Does occupational therapy reduce the need for surgery in carpometacarpal osteoarthritis? Protocol for a randomized controlled trial

CMC1-OA presents with a combination of reduced cartilage thickness, increased ligament laxity with resultant instability, and subluxation of the base of the metacarp on the trapezium, which in turn results in adduction contracture and decreased thumb web space [21]. As the severity of symptoms does not necessarily correspond with the radiographic stage of the disease, the main indications for surgery are pain and loss of function [22]. Surgical procedures most frequently comprise trapeziectomy with or without ligament reconstruction and tendon interposition (LRTI), and in the most severe cases trapeziometacarpal arthrodesis. Two systematic reviews comparing the effect of different surgical techniques in CMC1-OA conclude that there is currently no evidence that one surgical procedure is superior to another [23, 24]. The results further show that 10% of the participants who underwent trapeziectomy and 22% who underwent trapeziectomy with LRTI had adverse effects, including scar tenderness, tendon adhesion or rupture, sensory change, or complex regional pain syndrome [23]. Complications and repeat surgeries were also frequently reported following CMC1 arthrodesis, with non-union rates between 8 and 21% [24]. A recent randomized controlled trial (RCT), comparing arthrodesis with trapeziectomy with LRTI, was prematurely terminated due to the high prevalence of complications: 71% in the arthrodesis group and 21% in the trapeziectomy with LRTI group [25]. Thus, even if CMC1 surgery may reduce pain and improve function, the risk of adverse effects is high, and RCTs comparing surgery with non-surgical interventions are warranted. In a recent review of surgery for thumb osteoarthritis, the authors did not identify any studies that compared surgery to sham surgery or to non-operative treatments [26]. However, in a small prospective study evaluating the effect of occupational therapy (assistive devices alone or with addition of one of two types of orthoses) for patients awaiting CMC1-surgery, 23 out of 33 participants (70%) no longer required an operation after 7 months, while only two of the remaining 19 participants wanted an operation after 7 years [27]. The results thereby indicate that occupational therapy may reduce the need for surgery. However, as this was a small study with a high risk of bias, robust RTCs comparing the effects of occupational therapy and surgery are needed.

Occupational therapy aims at enhancing activity, participation and health-related quality of life. Interventions in occupational therapy for HOA most frequently comprises patient education, hand exercises, and the provision of assistive devices and thumb orthoses [28].

Based on one RCT [29], the authors of a review of treatment for HOA conclude that educating patients about self-management is effective for alleviating pain and avoiding disability [14]. In two later studies, the authors of a Dutch RCT found no effectiveness of a group-based multidisciplinary treatment programme [30], while there was a significant improvement in pain self-efficacy and a higher proportion of people meeting the OARSI responder criteria [31] among those receiving patient education (joint protection) compared to those who did not in a recent British RCT [32]. Further, participants in qualitative studies report that they experience lack of help, and little and/or contradictory information with regard to management of their condition [18, 19]. Thus, there is a need for developing education material that is unique to HOA and based on high-quality published evidence for HOA treatments.

Results from an RCT demonstrated that use of assistive devices significantly improved activity performance and satisfaction with performance in patients with HOA [33]. In addition, the use of assistive devices was the most frequently reported self-management strategy in a study exploring self-management strategies in persons with HOA [34]. Furthermore, participants with HOA have reported in qualitative interviews that assistive devices help them remain independent in daily activities [18, 19]. However, despite the availability of such devices, the participants had seldom received any information or guidance from health professionals during the process of selecting appropriate devices.

Several systematic reviews conclude that there is some, although limited, evidence that hand exercises may reduce pain and increase joint mobility and grip strength in HOA [12‐14, 35, 36]. Furthermore, there were no differences among those who received the multidisciplinary programme that included hand exercises compared to those who did not receive this programme in a recent Dutch RCT [30], or between the groups that received hand exercises versus those who did not in a British RCT [32]. However, none of these interventions included monitoring during the study period, and the rationale for the design of the exercise programmes is unclear. Moreover, when comparing the cost-utility of joint protection only, hand exercises only, and joint protection plus hand exercises with leaflet and advice (control group) in the British study, the findings showed that hand exercises was the most cost-effective option [37].

In a recent Norwegian RCT, hand exercises led to significant and clinically important functional improvements in terms of decreased pain and fatigue, increased thumb web space, approximately 25% improvement in grip strength, and improved activity performance [38]. The selection and monitoring of exercises in the study was in accordance with recent research and recommendations for exercises for HOA in general [1, 13, 39], as well as for CMC1-OA [21, 40‐42]. As the intervention was well tolerated and delivered at a low cost with respect to time expended and exercise materials used, the authors conclude that the programme should be considered for inclusion in the standard care for people with HOA [38]. In another Norwegian study investigating the effect of the same exercise programme, but with three additional exercises for the shoulder girdle and upper arm, these findings were supported, although the effects were more limited [43]. It therefore seems that the less comprehensive programme may be easier to perform on a regular basis, thus making it probably the most effective.

Based on the best available evidence, the European League Against Rheumatism (EULAR) recommends that persons with CMC1-OA should be provided with orthoses to prevent or correct subluxation and deformity of the thumb [1]. The effect of CMC1 orthoses has been addressed in six systematic reviews [1, 12, 13, 35, 36, 44]. Based on a meta-analysis of two randomized controlled trials, the authors in one of the most recent reviews conclude that there is consistent evidence that orthoses reduce hand pain [13]. The results from single trials included in this review further indicate that subluxation was reduced when wearing orthoses and more reduced with a rigid orthosis than with a soft one [45]. In the two included studies with a low risk of bias, a long and rigid orthosis was found to be pain relieving and well tolerated for long-term night use [46], while a shorter orthosis used during activities of daily living significantly reduced pain [47]. An effective regimen may therefore be a combination of a small and rigid orthosis for pain relief during the day and a rigid nighttime orthosis for prevention of deformities.

Appropriate outcome measures, based on patients’ concerns, are necessary for the comprehensive evaluation of existing and new interventions in HOA [48]. Such instruments should also be feasible, valid, reliable, and responsive to change. Currently, two patient reported outcomes (PROs) have been developed specifically for HOA: the Australian/Canadian Hand Osteoarthritis Index (AUSCAN) [49] and the Functional Index of HOA (FIHOA) [50]. In a Norwegian study comparing the psychometric properties of these measures, the results indicated that both of these questionnaires contained items that were deemed irrelevant by the Norwegian respondents, such as gender-specific questions and items addressing activities that are not relevant in a Norwegian context [51]. The items are further limited to activities within self-care and house work, and do not include any questions related to activities within work or leisure. The Measure of Activity Performance of the Hand (MAP-Hand) is a PRO where the items were generated through interviews with people with arthritis [52]. In addition to 18 standardized items, the questionnaire includes five open questions, where patients can add other activities of importance to daily functioning. This provides clinicians with important information for goal-setting processes, planning and evaluation of interventions, and meet the need for outcome measures that incorporate the patient perspective as stated by the Outcome Measures in Rheumatology Group (OMERACT) [48]. MAP-Hand has been tested for validity and responsiveness in rheumatoid arthritis (RA) and HOA with good results, but studies of reliability, including estimating the smallest detectable difference and minimal important change are warranted [53].

The main objective will be to evaluate whether occupational therapy reduces or delays the need for surgery in carpometacarpal osteoarthritis during the waiting period prior to surgical consultation.

More specifically, our study will consider the following research questions:

The trial has been designed in line with the SPIRIT guidelines [54] in a project group consisting of the following key stakeholders: Occupational therapists with experience of treating patients with CMC1-OA, researchers with expertise in HOA, and a patient research partner with experience from living with HOA, including having undergone CMC1-surgery and receiving non-pharmacological interventions. Members of the project group will be engaged throughout each stage of the trial and will contribute in the process of integrating study results in clinical practice.

This is a Norwegian multicentre randomized controlled trial (Trial registration: NCT01794754). The trial will be conducted at the departments of rheumatology at St. Olav’s Hospital, Trondheim University Hospital in Trondheim, Haukeland University Hospital in Bergen and Haugesund Rheumatism Hospital in Haugesund, in collaboration with the National advisory unit on rehabilitation in rheumatology, which is located at Diakonhjemmet Hospital in Oslo. Participants will be allocated to receive either standard care (no intervention) or occupational therapy during the waiting period between referral for and the actual surgical consultation (see Fig. 1). The occupational therapy intervention represents a supplement to standard HOA care, and no participant will receive an intervention that is below the current standard of care.

All persons who are referred by their GP for surgical consultation due to their CMC1-OA at one of the three participating departments of rheumatology are eligible for the RCT. However, persons who do not speak the Norwegian language will be excluded, as will those with cognitive dysfunction or with other diseases or injuries that may negatively impact hand function. The local project coordinator at each centre will have access to lists of patients who are referred for surgical consultation and will mail written information about the study to potential participants. The local project coordinator will telephone those who agree to participate and will book them for the next convenient appointment for the baseline assessment, which is expected to last approximately one hour. A questionnaire will be sent to the participant, who will be encouraged to fill it out and bring it to the appointment. On arrival at the centre, the local project coordinator will screen the potential participant for eligibility before discussing the study with participants and obtaining their written informed consent prior to further assessment and randomization.

Those who do not respond to the mailed written information or who fail to appear at the appointment for baseline assessment will not be re-contacted regarding participation in the study.

Participant flow is shown in Fig. 1. Follow-up will be at 4 months (immediately before the surgical consultation), 18 months and 24 months after randomization to evaluate short, medium and long-term effects. Before each follow-up, a questionnaire will be sent to the participants, who will be asked to fill it in at home and bring it to the follow-up appointment.

Additionally, participants in the control group will receive a phone call 2 weeks after baseline assessments, and will complete the MAP-Hand in a telephone interview. Participants in the intervention group will be convened for a second consultation 2 weeks after baseline, at which time the orthoses and the hand exercise programme will be adjusted, and additional information regarding use of assistive devices will be given if needed.

The surgical consultation takes place immediately after the follow-up assessment at 4 months. Based on the patient’s symptoms, functional limitations and degree of CMC1-OA, the surgeon may reach one of the three following conclusions;

All participants who are operated on, regardless of group affiliation, will be given post-operative treatment according to surgical procedures and recommendations.

The third and fourth follow-up will be after 18 and 24 months, respectively. Baseline assessments will be repeated, and for participants in the occupational therapy group, exercises and orthoses will be adjusted if needed. Participants in the control group will receive the occupational therapy intervention after assessments at the control at 24 months.

Each participant will have a special numeric code, and all data will be stored in a fireproof and locked cupboard and in a digital file on the research server at Diakonhjemmet Hospital. Selected members of the project group (project leader, local coordinators, bio-statistician, study assistants, a PhD-student and a post doc research fellow) can access the file. The OT will check missing data in the questionnaires at each follow-up. An independent study assistant will check incorrect and double data entries.

The randomization schedule will be prepared by a biostatistician making a computer-generated, randomized list with a block size of 10.

A secretary will prepare sealed, opaque envelopes containing the patient’s assignment to either the occupational therapy or the control group. The envelopes will be stored in a locked closet at each centre and may be opened by the participant after baseline assessments and information about HOA are completed.

In this trial, participants and therapists delivering the intervention are aware of the treatment assigned. To achieve blinding in the surgical consultation, participants will be asked not to inform the surgeon about their group allocation during the consultation. Moreover, the statistician who will perform the main statistical analyses will be blinded to group allocation.

The Norwegian Regional Ethical Committee have approved the study (2012/2265/REK sør-øst C). The research will be carried out in compliance with the Helsinki Declaration. Personal confidentiality is guaranteed, and each participant, having received detailed information on the study processes and purposes and affirming their right to withdraw from the study at any time without any explanation, will sign a declaration of voluntary participation. No participant will receive an intervention that is below the current standard of care, since participants in the control group will receive standard HOA care, whereas participants in the intervention group will receive an occupational therapy intervention as a supplement to standard HOA care.

All data collected will be regarded as confidential and will be securely stored in paper formats in a locked closet in a locked room. The project leader shall regularly remain in touch with the local project leaders and provide assistance in the data collection process, in addition to ensuring that the data are collected, stored and quality assured in accordance with current guidelines.

Six occupational therapists (two at each centre) with at least 4 years of experience within rheumatology will provide the occupational therapy intervention. Before the start of the inclusion process, the therapists will meet together with the project leader and the patient research partner and go through assessment and treatment procedures to ensure that these are performed and delivered consistently. Meetings will be held regularly throughout the trial period. Additionally, the project leader will answer therapists’ questions by mail or telephone between meetings.

Both groups will receive medical treatment as usual.

Participants in the intervention group will document exercises and the use of orthoses in a treatment diary (see Table 1). The diary will have one page for each week in the 3-month treatment period. Each of these pages contains three sections in which the participants will report date and length of the individual exercise session, rate their pain immediately after exercising on an 11-point numeric rating scale (0 = no pain), and provide comments. On each page, there is also one section listing each day of the week where participants will record days and nights on which they used the orthosis, the length of use in hours and minutes, and add any comments they may have.

The primary outcome will be the number of participants in each group who have received surgical treatment after 2 years.

Secondary and tertiary study outcomes are documented in Table 4 and are based on previously validated measures.

Patient-reported outcome measures will be collected at baseline, 4, 18 and 24 months and will include the following symptoms, functional aspects and measures: Activity performance measured by the MAP-Hand [52, 53]; physical function and symptoms in the arm, shoulder and hand measured by the Quick-Dash [56]; current situation for main reason for wanting CMC1-surgery on a 0–10 numeric rating scale; pain at rest and pain following measures of grip and pinch strength in right and left hand, respectively, on 0–10 numeric rating scales; satisfaction with care for HOA at current centre on a 0–10 numeric rating scale; comfort with use of assistive devices reported on a 0–10 numeric rating scale; and health-related quality of life using the EuroQol EQ-5D [57]. Participants in the control group will also complete the MAP-Hand in a telephone interview 2 weeks after baseline assessments. At the time of rescoring, perceived change in disease activity will be recorded on a global rating scale by asking participants to compare their current status with how they felt 2 weeks ago and mark on a five-point scale whether the disease has become much worse, slightly worse, unchanged, slightly better, or much better.

Additionally, the following observer-reported outcome measures will be reported at baseline, 4, 18 and 24 months: number of painful joints in each hand; joint mobility of digits 2–5 recorded as flexion deficit in mm for each finger [58]; range of motion of thumb metacarpophalangeal and interphalangeal joints measured by a goniometer; CMC1 and palmar abduction of the thumb measured with the Pollexograph [59]; and grip and pinch strength measured in Newton by the Grippit electronic instrument [60]. The therapists will also document the time spent to provide the occupational therapy intervention at baseline, 2 weeks (intervention group only), 4, 18 and 24 months.

Further, the direct and indirect costs in the study period will be self-reported at 4, 18 and 24 months as the number of days of sick leave and absence from paid work, number of visits to a given list of health providers and the number of hospital visits or stays over the period since previous control (see table 5). All participants will also report on any medication taken for their HOA, as well as medical or technical equipment purchased.

Participants in the intervention group will report treatment adherence in the treatment period between baseline and 4 months in a treatment diary (see Tables 1 and 5).

To be able to describe and compare participants, we will, at baseline, record the participants’ age, gender, marital status, employment status, level of education, hand dominance, other diseases or previous injuries that may negatively impact hand function, previous hand surgery, in which hand they are referred for surgery, previous hand therapy, main reason for wanting surgery, motivation for surgery, current medication taken for HOA, and joints with osteoarthritis (including interphalangeal joints with bony enlargement in each hand). The degree of CMC1-OA and trapeziometacarpal subluxation and the presence of erosive HOA will be evaluated based on radiographs taken in relation to referral to surgical evaluation (Table 5) [61].

The primary outcome will be differences between the groups in the number of participants who have received surgical treatment after 2 years. Calculation of sample size is based on the assumptions that at least 70% of those who are currently deemed eligible for surgery will undergo operations, and that the number of operations in the intervention group will be equal to or less than that of the control group (one-sided calculation). To be able to detect a difference of 20% in surgery rate between the groups, with a power of 80%, a significance level of 0.05 and a 20% dropout rate, a total of 180 participants (90 in each group) will be needed.

The main treatment analysis will be conducted blinded to treatment allocation and will be analysed on an intention-to-treat basis with all randomized participants retaining their original randomized group. In the analysis of the primary outcome, the number of participants in each group who have received surgical treatment after 2 years will be compared using binary mixed models logistic regression and presented as odds ratios with a 95% CI.

We will also use mixed Cox proportional hazard models to describe and compare time to operation in the two groups, including establishment of the vital assumption of proportionality of the hazard rates.

In terms of secondary and tertiary outcomes, differences in mean values with 95% CI at each follow up will be analysed using mixed models analysis of covariance, adjusting for baseline levels of the outcome measure, while generalized linear mixed models will be used to estimate the overall effect for the total 2-year trial period, adjusting for the effect of clustering and repeated measures over time.

The assessment of the psychometric properties of MAP-Hand (research question 4) will adhere to the methods recommended by the international COSMIN group, using Kappa and ICC for examining test-retest reliability, and an anchor-based approach (based on the five-point global rating scale) to determine the minimal important change [62]. The smallest detectable change will be assessed by examining limits of agreement of repeated measures, computed as ± 1.96 SD of the difference between baseline scores and retest scores. The results from participants who report change in disease activity between baseline and rescoring will be excluded from the analysis of test-retest reliability and the smallest detectable change.

To assess the cost-effectiveness, we will need to estimate health outcomes and costs. The health outcomes will be measured using the EQ-5D and the costs will include the cost of the two interventions. This comprises the hours and frequency of occupational therapy, costs related to the provision of assistive devices, orthoses, and exercise material, and costs related to the surgical procedures, including post-operative treatment. Furthermore, costs related to medical or technical equipment purchased by participants and to the use of other health care services (home care services, rehabilitation, and institution) will be recorded for both the intervention and control group during the trial period.

Standard methods for economic evaluation will be applied and the cost-effectiveness will be calculated as the incremental cost-effectiveness ratio, which is defined by the cost per incremental QALY.

The level of significance will be set to 0.05 and we will use IBM SPSS Statistics 21 (IBM Corporation), R (R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.) and Statistical Analysis System (SAS Institute Inc., Cary, NC, USA).

We plan to publish at least seven articles in international peer-reviewed journals using the data collected in this study. Three articles will be written by a PhD-student (corresponding to the first three research questions in the project description) and four by a post doc research fellow (corresponding to the four last research questions in the project description).

The results will also be disseminated through the information channels of the project group members, including web sites, national and international multidisciplinary networks of health professionals, national and international research conferences and congresses, and through lectures in bachelor and master programmes.

To reach patients and their relatives, the results will be published in the yearly special issue of the journal of the Norwegian Rheumatism Association, which focuses on recent research, and at conferences and meetings arranged by the association.

We also plan to arrange a national workshop for health professionals, where the results are presented and where participants’ can learn how to perform the assessments and deliver the interventions used in the study.