9-Day Level 3 Courses Instruction Guidelines

Course Format:

• 9 days of face to face lab time • 1 day of synchronous / asynchronous online lecture (more details below) • 1 day of generic online lectures – www.orthodived.ca • Written examination on the last day of the course – 1.5 hours • Case history examination and clinical reasoning: new format and

evaluation (see below for details) Remuneration:

• Chief, senior and co-instructors will be paid for 10 days of teaching as per the instructor fee guidelines plus cost of the online lecture service

• Lab assistants are paid for the 9 days of lab time as per the instructor fee guidelines

• Expenses are covered as outlined in the PODCR policy and procedure manual

Teaching recommendations: 9 days face-to-face lab time: This component should be devoted to practical time and case based learning exercises. An example of how this can be facilitated can be found in the appendix of this document. 1 day of online lectures: Any theory lectures you would normally teach during course time we recommend be done online prior to the practical lab sessions. ie. Review of the biomechanics and the anatomy of the cervical spine etc. We are recommending that instructors use an online service called dim dim, or webex. The web address is www.dimdim.com or www.webex.com . There are online tutorials on the websites to help you familiarize you with the webinar process. We will be hosting an online tutorial in August. These lectures can be recorded and downloaded as a flash file that can be sent to participants to review in their own time. Generic online lectures level 3

• Myokinetics and myokinematics • Exercise physiology • Blood supply of the spinal cord • Effects and efficacy of manipulation • Basic biomechanics (review from level 2) • Contraindications and indications to manipulation • Vertebral artery

• Vestibular Rehabilitation (level 3 Upper) – Available for September • Differential Diagnosis of Groin pain – case based learning session (level 3

lower) – Available for September • Urinary Incontinence (level 3 Lower) – Available for September

Please email the course participants which lectures they should review prior and to and during the course. These lectures are located at www.orthodived.ca The student username is: student@orthodived.ca Password: pass Instructor Login: instructor@orthodived.ca Password: manipulation Written Examination: The written examination will continue in the same format. This will be administered on the last day of the course. Case history evaluation: We are developing a different form of evaluation for the case history exam. The level 3 case will be done in a small groups facilitated by the instructors and assistants. A clinical reasoning template will be developed to reflect the existing case history booklet. This will allow for synchronous evaluation and general feedback for the students. In order for a participant to pass this portion of the course they must be present and participate in the group exercise to get credit for this portion of the course. 2 hours should be allotted during the course for this group exercise. A tutorial will be created to facilitate the instruction process. This template will be made available prior to the instructors meeting in St. Johns. Case Based Learning: Practical sessions during level 3 should incorporate components of the clinical reasoning template. Instructors are responsible for creating their own scenarios to illustrate the clinical application of assessment and treatment techniques, using the best available evidence. see appendix for an example. Rationale: Changes to the level 3 courses are designed to reflect the current ICF (International classification of function and disability) model of health. The online format will allow more time for incorporating clinical reasoning into practical sessions, and to minimize the amount of travel time required for course participants.

Level 3 Lower Case Study Example

lelevesque@sympatico.calelevesque@sympatico.ca

Lumbopelvic Case

Subjective � 21 year old university student complaining of pain in the left hip

and sacroiliac region after a fall while playing ultimate frisbee� Mechanism of injury – landed with the right leg back stretching

the front of the right thigh while the left hip and knee were in a flexed position

� Very active – rollerblades, plays competitive frisbee, cycles –has had to stop playing ultimate frisbee because of the pain in the left lower back and SI regionhas had to stop playing ultimate frisbee because of the pain in the left lower back and SI region

� 4th year psychology student currently working on her thesis –sits at the computer for prolonged periods

� Otherwise healthy � Medications – birth control, flovent for allergies� No diagnostic imaging � FABQ-Work – 2 FABQ-Physical Activity - 26

P1 left sided low back and SI pain NPRS 5/10Aggravated by walking, running and twisting

Relieved with cessation of activity and heat

P2 left sided lateral buttock pain along the iliac crest NPRS 3/10Aggravated by stretching movements to the right and sleeping on that side

Relieved by a change in position or activity

Generate your hypotheses

Lumbar spine

SI joint

Motor control

Hip

Muscle imbalance

Consider the ICF Model

Impairments in Activities/Participation

Getting into and being in a squat or partial squat position

Running and twisting movements (ultimate frisbee)

Sitting for prolonged periods at the computer

Unable to participate in her recreational activities

Difficulty completing thesis work

Quantification of Functional LimitationsWhat Outcome measures would you use ??

� Patient Specific Functional Scale

� Quebec Back Pain Disability Scale

� 20 items covering areas of daily activities selected from 6 relevant sub domains of functional skills for patients with low back pain pain

� Numerical Pain Rating Scale

� Roland Morris Questionnaire

� Oswestry

� ASLR

� Functional Performance Testing

� Deep squat position: NPRS @ baseline and after 2 minutes

Treatment Intervention

The treatment intervention for the left SI joint

was a supine gap manipulation.

Level 4/ 5 Manual Orthopaedic Division CPA

Objective Examination

What could you have found on your objective

examination of the pelvis that would lead you

to this treatment intervention?to this treatment intervention?

What evidence based tests would you include

in your examination of the pelvis?

Did you consider these tests and if so what would you expect to find?

Did you consider these tests and if so what would you expect to find?

Functional Movements

� Besides observing performance of a squat,

what other functional performance tests will

you look at and what could you possibly find?

� Keep in mind this patient’s activity and

participation restrictions.

Evidence to Consider

Reliability of the Stork TestHungerford et al 2007

Reliability of manual therapists to detect a pattern of intrapelvic motion during a weight-bearing task.

2 point scale - +ve if the PSIS moved cephalad relative to the sacrum or –ve if the PSIS stayed relative to the sacrum or –ve if the PSIS stayed neutral or moved caudad relative to the sacrum.

Interrater reliability was good for both sides and the PCA (percent close agreement) was high.

Clinical Relevance

� The Stork test assesses the ability of a subject to maintain a stable alignment of the innominate bone relative to the sacrum during a functional load transfer task.

� Failure to maintain this alignment results in a cephalad motion of the PSIS or anterior rotation of the innominate relative to the of the PSIS or anterior rotation of the innominate relative to the sacrum.

� Positive test suggests an inability of the SIJ to engage the self-bracing mechanism.

� Suggests that the recognition of altered movement patterns is possible.

Hungerford BA. Gilleard W. Moran M. Emmerson C. Evaluation of the Ability of Physical Therapist to Palpate Intrapelvic Motion with the Stork Test on the Support Side. Physical Therapy 2007 87(7) 879-887

Differences in Standing and Forward Bending in Women with Chronic Low Back Pain or Pelvic Girdle

Pain van Wingerden JP et al Spine 2008

� PGP patients in erect stance - the pelvis is significantly tilted backwards (7 degrees) and slight flattened lumbar lordosis.

� During the first one third of forward bending, LBP patients tend to maintain lordosis, PGP patients emphasize lumbar flexion.maintain lordosis, PGP patients emphasize lumbar flexion.

� At maximal forward bending, the ROM of the trunk was limited in all patient groups, but only the PGP group had significantly limited hip motion.

� Specific, consistent distinct motion patterns

� Functional compensation strategies following altered neuromuscular control

Diagnostic Validity of Criteria for Sacroiliac Joint Pain: A Systematic Review

Karolina M. Szadek,* Peter van der Wurff,y Maurits W. van Tulder,zWouter W. Zuurmond,x and Roberto S. G. M. Perezk

The Journal of Pain, Vol 10, No 4 (April), 2009: pp 354-368Available online at www.sciencedirect.com

2 individual pain provocation tests – the compression and thigh thrust

test – if positive more likely to have SI joint pain

Using a threshold of 3 or more positive stressing tests, the DOR of 3

positive provocation tests is high (DOR, 17.2)

Cannot distinguish which structure – intra-articular, ligamentous,

muscular

Diagnostic Validity of Criteria for Sacroiliac Joint Pain: A Systematic Review

Karolina M. Szadek,* Peter van der Wurff,y Maurits W. van Tulder,zWouter W. Zuurmond,x and Roberto S. G. M. Perezk

The Journal of Pain, Vol 10, No 4 (April), 2009: pp 354-368Available online at www.sciencedirect.com

Pain mapping or pain referral patterns have the ability to correctly identify patients with SI joint pain however, they fail in discriminating patients without SI joint pain.

Pain from the SI joint can extend to the buttock, groin and to the lower Pain from the SI joint can extend to the buttock, groin and to the lower extremity not exclusive for SI joint structures.

Some investigators postulate that patients with presumed SI joint pain point out the area adjacent to the PSIS

Based on 1 study, this clinical test has reasonable sensitivity and specificity with a DOR of 2.745 - consistent with Fortin et al

Questions??

� Is there anything else you would consider in

your examination of the pelvis?

Are there any other manipulation/mobilization � Are there any other manipulation/mobilization

techniques you would have chosen instead

and if so why?

� Is there any evidence to support the

treatment intervention?

Now, let’s look at her lumbar spine.......Now, let’s look at her lumbar spine.......

Note*** Instructors – if you feel you would like to stop at this point and

make the case strictly SI joint you could and ask students how they ruled

out lumbar spine involvement.

The following slides incorporate a quadrant approach and include

examination and treatment of the lumbar spine.

Active ROM

Combined Movement Testing

What did you observe?

AROM:

Decreased right side flexion

Crease in extension

Combined movement testing

- decreased flexion/right side flexion/right

rotation with reproduction of pain on the left

- pain reproduced with the left extension

quadrant

PIVMs / PAVMs

� Decreased unilateral flexion on the left at

L4/L5 with a capsular end feel

� Decreased superoanterior glide on the left

L4/L5 with a capsular end feelL4/L5 with a capsular end feel

Stability Testing

� Increased anterior translation of L3 on L4 –

late capsular end feel – increased neutral

zone

� Improved with activation of the core (TA)

� What significance does this have with

regards to manipulation/mobilization of her

lumbar spine?

Motor Control

� Poor motor control – inability of the local stabilizing muscles to stabilize the SI and lumbar spine making the transfer of load through the pelvis and lumbar spine through the pelvis and lumbar spine difficult

� How would you assess this?

Findings of Altered Motor Control in Subjects with PGP

Activity Altered Motor Control Strategies

De Groot et al

Hungerford et al (2007)

ASLR

Standing hip flexion

Increased bilateral EO activityDelayed onset of IO and multifidus bilaterallyDelayed onset of gluteus maximus on the symptomatic sideEarly activation of biceps femoris on the symptomatic sidefemoris on the symptomatic side

O’Sullivan et al (2002) ASLR Decreased diaphragmatic excursionAltered respiratory patternsDescent of the pelvic floor

O’Sullivan and Beales (2007, 2009, 2010)

Voluntary pelvic floor contraction Increased pelvic floor activation

Pool-Goudzwaard et al (2009 2010)

Voluntary pelvic floor maneuevers

Testing Criteria and Outcome

� What will you use as your testing criteria

and outcome measures?

� Consider both self-report and functional � Consider both self-report and functional

performance observer based measures.

Treatment Interventions

� Manual therapy to the lumbar spine –

mobilization, manipulation – what techniques

would you use?

� Is she a candidate for manipulation?

Consider the CPR for the lumbar spine

manipulation subgroup.

� Outline your exercise program for this patient

as well as educational advice.

Predictors of favourable outcome following spinal manipulation(Flynn et al 2002, Childs et al 2004)

Duration of symptoms < 16 days

At least one hip with internal rotation of > 35 degrees

Clinical Prediction Rule – Lumbar Spine

Lumbar hypomobility

No symptoms distal to the knee

FABQW score of <19

Flynn T. Fritz J. Whitman J. Wainner R. Magel J. et al A clinical prediction rule for classifying patients with low back pain who demonstrate short-term improvement with spinal manipulation. Spine 2002;27:2835-2843

Childs J. Fritz J. Flynn T. Irrgang J. Johnson K. Majkowski G. Delitto A. A clinical prediction rule to identify patients with low back pain most likely to benefit from spinal manipulation: A validation study. 2004 Annals of Internal Medicine 141:920-928

December 2009

The results support the hypothesis that the CPR is

generalizable to additional thrust manipulation techniques,

but not to nonthrust techniques.

Provided some support that that the CPR can be

generalized to different settings however further research

is needed.

Practical  Outline  for  LumboPelvic  Case  

Review:  

Assessment  AROM  for  lumbar  spine  

• Single  plane  • Combined  movements  –  Hand  I  testing  

 AROM  for  the  pelvis  

• As  above  for  lumbar  spine  • Functional  load  transfer  tests  

o One  leg  standing  test  (stork  test)  o Squat  o Lunge  o Gait  

• Fwd  bending  test    PROM  Lumbar  Spine  and  Pelvis  

• Uniplanar  and  combined  PIVMS  lumbar  spine  • Anterior  /  posterior  rotation  of  the  innominate    • Nutation  /  counternutation  of  the  sacrum  

 Pain  Provocation  tests  

• Lasletts  compression  and  distraction  • Fabers  • Thigh  thrust  (P4)  • Sacral  thrust  • Gaenslens  test  

 Accessory  motions  of  Lumbar  spine  

• Central  and  unilateral  PA  angled  cranial  and  caudal    or  

• Superior  anterior  glide  /  posterior  inferior  glide    Accessory  motions  for  the  Pelvis  

• Joint  play  assessment  either  using  innominate  rotation  or  sacrum  nutation  /  counternutation  inducing  the  glides.  Posterior  inferior  glide,  superior  anterior  glide  

 Stability  assessment  

• lumbar  spine  –  Anterior,  Posterior,  Torsion,  Lateral  shear  • SI  joint:  A/P  plane,  superior,  inferior  shear  

 Muscle  

• Review  muscle  length  assessment  for  the  trunk  and  pelvis  

• Review  ALSR  and  specific  muscle  strength  tests  for  the  lumbopelvic  region    Treatment  

• Anterior  rotation  mobilization  • Counternutation  of  the  sacrum  • Supine  distraction  manipulation  • Long  leg  traction  manipulation  of  the  SI  joint  • Flexion  mobilization  • Oblique  Gap  manipulation  • Review  core  muscle  stabilization  with  progressions  • Soft  tissue  techniques  and  muscle  stretching    • Home  exercise  prescription    

         

 

Clinical Reasoning Reflection Form – Subjective

Questions to be completed after the subjective assessment (see example)

1. Identify the key activities and participation restrictions for this client (based on the International Classification of Function, Disability and Health – ICF)

2. In what stage of the healing process does this client present (acute inflammatory – 0-3 days, proliferation/sub-acute phase 3 days – 6 weeks, remodelling/maturation – 6 weeks – 18 months)?

3. Identify the dominant pain mechanism (nociceptive, peripherally evoked, centrally

evoked) and provide the subjective evidence to support it.

Input Mechanisms Processing Mechanisms Output Mechanisms

Nociceptive Symptoms

Peripheral Evoked

Neurogenic Symptoms

Centrally Evoked

Neurogenic Symptoms

Patient’s perspective (Cognitive/ Affective

Influences)

Motor & Autonomic

Mechanisms

% of Influence

4. What are all the potential structures at fault for each area of symptoms?

5. Have you screened for red flags and/or identified any potential risk factors for

recovery?

6. Do the symptoms fit into a clinical pattern or sub-group? Explain.

7. Are there any environmental (ergonomic, work/home) or psychosocial factors

(patient perspectives, health and cultural beliefs) associated with the patients symptoms? Explain.

8. What is the clinically perceived level of irritability and what are the implications of

this?

9. Based on your subjective examination (and your assumptions) how would you formulate your objective examination? Please state the most relevant tests and the order you would perform them and why.

10. What outcome measurement tools would you use to monitor this patient and why

would you choose to use it?

Clinical Reasoning Reflection Form – Objective Questions to be completed after the objective assessment

1. Do the physical examination findings correlate with the degree of severity of the presenting symptoms?

2. Are there any objective findings which would indicate the need for caution in your

management?

3. Is there a need to refer this client to another health care professional (family doctor, orthopaedic surgeon, psychologist)?

4. What are three potential hypotheses and give the supporting (and/or negating) evidence from the subjective and objective examination?

5. Consider the prognostic indicators and state your predictive outcome for this client (i.e. length of time for recovery, percentage of recovery or do you expect residual findings)

6. Identify the key physical impairments from the objective examination i.e. posture,

movement patterns, joint dysfunction, motor control, neural mobility/sensitivity. 7. Indicate your treatment goals for your key impairments: Goals: Short term-

Long term- Management Strategies (include manual therapy, exercise, education, other): Initial Treatment: Outcome measures: 8. Identify the key subjective and objective findings that would help you recognize

this disorder (i.e. clinical pattern) in the future.

Subjective Objective

9. Is there evidence to support your assessment procedures and treatment interventions? Please elaborate.

Reflections for Prognosis, Assessment and Treatment Management: (may be completed on future visits or on reflection of a prior patient)

a. Are the client’s needs being met? b. What clues (if any) can you now recognize that you initially missed,

misinterpreted, under-or-over-weighted? c. What would you do different next time? d. Did your outcome measure change significantly over time?

I"uCDCLto0 Use of the ICF Model as a Clinical

Problem-Solving Tool in PhysicalTherapy and Rehabilitation Medicine

The authors developed an instrument called the "RehabilitationProblem-Solving Form" (RPS-Form), which allows health care profes-sionals analyze patient problems, to focus on specific targets, and torelate the salient disabilities to relevant and modifiable variables. Inparticular, the RPS-Form was designed to address the patients' per-spectives and enhance their participation in the decision-makingprocess. Because the RPS-Form is based on the International Classifica-tion of Functioning, Disability, and Health (ICF) Model of Functioningand Disability, it could provide a common language for the descriptionof human functioning and therefore facilitates multidisciplinaryresponsibility and coordination of interventions. The use of theRPS-Form in clinical practice is demonstrated by presenting an appli-cation case of a patient with a chronic pain syndrome. [Steiner WA,Ryser L, Huber E, et al. Use of the ICF model as a clinical problem-solving tool in physical therapy and rehabilitation medicine. Phys Ther.2002;82:1098-1107.]

Key Words: Communications; ICF; International Classification of Functioning, Disability, and Health;

Interprofessional relations; Patient-centered care; Problem solving.

Werner A Steiner, Liliane Ryser, Erika Huber, Daniel Uebelhart, Andre Aeschlimann, Gerold Stucki

1098 Physical Therapy . Volume 82 . Number 11 . November 2002

The most effective health care interventions forcomplex medical conditions, such as thoseencountered in people with chronic diseases,are thought by many authors to probably be

delivered by multidisciplinary care teams. This teammodel originates from the belief that a comprehensivetherapeutic approach is required to fully address thecurrent health care needs of patients with complex orchronic diseases.^^ Such integrated care, in our view,requires an exchange of information among all peopleinvolved in the therapeutic process. Multidisciplinaryhealth care thus necessitates tools that function acrossprofessional boundaries^ and that can handle differ-ences in perspectives (eg, those shovwi to exist betweenphysicians and nurses) .''• Health care professionals, aswell as their patients, may perceive specific needs anddisorders and their overall management quite different-ly 4.6-8 Dissimilar points of view regarding a patient's

A patient-centered evaluation toal is

needed in order to acknowledge the

views, experience, and perspectives of

all participants involved in the health

care process.

health care needs and goals can lead to inappropriatetreatment strategies, can hamper communication,^ andcan decrease the patient's adherence.^" In order to avoidcritical differences between the patient's and the healthcare professional's treatment goals, the goals need to beclarified prior to planning interventions.'^

Another important aspect is that the consequences ofdisease manifest differently in different people.Although many patients may have the same disease, theirresponses to disease can be unique, and these particulars

WA Steiner, PhD, is Scientific Collaborator, Department of Rheumatology and Institute of Physical Medicine, University Hospital Zurich,Gloriastrasse 25, CH-8091 Zurich, Switzerland (wsteiner@rehabnet.ch). Address all correspondence to Dr Steiner.

L Ryser, lie phil, is Psychologist, Department of Rheumatology and Institute of Physical Medicine, University Hospital, Zurich, Svntzerland.

E Huber, MHSA, is Clinical Leader of Physiotherapy and Occupational Therapy, Department of Rheumatology and Institute of Physical Medicine,University Hospital, Zurich, Switzerland.

D Uebelhart is Privatdozent, University of Geneva, and Leading Physician, Department of Rheumatology and Institute of Physical Medicine,University Hospital, Zurich, Switzerland.

A Aeschlimann, is Professor, Rheumatology and Rehabilitation Clinic, Zurzach, Svritzerland.

G Stucki, is Professor, Department of Physical Medicine and Rehabilitation, University of Munich, Munich, Germany.

Dr Steiner, Ms Ryser, Ms Huber, Mr Aeschlimann, and Mr Stucki provided concept/idea/design. Dr Steiner and Ms Huber provided writing. MsRyser provided data collection. Dr Steiner, Ms Ryser, and Mr Uebelhart provided data analysis. Dr Steiner and Mr Aeschlimann provided projectmanagement. Mr Aeschlimann provided fund procurement and clerical support. Ms Huber, Mr Uebelhart, and Mr Aeschlimann providedinstitutional liaisons. Mr Uebelhart and Mr Aeschlimann provided consultation (including review of manuscript before submission). The authorsthank Leanne Pobjoy for her help in preparing the manuscript and Professor Beat A Michel, Director of the Department of Rheumatology andInstitute of Physical Medicine, for his continuous support.

The Rehab-CYCLE project has been supported, in part, by an unrestricted educational grant from the Zurzach Rehabilitadon Foundation.

This article was submitted May 23, 2001, and was accepted May 15, 2002.

Physical Therapy . Volume 82 . Number 11 . November 2002 Steiner et al . 1099

Rehab-CYCLEidentify problems

and needs

Assess effects

1Pian, impiement,and

coordinate interventions

\

Relate problems tomodifiabie andlimiting fectors

Define target probiemsand target mediators,

seiect appropriate measures

Figure 1 .The Rehab-CYCLE is a modified version of the Rehabilitation Cycle developed by Stucki andSangha.''' It guides the health care professional with a logical sequence of activities. Endpointsof this rehabilitation management system are successful problem solving or individual goalsachieved. The Rehab-CYCLE involves identifying the patient's problems and needs, relating theproblems to relevant factors of the person and the environment, defining therapy goals, planningand implementing the interventions, and assessing the effects.

can become crucial in the care of patients.'^ Patient-centered practice is thought by some authors^^ toimprove health status and increase the efficiency of care.

To summarize our thoughts, a patient-centered evalua-tion tool is needed in order to acknowledge the views,experience, and perspectives of all participants involvedin the health care process. Ideally, such a tool shouldfulfill the clinical needs of both the patient and thehealth care team, should be simple to use, and shouldhave a background that can be accepted by all involvedpartners.

Based on the framework of the Rehabilitation Cycle(and its modified version, the Rehab-CYCLE) developedby Stucki and Sangha'"* (Fig. 1), we developed a furtherextension that we called the "Rehabilitation Problem-Solving Form" (RPS-Form) (Fig. 2). This form is used toidentify specific and relevant target problems, discernfactors that cause or contribute to these problems, andplan the most appropriate interventions. In addition, theRPS-Form was designed to be used as a tool to facilitateboth intraprofessional and interprofessional communi-cations and to improve the communication betweenhealth care professionals and their patients.

The aims of our article are to present the theoreticalconstruct that underlies the recently developed RPS-Form and to advocate its use in rehabilitation. The

Rehab-CYCLE is used as a frameworkto present this clinical problem-solving tool because the rehabilitationteam followed the different steps ofthis approach and because we believethe RPS-Form and the World HealthOrganization's (WHO) InternationalClassification of Functioning, Disability,and Health (ICF) Model of Function-ing and Disability's that underlies thisapproach (Fig. 3) are integrated inthe Rehab-CYCLE.

Materials and Methods

The Rehab-CYCLERehabilitation, in our view, is a con-tinuous process that involves identify-ing the problems and needs of indi-viduals, relating the problems torelevant factors of the person and theenvironment, defining therapy goals,planning and implementing the inter-ventions, and assessing the effects ofinterventions using measurements ofrelevant variables. To guide healthcare professionals in successful reha-

bilitation management, Stucki and Sangha "* developedthe Rehab-CYCLE. The ultimate goal of the Rehab-CYCLE is to improve a patient's health status and qualityof life by minimizing the consequences of disease.

The Rehab-CYCLE (Fig. 1) is a structured approach torehabilitation management that includes all tasks fromproblem analysis to the assessment of the effects, therebyinvolving the patient in clinical decision making. Theemphasis is on the patient's perspective (eg, throughpatient-rated questionnaires), taking into account thepatient's needs and preferences, and discussing therapygoals by means of the RPS-Form, which will be presentedin this article.

Because the consequences of disease manifest differentlyin different people, it is necessary to have a conceptualframework for ordering and understanding what diseasemeans to a patient. At the Institute of Physical Medicineof the University Hospital Zurich (Zurich, Switzerland),the WHO'S ICF Model of Functioning and Disability'^was recently implemented for this purpose by using theRPS-Form.

The ICFThe RPS-Form consists of a single data sheet that is basedon the ICF. The ICF classifies health and health-relatedcomponents (such as education and labor) that describe

1100 . Steiner et al Physical Therapy . Volume 82 . Number 1 1 . November 2002

body functions and structures, activi-ties, and participation.

The overall aim of the developers ofthe ICF was to provide a unified andstandard language and framework forthe description of all aspects of humanhealth and some health-relevantaspects of well-being.'^ The ICF pro-vides a structure to present this infor-mation in a meaningful, interrelated,and easily accessible way. The informa-tion is organized in 2 parts, with eachpart having 2 components. Part 1 ofthe ICF (functioning and disability)consists of (1) body functions andstructures and (2) activities and partic-ipation, and part 2 of the ICF (contex-tual factors) consists of (1) environ-mental factors and (2) personalfactors.

Pat.No., Disorder / Disease; Medication: ,

Coonlinator: i

Patient (or Relatives): P -oblems and Disabilitie

Body Structures / Functions o Activities

Health ProfessipqalsLMediatorsReievantJpJarge

Participation

t Problems

Personal Factors: Environmental Factors:

Each ICF component can be expressedin both positive and negative terms. Atone end of this scale are the terms thatindicate nonproblematic (ie, neutraland positive) aspects of health andhealth-related states, and at the other end are the termscan be used to indicate problems. Nonproblematicaspects of health are summarized under the umbrellaterm "functioning," whereas "disability" serves as anumbrella term for impairment, activity limitation, orparticipation restriction.

An ICF component consists of various domains and,within each domain, categories, which are the units ofthe ICF classification. All ICF categories are "nested" sothat broader categories are defined to include moredetailed subcategories of the parent category.

Health-related states of an individual are then recordedby selecting the appropriate category code or codes andthen adding qualifiers, which are numeric codes, andspecifying the extent or the magnitude of the function-ing or disability in that category or the extent to whichan environmental factor is a facilitator or barrier (fordetails, see the recently released full version of the

Figure 2.The Rehabilitation Problem-Solving Form (RPS-Form) is based on the International Classificationaf Functioning, Disability, and Health (ICF) Model of Functioning and Disability'^ (see Fig. 3).The main difference is that the RPS-Form is divided into 3 parts: (1) header for basic information,(2) upper part to describe the patient's perspective, and (3) lower part for the analysis of thehealth care professionals. Copyright 2000 by Dr Werner Steiner, Switzerland. Reprint allowedwith permission only.

"Health Condition"(disorder or disease)

Body Functionsand Structures

Activities

Personai EnvironmentaiFactors Factors

The ICF also provides a model of functioning anddisability, which reflects interactions between the com-ponents of the ICF (Fig. 3). The ICF Model of Function-ing and Disability^^ is a biopsychosocial model designedto provide a coherent view of various dimensions ofhealth at biological, individual, and social levels.

Figure 3.The International Classification of Functioning, Disability, and Health(ICF) Model of Functioning and Disability'^ visualizes the interactionsamong the various components in the "process" of functioning anddisability. The ICF provides a description of situations with regard tohuman functioning and disability and serves as a framework to organizeinformation. Functioning and disability ("Body Functions and Struc-tures," "Activities," and "Participation") are seen as an interactionbetween the "Health Condition" ("disorder/disease") and the contex-tual factors ("Personal Factors" and "Environmental Factors"). This figurehas been modified and reprinted with permission of the World HealthOrganization (WHO), and all rights are reserved by the Organization.

As illustrated in Figure 3, an individual's functioning ordisability in a specific domain represents an interactionbetween the "Health Condition" (eg, diseases, disorders,injuries, traumas) and the contextual factors (ie, "Envi-ronmental Factors" and "Personal Factors"). The inter-actions of the components in the model are in 2

Physical Therapy . Volume 82 . Number 11 . November 2002 Steiner etal . 1101

directions, and interventions in one component canpotentially modify one or more of the othercomponents.

The RPS-FormThe RPS-Form (Fig. 2) is constructed similarly to the ICFModel of Functioning and Disability (Fig. 3). Eachcomponent of the ICF model is graphically highlightedby a gray background. For instance, "Disorder/Disease"(or "Health Condition") is highlighted at the top of themodel; the main components of functioning or disabil-ities are highlighted in the middle of the model, with(left to right) body level ("Body Structures/Functions"),individual level ("Activities"), and societal level ("Partic-ipation"); and the contextual factors ("Personal Factors"and "Environmental Factors") are highlighted at thebottom of the model. As indicated in Figure 2 by the grayarrows pointing downward and upward, "Disorder/Dis-ease" as well as "Environmental Factors" and "PersonalFactors" may have an impact on all components of func-tioning and disabilities.

The RPS-Form is designed to distinguish between theperspectives held by the patient and those of the healthcare professional. The patient's view is recorded in theupper part of the form denoted with "Patient (or Rela-tives): Problems and Disabilities," and the health careprofessional's views are noted in the lower part denotedwith "Health Professionals: Mediators Relevant to TargetProblems." The header of the RPS-Form is reserved forbasic information: identification of the patient ("Pat.No."), form identification number ("Form No."), date("Date"), disorder/disease (eg, in words), current med-ication ("Medication") and case coordinator("Coordinator").

RPS-Form: Case of a Woman With Chronic PainA 49-year-old woman of Asian origin who had been livingin Switzerland for over 20 years, was married, had nochildren, and had worked for 10 years as a nurse wasreferred as an inpatient to the Department of Rheuma-tology and Institute of Physical Medicine at the Univer-sity Hospital Zurich for treatment for generalized pain-ful reactive arthritis. This referral followed an episode ofgastrointestinal infectious disease 2 years previously, witha positive stool identification of Yersinia enterocolitica asthe pathogenic agent. Upon entry, the patient had achronic pain syndrome that mostly affected her cervical(C5-C6 degenerative modifications) and thoracic spine.The patient also had pain at multiple locations such as inthe elbows, hands, knees, and feet. No additionalinvolvement of the axial skeleton could be found. Oneyear prior to referral as an inpatient she started receivingweekly injections of methotrexate (increasing the dosefrom 10 to 25 mg at time of referral), with hydroxychlo-roquine (2 X 200 mg/d) added after 7 months and

sulfasalazine (4 X 500 mg/d) added after 10 months.This mixed drug therapy did not alleviate the patient'ssymptoms.

At various times in the past, the patient received at theUniversity Hospital Zurich corticosteroid injections inboth feet and elbow (Kenacort* 20 mg/injection), whichtogether with numerous sessions of physical therapyhelped to reduce the symptoms. The patient's chronicsecondary depression was treated with antidepressiveagents (Surmontil^ 10 mg/d), which did not entirelyalleviate this condition. As a consequence of this chronicpain syndrome, the patient reduced her professionalactivity as a nurse to 60% 3 years before she was referredas an inpatient, and she stopped all professional activity2 years later.

The clinical laboratory investigations made at the begin-ning of her hospitalization were normal, with no indica-tion of any inflammatory or infectious activity, muscledegradation, or any other metabolic or biological abnor-mality. The diagnosis at the time of discharge from thehospital was chronic multifactorial pain syndrome withcervical and thoracic spondylarthritis and status afterreactive arthritis associated with secondary depression.The patient continued her basic medication, includingher antirheumatic drugs (weekly methotrexate injec-tions of 20 mg, hydroxychloroquine 2 X 200 mg/d). Thepatient was discharged from the bed unit after 2 weeksand then was admitted to the Interdisciplinary Outpa-tient Pain Program (IOPP), which is hosted at theDepartment of Rheumatology and Institute of PhysicalMedicine of the University Hospital Zurich.

Identification of problems and disabilities and reportingthem on the RPS-Form. According to Stucki andSangha,!* the identification of a patient's problems andneeds is the first step in rehabilitation management. Inthe case of our patient, a series of interviews were initiallyconducted with the rehabilitation team (physician, psy-chologist, physical therapist, and social worker). In addi-tion, questionnaires were used to comprehensively assessher experience with chronic pain. These questionnaireswere the Medical Outcomes Study 36-Item Short-FormHealth Survey (SF-36),i6 generic health status measure;the Hospital Anxiety and Depression Scale (HADS),'^which we used as a screening instrument for depressiveand anxiety disorders; and the Coping Strategies Ques-tionnaire (CSQ),i8 which we used to analyze thepatient's pain coping strategies.

* Bristol-Myers Squibb, La Grande Arche Nord, 92044 Paris, France.^ Wyeth-Ayerst Pharmaceuticals, Div of American Home Products Corp, PO Box8299, Philadelphia, PA 19101.

1102 . Steiner etal Physical Therapy . Volume 82 . Number 11 . November 2002

Pat.No.| 59834884.00176095 ,

Form-No.: L I J Date:i 28.07.2000

Disorder / Disease:Reactive arthritisChronic multifactoriai pain syndrome

i\/ledication: .NSAiD

Coordinator: Pr. M. Fisher

Patient (or Relatives): P

Often tired

Pain in neck, hands, feet

I

House keeping activities that invoiveiifting or carrying by hane.g. using a y a ^ l Sironing ci

Partial incapacity for work (60%)-^ Avoid sick leave

Anxious to iose her job

Missing social contacts previouslyfound in leisure dubs

Stopped to accompany herhusband in walking

Body Structures/Functions <-» Activities ParticipationMoferate impairment of joints:

inds, fingers s73021.2ikle, feet S75021.2

; Regulation of emotior^521.3I (anxiety and feelingsjn depression)

General physical b4550.2endurance

Hypertonia of neck b735.3Muscle power b7304.2

functions: arms / feet

alth Professionals: Medi

Difficutty in handlingstress + otherpsychological demands

a 10 r s Re I e ya n t t o Target Problems

Personal Factors: Environmental Factors:C[]_Copingstrategies

' " ^Social background -1

GemianlanguagePersonality

+3 Former medication-•-1 (pain killers)

e l 101.2

Figure 4.The Rehabilitation Problem-Solving Form (RPS-Form) applied to a patient with chronic pain. The form visualizes the current understanding of thepatient's state of functioning and disability, her target problems, and hov/ the health care professional team relates them to hypothetical mediatorsand contextual factors. NSAID=nonsteroidal anti-inflammatory drugs. Copyright 2000 by Dr Werner Steiner, Switzerland. Reprint allowed withpermission only.

The concerns of the patient, as compiled by the variousmembers of the health care team and supported by theanalysis of the initial SF-36, HADS, and CSQ results, werethen reported by the case coordinator on the upper partof the RPS-Form (Fig. 4). In order to avoid interpreta-tion that goes beyond the patient's statements, we arguethat it is essential to describe these concerns in thepatient's own words and to discuss these entries with thepatient.

As shown in Figure 4, the patient reported neck pain, aswell as pain in the hands and feet. She often felt tired,which she said prevented her from participating inleisure clubs as she had done 2 years before. Writing orhousekeeping activities that involved lifting and carryingobjects with the hands (eg, using a vacuum cleaner) werevery difficult tasks for her. Walking long distancesbecame almost impossible due to the pain in her feet.

and she could no longer join her husband on his walks.Above all, she was anxious about losing her job as a nurseas a consequence of the further degeneration of herhealth and that this would lead to financial dependencyon her husband.

Relate problems to relevant and modifiable factors. Sofar, the problem analysis that occurred was a compila-tion of the patient's problems and nieeds. Each specialistthen examined the patient, keeping in mind concernsstated by the patient on the RPS-Form.

Through this process, the rehabilitation team tried torelate these problems to impairments, activity limita-tions, participation restrictions, or personal and environ-mental factors. All team members were requested togenerate hypotheses about cause and effects. That is, therehabilitation team attempted to identify those charac-

Physical Therapy . Volume 82 . Number 11 . November 2002 Steiner etal . 1103

Figure 5.The concept of mediation as explained by Baron and Kenny." (A)Variable X is assumed to affect another variable (variable Y). VariableX is called the initial (or independent) variable, and the variable that itaffects (variable Y) is called the outcome variable. The direct impact ofthe independent variable is indicated by path c. (B) The effect of variableX on variable Y is mediated by a process or mediating variable (variableM), with path b indicating the impact of the mediator. The variable Xmay still affect variable Y (path c'). The mediator also has been called anintervening or process variable. Complete mediation can occur whenvariable X na longer affects variable Y (path c' = zero) after variable Mhas been controlled.

teristics of the patient or her environment that caused orcontributed to her problems, either directly or indirectlyby transmission. The multiple interactions betweenpatient and environment, and between all componentsof the patient's organism, require thinking in terms ofcausal networks, rather than in straight lines where Acauses B, which leads to C.'

Because it is often unclear whether a variable is directlyresponsible for a disability or whether it is a trigger thatreleases certain processes linked with the disability, theumbrella term "mediator" is used on the RPS-Form todescribe such variables. The concept of mediation'^ isexplained in Figure 5.

Mediators, as identified by the various members of therehabilitation team, are reported by the case coordina-tor on the lower part of the RPS-Form (Fig. 4). At thisstage, the RPS-Form is completed to be discussed at thenext interdisciplinary treatment team meeting.

Which terms should be used to denote mediators? Inorder to ensure a common language for interdiscipli-nary teams, we recommend that health care profession-als specify the mediators on the RPS-Form, listingcorresponding terms that are listed in the ICF'^(Fig. 4). Only the well-defined ICF items, we believe,can ensure consistency in the use of terminologyacross disciplines, and inconsistency can pose a bar-rier to effective communication.20

The ICF terms can be interpreted by means of 3 separatebut related constructs,'^ all using "qualifiers" for opera-tionalization. Body functions and body structures can bedescribed by a qualifier, with the negative scale used toindicate the extent or magnitude of an impairment(eg, the qualifier "s73021.2" can be used to indicatemoderate impairment of the joints of the hands andfingers). For the activities and participation component.

2 constructs are available: capacity and performance.'^The capacity qualifier describes an individual's ability toexecute a task or an action, and the performancequalifier describes what an individual does in his or hercurrent environment. Both qualifiers can be used withand without personal assistance or assistive devices. Forsimplicity, these 2 constructs are not differentiated fur-ther in this article. Therefore, the activities and partici-pation classification results in a single list of items,denoted by a leading "d." The item code "d240.3"(Fig. 4), for example, refers to the ICF item d240("Difficulty in handling stress and other psychologicaldemands"), without differentiating between activitiesand participation. The qualifier "3" denotes a severedifficulty in accomplishing this task, disregarding aspectsof capacity and performance.

Qualifiers also can be added to environmental factors. Adecimal point denotes a barrier (.1 to .4), whereas a plussign denotes a facilitator (-1-1 to -1-4).' For our patient(Fig. 4), the former medication (ie, chronic abuse ofpain killers [ellOl.2]), was considered a moderate bar-rier for her rehabilitation.

Although personal factors or resources are extremelyimportance in the rehabilitation process,^' they are notclassified in the ICF because of the large social andcultural variance associated with them.'^ j ^ our opinion,however, this should not hinder the rehabilitation teamin addressing personal factors relevant to the targetproblems or in describing their quantitative property inanalogy to the qualifier system applied to environmentalfactors (Fig. 4). In our case, personal factors consideredrelevant to problem solving were: command of theGerman language (-1-3 denotes a severe facilitator),personality ( + 1), social background (-1), and copingstrategies (-2).

Define target problems and target mediators on the RPS-Form, After clinical examination of the patient andcompilation of all limiting and modifiable mediators onthe RPS-Form, a revision process is needed to exchangeinformation within the rehabilitation team as well as withthe patient in order to define realistic therapy goals andto plan the most appropriate interventions. The clinicalexamination may have revealed underlying conditionsthat force the health care professional to set therapygoals that differ from the personal preferences andbeliefs of the patient. When using the Rehab-CYCLE torevise the problems mentioned by the patient (Fig. 4,upper part), there is a desire to meet the patient'sexpectations and to achieve his or her commitment, butalways taking into account practical and evidence-basedknowledge of the rehabilitation team (eg, aspects ofsecondary and tertiary prevention). Thus, this process ofdefining the target problems is usually the result of

1104 . Steiner et al Physical Therapy . Volume 82 . Number 1 1 . November 2002

consent between the patient and the health care team.The target problems are visualized on the RPS-Form bycircling, the corresponding items. In the case of ourpatient, they include the alleviation of pain in the neck,hands, and feet and the avoidance of sick leave (Fig. 4).

The importance of the mediators compiled by the healthcare professional on the RPS-Form might vary from lowto high, as does their potential to be modified during theintervention period. It is the "art of rehabilitation" todiscern the target mediators (ie, those mediators sup-posed to have the greatest potential to solve the targetproblems). This process generally takes place at theinterdisciplinary team meeting, where the RPS-Formserves as a basis for team members to discuss findingsand hypotheses in the framework of the ICF Model ofFunctioning and Disability. According to the targetproblems, the resulting target mediators are marked onthe RPS-Form by circling the corresponding items(Fig. 4). Lines can be then drawn to each of thecorresponding target problems (dark connecting linesin Fig. 4).

The case model. Once these hypothetical relationshipsare stated by the health care team, the RPS-Form repre-sents an explicit and interdisciplinary elaborated "casemodel," explaining by which mediators the target prob-lems can be solved. Because this case model is based onassumptions, the rehabilitation team must carefullyexplore associations or causal links between mediatorsand target problems during the intervention period.Individual therapy goals can now be formulated, usuallyincluding both target problems and target mediators,and the RPS-Form serves as an excellent tool for com-municating these goals to the patient.

The model of our case (Fig. 4) shows that the targetproblem "pain in neck, hands, feet" is mediated by 3pathways: (1) by mechanisms related to the body func-tions (ie, "general physical endurance," "hypertonia ofneck," and "muscle power functions: arms/feet"), (2) byan indirect mechanism (ie, "difficulty in handling stressand other psychological demands," and (3) by poorcoping strategies. Similarly, the target problem "partialincapacity for work (60%) -^ avoid sick leave" wasrelated to the same mediators except coping strategies.

Mediators considered by the health care professional tohave a low potential to solve the patient's problemsduring the treatment process are not directly included inthis initial case model (Fig. 4). However, their impor-tance might change with progress of the rehabilitationprocess.

Plan, implement, and coordinate interventions. The con-cept of the IOPP emphasizes the active participation of

the patient and a multidisciplinary team approach totreatment. According to the target mediators specifiedon the RPS-Form, the initial program for our patientincluded physical therapy to decrease the muscle con-tractures of the neck, to improve her general physicalendurance, and to increase the muscle force in her armsand feet. The psychological therapy focused on learningdaily living strategies to manage pain (ie, coping strate-gies) , to better handle stress (eg, at work), and to dealwith other demands (eg, her husband wanted her toaccompany him on long walks). Medical therapyincluded the use of nonsteroidal anti-infiammatorydrugs to manage her joint problems.

Assess effects of interventions. A rehabilitation manage-ment program for complex medical conditions, webelieve, needs a routine check of goal attainment bycomparing outcomes with target problems (Fig. 1,"Assess effects"). Use of qualifiers, as shown in Figure 4,allows the rehabilitation team to quantitatively monitorresults in longitudinal care. However, information aboutthe sensitivity and reliability of this new qualifier mea-surement system is still absent. A better solution tomeasure longitudinal changes of outcomes would be touse validated instruments such as psychometricallysound questionnaires and standardized clinical parame-ters.22 In the case of our patient, outcomes have beenmeasured by a battery of instruments. The correspond-ing assessment information was presented in the inter-disciplinary team meeting. The presentation of theseresults would go beyond the aims of this article and isdisregarded here.

According to the patient's progress in the rehabilitationprocess, it might be necessary to adjust treatment. Afterthe assessment of effects (Fig. 1), or when a full round ofthe Rehab-CYCLE is completed, the rehabilitation teamcompares changes of target problems and target media-tors with therapy goals. The degree of problem solving,among other topics, is then a key factor for the rehabil-itation team to decide whether a new "problem-solvingcycle" should be completed.

The Rehab-CYCLE is therefore an evolutionary andinteractive approach that implies continuous survey anda dynamic handling of all elements of the problem-solving process. Each RPS-Form represents a snapshotmodel of a patient's functioning and disabilities. Wetherefore advocate that, for every patient with complexhealth problems, several consecutive RPS-Forms shouldbe used in longitudinal care, these forms should becollected as comprehensive documentation of the treat-ment process, and this process should be related toobserved outcomes. Thereby, sound instruments such asinternationally validated, patient-rated questionnaires

Physical Therapy . Volume 82 . Number 1 1 . November 2002 Steiner etal . 1 105

can help the health care professional to measurewhether therapy goals are achieved.

Discussion and ConclusionThe RPS-Form described in this article has been appliedto many different health conditions (eg, cardiovasculardisease, neurologic problems such as hemiplegia, mus-culoskeletal problems such as arthritis and low backpain) at the Institute of Physical Medicine of the Uni-versity Hospital Zurich. This tool, we believe, is simple touse, helps to fully address the patients' perspectives, andserves as a platform where multidisciplinary care teamscan exchange information using a common language.The RPS-Form supports care teams in offering a visualrepresentation of the salient aspects of a disease, as wellas of the relationships between disabilities and underly-ing factors. Therefore, this tool also forms a basis fortreatment team meetings to discuss the individual goalsof the interventions. The underlying ICF Model ofFunctioning and Disability provides both the commonlanguage and the rational framework for the descriptionof health states associated with diseases and disorders.

The ICF (and the preceding ICIDH-2: International Clas-sification of Disability and Health^^) is the result of an effortthat started in 1993 and that focused on cross-culturaland multisectoral issues and involved the active partici-pation of 1,800 experts from 65 countries. ^ Studies havebeen undertaken in an effort to ensure that the ICF isapplicable across cultures, age groups, and genders, andit can be used to collect reliable and comparable data onhealth outcomes of individuals and populations. TheICF was accepted in November 2001 by 191 countries asthe international standard to describe and measurehealth and disability. At present, the ICF is available in 6languages (English, French, Spanish, Arabic, Chinese,and Russian); translations into other languages (eg, Cer-man) will follow in 2002. Because the ICF contains thecollective views of an international group of experts,^^ webelieve the RPS-Form permits international communica-tion about clinical problem solving at any level of health.

There are other major conceptual models that can guidehealth care professionals in understanding disabilitiesand functioning. Earlier conceptual models used in thesame context were reviewed by Jette.^* One of the firstmodels was developed by the sociologist Nagi. ^ Nagi'sclassification scheme varies from that of the WHO^ - ^primarily by suggesting the concept of functional limita-tions, that is, the physical manifestation of functionalproblems at the level of the organism as a whole.According to Nagi, ^ a functional limitation represents adirect way through which impairments contribute todisability. This conceptualization often is considereduseful for differentiating between performance-basedmeasures of function and self-reports, an important

aspect that was not explicitly integrated in the WHO'sInternational Classification of Impairments, Disabilities, andHandicaps (ICIDH).^^ However, this aspect is handled bythe ICF with the introduction of the concept of capacityand performance.

In the history of rehabilitation management at theInstitute of Physical Medicine of the University HospitalZurich, the first model used in physical therapy was themodel of Pope and Tarlov, ' which is an extension ofNagi's basic disablement formulation.^^ With progress inthe initial version of the ICIDH, ^ a revised version, theICIDH-2,2^ was introduced as a tool for thinking aboutand describing health and health-related states such asfunctioning and disabilities. The ICIDH-2 differs sub-stantially from the 1980 version of ICIDH in the depic-tion of the interrelationships between functioning anddisability. The ICIDH-2 was found to be useful in reha-bilitation, because the underlying model allows healthcare professionals to state the complex relationshipsbetween "Health Condition" (eg, diseases, disorders), thecomponents of health (body structures and functions,activities, and participation), and the contextual (ie, envi-ronmental and personal) factors. Along with the growinginternational acceptance of the ICF, the ICF Model ofFunctioning and Disability and the corresponding classifi-cation scheme are considered the future tools for organiz-ing information about functioning and disabilities.

Each model mentioned can be used to generate hypoth-eses about the interrelationships of different compo-nents included in the model. The key to successfulrehabilitation management, however, is understandingthe relationship between target problems and the com-ponents (impairments, functional limitations, and psy-chosocial and environmental factors) that affect themand addressing those (ie, the target mediators) with themost potential for improvement. In this process, theRehab-CYCLE is open to all ideologies of hypothesisgenerating, clinical reasoning, and decision making.

The Rehab-CYCLE (Fig. 1) is a structured approach torehabilitation management that should help to system-atically review disease consequences, to define therapygoals, to relate problems to mediators, and to optimizetreatment by relating interventions to results during therehabilitation process. It is thus similar to the hypothesis-oriented algorithm for clinicians (HOAC) described byRothstein and Echternach^s in that it guides the healthcare professional with a logical sequence of activities andrelies on the patient to describe his or her problems andon the health care professional to generate testablehypotheses. Both approaches are open to any treatmentstrategy. The main difference, we believe, is that theRehab-CYCLE is a more patient-centered approach witha biopsychosocial perspective.

1106 . Steiner et al Physical Therapy . Volume 82 . Number 11 . November 2002

In the case management at the Institute of PhysicalMedicine of the University Hospital Zurich, applying thenew and unfamiliar systematic coding scheme of ICFclassification initially hampered teamwork and commu-nication among the health care team. It is not thecorrect coding of ICF items, but the problem-solvingtechnique that can lead to better care for the patient.Therefore, when introducing the RPS-Form, health careprofessionals should feel free in how to describe healthand disability (eg, permit initial flexibility in wording,neglect alphanumeric ICF codes and "qualifiers"). Eventhis simple version of the RPS-Form trains health careprofessionals in proceeding through the Rehab-CYCLE,permits both health care professionals and patients tofocus on salient aspects of the disease, and assists withtreatment decision making shared by both the patientsand the health care professionals. Because understandingand motivation of the patient is usually a requirement forhis active involvement in the rehabilitation process, thesimple version of the RPS-Form could even be advanta-geous in the communication with certain patients.

After establishing the procedures associated with theRPS-Form, we believe the time may have come to intro-duce the standardized terms of the ICF classification. ^Use of these terms then can ensure consistency interminology across disciplines, improve interprofes-sional communication, and facilitate multidisciplinaryresponsibility and coordination of interventions in phys-ical therapy and rehabilitative medicine.

References1 Wagner EH. The role of patient care teams in chronic diseasemanagement. BMJ. 2000;320:569-572.

2 Minsky BD. Multidisciplinary case teams: an approach to the futuremanagement of advanced colorectal cancer. BrJ Cancer. 1998;77(suppl2):l-4.

3 Kole Snijders AM, Vlaeyen JW, Goossens ME, et al. Chronic low-backpain: what does cognitive coping skills training add to operant behav-ioral treatment? Results of a randomized clinical trial. / Consult ClinPsychol. 1999;67:931-944.

4Jacobson DH, Winograd CH. Psychoactive medications in the long-term care setting: differing perspectives among physicians, nursingstaff, and patients. J Geriatr Psychiatry Neurol. 1994;7:176-183.

5 Adamek ME, Kaplan MS. Caring for depressed and suicidal olderpatients: a survey of physicians and nurse practitioners. Int J PsychiatryMed. 2000;30:lll-125.

6 Donovan JL, Blake DR. Qualitative study of interpretation of reassur-ance among patients attending rheumatology clinics: "just a touch ofarthritis, doctor?" BMJ. 2000;320:541-544.

7 Potts M, Weinberger M, Brandt KD. Views of patients and providersregarding the importance of various aspects of an arthritis treatmentprogram. J Rheumatol. 1984;11:71-75.

8 Silvers IJ, Hovell MF, Weisman MH, Mueller MR. Assessing physi-cian/patient perceptions in rheumatoid arthritis: a vital component inpatient education. Arthritis Rheum. 1985;28:300-307.

9 Neville C, Fortin PR, Fitzcharles M, et al. The needs of patients witharthritis: the patient's perspective. Arthritis Care and Research. 1999;12:85-95.

10 Gopinath B, Radhakrishnan K, Sarma PS, etal. A questionnairesurvey about doctor-patient communication, compliance, and locus ofcontrol among south Indian people with epilepsy. Epilepsy Res. 2000;39:73-82.

11 Suarez Almazor ME, Conner Spady B, Kendall CJ, etal. Lack ofcongruence in the ratings of patients' health status by patients andtheir physicians. Med Deds Making. 2001;21:113-121.

12 McWhinney IR. Being a general practitioner: what it means. Euro-pean Journal of Family Practice. 2001;6:135-139.

13 Stewart M, Brown JB, Donner A, McWhinney IR, et al. The impactof patient<entered care on outcomes.//am Proc^ 2000;49:796-804.

14 Stucki G, Sangha O. Principles of rehabilitation. In: Klippel JH,Dieppe PA, eds. Rheumatology. 2nd ed. London, England: Mosby;1998:11.1-11.14.

15 International Classijication of Functioning, Disability, and Health (ICF).IGF full version. Geneva, Switzerland: World Health Organization;2001.

16 Ware JE Jr, Sherboume GD. The MOS 36-item short-form healthsurvey (SF-36), I: conceptual framework and item selection. Med Care.1992;30:473-483.

17 Zigmond AS, Snaith RP. The hospital anxiety and depression scale.Acta Psychiatr Scand. 1983;67:361-370.

18 Rosenstiel AK, Keefe FJ. The use of coping strategies in chronic lowback pain patients: relationship to patient characteristics and currentadjustment. Pain. 1983;17:33-44.

19 Baron RM, Kenny DA. The moderator-mediator variable distinctionin social psychological research: conceptual, strategic, and statisticalconsiderations. 7 Peri Soc Psychol. 1986;51:1173-1182.

20 Wanlass RL, Reutter SL, Kline AE. Communication among rehabil-itation staff: "mild," "moderate," or "severe" deficits? Arch Phys MedRehabii 1992;73:477-481.

21 Lorish GD, Abraham N, Austin J, et al. Disease and psychosocialfactors related to physical functioning in rheumatoid arthritis.JRheumatol. 1991;18:1150-1157.

22 Stucki G, Sangha O. Glinical quality management: putting thepieces together. Arthritis Care Res. 1996;9:405-412.

23 ICIDH-2: International Classification of Disability and Health. Prefinaldraft. Geneva, Switzerland: World Health Organization; 2000.

24Jette AM. Physical disablement concepts for physical therapyresearch and practice. Phys Ther. 1994;74:380-386.

25 Nagi S. Some conceptual issues in disability and rehabilitation.In: Sussman MB, ed. Sociology and Rehabilitation. Washington, DG:American Sociological Association; 1965:100-113.

26 ICIDH—International Classification of Impairments, Disabilities, andHandicaps: A Manual of Classification Relating to the Consequences ofDisease. Geneva, Switzerland: World Health Organization; 1980.

27 Pope AM, Tarlov AR. A model for disability and disability preven-tion. In: Pope AM, Tarlov AR, eds. Disability in America: Toward aNational Agenda for Prevention. Washington, DG: National AcademicPress; 1991:76-108.

28 ICIDH-2: International Classification of Functioning and Disability. Beta-2draft, fiill version. Geneva, Switzerland: World Health Organization; 1999.

29 Rothstein JM, Echtemach JL. Hypothesis-oriented algorithm forclinicians: a method for evaluation and treatment planning. Phys Ther.1986;66:1388-1394.

Physical Therapy . Volume 82 . Number 11 . November 2002 Steiner etal .1107

RDQ

Name: ________________________________________________ Date: _________________

Age: ______________ Score: _______________________

When your back hurts, you may find if difficult to do some of the things you normally do.

Mark only the sentences that describe you lately…. 1. [ ] I stay at home most of the time because of my back. 2. [ ] I walk more slowly than usual because of my back. 3. [ ] Because of my back, I am not doing any jobs that I usually do around the house. 4. [ ] Because of my back, I use a handrail to get upstairs. 5. [ ] Because of my back, I lie down to rest more often. 6. [ ] Because of my back, I have to hold onto something to get out of an easy chair. 7. [ ] Because of my back, I try to get other people to do things for me. 8. [ ] I get dressed more slowly than usual because of my back. 9. [ ] I stand up only for short periods of time because of my back. 10. [ ] Because of my back, I try not to bend or kneel down. 11. [ ] I find it difficult to get out of a chair because of my back. 12. [ ] My back or leg is painful almost all of the time. 13. [ ] I find it difficult to turn over in bed because of my back. 14. [ ] I have trouble putting on my socks (or stockings) because of pain in my back. 15. [ ] I sleep less well because of my back. 16. [ ] I avoid heavy jobs around the house because of my back. 17. [ ] Because of back pain, I am more irritable and bad tempered with people than usual. 18. [ ] Because of my back, I go upstairs more slowly than usual.

Roland Morris Disability Questionnaire

Scoring: Instructions for Roland-Morris :

The patient is instructed to put a mark next to each appropriate

statement. The total number of marked statements are added by the clinician.

Unlike the authors of the Oswestry Disability Questionnaire, Roland and Morris did not provide descriptions of the varying degrees of disability (e.g. 40%-60% is severe disability).

Clinical improvements over time can be graded based on the analysis of serial questionnaire scores. If, for example, at the beginning of treatment, a patient’s score was 12 and, at the conclusion of treatment, her score was 2 (10 points of improvement), we would calculate an 83% 910/12 x 100) improvement.

References

1. Deyo RA, Battie M, Beurskens AJ, Bombardier C, Croft P, Koes B, et al. Outcome measures for low back pain research. Spine 1998;23:2003-2013.

2. Roland M, Morris R. A study of the natural history of back pain: part I: development of a reliable and sensitive measure of disability in low-back pain. Spine 1983;8:141-144.

3. Deyo RA. Comparative validity of the sickness impact profile and shorter scales for functional assessment in low back pain. Spine 1986:11;951-0954.

4. Jensen MP, Strom SE, Turner JA, Romano JM. Validity of the Sickness Impact Profile Roland scale as a measure of dysfunction in chronic pain patients. Pain 1992:50;157-162.

5. Patrick DL, Deyo RA, Atlas SJ, Singer DE, Chapin A, Keller RB. Assessing health related quality of life in patients with sciatica. Spine 1995;20:1899-909.

6. Roberts A. The conservative treatment of low back pain. MD thesis, University of Nottingham, 1991.

7. Waddell G. The Back Pain Revolution. Edinburgh: Churchill Livingstone, 1998. 8. Baker CD, Pynsent PB, Fairbank JCT. The Oswestry Disability Index revisited:

its reliability, repeatability and validity, and a comparison with the St. Thomas's Disability Index. In: Roland MO, Jenner JR, eds. Back Pain: New Approaches to Education and Rehabilitation. Manchester University Press, 1989:174-86.

9. Stratford PW, Binkley JM. Measurement properties of the RM 18: a modified version of the Roland-Morris disability scale. Spine 1997;22:2416-2421.

10. CareTrak outcomes software. Grand Rapids, MN; (800) 393-7255, www.caretrak-outcomes.com.

Oswestry Disability Questionnaire This questionnaire has been designed to give us information as to how your back or leg pain is affecting your ability to manage in everyday life. Please answer by checking one box in each section for the statement which best applies to you. We realise you may consider that two or more statements in any one section apply but please just shade out the spot that indicates the statement which most clearly describes your problem.

Section 1: Pain Intensity

£ I have no pain at the moment £ The pain is very mild at the moment £ The pain is moderate at the moment £ The pain is fairly severe at the moment £ The pain is very severe at the moment £ The pain is the worst imaginable at the moment

Section 2: Personal Care (eg. washing, dressing)

£ I can look after myself normally without causing extra pain

£ I can look after myself normally but it causes extra pain £ It is painful to look after myself and I am slow and careful £ I need some help but can manage most of my personal

care £ I need help every day in most aspects of self-care £ I do not get dressed, wash with difficulty and stay in bed

Section 3: Lifting

£ I can lift heavy weights without extra pain £ I can lift heavy weights but it gives me extra pain £ Pain prevents me lifting heavy weights off the floor but I

can manage if they are conveniently placed eg. on a table £ Pain prevents me lifting heavy weights but I can manage

light to medium weights if they are conveniently positioned

£ I can only lift very light weights £ I cannot lift or carry anything

Section 4: Walking*

£ Pain does not prevent me walking any distance £ Pain prevents me from walking more than 2 kilometres £ Pain prevents me from walking more than 1 kilometre £ Pain prevents me from walking more than 500 metres £ I can only walk using a stick or crutches £ I am in bed most of the time

Section 5: Sitting

£ I can sit in any chair as long as I like £ I can only sit in my favourite chair as long as I like £ Pain prevents me sitting more than one hour £ Pain prevents me from sitting more than 30 minutes £ Pain prevents me from sitting more than 10 minutes £ Pain prevents me from sitting at all

Section 6: Standing

£ I can stand as long as I want without extra pain £ I can stand as long as I want but it gives me extra pain £ Pain prevents me from standing for more than 1 hour £ Pain prevents me from standing for more than 30

minutes £ Pain prevents me from standing for more than 10

minutes £ Pain prevents me from standing at all

Section 7: Sleeping

£ My sleep is never disturbed by pain £ My sleep is occasionally disturbed by pain £ Because of pain I have less than 6 hours sleep £ Because of pain I have less than 4 hours sleep £ Because of pain I have less than 2 hours sleep £ Pain prevents me from sleeping at all

Section 8: Sex Life (if applicable)

£ My sex life is normal and causes no extra pain £ My sex life is normal but causes some extra pain £ My sex life is nearly normal but is very painful £ My sex life is severely restricted by pain £ My sex life is nearly absent because of pain £ Pain prevents any sex life at all

Section 9: Social Life

£ My social life is normal and gives me no extra pain £ My social life is normal but increases the degree of pain £ Pain has no significant effect on my social l ife apart from limiting my more energetic interests e.g. sport £ Pain has restricted my social life and I do not go out as

often £ Pain has restricted my social life to my home £ I have no social life because of pain

Section 10: Travelling

£ I can travel anywhere without pain £ I can travel anywhere but it gives me extra pain £ Pain is bad but I manage journeys over two hours £ Pain restricts me to journeys of less than one hour £ Pain restricts me to short necessary journeys under 30

minutes £ Pain prevents me from travelling except to receive

treatment

Score: / x 100 = % Scoring: For each section the total possible score is 5: if the first statement is marked the section score = 0, if the last statement is marked it = 5. If all ten sections are completed the score is calculated as follows: Example: 16 (total scored) 50 (total possible score) x 100 = 32% If one section is missed or not applicable the score is calculated: 16 (total scored) 45 (total possible score) x 100 = 35.5% Minimum Detectable Change (90% confidence): 10%points (Change of less than this may be attributable to error in the measurement) Source: Fairbank JCT & Pynsent, PB (2000) The Oswestry Disability Index. Spine, 25(22):2940-2953.

Davidson M & Keating J (2001) A comparison of five low back disability questionnaires: reliability and responsiveness. Physical Therapy 2002;82:8-24.

*Note: Distances of 1mile, ½ mile and 100 yards have been replaced by metric distances in the Walking section.

Neck Disability Index

Source: Vernon H, Mior S. The Neck Disability Index: a study of reliability and validity. J Manipulative Physiol Ther. 1991 Sep;14(7):409-15.

Neck disorders are a significant source of pain and activity limitation in workers and those involved in motor vehicle collisions. The Neck Disability Index (NDI) [1] is designed to measure neck-specific disability. The questionnaire has 10 items concerning pain and activities of daily living including personal care, lifting, reading, headaches, concentration, work status, driving, sleeping and recreation. The measure is designed to be given to the patient to complete, and can provide useful information for management and prognosis of those with neck pain.

Scoring and interpretation Each item is scored out of five (with the no disability response given a score of 0) giving a total score for the questionnaire out of 50. Higher scores represent greater disability. The result can be expressed as a percentage (score out of 100) by doubling the total score.

The ‘Clinical guidelines for best practice management of acute and chronic whiplash-associated disorders’ [2] indicate that about 40% of patients with whiplash recover in less than four weeks, and that by six weeks about 50% have recovered. The guidelines recommend the use of the NDI to screen for risk factors and evaluate treatment effectiveness. An NDI score of >40/100 at initial assessment (first consultation following an injury) is associated with ongoing pain and disability after whiplash. This can alert a practitioner to the potential need for more regular review, or early referral to a specialised health provider such as a physiotherapist, chiropractor or psychologist. The guidelines indicate that ‘recovery’ is represented by an NDI score of less than 8/100, at which time treatment should be ceased.

Page 1

Neck Disability Index

Page 2

Neck Disability Index

Instructions

This questionnaire has been designed to give your health practitioner information as to how your neck pain has affected your ability to manage in everyday life. Please answer every section and mark in each section only the ONE box which applies to you. We realise you may consider that two of the statements in any one section relate to you, but please just mark the box which most closely describes your problem.

Section 1 – Pain intensity

I have no pain at the moment.

The pain is very mild at the moment.

The pain is moderate at the moment.

The pain is fairly severe at the moment.

The pain is very severe at the moment.

The pain is the worst imaginable at the moment.

Section 2 – Personal care (washing, dressing)

I can look after myself normally without causing extra pain.

I can look after myself normally but it causes extra pain.

It is painful to look after myself and I am slow and careful.

I need some help but manage most of my personal care.

I need help every day in most aspects of self-care.

I do not get dressed, I wash with difficulty and stay in bed.

Section 3 – Lifting

I can lift heavy weights without extra pain.

I can lift heavy weights but it gives extra pain.

Pain prevents me from lifting heavy weights off the floor, but I can manage if they are conveniently positioned, for example on a table.

Pain prevents me from lifting heavy weights, but I can manage light to medium weights if they are conveniently positioned.

I can lift very light weights.

I cannot lift or carry anything at all.

Section 4 – Reading

I can read as much as I want to with no pain in my neck.

I can read as much as I want to with slight pain in my neck.

I can read as much as I want with moderate pain in my neck.

I cannot read as much as I want because of moderate pain in my neck.

I can hardly read at all because of severe pain in my neck.

I cannot read at all.

Neck Disability Index

Section 5 – Headaches

I have no headaches at all.

I have slight headaches which come infrequently.

I have moderate headaches which come infrequently.

I have moderate headaches which come frequently.

I have severe headaches which come frequently

I have headaches almost all the time.

Section 6 – Concentration

I can concentrate fully when I want to with no difficulty.

I can concentrate fully when I want to with slight difficulty.

I have a fair degree of difficulty in concentrating when I want to.

I have a lot of difficulty in concentrating when I want to.

I have a great deal of difficulty in concentrating when I want to.

I cannot concentrate at all.

Section 7 – Work

I can do as much work as I want to.

I can only do my usual work, but no more.

I can do most of my usual work, but no more.

I cannot do my usual work.

I can hardly do any work at all.

I cannot do any work at all.

Section 8 – Driving

I can drive my car without any neck pain.

I can drive my car as long as I want with slight pain in my neck.

I can drive my car as long as I want with moderate pain in my neck.

I cannot drive my car as long as I want because of moderate pain in my neck.

I can hardly drive at all because of severe pain in my neck.

I cannot drive my car at all.

Section 9 – Sleeping

I have no trouble sleeping.

My sleep is slightly disturbed (less than 1 hr sleepless).

My sleep is mildly disturbed (1-2 hrs sleepless).

My sleep is moderately disturbed (2-3 hrs sleepless).

My sleep is greatly disturbed (3-5 hrs sleepless).

My sleep is completely disturbed (5-7 hrs sleepless).

Section 10 – Recreation

I am able to engage in all my recreation activities with no neck pain at all.

I am able to engage in all my recreation activities, with some pain in my neck.

I am able to engage in most, but not all of my usual recreation activities because of pain in my neck.

I am able to engage in a few of my usual recreation activities because of pain in my neck.

I can hardly do any recreation activities because of pain in my neck.

I cannot do any recreation activities at all.

Page 3

Neck Disability Index

References 1. Vernon H, Mior S. The Neck Disability Index: a study of reliability and validity. J Manipulative Physiol

Ther 1991 Sep;14(7):409-15.

2. TRACsa Trauma Injury and Recovery. Clinical guidelines for best practice management of acute and chronic whiplash-associated disorders. Canberra: National Health and Medical Research Council; 2008.

Page 4

The Patient-Specific Functional Scale This useful questionnaire can be used to quantify activity limitation and measure functional outcome for patients with any orthopaedic condition.

Clinician to read and fill in below: Complete at the end of the history and prior to physical examination.

Initial Assessment:

I am going to ask you to identify up to three important activities that you are unable to do or are having difficulty with as a result of your _________________ problem. Today, are there any activities that you are unable to do or having difficulty with because of your _________________ problem? (Clinician: show scale to patient and have the patient rate each activity).

Follow-up Assessments:

When I assessed you on (state previous assessment date), you told me that you had difficulty with (read all activities from list at a time). Today, do you still have difficulty with: (read and have patient score each item in the list)?

Patient-specific activity scoring scheme (Point to one number):

0 1 2 3 4 5 6 7 8 9 10

(Date and Score)

Activity Initial

1. 2. 3. 4. 5. Additional Additional

Total score = sum of the activity scores/number of activities Minimum detectable change (90%CI) for average score = 2 points Minimum detectable change (90%CI) for single activity score = 3 points

PSFS developed by: Stratford, P., Gill, C., Westaway, M., & Binkley, J. (1995). Assessing disability and change on individual patients: a report of a patient specific measure. Physiotherapy Canada, 47, 258-263.

Reproduced with the permission of the authors.

Unable to perform activity

Able to perform activity at the same level as before injury or problem

Screening for Elevated Levels of Fear-Avoidance Beliefs Regarding Work orPhysical Activities in People ReceivingOutpatient TherapyDennis L. Hart, Mark W. Werneke, Steven Z. George, James W. Matheson,Ying-Chih Wang, Karon F. Cook, Jerome E. Mioduski, Seung W. Choi

Background. Screening people for elevated levels of fear-avoidance beliefs isuncommon, but elevated levels of fear could worsen outcomes. Developing shortscreening tools might reduce the data collection burden and facilitate screening,which could prompt further testing or management strategy modifications to im-prove outcomes.

Objective. The purpose of this study was to develop efficient yet accurate screen-ing methods for identifying elevated levels of fear-avoidance beliefs regarding workor physical activities in people receiving outpatient rehabilitation.

Design. A secondary analysis of data collected prospectively from people with avariety of common neuromusculoskeletal diagnoses was conducted.

Methods. Intake Fear-Avoidance Beliefs Questionnaire (FABQ) data were col-lected from 17,804 people who had common neuromusculoskeletal conditions andwere receiving outpatient rehabilitation in 121 clinics in 26 states (in the UnitedStates). Item response theory (IRT) methods were used to analyze the FABQ data,with particular emphasis on differential item functioning among clinically logicalgroups of subjects, and to identify screening items. The accuracy of screening itemsfor identifying subjects with elevated levels of fear was assessed with receiveroperating characteristic analyses.

Results. Three items for fear of physical activities and 10 items for fear of workactivities represented unidimensional scales with adequate IRT model fit. Differentialitem functioning was negligible for variables known to affect functional statusoutcomes: sex, age, symptom acuity, surgical history, pain intensity, condition se-verity, and impairment. Items that provided maximum information at the median forthe FABQ scales were selected as screening items to dichotomize subjects by highversus low levels of fear. The accuracy of the screening items was supported for bothscales.

Limitations. This study represents a retrospective analysis, which should bereplicated using prospective designs. Future prospective studies should assess thereliability and validity of using one FABQ item to screen people for high levels offear-avoidance beliefs.

Conclusions. The lack of differential item functioning in the FABQ scales in thesample tested in this study suggested that FABQ screening could be useful in routineclinical practice and allowed the development of single-item screening for fear-avoidance beliefs that accurately identified subjects with elevated levels of fear.Because screening was accurate and efficient, single IRT-based FABQ screening itemsare recommended to facilitate improved evaluation and care of heterogeneous pop-ulations of people receiving outpatient rehabilitation.

D.L. Hart, PT, PhD, is Director ofConsulting and Research, FocusOn Therapeutic Outcomes, Inc,PO Box 11444, Knoxville, TN37939 (USA). Address all corre-spondence to Dr Hart at: hart@fotoinc.com.

M.W. Werneke, PT, MS, Dip MDT,is Physical Therapist, Spine Reha-bilitation at CentraState MedicalCenter, Freehold, New Jersey.

S.Z. George, PT, PhD, is AssociateProfessor, Department of PhysicalTherapy, Center for Pain Researchand Behavioral Health, BrooksCenter for Rehabilitation Studies,University of Florida, Gainesville,Florida.

J.W. Matheson, PT, DPT, MS, SCS,OCS, CSCS, is Physical Therapist,Minnesota Sport and Spine Reha-bilitation, Burnsville, Minnesota.

Y.-C. Wang, OT, PhD, is ResearchAssistant, Focus On TherapeuticOutcomes, Inc, Knoxville, Tennes-see, and Postdoctoral Fellow, Re-habilitation Institute of Chicago,Chicago, Illinois.

K.F. Cook, PhD, is Research Asso-ciate Professor, Department ofRehabilitation Medicine, Univer-sity of Washington, Seattle,Washington.

J.E. Mioduski, MS, is Programmer,Focus On Therapeutic Outcomes,Inc, Knoxville, Tennessee.

S.W. Choi, PhD, is Research Assis-tant Professor, Department ofMedical Social Sciences and Cen-ter on Outcomes, Research andEducation, Feinberg School ofMedicine, Northwestern Univer-sity, Chicago, Illinois.

[Hart DL, Werneke MW, GeorgeSZ, et al. Screening for elevatedlevels of fear-avoidance beliefs re-garding work or physical activitiesin people receiving outpatienttherapy. Phys Ther. 2009;89:770–785.]

© 2009 American Physical TherapyAssociation

Research Report

Post a Rapid Response orfind The Bottom Line:www.ptjournal.org

770 f Physical Therapy Volume 89 Number 8 August 2009

Clinicians and researchers haverecognized the role that psy-chosocial factors play in the

development of chronic disability inpeople with low back pain.1–3

Among the psychosocial risk factorsare fear-avoidance beliefs,4 whichare embodied in the fear-avoidancemodel of musculoskeletal pain.5 Themodel posits that an individual’s re-sponse to an episode of pain fallsalong a continuum ranging fromavoidance (maladaptive) to confron-tation (adaptive) and provides oneexplanation for why some peoplewith acute low back pain syndromesdevelop chronic disability.6–9

On the basis of theories of fear andavoidance of activities, Waddell et al4

developed the Fear-Avoidance Be-liefs Questionnaire (FABQ) to assessthe association between fear-avoidance beliefs and work disabilityfor people with chronic low backpain syndromes. The FABQ is a self-report questionnaire with 2 scales: 1assessing fear-avoidance beliefs re-garding work activities (FABQ-W)and 1 assessing fear-avoidance be-liefs regarding physical activities(FABQ-PA).4 Evidence supported anassociation between fear-avoidancebeliefs regarding work and absencefrom work because of low backpain.4 Thus, Waddell et al4 recom-mended that clinicians considerscreening for fear-avoidance beliefswhen managing low back pain. Sub-sequent studies indicated that ele-vated levels of fear were associatedwith10,11 and were predictive of12,13

disability and absence from work inpeople with low back and cervicalspine pain syndromes. There is evi-dence that identifying people withelevated levels of fear-avoidance be-liefs and managing those beliefs ac-cordingly may reduce fear and pre-dict or improve outcomes.1,4,5,10,12–22

Fear-avoidance beliefs may affectpeople with conditions other thanlow back pain. Evidence5,23 sup-

ported the possible existence of fear-avoidance beliefs or pain-related fearin people who have other impair-ments or who may not have pain,perhaps because of learned behaviorafter previous painful episodes ormisconceptions about pain.24 Pain-related fear scales, including theFABQ scales, have been used to as-sess the levels of fear in people withacute16 and chronic4,10,11 low backpain syndromes, cervical spine painsyndromes,11,25–27 cervical spine andshoulder28 pain syndromes, hip im-pairments,29 knee impairments,29–31

chronic headache,32 fibromyalgia,33

and chronic fatigue syndrome.33 It isreasonable to believe that pain-related fear would be applicable topeople with other conditions includ-ing, but not limited to, osteoarthri-tis,34 knee impairments,30,31 andneuropathic pain.35 These studiessuggested that pain-related fear is notuncommon in people with a widevariety of neuromusculoskeletal con-ditions, with and without pain, andanother study reported the preva-lence of elevated levels of fear-avoidance beliefs to be more than40% in specific samples.36

George17 described several screen-ing methods designed to identifypeople with elevated levels of fear,including the FABQ. Despite theavailability of these methods, thera-pists do not routinely screen for ele-vated levels of fear, a fact that may beattributable partly to the burden ofcollecting data or the difficulty in in-terpreting measures. In response tothese concerns, George17 challengedclinicians and researchers to refinescreening techniques by makingthem more efficient and accurate totry to improve acceptance and clini-cal use. Developing efficient and ac-curate screening methods is particu-larly important for therapistsassuming first-contact roles in pa-tient care,37 who need to identifyconfounding conditions that couldreduce the effectiveness of their

management strategies in diverse pa-tient populations.38 Screening resultsindicating elevated levels of fearwould alert therapists to the likeli-hood that patients might be fearful ofactivities that might be part of theirtherapeutic interventions; such a sit-uation might portend worse out-comes.39 Because short tests com-monly are associated with increasedmeasurement error,40 definitive test-ing often is recommended to con-firm the presence of the condition.41

Given that there is preliminary evi-dence of effective interventions forpeople with elevated levels of pain-related fear,19,42 the challenge ap-pears to be relevant to improved pa-tient care and outcomes.

In an effort to minimize the measure-ment error related to short tests,some authors have recommendedmodern psychometric techniques,such as item response theory (IRT)methods.43 Such methods are usefulfor assessing patient-report screen-ing surveys because they facilitateboth the evaluation of whether itemsmean the same thing to different re-spondents (ie, differential item func-tioning [DIF], described in theMethod section)44 and the identifica-tion of screening items by use ofitem information functions (de-scribed in the Method section).45

The absence of DIF is important ifFABQ scales are to be used to screendiverse populations for elevated lev-els of fear-avoidance beliefs. The use

Available WithThis Article atwww.ptjournal.org

• The Bottom Line clinicalsummary

• The Bottom Line Podcast

• Audio Abstracts Podcast

This article was published ahead ofprint on June 18, 2009, atwww.ptjournal.org.

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 771

of item information functions facili-tates the selection of single screen-ing items associated with the lowestmeasurement error related to a givenlevel of fear.45

The overall purpose of this study wasto develop an efficient yet accuratescreening method for identifyingpeople who have elevated levels offear-avoidance beliefs regardingwork or physical activities and whoare receiving outpatient rehabilita-tion. The specific purposes were: (1)to use IRT methods to analyze FABQitems, with particular emphasis onDIF among clinically logical groupsof people and identify screeningitems for each FABQ scale and (2) toassess the accuracy of screeningitems for identifying people with el-evated levels of fear-avoidance be-liefs. If the results suggest thatscreening items can identify peoplewith elevated levels of fear accu-rately, then more-precise fear-avoidance testing could be initiatedor management strategies could beused to reduce fear and improve out-comes. In addition, accurate screen-ing would reduce costly testing ofpeople not likely to be at risk of hav-ing elevated levels of fear.

MethodDesignWe conducted a secondary analysisof data collected prospectively frompeople with a variety of commonneuromusculoskeletal diagnoses.

Setting and ParticipantsWe analyzed data from 17,804 peo-ple (a sample of convenience)treated for common neuromusculo-skeletal conditions in 121 outpatientrehabilitation clinics in 26 states (inthe United States) between May2002 and December 2006 (Tab. 1).Clinics were participating with Fo-cus On Therapeutic Outcomes, Inc(Knoxville, Tennessee), an interna-tional medical rehabilitation data-base management company.46,47

Table 1.Characteristics of Subjects at Rehabilitation Intake (N�17,804)

Characteristic Value Characteristic Value

Clinical impairment groupingsof subjects (%)a

Surgical history (%)

Medical conditions 0.9 None 73.5

Neurological conditions 4.2 1 18.1

Orthopedic conditions 94.9 2 4.6

Upper extremity 25.5 3 1.5

Shoulder 70.3 4 or more 1.6

Elbow 8.9 Missing 0.7

Wrist or hand 20.8 No. of functionalcomorbidities (%)

Lower extremity 29.3 None 9.5

Hip 27.4 1 12.6

Knee 46.9 2 10.8

Foot or ankle 25.7 3 or more 21.4

Spine 45.2 Missing 45.7

Cervical 30.4 Pain intensityb

Lumbar 69.6 X (SD) 5.7(2.4)

Missing 0 Median 6

Age, y Minimum 0

X (SD) 50.0(16.1)

Maximum 10

Median 49 % at median or below 37.1

Minimum 18 % above median 24.7

Maximum 96 Missing 38.2

18–�45 (%) 38.1 Fear-avoidance scale scoresc

45–�65 (%) 42.1 Work activities (n�5,517)

65–�75 (%) 11.6 X (SD) 16.2(13.4)

75 or older (%) 8.2 Median 15

Missing (%) 0 Minimum 0

Sex (% female) 62.0 Maximum 42

Missing 0 Physical activities(n�16,243)

Acuity of symptoms (%) X (SD) 13.6(6.4)

Acute (0–21 d) 17.4 Median 14

Subacute (22–90 d) 28.2 Minimum 0

Chronic (�90 d) 53.7 Maximum 24

Missing 0.7 Missing 0

a Clinical impairment groupings of subjects. First: general—medical, neurological, and orthopedic;second: if orthopedic—upper extremity, lower extremity, and spine; third: if orthopedic and upperextremity—shoulder, elbow, and wrist or hand; fourth: if orthopedic and lower extremity—hip, knee,and foot or ankle; fifth: if orthopedic and spine—cervical and lumbar. Values in each of the 5groupings sum to 100%.b Pain intensity was rated from 0 to 10 with a numeric rating scale.c Summative scores from original scales.4

Screening for Fear of Work and Physical Activities

772 f Physical Therapy Volume 89 Number 8 August 2009

People were selected from the data-base of Focus On Therapeutic Out-comes because they had answeredthe FABQ for physical or work activ-ities (see below): 16,243 people hadanswered the FABQ for physical ac-tivities, 5,517 people had answeredthe FABQ for work activities, and3,956 people had answered both thework activity and the physical activ-ity surveys. Although diagnostic in-formation was available for only 68%of the people, the most prevalentgroupings of ICD-9-CM codes48 wererelated to soft-tissue disorders ofmuscle, synovium, tendon, or bursa(ICD-9-CM codes 725–729; 25% ofpeople) and pathologies of the spine(ICD-9-CM codes 720–724; 18% ofpeople). Most people were receivingpayment benefits from health main-tenance organizations (17%), pre-ferred provider organizations (11%),workers’ compensation (10%), andMedicare Part B (9%). Data on payerswere missing for 38% of the people.

Data CollectionAs described previously,46,49–52 datawere collected by use of Patient In-quiry computer software,* whichparticipating clinics used for routinecollection of data as part of their pa-tient care strategies. People seekingrehabilitation provided demographicdata before the initial evaluationand functional status informationthrough the use of condition-specificcomputerized adaptive tests49–55 atthe initial evaluation (intake) and atthe end of rehabilitation (discharge).Therapists also could elect to ask pa-tients to complete the FABQ physicalsubscale, the FABQ work subscale,or both; when selected, these sur-veys were administered via com-puter at intake and at discharge (butwere not computerized adaptivetests). Clinical staff entered demo-

graphic data at intake and at dis-charge. Only intake data wereanalyzed.

Fear-Avoidance Beliefs ItemsThe items in the FABQ describe therelationship between pain and phys-ical activities or work activities; forexample: “Physical activity mightharm my back” or “I cannot do mynormal work with my present pain.”For each item, a scale with ratings of0 to 6 (0�“completely disagree,”3�“unsure,” and 6�“completelyagree”) is used. There are no worddescriptors for responses 1, 2, 4, and5. Responses from 4 items aresummed to produce a score repre-sentative of the level of fear of phys-ical activities, and responses from 7items are summed to produce ascore representative of the level offear of work activities.4 Researchfindings have supported good iteminternal consistency and reliabilityand the presence of 2 factors in the

* Focus On Therapeutic Outcomes, Inc, POBox 11444, Knoxville, TN 37939-1444 (Website: www.fotoinc.com).

Table 2.Fear-Avoidance Belief Item Banks and Item Parameter Estimates

Scale andItema Slopeb (SE) Locationc (SE) Imax

dCategory

Parametere 1Category

Parametere 2Category

Parametere 3Category

Parametere 4

Physical

HARM 1.07 (0.01) 0.12 (0.01) 0.80 1.19 0.77 �0.70 �1.26

SHLDNOT 2.53 (0.02) �0.17 (0.01) 3.17 0.87 0.49 �0.48 �0.88

CANNOT 1.98 (0.02) �0.01 (0.01) 2.62 0.95 0.46 �0.47 �0.94

Work

WORK 0.91 (0.02) �0.18 (0.02) 0.67 0.59 0.37 �0.21 �0.74

COMPENS 0.89 (0.02) 0.86 (0.02) 0.61 0.32 0.27 �0.25 �0.34

WRKHVY 1.54 (0.03) 0.70 (0.01) 1.91 0.53 0.30 �0.25 �0.58

WRKWRSE 1.52 (0.03) �0.02 (0.01) 2.02 0.63 0.31 �0.21 �0.73

WRKHARM 1.29 (0.02) 0.41 (0.01) 1.35 0.62 0.39 �0.29 �0.72

WRKSHNT 2.53 (0.05) 0.32 (0.09) 4.79 0.49 0.29 �0.28 �0.50

WRKCANT 2.52 (0.05) 0.15 (0.01) 5.48 0.46 0.17 �0.16 �0.45

TREATED 2.17 (0.04) 0.20 (0.01) 3.96 0.45 0.19 �0.17 �0.47

MONTHS 1.19 (0.02) 0.95 (0.02) 1.03 0.70 0.52 �0.54 �0.68

GOBACK 1.05 (0.02) 1.46 (0.02) 1.84 0.62 0.44 �0.46 �0.60

a For item labels, see Appendix. Items in bold type were selected as screening items because their item information functions were the most informative atthe median theta value for each scale.b Slope�item discrimination parameter, which is unitless; see text for definition.c Location�item difficulty parameter in logits; see text for definition.d Imax�maximum item information at item location in Fisher information units.e Category parameter from graded-response model in logits; see text for definition.

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 773

FABQ (fear of work activities andfear of physical activities),4 FABQmeasure test-retest reliability,4 andan association of fear-avoidance be-liefs with absence from work anddisability.8,10,11,56 Because of interestin assessing the fear-avoidance be-liefs of people receiving outpatientrehabilitation regardless of impair-ment, 2 items were reworded toeliminate references to the back(Appendix). We believed that the re-sulting scale was appropriate for any-one with pain or fear of pain, such asthe people seeking outpatientrehabilitation.

Data AnalysesDistribution of response choices.The frequency distribution of re-sponses to each item was evaluated.

IRT analyses. We used unidimen-sional IRT methods to analyze thedata43,57–59 to determine how wellthe IRT model fit the data and howwell IRT assumptions were met.58

For unidimensional IRT models to beappropriate for analyzing FABQitems, the items must measure onlyone construct; that is, the scale mustbe unidimensional.45,59 In addition,the items must be locally indepen-dent; that is, any 2 items must not becorrelated when the latent trait isfixed.45 We used modern factor ana-lytic methods50–52,60 to investigateunidimensionality and local indepen-dence assumptions. The presence of

a dominant factor in the FABQ itemswas assessed with exploratory factoranalysis (EFA) and then confirmatoryfactor analysis (CFA),61 eliminatingitems with factor loadings of lessthan 0.40.62 Pairs of items with abso-lute residual correlations of greaterthan 0.25 were considered locallydependent.62 All 16 FABQ itemswere used for the initial IRT analysesbecause we wanted to test the factorstructure of all FABQ items.4 The EFAwas used to explore the generalstructure of the FABQ items withoutthe imposition of a preconceivedstructure to determine whether 1 ormore factors were present in thedata. The CFA was used to verify thefactor structure once the factorswere identified with the EFA.63 TheCFA model fit was evaluated with thecomparative fit index (CFI),64 theTucker-Lewis index (TLI),65 and theroot-mean-square error of approxi-mation (RMSEA).63,66 The TLI andthe CFI range from 0 (poor fit) to 1(good fit). Values for the CFI and theTLI of greater than 0.90 are indica-tive of good model fit. Values for theRMSEA of less than 0.08 suggest ad-equate fit.64

IRT model selection, item infor-mation function analysis, anditem fit. We fitted items remainingafter unidimensionality and local in-dependence testing to the graded-response IRT model (GRM)67,68 byusing PARSCALE software (version

4.1).†,69 The GRM was chosen be-cause it is appropriate for orderedresponses (such as FABQ items), itallows item discrimination parame-ters to vary, and it can be used toestimate ability parameters (thetavalues) that represent a subject’slevel of fear. We used PARSCALEsoftware to fit the data to the GRMand to estimate discrimination pa-rameters and category responsefunctions for each item.68,70 Cate-gory response functions representthe probability that an examinee willsuccessfully complete a particular re-sponse category. The category char-acteristic curve for each item in aresponse category is used to estimatethe operating characteristic curvefor each item, which represents theprobability of endorsing a responsecategory for the item at a given sub-ject’s ability (theta value).70 The cat-egory response functions are re-solved into an item location ordifficulty parameter and a set of cat-egory parameters. Therefore, PAR-SCALE produces an item discrimina-tion parameter, an item difficultyparameter, and a set of category pa-rameters for each item. PARSCALEestimates a subject’s level of fear(theta value), category characteristiccurves and parameters, and item dif-ficulty parameters, all of which are

† Scientific Software International Inc, 7383 NLincoln Ave, Suite 100, Lincolnwood, IL60712-1747.

Table 3.Diagnostic Accuracy of Using 1 or 2 Screening Items to Identify Subjects With High Levels of Fear-Avoidance Beliefs RegardingPhysical Activities (PA) or Work Activities (WA)

ScaleNo. of

Screening Items Cut Scorea AUC (95% CI)b Sensitivity Specificity% of Subjects

Correctly Classified �LR/�LRc

PA 1 4 0.94 (0.94, 0.95) 0.82 0.98 91 34.88/0.18

PA 2 7 0.97 (0.97, 0.97) 0.88 0.99 94 446.53/0.12

WA 1 3 0.97 (0.96, 0.97) 0.92 0.93 92 13.11/0.09

WA 2 5 0.99 (0.99, 0.99) 0.95 0.94 95 15.60/0.05

a Cut score�number of response choices (maximum of 5 choices with 1 screening item and 10 choices with 2 screening items) obtained with highestaverage sensitivity divided by specificity.b AUC�area under the receiver operating characteristic curve, CI�confidence interval.c �LR�positive likelihood ratio, �LR�negative likelihood ratio.

Screening for Fear of Work and Physical Activities

774 f Physical Therapy Volume 89 Number 8 August 2009

Table 4.Cross-Walk Table for Scoring Fear-Avoidance Scales With Item Response Theory (IRT) and Original Summative Methodsa

Fear of Physical Activities Fear of Work Activities

IRT-BasedFABQ Measure

in Logitsb

FABQ SummativeScore, X(95% CI) n

IRT-BasedFABQ Measure

in Logits

FABQ SummativeScore, X(95% CI) n

�1.6 2.1 (2.0, 2.2) 1,599 �1.5 0.1 (0.1, 0.2) 1,051

�1.3 4.4 (4.1, 4.8) 136 �1.2 2.1 (1.6, 2.5) 59

�1.2 7.1 (6.8, 7.3) 233 �1.1 4.3 (3.9, 4.7) 145

�1.1 7.6 (7.1, 8.2) 179 �1.0 3.9 (3.2, 4.6) 103

�1.0 4.5 (4.1, 4.9) 96 �0.9 4.0 (3.5, 4.5) 79

�0.9 6.2 (5.9, 6.5) 209 �0.8 6.1 (5.6, 6.6) 176

�0.8 6.9 (6.5, 7.2) 239 �0.7 6.7 (6.0, 7.4) 125

�0.7 7.5 (7.3, 7.6) 642 �0.6 9.5 (8.8, 10.2) 148

�0.6 9.9 (9.7, 10.1) 388 �0.5 9.1 (8.3, 10.0) 130

�0.5 10.3 (10.1, 10.5) 444 �0.4 10.2 (9.2, 11.1) 155

�0.4 10.8 (10.5, 11.1) 196 �0.3 10.9 (9.9, 12.0) 140

�0.3 12.0 (11.8, 12.2) 501 �0.2 10.4 (9.4, 11.4) 146

�0.2 10.1 (9.9, 10.4) 556 �0.1 11.8 (10.8, 12.7) 159

�0.1 11.7 (11.4, 12.0) 322 0 14.2 (13.4, 15.0) 177

0 13.2 (13.2, 13.3) 2,957 0.1 15.0 (14.1, 15.8) 194

0.1 14.2 (14.1, 14.4) 492 0.2 16.2 (15.4, 17.0) 193

0.2 15.2 (15.1, 15.4) 842 0.3 19.3 (18.6, 20.0) 204

0.3 14.9 (14.7, 15.1) 284 0.4 21.4 (20.7, 22.1) 213

0.4 15.7 (15.4, 15.9) 629 0.5 22.9 (22.1, 23.7) 202

0.5 16.1 (15.9, 16.2) 518 0.6 24.0 (23.1, 24.9) 177

0.6 17.5 (17.4, 17.7) 923 0.7 25.6 (24.6, 26.5) 171

0.7 19.0 (18.8, 19.3) 325 0.8 26.5 (25.4, 27.6) 142

0.8 18.7 (18.4, 18.9) 208 0.9 28.3 (27.5, 29.1) 153

0.9 20.9 (20.6, 21.1) 167 1.0 29.6 (28.8, 30.5) 134

1.0 20.2 (19.9, 20.4) 142 1.1 30.8 (30.1, 31.5) 130

1.1 18.1 (17.7, 18.4) 410 1.2 32.7 (32.0, 33.4) 103

1.2 19.8 (19.7, 19.9) 708 1.3 34.0 (33.3, 34.7) 90

1.4 21.8 (21.6, 21.9) 258 1.4 35.2 (34.4, 36.0) 74

1.7 23.8 (23.8, 23.8) 1,640 1.5 35.8 (35.2, 36.4) 100

1.6 38.0 (37.4, 38.5) 81

1.7 38.2 (37.7, 38.8) 102

1.8 39.1 (38.1, 40.0) 32

1.9 40.2 (39.5, 40.9) 42

2.0 39.8 (39.2, 40.4) 87

2.1 42.0 (42.0, 42.0) 3

a FABQ�Fear-Avoidance Beliefs Questionnaire, CI�confidence interval.b If a logit measure is missing, it is because there were no data for analysis.

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 775

placed on the same normal (X�0,SD�1) fear metric in logits.67

PARSCALE (with the GRM) also esti-mates item information functions,which quantify the capability of agiven item to adequately estimate asubject’s ability across the fear-avoidance scale range.45,70 An iteminformation function describes eachitem’s contribution to overall testprecision. The sum of the item infor-mation functions defines the idealprecision of the test (ie, test informa-tion function) at a given ability, facil-itating evaluation of the expectedstandard error. The standard error ofa subject’s ability estimate is in-versely proportional to the test infor-mation function: SE�1/square rootof the test information function. Forsamples with an observed varianceof 1, a standard error of less than0.23 is comparable to a reliability ofgreater than 0.95 (reliability�1�SE2).71,72

Item discrimination parameters andoperating characteristic curves wereassessed to determine how well theitems were modeled with the GRM.Because there is no recognized bestway to assess the fit of data to theGRM, particularly for samples ex-ceeding 1,500,71 we used 3 basic ap-proaches to assess the fit of our datato the GRM. First, we assessed em-pirical operating characteristiccurves to ensure that they pro-gressed from less difficult to moredifficult along the fear-avoidance axisand that each curve reached a max-imum at a unique interval of thescale.67,70 Second, we assessed itemdiscrimination parameters (ie,slopes) for an estimation of the dis-crimination power for each item.Items with larger discrimination pa-rameters (higher slopes) differenti-ate subjects with fear levels varyingover the range of theta values appro-priate for the item better than doitems with lower slopes; therefore,items with slopes of greater than

0.70 are preferred.62 Third, we as-sessed theoretical versus empiricaloperating characteristic curves for aqualitative determination of the fit ofitems to the GRM. Visual inspectionof empirical operating characteristiccurves can reveal the extent and na-ture of item fit or misfit.

DIF. The remaining items were as-sessed for DIF by selection of clini-cally logical groups of subjects: sex(male/female), surgical history (yes/no), acuity of symptoms (number ofcalendar days between date of on-set of symptoms and date of initialevaluation: acute�21 days or less,subacute�22–�90 days, and chronic�90 days or more), age group (18–�45, 45–�65, 65–�75, and 75 yearsor older), number of comorbidities(0, 1, 2, 3, or more), pain intensity(below median, median, or abovemedian, as indicated with a numericrating of 0 [“no pain”] to 10 [“painas bad as it can be”]), and impair-ment grouping (Tab. 1). Differentialitem functioning is present whenthe relationship between item re-sponses and the trait measured bythe test differs systematically be-tween groups of subjects after thesubjects’ underlying abilities are con-trolled for.44 The variables selectedfor testing have been shown to af-fect functional status outcomes.53,54

Associations between fear and theseindependent variables have only be-gun to be investigated, and prelimi-nary results suggest the need for fur-ther investigation.39

For FABQ measures to be used asscreening tools regardless of a sub-ject’s impairment, DIF must be ab-sent or negligible in as many inde-pendent variables as possible—butmost importantly, in impairment. Be-cause confirmable diagnoses formany subjects receiving outpatientrehabilitation often are not avail-able,73 we elected to group subjectsby impairment (ie, the problem di-recting patient management). Differ-

ential item functioning testing for im-pairment grouping was performed in3 ways. First, all subjects weregrouped by general impairment (ie, amedical, neurological, or orthopedicproblem). Second, subjects withan orthopedic impairment weregrouped by area treated (ie, upperextremity, spine, or lower extrem-ity). Third, subjects with an orthope-dic impairment were grouped byspecific body part treated withineach area treated (upper extremity:shoulder, elbow, and wrist or hand;spine: cervical and lumbar; lower ex-tremity: hip, knee, and foot orankle).

Each item was assessed for DIF withdifwithpar software (version1.0),‡,74–79 which combines IRT cal-ibration estimated by the GRM67

with PARSCALE software69 with mul-tiple ordinal logistic regression mod-els for each item and demographiccategory by use of Stata software(version 9.2).§,80 Using methods de-scribed by Crane et al,75 we evalu-ated items for the presence of uni-form DIF (ie, the interference relatedto demographic groups betweenability and item responses is thesame across the entire range mea-sured by the test) by examining therelative difference between beta co-efficients in the regression modelsand nonuniform DIF (ie, the interfer-ence varies at different levels of thetrait being measured) by comparingthe �2 log likelihoods of 2 of theregression models.79 For nonuniformDIF, we used Bonferroni adjustmentfor � values on the basis of the num-ber of items in the scale. The processis sequential (ie, it starts with oneindependent variable and progressesto subsequent variables) and itera-tive (ie, decisions are made at eachstep during the difwithpar process).

‡ Crane P, Gibbons LE, Jolley L, van Belle G,University of Washington, Seattle, WA, 2005.§ StataCorp LP, 4905 Lakeway Dr, College Sta-tion, TX 77845.

Screening for Fear of Work and Physical Activities

776 f Physical Therapy Volume 89 Number 8 August 2009

For example, when an item wasidentified with DIF, the software cre-ated a new item. Thus, items foundto have DIF related to an indepen-dent variable, such as sex, were splitinto 2 new items. For the first newitem, responses for women werecoded as in the original data set,whereas for men, all responses wereset to missing. For the second newitem, responses for men were codedas in the original data set, whereasfor women, all responses were set tomissing. We thus calibrated item pa-rameters independently in the 2groups for items identified with DIF.Items free of DIF served as anchoritems, ensuring that ability estimates(ie, levels of fear) were calibrated onthe same metric for the 2 sexes. Thepresence of possible false-positiveor false-negative DIF results wasassessed.75

In some samples, particularly largesamples, DIF might be detected (sig-nificant) but might be of little prac-tical importance.52,78 Therefore, be-fore progressing sequentially to thenext variable for DIF assessment, weassessed the correlation between un-adjusted ability estimates and DIF-adjusted ability estimates, and we as-sessed the magnitude of thedifference between unadjusted andDIF-adjusted ability estimates. We re-peated the entire procedure for sur-gical history, severity, age group, andimpairment grouping.

Screening item selection andaccuracy of the screening items.We wanted to select screening itemsthat provided the most informationfor the center of the fear continuum.We expected FABQ measures not tobe normally distributed14; therefore,we used the median for each fearscale as the measure of centraltendency.

For each scale, we examined iteminformation functions and selected 2screening items that provided the

most information (ie, the lowestmeasurement error) at the medianfear level. Using the median, we di-chotomized subjects by low versushigh levels of fear of physical activi-ties and fear of work activities withthe IRT-based theta values estimatedfrom all items for each scale.

We used nonparametric receiver op-erating characteristic (ROC) curveanalyses to quantify the accuracy ofthe responses to the screening itemor items (ie, 1 or 2 screening itemsper scale) for discriminating subjectswith fear levels below the median(low) or above the median (high).81

Such analyses produce plots of sen-sitivity/(1 � specificity) for the diag-nostic test (ie, the screening items).For each ROC, a diagnostic cut scorewas identified by selecting the itemresponse (or sum of 2 item re-sponses) with the largest averagespecificity/sensitivity. Positive likeli-hood ratios (�LRs) and negative like-lihood ratios (�LRs)82 and the per-centages of subjects correctlyidentified were produced for eachcut score. Positive likelihood ratioswere calculated as sensitivity/(1 �specificity), and negative likelihoodratios were calculated as (1 � sensi-tivity)/specificity.83 Likelihood ratiosare summary measures of diagnostictest performance (ie, classification)that indicate how much a given clas-sification will raise or lower the pre-test probability of the target disorderof interest (ie, level of fear).83–85 Ac-ceptable �LRs are 2 or higher, andacceptable �LRs are 0.5 or lowerbecause they generate at least smallbut possibly important changes inthe predictive value of the test.85 Ar-eas under the ROC curves, standarderrors, and 95% confidence intervalswere used to describe the ROC re-sults. To determine whether using 1versus using 2 screening items wasmore accurate for discriminatingsubjects with low versus high levelsof fear, we assessed the equality ofthe area under the curves by using an

algorithm suggested by DeLonget al.86

Mapping IRT-based measures tooriginal summative scores. Toassist clinicians in relating new IRT-based FABQ measures to originalFABQ summative scores,4 wemapped the new IRT-based FABQmeasures to the original FABQ sum-mative scores4 by aggregating theoriginal summative scores by eachtenth of a logit of the IRT-based mea-sures. Using the original 0 to 6 itemresponses,4 we summed the re-sponses for items 2 through 5 (Ap-pendix) to produce a summativescore (0–24) for fear of physical ac-tivities, and we summed the re-sponses for items 6, 7, 9, 10, 11, 12,and 15 to produce a summativescore (0–42) for fear of work activ-ities. At each tenth of a logit for eachFABQ-PA and FABQ-W IRT-basedtheta value, the mean and 95% con-fidence interval for the original sum-mative scores were calculated.

ResultsDistribution of Response ChoicesNo item had greater than 95% re-sponses in any one category. Exami-nation of proportions of responsesper item for the 11 items in the orig-inal FABQ-W scale and the 5 items inthe original FABQ-PA scale demon-strated that subjects selected re-sponses with word descriptors more(proportions of between .16 and .47for response choices 0, 3, and 6)than responses without word de-scriptors (proportions of between0.02 and 0.05 for response choices 1,2, 4, and 5), regardless of the FABQscale.

IRT AnalysesThe EFA results indicated that a2-factor solution for the 16 FABQitems (n�3,956, CFI�0.92, TLI�0.97, RMSEA�0.19) fit the data well.Items were loaded on the FABQ-PAand FABQ-W scales originally de-scribed by Waddell et al,4 and the

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 777

2-factor solution controlled 69.1% ofthe variance in the data. The 2-factorCFA results supported the presenceof 2 factors (CFI�0.93, TLI�0.97,RMSEA�0.18). Items were separatedinto respective scales (for the 11-item FABQ-W, n�5,517; for the5-item FABQ-PA, n�16,243), andseparate 1-factor CFAs were run.

The CFA results for the 11-itemFABQ-W suggested that 2 of the 3 fitstatistics supported the fit of the1-factor solution (CFI�0.94, TLI�0.97, RMSEA�0.25), all items wereloaded on 1 factor (loadings of�.75), but the items “My pain was

caused by my work or by an accidentat work” and “I do not think that Iwill ever be able to go back to thatwork” had a residual correlation of�0.26. The former item was deletedbecause it was associated with themost pairs of items with higher ab-solute residual correlations, and an-other CFA was run on the 10 remain-ing items. The CFA results for the10-item FABQ-W supported a slightlyimproved model fit (CFI�0.95,TLI�0.98, RMSEA�0.23), all itemswere loaded on 1 factor (loadings of0.68), there was no absolute residualcorrelation of greater than 0.25,there was 1 residual correlation of

less than �0.20, and there was a re-duction in absolute residual correla-tions of greater than 0.10, from36.4% (11-item scale) to 28.9% (10-item scale). With the exception ofthe RMSEA, the results supported aunidimensional scale with good localindependence.

The CFA results for the 5-itemFABQ-PA suggested that 2 of the 3 fitstatistics supported the fit of the1-factor solution (CFI�0.95,TLI�0.93, RMSEA�0.23), all itemsbut 1 were loaded on 1 factor (forthe item “My pain was caused byphysical activity,” the loading was0.34), and all absolute residual corre-lations were less than 0.25. This itemwas deleted, and another CFA wasrun on the 4 remaining items. TheCFA results for the 4-item FABQ-PAsuggested a questionably improvedmodel fit (CFI�0.96, TLI�0.95,RMSEA�0.26), all items were loadedon 1 factor (loadings of �0.60),there was no absolute residual corre-lation of greater than 0.25, there was1 residual correlation of greater than0.20, and there was a reduction inabsolute residual correlations ofgreater than .10, from 15.0% (5-itemscale) to 8.3% (4-item scale). Withthe exception of the RMSEA, the re-sults supported a unidimensionalscale with good local independence.

IRT Modeling, Item InformationFunction Analysis, and Item FitWe fitted items from both FABQscales separately to the GRM. Initialinspection of the operating charac-teristic curves demonstrated thatthere were no distinct maximum val-ues of the item response curves forthe second and third as well as thefifth and sixth response categories(ie, responses without word descrip-tors) for both scales. Therefore,these response categories (ie, sec-ond and third as well as fifth andsixth) were collapsed, and the datawere refit to the GRM. Subsequentinspection of operating characteris-

Figure 1.(A) Test information function for fear-avoidance beliefs regarding physical activities. (B)Test information function for fear-avoidance beliefs regarding work activities.Information�test information function.

Screening for Fear of Work and Physical Activities

778 f Physical Therapy Volume 89 Number 8 August 2009

tic curves supported an improvedshape for each curve, with clearmaximum values. One item (“Physi-cal activity makes my pain worse”)had a discrimination parameter ofless than 0.7 (actual value�0.67) andwas deleted. Examination of empiri-cal operating characteristic curveplots suggested that all items fit theGRM. The 3-item FABQ-PA data wererefit to the GRM (Tab. 2).

Test information functions with stan-dard errors for both FABQ scales aredisplayed in Figure 1. If a measure offear of physical activities is estimatedwith the 3-item scale, then the plotof standard errors (SEs of � 0.23 rep-resent a measure reliability of �.95)demonstrates that the measure canbe estimated with high precision be-tween �1.1 and 1.5, or 29.5% of therange for the FABQ-PA trait in oursample. Similarly, with the 10-itembank for fear of work activities, themeasure of fear can be estimatedwith high precision (reliability of�.95) between �0.2 and 0.4, or11.5% of the range for the FABQ-Wtrait in our sample.

DIFThe DIF results for fear of physicalactivities (3-item scale) indicatedthat there were no items with DIFfor the variables sex, symptom acu-ity, surgical history, number of co-morbidities, level of pain, and anyimpairment grouping. The item “Ishould not do physical activitieswhich (might) make my pain worse”was significant (P�.001) for nonuni-form DIF for age, but the unadjustedand adjusted levels of fear werehighly correlated (r�.99), and theaverage difference between the un-adjusted and adjusted measures was�.001 (SD�.02, range�.14–.07)—avalue that was �.001 standard devi-ation from the full 3-item scale.Therefore, the DIF was consideredto be of no practical importance.

The DIF results for fear of work ac-tivities (10-item scale) provided sim-ilar results. No items with DIF wereidentified for the variables sex, age,number of comorbidities, level ofpain, overall impairment grouping(medical, neurological, and orthope-dic), and orthopedic impairments ofthe upper or lower extremity. Sev-eral items were shown to have non-uniform DIF for the variables symp-tom severity; surgical history;orthopedic impairment grouping byupper extremity, lower extremity, orspine; and orthopedic impairmentgrouping by cervical or lumbarspine. However, the unadjusted andadjusted levels of fear were highlycorrelated (r�.99), and the averagedifferences between the unadjustedand adjusted measures ranged from�.12 to .02—values that representeda range of standard deviations of .01to .11. Therefore, the identified DIFwas considered to be of little practi-cal importance.

Screening Item Selection andAccuracy of the Screening ItemsBoth the FABQ-PA and the FABQ-Wwere distributed nonnormally(Shapiro-Wilks W statistics, P�.05).The median for the FABQ-PA was�.07, and the median for theFABQ-W was �.02; these valueswere used to dichotomize subjectsby level of fear.

We identified 2 items (Tab. 2) withthe highest slopes as the first 2screening items per scale: screeningitem 1 [SHLDNOT—“I should not dophysical activities which (might)make my pain worse”] and screeningitem 2 [CANNOT—“I cannot dophysical activities which (might)make my pain worse”] for fear ofphysical activities and screeningitem 1 (WRKCANT—“I cannot domy normal work with my presentpain”) and screening item 2 (WRK-SHNT—“I should not do my normalwork with my present pain”) for fearof work activities. Although

WRKSHNT provided slightly moreinformation (��2.53) than WRK-CANT (��2.52), the WRKCANTitem information function providedmore information at the mediantheta value and therefore was se-lected as the most informative forthe work scale at the cut score forhigh levels of fear.

The ROC results describing the accu-racy of using 1 or 2 items to predictsubjects with high levels of fear ofphysical or work activities are shownin Table 3. Although the areas underthe ROCs were similar when 1 or 2screening items were used to iden-tify high levels of fear, the use of 2items produced larger areas(�2�402.6, df�1, P�.001, for theFABQ-PA; �2�139.6, df�1, P�.001,for the FABQ-W) (Fig. 2). However,because the use of 1 screening itemproduced strong values for areas un-der the curves, sensitivity, specific-ity, �LR, �LR, and percentages ofsubjects correctly classified and be-cause the addition of a secondscreening item did not substantiallyimprove all of these values overthose obtained with 1 screeningitem, we decided to use only 1 item(ie, the most informative at the me-dian theta value) as the screeningitem to identify subjects with highlevels of fear for both scales.

Mapping IRT-Based Measures toOriginal Summative ScoresA cross-walk table for both fearscales is shown in Table 4. With thetable, once an IRT-based measure isknown, the original FABQ summa-tive score can be identified.

DiscussionThe 2 most important findings ofthese analyses were that the items inthe scales for fear-avoidance beliefsregarding physical and work activi-ties had negligible DIF across manyvariables describing people com-monly seen in outpatient rehabilita-tion. The lack of practically impor-

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 779

tant DIF allowed us to identify forIRT-based FABQ scales single screen-ing items that could be used to effi-ciently classify people with elevatedlevels of fear-avoidance beliefs in anaccurate manner regardless of theimpairment being treated. Thesefindings are consistent with reportsthat fear influences outcomes forpeople with hip,29 knee,29–31 cervi-cal spine and shoulder,28 andneck11,25–27 pain, as well as for peo-ple with lumbar spine impair-ments,4,10,11,16 for whom the FABQwas designed.4 Therefore, when ap-propriate, clinicians could use the

single screening items identified inthese analyses to identify peoplewith elevated levels of fear acrossa wide variety of impair-ments.1,12,15,16,18,87 If elevated levelsof fear were detected, people couldbe tested further to estimate more-precise measures of fear-avoidancebeliefs. There is evidence to suggestthat management strategies can beused to reduce fear19–21 and improveoutcomes for people with low backpain.14,19–21,35,88–90 In addition, thereis evidence that management strate-gies are evolving for other musculo-

skeletal conditions, such as shoulderimpairments.42

The present study was performed indirect response to a challenge to re-fine current screening techniquesfor elevated levels of fear-avoidancebeliefs, so that people can be accu-rately and efficiently classified andtheir conditions can be managed ac-cordingly.17 Because our IRT-basedscreening requires only 1 item to ac-curately classify people, the methodis efficient. Improved efficiency, thatis, a reduced burden of collectingdata, may be the catalyst for morewidespread screening for fear-avoidance beliefs in routine outpa-tient therapy and may facilitate con-current screenings of multiplepsychosocial prognostic indicators,such as depression38 and pain-relatedfear.4 As more therapists assume afirst-contact role,37 efficient yet accu-rate screening of multiple constructswill be developed to meet therapists’needs, which will allow rapid identi-fication of people who may requirecertain types of help as early as pos-sible. The use of IRT methods canfacilitate such development becauseIRT methods are well suited to thedevelopment of new scales and thereassessment of existing scales, in-cluding the identification of singlescreening items and the assessmentof measurement precision.

The results of the present study in-dicated that subjects selected theoriginal FABQ item responses4 withword descriptors more than re-sponses without word descriptors.In addition, approximately 4% (floor)and 9% (ceiling) of the subjects se-lected “completely disagree” and“completely agree” responses for allitems of the FABQ-PA scale, respec-tively; the corresponding values forthe FABQ-W scale were 20% (floor)and 3% (ceiling). Therefore, FABQscores tended not to be normally dis-tributed, and subjects might clusterat the 2 scale extremes; these find-

Figure 2.(A) Receiver operating characteristic (ROC) curves for use of 1 and 2 screening items toidentify subjects with high levels of fear of physical activities. (B) ROC curves for use of1 and 2 screening items to identify subjects with high levels of fear of work activities.One ROC area�1 screening item was used to estimate the area under the ROC curve,Two ROC area�2 screening items were used to estimate the area under the ROC curve.

Screening for Fear of Work and Physical Activities

780 f Physical Therapy Volume 89 Number 8 August 2009

ings support the results of previousstudies14,91 in which medians wereused as measures of central tenden-cies for both FABQ scales to dichot-omize subjects.

The results of the factor analyses in-dicated that the FABQ items wereunidimensional, with good local in-dependence, in the original item for-mat4 once the item CAUSED was de-leted from the work scale and theitem PHYSACTV was deleted fromthe physical activity scale becausethey were not loaded strongly on therespective scales. The loss of theitem PHYSACTV because of low fac-tor loading is consistent with the re-ports of Waddell et al4 and Staerkleet al.13 However, beyond the loss ofthe item PHYSACTV, our results can-not be compared directly with thoseof Waddell et al4 and Staerkle et al13

because we used factor analyses de-signed for categorical data, the sam-ples differed in size and diversity,and we edited 2 FABQ items to elim-inate references to the back(Appendix).

To our knowledge, no other re-search group has analyzed FABQdata by using IRT methods, whichallowed subjects’ FABQ responsesto be described in probabilisticterms.43,57–59,92 Specifically, operat-ing characteristic curves, whichgraphically depict the correspon-dence between the predicted re-sponses to an item and the latenttrait,59 demonstrated that subjectsdid not differentiate the originalFABQ responses well for responseswith no word descriptors; this find-ing supports the frequency distribu-tion results and calls into questionthe use of response categories with-out word descriptors. When we col-lapsed the 7 responses to 5, themonotonic nature of the operatingcharacteristic curves was restored;this finding implies that the originalresponse anchoring adds error to themeasurement of levels of fear-

avoidance beliefs and that subjectsmay be able to better differentiateamong 5 responses, thus supportingrecent recommendations.93,94 Agood fit of items to the 2-parameterIRT model was obtained with the10-item work activity scale and thecollapsed response choices, but agood item fit with the physical activ-ity scale and the collapsed responsechoices was obtained only after thedeletion of 1 more item for fear ofphysical activities: WORSE. There-fore, the final IRT-based FABQ scalescontained 3 items for physical activ-ities and 10 items for work activities.These results suggest that theFABQ—in its original format of 7 re-sponse categories for 4 items in thephysical activity scale and 7 items inthe work activity scale—could beimproved through the use of IRTmethods, which some authors sug-gest are more exacting than the clas-sical test theory method43,58,92 origi-nally used to analyze FABQ data anddevelop the original FABQ scales.

Once the FABQ data were analyzedwith IRT methods, screening itemsthat provided maximum informa-tion45 at the median for the FABQscales could be easily selected. Itemresponse theory methods are ideallysuited to this task because plots ofitem information functions allowidentification of the amount of infor-mation or discriminating ability ofeach item at any level of fear. Wewanted to develop a test (ie, thescreening items) that accurately di-chotomized subjects into groupswith low versus high levels of fear(ie, high levels of fear are diseasepositive),83 and selecting screeningitems that provided maximum infor-mation at the median fear level pro-duced strong diagnostic test results.

According to Sackett et al,83 when atest with high specificity is positive,the result effectively rules in the di-agnosis. The specificities for thephysical activity and work activity

scales with 1 screening item werestrong, 0.98 and 0.93, respectively.In addition, the �LRs, which can beinterpreted as the ratio of true-positive results to false-positive re-sults,84 also were strong. Althoughthe use of 2 screening items dramat-ically improved the �LR for thephysical activity scale, the alreadystrong specificity improved little; inaddition, the work activity scalespecificity and �LR did not improveappreciably with 2 screening items.The �LR can be interpreted as acost-to-benefit ratio, in which therate of true-positive results repre-sents a benefit criterion and the rateof false-positive results represents acost criterion.84 To minimize unnec-essary testing and inappropriatetreatment, �LR should be high.Here, we elected to use one screen-ing item per scale, and this methodwas accurate and efficient and pro-duced high �LRs. Therefore, withthe IRT-based fear-avoidance beliefscales, a subject who selected theunlabeled response “unsure” orhigher for the SHLDNOT screeningitem on the FABQ-PA scale wasabout 35 times more likely to havehigh levels of fear; a subject whoselected “unsure” or higher on theWRKCANT triage item on theFABQ-W scale was about 13 timesmore likely to have high levels offear. High levels of fear representedFABQ scores higher than the medianfear level, which has been associatedwith poorer functional statusoutcomes.14,19–21,35,88–90

The original FABQ scales scoredwith summative methods as de-scribed by Waddell et al4 are com-mon. However, summative scoringof categorical data typically pro-duces nonlinear scores, whereas IRT-based measures produce linearscores, as evidenced by the data inTable 4. Summative scores are easyto obtain in clinics, but scores fromIRT-based measures require com-puter technology to obtain. The va-

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 781

lidity of using parametric statisticaltechniques for nonlinear summativemeasures has been questioned.95,96

Clinicians who wish to transformnew IRT-based measures to originalFABQ summative scores4 can use thecross-walk table. For example, if theIRT-based measure of fear of physicalactivities were 0.6 logit, then theoriginal summative score, estimatedfrom the cross-walk table, would be17.5 (95% confidence interval�17.4–17.7)—a value considered tobe elevated.

Finally, predictions of elevated levelsof fear relate to intake fear. Medianintake FABQ scores have been usedto classify subjects into groups withhigh versus low levels of fear.14,19

Findings from these randomized con-trolled trials14,19 suggested that mod-ifications of management strategiesdesigned to reduce the effects offear-avoidance beliefs for subjectswith elevated levels of fear tend todecrease disability (ie, improve func-tional status). However, as describedby George et al,1 dichotomizing sub-jects on the basis of a median cutscore at intake does not necessarilyrepresent an increased probability ofdeveloping chronic symptoms. In ad-dition, FABQ items demonstrated noDIF by level of pain intensity, a find-ing that could facilitate future studiesexamining the relationship betweenpain intensity and activity-relatedfear. Further studies with longitudi-nal designs and external criteria arerecommended to test the predictivepower of the cut scores identifiedwith our data, as well as the use ofscreening items to assess improve-ments in functional status or qualityof life associated with changes infear-avoidance beliefs or even im-provements in fear-avoidance beliefsas a treatment outcome.

Limitations and Future StudiesThe present study is not without lim-itations. The RMSEA values werehigher than desired for assessing the

fit of the data to the CFAs. All otherCFA fit indexes were strong, as wereassessments of the fit of the data tothe GRM. High RMSEA values implythat the data do not fit the CFA;therefore, further testing to validatethe notion that FABQ items repre-sent unidimensional scales worthy ofIRT analyses is recommended. Sub-ject grouping by impairment to as-sess fear-avoidance belief screeningmay not be as discriminating as othermethods of grouping subjects, in-cluding grouping by diagnosis; how-ever, the impairment data appearedto be clinically logical, and obtainingconfirmable, reliable, and valid diag-noses for many people seeking out-patient rehabilitation is difficult.Other methods of grouping subjectsshould be explored.

The present study represents a retro-spective analysis of an existing effec-tiveness database. The researchershad no control over which subjectswere asked to complete the FABQsurveys; therefore, the potential forbiased results exists. However, be-cause the sample was large, it can beargued that the results representedadequate estimates of the FABQscores. However, it would prudentto investigate the effect of not allsubjects answering the FABQ sur-veys. Future prospective studies re-lated to the reliability and validity ofscreening FABQ measures are en-couraged. Future studies should con-sider screening for levels of fear withFABQ cut scores that are not basedon a median split and should explorepotentially informative associationsbetween clinical variables and otherpsychosocial factors, including lev-els of fear, false-positive results, andfloor and ceiling effects. The use ofscreening information by cliniciansto modify management and interven-tions to improve outcomes is en-couraged. Because IRT-based screen-ing for elevated levels of fear-avoidance beliefs was accurate, theuse of elevated levels of fear-

avoidance beliefs as a risk adjust-ment variable in longitudinal studiesof changes in functional statusshould be explored. Finally, efficientcollection of data is facilitated by theuse of computers. Exploration of theefficiency and accuracy of combin-ing IRT-based FABQ screening withcomputerized adaptive testing offunctional status is encouraged.

ConclusionUsing IRT methods, we analyzedscales commonly used for fear-avoidance beliefs regarding physicaland work activities with a large sam-ple of people being treated for com-mon neuromusculoskeletal impair-ments in outpatient rehabilitation.The results indicated that IRT meth-ods can improve assessments withFABQ scales and can be used to de-termine single screening items thatcan accurately identify people withhigh levels of fear at rehabilitationintake. Because the items in the IRT-based FABQ scales had negligibleDIF, particularly for a subject’s im-pairment, the results support the useof the screening items to identifypeople with high levels of fear inroutine practice in a variety of sub-groups of people seeking outpatientrehabilitation. The use of IRT-basedFABQ scales might prove beneficialby alerting therapists to the likeli-hood of elevated levels of fear,which could prompt further testingor modifications of managementstrategies designed to produce im-proved outcomes.

Dr Hart and Mr Werneke provided concept/idea/research design. Dr Hart, Mr Werneke,Dr George, Dr Matheson, and Dr Cook pro-vided writing. Dr Hart, Mr Werneke, DrMatheson, and Mr Mioduski provided datacollection. Dr Hart and Dr Cook provideddata analysis. Dr Hart provided project man-agement. Dr George, Dr Matheson, DrWang, Dr Cook, and Dr Choi provided con-sultation (including review of manuscript be-fore submission. Dr Hart is an employee ofand investor in Focus On Therapeutic Out-comes, Inc (FOTO), the database manage-ment company that manages the data ana-

Screening for Fear of Work and Physical Activities

782 f Physical Therapy Volume 89 Number 8 August 2009

lyzed in the study. Mr Mioduski wrote thesoftware used to collect the data and man-age the aggregated database from whichthe data were drawn for the analyses. MrWerneke and Dr Matheson work for clinicalfacilities that use the FOTO data collectionsystem for routine data collection and casemanagement. The authors thank Dr PaulCrane and Dr Laura Gibbons for their in-sightful comments regarding differentialitem functioning analyses.

The Institutional Review Board for the Pro-tection of Human Subjects, Focus On Ther-apeutic Outcomes, Inc, approved theproject.

Part of this research was presented at theInternational Conference on OutcomesMeasurement; September 11–13, 2008; Be-thesda, Maryland; and at the Combined Sec-tions Meeting of the American Physical Ther-apy Association; February 9–12, 2009; LasVegas, Nevada.

This article was received July 24, 2008, andwas accepted April 10, 2009.

DOI: 10.2522/ptj.20080227

References1 George SZ, Fritz JM, Childs JD. Investiga-

tion of elevated fear-avoidance beliefs forpatients with low back pain: a secondaryanalysis involving patients enrolled inphysical therapy clinical trials. J OrthopSports Phys Ther. 2008;38:50–58.

2 Linton SJ. A review of psychological riskfactors in back and neck pain. Spine.2000;25:1148–1156.

3 Pincus T, Burton AK, Vogel S, Field AP. Asystematic review of psychological factorsas predictors of chronicity/disability inprospective cohorts of low back pain.Spine. 2002;27:E109–E120.

4 Waddell G, Newton M, Henderson I, et al.A Fear-Avoidance Beliefs Questionnaire(FABQ) and the role of fear-avoidance be-liefs in chronic low back pain and disabil-ity. Pain. 1993;52:157–168.

5 Leeuw M, Goossens ME, Linton SJ, et al.The fear-avoidance model of musculoskel-etal pain: current state of scientific evi-dence. J Behav Med. 2007;30:77–94.

6 Lethem J, Slade PD, Troup JD, Bentley G.Outline of a Fear-Avoidance Model of ex-aggerated pain perception—I. Behav ResTher. 1983;21:401–408.

7 Vlaeyen JW, Koel-Snijders AMJ, RotteveelAM, et al. The role of fear of movement/(re)injury in pain disability. J Occup Reha-bil. 1995;5:235–252.

8 Vlaeyen JW, Kole-Snijders AM, Boeren RG,van Eek H. Fear of movement/(re)injury inchronic low back pain and its relation tobehavioral performance. Pain. 1995;62:363–372.

9 Vlaeyen JW, Linton SJ. Fear-avoidance andits consequences in chronic musculoskel-etal pain: a state of the art. Pain.2000;85:317–332.

10 Crombez G, Vlaeyen JW, Heuts PH, LysensR. Pain-related fear is more disabling thanpain itself: evidence on the role of pain-related fear in chronic back pain disability.Pain. 1999;80:329–339.

11 George SZ, Fritz JM, Erhard RE. A compar-ison of fear-avoidance beliefs in patientswith lumbar spine pain and cervical spinepain. Spine. 2001;26:2139–2145.

12 Fritz JM, George SZ, Delitto A. The role offear-avoidance beliefs in acute low backpain: relationships with current and futuredisability and work status. Pain. 2001;94:7–15.

13 Staerkle R, Mannion AF, Elfering A, et al.Longitudinal validation of the Fear-Avoidance Beliefs Questionnaire (FABQ)in a Swiss-German sample of low backpain patients. Eur Spine J. 2004;13:332–340.

14 Burton AK, Waddell G, Tillotson KM, Sum-merton N. Information and advice to pa-tients with back pain can have a positiveeffect: a randomized controlled trial of anovel educational booklet in primary care.Spine. 1999;24:2484–2491.

15 Cleland JA, Fritz JM, Brennan GP. Predic-tive validity of initial fear avoidance beliefsin patients with low back pain receivingphysical therapy: is the FABQ a usefulscreening tool for identifying patients atrisk for a poor recovery? Eur Spine J.2008;17:70–79.

16 Fritz JM, George SZ. Identifying psychoso-cial variables in patients with acute work-related low back pain: the importance offear-avoidance beliefs. Phys Ther. 2002;82:973–983.

17 George SZ. Fear: a factor to consider inmusculoskeletal rehabilitation. J OrthopSports Phys Ther. 2006;36:264–266.

18 George SZ, Bialosky JE, Donald DA. Thecentralization phenomenon and fear-avoidance beliefs as prognostic factors foracute low back pain: a preliminary inves-tigation involving patients classified forspecific exercise. J Orthop Sports PhysTher. 2005;35:580–588.

19 George SZ, Fritz JM, Bialosky JE, DonaldDA. The effect of a fear-avoidance-basedphysical therapy intervention for patientswith acute low back pain: results of a ran-domized clinical trial. Spine. 2003;28:2551–2560.

20 George SZ, Fritz JM, McNeil DW. Fear-avoidance beliefs as measured by the Fear-Avoidance Beliefs Questionnaire: changein Fear-Avoidance Beliefs Questionnaire ispredictive of change in self-report of dis-ability and pain intensity for patients withacute low back pain. Clin J Pain. 2006;22:197–203.

21 Linton SJ, Boersma K, Jansson M, et al. Theeffects of cognitive-behavioral and physi-cal therapy preventive interventions onpain-related sick leave: a randomized con-trolled trial. Clin J Pain. 2005;21:109–119.

22 Linton SJ, Buer N, Vlaeyen J, Hellsing AL.Are fear-avoidance beliefs related to theinception of an episode of back pain? Aprospective study. Psychol Health. 2000;14:1051–1059.

23 Houben RM, Leeuw M, Vlaeyen JW, et al.Fear of movement/injury in the generalpopulation: factor structure and psycho-metric properties of an adapted version ofthe Tampa Scale for Kinesiophobia. J Be-hav Med. 2005;28:415–424.

24 Goubert L, Crombez G, De BourdeaudhuijI. Low back pain, disability and back painmyths in a community sample: prevalenceand interrelationships. Eur J Pain. 2004;8:385–394.

25 Landers MR, Creger RV, Baker CV, Stutel-berg KS. The use of fear-avoidance beliefsand nonorganic signs in predicting pro-longed disability in patients with neckpain. Man Ther. 2008;13:239–248.

26 Linton SJ, Ryberg M. A cognitive-behavioral group intervention as preven-tion for persistent neck and back pain in anon-patient population: a randomized con-trolled trial. Pain. 2001;90:83–90.

27 Nederhand MJ, Ijzerman MJ, Hermens HJ,et al. Predictive value of fear avoidance indeveloping chronic neck pain disability:consequences for clinical decision mak-ing. Arch Phys Med Rehabil. 2004;85:496–501.

28 Huis ’t Veld RM, Vollenbroek-Hutten MM,Groothuis-Oudshoorn KC, Hermens HJ.The role of the fear-avoidance model infemale workers with neck-shoulder painrelated to computer work. Clin J Pain.2007;23:28–34.

29 van Baar ME, Dekker J, Oostendorp RA,et al. The effectiveness of exercise therapyin patients with osteoarthritis of the hip orknee: a randomized clinical trial. J Rheu-matol. 1998;25:2432–2439.

30 Chmielewski TL, Jones D, Day T, et al. Theassociation of pain and fear of movement/reinjury with function during anterior cru-ciate ligament reconstruction rehabilita-tion. J Orthop Sports Phys Ther. 2008;38:746–753.

31 Kvist J, Ek A, Sporrstedt K, Good L. Fear ofre-injury: a hindrance for returning tosports after anterior cruciate ligament re-construction. Knee Surg Sports Trauma-tol Arthrosc. 2005;13:393–397.

32 Nash JM, Williams DM, Nicholson R, TraskPC. The contribution of pain-related anxi-ety to disability from headache. J BehavMed. 2006;29:61–67.

33 Turk DC, Robinson JP, Burwinkle T. Prev-alence of fear of pain and activity in pa-tients with fibromyalgia syndrome. J Pain.2004;5:483–490.

34 Heuts PH, Vlaeyen JW, Roelofs J, et al.Pain-related fear and daily functioning inpatients with osteoarthritis. Pain. 2004;110:228–235.

35 de Jong JR, Vlaeyen JW, Onghena P, et al.Reduction of pain-related fear in complexregional pain syndrome type I: the appli-cation of graded exposure in vivo. Pain.2005;116:264–275.

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 783

36 Werneke MW, Hart DL. Centralization: as-sociation between repeated end-rangepain responses and behavioral signs in pa-tients with acute non-specific low backpain. J Rehabil Med. 2005;37:286–290.

37 Leemrijse CJ, Swinkels IC, Veenhof C. Di-rect access to physical therapy in theNetherlands: results from the first year incommunity-based physical therapy. PhysTher. 2008;88:936–946.

38 Haggman S, Maher CG, Refshauge KM.Screening for symptoms of depression byphysical therapists managing low backpain. Phys Ther. 2004;84:1157–1166.

39 Werneke MW, Hart DL, George SZ, et al.Clinical outcomes for patients classified byfear-avoidance beliefs and centralizationphenomenon. Arch Phys Med Rehabil.2009;90:768–777.

40 DeSalvo KB, Fisher WP, Tran K, et al. As-sessing measurement properties of twosingle-item general health measures. QualLife Res. 2006;15:191–201.

41 Rost K, Burnam MA, Smith GR. Develop-ment of screeners for depressive disordersand substance disorder history. Med Care.1993;31:189–200.

42 Geraets JJ, Goossens ME, de Groot IJ, et al.Effectiveness of a graded exercise therapyprogram for patients with chronic shoul-der complaints. Aust J Physiother. 2005;51:87–94.

43 van der Linden WJ, Hambleton RK, eds.Handbook of Modern Item Response The-ory. New York, NY: Springer-Verlag; 1997.

44 Millsap RE, Everson HT. Methodology re-view: statistical approaches for assessingmeasurement bias. Appl Psychol Meas.1993;17:287–334.

45 Lord FM. Applications of Item ResponseTheory to Practical Testing Problems.Hillsdale, NJ: Lawrence Erlbaum Associ-ates; 1980.

46 Deutscher D, Hart DL, Dickstein R, et al.Implementing an integrated electronicoutcomes and electronic health recordprocess to create a foundation for clinicalpractice improvement. Phys Ther. 2008;88:270–285.

47 Swinkels ICS, van den Ende CHM, de Bak-ker D, et al. Clinical databases in physicaltherapy. Physiother Theory Pract. 2007;23:153–167.

48 Hart AC, Stegman MS. ICD-9-CM 2008 Ex-pert. 6th ed. Salt Lake City, UT: Ingenix;2007.

49 Hart DL, Connolly JB. Pay-for-Perfor-mance for Physical Therapy and Occupa-tional Therapy: Medicare Part B Services.Final report. Grant #18-P-93066/9–01.Baltimore, MD: Centers for Medicare andMedicaid Services, Department of Healthand Human Services; 2006. Available at:http://www.cms.hhs.gov/TherapyServices/downloads/P4PFinalReport06-01-06.pdf. Ac-cessed May 18, 2009

50 Hart DL, Cook KF, Mioduski JE, et al. Sim-ulated computerized adaptive test for pa-tients with shoulder impairments was ef-ficient and produced valid measures offunction. J Clin Epidemiol. 2006;59:290–298.

51 Hart DL, Mioduski JE, Stratford PW. Simu-lated computerized adaptive tests for mea-suring functional status were efficientwith good discriminant validity in patientswith hip, knee, or foot/ankle impairments.J Clin Epidemiol. 2005;58:629–638.

52 Hart DL, Mioduski JE, Werneke MW, Strat-ford PW. Simulated computerized adap-tive test for patients with lumbar spineimpairments was efficient and producedvalid measures of function. J Clin Epide-miol. 2006;59:947–956.

53 Hart DL, Wang YC, Stratford PW, MioduskiJE. Computerized adaptive test for patientswith foot or ankle impairments producedvalid and responsive measures of function.Qual Life Res. 2008;17:1081–1091.

54 Hart DL, Wang YC, Stratford PW, MioduskiJE. A computerized adaptive test for pa-tients with hip impairments producedvalid and responsive measures of function.Arch Phys Med Rehabil. 2008;89:2129–2139.

55 Hart DL, Wang YC, Stratford PW, MioduskiJE. Computerized adaptive test for patientswith knee impairments produced validand responsive measures of function.J Clin Epidemiol. 2008;61:1113–1124.

56 Asmundson GJ, Norton GR, AllerdingsMD. Fear and avoidance in dysfunctionalchronic back pain patients. Pain. 1997;69:231–236.

57 Hambleton RK. Emergence of item re-sponse modeling in instrument develop-ment and data analysis. Med Care. 2000;38(9 suppl):II60–II65.

58 Hambleton RK, Swaminathan H, RogersHJ. Fundamentals of Item Response The-ory. Newbury Park, CA: Sage; 1991.

59 Hays RD, Morales LS, Reise SP. Item re-sponse theory and health outcomes mea-surement in the 21st century. Med Care.2000;38(9 suppl):II28–II42.

60 Muthen LK, Muthen BO. Mplus User’sGuide. 4th ed. Los Angeles, CA: Muthen &Muthen; 2006.

61 Bjorner JB, Kosinski M, Ware JE Jr. Thefeasibility of applying item response the-ory to measures of migraine impact: a re-analysis of three clinical studies. Qual LifeRes. 2003;12:887–902.

62 Fliege H, Becker J, Walter OB, et al. Devel-opment of a computer-adaptive test fordepression (D-CAT). Qual Life Res. 2005;14:2277–2291.

63 McDonald RP. Test Theory: A UnifiedTreatment. Mahwah, NJ: Lawrence Erl-baum Associates; 1999.

64 Hu LT, Bentler P. Cutoff criteria for fitindices in covariance structure analysis:conventional criteria versus new alterna-tives. Structural Equation Modeling.1999;6:1–55.

65 Tucker L, Lewis C. A reliability coefficientfor maximum likelihood factor analysis.Psychometrika. 1973;38:1–10.

66 Browne MW, Cudeck R. Alternative waysof assessing model fit. In: Bollen KA, LongJA, eds. Testing Structural Equation Mod-els. Newbury Park, CA: Sage Publications;1993:136–172.

67 Samejima F. Estimation of ability using aresponse pattern of graded responses. Psy-chometrika. 1969. Monograph 17.

68 Samejima F. Graded Response Model. In:van der Linden WJ, Hambleton RK, eds.Handbook of Modern Item Response The-ory. New York, NY: Springer-Verlag;1997:85–100.

69 PARSCALE for Windows, version 4.1. Lin-colnwood, IL: Scientific Software Interna-tional, Inc; 2003.

70 Dodd BG, Koch WR, De Ayala RJ. Opera-tional characteristics of adaptive testingprocedures using the Graded ResponseModel. Appl Psychol Meas. 1989;13:129–143.

71 Rose M, Bjorner JB, Becker J, et al. Evalu-ation of a preliminary physical functionitem bank supported the expected advan-tages of the Patient-Reported OutcomesMeasurement Information System (PRO-MIS). J Clin Epidemiol. 2008;61:17–33.

72 Thissen D. Reliability and measurementprecision. In: Wainer H, ed. Computer-ized Adaptive Testing: A Primer. 2nd ed.Mahwah, NJ: Lawrence Erlbaum Associ-ates; 2000:159–184.

73 Delitto A, Erhard RE, Bowling RW. Atreatment-based classification approach tolow back syndrome: identifying and stag-ing patients for conservative treatment.Phys Ther. 1995;75:470–489.

74 Crane PK, Cetin K, Cook KF, et al. Differ-ential item functioning impact in a modi-fied version of the Roland-Morris DisabilityQuestionnaire. Qual Life Res. 2007;16:981–990.

75 Crane PK, Gibbons LE, Jolley L, van BelleG. Differential item functioning analysiswith ordinal logistic regression tech-niques: DIFdetect and difwithpar. MedCare. 2006;44(11 suppl 3):S115–S123.

76 Crane PK, Gibbons LE, Narasimhalu K,et al. Rapid detection of differential itemfunctioning in assessments of health-related quality of life: the Functional As-sessment of Cancer Therapy. Qual LifeRes. 2007;16:101–114.

77 Crane PK, Gibbons LE, Ocepek-WeliksonK, et al. A comparison of three sets of cri-teria for determining the presence of dif-ferential item functioning using ordinal lo-gistic regression. Qual Life Res. 2007;16(suppl 1):69–84.

78 Crane PK, Hart DL, Gibbons LE, Cook KF.A 37-item shoulder functional status itempool had negligible differential item func-tioning. J Clin Epidemiol. 2006;59:478–484.

79 Crane PK, van Belle G, Larson EB. Test biasin a cognitive test: differential item func-tioning in the CASI. Stat Med. 2004;23:241–256.

80 Stata Statistical Software, release 9.2. Col-lege Station, TX: StataCorp LP; 2007.

81 Hanley JA, McNeil BJ. The meaning anduse of the area under a receiver operatingcharacteristic (ROC) curve. Radiology.1982;143:29–36.

82 Choi BC. Slopes of a receiver operatingcharacteristic curve and likelihood ratiosfor a diagnostic test. Am J Epidemiol.1998;148:1127–1132.

Screening for Fear of Work and Physical Activities

784 f Physical Therapy Volume 89 Number 8 August 2009

83 Sackett DL, Straus SE, Richardson WS,et al. Evidence-Based Medicine: How toPractice and Teach EBM. 2nd ed. NewYork, NY: Churchill Livingstone Inc; 2000.

84 Dujardin B, Van den Ende J, Van GompelA, et al. Likelihood ratios: a real improve-ment for clinical decision making? Eur JEpidemiol. 1994;10:29–36.

85 Jaeschke R, Guyatt G, Sackett DL;Evidence-Based Medicine Working Group.Users’ guides to the medical literature, III:how to use an article about a diagnostictest, A: Are the results of the study valid?JAMA. 1994;271:389–391.

86 DeLong ER, DeLong DM, Clarke-PearsonDL. Comparing the areas under two ormore correlated receiver operating char-acteristic curves: a nonparametric ap-proach. Biometrics. 1988;44:837–845.

87 Woby SR, Watson PJ, Roach NK, UrmstonM. Are changes in fear-avoidance beliefs,catastrophizing, and appraisals of control,predictive of changes in chronic low backpain and disability? Eur J Pain. 2004;8:201–210.

88 de Jong JR, Vlaeyen JW, Onghena P, et al.Fear of movement/(re)injury in chroniclow back pain: education or exposure invivo as mediator to fear reduction? Clin JPain. 2005;21:9–17; discussion 69–72.

89 Godges JJ, Anger MA, Zimmerman G, Del-itto A. Effects of education on return-to-work status for people with fear-avoidancebeliefs and acute low back pain. PhysTher. 2008;88:231–239.

90 Klaber Moffett JA, Carr J, Howarth E. Highfear-avoiders of physical activity benefitfrom an exercise program for patientswith back pain. Spine. 2004;29:1167–1172.

91 Werneke MW, Hart DL. Categorizing pa-tients with occupational low back pain byuse of the Quebec Task Force Classifica-tion system versus pain pattern classifica-tion procedures: discriminant and predic-tive validity. Phys Ther. 2004;84:243–254.

92 Embretson SE, Reise SP. Item ResponseTheory for Psychologists. Mahwah, NJ:Lawrence Erlbaum Associates; 2000.

93 Bode RK, Lai JS, Cella D, Heinemann AW.Issues in the development of an item bank.Arch Phys Med Rehabil. 2003;84(4 suppl2):S52–S60.

94 DeWalt DA, Rothrock N, Yount S, StoneAA. Evaluation of item candidates: thePROMIS qualitative item review. MedCare. 2007;45(5 suppl 1):S12–S21.

95 Tennant A, Penta M, Tesio L, et al. Assess-ing and adjusting for cross-cultural validityof impairment and activity limitationscales through differential item function-ing within the framework of the Raschmodel: the PRO-ESOR project. Med Care.2004;42(1 suppl):I37–I48.

96 Wright BD, Linacre JM. Observations arealways ordinal; measurements, however,must be interval. Arch Phys Med Rehabil.1989;70:857–860.

Appendix.Modified Fear-Avoidance Beliefs Questionnairea

Scale and Item Item Label

Fear-avoidance of physical activities

1. My pain was caused by physical activity. PHYSACTV

2. Physical activity makes my pain worse. WORSE

3. Physical activity might harm me.b HARM

4. I should not do physical activities which (might) make my pain worse. SHLDNOT

5. I cannot do physical activities which (might) make my pain worse. CANNOT

Fear-avoidance of work activities

6. My pain was caused by my work or by an accident at work. CAUSED

7. My work aggravated my pain. WORK

8. I have a claim for compensation for my pain. COMPENS

9. My work is too heavy for me. WRKHVY

10. My work makes or would make my pain worse. WRKWRSE

11. My work might harm me.b WRKHARM

12. I should not do my normal work with my present pain. WRKSHNT

13. I cannot do my normal work with my present pain. WRKCANT

14. I cannot do my normal work until my pain is treated. TREATED

15. I do not think that I will be back to my normal work within 3 months. MONTHS

16. I do not think that I will ever be able to go back to that work. GOBACK

a Modified and reprinted with permission of the International Association for the Study of Pain from:Waddell G, Newton M, Henderson I, et al. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the roleof fear-avoidance beliefs in chronic low back pain and disability. Pain. 1993:52:157–168.b Item was modified from the original wording by eliminating references to the back.

Screening for Fear of Work and Physical Activities

August 2009 Volume 89 Number 8 Physical Therapy f 785

F-A B Questionnaire

FABQ — PA Here are some of the things which other patients have told us about their pain. For each statement circle any number from 0 to 6 to say how much physical activities, such as bending, lifting, walking, or driving, affect or would affect your back pain.

Completely Disagree

Unsure Completely Agree

1. My pain was caused by physical activity. 0 1 2 3 4 5 6

2. Physical activity makes my pain worse. 0 1 2 3 4 5 6

3. Physical activity might harm my back. 0 1 2 3 4 5 6

4 I should not do physical activities which (might) make my pain worse. 0 1 2 3 4 5 6

5 I cannot do physical activities which (might) make my pain worse. 0 1 2 3 4 5 6

FABQ — WORK

The following statements are about how your normal work affects or would affect your back.

Completely Disagree

Unsure Completely Agree

6 My pain was caused by my work or by an accident at work. 0 1 2 3 4 5 6

7. My work aggravated my pain. 0 1 2 3 4 5 6

8. I have a claim for compensation for my pain. 0 1 2 3 4 5 6

9. My work is too heavy for me. 0 1 2 3 4 5 6

10. My work makes or would make my pain worse. 0 1 2 3 4 5 6

11. My work might harm my back. 0 1 2 3 4 5 6

12 I should not do my regular work with my present pain. 0 1 2 3 4 5 6

13. I cannot do my normal work with my present pain. 0 1 2 3 4 5 6

14. I cannot do my normal work until my pain is treated. 0 1 2 3 4 5 6

15. I do not think that I will be back to my normal work within 3 months. 0 1 2 3 4 5 6

16. I do not think that I will ever be able to go back to work. 0 1 2 3 4 5 6

Print your name _____________________________ Signature ________________________Date ___________

FABQ-PA ____________ FABQ-WORK _________

Fear-Avoidance Beliefs Questionnaire (FABQ) for Patients with Back Pain

Overview:

The Fear-Avoidance Beliefs Questionnaire (FABQ) can help measure how much fear and avoidance are affecting a patient with low back pain. This can help identify those patients for whom psychosocial interventions may be beneficial. The authors are from the Western Infirmary in Glasgow (Scotland) and the Hope Hospital in Salford (England).

NOTE: This scale can be modified to apply to patients with other types of chronic pain. Only items 3 and 11 mention "back".

Instructions: Here are some of the things which other patients have told us about their pain. For each statement please circle the number from 0 to 6 to say how much physical activities such as bending lifting walking or driving affect or would affect your back pain.

Statements:

(1) My pain is caused by physical activity.

(2) Physical activity makes my pain worse.

(3) Physical activity might harm my back.

(4) I should not do physical activities which (might) make my pain worse.

(5) I cannot do physical activities which (might) make my pain worse.

The following statements are about how your normal work affects or would affect you back pain:

(6) My pain was caused by my work or by an accident at work.

(7) My work aggravated my pain.

(8) I have a claim for compensation for my pain.

(9) My work is too heavy for me.

(10) My work makes or would make my pain worse.

(11) My work might harm my back.

(12) I should not do my normal work with my present pain.

(13) I cannot do my normal work with my present pain.

(14) I cannot do my normal work till my pain is treated.

(15) I do not think that I will be back to my normal work within 3 months.

(16) I do not think that I will ever be able to go back to that work.

Response Points

completely disagree 0

1

2

unsure 3

4

5

completely agree 6

fear-avoidance beliefs about work (scale 1) =

= (points for item 6) + (points for item 7) + (points for item 9) + (points for item 10) + (points for item 11) + (points for item 12) + (points for item 15)

fear-avoidance beliefs about physical activity (scale 2) =

= (points for item 2) + (points for item 3) + (points for item 4) + (points for item 5)

items not in scale 1 or 2: 1 8 13 14 16

Interpretation:

• minimal scale scores: 0

• maximum scale 1 score: 42 (7 items)

• maximum scale 2 score: 24 (4 items)

• The higher the scale scores the greater the degree of fear and avoidance beliefs shown by the patient.

Performance:

• Internal consistency (alpha) 0.88 for scale 1 and 0.77 for scale 2

References:

Waddell G Newton M et al. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain. 1993; 52: 157-168 (Appendix page 166).

The Lower Extremity Functional Scale(LEFS): Scale Development, MeasurementProperties, and Clinical Application

Background and Purpose. The purpose of this study was to assess thereliability, construct validity, and sensitivity to change of the LowerExtremity Functional Scale (LEFS). Subjects and Methods. The LEFS wasadministered to 107 patients with lower-extremity musculoskeletal dys-function referred to 12 outpatient physical therapy clinics. Methods. TheLEFS was administered during the initial assessment, 24 to 48 hoursfollowing the initial assessment, and then at weekly intervals for 4 weeks.The SF-36 (acute version) was administered during the initial assessmentand at weekly intervals. A type 2,1 intraclass correlation coefficient wasused to estimate test-retest reliability. Pearson correlations and one-wayanalyses of variance were used to examine construct validity. Spearmanrank-order correlation coefficients were used to examine the relationshipbetween an independent prognostic rating of change for each patient andchange in the LEFS and SF-36 scores. Results. Test-retest reliability of theLEFS scores was excellent (R5.94 [95% lower limit confidence interval(CI)5.89]). Correlations between the LEFS and the SF-36 physicalfunction subscale and physical component score were r 5.80 (95% lowerlimit CI5.73) and r 5.64 (95% lower limit CI5.54), respectively. Therewas a higher correlation between the prognostic rating of change and theLEFS than between the prognostic rating of change and the SF-36 physicalfunction score. The potential error associated with a score on the LEFS ata given point in time is 65.3 scale points (90% CI), the minimaldetectable change is 9 scale points (90% CI), and the minimal clinicallyimportant difference is 9 scale points (90% CI). Conclusion and Discus-sion. The LEFS is reliable, and construct validity was supported bycomparison with the SF-36. The sensitivity to change of the LEFS wassuperior to that of the SF-36 in this population. The LEFS is efficient toadminister and score and is applicable for research purposes and clinicaldecision making for individual patients. @Binkley JM, Stratford PW, LottSA, et al. The Lower Extremity Functional Scale (LEFS): scale develop-ment, measurement properties, and clinical application. Phys Ther.1999;79:371–383.#

Key Words: Disability, Function, Functional scales, Sensitivity to change, Tests and measurements.

Physical Therapy . Volume 79 . Number 4 . April 1999 371

Rese

arch

Repo

rt

Jill M BinkleyPaul W StratfordSue Ann LottDaniel L RiddleThe North AmericanOrthopaedic RehabilitationResearch Network*

v

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

Proposed new measures of health status shouldbe viewed with increasing rigor and sophistica-tion with respect to scale development.1,2

Numerous generic and disease-specificself-report measures that are suitable for physical ther-apy clinical practice and that have adequate measure-ment properties now exist. Generic health status mea-sures assess overall health, including social, emotional,and physical health status, and are intended to beapplicable across a broad spectrum of diseases, interven-tions, and demographic and cultural subgroups.1,2

Condition-specific measures, also termed “disease-specificmeasures,” are designed to assess attributes that are mostrelevant to the disease or condition of interest.1,2 Ideally,disease-specific measures are composed of items thatare frequently affected by the condition of interest andthat are likely to demonstrate clinically importantchange.

Several barriers to the widespread adoption of genericand condition-specific measures in clinical practice havebeen identified, including (1) difficulty of administeringthe scales and of scoring, (2) difficulty in administeringscales for different conditions and anatomical sites,(3) the practitioner’s belief that there is a lack of clinical

meaningfulness for the scores, and (4) inadequate mea-surement properties for application to individualpatients.1,3,4 A limitation often cited by clinicians is thatmany measures are required for a practice that serves avaried caseload. For example, there are numerouscondition-specific measures available for people withknee conditions. Measures exist for people with generalconditions of the knee,5 patellofemoral joint disor-ders,6,7 ligamentous deficiency,8–10 and joint replace-ment.11,12 It is conceivable that 4 or 5 measures would berequired to accommodate people with knee dysfunction,not to mention the number of measures that may berequired to assess people with other lower-extremity,upper-extremity, and spinal problems.

One approach to overcoming the need for multiplemeasures of health status in clinical practice is to explorewhether the measurement properties of condition-specific measures are superior to those of generic mea-sures. Should the measurement properties be similar, asingle generic measure or subscale of that measurecould be used in place of a number of condition-specificmeasures. Several generic measures have been appliedto a variety of patients with lower-extremity musculoskel-etal conditions, including the SF-36,13,14 the SF-12,15 the

JM Binkley, PT, FCAMT, FAAOMPT, is Physical Therapist, Appalachian Physical Therapy, 109A Tipton Dr, Dahlonega, GA 30534 USA(binkley01@sprynet.com), and Assistant Professor (Physical Therapy), School of Rehabilitation Science, McMaster University, Hamilton, Ontario,Canada. Address all correspondence to Ms Binkley at the first address.

PW Stratford, PT, is Associate Professor, School of Rehabilitation Science, and Associate Member, Department of Clinical Epidemiology andBiostatistics, McMaster University, Hamilton, Ontario, Canada.

SA Lott, PT, is Physical Therapist, Appalachian Physical Therapy, Dahlonega, Ga, and Clinical Professor, Physical Therapy Program, North GeorgiaCollege, Dahlonega, Ga.

D Riddle, PhD, PT, is Associate Professor, Department of Physical Therapy, Virginia Commonwealth University, Richmond, Va.

*North American Orthopaedic Rehabilitation Research Network is: Brad Balsor, PT, St Joseph’s Hospital, Hamilton, Ontario, Canada; Paul Beattie, PhD, PT, Department ofPhysical Therapy, University of Rochester, Rochester, NY; Andrew Berk, PT, Summit Injury Management, Victoria, British Columbia, Canada; Jill Binkley, PT, FCAMT,FAAOMPT, Appalachian Physical Therapy, Dahlonega, Ga; Susan Brenneman, PT, Penn Therapy and Fitness, Philadelphia, Pa; Linda Brett, PT, KakabekaPhysiotherapy, Kakabeka Falls, Ontario, Canada; Jane Burns, PT, Pacific Coast Rehabilitation Center, Vancouver, British Columbia, Canada; Bert Chesworth, PT,FCAMT, University of Western Ontario, London, Ontario, Canada; Doug Conroy, PT, ATC, Conroy Orthopaedic and Sports Physical Therapy, Flossmoor, Ill; RobertFeehley, PT, OCS, Baltimore Sports Rehab, Baltimore, Md; Karen Hayes, PhD, PT, Program in Physical Therapy, Northwestern University Medical School, Chicago,Ill; Scott Hyams, PT, Heartland Healthcare, Sunrise, Fla; Michael Kelo, PT, OCS, Sheltering Arms Physical Rehabilitation Hospital, Chester, Va; Carmen Kirkness, PT,McGill University, Montreal, Quebec, Canada; Kim Kramer, PT, Sartori Hospital, Cedar Falls, Iowa; Jim Krzaczek, PT, OCS, Life Care Medical Center, Glassboro, NJ;Sue Ann Lott, PT, Appalachian Physical Therapy, Dahlonega, Ga; Jane Mennie, PT, Fannin Regional Hospital, Blue Ridge, Ga; Jay Neel, Appalachian Physical Therapy,Dahlonega, Ga; Karen Orlando, PT, Physiotherapy on Bay, Toronto, Ontario, Canada; Beverly Padfield, PT, FCAMT, Four Counties Health Services, Newbury, Ontario,Canada; Corinne Roos, Kettle Creek Physiotherapy and Sports Injuries Clinic, St Thomas, Ontario, Canada; Linda Nolte Smith, PT, MTC, Park at Stony Point PhysicalTherapy, Richmond, Va; Dan Riddle, PhD, PT, Virginia Commonwealth University, Richmond, Va; Gregory Spadoni, PT, ProActive Physiotherapy, Hamilton, Ontario,Canada; Diane Stratford, PT, West End Physiotherapy Clinic, Hamilton, Ontario, Canada; Paul Stratford, PT, McMaster University, Hamilton, Ontario, Canada; MarcusWalser, PT, Walser Physiotherapy, Thunder Bay, Ontario, Canada; Linda Watts, PT, Algoma Physical Rehabilitation Clinic, Sault Ste Marie, Ontario, Canada; MichaelWestaway, PT, FCAMT, Canadian Sport Rehabilitation Institute, Calgary, Albert, Canada; Myra Westaway, PT, HSC, Lindsay Park Sport Rehab, Calgary, Albert, Canada.

Approval for this study was obtained from the institutional review board associated with the North American Orthopaedic Rehabilitation ResearchNetwork based in Dahlonega, Ga. In addition, local institutional review board approval was obtained by clinicians and clinics participating in thestudy, where necessary.

This project was funded in part by a grant from the Section on Research of the American Physical Therapy Association.

This article was submitted February 24, 1998, and was accepted January 4, 1999.

372 . Binkley et al Physical Therapy . Volume 79 . Number 4 . April 1999

Functional Status Index,16 and the MusculoskeletalFunctional Assessment Questionaire.4,17,18 The SF-36,SF-12, and Musculoskeletal Functional Assessment Ques-tionnaire have been validated for group decision makingonly.3,4,15,18 The Functional Status Index has been doc-umented to yield reliable and valid measurements inpatients with total hip replacement, but sensitivity tochange and application to other orthopedic conditionshave not been reported for the Functional StatusIndex.16

The SF-36 has served as the principal generic measure forcomparisons with condition-specific measures.4,8,12,19–25

The SF-36 consists of 8 health concept subscales (physicalfunction, role physical, bodily pain, general health, vitality,social function, role emotional, and mental health) and 2component summary scores. Each subscale score can varyfrom 0 to 100, with higher scores representing moredesirable health states. The physical and mental compo-nent summary scores represent weighted composite scoresderived from the 8 health concept scales. Each of thecomponent summary scores is scaled to have a mean of 50and a standard deviation of 10 for the general populationof the United States. To date, the responsiveness of severalof the SF-36 subscales and the physical component sum-mary score have been shown to be superior or equivalent tocondition-specific scales relevant to the lower extremi-ty.20,23,25–27 Due to the utilization of designs that are notthe most rigorous available28 and because formal statis-tical comparisons between the observed change indexeswere not reported, it is not known whether theseobserved differences represent true differences in themeasures’ capacities or whether they are merely a resultof sampling variation. There is no strong evidence tosuggest that existing condition-specific scales designedfor the lower extremity are superior to the SF-36. TheSF-36, however, is time-consuming to administer andscore in the clinic and was not designed for individualpatient decision making.

It is critical that measures of health status be reliable,valid, and responsive to clinical change that occurs overtime.2,29 The terms “responsiveness” and “sensitivity tochange” are often used interchangeably to describe theability of a measure to detect clinical change.2,30–32

Responsiveness, as defined by Kirshner and Guyatt,33

denotes the ability of a scale to detect change. WithinKirshner and Guyatt’s taxonomy, responsiveness existsindependent of validity. This position has been chal-lenged by Hays and Hadorn,34 who suggested thatresponsiveness is actually one indication of a measure’svalidity. An external standard of change was introducedto examine the extent to which a health status measuretruly differentiates among patients who have improved,deteriorated, or remained stable and subsequently usedin our study and by other authors.27,29,34–39 We used the

term “sensitivity to change” to denote the ability of ameasure to detect true change in patients’ status overtime, as we have done in our recent publications.35,36

Using this definition, sensitivity to change is a form ofvalidity.33,34 Because no criterion standard exists forassessing change in health status, construct validation isused to identify patients or groups of patients who areexpected to change by differing amounts. Several meth-ods have been used by us and by other researchers in anattempt to distinguish among patients’ levels of change,including other clinical measures (eg, spinal flexion),29

retrospective global rating of change,27,29,37–41 theachievement of treatment goals,35 and an external prog-nostic rating of change.42 In a report co-authored by oneof the present investigators, the authors contended thata bias can be introduced when the retrospective globalrating is performed at the time of the follow-upassessment.42

We believe that there is a need for a functional measurethat is easy to administer and score and applicable toa wide range of patients with lower-extremity ortho-pedic conditions. Our goal was to develop a self-reportcondition-specific measure that would yield reliable andvalid measurements and that would be appropriate foruse as a clinical and research tool. Accordingly, the scaledevelopment process took into account the barriersidentified for clinical implementation of self-report mea-sures. The purpose of this article is to report on thedevelopment and initial validation of a newly developedcondition-specific measure, the Lower Extremity Func-tional Scale (LEFS), including the determination ofinternal consistency, reliability, construct validity, sensi-tivity to change, and clinical application.

Method

SubjectsSubjects were consecutive patients referred for physicaltherapy with any lower-extremity musculoskeletal condi-tion (defined as any condition of the joints, muscles, orother soft tissues of the lower extremity). Patients whodid not speak English or were unable to read wereexcluded from the study. Data were collected over a4-month period. A total of 107 patients were enteredinto the study. A description of the patients is presentedin Table 1.

Data were collected in physical therapy clinics affiliatedwith the North American Orthopaedic RehabilitationResearch Network (NAORRN). At the time of the study,the NAORRN consisted of 19 physical therapy cliniciansand 5 physical therapy researchers in the United Statesand Canada. The NAORRN was designed to supportmulticentered research in the orthopedic field. Involve-ment in the NAORRN is voluntary and unfunded.

Physical Therapy . Volume 79 . Number 4 . April 1999 Binkley et al . 373

IIIIII

IIIIII

IIIIII

IIIIII

I

Twelve of the 19 clinicians contributed data to the study(Fig. 1). Informed consent was obtained from allpatients.

The LEFSThe conceptual framework that guided the developmentof the LEFS included that the scale (1) be based on theWorld Health Organization’s model of disability andhandicap,43 (2) be efficient to administer, score,and record in the medical record with respect to patientand clinician time, (3) be applicable to a wide variety ofpatients with lower-extremity orthopedic conditions,including patients with a range of disability levels, con-ditions, diseases, treatments, and ages, (4) be applicablefor documenting function on an individual patient basisas well as in groups, such as for clinical outcomesassessment and clinical research purposes, (5) be devel-oped using a systematic process of item selection anditem scaling,2 (6) yield reliable measurements (haveinternal consistency and test-retest reliability), and(7) yield valid measurements (at a single point in timeand sensitive to valid change).

Items were generated for the LEFS by a process ofreviewing existing questionnaires as well as surveyingclinicians and patients. The World Health Organiza-tion’s model of disability43 served as the basis for theitem generation phase of the scale development. Theterminology used to define disability and handicap wasused as the basis of questions posed about functionallimitations to patients. Thirty-five patients with a varietyof lower-extremity orthopedic conditions were surveyedto determine important functional limitations associatedwith their problem. Patients were asked to “identify up to3 important activities that you are unable to do or arehaving difficulty with because of your lower-limb prob-lem?” From this survey, we selected 75 items and col-lapsed them to 22 items by grouping similar activities.“Walking on uneven ground” and “walking on grass,” forexample, were 2 activities that were grouped together.Three orthopedic physical therapists, each with at least10 years of experience in orthopedic physical therapypractice, reviewed the 22 items and were given theopportunity to add additional items. We surveyed exist-ing questionnaires. No additional items were identifiedas important to include in the LEFS by these additionalprocesses.

The initial version of the scale consisted of 22 items. Theintroductory statement of the questionnaire states:“Today, do you or would you have any difficulty at allwith:” followed by a listing of the functional items. Itemsare rated on a 5-point scale, from 0 (extreme difficulty/unable to perform activity) to 4 (no difficulty). The5-point difficulty rating scale was selected to maximizethe capacity of the scale to measure change (Appendix).

Table 1.Demographic Characteristics of Patients (N5107)

Characteristic Summary

Age (y)X 44.0SD 16.2

SexMale 46Female 58Missing 3

Time since onset (wk)X# 6Minimum 0Maximum 2501st quartile 33rd quartile 11n 98

Site of problemHip 2Thigh 1Knee 71Leg 8Ankle 14Foot 8Missing 3

SurgeryYes 62No 44Missing 1

ConditionHip

Osteoarthritis 1Muscle strain 1Missing 0

Knee/thigh/legLigament sprain 22Muscle strain 5Meniscal injury 10Osteoarthritis 9Patellofemoral pain 6Fracture 3Total joint replacement 8Nonspecific sprain/strain 12Missing 5

Foot/ankleLigament sprain 9Muscle strain 1Osteoarthritis 2Fracture 8Nonspecific sprain/strain 2Missing 0

Educational level completedElementary 13Secondary 46University 41Missing 7

SmokerYes 19No 85Missing 3

374 . Binkley et al Physical Therapy . Volume 79 . Number 4 . April 1999

The initial version of the scale wasadministered to 57 patients whowere referred for physical therapywith lower-extremity dysfunction.The anatomical sites represented inthis group of patients were: foot andankle (n512), knee (n529), hip(n58), multiple trauma, (n55), andmissing (n53). The broad catego-ries of orthopedic conditions in thisgroup were: sprains and strains(n532), fractures and bone disor-ders (n59), osteoarthritis (n58),articular subluxation or dislocation(n57), and missing (n51). TotalLEFS scale scores, means, and scoredistribution were determined forthis group. The mean LEFS score,out of a possible score of 88, was 39(SD518.0, median score540.0,range52–85).

At the individual item level, meanscore, median score, standard devia-tion, range, and frequency ofendorsement of each level (0–4) ofall items were determined. Inter-item correlations and correcteditem-item total correlations were cal-culated. The corrected item-item total correlation is anestimate of the degree to which a single item scorecorrelates with the total scale score with that itemremoved. The alpha coefficient, a measure of internalconsistency, was determined for the scale and calculatedwith each of the items removed. The overall goal of thisanalysis was to ensure that individual item scores werereasonably normally distributed, with mean initial scoresof about 50% of the items at approximately the mid-point44 of the scale. In order to develop a measure thatis applicable to a spectrum of conditions and levels ofdisabilities, the remaining items were selected to repre-sent different difficulty levels, as indicated by item meanscores that were higher and lower than the midpoint. Asa result of the item analysis, 2 items were removed fromthe original LEFS and 1 item was reworded. A factoranalysis performed on the final 20-item questionnaireindicated that all items loaded on a single factor. Thefactor loadings varied from .44 (walking between rooms)to .86 (performing heavy activities around the house),with 19 of the factor loadings between .58 and .86.

The final version of the LEFS consists 20 items, each witha maximum score of 4. The total possible score of 80indicates a high functional level (Appendix). The scale isone page, can be filled out by most patients in less than2 minutes, and is scored by tallying the responses for allof the items. Scoring is performed without the use of a

calculator or computer and requires approximately 20seconds.

Procedure

Reliability. The LEFS was administered during the ini-tial assessment to patients with lower-extremity musculo-skeletal dysfunction referred for physical therapy. TheLEFS was readministered to patients 24 to 48 hoursfollowing the initial administration in order to examinetest-retest reliability. The LEFS was then administered atweekly intervals (within 7 days, 61 day) for 4 weeks oruntil patients were discharged (in cases where dischargeoccurred prior to 4 weeks). In addition, the SF-36 (acuteversion) was administered during the initial assessmentand at the weekly follow-up assessments. These follow-upintervals allowed examination of the validity of the LEFSmeasurements as well as comparison of sensitivity tochange between the LEFS and SF-36.

Construct validity. In the absence of an accepted mea-sure of function, determination of the validity of func-tional scales has relied heavily on the concept of con-struct validity. One or more theories are developed, andthe extent to which a measure yields results concordantwith the theory provides support for the validity of themeasure. In this study, we believed that validity for ourmeasure would be supported if:

Figure 1.Clinic locations and numbers of patients entered in study.

Physical Therapy . Volume 79 . Number 4 . April 1999 Binkley et al . 375

IIIIII

IIIIII

IIIIII

IIIIII

I

1. There would be a moderate correlation (r..6)between LEFS scores and SF-36 physical functionsubscale and SF-36 physical component summaryscores at the initial assessment and at the 3-weekfollow-up assessment.

2. There would be a low correlation (r ,.5) betweenLEFS scores and SF-36 mental health subscale andSF-36 mental component summary scores at the ini-tial assessment and at the 3-week follow-up assess-ment.

3. Patients who had recently undergone surgery (sur-gery less than 2 weeks prior to initial assessment)would have lower LEFS and SF-36 physical functionsubscale and physical component summary scoresthan would patients who did not have recent surgery(no surgery or surgery greater than 2 weeks prior toassessment).

4. Patients with acute conditions would demonstratelower LEFS scores and SF-36 physical function sub-scale and physical component summary scores thanwould patients with chronic conditions.

The SF-36 was selected as the comparison scale forexamination of the construct validity of the LEFS scores.The selection of the SF-36 was based on the literaturedocumenting the measurement properties of theSF-36, including its applicability to patients with lower-extremity dysfunction. The reliability, validity, and respon-siveness of measurements obtained with the SF-36 havebeen documented in diverse patient groups. The physicalfunction and pain dimensions appear to be most relevantto orthopedic outpatients.20,37 Although several of theSF-36 subscales have the capacity to measure change onoutpatients with musculoskeletal conditions, several of thesubscales do not change or change minimally in thispopulation.20,37 The mental health subscale of the SF-36demonstrates minimal change in outpatients with muscu-loskeletal conditions.20,37

In order to examine our argument for validity, whichspecified that patients with acute conditions would dem-onstrate more functional limitation than patients withchronic conditions, all patients were assigned a chronic-ity rating on a 3-point scale by 2 orthopedic physical

therapists blinded to patients’ functional scale scores.The ratings were performed allowing discussion of thepatient profile, and a single agreed-on score was deter-mined. Patients were placed in one of the followingcategories based on a review of documentation, whichincluded diagnosis and the time since onset of condition(or the time since surgery or cast removal): (1) acute—less than 4 weeks since onset of condition, surgery, orpostfracture immobilization, (2) moderate/unclear, or(3) chronic—more than 4 weeks since onset of condi-tion or having a chronic condition such as osteoarthritis.The basis for the selection of 4 weeks was the judgmentof the investigators.

Sensitivity to change. Sensitivity to change was exam-ined using a prognosis rating. Each patient was given arating of prognosis using a 7-point scale (Fig. 2). Twoorthopedic physical therapists who were blind to thepatient’s functional scale scores performed independentprognostic ratings on each patient, which were subse-quently averaged. Prognostic ratings were based ondocumentation review of patients’ diagnoses, age, chro-nicity, number of comorbid conditions, and type ofsurgery and time since surgery, where applicable. Ratersanswered the questions “How much change would youexpect in this patient at 1 week following the initialassessment?” and “How much change would you expectin this patient at 3 weeks following the initial assess-ment?” We believed that, if our assumption about valid-ity was correct, there would be a correlation between (1)the 1-week LEFS and SF-36 scores and the 1-weekprognostic ratings and (2) the 3-week LEFS and SF-36scores and the 3-week prognostic ratings. This approachwas based on clinical judgment and previous work byWestaway et al,45 whose data suggested that experiencedclinicians can make prognoses about patients. Thecapacity of the LEFS and the SF-36 physical functionsubscale and physical component summary scores tomeasure valid change was compared at 1 week and at 3weeks using this theory for change.

The interrater reliability for the prognostic ratings wasdetermined using a type 3,2 intraclass correlation coef-ficient (ICC). This class of ICC is appropriate whenratings are averaged and an adjustment has been appliedto address a systematic difference between raters.46 Theinterrater reliability of the prognostic rating was R5.84

Figure 2.Prognostic rating of change scale.

376 . Binkley et al Physical Therapy . Volume 79 . Number 4 . April 1999

(95% lower limit confidence interval @CI#5.78). Becausethe goal of the analysis was to examine change, ratherthan to evaluate intervention, we made no attempt tocontrol interventions.

Data Analysis

Internal consistency, reliability, and minimal detectablechange. We used the alpha coefficient to estimateinternal consistency, a measure of homogeneity ofitems.2 A type 2,1 intraclass correlation coefficient (ICC)was used to estimate test-retest reliability.46 Becausemany patients with musculoskeletal problems demon-strate true change over a short period of time, test-retestreliability was estimated over a 24- to 48-hour periodusing the entire patient sample and a subset of patientswho were deemed to have more chronic conditions, asdetermined by the chronicity rating described above,and who were presumably more stable.

Reliability of the LEFS scores was also quantified usingthe standard error of measurement (SEM), a represen-tation of measurement error expressed in the same unitsas the original measurement, in this case, LEFS points.Two estimates of the SEM were obtained. The firstestimate, based on the alpha coefficient, was used toquantify measurement error at the 90% confidence levelabout a patient’s score at a single point in time. Thisquantification was achieved by multiplying the SEM bythe z value associated with the 90% confidence level(ie, z51.65). The test-retest reliability coefficientobtained for the subset of patients with more chronicconditions was used to estimate the SEM that was used tocalculate minimal detectable change (MDC) at the 90%confidence level. To obtain this estimate, the SEM ismultiplied by the z value for the confidence level ofinterest, and this quantity is multiplied by the squareroot of 2.47

Construct validity. Pearson correlation coefficients and95% one-sided lower limit confidence intervals werecalculated to examine the relationship between theLEFS scores and the SF-36 subscale and componentsummary scores at the initial assessment. One-way anal-yses of variance were used to examine the hypothesesabout validity that specified that there would be adifference in initial LEFS scores and SF-36 physicalfunction subscale and physical component summaryscores between: (1) patients with recent surgery andpatients without recent surgery and (2) patients withacute conditions and patients with chronic conditions.

Sensitivity to change and minimal clinically importantdifference. Spearman rank-order correlation coeffi-cients were used to examine the relationship betweenthe prognostic rating and change in the following func-

tional status scores at 1 week and 3 weeks: LEFS score,SF-36 physical function subscale score, SF-36 physicalcomponent summary score, and SF-36 mental compo-nent summary score. The magnitudes of the correlationsbetween the prognostic ratings and the LEFS, SF-36physical function subscale, and SF-36 physical compo-nent summary scores were formally compared using themethod of Williams48 for dependent data.

The minimal clinically important difference (MCID),defined as the minimal amount of change on the scalerequired to be considered a clinically important change,was determined using 2 methods. In the first approach,we used the prognostic ratings of change to separatepatients into those who were predicted to undergoimportant change (prognostic ratings of 2, 3, and 4) andthose who were predicted to undergo no importantchange at 3 weeks (prognostic ratings of 0 and 2). Thecutpoint of change on the LEFS that maximized the areaunder a receiver operating characteristic (ROC) curvewas determined as the estimate of the amount of changeon the LEFS that best classified patients who hadchanged an important amount from those who hadnot.35,36 Sensitivity and specificity for this cutpoint valuewere determined. The second approach was a survey of5 clinicians who reported that they had used the LEFSfor an minimum of 4 months and on at least 10 patientsas a clinical decision-making tool. Clinicians were askedto estimate the amount of change that they wouldconsider to be clinically important for initial LEFS scoresof 10, 25, 40, 55, and 70. Clinicians were asked to identifythe minimal amount of change on the LEFS, in scalepoints, that would suggest that improvement hadoccurred. The same question was posed to clinicians interms of deterioration. Clinicians’ judgments of MCIDwere compared with the statistical approach.

Results

Descriptive StatisticsPatients’ characteristics are shown in Table 1. Descrip-tive statistics for the patients by measure are presented inTable 2. None of the patients received the minimum ormaximum scores for the LEFS at any of the assessments.Minimum and maximum SF-36 physical function sub-scale scores were obtained for 1 and 4 patients, respec-tively, at the initial assessment. Minimum and maximumSF-36 physical function subscale scores were eachobtained for 1 patient at the 3-week follow-upassessment.

ReliabilityInternal consistency was a5.96 (N5107). Test-retestreliability estimates were R5.86 (95% lower limitCI5.80) for the entire sample (n598) and R5.94 (95%

Physical Therapy . Volume 79 . Number 4 . April 1999 Binkley et al . 377

IIIIII

IIIIII

IIIIII

IIIIII

I

lower limit CI5.89) for the subset of patients with morechronic conditions (n531).

Construct ValidityTable 3 includes the validity coefficient estimates at theinitial assessment between the LEFS and SF-36 scores.Correlations between the LEFS scores and the SF-36physical function subscale and physical component sum-mary scores were r 5.80 (95% lower limit CI5.73) andr 5.64 (95% lower limit CI5.54). The correlationbetween the LEFS scores and the SF-36 mental compo-

nent summary scores was r 5.30 (95% lower limitCI5.14). There was a difference in LEFS scores betweenthe patients with recent surgery and the patients withoutrecent surgery at the initial assessment (P5.006)(Tab. 4). There was a difference in LEFS scores betweenthe patients with acute conditions and the patients withchronic conditions (P5.027). There was no difference inSF-36 physical function subscale, physical componentsummary, and mental component summary scoresbetween the patients with recent surgery and thepatients without recent surgery (P5.117) or between the

Table 2.Descriptive Data for Lower Extremity Functional Scale (LEFS) and SF-36 Physical Function Subscale and Physical and Mental ComponentSummary Scores

SampleSize X SD Median Minimum Maximum

Skewness(SE) Kurtosis (SE)

LEFS (/80)Initial 107 34 16 34 5 74 0.34 (0.23) 20.57 (0.46)Retest (24–48 h) 98 37 17 36 5 76 0.23 (0.24) 20.56 (0.48)Week 1 101 43 18 42 2 77 0.02 (0.24) 20.84 (0.47)Week 2 90 45 16 45 9 77 0.16 (0.26) 20.70 (0.50)Week 3 58 49 17 50 6 78 20.21 (0.31) 20.81 (0.61)Week 4 35 51 18 50 18 78 20.33 (0.40) 21.01 (0.78)

SF-36 physical function subscalescores(/100)

Initial 101 35 22 35 0.0 100 0.56 (0.24) 20.01 (0.47)Week 1 101 46 24 40 10 100 0.46 (0.24) 20.98 (0.47)Week 2 88 50 25 45 5 100 0.26 (0.26) 20.96 (0.51)Week 3 55 53 24 52.5 0 100 0.04 (0.32) 20.83 (0.63)Week 4 35 56 24 60 10 100 20.27 (0.40) 20.92 (0.78)

SF-36 physical component summaryscoresa

Initial 101 31 8 30 19 58 0.78 (0.78) 0.17 (0.47)Week 1 101 36 9 34 19 58 0.61 (0.24) 20.37 (0.48)Week 2 88 37 10 35 20 58 0.33 (0.26) 20.91 (0.51)Week 3 55 38 9 36.5 21 58 0.25 (0.32) 20.94 (0.63)Week 4 35 40 9 39 26 58 0.11 (0.40) 21.16 (0.78)

SF-36 mental component summaryscoresa

Initial 101 52 11 54 20 73 20.68 (0.24) 0.02 (0.47)Week 1 101 54 9 56 31 70 20.48 (0.24) 20.31 (0.48)Week 2 87 55 9 56 34 70 20.52 (0.26) 20.29 (0.51)Week 3 55 56 9 59 29 72 20.91 (0.32) 0.24 (0.63)Week 4 35 56 9 61 34 69 20.95 (0.40) 20.33 (0.78)

a Component summary scores are scaled to have a mean of 50 and a standard deviation of 10 in the general population of the United States.

Table 3.Validity Coefficient Estimatesa for Lower Extremity Functional Scale (Single Point in Time) (n5100)

SF-36 Subscale and Component Summary Scores

PhysicalFunction

RolePhysical

BodilyPain

GeneralHealth Vitality

SocialFunction

RoleEmotional

MentalHealth

Physical ComponentSummary Score

Mental ComponentSummary Score

.80 .51 .49 .09 .43 .62 .32 .23 .64 .30(.73) (.38) (.35) (2.08) (.28) (.51) (.16) (.07) (.54) (.14)

a Lower 1-sided 95% confidence interval shown in parentheses.

378 . Binkley et al Physical Therapy . Volume 79 . Number 4 . April 1999

patients with acute conditions and the patients withchronic conditions (P5.471) (Tab. 4).

Sensitivity to ChangeThe correlations relating to change scores are presentedin Table 5. There was no difference between the corre-lations of the prognostic rating with the LEFS and theprognostic rating with the SF-36 physical function sub-scale score at the initial assessment (t(95)51.24,P(1)5.106). There was a difference between the correla-tions of the prognostic rating with the LEFS and theprognostic rating with the SF-36 physical componentsummary score at the initial assessment (t(95)51.67,P(1)5.05). There was a difference between the correla-tions of the prognostic rating with the LEFS and theprognostic rating with the SF-36 physical function sub-scale score at week 3 (t(95)53.05, P(1)5.002). There was

also a difference between the correlations of the prog-nostic rating with the LEFS and the prognostic ratingwith the SF-36 physical component summary score atweek 3 (t(95)52.13, P(1)5.019).

Individual Patient Decision MakingWith respect to individual patient decision making, thepotential error associated with a score on the LEFS at agiven point in time is 65.3 scale points on the 80-pointscale (90% CI) (Tab. 6). The MDC is 69 scale points(90% CI). The MCID is approximately 9 scale points.The area under the ROC curve associated with this valueis .76, and the sensitivity and specificity are .81 and .70,respectively. The average of the 5 clinician estimates forMCID was 10 scale points, suggesting that the statisticalapproach has resulted in a reasonable estimate ofthe MCID.

Table 4.Construct Validity: Comparison of Lower Extremity Functional Scale (LEFS) and SF-36 Physical Function Subscale Scores for Subgroups of PatientsWith and Without Surgery and Patients With Acute, Moderate, and Chronic Conditions at Initial Assessment

Subgroup N X SD F P

LEFS No surgery 58 38.2 15.8 7.8 .006Surgery 42 29.6 14.2

SF-36 physical function subscale No surgery 58 38.8 22.4 2.49 .117Surgery 42 31.9 20.2

LEFS Acute 29 27.7 13.6 3.8 .027Moderate/unclear 37 36.9 15.7Chronic 30 37.4 16.9

SF-36 physical function subscale Acute 29 31.6 21.5 0.76 .471Moderate/unclear 37 37.7 21.3Chronic 30 37.5 23.7

Table 5.Validity Coefficient Estimates of Sensitivity to Change for Lower Extremity Functional Scale (LEFS) and SF-36 Physical Function Subscale andPhysical and Mental Component Summary Scoresa

LEFSChange

SF-36 PhysicalFunctionSubscale Change

SF-36 PhysicalComponentSummary ScoreChange

SF-36 MentalComponentSummary ScoreChange

Prognosisat Week 1

LEFS change .57 .26 .13 .36(.44) (.10) (2.04) (.20)

SF-36 physical .77 .57 2.05 .25function subscale change (.66) (.44) (2.22) (.09)

SF-36 physical .64 .81 2.45 .16component summary score change (.48) (.71) (2.57) (2.01)

SF-36 mental .25 .16 2.12 .14component summary score change (.02) (2.07) (2.34) (2.03)

Prognosis at week 3 .64 .42 .45 .28(.48) (.21) (.25) (2.06)

a Upper diagonal represents correlations (95% confidence interval) at week 1 (n598); lower diagonal represents correlations (95% confidence interval) at week 3(n554).

Physical Therapy . Volume 79 . Number 4 . April 1999 Binkley et al . 379

IIIIII

IIIIII

IIIIII

IIIIII

I

Discussion and ConclusionsMeasurement of functional status in our patients served2 important and distinct purposes: (1) documentation ofphysical therapy outcome in groups of patients for qualityassurance, establishment of clinical standards, orresearch purposes and (2) documentation of functionallevel used to set goals and measure functional progressand outcome for individual patients. The capacity of theLEFS to detect change in lower-extremity functionappears to be superior to that of the SF-36 physicalfunction subscale, as indicated by higher correlationswith an external prognostic rating of change. In light ofthis finding as well as the greater ease of administeringand scoring the LEFS, this scale appears to be a goodchoice for documenting lower-extremity function. Onlyprospective research, however, validates the use of thismeasure in clinical decision making. Because the LEFSmeasures physical function but not overall health, webelieve that a generic health status measure such as theSF-36 should be used to supplement the LEFS when thegoal is to measure the overall health status of ourpatients. The LEFS appears to overcome, to some extent,the barriers identified for implementation of a healthstatus measure in clinical practice.

The LEFS is easy to administer and score and is applica-ble to a wide range of disability levels and conditions andall lower-extremity sites. In our view, the LEFS is moreinterpretable with respect to understanding error asso-ciated measurement and for determining minimallyclinically important score changes and is a sufficientmeasure of reliability, validity, and sensitivity to change,at a level that is commensurate with utilization at anindividual patient level.

The LEFS can be used by clinicians as a measure ofpatients’ initial function, ongoing progress, and out-come as well as to set functional goals. For an outpatientorthopedic population, for example, initial and weeklyfollow-up administration may be considered appropri-ate. In order to set short- and long-term goals based ona self-report functional scale such as the LEFS, theclinician, in our view, should synthesize the patient’sclinical history and findings, as well the measurementproperties of the scale (ie, the error associated with a

single-scale measure, MDC, and MCID). The error asso-ciated with a given measure on the LEFS is about 65scale points (90% CI). Clinicians, therefore, can bereasonably confident that an observed score is within 5points of the patient’s “true” score. The MDC of theLEFS is 69 scale points (90% CI). Clinicians can also bereasonably confident that change on the LEFS of greaterthan 9 scale points is a true change. This informationcan be used to base short- and long-term goals forfunctional change that are at least greater than theMDC. The MCID of the LEFS is about 9 scale points.Clinicians can be reasonably confident that a change ofgreater than 9 scale points is not only a true change butis also a clinically meaningful functional change.Whether short- or long-term goals are set that are just ator greater than the MDC and MCID for the LEFS willdepend on the patient’s initial functional level, clinicalhistory and findings, and time frames for the goals.

An example of the application of the LEFS to establishfunctional level, set goals, and track progress and out-come. Consider a patient with an initial LEFS score of46/80. Based on the error at a given point in time for theLEFS of 5 points, the clinician can be 90% confident thatthe actual scale score is between 41 and 51. If thepatient’s condition is deemed to be relatively chronicand is expected to change slowly, the clinician mightselect a 2-week time frame for a change in score of just atthe MDC and MCID of 9 scale points. The short-termgoal, therefore, could be: “Increase LEFS score to lessthan or equal to 54/ 80.” In setting a short-term goal fora patient with a relatively acute condition who is pre-dicted to experience change quickly, a shorter timeframe of, for example, 1 week with a greater change thanthe MDC and MCID may be selected. In this case, thegoal may be: “Increase LEFS score to greater than orequal to 60/80.” On follow-up, for example, 1 week later,progress is could be determined by the amount ofchange on the scale. In cases where improvementgreater than the MDC and MCID occur, clinicians canbe reasonably confident that true (MDC) and important(MCID) change has occurred. In cases where there is nochange or change less than the MDC on follow-up,clinicians may be confident that true clinical change hasnot occurred. In this case, depending on the clinicalpicture and time frame since the previous assessment, achange in intervention, referral, or discharge of thepatient may be considered.

Ceiling and floor effects exist for a health status measurewhen patients often score at the extremes of normalfunction or severely restricted function. In the case of aceiling effect, there is restricted range for improvementbecause patients begin at the high level of function onthe scale. In the case of a floor effect, there is a restrictedrange for deterioration in functional status. For

Table 6.Measurement Properties of the Lower Extremity Functional Scale(LEFS): Application to Individual Patients

SEM

Error atSingle Pointin Time

MinimalDetectableChange

MinimalClinicallyImportantDifference

LEFS points 63.9 65.3 9 9

380 . Binkley et al Physical Therapy . Volume 79 . Number 4 . April 1999

example, the existence of both a ceiling effect and afloor effect has been reported for patients with lower-extremity dysfunction for the MFA4,18 and for the SF-36in our study. None of the patients in our study scored at0 or 80 on the scale at admission or at the 3-weekfollow-up assessment, indicating that there is no ceilingor floor effect associated with the LEFS in this type ofpatient population. The implication of ceiling and flooreffects is to lower the capacity to detect clinically impor-tant change in all patients.

In our study, we also used a rating of expected change asthe theory for change. Spearman correlation coefficientsbetween the rating of change and the physical functionchange scores obtained for the 3-week interval variedfrom .42 to .64. Our results, coupled with those ofWestaway et al,45 provide support for using a prognosticrating of change as a theory for evaluating a measure’ssensitivity to change. The results of the studies suggestthat a correlation coefficient of approximately .50 can beexpected. This information may be useful for estimatingthe sample size for subsequent studies, where a prognos-tic rating is used as a theory for change.

There are 2 major limitations to this study. The LEFS wasconceived as a measure applicable to a broad spectrumof lower-extremity problems. Our sample included only3 patients with hip and thigh conditions. In addition, allpatients in the study were outpatients. Further investiga-tion is needed to document the measurement propertiesof the LEFS in patients with hip conditions and in othersettings, including inpatient orthopedics.

We conceived the LEFS as a measure applicable forpeople with a broad spectrum of lower-extremity prob-lems. Accordingly, it was necessary that the first studycompare the LEFS with a measure of established validity.The SF-36 has been used to assess outcomes in peoplewith hip, knee, and ankle dysfunction.4,7,12,19–25 It was forthis reason that we chose the SF-36 as the comparisonmeasure. Deyo, when reflecting on the proliferation ofoutcome measures, stated, “while the development ofnew instruments would be encouraged where necessary,we may hope that investigators will not reinvent thewheel.”49(p1052) New measures are appropriate whentheir measurement properties or efficiency—in terms ofthe burden on both the respondent and those requiredto score the measure—are superior to existing mea-sures.50 It is for this reason that a one-sided researchquestion was posed: Is the LEFS superior to the SF-36?The results of our study, in our opinion, provide evi-dence supporting the superiority of the LEFS over theSF-36 for assessing lower-extremity function. Subsequentinquiry concerning the LEFS should center on head-to-head comparisons with condition-, disease-, or region-specific measures. Rather than asking whether the LEFS

is superior to existing measures, future research shouldinquire about the equivalence of the LEFS and thecompeting measures of interest.

Selection of self-report measures suitable for document-ing outcomes in clinical practice and in clinical trials andchoosing a condition-specific or generic health statusmeasure should be dependent, in part, on the goals ofmeasurement. Condition-specific measures, such as theLEFS, often do not include measures of psychosocialfunction and tend to be less influenced by comorbidstates.8,19 The LEFS, however, is superior to the SF-36 interms of clinical efficiency and sensitivity to change forthe documentation of physical function in patients withlower-extremity dysfunction. Generic measures, such asthe SF-36, are not generally practical for application atan individual patient level due to the length of the scaleand complexity of scoring. Because the conceptualframeworks for generic and disease-specific measures—such as the LEFS—differ, we believe that they can beviewed as being complementary rather than competingmeasures. Indeed, there is considerable agreement thata comprehensive assessment should include the admin-istration of both generic and disease-specific mea-sures.8,19 In clinical practice, the administration of botha generic measure, such as the SF-36, and a condition-specific measure, such as the LEFS, at admission anddischarge, with weekly re-evaluation of using the condition-specific measure, would achieve the benefits offered byboth types of measures.

AcknowledgmentsWe acknowledge the contributions to this project ofGregory Alcock, PT, and Aly Mawani, PT, who werestudents in the School of Rehabilitation Science atMcMaster University, Hamiliton, Ontario, Canada, at thetime of this study.

References1 Nelson EC, Berwick DM. The measurement of health status inclinical practice. Med Care. 1989;27(suppl):S77–S90.

2 Streiner DL; Norman GR. Health Measurement Scales: A Practical Guideto Their Development and Use. Oxford, England: Oxford University Press;1995.

3 McHorney CA, Tarlov AR. Individual-patient monitoring in clinicalpractice: Are available health status surveys adequate? Qual Life Res.1995;4:293–307.

4 Martin DP, Engelberg R, Agel J, Swiontkowski MF. Comparison ofthe Musculoskeletal Function Assessment Questionnaire with theShort Form-36, the Western Ontario and McMaster Universities Osteo-arthritis Index, and the Sickness Impact Profile health-status measures.J Bone Joint Surg Am. 1997;79:1323–1335.

5 Irrgang JJ, Snyder-Mackler L, Wainner RS, et al. Development of apatient-reported measure of function of the knee. J Bone Joint Surg Am.1998;80:1132–1145.

6 Kujala UM, Jaakkola LH, Koskinen SK, et al. Scoring of patellofemo-ral disorders. Arthroscopy. 1993;9:159–163.

Physical Therapy . Volume 79 . Number 4 . April 1999 Binkley et al . 381

IIIIII

IIIIII

IIIIII

IIIIII

I

7 Chesworth BM, Culham EG, Tata GE, Peat M. Validation of outcomemeasures in patients with patellofemoral syndrome. J Orthop Sports PhysTher. 1989;10:302–308.

8 Kantz ME, Harris WJ, Levitsky K, et al. Methods for assessingcondition-specific and generic functional status outcomes after totalknee replacement. Med Care. 1992;30:MS240–MS252.

9 Lysholm J, Gillquist J. Evaluation of knee ligament surgery resultswith special emphasis on use of a scoring scale. Am J Sports Med.1982;10:150–154.

10 Tegner Y, Lysholm J. Rating systems in the evaluation of kneeligament injuries. Clin Orthop. 1985;198:43–49.

11 Bellamy N, Buchanan WW, Goldsmith C, et al. Validation study ofWOMAC: a health status instrument for measuring clinically importantpatient relevant outcomes to antirheumatic drug therapy in patientswith osteoarthritis of the hip and knee. J Rheumatol. 1988;15:1833–1840.

12 Hawker G, Melfi CA, Paul J, et al. Comparison of a generic (SF-36)and a disease-specific (WOMAC) instrument in the measurement ofoutcomes after knee replacement surgery. J Rheumatol. 1995;22:1193–1196.

13 Ware JE Jr, Sherbourne CD. The MOS 36-Item Short-Form HealthSurvey (SF-36), I: conceptual framework and item selection. Med Care.1992;30:473–483.

14 Ware JE Jr, Kosinski M, Bayliss MS, et al. Comparison of methods forthe scoring and statistical analysis of SF-36 health profile and summarymeasures: summary of results from the Medical Outcomes Study. MedCare. 1995;33:AS264–AS279.

15 Ware JE Jr, Kosinski M, Keller SD. A 12-item short-form healthsurvey: construction of scales and preliminary tests of reliability andvalidity. Med Care. 1996;34:220–232.

16 Jette AM. The Functional Status Index: reliability and validity of aself-report functional disability measure. J Rheumatol. 1987;14:15–21.

17 Engelberg R, Martin DP, Agel J, et al. Musculoskeletal FunctionAssessment Instrument: criterion and construct validity. J Orthop Res.1996;14:182–192.

18 Martin DP, Engelberg R, Agel J, et al. Development of a musculoskel-etal extremity health status instrument: the Musculoskeletal FunctionInstrument. J Orthop Res. 1996;14:173–181.

19 Bombardier C, Melfi CA, Paul J, et al. Comparison of a generic anda disease-specific measure of pain and physical function after kneereplacement surgery. Med Care. 1995;4:AS131–AS144.

20 Jette DU, Jette AM. Physical therapy and health outcomes inpatients with knee impairments. Phys Ther. 1996;76:1178–1187.

21 Dawson J, Fitzpatrick R, Carr A, Murray D. Questionnaire on theperception of patients about total hip replacement. J Bone Joint Surg Br.1996;78:185–190.

22 Shapiro ET, Richmond JC, Rockett SE, et al. The use of a generic,patient-based health assessment (SF-36) for evaluation of patients withanterior cruciate ligament injuries. Am J Sports Med. 1996;24:196–200.

23 Katz J, Harris TM, Larson MG, et al. Predictors of functionaloutcomes after arthoscopic partial menisectomy. J Rheumatol. 1992;19:1938–1942.

24 Bennett KJ, Torrance GW, Moran LA, et al. Health state utilitiesin knee replacement surgery: the development and evaluation ofMcKnee. J Rheumatol. 1997;24:1796–1805.

25 Dawson J, Fitzpatrick R, Murray D, Carr A. Comparison of measuresto assess outcomes in total hip replacement surgery. Quality in HealthCare. 1996;5:581–588.

26 Weinberger M, Samsa GP, Tierney WM, et al. Generic versusdisease-specific health status measures: comparing the Sickness ImpactProfile and the Arthritis Impact Measurement scales. J Rheumatol.1992;19:543–546.

27 Wright JG, Young NL. A comparison of different indices of respon-siveness. J Clin Epidemiol. 1997;50:239–246.

28 Stratford PW, Binkley JM, Riddle DL. Health status measures:strategies and analytic methods for assessing change scores. Phys Ther.1996;76:1109–1123.

29 Deyo RA, Centor RM. Assessing the responsiveness of functionalscales to clinical change: an analogy to diagnostic test performance.J Chronic Dis. 1986;39:897–906.

30 Liang MH. Evaluating measurement responsiveness. J Rheumatol.1995;22:1191–1192.

31 Deyo RA, Diehr P, Patrick DL. Reproducibility and responsivenessof health status measures: statistics and strategies for evaluation. ControlClin Trials. 1991;12(suppl):142S–158S.

32 Beaton DE, Hogg-Johnson S, Bombardier C. Evaluating changes inhealth status: reliability and responsiveness of five generic health statusmeasures in workers with musculoskeletal disorders. J Clin Epidemiol.1997;50:79–93.

33 Kirshner B, Guyatt G. A methodological framework for assessinghealth indices. J Chronic Dis. 1985;38:27–36.

34 Hays RD, Hadorn D. Responsiveness to change: an aspect of validity,not a separate dimension. Qual Life Res. 1992;1:73–75.

35 Riddle DL, Stratford PW, Binkley JM. Sensitivity to change of theRoland-Morris Back Pain Questionnaire: part 2. Phys Ther. 1998;78:1197–1207.

36 Stratford PW, Binkley JM, Riddle DL, Guyatt GH. Sensitivity tochange of the Roland-Morris Back Pain Questionnaire: part 1. PhysTher. 1998;78:1186–1196.

37 Chatman AB, Hyams SP, Neel JM, et al. The Patient-Specific Func-tional Scale: measurement properties in patients with knee dysfunc-tion. Phys Ther. 1997;77:820–829.

38 Stratford PW, Gill C, Westaway M, Binkley JM. Assessing disabilityand change on individual patients: a report of a patient-specificmeasure. Physiotherapy Canada. 1995;47:258–263.

39 Beurskens AJHM, de Vet HCW, Koke AJA. Responsiveness offunctional status in low back pain: a comparison of different instru-ments. Pain. 1996;65:71–76.

40 Stratford PW, Binkley JM. Measurement properties of the RM-18:a modified version of the Roland-Morris disability scale. Spine. 1997;22:2416–2421.

41 Kopec JA, Esdaile JM, Abrahamowicz M, et al. The Quebec BackPain Disability Scale: measurement properties. Spine. 1995;20:341–352.

42 Norman GR, Stratford PW, Regehr G. Methodological problems inthe retrospective computation of responsiveness to change: the lessonof Cronbach. J Clin Epidemiol. 1997;50:869–879.

43 Constitution of the World Health Organization. Geneva, Switzerland:World Health Organization; 1948.

44 Jette DU, Jette AM. Health status assessment in the occupationalhealth setting. Orthop Clin North Am. 1998;27:891.

45 Westaway M, Stratford PW, Binkley JM. The Patient-Specific Func-tional Scale: validation of its use in persons with neck dysfunction.J Orthop Sports Phys Ther. 1998;27:331–338.

46 Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing raterreliability. Psychol Bull. 1979;86:420–428.

382 . Binkley et al Physical Therapy . Volume 79 . Number 4 . April 1999

47 Stratford PW, Binkley J, Solomon P, et al. Defining the minimumlevel of detectable change for the Roland-Morris Questionnaire. PhysTher. 1996;76:359–365.

48 Williams EJ. The comparison of regression variables. Journal of theRoyal Statistical Society, Series B. 1959;21:396–399.

49 Deyo RA. Measuring the functional status of patients with low backpain. Arch Phys Med Rehabil. 1988;69:1044–1053.

50 Armadio P, Beaton D, Bombardier C, et al. Measuring disability andsymptoms of the upper limb: a validation study of the DASH question-naire @Abstract#. Arthritis Rheum. 1996;39:S112.

Appendix.Lower Extremity Functional Scale

We are interested in knowing whether you are having any difficulty at all with the activities listed below because of your lower limb problem forwhich you are currently seeking attention. Please provide an answer for each activity.

Today, do you or would you have any difficulty at all with:

(Circle one number on each line)

Activities

ExtremeDifficulty orUnable toPerformActivity

Quite a BitofDifficulty

ModerateDifficulty

A LittleBit ofDifficulty

NoDifficulty

a. Any of your usual work, housework, or school activities. 0 1 2 3 4b. Your usual hobbies, recreational or sporting activities. 0 1 2 3 4c. Getting into or out of the bath. 0 1 2 3 4d. Walking between rooms. 0 1 2 3 4e. Putting on your shoes or socks. 0 1 2 3 4f. Squatting. 0 1 2 3 4g. Lifting an object, like a bag of groceries from the floor. 0 1 2 3 4h. Performing light activities around your home. 0 1 2 3 4i. Performing heavy activities around your home. 0 1 2 3 4j. Getting into or out of a car. 0 1 2 3 4k. Walking 2 blocks. 0 1 2 3 4l. Walking a mile. 0 1 2 3 4m. Going up or down 10 stairs (about 1 flight of stairs). 0 1 2 3 4n. Standing for 1 hour. 0 1 2 3 4o. Sitting for 1 hour. 0 1 2 3 4p. Running on even ground. 0 1 2 3 4q. Running on uneven ground. 0 1 2 3 4r. Making sharp turns while running fast. 0 1 2 3 4s. Hopping. 0 1 2 3 4t. Rolling over in bed. 0 1 2 3 4Column Totals:

SCORE: /80Error (single measure): 65 scale pointsMDC: 9 scale pointsMCID: 9 scale points

Physical Therapy . Volume 79 . Number 4 . April 1999 Binkley et al . 383

IIIIII

IIIIII

IIIIII

IIIIII

I

Lower extremity injuries Program of Care | february 2008

Measurement Tools and Instructions

Lower Extremity Functional Scale (LEFS)

The Lower extremity functional Scale will be used as a subjective outcome measure in this Program of Care. This scale was developed for a variety of lower ex-tremity conditions. It was based on the World Health organization’s model of impairment, disability and handicap at the time of its development in 1999 (bin-kley et al, 1999). The measure is comprised of 20 items asking about difficulty performing a variety of every-day activities. each item is scored by the subject as 0 (unable to perform) to 4 (no difficulty). LefS score vary from 0 (low) to 80 (normal function).

The scale has high test-retest reliability and appears to be moderately responsive to clinical change in pa-tients with anterior knee pain (Watson, 2005). The reli-ability has also been established in acute ankle sprains (alcock, Stratford, 2002). There has been no floor or ceiling effect reported and it appears to be applicable to all levels of function (finch et al, 2002). Construct validity was supported by comparison with the Sf-36 (binkley et al, 1999).

overall, the minimum clinically important difference (mCID) is nine points; “Clinicians can be reasonably confident that a change of greater than 9 points is… a clinically meaningful functional change” (binkley et al, 1999).

The scale takes 3-5 minutes to complete by the work-er and 30 seconds to score by the health practitioner (finch et al, 2002).

Numeric Pain Rating Scale (NPRS)

The Numeric Pain rating Scale will be used as an out-come measure within this Program of Care. This is a subjective pain measure that is widely used in clinical practice and research. Workers indicate their present pain level on a scale with numeric indicators from 0 (no pain) to 10 (worst possible pain).

It has been established as valid, reliable and appropri-ate for clinical practice (Williamson, 2005). The NPrS proved more reliable than another pain scale for pa-tients with pain of traumatic origin (berthier, 1998). from a practical standpoint, this scale has good sensi-tivity and generates data that can be statistically ana-lyzed for audit purposes (Williamson, 2005). However, there is little data on psychometric properties within the lower limb.

on average, a reduction of one point or a reduction of 15.0% in the NPrS represented a minimally Clinically Important Difference for the patient. a NPrS change score of -2.0 and a percent change score of -33.0% were best associated with the concept of “much bet-ter” improvement. for this reason these values can be considered as appropriate cut-off points for this mea-sure (Salaffi, 2004). These results support the use of a “much better” improvement on the pain relief as a clinically important outcome that exceeds the bounds of measurement error (Salaffi, 2004; Childs, 2005).

The NPrS should be recorded on the Initial assess-ment report and Care & outcomes Summary.

measurement tools and instructions 1

mEa

SuR

EmEN

t t

oo

L

Lower extremity injuries Program of Care | february 2008

activities Exte

me

diffi

cult

y or

una

ble

to

perfo

rm a

ctiv

ity

Qui

te a

bit

of

diffi

culty

mod

erat

e

diffi

culty

a li

ttle

bit

of

diffi

culty

No

diffi

cult

y

a. any of your usual work, housework, or school activities. 0 1 2 3 4

B. your usual hobbies, recreational or sporting activities. 0 1 2 3 4

C. getting into or out of the bath. 0 1 2 3 4

D. Walking between rooms. 0 1 2 3 4

E. Putting on your shoes or socks. 0 1 2 3 4

F. Squatting. 0 1 2 3 4

G. Lifting an object, like a bag of groceries from the floor. 0 1 2 3 4

H. Performing light activities around your home. 0 1 2 3 4

I. Performing heavy activities around your home. 0 1 2 3 4

J. getting into or out of a car. 0 1 2 3 4

K. Walking 2 blocks. 0 1 2 3 4

L. Walking a mile. 0 1 2 3 4

m. going up or down 10 stairs (about 1 flight of stairs). 0 1 2 3 4

N. Standing for 1 hour. 0 1 2 3 4

o. Sitting for 1 hour. 0 1 2 3 4

P. running on even ground. 0 1 2 3 4

Q. running on uneven ground. 0 1 2 3 4

R. making sharp turns while running fast. 0 1 2 3 4

S. Hopping. 0 1 2 3 4

t. rolling over in bed. 0 1 2 3 4

Column totals:

minimum level of detectable change (90% Confidence): 9 points SCoRE ______ / 80

Lower Extremity Functional Scale

We are interested in knowing whether you are having any difficulty at all with the activities listed below because of your lower limb problem for which you are currently seeking attention. Please provide an answer for each activity.

Today, do you or would you have any difficulty with:(Circle one number on each line)

Measurement Tools and Instructions

measurement tools and instructions

mEa

SuR

EmEN

t t

oo

L

Source: binkley et al (1999): The Lower extremity functional Scale (LefS): Scale development, measurement properties, and clinical application. Physical Therapy. 79: 371-383

2