Occupational therapy for rheumatoid arthritis: A systematic review

Authors


Patients with rheumatoid arthritis (RA) show a reduction in physical capacities compared with healthy persons. Symptoms such as pain, fatigue, stiffness, and decreased muscle strength cause difficulties with daily activities such as grooming and dressing, cooking a meal, cleaning, shopping, work, and leisure activities. The physical, personal, familial, social, and vocational consequences of RA are extensive. Occupational therapy (OT) is concerned with facilitating people in performing their activities of daily living overcoming barriers by maintaining or improving abilities, or compensating for decreased ability in the performance of occupations (1). The most important interventions in OT are training of skills, counseling, education about joint protection, prescription of assistive devices, and the provision of splints (2). Advice/instruction in the use of assistive devices, training in self-care activities, and training in productivity activities are the 3 interventions for RA patients chosen most often by occupational therapists (3).

So far, one narrative review (4) discussed the effectiveness of splinting, joint protection, and provision of aids/equipment for several rheumatic diseases on the basis of the results of only a few studies on OT. However, evidence on the effects of OT on the functional performance and social participation of RA patients has not been reviewed systematically. Therefore, we conducted a systematic review of published studies evaluating occupational therapy for rheumatoid arthritis.

Methods

Search strategy.

We conducted an extensive search in the following databases: Medline (1966 to January 2001), Cinahl (1982 to March 2001), Embase (1988 to April 2000), Scisearch (1974 to April 2000), Cochrane Controlled Trials Register, the databases of the libraries of medical and rehabilitation literature of 2 Dutch institutes (Dutch National Institute Allied Health Professions, Netherlands Institute for Health Services Research) (May 2000), the database of the Rehabilitation and Related Therapies Field of the Cochrane Collaboration (August 2000), and the specialized trials register of the Cochrane Musculoskeletal Group (August 2000). Our broad computerized search strategy was built upon the following components: a) the search strategy for randomized controlled trials (RCTs) and controlled clinical trials (CCTs) recommended by the Cochrane Musculoskeletal Group (5), b) a search strategy for other designs (ODs) using the keywords “epidemiologic studies,” “evaluation studies,” “program evaluation,” “questionnaires,” “patient series,” “case series,” “program,*” “experiment,*” “observation,*” “method,*” and “effect,*” c) a search strategy for the identification of studies involving RA patients using the terms “arthritis” and “rheumatoid arthritis,” and d) a search strategy for the identification of studies involving occupational therapy interventions using the terms “occupational therapy,” “training,” “education,” “splints,” “assistive devices,” “counseling,” “joint protection,” “dexterity,” “activities of daily living,” and “self-care.”

The search strategy was formulated in WinSpirs (Medline, Cinahl) and was adapted by an experienced medical librarian to make it applicable to the other databases. The same databases were searched to identify reviews about the effectiveness of OT, in order to find more studies. Additionally, the reference lists of all identified studies and reviews were scanned. Finally, the corresponding authors of reports eligible for inclusion in this review were contacted by mail and were asked to provide any additional published studies relevant to this systematic review.

Selection for inclusion.

Because occupational therapy is a relatively young profession, and because literature about the efficacy of OT was expected to be sparse, study designs other than controlled clinical trials were also included in this systematic review. As will be explained below, uncontrolled studies could only restrictedly contribute to the best evidence syntheses (indicative findings; see Data Analysis). Studies with the following designs were included: 1) RTCs. Experiments in which investigators randomly allocate eligible subjects into treatment and control groups. Crossover trials were considered as RCTs, according to the Cochrane Collaboration guidelines (6); 2) CCTs. Experiments in which eligible subjects are allocated in a nonrandomized manner to the treatment and the control groups; 3) ODs. Patient series, pre–post studies, as well as studies comparing interventions on different hands of the same person. Only full-length articles or full written reports were considered for inclusion in the review.

Studies of patients with a diagnosis of rheumatoid arthritis were included. Occupational therapy interventions were either regarded as comprehensive OTs (combination of all interventions) or were classified into 6 specific intervention categories as follows: 1) training of motor function, 2) training of skills, 3) instruction on joint protection, 4) counseling, 5) advice and instruction in using assistive devices, and 6) provision of splints. Studies focusing on a contrast between the experimental and control groups consisting of a multidisciplinary intervention (with OT as part of it) were excluded.

The outcome measures pain, fatigue, functional ability (including dexterity), and social participation were included. Occupational therapy also focuses on measures considered to be mediators of a favorable outcome, such as knowledge about disease management, compliance, self-efficacy, grip strength, and range of motion. These process measures are considered to be indicators of a successful treatment and are therefore also covered in this review. As will be explained below, process measures could only restrictedly contribute to the best evidence syntheses (indicative findings; see Data Analysis). Studies with one or more of the specified outcome or process measures were included.

Procedure for inclusion.

The procedure for inclusion of the studies was based on the recommendations described by Van Tulder et al (7). The first selection, based on titles and abstracts, was independently performed by 2 reviewers (EMJS and CHME), who considered the criteria for type of study, type of participants, and type of outcome measures. This first selection resulted in inclusion of the study, exclusion of the study, or was considered indecisive. The second step for inclusion was done independently by 2 reviewers (EMJS and CHME), using full reports of all included and indecisive studies and considering all the criteria described above. Disagreements regarding inclusion status were resolved by discussion. If no consensus was met, a third reviewer (MAHK) decided. Finally, a group of 4 occupational therapists and reviewer CHME assessed the criteria for type of intervention and, if appropriate, classified the type of intervention into comprehensive OT or one of the 6 different interventions. Consensus was reached by discussion.

Methodologic quality.

The methodologic quality of RCTs and CCTs was rated using a list recommended by Van Tulder et al (7). The list, containing specified criteria proposed by Jadad et al (8) and Verhagen et al (9), consists of 11 criteria for internal validity, 6 descriptive criteria, and 2 statistical criteria (Appendix 1). One modification was made in the specification of the eligibility criterion: the condition of interest (impairment or disability that indicated referral to OT) was added as an eligibility criterion, as proposed by Wells et al (10). All criteria were scored as yes, no, or unclear. Studies were considered to be of high quality if at least 6 criteria for internal validity, 3 descriptive criteria, and 1 statistical criterion were scored positively.

The methodologic quality of the other designs has been rated using an adapted version of the list described by Van Tulder et al (Appendix 1). Some items (concerning randomization, similarity of patient groups, blinding of care provider, blinding of patient) were considered inapplicable to ODs and were removed from the list. Some items were reformulated to make them applicable to one patient group (e.g., the item “Were co-interventions avoided or comparable?” was reformulated into “Were co-interventions avoided?”) or to make the item applicable to the design of the study (e.g., the item “Was the outcome assessor blinded to the intervention” was reformulated into “Was the care provider not involved in the outcome assessment?”). The final list of criteria used in ODs consists of 7 criteria for internal validity, 4 descriptive criteria, and 2 statistical criteria (Appendix 1). All criteria were scored as yes, no, or unclear. Studies were considered to be of sufficient quality if at least 4 of 7 criteria for internal validity, 2 descriptive criteria, and 1 statistical criterion were scored positively.

The methodologic quality of the included trials was independently assessed by 2 reviewers (EMJS, MAHK). Disagreements were resolved by discussion. If no consensus was met, a third reviewer (CHME) decided.

Data analysis.

A predefined data extraction form, with study characteristics, patient characteristics, and baseline, posttest, and followup data of outcome and process measures, was used. For continuous variables, the standardized mean difference (Hedges' g) was calculated, if possible (11). For dichotomous variables, odds ratios with corresponding 95% confidence intervals were computed.

Analysis of the results was performed separately for each intervention category. For crossover trials without a washout period between interventions, data were not further analyzed. If studies compared the effect of 3 or more intervention groups with each other, 2 reviewers (EMJS, CHME) decided by consensus which 2 groups had the largest contrast, making comparison between 2 groups possible. A contrast between the OT intervention group and a nontreated control group was preferred. If 2 interventions were compared, the predominant contrast was the OT provided.

In advance, we expected to find too much diversity among the studies with regard to patients (severity of the disease), interventions (duration, frequency, and setting), and outcome measures (diversity, presentation of results) to make quantitative analysis (meta-analysis) appropriate. Therefore, we formulated a best evidence synthesis by attributing various levels of evidence to the effectiveness of OT, taking into account the design of the studies, the methodologic quality, the type of outcome measures, and the statistical significance of findings. The best evidence synthesis (Table 1) was based upon the one proposed by Van Tulder et al (12) and was adapted for the purpose of this review.

Table 1. Best evidence synthesis*
  • *

    RTCs = randomized controlled trials; CCTs = controlled clinical trials; ODs = other designs.

  • If the proportion of studies that show evidence is <50% of the total number of studies within the same category of methodological quality and study design (RCTs, CCTs or ODs), we state no evidence.

Strong evidenceProvided by consistent, statistically significant findings in outcome measures in at least two high quality RCTs
Moderate evidenceProvided by consistent, statistically significant findings in outcome measures in at least one high quality RCT and at least one low quality RCT or high quality CCT
Limited evidenceProvided by statistically significant findings in outcome measures in at least one high quality RCT, or provided by consistent, statistically significant findings in outcome measures in at least two high quality CCTs (in the absence of high quality RCTs)
Indicative findingsProvided by statistically significant findings in outcome and/or process measures in at least one high quality CCT or one low quality RCT (in the absence of high quality RCTs), or provided by consistent, statistically significant findings in outcome and/or process measures in at least two ODs with sufficient quality (in absence of RCTs and CCTs)
No evidenceIn cases of results of eligible studies that do not meet the criteria for one of the above-stated levels of evidence, or in case of conflicting results among RCTs and CCTs, or in case of no eligible studies

Only results of studies contributing to the outcome of the best evidence synthesis (e.g., RCTs with a high methodologic quality, low quality RCTs with significant findings, high quality CCTs with significant findings, and high quality ODs with significant findings) are presented. Characteristics and comprehensive results of all included studies on outcome and process measures are on request available from the corresponding author.

Three sensitivity analyses were performed: 1) excluding low quality studies, 2) considering studies to be of high quality if 4 or more criteria of internal validity were met, and 3) excluding studies that did not use the American College of Rheumatology criteria for the diagnosis of RA (13).

Results

Selection of studies.

The search strategy resulted in a list of 2,137 citations. After selection by title and abstract, 149 full articles were obtained. Fifty-seven publications concerned the effectiveness of occupational therapy for RA. Of these, 42 articles, presenting 37 studies (15 RCTs, 6 CCTs, 16 ODs), fulfilled all inclusion criteria. Data from 4 studies were presented in more than one article (14–23). One publication (24) presented 2 studies. Fifteen studies (25–39) were excluded for the following reasons: because treatment contrast was a multidisciplinary intervention, because patients other than those with rheumatic diseases participated in the study, or because outcome measures were beyond the scope of our review (Appendix 2).

Methodologic quality.

The methodologic quality of 21 RCTs/CCTs and in 16 ODs was assessed (Table 2). Five RCTs (40–44) had high methodologic quality, and all CCTs had low quality. In particular, the following criteria were fulfilled in fewer than one-third of the RCTs/CCTs: adequate allocation concealment, blinded care provider, blinding of patients, information on co-interventions, blinded outcome assessor, intention to treat analysis, and long-term followup. Given the methodologic constraints of other designs, 9 ODs (45–53) had sufficient methodologic quality. The following criteria were fulfilled in one-third or fewer of the ODs: outcome assessor not involved in treatment and long-term followup.

Table 2. Characteristics and quality of randomized clinical trials, controlled clinical trials, and other designs*
Authors (ref.)Internal validityDescriptiveStatisticalMethodologic quality
Randomized clinical trials    
 Kraaimaat et al (14)b1, g, j, l, n,c, d, m1, m2o, qlow
 van Deusen and Harlowe (16)b1, l, nm1, m2olow
 Stern et al (21)b1, g, ja, c, d, k, m1o, qlow
 Neuberger et al pilot (24)b1, jd, m1olow
 Hammond et al (40)b1, g, i, j, n, pa, c, d, k, m1o, qhigh
 Helewa et al (41)b1, e, f, i, j, l, n, pa, c, d, m1o, qhigh
 Hoenig et al (42)b1, e, f, g, i, j, nc, d, k, m1ohigh
 Ter Schegget and Knipping (43)b1, g, j, l, n, pd, k, m1o, qhigh
 Tijhuis et al (44)b1, g, j, l, n, pa, c, d, k, m1o, qhigh
 Mowat et al (54)b1, f, i, j, l, n,a, c, d, m1, m2 low
 Brighton et al (56)b1, g, i, nd, m1oLow
 Wagoner and le Lieuvre (59)b1, g, j, l, nd, m1olow
 Anderson and Maas (62)b1, f, g, nc, d, m1o, qLow
 Callinan and Mathiowetz (63)b1, g, j, l, n,a, d, k, m1o,low
 Palchik et al (65)b1, la, k, m1olow
Controlled clinical trials    
 Furst et al (19)j, n, pa, c, d, m1, m2o, qlow
 Neuberger et al (24) follow-upjc, d, m1o, qlow
 Dellhag et al (57)f, j, nc, d, m1 low
 Ring et al (58)i, ld, k, m1o, qlow
 Hass et al (61)jd, m2olow
 Feinberg (64)h, j, l, na, c, d, m1o, qlow
Other designs    
 Barry et al (45)g, i, l, nm1, m2o, qsufficient
 Cartlidge et al (46)f, j, l, n, pa, d, m1osufficient
 Hammond (47)g, i, j, l, na, d, m1o, qsufficient
 Hammond and Lincoln (48)g, j, l, na, d, m1o, qsufficient
 McKnight and Schamburg (49)f, j, l, na, d, k, m1osufficient
 Nordenskiöld (50)f, g, j, l, n, pd, m1o, qsufficient
 Nordenskiöld (51)f, g, j, nd, m1osufficient
 Pagnotta et al (52)f, g, j, l, n, pa, d, k, m1o, qsufficient
 Rennie (53)g, j, l, n, pd, m1o, qsufficient
 McAlphine et al (55)j, nm1o, qlow
 Schaufler et al (60)j, n, pd, m1olow
 Agnew and Maas (66)l, pk, m2o, qlow
 Feinberg and Brandt (67)ja, d, k, m2olow
 Malcus et al (68)f, j, la, d, k, m1olow
 McKnight and Kwoh (69)f, j, na, d, k, m1o, qlow
 Spoorenberg et al (70)ja, d, ko, qlow

Outcome of interventions.

Comprehensive occupational therapy.

Four studies on the effectiveness of comprehensive OT were identified (Table 3): 3 RCTs (14, 41, 54) and one OD (55). One of the RCTs (41) had high methodologic quality. The results on outcome measures are shown in Table 4. In the high quality RCT, Helewa et al (41) reported a significant positive effect of comprehensive OT on functional ability. The process measure “knowledge” was assessed in one study (54), in which no difference in gain in knowledge between the intervention and the control groups was reported. Thus, on the basis of one RCT (41), there is limited evidence for the effectiveness of comprehensive OT on functional ability (Table 5).

Table 3. Characteristics of included RCTs/CCTs
Authors (ref.)No. of participantsMethodsInclusion criteria/settingInterventionFrequency and duration of interventionOutcome measure
All OT interventions combined      
 Kraaimaat et al (14)77RCTRA, class I, II, III, min age 20, duration of illness ≥1 yr. OutpatientsI: group OT treatment§ R: waiting list/no treatment2 hours for 10 weeksPain: IRGL Functional ability: IRGL Anxiety, depression: IRGL Knowledge: questionnaire
 Helewa et al (41)105RCTRA, limitation in physical function, clinical stable, stable drug therapy. In communityI: individual OT treatment§ R: no treatment6 weeksPain: VAS Functional ability: questionnaire, HAQ Depression: Beck scale
 Mowat et al (54)137RCTDefinite RA, treated at RA-unit for a minimum of 14 days. OutpatientsI: individual OT followup§ at home R1: follow-up by general practicioner R2: follow-up by routine hospital careevery 3 months or more if neededFunctional ability: ADL-list Participation: ADL-list social scale Knowledge: questionnaire
Training of motor function      
 van Deusen and Harlowe (16)46RCTRA, ambulatory, recommendation for home rest and exercise. OutpatientsI: group instruction on expressive dance and discussion§ R: traditional exercise and rest90 minutes for 8 weeksCompliance: exercise-rest rating scales Range of motion: goniometer
 Hoenig et al (42)57RCTRA, Class II and III In communityI1: ROM tendon gliding exercises I2: resistive theraputty 85 I3: I1 and I2 R: no treatmenttwice daily 10–20 minutes for 3 monthsPain: articular index Dexterity: 9 hole peg test Grip strength: aneroid manometer Range of motion: MCP-PIP
 Brighton et al (56)44RCTRA, sero-positive rheumatoid factor >1 yr, erosion in the MCP and/or PIP joints of the hand In communityI: exercises of the hand at home plus reinforcement by therapist R: no exerciseon daily basis for 4 yearsGrip strength: sphygmomanometer Range of motion: goniometer
 Dellhag et al (57)52CCTRA, <70 yrs, disease duration >6–<10 yrs Class I, II, decreased ROM and/or grip strength, OutpatientsI1: wax bath and hand exercise I2: wax bath only I3: exercise only R: no treatment3× week, for 4 weeksPain: VAS Dexterity: Sollermantest Grip strength: Grippit Range of motion: goniometer
 Ring et al (58)24CCTRA, MCP silicone rubber interposition arthroplasty for all fingers recommended OutpatientsI: the continuous passive motion machine R: 10 repetitions extension/flexionI: as tolerated for 6 weeks R: Every 2 hoursGrip strength: dynamometer Range of motion: goniometer
 Wagoner and leLieuvre (59)12RCTRA, with hand involvement OutpatientsI: 10 squeezes Hand Helper both hand visual display R: same as I: without visual displaytwo/three times a day for 6 weeksCompliance: percentage exercises Grip strength: sphygmomanometer Range of motion: goniometer
Authors (ref.)No. of participantsMethodsInclusion criteria/settingInterventionFrequency and duration of interventionOutcome measure
Instruction on joint protection      
 Furst et al (19)28CCTRA >1 year, not receiving energy conservation training, no mobility limitations, ≥18 yrs In- and outpatientsI: group/individual OT education program using specific didactic format,§ R: individual routine OT treatmentI: 1.5 hour for 6 weeks R: 50-min/2 hrs, 1 to 3 timesPain: VAS Fatigue: VAS Functional ability: HAQ Participation: PAIS Knowledge: questionnaire Grip strength: sphygmomanometer
 Neuberger et al (24) pilot45RCTRA, OutpatientsI: self instructional OT§ programme plus feedback R: no intervention25 minutes, 4 sessionsFunctional abilities: observation Knowledge: questionnaire
 Neuberger et al (24) follow-up98CCTRA, OutpatientsI1: self instructional OT§ programme I2: I1 + ROM exercise and JPP I3: I1+I2 + nurse/pt contact R: no intervention30–45 minutes, 4 sessionsPain: VAS Functional abilities: observation Participation: depression Knowledge: questionnaire
 Hammond et al (40)35RCTRA, class III, wrist and hand involvement OutpatientsI: group OT education§ based on health belief model and self-efficacy theory R: no intervention2 hours, 4 sessionsPain: VAS, HAQ Functional abilities: JPBA, HAQ Knowledge: questionnaire Self-efficacy: Arthritis self-efficacy scale Grip strength: dynamometer Range of motion: HJAM
Advice/assistive devices      
 Hass et al (61)190CCTRA, In communityI: group OT sessions, improved user information and altered selection process§ R: routine for prescription devices4 sessionsPain: FSI (pain) Functional abilities: FSI Participation: SIP
Provision of splints      
 Stern et al (21)42RCT, cross over trialRA, class II, III, wrist involvement dominant hand OutpatientsI1: Alimed working splint, wash-out period of 1 week I2: Rolyan working splint I3: Futuro working splint4 hours, 5–7 daysPain: interview Dexterity: Jebsen Taylor test Grip strength: dynamometer
 Tijhuis et al (44)10RCT, cross over trialRA, swollen/painful wrist of the dominant hand, OutpatientsI1: Thermolynn orthosis I2: Futuro orthosisas much as possible, 2 weeksPain: VAS Grip strength: vigorimeter Range of motion: goniometer
 Ter Schegget and Knipping (43)18RCT, cross over trialSwanneck deformities OutpatientsI1: SRS othosis, I2: custom made orthosiseach day, 6 monthsPain: VAS Dexterity: questionnaire Grip strength: My gripper Range of motion: goniometer
Authors (ref.)No. of participantsMethodsInclusion criteria/settingInterventionFrequency and duration of interventionOutcome measure
  • * RCT = randomized clinical trial; CCT = controlled clinical trial; OD = other design; OT = Occupational therapy; RA = rheumatoid arthritis; I = Index group; R = reference group; ROM = range of motion; MCP = metacarpal phalangeal; PIP = proximal interphalangeal; ADL = activities of daily living; SRS = silver ring splint; VAS = visual analogue scale; HAQ = Health Assessment Questionnaire; IRGL = Impact of Rheumatic Diseases on Health and Lifestyle; PAIS = Psychosocial Adjustment to Illness scale; JPBA = Joint Protection Behavior Assessment; HJAM = Hand Joint Alignment and Motion scale; FSI = Functional Status Index; SIP = Sickness Impact Profile; AIMS2 = Arthritis Impact Measurement Scales 2.

  • Criteria for definite or classical rheumatoid arthritis according to the American Rheumatism Association (ARA) (13).

  • ARA functional classification.

  • §

    Occupational therapy treatment—biomedical information, energy conservation, joint protection, use of devices and splints, exercise for hands.

 Anderson and Maas (62)92RCTRA, OutpatientsI1: dorsal working splint I2: palmar splint I3: gauntlet splint I4: fabric ready made splint R: no splint1 sessionGrip strength: sphygmomanometer
 Callinan and Mathiowetz (63)45RCT cross over trialRA, presence of hand pain/morning stiffness, OutpatientsI1: soft resting splint, I2: hard resting splint R: no splint28 nightsPain: AIMS2 Functional abilities: AIMS2, compliance: diary Grip strength: dynamometer
 Feinberg (64)46CCTRA, class I, II, OutpatientsI: resting splint extensive compliance§ R: resting splint1 sessionPain: 5 pt. scale Compliance: dairy
 Palchik et al (65)7RCTBoutonierre deformity RA origin, passive correctable, loss of PIP extension <25°. OutpatientsI: gutter splint R: no intervention24 hours for 6 weeksRange of Motion: goniometer
Table 4. Results on pain and functional abilities per intervention category; studies contributing to the outcome of the best evidence synthesis*
First author (ref.)DesignMethodologic qualityPainFunctional ability
SMD (95% CI)SMD (95% CI)
  • *

    SMD = standardized mean difference; 95% CI = 95% confidence interval; OT = occupational therapy; RCT = randomized clinical trial; CCT = controlled clinical trial; OD = other design; ns = no significant differences between groups; ne = standardized mean difference (SMD) not estimable.

  • 1 = high; 0 = low.

  • Calculation of standardized mean difference (Hedges' g) based on, P F, or t-value, a positive SMD (a SMD with a 0.95 Ci indicates a decrease in pain and an increase in functional ability; if more than two groups were evaluated, the two groups with the greatest contrast in intervention were compared. For further information about the methodologic quality see Table 1.

Comprehensive OT    
 Kraaimaat (14)RCT0−0.17 (−0.76, 0.41)−0.26 (−0.84, 0.33)
 Helewa (41)RCT1ns0.49 (0.10, 0.89)
 Mowat (54)RCT0not measuredne
Training of motor function    
 Hoenig (42)RCT1nsns
 Dellhag (57)CCT00.81 (0.10, 1.61)ns
Instruction on joint protection    
 Furst (19)CCT01.2 (0.20, 7.18)0.17 (−0.68, 1.02)
 Neuberger (24) pilotRCT0not measured0.65 (−0.03, 1.33)
 Neuberger (24) follow upCCT00.31 (−0.49, 1.10)1.45 (0.55, 2.34)
 Hammond (40)RCT10.56 (−0.12, 1.24)1.79 (0.98, 2.61)
Advice assistive devices    
 Nordenskiöld (51)OD1P < 0.001not measured
 Hass (61)CCT0nene
Provision of splints    
 Stern (21)RCT0neimmediately, P < 0.001, 1 week, P = 0.015
 Ter Schegget (43)RCT10.34 (−0.59, 1.27)1.02 (−0.03, 2.01)
 Tijhuis (44)RCT1−0.44 (−1.33, 0.44)not measured
 Nordenskiöld (50)OD1immediately, P < 0.001not measured
 Pagnotta (52)OD1immediately, P < 0.001P < 0.01
 Callinan (63)RCT0P < 0.001not measured
 Feinberg (64)CCT0P = 1.00not measured
Table 5. Outcome of best evidence synthesis and sensitivity analyses per intervention.
 Best evidence synthesisSensitivity analysis 1Sensitivity analysis 2Sensitivity analysis 3
Comprehensive occupational therapyLimited evidence on functional abilityLimited evidence on functional abilityIndicative findings for effectiveness on functional abilityLimited evidence on functional ability
Training of motor functionNo evidenceNo evidenceNo evidenceNo evidence
Instruction on joint protectionLimited evidence on functional abilityLimited evidence on functional abilityLimited evidence on functional abilityLimited evidence on functional ability
Advice/instruction in the use of assistive devicesNo evidenceNo evidenceNo evidenceNo evidence
Provision of splintsIndicative findings for effectiveness on painIndicative findings for effectiveness only for the immediately assessed pain after provision of the splintIndicative findings for effectiveness on painIndicative findings for effectiveness on pain

Training of motor functions.

Seven studies (6 RCTs/CCTs [16, 42, 56–59] and one OD [60]) focused on the intervention “training of motor function” (Table 3). Interventions varied from group instruction on expressive dance (16), use of a continuous passive motion machine after arthroplasty (58), to hand exercises (42, 56, 57, 59, 60). The interventions on hand exercises varied widely with regard to type of exercises, type of device used, type of setting for therapy (at home without supervision or at an OT department with supervision), and duration of the intervention. One study (42) had high methodologic quality.

The results on outcome measures are shown in Table 4. The outcome measures “pain” and “functional ability” were assessed in two (42, 57) and three (42, 57, 60) studies, respectively. In the RCT with high methodologic quality (42), no significant differences between groups with regard to pain and functional ability after training of hand function were reported.

All studies measured 1 or 2 of the following process measures: compliance (16, 59), grip strength (42, 56–58, 60), and range of motion (16, 42, 56–58, 60). In the high quality RCT (42), no significant differences in grip strength between groups were found, whereas in the low quality RCT (56), significant changes in grip strength after training of hand function were reported. Thus, there is no evidence for the effectiveness of training of motor function on the outcome and process measures (Table 5).

Instruction on joint protection and energy conservation.

Four RCTs/CCTs (19, 24, 40) and four ODs (45–48) (Table 3) were identified for the intervention “instruction on joint protection and energy conservation.” One of the RCTs (40) had high methodological quality.

The results on outcome measures are shown in Table 4. Seven studies (19, 24, 40, 45, 47, 48) assessed functional ability. Hammond et al (40) reported significant improvement in functional ability. This finding was supported by Neuberger et al (24), who reported significant improvement in their CCT. Four studies (19, 24, 40, 48) measured pain. Hammond et al (40) reported no significant differences between groups.

All but one of the 8 studies (47) measured one or more process measures. Of the 7 studies (19, 24, 40, 45, 46, 48) that assessed knowledge, 2 RCTs/CCTs (24, 40) demonstrated a significant increase in knowledge after patients received instruction on joint protection. All ODs with sufficient methodologic quality (45, 46, 48) supported these findings. Thus, on the basis of the results of one high quality RCT (40), there is limited evidence that instruction on joint protection leads to an improvement of functional ability (Table 5).

Assistive devices.

One CCT (61) and one OD (51) were included for the intervention “advice/instruction in the use of assistive devices” (Table 3). Only the latter study had sufficient methodologic quality.

The results on outcome measures are shown in Table 4. Nordenskiöld (51) reported a significant decrease in pain when assistive devices were used while performing kitchen tasks. Thus, there is insufficient data to determine the effectiveness of advice/instruction in the use of assistive devices (Table 5).

Splints.

Sixteen studies (Table 3) focused on the intervention “provision of splints.” Seven of these studies were RCTs/CCTs (21, 43, 44, 62–65), and 9 were ODs (49, 50, 52, 53, 66–70). Within these 16 studies, six different types of splints were evaluated (working splint, resting splint, three types of antideformity splints, air-pressure splint). In 4 of the RCTs/CCTs (21, 43, 44, 64), 2 splints were compared with each other. Three of the RCTs/CCTs (62, 63, 65) compared patients receiving splint treatment with a nontreated control group. Two of the RCTs (43, 44) had high methodologic quality and 4 of the ODs (49, 50, 52, 53) had sufficient methodologic quality.

The results on outcome measures are shown in Table 4. Pain was assessed with regard to two aspects. The effect on pain immediately after provision of the splint was evaluated in three studies (50, 52, 53). Nordenskiöld (50) and Pagnotta et al (52) reported a significant decrease in pain when patients wore working splints. The effect on pain after splinting for a period of 1 week to 1.5 year was assessed in ten studies (21, 43, 44, 49, 63, 64, 67–70). Only the 2 studies that compared splinting with no treatment (49, 63) presented positive significant results.

Five studies (21, 43, 52, 53, 70) assessed measures of functional ability (dexterity). In one of these studies, a low quality RCT (21), a significant decline in dexterity after 1 week of wearing a working splint was reported.

Fifteen studies measured one or more process measures. Compliance with splinting was assessed by 5 studies (63, 64, 66, 67, 70), all of which had a low methodological quality. In one RCT (64), positive significant results on compliance were reported.

Grip strength was assessed with regard to 2 aspects. The effect on grip strength immediately after provision of the splint was evaluated in 6 studies (21, 43, 44, 50, 53, 62). In two high quality studies (50, 53) it was reported that patients had an increase in grip strength while wearing a splint. The effect of splinting on grip strength after a period of time was measured in four RCTs/CCTs (21, 43, 44, 63). The 2 high quality RCTs (43, 44) found no significant differences between groups. Four studies (43, 44, 65, 67) measured range of motion, and in the 2 high quality RCTs (43, 44), no significant differences between groups was reported. One low quality RCT (65) demonstrated significant improvement after patients wore an anti-boutonniere splint for 6 weeks.

Thus, there are indicative findings that splints are effective in reducing pain both immediately after provision of the splint and after splinting over a period of time (Table 5). Also, there are indicative findings that splinting has a negative effect on dexterity (Table 5). Furthermore, indicative findings for a gain in grip strength immediately after provision of the splint have been reported.

Training of skills, counseling.

No studies concerning the interventions “training of skills” and “counseling” were identified.

Sensitivity analyses.

Three sensitivity analyses were performed to determine the robustness of the outcome of the best evidence syntheses (Table 5).

Considering only studies that scored a high or sufficient methodologic quality, the outcome of the best evidence syntheses for all interventions (except provision of splints) is the same as the results presented. Within the category “provision of splints,” only the indicative findings for evidence of splinting on the immediate decrease in pain will hold.

Analyzing the results with incorporation of studies with a score of 4 items or more on the internal validity criteria, the outcome of the best evidence synthesis is, for all interventions except “comprehensive OT,” the same as the results presented. Within the category “comprehensive OT,” the results of three studies (14, 41, 54) instead of one contribute to the best evidence synthesis. Two studies (14, 54) found no significant results on functional abilities, whereas one (41) did. As a result, the findings of “limited evidence” changes to “indicative findings” for the evidence of effectiveness of OT on functional ability.

In 19 studies (14, 19, 21, 40–42, 47–50, 52, 55, 63, 64, 67–70), investigators explicitly reported use of the ACR criteria for diagnosis of RA as inclusion criteria for the patients. Considering only those studies in the analysis, results are the same as the best evidence synthesis for all the interventions categories.

Discussion

In this review, the efficacy of several OT interventions for rheumatoid arthritis was explored. Seven different intervention categories were distinguished (comprehensive OT, training of motor function, training of skills, instruction on joint protection, counseling, advice/instruction in the use of assistive devices, and provision of splints). The outcome measures were pain, fatigue, functional ability, and social participation. Process measures such as knowledge about disease management, compliance, self-efficacy, grip strength, and range of motion were also taken into account. This systematic review established limited evidence for the effectiveness of 2 intervention categories on functional ability. Both comprehensive OT and instruction on joint protection resulted in an increase in functional ability. For the intervention “provision of splints,” indicative findings for a decrease in pain were demonstrated. Indicative findings for a negative effect of splinting on dexterity were discovered, as were indicative findings for evidence that grip strength increases after provision of splints.

Randomized controlled trials, controlled clinical trials, and studies with other designs were included in this review. Sixteen ODs were identified. A distinction was made between ODs with sufficient methodological quality and those that lacked sufficient methodological quality. Because of the weakness of the internal validity of ODs, those with sufficient methodologic quality could demonstrate “indicative findings” only in the best evidence synthesis. Incorporation of the outcomes of ODs resulted in indicative findings for a decrease in pain immediately after provision of a splint. Within the other intervention categories, results of ODs did not contribute to the outcome of the best evidence synthesis, because RCTs and/or CCTs were available. However, in most categories of interventions, the results of ODs supported the findings of RCTs/CCTs. Therefore, in emerging fields of research such as occupational therapy, results of studies other than controlled trials may have some value in assessing the effectiveness of interventions when RCTs and CCTs are lacking.

Overall, the methodologic quality of the studies was rather poor. Only 5 of the 15 RCTs had a high methodologic quality. No CCTs with a high methodologic quality were identified, and only one-half of the 16 ODs were considered of sufficient methodologic quality. Bias was possible, because most studies did not include information on blinding of patients, blinding of care providers, and blinding of outcome assessors. Because blinding of patients and care providers is rather difficult in allied health interventions, the blinding of the outcome assessor is of paramount importance to avert detection bias (71, 72).

The nature of the OT interventions varied widely. Even within intervention categories, large differences in interventions with regard to type of treatment, duration, and setting precluded comparing results. Furthermore, poor data presentation impeded comparison of results among studies. Only five RCTs presented sufficient data to compute effect sizes. In future research, special attention should be given to the presentation of study results according to international standards (73). Finally, outcome measures were very heterogeneous. For each outcome and process measure, several measurement instruments were used. To overcome this problem, international consensus about a core set of outcome measures for the outcome of occupational therapy for rheumatoid arthritis is needed. The first question to be addressed should be which outcomes are most important for OT. The second question concerns which outcome instruments are most reliable, valid, responsive, and easy to obtain.

The power of the studies included in this review was rather poor. To detect a medium effect size of 0.5 (with α = 0.05 and power at 80%), the sample size per group needs to be at least 50 (74). Only 2 controlled studies had a sample size of ≥50 participants per group (41, 61). The findings of this review could be an underestimation of the real evidence for the effectiveness of OT, due to the limited power of the studies. Conversely, the results of this review could also be an overestimation because of publication bias by unpublished small negative studies.

In future research, several items about the efficacy of occupational therapy should be considered. To improve the methodologic quality of studies, proper randomization procedures should be performed after baseline assessment, with special attention to the concealment of allocation. Another important issue is the blinding of the outcome assessor. Because blinding of patients and care providers is almost impossible in OT interventions, procedures to guarantee the blinding of the outcome assessors are needed to prevent bias. Statistically significant differences are more likely to occur in studies with sufficient power. This means that large groups of rather homogenous participants should be included in trials that compare the experimental intervention with no treatment or, if that is not possible, with a treatment with a clear contrast. Furthermore, outcome measures should be carefully chosen with regard to the aim of the intervention. Studies in which outcome measures that are relevant and responsive are applied are more likely to result in statistically significant differences between groups.

The inventory of studies in this review reveals important gaps in OT research. No studies were found for the category “training of skills,” and only two studies were found for the intervention “instruction/advice assistive devices.” This is remarkable, because “training of skills” and “instruction/advice assistive devices” are very common OT interventions (3). Another finding is the lack of data on the outcome measure “social participation.” The ultimate goal of OT is to restore/maintain full participation in all social activities. Outcome measures should reflect this aim.

In conclusion, we found limited evidence for the effectiveness of occupational therapy for functional ability and pain in patients with RA. Studies that evaluated comprehensive OT and those that evaluated instruction on joint protection interventions showed limited evidence for the effectiveness of these interventions on functional ability. Studies that evaluated splint interventions reported indicative findings for the effectiveness in reducing pain. These results are encouraging in terms of occupational therapy being an important part of treatment for patients with rheumatoid arthritis. Also, this review revealed that important fields of occupational therapy, such as “training of skills” and “advice in the use of assistive devices,” are underresearched and should get more attention. On the basis of this review, we suggest that further clinical trials for each category of intervention are necessary. In future studies, special attention should be given to the design of trials, the use of responsive, reliable, and valid outcome measures, inclusion of a sufficient number of patients to create statistical power, and presentation of trial results according to international standards.

Appendix

Table Appendix 1. Criteria of Methodological Quality*
  • *

    Internal validity = b, e, f, g, h, i, j, l, n, p; descriptive criteria = a, c, d, k, m; statistical criteria = o, q.

Randomized Clinical Trials, Controlled Clinical Trials
Patient selection
 a) Were the eligibility criteria specified?
 b) Treatment allocation:
  1) Was a method of randomization performed?
  2) Was the treatment allocation concealed?
 c) Were the groups similar at baseline regarding the most important prognostic indicators?
Interventions
 d) Were the index and control interventions explicitly described?
 e) Was the care provider blinded for the intervention?
 f) Were co-interventions avoided or comparable?
 g) Was the compliance acceptable in all groups?
 h) Was the patient blinded to the intervention?
Outcome measurement
 i) Was the outcome assessor blinded to the interventions?
 j) Were the outcome measures relevant?
 k) Were adverse effects described?
 l) Was the withdrawal/drop out rate described and acceptable?
 m) Timing followup measurements:
  1) Was a short-term followup measurement performed?
  2) Was a long-term followup measurement performed?
 n) Was the timing of the outcome assessment in both groups comparable?
Statistics
 o) Was the sample size for each group described?
 p) Did the analysis include an intent-to-treat analysis?
 q) Were point estimates and measures of variability presented for the primary outcome measures?
Other design
Patient selection
 a) Were the eligibility criteria specified?
Interventions
 d) Was the intervention explicitly described?
 f) Were cointerventions avoided?
 g) Was the compliance acceptable?
Outcome measurement
 i) Was the outcome assessor not involved in the treatment?
 j) Were the outcome measures relevant?
 k) Were adverse effects described?
 l) Was the withdrawal/drop out rate described and acceptable?
 m) Timing follow-up measurements:
  1) Was a short-term followup measurement performed?
  2) Was a long-term followup measurement performed?
 n) Was the timing of the outcome assessment in all patients comparable?
Statistics
 o) Was the sample size of the patient group described?
 p) Did the analysis include an intent-to-treat analysis?
 q) Were point estimates and measures of variability presented for the primary outcome measures?

Appendix

Table Supplement to Appendix 1. Specification of the criteria for methodological quality
a.In order to score a ‘yes’ explicit classification criteria for RA should be described. An established set of criteria (ARA classification) or clinical criteria including disease duration. The condition of interest (for example: pain when treatment is a resting splint) is described as present at the start of the study.
b1A random (unpredictable) assignment sequence. Methods of allocation using date of birth, date of admission, hospital numbers, or alternation should not be regarded as appropriate.
b2Assignment generated by an independent person not responsible for determining eligibility of the patients. This person has no information about the persons included in the trial and has no influence on the assignment sequence or the decision about eligibility of the patient.
cIn order to receive a ‘yes’ groups have to be similar regarding four of the most important prognostic factors: age, duration of disease, severeness of disease, baseline main outcome measure(s). If a baseline difference exists in one of the these factors, a no applies.
dAdequate description of type, modality, application technique, intensity, duration, number of frequency of sessions for both the experimental interventions and (only for B4) control intervention(s) in order to replicate the study.
eThe reviewer determines when enough information about the blinding is given in order to score a ‘yes’. For Occupational therapy this item scores always ‘no’.
fCo-interventions concerning other similar to occupational therapy interventions are avoided or either standardised.
gThe reviewer determines when the compliance to the interventions is acceptable when based on the reported intensity, duration, number and frequency of sessions for the experimental intervention and (only B7) the control intervention(s). Criterion compliance >70% in all groups.
hThe reviewer determines (per outcome parameter) when enough information about blinding is given to score a ‘yes’. For occupational therapy this item always scores a ‘no’.
iThe reviewer determines (per outcome parameter) when enough information about independency/blinding is given to score a ‘yes’.
jAt least one of the important outcome parameters. For this review; pain, fatigue, functional abilities (including dexterity), physical independence and quality of life (including well being).
kEach event described and correctly attributed to (allocated) treatment; if explicit report of ‘no adverse effect’ a ‘yes’ applies. Scores either a ‘yes’ or a ‘no’, a don't know doesn't exist.
lParticipants who were included in the study but did not complete the observation period or were not included in the analysis must be described. If the percentage of withdrawals and drop-outs does not exceed 20% for short-term follow-up and 30% for long-term follow-up and does not lead to substantial bias a ‘yes’ is scored. No drop-outs reported scores as don't know.
m1Outcome assessment at the end of the intervention period.
m2Outcome assessment ≥6 months after pre-test.
nTiming of outcome assessment identical for all patients (A10) or identical for all intervention groups (B14); for all important outcome assessments.
oTo be presented per group at pre-test and for most important outcome assessments.
pAll patients are reported/analysed for the most important moments of effect measurement (minus missing values) irrespective of non-compliance and co-interventions.
qBoth point estimates and measures of variability should be presented (to be scored for each important outcome parameter separately). Point estimates are: means, medians, modes etc. Measures of variability are; standard deviations, 95% confidence intervals, etc. For dichotomous or categorical data proportions have to be presented.

Appendix

Table Appendix 2. Characteristics of excluded studies*
Author (ref.)Design and reason for exclusion
  • *

    RCT = randomized clinical trial; CCT = controlled clinical trial; RA = rheumatoid arthritis.

Alderson, et al (25)Pre–post test, multidiscipline intervention
Brattström, et al (26)Cohort study, participants with RA and other diseases, multi-discipline intervention
Chen, et al (27)Patient series, participants with RA and other diseases
Cytowicz, et al (28)Pre–post test, outcome measures not included in review
Gault, et al (29)Pre–post test, outcome measures grip strength and range of motion only measured as adverse effects of immobilisation intervention
Karten, et al (30)Patient series, multi-discipline intervention
Kjeken, et al (31)RCT, participants with RA and other diseases
Löfkvist, et al (32)Patient series, outcome measures not included in review
Maggs, et al (33)RCT, participants with RA and other diseases
Mann, et al (34)Cohort study, participants with RA and other diseases
Nicholas, et al (35)Patient series, outcome measures not included in review
Schulte, et al (36)Patient series, participants with RA and other diseases, multi-discipline intervention
Stern, et al (37)CCT, participants in study are women with no physical disability
Stern, et al (38)CCT, participants in study are women with no physical disability
Stewart, et al (39)Cohort study, outcome measures not included in review

Ancillary