Was the compliance of patients acceptable in all groups?

n progn

d Was the drop-out rate described and acceptable?e Did the analysis include an intention-to-treat analysis.f Are adjustments made for confounding variables and/or differences i

deteriorated in the usual care group [79]. Two studies didnot compare comprehensive care with usual care [55,60].Mitton et al. [60] observed no significant change in cog-nitive functioning over time, whereas Duke et al. [55]suggested a deterioration of cognitive functioning overtime. Considering these findings, there is no evidence for abeneficial effect of comprehensive care on cognitive func-tioning.

3.4.1.2. Depressive symptoms. Seven studies measured theeffect of comprehensive care on depressive symptomsof which three had relatively higher quality sum-scores [50,56,58]. Five studies reported no differencesbetween comprehensive care and usual care groups
[50,56,61,68,76]. Two studies, however, reported a signif-icantly larger reduction of depressive symptoms in thecomprehensive care group than in the usual care group[58,66]. Considering these findings, there is no evidence
ostic indicators at baseline?

for a beneficial effect of comprehensive care on depressivesymptoms.

3.4.1.3. Functional status. Nine studies evaluated theeffects of comprehensive care on activities of daily liv-ing (ADL), of which three had relatively higher qualitysum-scores [51,56,74]. Six studies reported no differ-ences between comprehensive care and usual care groups[51,61,66,68,73,74]. Two studies reported a significantlylarger improvement of ADL in the comprehensive caregroup compared to the usual care group [56,79]. Mittonet al. [60] presented the results of comprehensive care only,showing no effects on ADL.

The effect of comprehensive care on instrumental activ-
ities of daily living (IADL) was evaluated in six studies, ofwhich one had a relatively higher quality sum-score [51].Five studies observed no effect of comprehensive care onIADL [51,61,66,68,73] while one study observed significant

134S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Table 3aClinical patient outcomes (mean (SD)) of studies evaluating impact of comprehensive care programs for multimorbid patients.*

Study Study design (N)a Measuring instrumentb Comprehensive care groupc Usual care group p

Baseline Follow-up Baseline Follow-up

Cognitive functioningBurns et al. [66] RCT (N = 130) MMSE 24.6 (3.6) 25.8 (3.2) 24.5 (3.1) 24.7 (3.8) >0.05, for group difference

<0.01, for time change>0.05, for interaction effect

Chumbler et al. [79] Case control design inwhich subjectsreceivingcomprehensive carewere compared withsubjects from anotherlongitudinal study(N = 226)

MMSE 20.8 (6.8) 25.8 (4.7) 26.2 (6.4) 28.2 (6.4) 0.59, for group differenceFIM cognitive subscale 29.2 (5.2) 31.8 (5.3) 31.3 (6.8) 31.0 (6.7) <0.0001, idem

Duke et al. [55] Pretest-posttest design(N = 107)

MMSE N.R. N.R. – – N.R.

Mitton et al. [60] Pretest-posttest design(N = 37)

MMSE N.R. N.R. – – 0.32, for time change

Saltvedt et al. [68,69] RCT (N = 254) MMSE ∼25.0 (N.R.) ∼25.0 (N.R.) ∼25.1 (N.R.) ∼24.9 (N.R) >0.05, for group difference

Depressive symptomsBurns et al. [66] RCT (N = 130) CES-D 10.9 (11.0) 3.9 (5.2) 11.7 (10.6) 7.6 (7.1) <0.01, for group difference

<0.001, for time change<0.05, for interaction effect

Coleman et al. [50] RCT (N = 169) CES-D 11.4 (N.R.) 14.8 (N.R.) 15.9 (N.R.) 12.4 (N.R.) 0.19, for group differenceLeveille et al. [56] RCT (N = 201) CES-D 10.1 (8.0) N.R. 8.7 (7.3) N.R. 0.526, for group differenceMarkle – Reid et al. [58] RCT (N = 288) CES-D 15.8 (10.0) 11.9 (10.2) 12.9 (8.5) 11.8 (10.7) 0.022, for group differenceRubenstein et al. [61] RCT (N = 792) GDS-15 5.0 (3.6) 3.9 (3.2) 5.2 (3.7) 3.4 (3.4) >0.05, for group differenceSaltvedt et al. [68,69] RCT (N = 254) MADRS N.R. N.R. N.R. N.R. >0.05, for group differenceTaylor et al. [76] RCT (N = 169) BDI N.R. N.R. N.R. N.R. N.R.

Functional statusBurns et al. [66] RCT (N = 130) Katz ADL 0.3 (0.8) 0.2 (0.7) 0.5 (1.1) 0.4 (1.0) >0.05, for group difference

<0.01, for time change>0.05, for interaction effect

IADL 11.6 (5.4) 11.9 (6.2) 10.9 (4.1) 12.7 (6.1) >0.05, for group difference<0.01, for change over time>0.05 for interaction effect

Chumbler et al. [79] Case control design inwhich subjectsreceivingcomprehensive carewere compared withsubjects from anotherlongitudinal study(N = 226)

FIM motor subscaleIADL

85.3 (21.9)9.0 (4.3)

99.2 (15.2)11.4 (3.1)

73.9 (14.2)8.6 (4.1)

73.3 (14.5)8.3 (4.2)

<0.0001, for group difference<0.0001, idem

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145135

Table 3a (Continued )



Counsell et al. [51,52] cRCT (N = 951) AHEAD (ADL items) 1.6 (3.1) N.R. 1.3 (2.6) N.R. 0.37, for group differenceAHEAD (IADL items) 2.7 (4.2) N.R. 2.5 (3.9) N.R. 0.77, idem

Douglas et al. [73] RCT (N = 334) OASIS N.R. 24.5 (11.6) N.R. 22.5 (13.5) 0.33, for group differenceLeveille et al. [56] RCT (N = 201) HAQ 0.24 (0.32) N.R. 0.23 (0.34) N.R. 0.014, for group differenceMitton et al. [60] Pretest-posttest design

(N = 37)Barthel index N.R. N.R. – – N.R.

Naylor et al. [74] RCT (N = 239) ESDS 3.3 (1.1) 3.1 (1.5) 3.3 (1.1) 2.9 (1.6) >0.10, for group difference

Rubenstein et al. [61] RCT (N = 792) FSQ (ADL items) 84.1 (19.6) 82.4 (22.1) 82.8 (20.4) 85.2 (20.6) >0.05, for group differenceFSQ (IADL items) 53.9 (26.3) 56.5 (32.9) 53.4 (28.6) 58.2 (31.7) Idem

Saltvedt et al. [68,69] RCT (N = 254) Barthel index ∼15.5 (N.R.) ∼15.0 (N.R.) ∼16.0 (N.R.) ∼15.0 (N.R.) >0.05, for group differenceIADL ∼4.0 (N.R.) ∼4.2 (N.R.) ∼4.2 (N.R.) 4.1 (N.R.) >0.05, idem

Health behaviorElzen et al. [77] RCT (N = 144) Exercise (Lorig) N.R. N.R. N.R. N.R. 0.47, for group difference

Cognitive symptommanagement (Lorig)

N.R. N.R. N.R. N.R. 0.14, idem

Communication withphysician (Lorig)

N.R. N.R. N.R. N.R. 0.48, idem

Kobb et al. [40] CCT (N = 1401) Immunizationcompliance

N.R. N.R. N.R. N.R. Not tested for significance

Medication compliance N.R. N.R. N.R. N.R. Not tested for significance

Krause et al. [64] Pretest-posttestdesign (N = 39)

PPCQ (exercise) 2.49 (1.49) 3.64 (1.29) – – <0.001, for time changePPCQ (healthy eating) 3.69 (1.15) 4.08 (1.09) – – 0.030, idem

Leveille et al. [56] RCT (N = 201) PACE 5.60 (N.R.) N.R. 4.97 (N.R.) N.R. 0.028, for group difference

Rose et al. [78] Pretest-posttest design(N = 175)

Stretching andstrengthening (Lorig)

1.07 (1.2) 1.42 (1.2) – – 0.03, for time change

Aerobic exercise (Lorig) 0.68 (0.9) 0.64 (0.7) – – >0.05, idemCognitive symptommanagement (Lorig)

1.48 (1.0) 1.71 (1.0) – – 0.06, idem

Communication withphysician (Lorig)

2.49 (1.4) 2.80 (1.0) – – 0.03, idem

Health related quality of lifeBrand et al. [65] CCT (N = 166) AQoL 8.7 (0.57) N.R. 9.2 (0.48) N.R. >0.05, group differenceColeman et al. [50] RCT (N = 169) SF-36, PCS 47.7 (N.R.) 37.5 (N.R.) 43.8 (N.R.) 37.5 (N.R.) 0.97, for group difference

Counsell et al. [51,52] cRCT (N = 951) SF-36, PCS 35.8 (10.8) N.R. 36.5 (11.2) N.R. 0.38, for group differenceSF-36, MCS 51.0 (10.2) N.R. 51.7 (10.4) N.R. <0.001, idem

Courtney et al. [67] RCT (N = 128) SF-12, PCS 32.6 (10.3) 43.8 (9.4) 34.5 (10.6) 26.0 (9.9) <0.001SF-12, MCS 46.2 (12.7) 59.4 (5.1) 46.4 (10.6) 48.3 (7.7) <0.001

Douglas et al. [73] RCT (N = 334) SF-8, PCS ∼44.0 ∼41.0 ∼44.0 ∼43.0 0.12, for group differenceSF-8, MCS N.R. N.R. N.R. N.R. 0.22, idem

Duke et al. [55] Pretest-posttest design(N = 107)

MGDS N.R. N.R. – – N.R.

136S.R

. de

Bruin et

al. /

Health

Policy 107 (2012) 108– 145

Elzen et al. [77] RCT (N = 144) RAND-36, PCS N.R. N.R. N.R. N.R. >0.05, for group differenceRAND-36, MCS N.R. N.R. N.R. N.R. >0.05, idem

Kobb et al. [40] CCT (N = 1401) SF-36 N.R. N.R. N.R. N.R. Not tested for significanceKrause et al. [64] Pretest-posttest design

(N = 39)SF-36 (physicalfunctioning subscale)

2.46 (1.02) 2.85 (0.81) – – 0.020, for time change

Leveille et al. [56] RCT (N = 201) SF-36 (physicalfunction subscale)

66.4 (22.7) N.R. 62.9 (22.7) N.R. 0.461, for group difference

SF-36 (role limitationscaused by physicalhealth problemssubscale)

53.2 (40.9) N.R. 50.0 (42.2) N.R. 0.610, idem

SF-36 (role limitationscaused by emotionalhealth problemssubscale)

70.6 (40.1) N.R. 76.1 (36.6) N.R. 0.530, idem

SF-36 (general healthsubscale)

66.2 (17.4) N.R. 64.4 (19.2) N.R. 0.751, idem

Markle – Reid et al. [58] RCT (N = 288) SF-36, PCS 37.9 (17.8) 49.1 (24.6) 37.5 (17.7) 46.7 (23.4) 0.522, for group differenceSF-36, MCS 54.3 (19.5) 65.1 (22.4) 60.7 (18.7) 65.2 (22.1) 0.009, idem

Mitton et al. [60] Pretest-posttestdesign (N = 37)

SF-8, PCS 45.5 (N.R.) N.R. – – 0.314, for time changeSF-8, MCS 52.0 (N.R.) N.R. – – 0.968, idemEQ5D N.R. N.R. – – N.R.

Naylor et al. [74] RCT (N = 239) MLHFQ 2.4 (0.7) 2.8 (1.8) 2.3 (0.7) 2.6 (1.7) >0.10, for group differenceMLHFQ (physicaldimension)

2.8 (0.9) 3.1 (1.9) 2.8 (0.9) 2.9 (1.9) >0.10, idem

MLHFQ (emotionaldimension)

3.3 (1.3) 3.1 (1.9) 3.3 (1.2) 3.0 (1.9) >0.10, idem

Rose et al. [78] Pretest-posttest design(N = 175)

Modified version ofHAQ and SF-36:

For time change

Self-rated health 3.5 (0.8) 3.4 (0.7) – – >0.05Disability 0.4 (0.6) 0.4 (0.5) – – >0.05Social/roleactivities/limitations

1.4 (1.3) 1.2 (1.1) – – >0.05

Pain 4.8 (3.4) 4.5 (3.5) – – >0.05Illness intrusiveness 33.0 (16.9) 33.3 (17.5) – – >0.05Energy/fatigue 3.2 (1.0) 3.0 (0.8) – – 0.04Health distress 1.7 (1.3) 1.6 (1.2) – – >0.05Shortness of breath 3.3 (3.1) 3.3 (3.3) – – >0.05

Rubenstein et al. [61] RCT (N = 792) SF-36 (general healthperception subscale)

33.5 (13.8) 35.6 (13.5) 33.7 (12.4) 36.2 (13.1) >0.05, for group difference

Taylor et al. [76] RCT (N = 169) SF-36 N.R. N.R. N.R. N.R. N.R.

Zhang et al. [75] CCT (N = 65132) DUKE For group differencePhysical health 27.2 (23.3) 25.9 (21.9) 33.2 (23.0) 31.1 (25.0) >0.05Mental health 57.8 (25.0) 56.6 (24.1) 59.3 (25.0) 51.8 (24.3) <0.01Social health 61.3 (21.9) 59.1 (21.4) 62.6 (22.9) 21.6 (21.6) <0.01

S.R.

de Bruin

et al.

/ H

ealth Policy

107 (2012) 108– 145137

Table 3a (Continued )



General health 48.4 (18.6) 46.9 (17.6) 51.3 (19.5) 47.4 (19.1) <0.05Perceived health 26.7 (35.1) 23.8 (32.9) 35.7 (38.8) 29.5 (38.4) <0.05Self-esteem 72.6 (22.9) 70.4 (22.5) 71.8 (23.3) 67.6 (23.3) <0.05Anxiety 50.6 (23.0) 52.0 (22.2) 48.5 (21.8) 51.9 (22.2) <0.05Depression 52.5 (24.0) 53.5 (22.8) 49.1 (24.3) 55.2 (23.8) <0.01Anxiety/depression 50.1 (23.7) 51.3 (22.9) 47.3 (23.9) 52.6 (23.0) <0.01Pain 74.2 (32.1) 76.5 (31.5) 64.2 (34.3) 69.0 (34.4) >0.05Disability 27.6 (25.4) 27.7 (25.4) 24.1 (25.1) 28.1 (25.6) <0.01

Medication useBurns et al. [66] RCT (N = 130) N.R. 6.2 (3.3) 5.8 (3.8) 6.9 (3.3) 7.2 (3.5) <0.01, for group difference

>0.05, for time change>0.05, for interaction

Coleman et al. [50] RCT (N = 169) Use of high riskmedication (Pharmacydatabase)

2.0 (N.R.) 1.9 (N.R.) 3.9 (N.R.) 2.5 (N.R.) 0.20, for group difference

Kobb et al. [40] CCT (N = 1401) Prescription refills andinformation of patientsand informal caregivers

– Medication complianceincreased from 68% to93%

– – Not tested for significance

Schmader et al. [70] RCT (N = 864) Inpatient period Group differenceNo. of unnecessarydrugs

0.9 (1.1) 0.4 (0.7) 0.7 (0.9) 0.8 (1.1) <0.0001

Medicationappropriateness(medicationappropriateness index)

10.0 (7.8) 5.3 (4.9) 7.7 (7.2) 9.6 (8.2) <0.0001

No. of inappropriatedrugs

0.5 (0.7) 0.2 (0.5) 0.5 (0.7) 0.4 (0.6) 0.03

No. of conditions withomitted drugs

1.4 (1.3) (1.1) (1.1) 1.1 (1.3) <0.0001

Outpatient period 0.7 (1.0) 0.7 (1.0) 0.6 (0.9) 0.7 (0.9) 0.47No. of unnecessarydrugs

7.5 (7.3) 7.6 (7.9) 7.5 (6.9) 8.3 (6.6) 0.26

Medicationappropriateness(medicationappropriateness index)

0.3 (0.6) 0.3 (0.5) 0.3 (0.6) 0.3 (0.5) 0.78

No. of inappropriatedrugs

1.0 (1.1) 0.8 (0.9) 1.1 (1.2) 1.2 (1.3) 0.0004

No. of conditions withomitted drugs

Sorrento et al. [62] Pretest-posttest design(N = 273)

DAM N.R. N.R. – – Not tested for significance

* For these tables we used information that was provided in the original papers. In several papers, however, information was not complete, which explains why baseline, follow-up, and/or p-values aresometimes missing. It should further be noted that outcomes related to patient satisfaction and quality of care are not included in these tables since they were mostly measured in a qualitative way. For the samereason informal caregiver and professional caregiver related outcomes are not included either.

138 S.R. de Bruin et al. / Health Policy 107 (2012) 108– 145

Table 3bClinical patient outcomes of studies evaluating impact of comprehensive care programs for multimorbid patients.*

Study Study design (N)a Measuring instrumentb Comprehensive caregroupc

Usual care groupc p

BaselineN or %

Follow-upN or %

BaselineN or %

Follow-upN or %

Medication useTibaldi et al. [71] RCT (N = 109) Number of patients

using antipsychoticdrugs

N = 26 N = 6 N = 17 N = 13 0.001, for groupdifference

MortalityCounsell et al. [51,52] cRCT (N = 951) Number of deaths – N = 33 – N = 37 0.64, for group

difference

Dorr et al. [53,54] CCT (N = 3432, ofwhom 1671 hadcomplex illnessincluding diabetes)

Rate of deaths – –All patients withcomplex illness

13.1% 16.6% 0.07, for groupdifference

Patients with complexillness includingdiabetes

N.R. N.R. 0.03, idem

Douglas et al. [73] RCT (N = 334) Number of deaths – N.R. – N.R. >0.05, for groupdifference

Naylor et al. [74] RCT (N = 239) Number of deaths – N = 11 – N = 13 0.830, for groupdifference

Saltvedt et al. [68,69] RCT (N = 254) Number of deaths – N = 35 – N = 43 >0.05, for groupdifference

Tibaldi et al. [71] RCT (N = 109) Number of deaths N.R. N.R. N.R. N.R. >0.05, for groupdifference

a CCT = controlled clinical trial; cRCT = cluster-randomized controlled trial; RCT = randomized controlled trial.b ADL = Activities of Daily Living; AHEAD = Assets and Health Dynamics of the Oldest-Old; AQoL = Assessment of Quality of Life Instrument; BDI = Beck

Depression Inventory; CES-D = Centre for Epidemiologic Studies Depression Scale; DAM = Drug Appropriateness Measure; DUKE = Duke Health Pro-file; EQ5D = A measure of health status of the EuroQol Group; ESDS = Enforces Social Dependency Scale; FIM = Functional Independence Measure;FSQ = Functional Status Questionnaire; GDS = Geriatric Depression Scale; HAQ = Health Assessment Questionnaire; IADL = instrumental activities ofdaily living; MADRS = Montgomery Asberg Depression Rating Scale; MCS = Mental Component Scale; MGDS = Modified Geriatric Depression Scale;MLHFQ = Minnesota Living with Heart Failure Questionnaire; MMSE = Mini Mental State Examination; N.R. = not reported OASIS = Outcomes and Assess-ment Information Set; PACE = Physician-based Assessment and Counseling for Exercise scoring form; PCS = Physical Component Scale PPCQ = Perception ofPersonal Control Questionnaire; SF-8, SF-12, SF-36 = Medical Outcomes 8-item, 12-item, and 36-item short-form.

c N.R. = not reported.* For these tables we used information that was provided in the original papers. In several papers, however, information was not complete, which explains

w furthera alitativer

isqsper

bt

3cwobc[ctcin

hy baseline, follow-up, and/or p-values are sometimes missing. It shouldre not included in these tables since they were mostly measured in a quelated outcomes are not included either.

mprovement of IADL in the comprehensive care group andignificant deterioration in the usual care group [79]. Fre-uency of falls was evaluated in two studies [50,61]. Bothtudies found no difference between the effect of com-rehensive care and usual care. One study evaluated theffect of comprehensive care on physical performance andeported no effect of comprehensive care [56].

Considering these findings, there is no evidence for aeneficial effect of comprehensive care on functional sta-us.

.4.1.4. Health behavior. Health behavior, including physi-al activity, dietary habits, and immunization compliance,as evaluated in five studies. Overall, quality sum-scores

f these studies were low. One study observed that healthehavior (i.e. exercise) improved significantly more in theomprehensive care group than in the usual care group56], whereas one study found no effect of comprehensiveare on health behavior (i.e. exercise and cognitive symp-
om management) [77]. The studies that did not compareomprehensive care with usual care found a (significant)mprovement of behavior towards, diet, exercise, immu-ization or medication but not towards cognitive symptom
be noted that outcomes related to patient satisfaction and quality of care way. For the same reason informal caregiver and professional caregiver

management [40,64,78]. Considering these findings, thereis moderate evidence for a beneficial effect of comprehen-sive care on health behavior.

3.4.1.5. Health-related quality of life. Effectiveness of com-prehensive care regarding health-related quality of lifewas reported in seventeen studies including six withrelatively higher quality sum-scores [50,51,56,58,67,74].Studies differed widely in the definition and measure-ment of health-related quality of life. Thirteen studiescompared health-related quality of life between patientsreceiving comprehensive care and patients receivingusual care. Nine of these studies reported no differ-ences between comprehensive care and usual care groups[40,50,56,61,65,73,74,76,77]. Two studies reported thatcomprehensive care favored usual care in terms of mentalfunctioning but not in terms of physical functioning [51,58].Courtney et al. [67] observed that comprehensive carefavored usual care in terms of mental as well as in terms of
physical functioning. Also Zhang et al. [75] observed for themajority of aspects of quality of life a significant effect ofcomprehensive care (i.e. significantly slower deteriorationthan in the usual care group). Four studies did not compare

lth Polic
comprehensive care with usual care [55,60,64,78], withtwo studies observing an improvement in quality of lifeover time due to comprehensive care [55,64]. Consideringthe inconsistent findings across studies, there is insufficientevidence for a beneficial effect of comprehensive care onquality of life in terms of mental functioning. Almost noneof the studies, including those with relatively higher qual-ity sum-scores, reported positive effects of comprehensivecare on health-related quality of life in terms of physicalfunctioning. Considering these findings, there is no evidencefor a beneficial effect of comprehensive care on quality oflife in terms of physical functioning.

3.4.1.6. Medication use. Six studies evaluated the effectof comprehensive care on medication use. Most of thesestudies had relatively low quality sum-scores. Two stud-ies found no differences between comprehensive careand usual care groups [50,66]. Three studies, however,observed that comprehensive care favored usual care interms of medication compliance, medication appropriate-ness, antipsychotic and unnecessary drug use, detectionof adverse drug reactions, and conditions with omitteddrugs [40,70,71]. One study did not compare comprehen-sive care with usual care but also reported an improvementin medication management due to comprehensive care[62]. Considering the inconsistent findings across studies,there is insufficient evidence for a beneficial effect of com-prehensive care on medication use.

3.4.1.7. Mortality. Mortality was evaluated in six studies(Table 3b) [51,54,69,71,73,74], of which most had rela-tively low quality sum-scores. Two of these studies founda significantly lower mortality rate and/or a significantlylonger time to death in the comprehensive care group thanin the usual care group [54,74]. In one of these studies,however, this effect was only observed for a subgroup ofpatients with complex diseases including diabetes mel-litus [54]. Considering the fact that almost none of thestudies, including those with relatively higher quality sum-scores, reported positive effects of comprehensive care onmortality, there is no evidence for a beneficial effect of com-prehensive care on mortality.

3.4.1.8. Perceived quality of care and satisfaction. Fourteenstudies measured perceived quality of care and patient sat-isfaction. Five studies compared effects of comprehensivecare and usual care [44,48,50,51,74]. Four had relativelyhigh quality sum-scores. Two studies found significantlyhigher perceived quality of care or satisfaction with care inthe comprehensive care group than in the usual care group[44,74]. Two studies observed no differences between com-prehensive care and usual care groups [50,51]. One studyfound that comprehensive care favored usual care for someaspects of care (e.g. quality of patient-physician commu-nication) whereas for other aspects no differences werefound (e.g. physician’s knowledge of patient, quality ofintegration of care, trust in physician) [48].

Nine studies only reported outcomes for patients receiv-ing comprehensive care and found that the majority ofpatients were satisfied with the program. Patients foundthe comprehensive care program helpful and user-friendly,

y 107 (2012) 108– 145 139

and appreciated the holistic approach. They furtherbelieved that the program helped them to manage their dis-ease, had improved their health, and made it easier to getaccess to healthcare services [11,40–42,60,62,64,76,77].Considering the fact that two studies with relativelyhigh quality sum-scores reported positive effects of com-prehensive care on quality of care and satisfaction andthat these findings were supported by qualitative find-ings, there is moderate evidence for a beneficial effectof comprehensive care on quality of care and patientsatisfaction.

3.4.2. Patient related outcomes—economic outcomes3.4.2.1. Healthcare utilization and healthcare costs. Twenty-five studies evaluated the effect of comprehensive care onhealthcare utilization and/or healthcare costs (Table 1). Wedistinguish the effect on inpatient healthcare utilizationand outpatient healthcare utilization. Twenty-one studiesmeasured the utilization of inpatient healthcare services.

Sixteen studies compared inpatient healthcare uti-lization between patients receiving comprehensive careand patients receiving usual care, including seven withrelatively high quality sum-scores [43,46,50,52,56,67,74].Studies differed widely in the definition and measure-ment of utilization of inpatient healthcare services (e.g.number of ER visits, hospital readmission rate, length ofhospital stay, number of hospitalizations, and percentageof patients using nursing home care). Comprehensive carefavored usual care in six studies [51,56,67,72,74,75]. Intwo of these studies the ER visit rate was significantlylower in the comprehensive care group than in the usualcare group [51,75]. In three studies the number of hospital(re)admissions was significantly lower and/or the time toreadmission significantly longer in the comprehensive caregroup than in the usual care group [67,74,75]. In one studylength of hospital stay significantly decreased in the com-prehensive care group whereas it significantly increasedin the usual care group [56]. Two studies observed sig-nificant differences between comprehensive are and usualcare groups that were not in favor of comprehensive care[43,54]. One of these studies found significantly more ERvisits in the comprehensive care group than in the usualcare group [54]. The other study found significantly higherutilization of nursing home care in the comprehensive caregroup than in the usual care group [43]. No significant dif-ferences were observed for other parameters measured inthese studies.

Eight studies that compared utilization of inpatienthealthcare services between patients receiving compre-hensive care and patients receiving usual care observed nosignificant differences between the groups for any of theparameters that were measured [40,46,49,50,61,65,66,76].Although differences in the utilization of inpatient health-care services were mostly not significant, five studies thatalso measured healthcare costs, reported negative incre-mental direct healthcare costs (i.e. lower direct healthcarecosts for patients receiving comprehensive care than
for patients receiving usual care). Across these studies,of which four had relatively high quality sum-scores[43,46,50,74], incremental healthcare costs ranged from−$5708 to −$204 per patient per year [43,46,49,50,74]. This

1 lth Polic

wc

ohsvIsiht[fth[iTa(driist

crcevcspfitygphoigogncg

opwchol

ah


as primarily due to lower inpatient healthcare costs in theomprehensive care group than in the usual care group.

Five of the twenty-one studies that measured utilizationf inpatient healthcare services did not compare compre-ensive care with usual care [42,55,57,60,78]. One of thesetudies reported no significant change over time in ERisits, hospitalizations, and length of hospital stay [78].n the other studies change over time was not tested forignificance. These studies, however, observed reductionsn the use of hospital care and ER visits. Reductions inospitalizations varied from 38% to 69% [42,55,60]. Reduc-ions in the length of hospital stay varied from 37% to 64%55,57,60]. Reductions in the number of ER visits rangedrom 32% to 54% [55,57,60]. These findings are in line withhe findings of the studies that reported direct inpatientealthcare costs for the comprehensive care group only11,55,60,63,72,75]. Two of these studies reported cost sav-ngs ranging from $3029 to $6198 per patient per year.hese savings were primarily due to less hospitalizationsnd ER visits, and reduced length of hospital stay duringre)admission [60,72]. Four other studies that only reportedirect healthcare costs for the comprehensive care groupeported cost reductions due to comprehensive care rang-ng from $147,000 to $1.8 million per year for patientsncluded in the study which again were primarily due toignificant reduced numbers of ER visits and hospitaliza-ions or a reduced length of hospital stay [11,55,63,75].

Six studies measured utilization of outpatient health-are services [40,43,46,56,65,76], of which three hadelatively high quality sum-scores [43,46,56]. All studiesompared comprehensive care with usual care. Bélandt al. [43] observed significantly more GP visits, home careisits, and visits to day hospitals in the comprehensiveare group than in the usual care group. This resulted inignificantly higher outpatient healthcare costs in the com-rehensive care group than in the usual care group. Thesendings are supported by the study that reported posi-ive incremental healthcare costs of $1321 per patient perear (i.e. higher healthcare costs in the comprehensive careroup) [52]. In this study, costs for chronic (including com-rehensive care) and preventive care were significantlyigher in the comprehensive care group. Leveille et al. [56]bserved significantly higher senior centre participationn the comprehensive care group than in the usual careroup, whereas Kobb et al. [40] observed significantly lessutpatient and pharmacy visits in the comprehensive careroup than in the usual care group. Four studies observedo differences in utilization of (other) outpatient health-are services between comprehensive care and usual careroups [46,56,65,76].

Two studies did not specify costs for inpatient andutpatient healthcare services [58,64]. One study com-ared healthcare costs of the comprehensive care groupith these of the usual care group, and found no effect of

omprehensive care [58]. The other study reported onlyealthcare costs for the comprehensive care group andbserved that the actual healthcare costs were significantly
ower than the projected healthcare costs [64].
Although quality sum-scores and findings differedcross studies, there were six studies with relativelyigh quality scores and five studies with relatively lower

y 107 (2012) 108– 145

quality scores that reported decreased inpatient health-care utilization and/or decreased healthcare costs dueto comprehensive care. Based on these findings, there ismoderate evidence for a beneficial effect of comprehensivecare on inpatient healthcare utilization and healthcarecosts. Two studies with relatively high quality sum-scoresobserved higher outpatient healthcare utilization due tocomprehensive care groups. These findings were, however,not supported by the studies with relatively lower qualitysum-scores. Moreover, the four studies that observeddifferences between comprehensive care and usual caregroups widely differed in the measurement of utilization ofoutpatient healthcare services. Considering these findings,there is insufficient evidence for an effect of comprehensivecare on outpatient healthcare utilization.

3.4.3. Informal caregiver related outcomesSix studies reported the effect of comprehensive care

on informal caregivers. Outcomes varied from depressivesymptoms and caregiver strain to caregiver’s productivityloss, satisfaction, and perceived quality of care. Two stud-ies compared effects of comprehensive care with usual care[47,71]. Quality sum-scores of these studies were ratherlow. One of these studies reported significantly higher per-ceived quality of care in caregivers of patients enrolled ina comprehensive care program than in those of patientsthat received usual care, but found no effect of compre-hensive care on caregiver depression, caregiver strain, andproductivity loss [47]. Tibaldi et al. [71], however, observeda significantly larger reduction of caregiver strain in care-givers of patients that received comprehensive care than incaregivers of patients that received usual care. Consideringthese findings, there is no evidence for a beneficial effect ofcomprehensive care on caregiver burden.

Four studies only reported outcomes for caregivers ofpatients that received comprehensive care and observedthat caregivers were satisfied with the program [41,60],experienced an improved ability to cope with the diseaseof their family member, and experienced less anxiety andcaregiver burden [42,60,62]. Although quality sum-scoresof these studies were rather low, these studies’ findingswere rather consistent. Considering these findings, there ismoderate evidence for a beneficial effect of comprehensivecare on satisfaction of informal caregivers.

3.4.4. Professional caregiver related outcomesNine studies reported the effect of comprehensive care

on professional caregivers. Outcomes varied from satis-faction and knowledge to experiences with processes ofcare. Two studies compared outcomes of caregivers provid-ing comprehensive care with those of caregivers providingusual care [45,59]. Quality sum-scores of these studieswere relatively low. Both studies observed that care-givers providing comprehensive care were significantlymore satisfied with communication with patients and fam-ily members and some aspects of their knowledge thancaregivers providing usual care. They did not report any
differences in time spent on patient care and care coordi-nation.
Seven studies only reported outcomes for caregiversproviding comprehensive care and observed that the

lth Polic
majority of caregivers expressed high levels of satisfactionwith e.g. communication with other healthcare providersand newly established partnerships. They further indicatedto be better able to meet the needs of frail older patientsand that the program decreased time spent on patientcare [11,40,41,50,60,76]. One study suggested that com-prehensive care enabled early detection and treatment ofproblems [42]. Although quality sum-scores of the twostudies that compared outcomes of caregivers providingcomprehensive care with those of caregivers providingusual care were rather low, their findings were in line withthe qualitative findings of the other studies. Consideringthese findings, there is moderate evidence for a beneficialeffect of comprehensive care on satisfaction of professionalcaregivers.

4. Discussion

This systematic literature review summarizes availableinformation about the characteristics of comprehensivecare programs for people with multiple chronic conditionsand about their impact on patients and caregivers. There is avast amount of literature on comprehensive care. However,most literature focuses on comprehensive care programsfor people with a single disease [e.g. 19,29–31,82,83–85].Our work focuses on comprehensive care programs forpeople with multiple diseases, and therefore adds to thecurrent literature. Research on multimorbidity is still in itsinfancy [6,14,75]. With this review we therefore aimed tocontribute to a scientific basis for the improvement of carefor patients with multimorbidity.

We conducted a thorough systematic review ofthirty-three studies that examined twenty-eight differentcomprehensive care programs for multimorbid patients.Although our study includes more recent studies thanearlier review studies on the same subject [17,20–25], evi-dence regarding the effectiveness of comprehensive careremains inconsistent. Overall, the effects of comprehen-sive care were comparable to or more positive than thoseof usual care. Moderate evidence was found for a benefi-cial effect of comprehensive care on inpatient healthcareutilization and healthcare costs. Moderate evidence wasalso found for a beneficial effect of comprehensive careon health behavior of patients, perceived quality of care,and satisfaction of patients and caregivers. Insufficient evi-dence was found for a beneficial effect of comprehensivecare on health-related quality of life in terms of mentalfunctioning, medication use, and outpatient healthcare uti-lization and healthcare costs. No evidence was found for abeneficial effect of comprehensive care on cognitive func-tioning, depressive symptoms, functional status, mortality,quality of life in terms of physical functioning, and caregiverburden.

The majority of comprehensive care programs incor-porated interventions related to three or more CCMcomponents. As observed in recent studies on compre-hensive care programs for people with a single disease
[27–30,86], also almost all comprehensive care programsfor people with multiple chronic diseases included inter-ventions related to the CCM components delivery systemdesign and self-management support. In contrast to these
y 107 (2012) 108– 145 141

studies, however, several programs reviewed in our studyincluded interventions promoting access to communityresources. Such interventions were particularly part of theprograms for frail older people and were mostly designedas partnerships between primary care practices or hospi-tals and local community service centres to establish easyaccess to community resources (e.g. transportation, meals,and adult day care) for patients and assessment of patientneeds regarding community services [e.g. 43,45,53,56,60].This may well reflect an increasing attention for socialissues of multimorbid patients. This is an important findingsince the integration of medical and social services is advo-cated to optimize care for patients with multiple chronicconditions [11,87].

The diversity in the effects of comprehensive care pro-grams may be related to the substantial variation in theprograms in terms of target patient groups, design of theintervention/program, and settings in which they wereimplemented. It is hard to determine to which programcomponents positive effects could be attributed and underwhich circumstances comprehensive care programs maybe most effective.

The diversity in the effects of comprehensive care pro-grams may also be related to whether the programs werecorrectly implemented, whether the program componentswere integrated, and whether they were fully adoptedby the patients and the caregivers involved. The level ofimplementation of the comprehensive care programs wasdiscussed in few studies only. Explanations proposed for alack of positive effects of comprehensive care were inad-equate integration of newly developed interventions indaily practice, insufficient understanding of the care inno-vations among caregivers involved, inadequate skills forimplementing the innovations, insufficient time to adaptto newly implemented care models, and non-complianceof patients [43,51,65,67].

Finally, the wide variety in the study results may also berelated to differences in usual care conditions. Our studyyielded comprehensive care programs from seven coun-tries. As was also suggested in studies on comprehensivecare programs for patients with a single-disease [83,88],differences in healthcare systems between countries andbetween settings within countries, and differences in usualcare settings are to be expected. Information about theusual care conditions tended to be poor, which made ithard to distinguish the usual care conditions from the com-prehensive care conditions. Consequently, it is often notpossible to determine whether differences in study resultscould be related to differences in comprehensive care pro-grams or to differences between comprehensive care andusual care.

Although our study provided insight into the char-acteristics of a large number of available programs formultimorbid patients and into their potential effects, weshould acknowledge its limitations. A limitation of oursearch strategy is that we included scientific papers writ-ten in the English language only. Therefore, we may have
missed relevant comprehensive care programs.
Another limitation was that we could not use a completestandardized tool to assess the methodological qualityof the included studies. The usefulness and relevance of

1 lth Polic

etoA[ohtcfpishptqsbnsositc

tawhiitbcvihw

pLpicbwcnFai[oorb

rC


xisting assessment instruments are limited as several ofhe quality assessment criteria are not applicable to studiesn comprehensive care [89]. Also the Health Technologyssessment—Disease Management (HTA-DM) instrument

89] which was specifically developed to assess the qualityf complex interventions such as comprehensive careas limitations and needs further validation [28,90]. Weherefore decided to select six criteria from existing qualityriteria lists [35,36] that in our opinion were most relevantor the assessment of studies on comprehensive carerograms. With this approach, we aimed to provide some

nsight into the quality of the included studies and into thetrength of evidence. Quality assessment was sometimes,owever, hampered by poor or lacking information in theapers. Furthermore, four quality criteria did not applyo studies with non-experimental study designs. Conse-uently, these studies have been assigned lower qualityum-scores, suggesting lower study quality, which may note justifiable. It should further be noted that our study didot only aim to determine the effectiveness of comprehen-ive care programs but also to determine the characteristicsf comprehensive care programs. Like papers describingtudies with higher quality sum-scores, the papers describ-ng studies with relatively lower quality sum-scores couldherefore also provide relevant information regarding theharacteristics of comprehensive care programs.

The consequences of including studies irrespective ofheir methodological quality were that we observed vari-tion in study designs and length of follow-up. Moreover,e observed a wide variety in outcome measurements andow they were made operational, and a high rate of miss-

ng data (e.g. standard deviations for mean values). Thesessues are generally acknowledged by researchers inves-igating comprehensive care [21–23,91–93]. Against thisackdrop we considered estimating an average effect ofomprehensive care by means of meta-analysis of littlealue. Hence, we limited ourselves to qualitative compar-sons of the results of the included studies. It was furtherard to determine which number of CCM components andhich combinations are most promising.

These findings are not unique to comprehensive carerograms for patients with multiple chronic conditions.ikewise, recent studies focusing on comprehensive carerograms for people with a single-disease suggest that

t is hard to determine to which number and to whichomponents positive effects of comprehensive care cane attributed [27,30,31,84]. A study focusing on programsith a single-disease approach suggests that programs that

ontain interventions related to at least one CCM compo-ent improve clinical outcomes and processes of care [29].or COPD patients significant effects on hospitalizationsnd ER visits have been found for programs that containednterventions related to at least two CCM components28,31]. For other outcomes or in other studies focusingn comprehensive care programs for COPD patients it wasbserved that programs should include interventions thatelated to at least three or even four CCM components to
e effective [28,93].
Also with regard to the type of CCM components mixedesults have been reported. Some studies suggest that theCM components self-management support, delivery system

y 107 (2012) 108– 145

design, and decision support are associated with better clin-ical outcomes and/or processes of care [29,30]. Anotherstudy, however, failed to find an association between theelements self-management support or delivery system designand clinical outcomes [94]. The CCM has, however, beenpromoted as a unified package; the different CCM compo-nents are supposed to have a synergistic effect in which thewhole is greater than the sum of the parts [29,84].

The finding that comprehensive care programs maybe effective in terms of reducing healthcare utilizationand as a result in reducing healthcare costs is promising.The increasing prevalence of chronic diseases has majoreconomic consequences for healthcare systems, and it istherefore important to have insight into strategies thatmay reduce healthcare costs. However, comprehensivecare programs are not only implemented to make efficientuse of healthcare resources, but also to improve patienthealth outcomes. In line with other recent studies, wefound hardly any studies that suggested a positive effectof comprehensive care on clinical patient outcomes (i.e.cognition, depressive symptoms, functional status, mortal-ity). An important explanation for this finding, which wasalso suggested by others [21], may be that multimorbidpatients are too a heterogeneous group with large varia-tions in the progression of their diseases and impairments.Conditions may be too severe to be affected by the inter-ventions offered and the time too short to establish aneffect [51,95,96]. Moreover, in most studies standardizedassessment tools were used to evaluate effects of com-prehensive care on patient health outcomes. It has beensuggested that such instruments are not sufficiently sensi-tive to capture clinically important changes over time thatcan be attributed to comprehensive care programs [97,98].

To support decision-making in the field of chronic carefor patients with multiple chronic conditions, it is nec-essary to enhance comparability between studies. It istherefore recommended to better describe the compre-hensive care programs, their level of implementation, andthe usual care conditions to which their effectiveness iscompared. Only then more consistent evidence for theireffectiveness may be obtained. Since there are indica-tions that comprehensive care for patients with multiplechronic conditions may be cost-saving, it is further recom-mended to perform more rigorous economic evaluationsthat include similar cost dimensions and take both costsand effects of comprehensive care into account. Our studyrevealed substantial variation in the costs that were takeninto account in the included studies (e.g. inpatient costs,outpatient costs, costs of comprehensive care program),which hampers comparability of their results. Insight intothe economic effects of comprehensive care, in particu-lar the incremental costs of the programs, are of primaryinterest, since they provide information about the potentialadded value of comprehensive care as compared to usualcare.

In addition, more insight is needed into the impactof comprehensive care on caregiver burden. Literature
suggests that caregiver burden is a predictor of institu-tionalization of frail older people, particularly those withdementia [99]. Older people are increasingly stimulated tolive independently for as long as possible. Future studies

lth Polic

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[


should therefore examine the effect of comprehensive careon predictors of institutionalization, such as caregiver bur-den, to be able to establish whether comprehensive careprograms may delay or prevent institutionalization. It isfurther recommended to gain more insight into patientand caregiver satisfaction, and perceived quality of care.Most studies included in this review measured these out-comes only after the implementation of the comprehensivecare program and did not compare these outcomes with(family caregivers of) patients receiving usual care and pro-fessional caregivers providing usual care. It is thereforerecommended to measure satisfaction and perceived qual-ity of care both before and after the implementation of thecomprehensive care program and thus be able to determinethe effect of comprehensive care. Such information is valu-able since it may provide insight into processes underlyingpositive or negative outcomes of care.

Finally, in line with the client-centered approach inhealthcare it is recommended to incorporate individualgoals and individual preferences and needs in the eval-uation of comprehensive care programs for multimorbidpatients. Such methods (e.g. Goal Attainment Scaling) maybe more meaningful and more sensitive to detect clini-cally important change related to comprehensive care thanstandard assessment tools that are mostly used [97,98].

5. Conclusion

There is a broad array of comprehensive care programsavailable to patients with multiple chronic conditions thatoften integrate both medical and social services. Despiteindications that comprehensive care programs for multi-morbid patients decrease inpatient healthcare utilizationand healthcare costs, improve health behavior of patientsand perceived quality of care, and realize satisfaction ofpatients and caregivers because of the heterogeneity of theprograms it is as yet too early to draw firm conclusionsregarding their effectiveness. It remains hard to determinewhich patient groups benefit most from comprehensivecare and which program characteristics contribute most topositive effects of comprehensive care. More consistentlyperformed evaluation studies are necessary to determinewhat constitutes best care for the increasing number ofpeople with multiple chronic conditions.

References

[1] World Health Organization. The global burden of disease, 2004update. Retrieved 2 February 2011, from: http://www.who.int/healthinfo/global burden disease/GBD report 2004update full.pdf.

[2] World Health Organization. The European health report 2009: healthand health systems. Copenhagen: World Health Organization; 2009.

[3] World Health Organization. Preventing chronic diseases: avital investment. Retrieved 2 February 2011, from: http://whqlibdoc.who.int/publications/2005/9241563001 eng.pdf.

[4] Marengoni A, Winblad B, Karp A, Fratiglioni L. Prevalence of chronicdiseases and multimorbidity among the elderly population inSweden. American Journal of Public Health 2008;98:1198–200.

[5] Boyd C, Fortin M. Future of multimorbidity research: how should
understanding of multimorbidity inform health system design? Pub-lic Health Reviews 2010;32:451–74.
[6] Marengoni A, Angleman S, Melis R, Mangialasche F, Karp A, Gar-men A, et al. Aging with multimorbidity: a systematic review of theliterature. Ageing Research Reviews 2011;10:430–9.

[

y 107 (2012) 108– 145 143

[7] Feinstein AR. The pre-therapeutic classification of co-morbidity inchronic disease. Journal of Chronic Disease 1970;23:455–68.

[8] Vogeli C, Shields AE, Lee TA, Gibson TB, Marder WD, Weiss KB, et al.Multiple chronic conditions: prevalence, health consequences, andimplications for quality, care management, and costs. Journal of Gen-eral Internal Medicine 2007;22(Suppl. 3):391–5.

[9] Van den Akker M, Buntinx F, Metsemakers JF, Roos S, KnottnerusJA. Multimorbidity in general practice: prevalence, incidence, anddeterminants of co-occurring chronic and recurrent diseases. Journalof Clinical Epidemiology 1998;51:367–75.

10] Fortin M, Bravo G, Hudon C, Lapointe L, Almirall J, Dubois MF,et al. Relationship between multimorbidity and health-related qual-ity of life of patients in primary care. Quality of Life Research2006;15:83–91.

11] Wright K, Hazelett S, Jarjoura D, Allen K. The AD-LIFE trial: working tointegrate medical and psychosocial care management models. HomeHealthcare Nurse 2007;25:308–14.

12] Clarfield AM, Bergman H, Kane R. Fragmentation of care for frail olderpeople—an international problem. Experience from three countries:Israel, Canada, and the United States. Journal of the American Geri-atrics Society 2001;49:1714–21.

13] Greß S, Baan CA, Clanan M, Dedeu T, Groenewegen P, Howson H, et al.Co-ordination and management of chronic conditions in Europe: therole of primary care—position paper of the European forum for pri-mary care. Quality in Primary Care 2009;17:75–86.

14] Fortin M, Soubhi H, Hudon C, Bayliss EA, Van den Akker M. Mul-timorbidity’s many challenges. British Medical Journal 2007;334:1016–7.

15] Westert GP, Satariano WA, Schellevis FG, Van Den Bos GAM. Patternsof comorbidity and the use of health services in the Dutch population.European Journal of Public Health 2001;11:365–72.

16] Versnel N, Welschen LM, Baan CA, Nijpels G, Schellevis FG. The effec-tiveness of case management for comorbid diabetes type 2 patients;the CasCo study. Design of a randomized controlled trial. BMC FamilyPractice 2011;12, http://dx.doi.org/10.1186/471-2296-12-68.

17] Boult C, Green AF, Boult LB, Pacala JT, Snyder C, Leff B. Successfulmodels of comprehensive care for older adults with chronic con-ditions: evidence for the institute of medicine’s “retooling for anAging America” report. Journal of the American Geriatrics Society2009;57:2328–37.

18] Wagner EH, Bennett SM, Austin BT, Greene SM, Schaefer JK, VonkorffM. Finding common ground: patient-centeredness and evidence-based chronic illness care. Journal of Alternative and ComplementaryMedicine 2005;11(Suppl. 1):S7–15.

19] Mattke S, Seid M, Ma S. Evidence for the effect of disease manage-ment: is $1 billion a year a good investment? American Journal ofManaged Care 2007;13:670–6.

20] Boult C, Wieland GD. Comprehensive primary care for older patientswith multiple chronic conditions: “Nobody rushes you through”.JAMA 2010;304:1936–43.

21] Eklund K, Wilhelmson K. Outcomes of coordinated and integratedinterventions targeting frail elderly people: a systematic review ofrandomised controlled trials. Health and Social Care in the Commu-nity 2009;17:447–58.

22] Latour CH, van der Windt DA, de Jonge P, Riphagen II, de Vos R,Huyse FJ, et al. Nurse-led case management for ambulatory com-plex patients in general health care: a systematic review. Journal ofPsychosomatic Research 2007;62:385–95.

23] Oeseburg B, Wynia K, Middel B, Reijneveld SA. Effects of case manage-ment for frail older people or those with chronic illness: A systematicreview. Nursing Research 2009;58:201–10.

24] Peikes D, Chen A, Schore J, Brown R. Effects of care coordinationon hospitalization, quality of care, and health care expendi-tures among Medicare beneficiaries: 15 randomized trials. JAMA2009;301:603–18.

25] Low LF, Yap M, Brodaty H. A systematic review of different modelsof home and community care services for older persons. BMC HealthServices Research 2011;11:93.

26] Norris SL, Nichols PJ, Caspersen CJ, Glasgow RE, Engelgau MM, Jack L,et al. The effectiveness of disease and case management for peoplewith diabetes. A systematic review. American Journal of PreventiveMedicine 2002;22:15–38.

27] De Bruin SR, Heijink R, Lemmens L, Struijs JN, Baan CA. Impact of dis-ease management programs on healthcare expenditures for patients
with diabetes, depression, heart failure or chronic obstructive pul-monary disease: a systematic review of the literature. Health Policy2011;101:105–21.
28] Lemmens K, Lemmens L, Boom J, Drewes H, Meeuwissen J, SteutenL, et al. Chronic care management for patients with COPD: a critical

http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf

http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf

http://whqlibdoc.who.int/publications/2005/9241563001_eng.pdf

http://whqlibdoc.who.int/publications/2005/9241563001_eng.pdf

dx.doi.org/10.1186/471-2296-12-68

1 lth Polic

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[


review of available evidence. Journal of Evaluation in Clinical Practice2011, http://dx.doi.org/10.1111/j.1365-2753.2011.01805.x.

29] Tsai AC, Morton SC, Mangione CM, Keeler EB. A meta-analysis ofinterventions to improve care for chronic illnesses. American Journalof Managed Care 2005;11:478–88.

30] Zwar N, Harris M, Griffiths R, Roland M, Dennis S, Powell Davis G, et al.APHCRI stream four: a systematic review of chronic disease manage-ment. Canberra: Australian Primary Health Care Research Institute(APHCRI); 2006.

31] Adams SG, Smith PK, Allan PF, Anzueto A, Pugh JA, Cornell JE. Sys-tematic review of the chronic care model in chronic obstructivepulmonary disease prevention and management. Archives of InternalMedicine 2007;167:551–61.

32] Wagner EH, Austin BT, Davis C, Hindmarsh M, Schaefer J, BonomiA. Improving chronic illness care: translating evidence into action.Health Affairs (Millwood) 2001;20:64–78.

33] Wagner EH. The role of patient care teams in chronic disease man-agement. British Medical Journal 2000;320:569–72.

34] Organisation for Economic Co-operation and Development. OECDHealth Data 2009. Paris: OECD; 2009.

35] Van Tulder M, Furlan A, Bombardier C, Bouter L. Updated methodguidelines for systematic reviews in the cochrane collaboration backreview group. Spine (Phila Pa 1976) 2003;28:1290–9.

36] Verhagen AP, de Vet HC, de Bie RA, Kessels AG, Boers M, BouterLM, et al. The Delphi list: a criteria list for quality assessmentof randomized clinical trials for conducting systematic reviewsdeveloped by Delphi consensus. Journal of Clinical Epidemiology1998;51:1235–41.

37] Proper KI, Singh AS, van Mechelen W, Chinapaw MJ. Sedentarybehaviors and health outcomes among adults: a systematic reviewof prospective studies. American Journal of Preventive Medicine2011;40:174–82.

38] Hoogendoorn WE, van Poppel MN, Bongers PM, Koes BW, Bouter LM.Systematic review of psychosocial factors at work and private life asrisk factors for back pain. Spine (Phila Pa 1976) 2000;25:2114–25.

39] Singh AS, Mulder C, Twisk J, Van Mechelen W, Chinapaw MJ. Trackingof childhood overweight into adulthood: a systematic review of theliterature. Obesity reviews 2008;9:474–88.

40] Kobb R, Hoffman N, Lodge R, Kline S. Enhancing elder chronic carethrough technology and care coordination: report from a pilot.Telemedicine Journal and e-Health 2003;9:189–95.

41] Liddy C, Dusseault JJ, Dahrouge S, Hogg W, Lemelin J, Humbert J. Tele-homecare for patients with multiple chronic illnesses: pilot study.Canadian Family Physician 2008;54:58–65.

42] Procter S, Single A. Home telehealthcare: findings from a pilot studyin North-east London. The British Journal of Healthcare Computing& Information Management 2006;23:10–3.

43] Béland F, Bergman H, Lebel P, Dallaire L, Fletcher J, ContandriopoulosAP, et al. Integrated services for frail elders (SIPA): a trial of a modelfor Canada. Canadian Journal on Aging 2006;25:5–42.

44] Boyd C, Reider L, Frey K, Scharfstein D, Leff B, Wolff J, et al. Theeffects of guided care on the perceived quality of health care formulti-morbid older persons: 18-month outcomes from a cluster-randomized controlled trial. Journal of General Internal Medicine2010;25:235–42.

45] Boult C, Reider L, Frey K, Leff B, Boyd CM, Wolff JL, et al. Early effects of“Guided Care” on the quality of health care for multimorbid older per-sons: a cluster-randomized controlled trial. Journals of GerontologySeries A-Biological Sciences & Medical Sciences 2008;63:321–7.

46] Leff B, Reider L, Frick KD, Scharfstein DO, Boyd CM, Frey K, et al.Guided care and the cost of complex healthcare: a preliminary report.The American Journal of Managed Care 2009;15:555–9.

47] Wolff JL, Giovannetti ER, Boyd CM, Reider L, Palmer S, Scharfstein D,et al. Effects of guided care on family caregivers. The Gerontologist2010;50:459–70.

48] Boyd C, Shadmi E, Conwell LJ, Griswold M, Leff B, Brager R, et al. Apilot test of the effect of guided care on the quality of primary careexperiences for multimorbid older adults. Journal of General InternalMedicine 2008;23:536–42.

49] Sylvia ML, Griswold M, Dunbar L, Boyd CM, Park M, Boult C. Guidedcare: cost and utilization outcomes in a pilot study. Disease Manage-ment 2008;11:29–36.

50] Coleman EA, Grothaus LC, Sandhu N, Wagner EH. Chronic care clin-ics: a randomized controlled trial of a new model of primary care
for frail older adults. Journal of the American Geriatrics Society1999;47:775–83.
51] Counsell SR, Callahan CM, Clark DO, Tu W, Buttar AB, Stump TE, et al.Geriatric care management for low-income seniors: a randomizedcontrolled trial. JAMA 2007;298:2623–33.

[

y 107 (2012) 108– 145

52] Counsell SR, Callahan CM, Tu W, Stump TE, Arling GW. Cost analysisof the Geriatric Resources for Assessment and Care of Elders caremanagement intervention. Journal of the American Geriatrics Society2009;57:1420–6.

53] Dorr DA, Wilcox A, Burns L, Brunker CP, Narus SP, Clayton PD. Imple-menting a multidisease chronic care model in primary care usingpeople and technology. Disease Management 2006;9:1–15.

54] Dorr DA, Wilcox AB, Brunker CP, Burdon RE, Donnelly SM. Theeffect of technology-supported, multidisease care management onthe mortality and hospitalization of seniors. Journal of the AmericanGeriatrics Society 2008;56:2195–202.

55] Duke C. The frail elderly community-based case managementproject. Geriatric Nursing 2005;26:122–7.

56] Leveille SG, Wagner EH, Davis C, Grothaus L, Wallace J, LoGerfoM, et al. Preventing disability and managing chronic illness in frailolder adults: a randomized trial of a community-based partner-ship with primary care. Journal of the American Geriatrics Society1998;46:1191–8.

57] Littleford A, Kralik D. Making a difference through integrated com-munity care for older people. Journal of Nursing and Healthcare ofChronic Illness 2010:178–86.

58] Markle-Reid M, Weir R, Browne G, Roberts J, Gafni A, HendersonS. Health promotion for frail older home care clients. Journal ofAdvanced Nursing 2006;54:381–95.

59] Marsteller JA, Hsu YJ, Reider L, Frey K, Wolff J, Boyd C, et al. Physiciansatisfaction with chronic care processes: a cluster-randomized trialof guided care. Annals of Family Medicine 2010;8:308–15.

60] Mitton C, O’Neil D, Simpson L, Hoppins Y, Harcus S. Nurse-PhysicianCollaborative Partnership: a rural model for the chronically ill. Cana-dian Journal of Rural Medicine 2007;12:208–16.

61] Rubenstein LZ, Alessi CA, Josephson KR, Trinidad Hoyl M, HarkerJO, Pietruszka FM. A randomized trial of a screening, case finding,and referral system for older veterans in primary care. Journal of theAmerican Geriatrics Society 2007;55:166–74.

62] Sorrento TA, Casselman MB, Noonan J. A consultant pharmacist’srole on a geriatric assessment service. The Consultant Pharmacist2007;22:935–43.

63] Anonymous. DM for complex patients tackles hospital utilization ina big way. Disease Management Advisor 2003;9:99–101.

64] Krause CM, Jones CS, Joyce S, Kuhn ME, Curtin K, Murphy LP, et al.The impact of a multidisciplinary, integrated approach on improv-ing the health and quality of care for individuals dealing withmultiple chronic conditions. American Journal of Orthopsychiatry2006;76:109–14.

65] Brand CA, Jones CT, Lowe AJ, Nielsen DA, Roberts CA, King BL, et al.A transitional care service for elderly chronic disease patients at riskof readmission. Australian Health Review 2004;28:275–84.

66] Burns R, Nichols LO, Graney MJ, Cloar FT. Impact of continued geri-atric outpatient management on health outcomes of older veterans.Archives of Internal Medicine 1995;155:1313–8.

67] Courtney M, Edwards H, Chang A, Parker A, Finlayson K, Hamil-ton K. Fewer emergency readmissions and better quality of life forolder adults at risk of hospital readmission: a randomized controlledtrial to determine the effectiveness of a 24-week exercise and tele-phone follow-up program. Journal of the American Geriatrics Society2009;57:395–402.

68] Saltvedt I, Jordhoy M, Opdahl Mo ES, Fayers P, Kaasa S, SletvoldO. Randomised trial of in-hospital geriatric intervention: impact onfunction and morale. Gerontology 2006;52:223–30.

69] Saltvedt I, Mo ES, Fayers P, Kaasa S, Sletvold O. Reduced mortalityin treating acutely sick, frail older patients in a geriatric evaluationand management unit. A prospective randomized trial. Journal of theAmerican Geriatrics Society 2002;50:792–8.

70] Schmader KE, Hanlon JT, Pieper CF, Sloane R, Ruby CM, Twersky J, et al.Effects of geriatric evaluation and management on adverse drug reac-tions and suboptimal prescribing in the frail elderly. The AmericanJournal of Medicine 2004;116:394–401.

71] Tibaldi V, Aimonino N, Ponzetto M, Stasi MF, Amati D, Raspo S, et al.A randomized controlled trial of a home hospital intervention forfrail elderly demented patients: behavioral disturbances and care-giver’s stress. Archives of Gerontology and Geriatrics Suppl 2004:431–6.

72] Daly BJ, Douglas SL, Kelley CG, O’Toole E, Montenegro H. Trial of adisease management program to reduce hospital readmissions of the
chronically critically ill. Chest 2005;128:507–17.
73] Douglas SL, Daly BJ, Kelley CG, O’Toole E, Montenegro H. Chroni-cally critically ill patients: health-related quality of life and resourceuse after a disease management intervention. American Journal ofCritical Care 2007;16:447–57.

dx.doi.org/10.1111/j.1365-2753.2011.01805.x

lth Polic

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[


74] Naylor MD, Brooten DA, Campbell RL, Maislin G, McCauley KM,Schwartz JS. Transitional care of older adults hospitalized with heartfailure: a randomized, controlled trial. Journal of the American Geri-atrics Society 2004;52:675–84.

75] Zhang NJ, Wan TT, Rossiter LF, Murawski MM, Patel UB.Evaluation of chronic disease management on outcomes andcost of care for Medicaid beneficiaries. Health Policy 2008;86:345–54.

76] Taylor CB, Miller NH, Reilly KR, Greenwald G, Cunning D, Deeter A,et al. Evaluation of a nurse-care management system to improveoutcomes in patients with complicated diabetes. Diabetes Care2003;26:1058–63.

77] Elzen H, Slaets JP, Snijders TA, Steverink N. Evaluation of thechronic disease self-management program (CDSMP) among chroni-cally ill older people in the Netherlands. Social Science & Medicine2007;64:1832–41.

78] Rose MA, Arenson C, Harrod P, Salkey R, Santana A, Diamond J.Evaluation of the Chronic Disease Self-Management Program withlow-income, urban, African American older adults. Journal of Com-munity Health Nursing 2008;25:193–202.

79] Chumbler NR, Mann WC, Wu S, Schmid A, Kobb R. The associationof home-telehealth use and care coordination with improve-ment of functional and cognitive functioning in frail elderly men.Telemedicine Journal and e-Health 2004;10:129–37.

80] Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practicalmethod for grading the cognitive state of patients for the clinician.Journal of Psychiatric Research 1975;12:189–98.

81] Keith RA, Granger CV, Hamilton BB, Sherwin FS. The functional inde-pendence measure: a new tool for rehabilitation. Advanced ClinicalRehabilitation 1987;1:6–18.

82] Ofman JJ, Badamgarav E, Henning JM, Knight K, Gano Jr AD, LevanRK, et al. Does disease management improve clinical and economicoutcomes in patients with chronic diseases? A systematic review.American Journal of Medicine 2004;117:182–92.

83] Singh D. How can chronic disease management programmes oper-ate across care settings and providers? Birmingham: World HealthOrganization; 2008.

84] Weingarten SR, Henning JM, Badamgarav E, Knight K, HasselbladV, Gano Jr A, et al. Interventions used in disease manage-ment programmes for patients with chronic illness-which oneswork? Meta-analysis of published reports. British Medical Journal2002;325:925.

85] Goetzel RZ, Ozminkowski RJ, Villagra VG, Duffy J. Return on invest-
ment in disease management: a review. Health Care FinancingReview 2005;26:1–19.
86] De Bruin SR, Baan CA, Struijs JN. Pay-for-performance in disease man-agement: a systematic review of the literature. BMC Health ServicesResearch 2011;11, http://dx.doi.org/10.1186/472-6963-11-272.

[

y 107 (2012) 108– 145 145

87] Allen KR, Hazelett SE, Jarjoura D, Wright K, Fosnight SM, Kropp DJ,et al. The after Discharge Care Management of Low Income FrailElderly (AD-LIFE) randomized trial: theoretical framework and studydesign. Population Health Management 2011;14:137–42.

88] Coleman K, Austin BT, Brach C, Wagner EH. Evidence on the ChronicCare Model in the new millennium. Health Affairs (Millwood)2009;28:75–85.

89] Steuten LMG, Vrijhoef HJM, Van Merode GG, Severens JL, Spreeuwen-berg C. The Health Technology Assessment-Disease Managementinstrument reliably measured methodologic quality of health tech-nology assessments of disease management. Journal of ClinicalEpidemiology 2004;57:881–8.

90] Drewes H, Steuten L, Lemmens L, Baan C, Boshuizen H, Elis-sen A, et al. The effectiveness of chronic care managementfor heart failure: meta-regression analysis to explain theheterogeneity in outcomes. Health Services Research 2012,http://dx.doi.org/10.1111/j.1475-6773.2012.01396.x.

91] Peytremann-Bridevaux I, Staeger P, Bridevaux PO, Ghali WA,Burnand B. Effectiveness of chronic obstructive pulmonary disease-management programs: systematic review and meta-analysis. TheAmerican Journal of Medicine 2008;121, 433-43 e4.

92] Seto E. Cost comparison between telemonitoring and usual care ofheart failure: a systematic review. Telemedicine Journal and e-Health2008;14:679–86.

93] Lemmens K, Nieboer A, Huijsman R. A systematic review of inte-grated use of disease management interventions in asthma andCOPD. Respiratory Medicine 2009;103:670–91.

94] Solberg LI, Crain AL, Sperl-Hillen JM, Hroscikoski MC, Engebret-son KI, O’Connor PJ. Care quality and implementation of thechronic care model: a quantitative study. Annals of Family Medicine2006;4:310–6.

95] Baumgarten M, Lebel P, Laprise H, Leclerc C, Quinn C. Adult day carefor the frail elderly: outcomes, satisfaction, and cost. Journal of Agingand Health 2002;14:237–59.

96] Zank S, Schacke C. Evaluation of geriatric day care units: effects onpatients and caregivers. The Journals of Gerontology Series B: Psy-chological Sciences and Social Sciences 2002;57:P348–57.

97] Rockwood K, Howlett S, Stadnyk K, Carver D, Powell C, StoleeP. Responsiveness of goal attainment scaling in a randomizedcontrolled trial of comprehensive geriatric assessment. Journal ofClinical Epidemiology 2003;56:736–43.

98] Bouwens SF, van Heugten CM, Verhey FR. Review of goal attain-ment scaling as a useful outcome measure in psychogeriatric patients
with cognitive disorders. Dementia and Geriatric Cognitive Disorders2008;26:528–40.
99] Luppa M, Luck T, Brähler E, König HH, Riedel-Heller SG. Prediction ofinstitutionalisation in dementia: a systematic review. Dementia andGeriatric Cognitive Disorders 2008;26:65–78.

dx.doi.org/10.1186/472-6963-11-272

dx.doi.org/10.1111/j.1475-6773.2012.01396.x

comprehensive care programs for patients with multiple chronic conditions: a systematic literature...

Documents