Manual and manipulative therapy in addition to rehabilitation for osteoarthritis of the knee: assessor-blind randomized pilot trial.

ORIGINAL ARTICLES MANUAL AND MANIPULATIVE THERAPY IN ADDITION TO REHABILITATION FOR OSTEOARTHRITIS OF THE KNEE: ASSESSOR-BLIND RANDOMIZED PILOT TRIAL Lauren Dwyer, MTech(Chiro), a Gregory F. Parkin-Smith, MTech(Chiro), MBBS, MSc, DrHC, b

James W. Brantingham, DC, PhD, c, 1 Charmaine Korporaal, MTech(Chiro), d Tammy K. Cassa, DC, e Gary Globe, PhD, MBA, f Debra Bonnefin, DC, MAppSc, g and Victor Tong, DC, MBA h ABSTRACT Objectives: The purpose of this study was to examine the methodological integrity, sample size requirements, and short-term preliminary clinical outcomes of manual and manipulative therapy (MMT) in addition to a rehabilitation program for symptomatic knee osteoarthritis (OA). Methods: This was a pilot study of an assessor-blinded, randomized, parallel-group trial in 2 independent universitybased outpatient clinics. Participants with knee OA were randomized to 3 groups: 6 MMT sessions alone, training in rehabilitation followed by a home rehabilitation program alone, or MMT plus the same rehabilitation program, respectively. Six MMT treatment sessions (provided by a chiropractic intern under supervision or by an experienced chiropractor) were provided to participants over the 4-week treatment period. The primary outcome was a description of the research methodology and sample size estimation for a confirmatory study. The secondary outcome was the short-term preliminary clinical outcomes. Data were collected at baseline and 5 weeks using the Western Ontario and McMasters Osteoarthritis Index questionnaire, goniometry for knee flexion/extension, and the McMaster Overall Therapy Effectiveness inventory. Analysis of variance was used to compare differences between groups. Results: Eighty-three patients were randomly allocated to 1 of the 3 groups (27, 28, and 28, respectively). Despite 5 dropouts, the data from 78 participants were available for analysis with 10% of scores missing. A minimum of 462 patients is required for a confirmatory 3-arm trial including the respective interventions, accounting for cluster effects and a 20% dropout rate. Statistically significant and clinically meaningful changes in scores from baseline to week 5 were found for all groups for the Western Ontario and McMasters Osteoarthritis Index (P ≤ .008), with a greater change in scores for MMT and MMT plus rehabilitation. Between-group comparison did not reveal statistically significant differences between group scores at week 5 for any of the outcome measures (P ≥ .46). Conclusions: This pilot trial suggests that a confirmatory trial is feasible. There were significant changes in scores from baseline to week 5 across all groups, suggesting that all 3 treatment approaches may be of benefit to patients with mild-tomoderate knee OA, justifying a confirmatory trial to compare these interventions. (J Manipulative Physiol Ther 2015;38:1-21.e2) Key Indexing Terms: Pilot Projects; Osteoarthritis; Knee; Musculoskeletal Manipulations; Rehabilitation

O

steoarthritis (OA) is the most common form of chronic arthritis, with radiologic evidence of joint degeneration in more than 50% of people older

a Private Practice, Department of Chiropractic, Durban University of Technology, Durban, South Africa. b Adjunct Senior Lecturer, Murdoch University, Murdoch, Australia. c Adjunct Associate Professor, School of Chiropractic and Sports Science, Murdoch University, Murdoch, Australia. d Senior Lecturer in Chiropractic, Department of Chiropractic, Durban University of Technology, Durban, South Africa. e Practicing Chiropractor, Erie, PA. f Director, Global Health Economics, Amgen Inc, Thousand Oaks, CA. g Associate Professor, Texas Chiropractic College, Pasadena, TX.

than 65 years, with a propensity for symptomatic OA to occur in women. 1,2 The knee joint is one of the most commonly affected by OA, leading to a burden on the h

Radiologist, Rowland Heights, CA. Submit requests for reprints to: James W. Brantingham, DC, PhD, Adjunct Associate Professor, 250 Whiteside Place, Thousand Oaks, CA 91362. (e-mail: [email protected]). 1 Formerly Director of Research and Associate Professor, Cleveland Chiropractic College, Los Angeles, CA.

Paper submitted September 28, 2013; in revised form October 10, 2014; accepted October 13, 2014. 0161-4754 Copyright © 2015 by National University of Health Sciences. http://dx.doi.org/10.1016/j.jmpt.2014.10.002

2

Dwyer et al Manual Therapy and Rehabilitation for Knee OA

individual and the community through reduced quality of life, diminished capacity for employment, and an increase in health care costs. 3 With an aging population globally, the problem of OA of the knee will likely worsen, which has prompted the development of clinical practice guidelines that now recommend treatment at an early stage of the disease, using health promotion/exercise to reduce progression, and optimization of both pharmacologic and nonpharmacologic treatment. 4 For mild to moderate OA, clinical practice guidelines promote a multimodal treatment approach that includes drug therapy and nonpharmacologic treatment, such as patient education, exercise, and weight optimization. 5 Physical therapies, such as manual and manipulative therapy (MMT), are also emerging as a viable treatment option, 6–8 particularly in combination with other interventions like exercise, 9 health promotion, and simple analgesia. 10 Manual and manipulative therapy can offer positive outcomes for knee OA 8,11,12 but needs to be explored further within the context of multimodal care. Combining treatments to form a package of care for a specific disorder implies the delivery of a complex intervention, the outcome of which may be confounded by internal and external factors. 13 Therefore, the trials testing complex interventions require modeling or piloting prior to a confirmatory trial with a view to optimize the research design and determine feasibility. 14–16 The purpose of this trial was to conduct an external pilot trial and compare MMT, a rehabilitation program, and the combination of MMT plus the rehabilitation program for mild to moderate OA of the knee. The primary outcome was to assess the design integrity, with a view to provide a critique of the research design, offer design recommendations, and provide a sample size estimate for a larger confirmatory trial. The secondary outcome was to gain insight into preliminary clinical outcomes. The research hypothesis was that the 3 treatment approaches would show significant change in scores from baseline to week 5 and that there would be a significant difference between treatment groups. To achieve the primary end-point, the following objectives were covered by descriptive analysis of (a) patient recruitment, enrollment, and allocation; (b) data collection and patient retention; (c) acceptability of interventions; and (d) sample size calculations.

METHODS Trial Design and Changes to the Methods After Commencement This study was initially designed as an assessor-blind, parallel-group clinical trial with 1-, 3-, and 6-month follow-up. However, toward the latter part of this study, access to funding became more difficult, which impacted recruitment and data collection. The authors realized that a pilot study was needed to inform the design of future trials to account for financial and

Journal of Manipulative and Physiological Therapeutics January 2015

recruitment problems. Therefore, this trial is presented as a 2-center, assessor-blind, parallel-group short-term clinical trial with balanced randomization (1:1). The results and recommendations of this pilot trial may be considered a template for a fully-powered confirmatory trial. The study protocol was not updated as regularly as was desired on ClinicalTrials.gov. However, third-party access resulted in premature cancellation/closure of the study on the website. The recommendation for future trials is to limit third-party access to the protocol on ClinicalTrials.gov and to ensure regular updating, to facilitate transparency in the conduct of the trial. This project received approval from the relevant institutional review board of Cleveland Chiropractic College (IRB No. 08132010) and the Research Ethics Committee of Durban University of Technology. The trial was registered with Clinical.Trials.gov (NCT01188837).

Evaluation of the Research Integrity and Methodological Rigor The primary outcomes of this study included the following: patient recruitment, enrollment, and allocation; data collection and patient retention; acceptability of interventions; sample size calculations; and research design alternative. Sample size calculations were performed and reported using the data from this external pilot trial. Initial calculations were based on the F test (analysis of covariance [ANCOVA]) for fixed effects, main effects, and interaction effects at a 95% level of confidence to achieve statistical Power of 80% (using G* Power software, www.psycho.unidusseldorf.de, Dusseldorf, Germany) taking into account the effect size, the number of groups, the number of covariates, and the degrees of freedom for the outcome measure (Western Ontario and McMasters Osteoarthritis Index [WOMAC]). For research design alternative, multisite or cluster randomized trials as a potential research design were discussed. Accordingly, a sample size estimate is offered with this design in mind.

Selection Criteria Inclusion Criteria. All eligible participants had a diagnosis of mild-moderate knee OA based on the diagnostic criteria of the American College of Rheumatology and the Kellgren-Lawrence grade for knee OA, suitable grades for this study being grades 0 to 3. 17,18 This diagnosis was reached after assessment by the lead clinician and radiographic investigation of each eligible participant. The radiographs were read and interpreted by a Diplomate in Chiropractic Radiology who was blind to the study criteria, data, and randomization. Inclusion and exclusion criteria are in Figure 1.

Setting, Practitioners, and Participants This study was implemented across 2 chiropractic university-based outpatient teaching clinics, 1 in the city of

Journal of Manipulative and Physiological Therapeutics Volume 38, Number 1

Inclusion Criteria: 1. Age ≥38 and ≤80 years; 2. Knee pain ≥ 1 year and able to stand and walk without severe varus/valgus deformity and/or severe instability (instability being defined as a significant increase in anterior drawer or varus/valgus movement when compared to the opposite knee); 3. A minimum of1of the 3 clinical criteria below for diagnosis of knee OA (sensitivity 89% and specificity 88%) 17 : a. Knee pain and crepitus with active motion and morning stiffness ≤30 min; or b. Knee pain and crepitus with active motion and morning stiffness >30 minutes and bony enlargement; or c. Knee pain and no crepitus and bony enlargement – bony enlargement being determined on palpation and supplemented by observations of radiographs. 4. No history of knee surgery in the past 6 months;19 5. Kellgren and Lawrence grade of 0-3 on plain-film radiographs; 18 6. Ability to stand and walk without assistance for most (≥¾) of the day, as keeping active and performing exercises would otherwise be difficult; 7. A participant was required to have a score of ≥720/2400 (≥30%) on the WOMAC; and 8. No previous manual and/or manipulative therapy for their knee pain. Exclusion Criteria: 1. Kellgren-Lawrence grade 4 knee degenerative changes on plain-film radiographs, indicating severe knee OA; 2. Possibility of serious pathological or psychiatric disorders; 3. Possibility of a disorder that would prevent the participant performing exercises or contraindications to MMT.

Fig 1. Inclusion and exclusion criteria. MMT, manual and manipulative therapy; OA, osteoarthritis; WOMAC, a 24-question survey used to measure pain, joint stiffness, and function (disability).

Durban, South Africa, at Durban University of Technology and the other in Los Angeles, California, at Cleveland Chiropractic College, Los Angeles, in the United States. Participants were recruited over a 6-month period, from September 2010 to February 2011, with the study being advertised on campus, on local radio, and in local newspapers. Potential study candidates with knee pain who were seeking care responded to the advertising by contacting the clinical lead at each study site. The lead clinician screened each potential participant by telephone or in person for eligibility, the screening criteria being the following: (a) knee pain ≥ 1 year, (b) no history of knee surgery in the past 6 months, (c) aged ≥ 38 and ≤ 80 years, (d) the knee pain was not to be due to an acute injury, and (e) had no previous manual and/or manipulative therapy for their knee pain. 17,19,20 The patient information sheet and consent form is offered in Appendix A which is available online. Potential patients made an appointment with the lead clinician at the study site to undergo a full assessment and complete an information sheet and were invited to participate


in the study. Eligible patients who volunteered for the study were offered a cooling-off period of up to 3 days, with those choosing to proceed giving written informed consent and randomly allocated to 1 of the treatment groups. 21 The rationale for the cooling-off period was 2-fold—firstly, to allow the potential participant to make an informed, considered decision about joining the study (ie, no coercion to join the study), and to ensure participants were genuinely experiencing an episode of symptoms, because knee OA often has a relapsing/remitting course or acute or chronic symptoms. Treatment was provided by a senior chiropractic intern under supervision or by an experienced chiropractor. Supervising and treating chiropractors had more than 5 years of clinical experience and specific experience in either sportstype and/or extremity injuries. Most patients were treated by the same treating clinician at each site.

Interventions Group 1: MMT Alone. Participants were required to undergo 6 treatments sessions over the 4-week treatment period, then a final consultation during week 5 for data collection only. The limit of absence was 3 or more missed appointments. The frequency of treatment was approximately twice per week. Each treatment session was approximately 20 minutes, during which time each participant was also assessed for adverse signs such as pain, joint swelling, or redness. The treatment would then be adjusted to participant tolerance, at the discretion of the clinician. Each participant in this group had 6 treatment sessions with a total treatment dose of around 120 minutes (2 hours). Patient education about the diagnosis and prognosis was provided, alongside lifestyle advice, as described in the informed consent and other handouts given to the patients. The treatment comprised joint mobilization (grades 1-4) and joint manipulation (grade 5; high-velocity, low-amplitude, thrust-type manipulation) of the affected kinematic chain (knee, hip, foot, and spine). Manipulation, mobilization, and soft tissue treatment were based on techniques previously described. 19,22,23 Manipulation was applied to joints with restricted range of motion (ROM), identified by joint motion palpation by the treating clinician, using the high-velocity low-amplitude manipulations noted above or, as described in textbooks and other peer-reviewed papers. 24–26 Forced end-ROM grade 4 mobilizations or grade 5 thrust manipulations were avoided, particularly in flexion and extension, where it was likely to worsen symptoms or could not be tolerated by the participant. 22,23 A description of the 3 most commonly used mobilization and manipulation techniques were as follows: Manual 1: second oscillating mobilization of the knee in flexion (and separately into extension) starting at the point of stiffness with lesser grades II to III and progressing to IV++ or rarely to grade V if safe and appropriate; 2 to 6 sets of 30 repetitions. As appropriate a light—varus (adduction stress), valgus, (abduction stress), or gentle—internal rotation

3

4


may be added if felt indicated per motion palpation technique (applied at any joint that might be a candidate for MMT) as described and noted in commonly used textbooks, now further supported by the recently published review of methods used by chiropractors to determine the site for applying manipulation. 23,27,28 Patellar mobilization used could be applied with various techniques developed at Macquarie University using their patellar myofascial mobilization procedure where the participant is seated on an examination table with their knee over the practitioner's ipsilateral knee and in relaxed flexion. The practitioner places a web contact (or pincer with thumb and index finger) against the superior pole of the patella while the participant extends the knee and the patella is gently stressed superior to inferior with a bias toward medial or lateral rotation. This is repeated up to 10 times (additionally a light thrust may be applied at any time throughout the mobilizations; see the published peer-reviewed paper for complete details). 19 Patellar mobilization was also performed as per a previous trial from Durban University of Technology (with the supine participant's knee slightly flexed and with the participant's knee between the practitioner's legs); the thumbs of the practitioner remove the slack from superior to inferior and either lateral to medial or medial to lateral against and slightly lifting up the edge of the patella. As the knee is brought into extension, the patella is stressed inferiorly or distally with a combined lateral or medial stress. Based on the experience of 1 of the authors (J.W.B.), it was suggested that the patella be most often mobilized with a bias of lateral to medial then superior to inferior. However, the patellar could also be simply mobilized from lateral to medial from the side of the participant and/or in any direction of apparent restriction. 23,28,29 If the technique developed at Macquarie was not used, then as per the 2005 trial previously noted, the patella was mobilized for 2 to 6 sets of 30 repetitions. 30 Axial elongation of the knee in extension was frequently performed as nonthrust (as a lesser grade of mobilization) but often initially as a grade IV or IV++ as a light thrust 2 to 3 times, and if no side effects were detected within a few treatments, it was delivered as a full high-velocity, low-amplitude (HVLA) grade V thrust manipulation, 1 up to 3 times. Axial elongation or distractive thrust (pulling/ tractioning or thrusting the femur and tibia apart) was done without any combined forced knee extension. 23,29,31 Other manipulations could be applied if strongly indicated (presenting with pain and/or functional deficit or both) per the lumbar spine, sacroiliac joint, the hip (per a specific protocol if OA was present), 32 or full lower extremity as per the literature previously indicated. 26,27,31

Group 2: Rehabilitation Program. The rehabilitation program included patient education, exercise prescription, soft tissue treatment, and passive stretches (on treatment days only) to the knee and elsewhere along the full kinetic chain, where needed, based on functional assessment.


Education consisted of information about the diagnosis and prognosis, and advice on health promotion and lifestyle. The content and timing of treatment were important in that advice, education, and training were provided to participants at the onset of their treatment program (week 1) and reinforced at 2 other points during the treatment period (weeks 2 and 4). This was to reinforce the need for rehabilitation and to encourage compliance. The home-exercise prescription was standardized for use across the trial centers, yet still allowing for minor case by case variations. The exercise program was similar to that described previously. 22 At the first treatment session during week 1, participants received training in the exercise program, which was to be performed at home for around 20 minutes, at least 3 times per week. Participants were required to return for a second and third treatment sessions during weeks 2 and 4 to reinforce the treatment approach, ensure exercises were being done correctly, and encourage compliance. Participants were provided a pamphlet of information with written descriptions and drawings of the exercises, alongside direct supervision/training. The aim of the rehabilitation program was to keep it simple, to encourage compliance, and to contain costs by keeping supervision to the minimum. This is likely to replicate clinical practice more closely because patients do not necessarily have the motivation, time, or finances to undertake regular supervised rehabilitation. Recommendations on the rehabilitation program are provided latterly. Each treatment session was approximately 20 minutes, during which time each participant was also assessed for adverse signs such as pain, joint swelling, or redness. The exercise regime would then be adjusted to participant tolerance, performing less exercise if there were adverse symptoms. Each participant in this group had 3 treatment sessions (of around 20 minutes each) with a total treatment dose of around 60 minutes (1 hour), and total exercise dose was approximately 240 minutes (4 hours; 12 exercise sessions of 20 minutes each) over the treatment period. A diary was kept by all patients in the exercise group to monitor at home exercise compliance. Group 2: Rehabilitation Program The rehabilitation program included patient education, exercise prescription, soft tissue treatment, and passive stretches (on treatment days) to the knee and elsewhere along the full kinetic chain, where needed, based on functional assessment. 22 Functional examination included active and passive ROM, manual muscle strength and length, reflex, sensory and myofascial restriction or tenderness and palpation of the lumbar spine, hip, ankle, foot, and knee joints. In addition, squatting, performance of step-ups, and similar evaluation were also checked to determine the ability of participants to participate in exercise and manual therapy protocols. Education consisted of information about the diagnosis and prognosis, and advice on health promotion and lifestyle.


The participant was reassured that knee OA is not an immediately life-threatening or serious problem, although it could be painful. The participant was advised to stay active as much as possible and that nonaggravating physical activity was helpful and to return to normal activities, including work, as soon as possible leads to better recovery and function. Participants were informed that exercise was important for OA of the knee as it helped to reduce pain and maintain general health. Low-impact activities, where there was less weight going through the hips and knees were recommended. Participants were advised that being overweight could lead to more pain and damage to hips and knees affected by OA. The content and timing of treatment were important in that advice, education, and training were provided to participants at the onset of their treatment program (week 1) and reinforced at 2 other points during the treatment period (weeks 2 and 4). This was to reinforce the need for rehabilitation and to encourage compliance. Most of the soft-tissue therapy and passive stretching, which was used on 3 occasions over the treatment period during weeks 1, 2 and 4, was applied in the following manner: focused soft tissue treatment was applied similar to a previous trial, including circular finger and palm pressure (forms of point pressure, petrissage, and effleurage), and stretch to the knee and other joints along the full kinetic chain based on functional assessment. 22,32 The home-exercise program was similar to that described in a previous trial, 22 which included strengthening exercises, stretching, supine ROMs exercises, and walking. At the first treatment session during week 1, participants received training in the exercise program, which was to be performed at home, at least 3 times per week. Participants were required to return for a second and third treatment sessions during weeks 2 and 4 to reinforce the treatment approach, ensure exercises were being done correctly, and encourage compliance. Each participant was also assessed for adverse signs such as pain, joint swelling, or redness. The exercise regime would then be adjusted to participant tolerance, performing less exercise if there were adverse symptoms. Each participant in all 3 groups had 3 treatment sessions. A diary was kept by all patients to monitor at home exercise compliance. Group 3: MMT Plus Rehabilitation. Participants in group 3 received a combination of the treatments described for group 1 and group 2. Participants were required to skip exercise on the days they received MMT. The duration of each session was around 20 minutes, with an additional 20 minutes for patient education and training during 3 sessions, with the total treatment dose approximately 180 minutes (3 hours) and total exercise dose of around 240 minutes (4 hours). Each participant in this group had 6 treatment sessions over the 4-week treatment period, with additional time during a treatment session in weeks 1, 2, and 4 to reinforce patient education and the exercise program. Participants were required to skip exercise on the days they received MMT.


Outcome Measures The outcome measures were as follows: The WOMAC, a 24-question survey, was used to measure pain, joint stiffness, and function (disability). Each parameter was scored out of 100, with a maximum total score of 2400 (pain 500, stiffness 200, function 1700). 33,34 The WOMAC index is a valid and reliable measure of changes in pain and function, and used in various knee OA trials, where a 20% overall change in score (ie, ≥ 480/2400) is reported as being clinically meaningful; 33,34 Knee ROMs (flexion/extension) were measured by the same research assistant at each study site. Knee flexion and extension were performed 3 times per each evaluation, as described by Norkin and White and using a Saunders digital inclinometer (Saunders Equipment Company 1998, Florida, CA). The mean of these measurements was then obtained. 35–37 Digital Inclinometry or goniometry has adequate intraexaminer reliability including for OA of the knee (intratester reliability with an interexaminer correlation coefficient of 0.75-0.88 or higher). 38–40 The McMaster Overall Therapy Effectiveness (OTE) inventory, which measures global patient-perceived satisfaction and improvement with care, is based on a 15-point Likert scale where 0 indicates no change; − 7, maximum deterioration; and + 7, maximum improvement. 41–43 The minimal important difference of quality of life score per item on a 15-point scale (− 7 to + 7) is very close to 0.5 (range, 0.42-0.58), where differences of approximately 1.0 represented a moderate change (range, 0.77-1.51) and differences greater than 1.5 represented large changes. 41 For the OTE, a clinically meaningful posttreatment score is a change of 14% (≥ 1/7). 41 Baseline demographic data and statistics were reported in Table 1, for total or cumulative WOMAC in Table 2, for knee ROM in Tables 3 and 4, and, finally, for the OTE in Table 5. Adverse events were monitored in each participant and determined at each treatment consultation by interview; patient notes were specifically designed to prompt such questions each visit. Participants had been previously informed that transient, benign side effects are often encountered with exercise and MMT, such as stiffness and discomfort and normally resolve. Although recorded in the patient notes, these were not reported as adverse effects. However, participants were informed that side effects of treatment, such as persistent or escalation of pain, or unintended effects, such as but not limited to, joint swelling/redness or inability to weight-bear, were considered adverse events and were to be immediately reported to the assessing/treating clinician because the participant was not to continue further treatment or perform exercises until it was determined safe to do so.

5

6



Table 1. Sample Characteristics and Descriptive Statistics Variable Age (y) Mean (SD) 95% CI Sex, female (frequency) Height (cm) Mean (S) 95% CI Weight (kg) Mean (SD) 95% CI BMI Mean (SD) 95% CI Onset (frequency) Insidious Traumatic Duration of symptoms (mo) Mean (SD) 95% CI Side (frequency) Left Right Both Occupation (frequency) Not working Nonmanual work Manual work Regular exercise (frequency) None Mild intensity Moderate intensity High intensity Grade of knee OA (frequency) Grade 0 Grade 1 Grade 2 Grade 3 No. of treatment sessions No. of exercise sessions during trial

Group 1 a (n b=26)

Group 2 c (n = 26)

Group 3 d (n = 26)

63.5 (10.9) 60.9 (10.3) 62.2 (11.8) 59.1-67.9 56.7-65.0 57.4-66.9 19 (73%) 15 (58%) 15 (58%) 165.9 (9.8) 167.1 (15.5) 165.9 (8.5) 162.0-169.9 160.8-173.4 162.4-169.3 78.6 (15.2) 85.8 (20.4) 84.3 (21.2) 72.5-84.8 77.6-94.1 75.7-92.8 28.6 (5.2) 26.5-30.7

30.8 (6.4) 28.2-33.4

30.6 (7.6) 27.5-33.7

22 (85%) 4

23 (88%) 3

20 (77%) 6

the full trial is likely to be inadequate, whereas another author suggests that a sample size for a pilot trial should be 30 or more participants per group, so that sufficient data are available to assess the study parameters accuracy. 44,45 Regarding sample size determination for pilot trials, “If the aim of a pilot study is to demonstrate intervention efficacy in a single group, a sample in the range of 20-25 will probably be adequate when population effect sizes are likely to be moderate or larger….” 44 The authors agreed on a minimum sample size of ≈ 25 per group (3 groups), while considering available funding, practical constraints, and the available time frame for recruitment. 44,45

Randomization and Data Management

89.3 (81.4) 58.4 (53.4) 104.1 (131.0) 56.4-122.1 36.8-80.0 51.2-157.0 13 (50%) 11 2

13 (50%) 10 3

17 (65%) 7 2

9 15 (58%) 2

14 (54%) 12 0

16 (62%) 7 3

14 (54%) 9 3 0

12 (46%) 7 7 0

16 (62%) 9 1 0

0 22 (85%) 4 0 6 n/a

0 16 (62%) 10 0 4 10

0 16 (62%) 10 0 6 9

BMI, body mass index; CI, confidence interval; OA, osteoarthritis. a Manipulative therapy alone. b Sample size. c Exercise prescription alone. d Manipulative therapy plus exercise.

The Pilot Trial Sample Size Usually, a definitive sample size is not required for a pilot trial, because hypothesis testing is not the primary endpoint. However, as a secondary outcome, this trial intended to offer insight into preliminary clinical outcomes to determine the need for a larger confirmatory trial. It has been suggested that the traditional sample size for a pilot trial of 10 per group or 10% of the expected sample size of

Randomization was conducted using a random numbers process where a sequence of random numbers with sequence boundaries was generated using computer software. This randomization process was conducted by 1 of the authors who did not participate as a clinician in the trial or collect data. 21 Each study site had its own randomization set/sequence, generated in the same way for each site. Each random treatment assignment was placed in a sealed opaque envelope, producing a series of sequentially numbered sealed envelopes, thus conferring allocation concealment. As each volunteer consented to participate, the clinic manager for each site, who was independent of the study, then opened an envelope and identified the treatment group to which the participant was assigned. This trial, therefore, followed an individually randomized design, with the clinical lead being informed of the allocation of participants, post hoc. This randomization process was adopted at both clinic sites. Data were collected by an independent research assistant who was not aware of the treatment groups or randomization. Treatment was provided by either the lead clinician or other vetted, participating clinicians, none of whom collected any data and were masked from the data and randomization.

Data Analysis Intention-to-treat analysis was performed by replacing missing data using a multiple imputations method (WinMICE prototype version 0.1; TNO Quality of Life, Leiden, the Netherlands, http://www.multiple-imputation.com). The calculations were performed using a multilevel regression model that created 5 multilevel imputations 46,47 . The mean of the 5 imputed values was then used to replace missing scores, and the data were then analyzed using standard statistical tests. To address the secondary outcome of this trial, clinical outcomes, data analysis was conducted. The data sets were amenable to the F test (MedCalc software; www.medcal.org) using weighted 2-tailed ANCOVA at a 95% confidence level, using data from baseline and the final consultation (week 5).



Table 2. Comparison of Cumulative or Total WOMAC Data of the 3 Treatment Groups Sphericity ε = 0.48 P ≤ .001

Greenhouse-Geisser Sphericity assumed Summary Statistics

Mean

SE

95% CI

Group 1 baseline Group 2 baseline Group 3 baseline Group 1 4th visit Group 2 4th visit Group 3 4th visit Group 1 1wFU Group 2 1wFU Group 3 1wFU

1113.1 997.1 1165.0 790.5 710.8 696.9 553.0 633.5 570.3

68.9 59.6 74.8 79.7 66.2 66.0 80.7 63.4 86.7

971.2-1255.0 874.2-1120.0 1011.1-1319.1 626.4-954.7 574.4-847.2 561.0-832.7 386.8-719.3 502.9-764.2 391.7-748.8

Mean Difference Within-group comparison Group 1 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Group 2 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Group 3 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Between-group comparison Group 1–Group 2 Baseline 4th visit 1wFU Group 2–Group 3 Baseline 4th visit 1wFU Group 1–Group 3 Baseline 4th visit 1wFU

SE

Pa

95% CI a

322.5 237.5 560.0

54.7 43.8 65.4

.0001 .0005 b.0001

126.0 to 519.1 80.0 to 395.0 324.8 to 795.2

286.3 77.2 363.5

75.5 38.4 78.1

.0305 1.0 .0033

14.8 to 557.9 − 60.7 to 215.1 82.8 to 644.3

374.5 126.6 570.3

110.8 45.2 109.6

.0860 .3462 .0003

− 24.0 to 773.0 − 35.7 to 289.0 218.1 to 1006.0

116.0 79.8 − 80.5

88.7 102.1 92.1

1.0 1.0 1.0

− 202.8 to 434.8 − 287.3 to 446.9 − 411.5 to 250.5

− 168.0 13.9 63.3

107.1 95.7 105.0

1.0 1.0 1.0

− 553.1 to 217.2 − 330.2 to 358.0 − 314.1 to 440.7

− 52.0 93.7 − 17.2

109.4 101.2 113.4

1.0 1.0 1.0

− 445.3 to 341.4 − 270.3 to 458.0 − 425.0 to 391.0

A significant change in cumulative or total WOMAC scores was found from baseline and to the 1-week follow-up for all 3 groups, suggesting that all 3treatments had a meaningful beneficial clinical effect over the study treatment period of 1 month and at the fifth-week follow-up. When comparing the 3 groups, no statistically significant differences were found in scores across all 3 time-points. P values less than 0.05 are statistically significant. 1wkFU, 1-week follow-up; CI, confidence interval; SE, standard error. a Boneferroni corrected.

Differences in baseline data and regression toward the mean were considered potential confounding factors and were accounted for statistically using the ANCOVA analysis. The data were statistically adjusted using the covariates of age, duration of symptoms, and body mass index. These covariates were chosen because they appeared dissimilar when examining the patient characteristics and confidence intervals, although they may not have been statistically different (Table 1). If significant differences between data sets were discovered, Tukey honestly significant

difference analyses were performed to establish where the differences lay.

RESULTS Evaluation of the Research Integrity and Methodological Rigor Recruitment, Enrollment, Allocation, and Patients. Flow diagram of participant recruitment and retention (CONSORT diagram) are found in Figure 2. Initial data analysis (Table 1) indicated that 13 and 70 participants were recruited at the

7

8



Table 3. Comparison of Knee Flexion ROM Data of the 3 Treatment Groups Sphericity ε = 0.39 P≤ .0029


Mean

SE

95% CI


99.0 100.8 93.8 105.5 106.0 103.6 106.1 105.5 109.7

3.8 4.0 5.3 3.5 3.6 2.9 3.6 3.5 3.0

91.1 to 106.9 92.6 to 109.0 83.0 to 105.0 98.3 to 112.6 98.5 to 113.4 97.7 to 109.5 98.7 to 113.5 98.4 to 112.6 103.5 to 115.9

SE

Pa

95% CI a

− 6.4 − 0.7 − 7.1

1.6 1.5 2.2

.0143 1.0 .106

− 12.095 to − 0.774 − 6.068 to 4.737 − 14.844 to 0.644

− 5.2 0.5 − 4.7

1.6 2.5 2.8

.149 1.0 1.0

− 11.074 to 0.720 − 8.666 to 9.590 − 14.905 to 5.475

− 9.7 − 6.2 − 15.9

4.8 3.0 4.7

1.0 1.0 .086

− 27.005 to 7.520 − 16.912 to 4.604 − 32.817 to 1.025

− 1.7 − 0.5 0.6

4.2 4.2 4.6

1.0 1.0 1.0

− 16.8 to 13.3 − 15.6 to 14.6 − 15.8 to 17.1

6.9 2.4 − 4.2

6.3 4.2 3.3

1.0 1.0 1.0

− 15.8 to 29.7 − 12.8 to 17.5 − 16.1 to 7.6

5.2 1.9 − 3.6

6.8 4.5 5.0

1.0 1.0 1.0

− 19.3 to 29.6 − 14.2 to 18.0 − 21.4 to 14.2


P values less than 0.05 are statistically significant. 1wkFU, 1-week follow-up; CI, confidence interval; SE, standard error. a Boneferroni corrected.

American and South African study sites, respectively. Between-group analysis of variance did not reveal any significant difference between the 2 study site groups in terms of age, height, grade of knee OA, or duration of symptoms (P ≥ .17), although there was a significant difference for weight and body mass index (P ≤ .01). This analysis was conducted prior to amalgamating the data into the 3 treatment groups (Table 1). The flow of participants is shown in the Consort Figure 2. Of the 111 participants screened, 91 were assessed, with 83 eligible participants volunteering for the study. Five participants dropped out of the study (3 “no shows”, 1 moved to another town, 1

underwent unrelated surgery), and there was a total of 10% (79/780) missing data, with the data from 78 participants available for analysis (26/group) after missing data were replaced. Examination of the descriptive and patient characteristics suggests that the 3 groups were similar at baseline (Table 1) as well for the WOMAC, ROM, and the patient satisfaction or global OTE score (see Tables 2, 3, 4, and 5). The baseline scores for the individual WOMAC domains of pain, stiffness, and disability can be found in Tables 6, 7, and 8. The average number of exercise sessions undertaken in group 2s and 3 over the treatment period was 9 and 10,



Table 4. Comparison of Knee Extension ROM Data of the 3 Treatment Groups Sphericity ε = 0.13 P = .006


Mean

SE

95% CI


4.7 4.5 3.6 5.8 5.1 8.0 6.2 5.2 9.0

0.8 0.7 0.8 0.8 0.9 3.8 0.9 0.8 4.2

3.1-6.3 3.0-6.0 2.0-5.2 4.1-7.5 3.3-7.0 0.2-15.7 4.4-8.0 3.5-6.9 0.2-17.7

Within-group comparison Group 1 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Group 2 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Group 3 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Between-group comparison Group 1–Group 2 Baseline 4th visit 1wFU Group 2–Group 3 Baseline 4th visit 1wFU Group 1–Group 3 Baseline 4th visit 1wFU

Mean Difference

SE

Pa

95% CI a

− 1.1 − 0.4 − 1.5

0.3 0.2 0.4

.044 1.0 .033

− 2.2 to 0.0 − 1.3 to 0.5 − 2.9 to − 0.1

− 0.6 − 0.1 − 0.7

0.3 0.3 0.3

1.0 1.0 1.0

− 1.7 to 0.5 − 1.3 to 1.0 − 1.9 to 0.5

− 4.3 − 1.0 − 5.4

4.0 0.5 4.5

1.0 1.0 1.0

− 18.9 to 10.3 − 3.0 to 0.9 − 21.8 to 11.0

0.2 0.7 1.0

0.5 0.5 0.6

1.0 1.0 1.0

− 1.6 to 2.0 − 1.1 to 2.6 − 1.1 to 3.1

0.9 − 2.8 − 3.8

0.5 3.9 4.5

1.0 1.0 1.0

− 1.0 to 2.8 − 17.0 to 11.4 − 20.1 to 12.5

1.1 − 2.1 − 2.8

0.6 4.1 4.6

1.0 1.0 1.0

− 1.0 to 3.1 − 16.8 to 12.6 − 19.4 to 13.9


respectively, based on participant diaries. The actual treatment doses, based on the actual number of treatments and exercise sessions undertaken in groups 1, 2, and 3, were 120, 240, and 360 minutes, respectively (Table 1). The average participant was a 62-years-old woman, was not working, was overweight, took very little exercise, and had mild-moderate OA of the knee. Participants were recruited over a 6-month period, from September 2010 to February 2011, delivering a patient screening rate of 20 potential patients per month (5/week) and a screening exclusion rate of 18% (3 per month). Monthly assessment/examination rate was 15 patients with an 8% exclusion rate (8/month).

Data Collection and Patient Retention. The WOMAC and the OTE surveys are easily accessible using an online search engine. Of the original 83 randomly allocated participants, there was a 6% dropout (5/83). Usually, a sample size for a proposed trial is increased by 20% to account for dropouts, so a 6% dropout rate was considered acceptable by the authors in this trial. Although the authors speculate that the dropout rate in a larger trial would be higher, it would be prudent to expand on a required sample size by 20% and account for this by extending the recruitment period. The number of eligible participants recruited was 81, being 75% of the 111 patients first presenting for screening and a recruitment rate of around 14 participants per month.

9

10



Table 5. Comparison of OTE Data of the 3 Treatment Groups Friedman P ≤ .01 Minimum required difference of mean rank: 1.7572 Summary Statistics

Mean Rank

Median

SE

95% CI

Group 1 Section A 4th visit Group 1 Section B 4th visit Group 2 Section A 4th visit Group 2 Section B 4th visit Group 3 Section A 4th visit Group 3 Section B 4th visit Group 1 Section A 1wFU Group 1 Section B 1wFU Group 2 Section A 1wFU Group 2 Section B 1wFU Group 3 Section A 1wFU Group 3 Section B 1wFU

4.6 7.5 4.3 6.3 5.4 8.4 6.0 8.1 6.1 7.3 6.0 8.2

4 5 4 5 4.5 5 4 5 4.5 5 4.5 5.5

0.3 0.3 0.3 0.3 0.5 0.3 0.4 0.2 0.4 0.3 0.5 0.3

3-5 4-6 3-5 4-5 3-5 5-6 4-5 5-6 4-5 4-6 3-5 5-6

Between-Group Comparison Mean Rank Difference Group 1–Group 2 Section A 4th visit 0.3 1wFU − 0.1 Group 1–Group 2 Section B 4th visit 1.2 1wFU 0.8 Group 2–Group 3 Section A 4th visit −1.1 1wFU 0.1 Group 2–Group 3 Section B 4th visit − 2.1 1wFU − 0.9 Group 1–Group 3 Section A 4th visit −0.8 1wFU −0.1 Group 1–Group 3 Section B 4th visit 0 1wFU −0.1

SE

P

95% CI

0.4 0.6

1.0 1.0

− 2 to 2 − 2 to 2

0.5 0.4

1.0 1.0

− 1 to 2 − 1 to 2

0.4 0.5

1.0 1.0

− 2 to 2 − 2 to 2

0.3 0.4

b .05 1.0000

− 2 to 1 − 2 to 1

0.7 0.7

1.0 1.0

− 3 to 3 − 2 to 3

0.5 0.5

1.0 1.0

− 2 to 2 − 2 to 2

P values less than 0.05 are statistically significant. 1wkFU, 1-week follow-up; CI, confidence interval; SE, standard error.

Acceptability of Interventions. No complaints or feedback from participants indicated a reluctance or objection to treatment provided. Participants were informed that they had the right to withdraw from the study at any time, without a need to provide a reason, and without prejudice. No adverse events or complications were reported (such as persistent, severe nonresolving pain, stiffness, or disability), and no patients were known to have left the trial because of side effects, but rather appeared to drop out by choice. Although there were isolated cases of transient, benign symptoms of stiffness, or pain reported, these benign side effects subsequently resolved in all cases and were not reported as adverse effects. Sample Size Calculation for a Confirmatory Trial Post hoc average effect sizes and power were ≤ 0.17 and ≤ 0.24, for WOMAC, respectively (Table 9), implying

the potential for significant type II error in this pilot trial. Using the average effect size across all 4 domains (pain, disability, stiffness, and cumulative) of the WOMAC of 0.1, the sample size required for a confirmatory trial with an individual-randomization design would be 962 total, or 323/group over 3 groups (power ≥ 80%; ANCOVA statistical test for 3 groups and 3 covariates; 2-tailed test; α = .05) 48 . For a given effect size of 0.2 and 0.3 (as reported or calculated from similar trials), the total sample size required would be 244 (82/group) and 111 (37/group), respectively. 12,49,50 If the recommendations proposed later in this article are implemented to improve on trial design, then an effect size of 0.2 would be reasonable. Therefore, a recommended sample size for an individual-randomized trial would be ≥ 82/group (total ≥ 244; Table 9). However, this sample size does not fully account for sample differences (cluster effects) if there are ≥ 2 study sites. Although individual-randomization is possible at each study site, as in this pilot trial, the variance between sites needs to be considered when estimating a sample size for a confirmatory trial, as with a cluster-randomized design. 51,52 The sample size requirement for a multisite or clusterrandomized trial would need to be inflated by a variance inflation factor (VIF) to account for the loss of power due to the clustering effect. Determining summaries of sample size estimation for cluster-randomized trials has been well described in the literature. 52,53 The VIF for a confirmatory trial, based on the WOMAC data of this trial, is 45.5 (rounded to 46), where VIF = 1 + (n − 1)p, where n = sample size; p = intraclass correlation coefficient (0.55). 54,55 Therefore, the required sample size per group for a multisite or clusterrandomized trial would be 82 + 46 + 26 = 154/group (sample estimate for an individually-randomized trial + VIF + dropout), or a total of 462 for 3 treatment groups. With sample sizes of this number it would be important to have a numerous clinician assessors, treating clinicians and research assistants to be able to manage the sample and the data collection, and careful consideration on this matter is required for future trials. The fundamental costing for a confirmatory trial, at ≈ US$240.00 per participant, would be in the region of ≈ US$110 000.00)—this excludes patient compensation costs.

Secondary Outcome: Clinical Outcomes Within-group analyses (Tables 2, 6, 7, and 8) demonstrated statistically significant (P ≤ .004) and clinically meaningful changes in scores for the WOMAC from baseline to week 5 across all 3 groups and all 3 domains (pain, stiffness, and disability).There were no statistically significant changes in scores for knee flexion/extension (P ≥ .08; Tables 3 and 4). Within-group post hoc power across all groups was ≥ 0.94. Treatment effectiveness was not explored further in this pilot trial, and the study did not



Potential Patients Responding to Advertising/word of mouth N = 111

Potential patients excluded after interview/screening n = 20

Patients who underwent assessment after being screened N = 91 Potential patients excluded after assessment n=8 Eligible patients that were randomly allocated to a treatment group N = 83 Group 1 MMT only n = 27

Drop outs n=1

Number of participants from whom data was available n = 26

Group 2 Exercise only n = 28

Drop outs n=2

Group 3 M MT + Exercise n = 28

Drop outs n=2



Fig 2. Flow diagram of participant recruitment and retention (CONSORT diagram). MMT, manual and manipulative therapy.

include a control group (no-treatment or placebo treatment). Figure 2 (WOMAC), Figure 3 (knee flexion and extension ROM), and Figure 4 (the OTE) bar charts give an overview of change in scores (Figs 2, 3, and 4). Between-group analysis of week 5 (final consultation) data did not reveal any significant differences in scores (P ≥ .46) across any of the treatment groups for any of the outcome measures (Table 9). The mean difference in scores between groups at week 5 was ≤80/2400 points (≤ 3%) for WOMAC and ≤ 4° (≤ 3%) for knee flexion/extension. However, the overall percent change in the WOMAC in group 1 or MMT alone (50%) in group 2 (37%) and group 3 combined MMT plus rehabilitation (51%) with similar percentage-like scores for the domains of pain, stiffness, and disability are notable and suggest at minimum that further research is justified (Tables 6, 7, 8, and 10). The median OTE score at week 5 across the groups was ≥ 5/7 (≥ 72%; range, 72%-79%), with the percentage score for groups 1, 2, and 3 being 79%, 75%, and 72%, respectively. This shows that participants in all 3 groups experienced relatively high levels of self-perceived

patient satisfaction and improvement, with no group demonstrating significantly higher scores over another (Table 5). Post hoc effect sizes and power were ≤ 0.17 and ≤ 0.24, respectively (Table 5). Although the patient satisfaction scale used in this study (OTE) cannot be directly compared (for short-term outcomes only), it is interesting to look at the long-term 1-year follow-up patient satisfaction scales in 2 other knee OA MMT (with and without exercise and/or usual, or medical practitioner, care) clinical trials. 22,56 The first delivered 8 treatments for 4 weeks and the next more recent study 9 total treatments for about 2 months. 22,56 Patients were trained to and continued home exercise for knee OA after release. 22,56 The first study listed patient satisfaction at 1 year using a Likert scale, with the MMT + exercise group reporting 52% “very much satisfied” compared with the exercise-only group reporting 25% “very much satisfied.” 22 Another recently published Knee OA study using a 0-10 scale (10 worst) demonstrated overall pain intensity decreased by − 1.1 points (for MMT + usual care) on a 0–10 scale vs exercise (+ usual care) of − 0.45 and combined

11

12



Table 6. Comparison of Pain Data (WOMAC) of the 3 Treatment Groups Sphericity ε = 0.51 P ≤ .001


Mean

SE

95% CI


227.3 205.8 216.8 150.4 127.7 133.5 102.3 129.2 97.7

17.2 15.2 17.0 21.3 12.2 14.5 17.1 12.9 16.4

191.8-262.8 174.5-237.1 181.9-251.8 106.5-194.2 102.6-152.8 103.7-163.3 67.0-137.5 102.5-155.8 64.0-131.5

Within-group comparison Group 1 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Group 2 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Group 3 Baseline–4th visit 4th visit–1wFU Baseline–1wFU Between-group comparison Group 1–Group 2 Baseline 4th visit 1wFU Group 2–Group 3 Baseline 4th visit 1wFU Group 1–Group 2 Baseline 4th visit 1wFU

Mean Difference

SE

Pa

95% CI a

76.9 48.1 125.0

17.8 11.2 17.6

b .0077 b .0085 b .0001

12.9 to 140.8 7.8 to 88.5 61.6 to 188.4

78.1 − 1.4 76.7

15.1 13.8 17.0

b .0009 1.0000 b .0048

23.8 to 132.3 − 50.9 to 48.0 15.5 to 137.8

83.4 35.7 119.1

15.8 9.3 15.5

b .0007 b .0276 b .0001

26.6 to 140.2 2.2 to 69.3 63.6 to 174.7

21.5 22.7 − 26.9

23.1 24.0 18.6

1.0 1.0 1.0

− 61.7 to 104.7 − 63.7 to 109.0 − 93.7 to 40.0

− 11.0 − 5.7 31.4

25.4 20.6 21.5

1.0 1.0 1.0

− 102.4 to 80.3 − 79.9 to 68.0 − 45.9 to 108.7

10.4 16.9 4.5

24.6 26.0 25.0

1.0 1.0 1.0

− 77.9 to 98.7 − 76.4 to 110.3 − 85.4 to 94.4


MMT + exercise (+ usual care) of − 0.84. 56 (Tables 5, 6, 7, 8, and 10; Fig 5).

DISCUSSION Primary Outcome: Evaluation of the Research Integrity and Methodological Rigor Patient Recruitment, Enrollment, and Allocation. This trial had a successful recruitment rate that could be improved in a confirmatory trial by having multiple study sites, 2 or more clinicians conducting the screening and examination at each site, and cooperation from local medical, physiotherapy, and orthopedic clinics to advertise or refer patients. This

may well increase recruitments rates to over double those rates in this pilot trial. The authors recommend 3 or more clinic sites, with referral from local medical and orthopedic clinics, to achieve the estimated sample size for a fully powered confirmatory study. Referral may be encouraged by providing continuing medical education workshops for physicians with a view to facilitate patient referral and lectures for family physicians, where presentations on research may be offered and may enhance referrals and understanding of these research activities. The screening criteria were sufficiently broad to allow the assessment of patients with knee pain, yet were likely to focus on those who probably had knee OA. Screening of



Table 7. Comparison of Stiffness Data (WOMAC) of the 3 Treatment Groups Sphericity ε = 0.51 P ≤ .001


Mean

SE

95% CI

Group 1 baseline Group 2 baseline Group 3 baseline Group 1 4th visit Group2 4th visit Group 3 4th visit Group 1 1wFU Group 2 1wFU Group 3 1wFU

96.5 96.5 119.3 70.5 63.0 65.9 46.6 48.5 41.1

9.6 9.6 8.7 9.3 6.2 8.0 6.4 7.1 7.7

76.9-116.2 76.9-116.2 101.4-137.3 51.3-89.6 50.1-75.8 49.5-82.3 33.5-59.7 33.9-63.2 25.4-56.9


SE

Pa

95% CI a

26.1 23.8 49.9

9.1 7.1 9.6

.3027 .0909 .0008

− 6.7 to 58.8 − 1.7 to 49.4 15.4 to 84.5

25.5 14.4 47.4

10.3 4.6 9.3

.7297 .1560 .0011

− 11.5 to 62.4 − 2.1 to 31.0 13.9 to 80.8

53.5 24.8 78.2

8.1 6.2 8.4

b.0001 .0188 b.0001

24.5 to 82.5 2.4 to 47.1 48.2 to 108.3

0.6 7.5 − 1.9

11.1 10.1 8.3

1.0 1.0 1.0

− 39.2 to 40.5 − 28.7 to 43.7 − 31.9 to 28.0

− 23.4 − 2.9 7.4

11.9 8.6 11.1

1.0 1.0 1.0

− 66.0 to 19.2 − 34.0 to 28.2 − 32.6 to 47.5

− 22.8 4.6 5.5

13.9 13.6 10.9

1.0 1.0 1.0

− 72.6 to 27.0 − 44.2 to 53.4 − 33.7 to 44.7


patients by the clinical lead at each study site and the consequent scheduling of appointments for further assessment of potential participants were logistically successful. The subsequent assessment and examinations of potential participants by the same clinical lead allowed for continuity of assessment and uniformity of participant selection. The inclusion criteria were thorough and facilitated the selection of participants with genuine knee OA. The diagnosis was reached after assessment and review of radiographs and the report by the primary treating clinician. However, the validity of the diagnosis would be bolstered by using 2 assessors to confirm the diagnosis, and this is recommended

for future trial. The descriptive statistics indicates that most participants in this trial were grade 1 knee OA, some with grade 0, and very few with other grades. A recommendation is that all grades of knee OA be represented (grades 1-4), excluding grade 0, and by having a large enough sample size to allow the representation of these grades. This approach would allow subgroup analysis based on knee OA grades, to explore which grade(s) the intervention(s) may be of benefit. Worth considering for a confirmatory trial are broader diagnostic criteria developed for knee OA through systematic review and Delphi panel analysis to take into account a variety of presentations: symptoms include

13

14



Table 8. Comparison of Disability Data (WOMAC) of the 3 Treatment Groups Sphericity ε = 0.37 P ≤ .001


Mean

SE

95% CI

Group 1 baseline Group 2 baseline Group 3 baseline Group 1 4th visit Group2 4th visit Group 3 4th visit Group 1 1wFU Group 2 1wFU Group 3 1wFU

759.0 731.7 411.7 573.8 521.1 514.4 389.7 411.7 378.9

47.6 50.8 52.0 56.4 39.5 61.6 49.4 52.0 62.0

660.9-857.2 627.0-836.3 304.5-518.9 457.7-689.9 439.8-602.4 387.6-641.2 288.0-491.4 304.5-518.9 251.1-506.6

SE

Pa

95% CI a

185.2 184.1 369.3

39.9 35.0 42.0

.0034 .0007 b.0001

41.8 to 328.6 58.4 to 309.8 218.5 to 520.2

210.5 109.4 320.0

43.0 32.2 60.0

.0017 .0820 .0006

56.2 to 364.8 − 6.3 to 225.2 104.2 to 535.7

317.5 135.5 453.0

53.1 31.3 63.2

.0001 .0076 b.0001

126.5 to 508.5 23.0 to 248.0 225.6 to 680.4

27.4 52.7 − 22.0

57.2 63.3 61.5

1.0000 1.0000 1.0000

− 178.3 to 233.1 − 174.9 to 280.3 − 243.0 to 199.0

− 100.2 6.7 32.8

81.9 68.9 69.1

1.0000 1.0000 1.0000

− 394.8 to 194.3 − 241.0 to 254.5 − 215.6 to 281.2

− 72.8 59.4 10.8

84.2 93.5 89.0

1.0000 1.0000 1.0000

− 375.5 to 229.8 − 277.0 to 395.7 − 309.4 to 331.0



persistent knee pain, limited knee stiffness (b 30 minutes), and reduced function, and signs include crepitus, restricted movement, and bony enlargement (the more symptoms and signs the more likely the diagnosis of knee OA (all 6 denoting a 99% confidence interval [CI]). 8,57 The patient information sheet provided to eligible participants, as approved by the respective ethics committee and internal review board, was deemed satisfactory. The format of the information sheet was similar to that of those found in clinical trials. It contained all the necessary information to enable eligible participants to make an informed choice about volunteering for the study and

included a cooling off period of up to 3 days to afford each eligible patient to reflect on their decision and change their mind if they so wished. The 3-day cooling off period also acted as a brief quantification period to ensure the patient's knee pain was not transient but a persistent knee pain in line with the diagnosis of knee OA. Data Collection and Patient Retention. In a trial testing manual therapy or exercise, neither participant nor therapist/ clinician blinding is practical or possible, respectively. However, to reduce the chance of treatment contamination and clinician bias, a blind assessor collected the data from each participant at the appropriate time-points in this pilot



Table 9. Comparison of the Data at Week 5 for All Outcome Measures Group 1 a

Group 2 b

Group 3 c

Homogeneity Of Analysis Post hoc Power (ANCOVA d) Regressions

Outcome Measure

Mean (SEM) [95% CI]

Mean (SEM) [95% CI]

Mean (SEM) [95% CI]

P Value (r)

P value

Power (Effect Size)

WOMAC Cumulative WOMAC Pain WOMAC Stiffness WOMAC Disability Knee flexion Knee extension OTE Part A (median) OTE Part B (median)

553.0 (80.7) [387-719] 102.3 (17.1) [67-138] 46.6 (6.4) [34-60] 389.7 (49.4) [288-491] 106.1 (3.6) [99-114] 6.2 (0.9) [4-8] 4 (0.4) [4-5] 5 (0.2) [5-6]

633.5 (63.4) [503-764] 129.2 (12.9) [103-156] 48.5 (7.1) [34-63] 411.7 (52.0) [305-519] 105.5 (3.5) [98-113] 5.2 (0.8) [4-7] 4.5 (0.4) [4-5] 5 (0.3) [4-6]

570.3 (86.7) [392-749] 97.7 (16.4) [64-132] 41.1 (7.7) [25-57] 378.9 (62.0) [251-507] 109.7 (3.0) [104-116] 9.0 (4.2) [0.2-18] 4.5 (0.5) [3-5] 5.5 (0.3) [5-6]

.78 (− 0.01) .34 (− 0.04) .75 (− 0.07) .95 (− 0.04) .67 (0.02) .61 (0.04) .97 (− 0.01) .58 (− 0.02)

.87 .64 .97 .57 .46 .87 .93 .80

0.10 (0.09) 0.24 (0.17) 0.10 (0.09) 0.06 (0.05) 0.16 (0.13) 0.22 (0.16) 0.06 (0.04) 0.11 (0.10)

ANCOVA, analysis of covariance; CI, confidence interval; OTE, McMaster Overall Therapy Effectiveness; WOMAC, a 24-question survey used to measure pain, joint stiffness, and function (disability). a Manipulative therapy alone. b Exercise prescription alone. c Manipulative therapy plus exercise. d With mean scores adjusted for the covariates of age, body mass index, and duration of symptoms.

trial. The data were collected by an independent research assistant who was not involved in the study otherwise. As with other staff, the research assistant was provided training on the outcome measures and how to administer them. No problems were encountered in this trial in administering these outcome measures post training. Logistically, it was more practical to book appointments for participants on specific days of the week so to coordinate with availability of the clinician(s) and research assistant, for treatment and data collection, respectively. The outcome measures, being the WOMAC, OTE, and knee ROMs, were straightforward to administer by the research assistant at both study sites. Clearly, the WOMAC is the preferred clinical outcome measure, being well-known, valid, reliable, and easy to administer. The WOMAC is recommended as a key clinical outcome measure for future trials because of its utility, ease of use, and the ability for researchers to amalgamate data from various trials using the WOMAC for meta-analysis, for example, systematic reviews and in clinical guideline development. In addition, the WOMAC collects data across 4 domains (pain, stiffness, disability, and cumulative), allowing for individual domain analysis and amenable to being used as a single outcome measure if a study is constrained by time/ practical issues. The utility and appropriateness of both the OTE and knee ROMs are uncertain. Manual goniometric measurement in the clinical setting is prone to measurement error, and as an outcome measure, the relationship between knee ROM and functional disability is questionable, particularly with only relatively small reductions in ROM in lower grades of knee OA. The OTE is encountered sporadically in the health care literature and may well be a useful participant self-report of the effectiveness of care, but it has been suggested that outcome measures for OA trials should include 4 key domains: pain, physical function, patient global assessment,

and, for studies of 1 year or longer, joint imaging. 58 In a systematic review, the most common outcome measures encountered in published trials for knee OA are the WOMAC, visual analog scale, the Arthritis Impact Measurement Scales, and the Knee injury and Osteoarthritis Outcome Score (KOOS). 12 In addition, Health-related quality of life assessment is receiving increased attention as an outcome measure in OA. 59 The RAND-36 Measure of Health-Related Quality of Life, as a well-known and tested questionnaire, could offer meaningful data in this regard. The overall impression of the authors is that the primary outcome measures for a knee OA trial should comprise the widely used tools that cover the domains now commonly stipulated for knee OA to include quality of life measures, amenable to meta-analysis in future systematic review. Therefore, the authors recommend the WOMAC and the RAND-36 Measure of Health-Related Quality of Life plus 1 other, either the Arthritis Impact Measurement Scales 2 or the KOOS, with the exclusion of the OTE and knee ROM. 60–63 Patient tracking occurred through appointment scheduling and follow-up. Clinic managers and clinical leads for this trial did not report any major issues with patient tracking, record keeping, or follow-up. However, for a larger trial, the following should be considered: a. A clear indication in the participant consent form that the agreement to participate implies that an appointment and instruction would be followed to the best of the participant's ability. b. The study should have a written flowchart indicating patient scheduling and time-points of follow-up, with a view to ensure appropriate timing of appointments, participant tracking, and follow-up. c. Ensure clinical and administrative record keeping in line with required standards of practice.

15

16



Fig 3. Changes in the WOMAC for cumulative, pain, disability, and stiffness scores from baseline, to the fourth visit and for the final or fifth visit, or 1-week follow-up for group 1 (manual therapy alone), group 2 (exercise only), and group 3 (manual therapy and exercise).

Fig 4. Changes in knee ROM in flexion and extension from baseline, to the fourth visit and for the final or fifth visit, or 1-week follow-up for group 1 (manual therapy alone), group 2 (exercise only), and group 3 (manual therapy and exercise).

To ensure participant compliance, scheduling of appointments occurred early in the treatment period, usually at baseline. Participants were followed up telephonically to remind them of appointments, or to reschedule, as necessary. Participants were reminded verbally, by the treating clinician at the respective appointments, of the treatment and the importance of compliance. For those participants undertaking home exercise, they were required to maintain a diary to encourage engagement with the exercise routine and compliance. Adherence to home exercise is often difficult to maintain, and lack of adherence limits the long-term effectiveness of exercise. Strategies to improve compliance include the following 64:

a. Education about exercise and the benefits of exercise b. Use of a graded progressive exercise or physical activity program that ensures that pain and discomfort are not excessive during or after exercise c. Initiate exercise under instruction and supervise exercise sessions if possible d. Supplement face to face instruction with other materials; e. Increase self-efficacy by incorporating goal setting, positive reinforcement, use of an exercise contract, and self-monitoring using a diary 63,64 Based on these recommendations, a trial requires a clear written protocol and explanation about knee and knee OA



Table 10. Change in Score From Baseline to Week 5 for WOMAC Group 1 (n b= 26)

Group 2 (n = 26)

d

Group 3 (n = 26)

Outcome Measure

Change in % Change in % Change in % c (Mean Difference) (Mean Difference) (Mean Difference)

WOMAC Cumulative WOMAC Pain WOMAC Stiffness WOMAC Disability

50 (560)

37 (364)

51 (595)

55 (125)

37 (77)

55 (119)

52 (50)

49 (47)

66 (78)

47 (369)

44 (320)

54 (453)

WOMAC, a 24-question survey used to measure pain, joint stiffness, and function (disability). a Manipulative therapy alone. b Sample size. c Change in scores from baseline to week 5 presented as % change and mean differences score. d Exercise prescription alone. e Manipulative therapy plus exercise.

so that the participants understand that the increased activity of the exercises to treat their knee OA are similar to simple activities such as increased walking and will not further hurt their knee if carried out cautiously as taught—and how to modify, reduce, or expand on these exercises (thus, a diary is required) based on participant response. 63,64

Overal Therapy Effectiveness

9

e

8

Score at Week 5

a

7 6 5 4 3 2 1 0 Grp1 PartA

Grp2 PartA

Grp3 PartA

Grp1 PartB

Grp2 PartB

Grp3 PartB

Groups

Fig 5. Scores at week 5 (final consultation, after the 4-week treatment period) for both Part A and Part B of the OTE index.

Design Alternative. Because of the complex nature of treatment and care delivery in community-based care, a design alternative is the cluster-randomized approach. This methodology is thoroughly discussed in the literature and is thought to be very useful for the testing of complex interventions. 51,55,59 However, this approach usually requires sample size inflation, as demonstrated in the sample size calculation previously, adding to the project cost and duration. 13 Clinical Outcomes

Acceptability of Interventions. No complaints or feedback from participants indicated a reluctance or objection to treatment provided. As indicated previously, benign side effects were transient and no adverse events were reported. Patients also were required to read instructions (Appendix A available online), which details possible adverse effects from all of the 3 treatments offered, and it was made clear that the most common side effect for any of the treatments was temporary stiffness and soreness. They were also informed that in Appendix A, per each protocol, they were to report to the treating chiropractor immediately if they felt any side effects, or injury from treatment, which then would have been dealt with as a serious adverse reaction. There were only mild, transient side effects of soreness and stiffness noted (in all 3 groups), but in retrospect, although data around adverse events were to be noted in patient files with a view to identify dropouts, the minimal side effects seen were not comprehensively reported (such as a percentage or mean) and, therefore, not fully explored in the data analysis. This oversight is described later as a limitation of the study. Acceptability of interventions was considered through synthesis of (a) the number of dropouts and the reasons for such, (b) frequency and type of adverse events, and (c) feedback from the participants, treating clinicians, and patient records related to treatment. There was no formal outcome measure or survey of acceptability, but rather a cumulative measure. A more formal feedback system, as part of reporting adverse events, should be considered in future trials.

The impression generated from the results of statistical analysis suggest that all 3 treatment approaches had a beneficial effect in the short term (4 weeks and 1-week follow-up) based on WOMAC scores (Table 10). The results of this trial correlate with those reported by another study that demonstrated significant short-term benefits of manual therapy and supervised exercise, also based on WOMAC scores. 30 Therefore, a reasonable inference is that any of the treatments tested in this study could improve symptoms over a 4-week treatment period. Broadly, the outcomes of this trial are also consistent with the positive WOMAC outcomes reported with the use of MMT for knee OA in other published (Table 10) trials. 22,23,56 The greatest change in WOMAC scores in this trial was in groups 1 and 3 (Table 2), where treatment was provided directly by a clinician (on 6 treatment occasions), as opposed to group 2, where direct supervision was provided on only 3 occasions. There was also a “hands-on” manual therapy component of treatment in groups 1 and 3, which was not offered in group 2. So, there may be a potential dose-response association related to the frequency of direct treatment or supervision. Although this association was not statistically explored in this study, it does bear similarity to the conclusions drawn by others where the greater the level/frequency of supervision, the better the clinical outcomes, particularly when exercise prescription is an intervention. 12,22,30 Plausibly, the higher levels of direct supervision or hands-on treatment, as offered in groups 1 and 3, resulted in better participant

17

18


compliance, particularly with self-directed home exercise, and therefore improved clinical outcome. The matter of attention bias will always be present in studies that test physical therapies where 1 treatment may be offered more frequently, for longer or have a more intense treatment “dose.” It is possible therefore that if the participants in group 2 (rehabilitation alone) received more direct supervision, the outcomes may have been better. Arguably, the rehabilitation group (group 2) should receive the same dose and level of supervision as the other groups in a confirmatory trial. Also, rehabilitation should perhaps be more comprehensive and consensus led (by a panel of clinical experts) with a view to create an intervention that is of equivalent in dose to other intervention groups being tested. Manual and manipulative therapy alone appeared to have a positive effect on outcomes, based on WOMAC scores (Table 2), possibly even better than the rehabilitation program as delivered in this pilot trial. Although there were no statistically significant differences between the groups (Tables 2 and 10), the positive effect of MMT corroborates with those reported in a systematic review of treatments for knee OA, where the authors recommend manual mobilization in addition to supervised exercise to gain better pain relief (over exercise alone). 12 Understandably then, one could speculate that the “right” combination of MMT plus exercise delivered in a better or best way may yield better clinical outcomes than either MMT or exercise alone. A potential confounding factor is the overlap of manual therapy in both the rehabilitation and MMT groups, specifically soft tissue therapy and passive stretching, which are arguably also manual therapies. With these 2 treatments included in the rehabilitation group, it is difficult to explain which of the treatment components in that group may have had an effect. In future trials, there should be clearer treatment “boundaries” for each group. Recent high-quality studies found significant and clinically important benefits with exercise and manual therapy for knee OA, but treatment was provided over a longer period in these trials, as opposed to the 4 weeks in this trial. 22,56 The implication is that conservative treatments, such as MMT and exercise, need to be offered at regular intervals over an extended period to obtain the best possible outcomes. For example, the first study provided treatment for 4 weeks (8 treatments) and the second for 2½ months (9 treatments), respectively, with long-term follow-up. However, based on the trends in outcomes in this trial, the authors speculate that the addition of MMT to a rehabilitation or exercise program (with a recent recommendation that initially, manual therapy and exercise be delivered—in a clinic—on opposite days) may indeed be more beneficial than either MMT or rehabilitation alone, but possibly only clinically demonstrable over a longer treatment period rather than in the


short term. 22,56 This hypothesis should be tested in a future fully powered confirmatory trial because definitive inferences cannot be drawn from the results of this external pilot trial.

Limitations The limitation of this trial is that there is no control group featuring a “wait-and-see” approach or a “minimumtreatment” group, so conclusions regarding the effectiveness of treatment in each group cannot be drawn. Furthermore, the natural history of knee OA, which is often characterized by fluctuating signs and symptoms, alongside the potential effect of the Hawthorne phenomenon, may account for some of the improvement seen in each treatment group. Also, as a pilot trial, this study also does not permit definitive conclusions regarding clinical outcomes due to its inherent low power. Adverse events and treatment side effects were not comprehensively reported in this trial and (only mild side effects were noted, but neither were they given as a percentage per group or descriptively described), therefore, not fully explored in the data analysis. Treatment safety and adverse events are important components to a pilot trial because these data inform the design and treatment protocols for a definitive trial. In future trials, there should be a protocol of noting, reporting, and analysis of adverse events and side effects per group either (at minimum) descriptively or, better, statistically in future trials. Practical and financial constraints limited this trial to the outcome measures described previously and excluded the initially planned neuromuscular outcomes such as balance and coordination. Insights into these proprioceptive outcomes would be valuable in future trials related to the potential neuromuscular component of the disorder, because balance and coordination are directly related to falls—falls being a major concern causing further injury and disability. Also, this trial did not include performance-based (such as the timed get up and go test, the 30-second sit to stand test, or the 40-m selfpaced walk test) or other patient-reported outcomes previously noted above, the former offering insight into outcomes related to patients that regularly participate in sport and, the latter, providing deeper insight into patient-perceived effect of treatment. 30,56 Both these types of outcomes warrant consideration in future trials. In retrospect, the overlap of manual therapy in both the rehabilitation and MMT groups (not boney manipulation and mobilization) but specifically, soft tissue therapy and passive stretching, considered by some (also) manual therapies, may have produced better outcomes in the rehabilitation group than the components of exercise and self-stretching alone would have accomplished. 30,56 As the authors of a recently completed trial put it in describing the use of exercise alone or exercise physiotherapy, they were adamant in stating that those who were to receive the “exercise therapy protocol” were strictly


(otherwise) not allowed therapist-applied manual forces. 56 In future trials, clear treatment boundaries should be defined for each group. Data on precisely how many clinicians provided treatment or the frequency of treatment were not available and were difficult to retrieve post hoc. Not all patients were treated by the same clinicians. Treating clinician demographics/data should be reported in future trials. A further limitation of this pilot trial was that participant recruitment was arguably slow, with essentially only 1 clinician performing the assessment of patients and 1 to 2 clinicians delivering the treatment. A larger confirmatory trial would benefit from a greater number of clinicians performing assessment and treatment, although this would potentially increase the study's costs. As in this pilot trial, there are inherent difficulties and confounding factors in testing of complex nonpharmacologic interventions, for example, characteristics of treatment such as timing and frequency of treatment and patient compliance.


WOMAC is an obvious primary clinical outcome measure for a confirmatory trial, but also an HRQoL measure (eg, RAND-36) and the KOOS (or AMS2). f. In addition, balance and coordination measures should be considered with a view to measure neuromuscular and proprioceptive functional changes with treatment, alongside patient-reported outcomes. g. Sufficiently large sample size to achieve more than 80% power, which is estimated at ≥ 154/group for a multicenter or cluster-randomized trial for 3 treatment groups. h. The short-, medium-, and long-term outcomes should be analyzed with data collection points at baseline, 2 weeks, 1 month, 3 months, 6 months, and 1 year—(a) these timescales are often reported in the literature for trials, and (b) the timescales enable analysis of the data at short-, medium, and long-term end-points and, possibly, allow for subgroup analysis and treatment dose-response. i. Cost analysis and dose-response assessment would be very useful adjuncts to data collection and analysis.

Recommendations for Future Studies Development of treatment approaches for each intervention based on the best-available research evidence, including statements on the timing and frequency of treatment, is recommended. Each group should have distinct treatment boundaries so that the treatments are not too similar. For example, the rehabilitation group should only include exercise and active (self) stretching, but no soft tissue or passive stretching. Broader inclusion criteria include grades 1 to 4 of knee OA, excluding grade 0, allowing for subgroup data analysis. Recommendations include: a. Have 2 or more assessors, arguably at least 3 per site, to confirm the diagnosis for each eligible participant as this would improve on diagnosis/selection validity. b. There should be a number of treating clinicians at each site, at least 3 to 4 per site if a sample of more than 200 per site is expected, all of whom having had training in the study's assessment and treatment protocols, with a view to standardize care. c. Treating clinician demographics/data, in terms of the number of clinicians and the frequency of treatment, should be reported in a confirmatory trial. d. There should be a protocol to report adverse events. Although not essential, inclusion of a control group to explore treatment effectiveness, that is, a wait-andsee or a minimum treatment group, is strongly recommended. e. Selection of the outcome measures that measure various domains, such as pain, disability, and so forth. The best-available outcomes are those that are valid, reliable, and believed to be the most appropriate by the researchers to achieve the research objectives.

CONCLUSIONS This pilot trial suggests that a confirmatory trial is feasible. There were significant changes in scores from baseline to week 5 across all groups, suggesting that all 3 treatment approaches may be of benefit to patients with mild-to-moderate knee OA, justifying a confirmatory trial to compare these interventions.

FUNDING SOURCES AND POTENTIAL CONFLICTS OF INTEREST We thank the NCMIC Foundation for the support in the development of this manuscript. No conflicts of interest were reported for this study.

CONTRIBUTORSHIP INFORMATION Concept development (provided idea for the research): L.D., G.F.P-S., J.W.B., C.K., T.K.C. Design (planned the methods to generate the results): L.D., G.F.P-S., J.W.B., C.K., T.K.C. Supervision (provided oversight, responsible for organization and implementation, writing of the manuscript): L.D., J.W.B., C.K. Data collection/processing (responsible for experiments, patient management, organization, or reporting data): L.D., J.W.B., C.K., T.K.C. Analysis/interpretation (responsible for statistical analysis, evaluation, and presentation of the results): L.D., G.F.P-S., J.W.B., C.K., T.K.C., G.G., D.B., V.T. Literature search (performed the literature search): L.D., G.F.P-S., J.W.B., C.K., T.K.C.

19

20


Writing (responsible for writing a substantive part of the manuscript): L.D., G.F.P-S., J.W.B., C.K. Critical review (revised manuscript for intellectual content, this does not relate to spelling and grammar checking): L.D., G.F.P-S., J.W.B., C.K., T.K.C., G.G., D.B., V.T. Rehabilitation consultation and overview and radiologic interpretation: D.B., V.T.


11. 12.

13.

Practical Applications • Outcomes from this study suggest that a confirmatory trial is feasible. • Significant changes in scores from baseline to week 5 across all groups suggest that all 3 treatment approaches may be of benefit to patients with symptomatic mild-to-moderate knee OA.

14. 15.

16.

17.

REFERENCES

18. 19.

1. Woolf A, Pfleger B. The burden of major musculoskeletal conditions. Bull World Health Organ 2003;81:646-56. 2. Felson DT, Zhang Y. An update on the epidemiology of knee and hip osteoarthritis with a view to prevention. Arthritis Rheum 1998;41:1343-55. 3. Australian Institute of Health and Welfare. Arthritis and musculoskeletal conditions in Australia, 2005. Canberra: AIHW; 2005. 4. RACGP Osteoarthritis Working Group. Guideline for the non-surgical management of hip and knee osteoarthritis. Royal Australian College of General Practitioners. [cited 2012] Available fromhttp://www.nhmrc.gov.au/_file_snhmrc/ publications/attachments/cp117-hip-knee-osteoarthritis2009. 5. Jordan KM, Arden NK, Doherty M, et al. EULAR Recommendations 2003: an evidence based approach to the management of knee osteoarthritis: report of a Task Force of the Standing Committee for International Clinical Studies Including Therapeutic Trials (ESCISIT). Ann Rheum Dis 2003;62:1145-55. 6. Jamtvedt G, Dahm KT, Christie A, et al. Physical therapy interventions for patients with osteoarthritis of the knee: an overview of systematic reviews. Phys Ther 2008;88:123-36. 7. Bjordal JM, Johnson MI, Lopes-Martins RA, Bogen B, Chow R, Ljunggren AE. Short-term efficacy of physical interventions in osteoarthritic knee pain. A systematic review and meta-analysis of randomised placebo-controlled trials. BMC Musculoskelet Disord 2007;8:51. 8. Bennell KL, Hunter DJ, Hinman RS. Management of osteoarthritis of the knee. BMJ 2012;345:e4934. 9. Pisters MF, Veenhof C, van Meeteren NL, et al. Long-term effectiveness of exercise therapy in patients with osteoarthritis of the hip or knee: a systematic review. Arthritis Rheum 2007; 57:1245-53. 10. Bjordal JM, Klovning A, Ljunggren AE, Slordal L. Short-term efficacy of pharmacotherapeutic interventions in osteoarthritic

20. 21. 22.

23.

24. 25. 26.

27. 28. 29.

knee pain: a meta-analysis of randomised placebo-controlled trials. Eur J Pain 2007;11:125-38. Bronfort G, Haas M, Evans R, Leiniger B, Triano J. Effectiveness of manual therapies: the UK evidence report. Chiropr Osteopat 2010;18:1-112. Jansen MJ, Viechtbauer W, Lenssen AF, Hendriks EJ, de Bie RA. Strength training alone, exercise therapy alone, and exercise therapy with passive manual mobilization each reduce pain and disability in people with knee osteoarthritis: a systematic review. J Physiother 2011;57:11-20. Lancaster GA, Campbell MJ, Eldridge S, et al. Trials in primary care: statistical issues in the design, conduct and evaluation of complex interventions. Stat Methods Med Res 2010;19:349-77. Campbell M, Fitzpatrick R, Haines A, et al. Framework for design and evaluation of complex interventions to improve health. BMJ 2000;321:694-6. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M. Developing and evaluating complex interventions: the new Medical Research Council guidance. BMJ 2008;337:a1655. Arain M, Campbell MJ, Cooper CL, Lancaster GA. What is a pilot or feasibility study? A review of current practice and editorial policy. BMC Med Res Methodol 2010;10:1-7 http:// dx.doi.org/10.1186/1471-2288-1110-1167. Altman R. Criteria for classification of clinical osteoarthritis. J Rheumatol Suppl 1991;27:10-2. Kellgren JH, Lawrence JS. Radiological assessment of osteoarthrosis. Ann Rheum Dis 1957;16:494-502. Pollard H, Ward G, Hoskins W, Hardy K. The effect of a manual therapy knee protocol on osteoarthritic knee pain: a randomised controlled trial. J Can Chiropr Assoc 2008;52: 229-42. Altman RD, Bloch DA, Bole GG, et al. Development of clinical criteria for osteoarthritis. J Rheumatol 1987;14 Spec No:3–6. Saghaei M. Random allocation software for parallel group randomized trials. BMC Med Res Methodol 2004;4:26. Deyle GD, Allison SC, Matekel RL, et al. Physical therapy treatment effectiveness for osteoarthritis of the knee: a randomized comparison of supervised clinical exercise and manual therapy procedures versus a home exercise program. Phys Ther 2005;85:1301-17. Fish D, Kretzmann H, Brantingham JW, Globe G, Korporaal C, Moen J. A randomized clinical trial to determine the effect of combining a topical capsaicin cream and knee joint mobilization in the treatment of osteoarthritis of the knee. J Am Chiropr Assoc 2008;45:8-23. Schafer RC, Faye LJ. Motion palpation and chiropractic technique:principles of dynamic chiropractic. Huntington Beach, CA: The Motion Palpation Institute; 1990. Peterson DH, Bergmann TF. Chiropractic technique: principles and procedures. 2nd ed. St. Louis, Missouri: Mosby; 2002. Tucker M, Brantingham J, Myburgh C. The relative effectiveness of a non-steroidal anti-inflammatory medication (meloxicam) versus manipulation in the treatment of osteoarthritis of the knee. Eur J Chiropr 2003;50:163-84. Bergmann T, Peterson DH. Chiropractic principles and procedures. St Louis: Mosby; 201035-381. Triano JJ, Budgell B, Bagnulo A, et al. Review of methods used by chiropractors to determine the site for applying manipulation. Chiropr Man Therap 2013;2013:36. Maitland G. Peripheral manipulation. 3rd ed. London, UK: Butterworth_Heinemann; 1999:1-258.


30. Deyle G, Henderson N, Matekel R, Ryder M, Barber M, Allison S. Effectiveness of manual physical therapy and exercise in osteoarthritis of the knee. Ann Intern Med 2000; 132:173-80. 31. Brantingham JW, Bonnefin D, Perle SM, et al. Manipulative therapy for lower extremity conditions: update of a literature review. J Manipulative Physiol Ther 2012;35:127-66. 32. Brantingham JW, Smith Parkin, Cassa T, et al. Full kinetic chain manual and manipulative therapy plus exercise compared with targeted manual and manipulative therapy plus exercise for symptomatic osteoarthritis of the hip—a randomized controlled trial. Arch Phys Med Rehabil 2012;93: 259-67. 33. Bellamy B, Buchanon WW, Goldsmith CH, Campbell J, Stitt L. Validation of WOMAC: a health status instrument for measuring clinical important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip and knee. J Rheumatol 1988;15:1833-40. 34. Bellamy N. WOMAC Osteoarthritis Index—a user's guide IV. Queensland: CONROD, Royal Brisbane Hospital; 1995. 35. Magee D. Orthopedic Physical Assessment. 3rd ed. Philadelphia: W.B. Saunders Company; 1997520-1. 36. Norkin C, White J. Measurement of joint motion: a guide to goniometry. Philadelphia: F.A. Davis Company; 198575-112. 37. Norkin CC, Levangie PK. Joint structure and function. 3rd edition. Philadelphia: F.A. Davis Company; 2001. p. 327-42. 38. Piriyaprasarth P, Morris ME, Winter A, Bialocerkowski AE. The reliability of knee joint position testing using electrogoniometry. BMC Musculoskelet Disord 2008;9:6. 39. Piriyaprasarth P, Morris ME. Psychometric properties of measurement tools for quantifying knee joint position and movement: a systematic review. Knee 2007;14:2-8. 40. Cibere J, Bellamy N, Thorne A, et al. Reliability of the knee examination in osteoarthritis: effect of standardization. Arthritis Rheum 2004;50:458-68. 41. Juniper EF, Guyatt GH, Willan A, Griffith LE. Determining a minimal important change in a disease-specific Quality of Life Questionnaire. J Clin Epidemiol 1994;47:81-7. 42. O'Meara E, Lewis E, Granger C, et al. Patient perception of the effect of treatment with candesartan in heart failure. Results of the Candesartan in Heart failure: Assessment of Reduction in Mortality and morbidity (CHARM) program. Eur J Heart Fail 2005;7:650-6. 43. Chan L, Mulgaonkar S, Walker R, Arns W, Ambühl P, Schiavelli R. Patient-reported gastrointestinal symptom burden and health-related quality of life following conversion from mycophenolate mofetil to enteric-coated mycophenolate sodium. Transplantation 2006;81:1290-7. 44. Hertzog MA. Considerations in determining sample size for pilot studies. Res Nurs Health 2008;31:180-91. 45. Browne RH. On the use of a pilot sample for sample size determination. Stat Med 1995;14:1933-40. 46. Donders A, van der Heijden G, Stijnen T, Moons K. Review: a gentle introduction to imputation of missing values. J Clin Epidemiol 2006;59:1087-91. 47. Moons K, Donders R, Stijnen T, Harrell FEJ. Using the outcome for imputation of missing predictor values was preferred. J Clin Epidemiol 2006;59:1092-101.


48. Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 2007;39:175-91. 49. Jensen AP, Dworkin RH, Gammaitoni AR, Olaleye DO, Oleka N, Galer BS. Assessment of pain quality in chronic neuropathic and nociceptive pain clinical trials with the Neuropathic Pain Scale. J Pain 2005;6:98-106. 50. Jensen MP, Karoly P, O'Riordan EF, Bland F, Burns RS. The subjective experience of acute pain. An assessment of the utility of 10 indices. Clin J Pain 1989;5:153-9. 51. Kerry SM, Bland JM. Sample size in cluster randomization. BMJ 1998;316:549. 52. Campbell MJ. Cluster randomized trials in general (family) practice research. Stat Methods Med Res 2000;9:81-94. 53. Kerry SM, Bland JM. The intracluster correlation coefficient in cluster randomization. BMJ 1998;316:1455. 54. Campbell MK, Thomson S, Ramsay CR, MacLennan GS, Grimshaw JM. Sample size calculator for cluster randomized trials. Comput Biol Med 2004;34:113-25. 55. Hemming K, Girling AJ, Sitch AJ, Marsh J, Lilford RJ. Sample size calculations for cluster randomized controlled trials with a fixed number of clusters. BMC Med Res Methodol 2011;11:102. 56. Abbott JH, Robertson MC, Chapple C, et al. Manual therapy, exercise therapy, or both, in addition to usual care, for osteoarthritis of the hip or knee: a randomized controlled trial. 1: clinical effectiveness. Osteoarthritis Cartilage 2013;21: 525-34. 57. Zhang W, Doherty M, Peat G, et al. EULAR evidence-based recommendations for the diagnosis of knee osteoarthritis. Ann Rheum Dis 2010;69:483-9. 58. Bellamy N. The WOMAC Knee and Hip Osteoarthritis Indices: development, validation, globalization and influence on the development of the AUSCAN Hand Osteoarthritis Indices. Clin Exp Rheumatol 2005;23(5 Suppl 39): S148-53. 59. Salaffi F, Carotti M, Stancati A, Grassi W. Health-related quality of life in older adults with symptomatic hip and knee osteoarthritis: a comparison with matched healthy controls. Aging Clin Exp Res Aug 2005;17:255-63. 60. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and Osteoarthritis Outcome Score (KOOS)— development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998;28:88-96. 61. Meenan RF, Mason JH, Anderson JJ, Guccione AA, Kazis LE. AIMS2. The content and properties of a revised and expanded Arthritis Impact Measurement Scales Health Status Questionnaire. Arthritis Rheum 1992;35:1-10. 62. Hays RD, Morales LS. The RAND-36 measure of healthrelated quality of life. Ann Med 2001;33:350-7. 63. Grime J, Richardson JC, Ong BN. Perceptions of joint pain and feeling well in older people who reported being healthy: a qualitative study. Br J Gen Pract 2010;60:597-603. 64. Bellamy N, Kirwan J, Boers M, et al. Recommendations for a core set of outcome measures for future phase III clinical trials in knee, hip, and hand osteoarthritis. Consensus development at OMERACT III. J Rheumatol 1997;24: 799-802.

21



21.e1

APPENDIX A. PATIENT INFORMATION & WRITTEN CONSENT follow up on your progress and/or for general communiManual and Manipulative Therapy in Addition to Rehabilitation for Osteoarthritis of the Knee: Assessor-blind Randomized Pilot Trial. We invite you to take part in a study at (XXX) clinic. Our purpose is to study the effect of manipulative therapy and rehabilitation on knee osteoarthritis (knee OA). The principal investigator of the study is (blinded). To join this study - you must not have not received chiropractic care or other manual treatment (massage or physical therapy) for your Knee OA, in the last month. You also cannot have started in the last month other types of care or treatment for your knee OA such as acupuncture, prescribed non-steroidal anti-inflammatory drugs, or medical care, etc. You must be able to stand and walk most of the time without constant dependency on canes, walkers or braces. Occasionally using a cane or other support is acceptable. Serious knee deformity such as “severe knockknees or bowed-legs” or instability (usually manifesting as a knee that “gives way” or “buckles”) may exclude you from this trial.

Procedures If you are eligible and agree to join the study, you will first undergo an interview, a physical exam and knee radiographs, unless you already have recent radiographs. Radiographs of the knee(s) are required for this study to rule our severe knee OA and other pathology that could contraindicate treatment. We will also ask you to perform activity tests (such as bending the knees or hips, standing on1leg, etc.). This visit will take about 2-2½ hours. You will then be randomly assigned to1of 3 different treatment groups. The treatments being compared in this study have been shown to be effective for knee OA, which means that whichever group you are allocated to, the treatment would be beneficial. This study is designed to find out if1treatment is better than the other. Your participation may help us find out which has the most benefit. The 3 treatments are not exactly the same and vary from each other. If you are in Groups A or C you will need to receive treatment around twice per week for a total of 6 visits. If you are in Group B, you will need to come in for a total of 3 visits over 3 weeks for education, as exercise mostly can be done at home. Both groups are examined at the end of the 3 weeks of visits. You must be examined at or just after 3 weeks from when you started and finished your care.

Follow Up After the last in clinic visit there is a 1 week follow up and a 3 month follow up at the (blinded) clinic you will receive a follow-up exam and complete questionnaires. You will then be asked to continue your exercise protocol at home. We will contact some of you by phone if necessary to

cation. At 6 months you will receive a questionnaire with a pre-paid and stamped return envelope with 2 questionnaires to complete and return to us (please see below).

Treatments Group A will receive manual and manipulative therapy alone, in the form of chiropractic care. Chiropractic care in this case consists of adjustive or manipulative therapy of the knee, spine, hip, ankle and foot, and any other indicated joints and soft tissue treatment and exercise. You will also receive advice and regular communication on exercise and health habits from licensed, experienced chiropractor(s). This will take up to 1 hour each visit. Group B will receive a rehabilitation program. In this case treatment consists initially of exercise therapy performed under supervision in the clinic for the knee OA. It also consists of training, education, advice and as needed communication about how to carry out home exercises for you knee OA and on health habits from licensed, experienced chiropractor(s). This will take up to 1 hour each visit. Group C will be treated with manual & manipulative therapy combined with rehabilitation. We will also ask you (whichever group you are assigned too) to keep a diary and record medication use. There is a 6 month by mail only follow up in which we ask you to fill out 2 forms and send them back to (blinded). These will have a self-addressed pre-postage paid stamped envelope.

Benefits You may benefit by getting the results of your exams. The treatment may also benefit your knee osteoarthritis. All exams, treatment and informational materials are free of charge. Future patients with knee OA may also benefit as the results of this study may inform clinical practice so that patients get the best available treatment for their problem.

Risks/Discomforts All tests and treatments used in this study are commonly used by chiropractors. These treatments are believed to involve little risk. Serious side effects are rare. Some people experience mild and temporary discomfort (soreness and or stiffness) after a chiropractic adjustment. Some rehabilitation procedures will require mild to moderate exertion. Some people may experience fatigue, muscle soreness and or stiffness after the rehabilitation procedures. We will make every effort to prevent injury during this study. No compensation (such as medical expenses, lost wages, or payments for emotional distress) is offered by (blinded), (as all treatments in this research have been independently or together long and commonly employed

21.e2


and are considered within the standard of care and safely delivered by chiropractors) although compulsory insurance covers matters of malpractice. (blinded) do not pay for medical treatment needed due to injury. It does not pay for non-related medical problems or financial losses due to injury. If an injury occurs, you do not give up any legal rights as a research participant by signing this consent form. If you think you have suffered a research-related injury, let the study physician or doctor know right away. Chiropractors will check you carefully at the beginning of the study. We will not allow you to take part in the study if we believe that you are at risk for serious injury from any of the study treatments. If you are in the study and the investigators feel that it is harmful for you to continue, they will remove you from the study.


without prejudice at any time For further information, concerns, complaints or withdrawal from the study, please contact (blinded).

Confidentiality of Records All records with your name in them will be kept private. Your name will only be available to people conducting the study. In any report we publish or present, we will not include information that will make it possible to identify you. Questions regarding this Study: Please ask the research assistant or doctor any questions you have now. If you have questions later, please contact (blinded). We will give you a copy of this form to keep for your records.

Alternatives to Participation

Statement of Consent

Your regular medical doctor can give you a similar exam and prescribe treatment. Many local chiropractors or some physical therapists can give you similar treatment. Many local exercise programs are also available. Voluntary Participation: You are free to take part or not take part in this study. You may withdraw for any reason

I, have read the above information and all Print Name Here my questions have been addressed. I freely and voluntarily consent to participate in this study. Signature:____________________________________ Date_________________

Manual physical therapy and perturbation exercises in knee osteoarthritis.

Yoga for managing knee osteoarthritis in older women: a pilot randomized controlled trial.

The addition of upper cervical manipulative therapy in the treatment of patients with fibromyalgia: a randomized controlled trial.

Addition of transcranial direct current stimulation to quadriceps strengthening exercise in knee osteoarthritis: A pilot randomised controlled trial.

Manipulative therapy and rehabilitation for recurrent ankle sprain with functional instability: a short-term, assessor-blind, parallel-group randomized trial.

Comparative Effectiveness of Tai Chi Versus Physical Therapy for Knee Osteoarthritis: A Randomized Trial.

A pilot randomised clinical trial of physiotherapy (manual therapy, exercise, and education) for early-onset hip osteoarthritis post-hip arthroscopy.

Comparison of the clinical effectiveness of thermal cure and rehabilitation in knee osteoarthritis. A randomized therapeutic trial.

The effect of exercise therapy on knee osteoarthritis: a randomized clinical trial.

Assessing the comparative effectiveness of Tai Chi versus physical therapy for knee osteoarthritis: design and rationale for a randomized trial.

Effect of eccentric isokinetic strengthening in the rehabilitation of patients with knee osteoarthritis: Isogo, a randomized trial.

Medium term effects of including manual therapy in a pulmonary rehabilitation program for chronic obstructive pulmonary disease (COPD): a randomized controlled pilot trial.

Effectiveness of massage therapy as co-adjuvant treatment to exercise in osteoarthritis of the knee: a randomized control trial.

Effects of the addition of transcranial direct current stimulation to virtual reality therapy after stroke: a pilot randomized controlled trial.

The Influence of Tai Chi Exercise on Proprioception in Patients with Knee Osteoarthritis: Results from a Pilot Randomized Controlled Trial.

Efficacy of manual and manipulative therapy in the perception of pain and cervical motion in patients with tension-type headache: a randomized, controlled clinical trial.

Exercise Training in Treatment and Rehabilitation of Hip Osteoarthritis: A 12-Week Pilot Trial.

Intra-Articular Corticosteroids in Addition to Exercise for Reducing Pain Sensitivity in Knee Osteoarthritis: Exploratory Outcome from a Randomized Controlled Trial.

Chiropractic spinal manipulative therapy for cervicogenic headache: a single-blinded, placebo, randomized controlled trial.

Randomized trial comparing robotic to manual ablation for atrial fibrillation.

Effectiveness of moxibustion treatment as adjunctive therapy in osteoarthritis of the knee: a randomized, double-blinded, placebo-controlled clinical trial.

Chinese Herbal Bath Therapy for the Treatment of Knee Osteoarthritis: Meta-Analysis of Randomized Controlled Trials.

Wearable sensor-based rehabilitation exercise assessment for knee osteoarthritis.

High-speed resistance training and balance training for people with knee osteoarthritis to reduce falls risk: study protocol for a pilot randomized controlled trial.