What is the clinical and cost-effectiveness of surgery with Medial Opening wedge high TIbial Osteotomy compared with Non-surgical treatment (MOTION) in the management of osteoarthritis of the knee in patients younger than 60 years?

Abstract

Aims

For patients aged < 60 years with symptomatic medial compartment knee osteoarthritis (OA), there is uncertainty regarding the most effective management (surgical or non-surgical). Surgical realignment with a high tibial osteotomy (HTO) can relieve pain and improve function, while also potentially avoiding or delaying a knee arthroplasty. However, HTO has never been compared with non-surgical treatment in this patient group. MOTION (Medial Opening wedge high TIbial Osteotomy compared with Non-surgical treatment) aims to determine the relative clinical effectiveness and cost-effectiveness of HTO compared with non-surgical management for patients aged < 60 years with medial compartment knee OA.

Methods

MOTION is a pragmatic, multicentre, prospective randomized open blinded endpoint (PROBE) parallel group-controlled superiority trial. A total of 224 patients (112 per arm) will be recruited from approximately 20 UK sites. Patients aged < 60 years with moderate to severe symptomatic knee OA localized to the medial compartment, in whom surgical intervention is indicated, are eligible. The intervention is surgery with medial opening wedge HTO followed by standard postoperative rehabilitation, compared with a non-surgical intervention in the form of personalized knee therapy (PKT) focused on a structured, progressive, and tailored exercise programme of rehabilitation. Participants will be randomly assigned to one of the two treatment arms in a 1:1 ratio. The primary outcome is the Knee Injury and Osteoarthritis Outcome Score (KOOS) at 24 months post randomization. Secondary outcomes include the Oxford Knee Score (OKS), Forgotten Joint Score (FJS), EuroQol five-dimension questionnaire (EQ-5D), Sleep Problem Scale, return to work, need for secondary surgical intervention, and adverse events at 12 and 24 months. Health economic evaluation will determine the cost-effectiveness of the interventions. MOTION is funded by the National Institute for Health and Care Research (Ref: NIHR129820).

Discussion

MOTION aims to determine whether HTO is a superior option to non-surgical management (PKT) for patients aged < 60 years with medial compartment knee OA with regard to both the clinical benefit for patients and cost-effectiveness for the NHS.

Cite this article: Bone Jt Open 2025;6(12):1598–1610.

Introduction

Osteoarthritis (OA) is one of the leading causes of joint pain and disability worldwide,^1,2 with the knee most commonly affected in people aged > 45 years. Epidemiological data from Versus Arthritis suggest that one in five people in the UK seeks treatment each year for knee OA, with this number estimated to increase over the next two decades.³ There were 2.36 million working-age people who sought treatment for knee OA in 2013,¹ and the annual incidence of new general practitioner (GP) consultations for knee pain in adults aged > 50 years is approximately 10%.⁴ Furthermore, about half the patients with severe knee pain may not visit their GP, due to a perception that ‘nothing can be done’, highlighting the unmet need in this patient group.⁵

The most common surgical treatment to address pain and disability from knee OA is with knee arthroplasty (either partial or total). However, the median age of patients in the UK having a knee arthroplasty is 70 years, with such surgery in patients aged < 60 years associated with an increased failure rate according to data from the National Joint Registry (NJR).⁶ Therefore, every effort to avoid joint arthroplasty at an early age is worthwhile for the younger active patient. The primary treatment options in this patient group are either non-surgical treatment, or joint realignment through a high tibial osteotomy (HTO) if the disease mainly affects the medial compartment of the knee. However, these two interventions have never been compared in a randomized controlled trial (RCT). Furthermore, the uncertainty in relation to the optimal treatment option has been recognized by the James Lind Alliance (JLA) which has designated the surgical and non-surgical options for the treatment of knee OA as a top ten priority-setting partnership.⁷

The MOTION (Medial Opening wedge high TIbial Osteotomy compared with Non-surgical treatment) trial aims to answer the following research question: for patients aged < 60 years with medial compartment knee OA, what is the relative clinical effectiveness and cost-effectiveness (modelled over a lifetime horizon) of HTO compared with non-surgical management at 12 and 24 months?

Methods

Study design

MOTION is a multicentre, prospective randomized open blinded endpoint (PROBE) parallel group-controlled trial. Patients will be randomly assigned to one of the two treatment arms, in a 1:1 ratio by the Edinburgh Clinical Trials Unit (ECTU) using a web-based system. A total of 224 patients (112 in each arm) will be recruited from approximately 20 UK sites over a 24-month period. To ensure the groups are balanced, a minimization algorithm will incorporate age, sex, BMI, and baseline Knee Injury and Osteoarthritis Outcome Score (KOOS).⁸ Patients will be randomized to either the surgical intervention of HTO or non-surgical intervention of personalized knee therapy (PKT). The setting will be within the orthopaedic surgical departments of the NHS hospitals in the UK that treat patients with medial compartment knee OA with facilities to support the trial. Patients will be followed up for 24 months following randomization.

MOTION is funded by the NIHR (Ref: NIHR129820). The study was given a favourable opinion by the Research Ethics Committee (REC number 22/SC/0046). The trial is registered with the ISRCTN register (ISRCTN: 99042491). The current version of the protocol is V5.0 dated 31 October 2024. The study opening date was August 2022 with a planned end date of July 2027.

The study design incorporates a nine-month internal pilot that will be reviewed by the Data Monitoring Committee (DMC), Trial Steering Committee (TSC), and the funder to determine whether the study progresses to the full trial. The aim is to open at least five sites during the pilot phase (Table I). An embedded process evaluation will record specific factors affecting the recruitment rate to provide further data to optimize recruitment.

Table I.

Stop/Go red, amber, green criterion for internal pilot phase.

Progression criterion – internal pilot	Red	Amber	Green
Trial recruitment
Number of sites opened during pilot	≤ 2	3 to 4	> 5
Total number of participants recruited	< 10	10 to 20	> 20
Recruitment rate/site/month	< 0.2	0.2 to 0.7	> 0.7
Non-adherence
Crossover rate, %	> 50	20 to 50	< 20
Off-protocol intervention rate, %	> 30	10 to 30	< 10

Recruitment

Inclusion and exclusion criteria are detailed in Table II. Potential patients will be identified and approached in outpatient or specialist knee clinics by the usual care team comprising the participating surgeon or their delegated specialist trainee. Patients will be provided with the MOTION patient information sheet (PIS) and consent form and asked if they are willing to be contacted by the local research team using a ‘MOTION consent to contact’ form. Equipoise will be established at this first clinic visit. If patients have received intensive nonoperative treatment before attendance, then they would be deemed unsuitable for recruitment as the conditions for equipoise would not be met. As part of the PIS, a detailed explanation of the PKT intervention is found so that patients can better understand how this may differ from other nonoperative treatments they may have received prior to attendance.

Table II.

Inclusion and exclusion criteria for patient recruitment.

Inclusion criteria	Exclusion criteria
Patients aged < 60 yrs with symptomatic medial compartment knee OA who the treating orthopaedic surgeon considers a suitable candidate for medial opening wedge HTO	Age < 18 yrs or > 60 yrs
	BMI > 40 kg/m2
	Patients considered for HTO but who DO NOT have any knee OA including: Offloading HTO for concomitant cartilage repair (No OA) Offloading HTO solely to treat ligamentous instability (ACL/PCL) Symptomatic avascular necrosis/osteonecrosis Correction of intra-articular or extra-articular post-traumatic knee deformity
	Patients requiring double-level knee osteotomy for correction of deformity
	History of inflammatory arthropathy including rheumatoid arthritis, ankylosing spondylitis, gouty arthropathy, and psoriatic arthropathy
	Previous high tibial or distal femoral osteotomy in same or contralateral knee
	Previous knee arthroplasty (partial or total) in the same or contralateral knee
	Cognitive impairment resulting in the inability to consent
	Inability to comply with study procedures
	Previous history of septic arthritis in the knee

ACL, anterior cruciate ligament; HTO, high tibial osteotomy; OA, osteoarthritis; PCL, posterior cruciate ligament.

A screening visit to assess their eligibility for the study will be arranged. If the patient is identified during an outpatient appointment, this screening visit could coincide with their outpatient clinic appointment, depending on local circumstances. Participants meeting the inclusion and exclusion criteria will be approached for consent by a member of the research team delegated to take consent, allowing sufficient additional time to consider the study information. Following consent, data will be collected to allow randomization. There is provision within the trial consent form to request patient consent for longer-term follow-up at five and ten years postintervention through data linkage, to measure survivorship of HTO and later requirement for knee arthroplasty. This will be a separate study and additional funding to conduct the study will need to be sought.

The trial incorporates a process evaluation. As part of the initial consent to be contacted, patients and delivery staff will also be asked for their consent to be contacted by the process evaluation team. A subgroup of patients and delivery staff (including patients who agree to participate in the MOTION study and those who do not) will be interviewed by the process evaluation team to explore their decision-making.

Withdrawals

Participants are free to withdraw from the study at any point, or a participant can be withdrawn by the investigator. If withdrawal occurs, the primary reason for withdrawal will be documented in the participant’s case report form. The participant will have the option of withdrawal from all aspects of the trial but continued use of data collected up to that point. To safeguard rights, minimum personally identifiable information will be collected, and a withdrawal form completed. The trial management and trial steering committee will monitor these withdrawals from the trial. The process evaluation team will also be informed of any patient who previously gave permission for re-contact but has since withdrawn from the study.

Randomization

After patients have consented and their baseline forms have been completed, the recruiting research nurse will contact the ECTU, either by telephone or via the internet, to access a secure randomization service, ensuring allocation concealment. The randomization service will record information and check patient eligibility once again to avoid inappropriate entry of patients into the trial. ECTU will then perform independent and concealed random allocation (1:1), stratified by centre. The local site research team will be informed of the allocation. Following randomization, patients will be added to the routine NHS waiting lists for physiotherapy (to receive PKT) or HTO (surgical intervention).

Interventions

Surgical intervention (high tibial osteotomy, HTO): Patients randomized to the surgical intervention group will a receive a medial opening wedge HTO. The aim of HTO is to alleviate arthritic knee pain by shifting the weightbearing axis of the lower limb away from the diseased medial compartment of the knee, ideally to a zone between 55% to 65% of the tibial width (as measured from the medial tibial border). Based on preoperative hip-knee-ankle (HKA) radiographs, the size of the opening wedge is calculated in millimetres using a variety of commercially available digital and computer-assisted measurement technologies which are already embedded within the Picture Archiving and Communication System (PACS) or similar software available for use within all NHS hospitals. Under general anaesthesia and using standardized surgical instrumentation with intraoperative imaging control, this wedge is created and opened to the preoperatively planned measured distance. The intended correction is then checked using intraoperative imaging or a ruler to validate the actual correction against the preoperatively planned intended correction. If satisfactory, the opening wedge is then stabilized with a locked plate and screws. The wound is closed and dressings applied. Depending on surgeon preference, the patient may or may not be permitted to bear weight in the postoperative period. The patient is mobilized with the assistance of walking aids as required (crutches, hinged braces) under the supervision of a physiotherapist in hospital. Most patients are discharged home within two days of the intervention. The patient is instructed to attend for follow-up during the postoperative period according to local protocols to monitor the operated knee with sequential clinical and radiological examination to ensure progress with rehabilitation until independently mobile. The postoperative protocols and frequency of clinical/radiological monitoring vary between different surgeons and hospitals, but the postoperative recovery is usually complete within three months of the intervention. During this period a final postoperative HKA radiograph is also taken and compared with the preoperative HKA radiograph to evaluate whether the intended correction was successfully achieved (Figure 1).

Fig. 1.

a) A black line is drawn from the centre of the hip to the centre of the ankle, and represents the weightbearing axis (how the body weight passes through the leg). Normally this weightbearing axis should also pass through the centre of the knee. In this patient suffering from knee osteoarthritis, the weightbearing axis is abnormal and passes through the inner diseased half of the knee instead of the centre of the knee. b) A wedge has been opened just below the knee joint and stabilized with plates and screws in the operation called high tibial osteotomy (HTO). By opening this wedge of bone at the proximal end of a long bone, the overall alignment of the lower leg and its weightbearing axis is altered, as shown in c). The lower leg has been realigned so that the weightbearing axis (black line) now passes through the middle of the knee, away from the diseased inner half of the knee. This shifting of the weightbearing axis away from the arthritic inner part of the knee following HTO relives pain, and reduces disability with preservation of the native knee joint.

Personalized knee therapy (PKT) is a package of evidence-based non-surgical interventions that are currently in use to varying extents across the NHS for managing medial compartment knee OA in patients aged < 60 years. The specific PKT intervention was developed for the trial and is detailed in the Supplementary Material as per the template for intervention description and replication (TIDieR) guidelines.⁹ The core MOTION nonoperative intervention is a progressive and tailored exercise intervention delivered over 12 to 16 weeks under the guidance of a physiotherapist with concomitant lifestyle behaviour modification and pain management advice provided by the treating physiotherapist (as per National Institute for Health and Care Excellence (NICE) guidance).² Additional to the core exercise-focused package (if already routinely delivered in local treatment pathways), patients may also be offered medial offloading bracing or shoe insoles, intra-articular steroid injections to the knee, manual therapy techniques (mobilizations, distractions etc.) performed by the trial physiotherapist. All additional interventions and additional exercise undertaken outwith the core trial intervention will be documented. The core intervention comprises three 30-minute structured exercise sessions per week over a minimum 12-week, maximum 16-week, intervention window. Six of these sessions are to be directly supervised by a trial physiotherapist to tailor and progress the exercises appropriately to the individual participant. Flexibility as to the specific exercises delivered will be built into the intervention to allow tailoring; however, the exercises are intended to address the presenting impairments in knee joint range of motion and neuromuscular control around the knee. Progression of exercises will be based on achievement of previous exercises and pain tolerances. A broad treatment aim will be to achieve 90% of the strength/neuromuscular control of the contralateral (unaffected) side. There is flexibility built into the programme delivery, however the intention is for three 30-minute sessions per week of focused exercise rehabilitation with a rest day between these sessions. Regular physiotherapy supervision should be offered across the exercise intervention, with the timing of contact sessions dependent on the needs of the participant and local logistics.

Follow-up and outcomes

Figure 2 and Table III detail the participant pathway and assessment timeline for recruited patients, respectively. As with many surgical trials, where the surgical site is clearly visible, it is not feasible to completely blind patients, surgeons, or outcome assessors. Surgeons and patients will not be blind to the interventions in this trial. Outcome assessors will be blinded because all outcome scores are patient-reported and collected over the telephone, email, or by post, avoiding problems with unblinding due to the presence of the surgical scar.

Fig. 2.

MOTION (Medial Opening wedge high TIbial Osteotomy compared with Non-surgical treatment) trial CONSORT participant pathway. EQ-5D-3L, EuroQol five-dimension three-level questionnaire; FJS, Forgotten Joint Score; KOOS, Knee Injury and Osteoarthritis Outcome Score; OKS, Oxford Knee Score.

Table III.

Tabulated summary of assessments.

Assessment	Screening	Baseline	6-mth MRAV proforma for HTO	6-mth MRAV proforma for PKT	12-mth PROMs	24-mth PROMs
Who is responsible for assessment?	Local site team	Local site team	Local site team	Local site team	ECTU	ECTU
Allowed variation in months			± 6	± 6	± 3	± 3
How is the assessment done?	Clinic/ Medical records	Clinic	Medical records	Clinic/Medical record	Email Postal Online	Email Postal Online
Assessment of eligibility criterion	X	X
Log of eligible patients excluded (decline, exclusion criterion, other reason)		X
Written informed consent		X
Baseline data and demographics		X
Randomization		X
MRAV proforma			X	X
Primary outcome score – KOOS		X			X	X
OKS10,11		X			X	X
FJS12		X			X	X
EQ-5D-5L13		X			X	X
Pittsburgh Sleep Problem Scale14		X			X	X
Return to work/employment questionnaire		X			X	X
Additional operative intervention in the study					X	X
Complications					X	X

EQ-5D-5L, EuroQol five-dimension five-level questionnaire; FJS, Forgotten Joint Score; HTO, high tibial osteotomy; KOOS, Knee Injury and Osteoarthritis Outcome Score; MRAV, minimum requirement/acceptable variation; OKS, Oxford Knee Score; PKT, personalized knee therapy; PROMs, patient-reported outcome measures.

Delays in the delivery of the intervention post-randomization due to NHS waiting lists (for both surgical and non-surgical interventions) and the time lag related to the recovery from the interventions (approximately three months) may not capture the 12-month KOOS timepoint post-intervention in all participants. The 24-month follow-up interval for the primary outcome measure (and all other secondary measures) will ensure data capture for all participants at least 12 months post-intervention, and will further support interpretation of the primary outcome measure.

Primary outcome

The primary outcome will be the 24-month Knee Injury and Osteoarthritis Outcome Score (KOOS), a validated patient reported outcome measure (PROM) intended for knee OA.^8,15-18 The KOOS score comprises five subscales: pain (nine items), activities of daily living (17 items), sport and recreation function (five items), knee-related quality of life (four items), and other symptoms (seven items). Each subscale is scored separately from 0 (extreme knee problems) to 100 (no knee problems) and a composite score derived from the five subscales is used for the primary outcome analysis.⁸ The clinical utility of the KOOS is highlighted by large international patient datasets (> 100,000 unique patient records) and frequent use in scientific publications.^16,19 The strength of its measurement properties are content validity, internal consistency, test-retest reliability, construct validity, and responsiveness for age- and condition-relevant subscales.¹⁶ Structural validity, cross-cultural validity, and measurement error have been reported to be potential weaknesses of its measurement properties.^8,16

Secondary outcomes are listed below.

Minimum requirements/acceptable variation

This proforma will be completed for each participant approximately six months post-intervention to record the fidelity of the trial interventions. The proforma will be completed by the local research team at each participating centre. It should be noted that this proforma is collected from approximately six months post-intervention, and this interval may be different from the time since randomization due to NHS waiting lists. In the PKT group, data will be collected and recorded at each one-to-one session. The date of the last one-to-one physiotherapy session (usually the sixth session) will be considered as the date for completion of the intervention.

The OKS is a widely used, valid, reliable and responsive 12-item PROM. Although it has been adopted by the NHS PROMs programme in England and Wales, as the primary outcome measure for knee arthroplasty to monitor and benchmark the performance of health providers, the OKS summary scale, together with its pain and functional component subscales, has excellent measurement properties when used with patients with knee OA undergoing nonoperative treatment.^10,11,20

The FJS-12 Knee is designed to assess outcomes in patients undergoing conservative or operative treatment of the knee and is based on the patients’ ability to forget about a joint as a result of successful treatment.^12,21,2212It has been designed specifically to reduce ceiling effects commonly associated with many patient-reported outcome measures (PROMs) in this patient group, such as the OKS.²⁰

The EQ-5D is a preference-based health-related quality of life measure with one question for each of the five dimensions that comprise mobility, self-care, usual activities, pain/discomfort, and anxiety/depression.²² The answers given to ED-5D model unique health states or can be converted into EQ-5D index and utility scores anchored at 0 for death and 1 for perfect health. It includes a visual analogue scale (VAS) from 0 (the worst possible health status) to 100 (the best possible health status). When used in economic evaluation EQ-5D preference weights are combined over time with mortality data to compute quality-adjusted life-years (QALY).

Our patient and public involvement (PPI) work indicated that a measurement of the quality of sleep would be important in assessing the efficacy of the interventions. The Sleep Problem Scale consists of four questions, rated on a six-point frequency rating scale, ranging from 0 (not at all) to 5 (22 to 31 days/month).¹⁴ Sleep disturbance is defined as a mean score ≥ 4, corresponding to at least 15 troubled nights per month. The lead centre has used and published this scale previously and patients found it very acceptable.²³

KOOS subscales at 12 and 24 months post-randomization

Calculating a composite KOOS score (primary outcome) from subscale scores ensures similar weight from all subscales. As indicated in the instructions for measuring and interpretation of the KOOS score for RCTs, each of the five KOOS subscales will also be included as secondary outcomes to enable clinical interpretation of the results.⁸

OKS at 12 and 24 months post-randomization

The OKS is a widely used, valid, reliable and responsive 12-item PROM. Although it has been adopted by the NHS PROMs programme in England and Wales, as the primary outcome measure for knee arthroplasty to monitor and benchmark the performance of health providers, the OKS summary scale, together with its pain and functional component subscales, has excellent measurement properties when used with patients with knee OA undergoing nonoperative treatment.^10,11,20

Forgotten Joint Score-12 (FJS-12)^12,21 at 12 and 24 months post-randomization: The FJS-12 Knee is designed to assess outcomes in patients undergoing conservative or operative treatment of the knee and is based on the patients’ ability to forget about a joint as a result of successful treatment. It has been designed specifically to reduce ceiling effects commonly associated with many patient-reported outcome measures (PROMs) in this patient group, such as the OKS.²⁰

EuroQol five-dimension, five-level questionnaire (EQ-5D-5L) score 12 and 24 months post-randomization: The EQ-5D is a preference-based health-related quality of life measure with one question for each of the five dimensions that comprise mobility, self-care, usual activities, pain/discomfort, and anxiety/depression.²² The answers given to ED-5D model unique health states or can be converted into EQ-5D index and utility scores anchored at 0 for death and 1 for perfect health. It includes a Visual Analogue Scale (VAS) from 0 (the worst possible health status) to 100 (the best possible health status). When used in economic evaluation EQ-5D preference weights are combined over time with mortality data to compute quality-adjusted life-years (QALY).

Pittsburgh Sleep Problem Scale¹⁴ at 12 and 24 months post-randomization: Our patient and public involvement (PPI) work indicated that a measurement of the quality of sleep would be important in assessing the efficacy of the interventions. The Sleep Problem Scale consists of four questions, rated on a six-point frequency rating scale, ranging from 0 (not at all) to 5 (22 to 31 days/month). Sleep disturbance is defined as a mean score ≥ 4, corresponding to at least 15 troubled nights per month. The lead centre has used and published this scale previously and patients found it very acceptable.²³

Return to work/employment questionnaire: Patients will be asked about employment status pre- and postintervention (returned to same job/other job, full-time/part-time, retired, sick leave, on welfare benefits) and the nature of this employment (sedentary, light manual, moderate manual, heavy manual). Those who return to work will be asked to document the amount of sick leave that they have taken. The lead centre has used and published this method of measuring return to work status following hip and knee arthroplasty.^24,25

Additional study-knee operative intervention at 12 and 24 months post-randomization: The percentage rate of secondary surgical intervention during the trial period due to off-protocol interventions (e.g. knee arthroplasty), hardware removal, and reoperation will be compared for the two interventions.

Intraoperative and postoperative complications at 12 and 24 months post-randomization/adverse events: Expected surgical complications that will be recorded include (but not be limited to) intraoperative fracture, superficial/deep wound infection, bleeding, thromboembolic complications, neurovascular injury, nonunion, and delayed union. Compared with the surgical intervention, the risks associated with non-surgical intervention for knee OA are minimal but will also be recorded including muscle soreness, flare-ups of knee OA pain resulting from a change in exercise behaviour, and risks associated with pain medications/steroid injections into the knee. Adverse events will be categorized as involving the index knee or sites other than the index knee, and serious adverse events will be identified according to established international definitions.

Health Economic Evaluation Outcomes

1. Health and social care resource use, associated NHS and personal social services (PSS), and QALYs at 24 months

2. Incremental cost per QALY at 24 months

3. NHS and PSS cost, and QALYs as modelled over a lifetime horizon to account for future impacts on the need for total knee arthroplasty (TKA) and associated revision surgery, and their timing relative to retirement

4. Incremental cost per QALY as modelled over a lifetime horizon to account for future impacts on the need for TKA and associated revision surgery, and their timing relative to retirement

There is provision within the trial with patient consent to allow longer-term follow-up at five and ten years post-intervention through data linkage, to measure survivorship of HTO and later requirement for knee arthroplasty. The outcome measures used in this study to quantify the effectiveness of the intervention mirror those used in the national joint registries including the National Joint Registry (NJR), which uses data from England, Wales, Northern Ireland, the Isle of Man, and the States of Guernsey, and the UK Knee Osteotomy Registry (UKKOR).^26,27 It is envisaged that the UKKOR dataset will serve as a useful comparator to benchmark and validate long-term outcome data from this trial in keeping with the IDEAL (idea, development, exploration, assessment, long-term study) framework.²⁸ The longer-term follow-up will also link with the economic modelling over a lifetime horizon with potential to update the economic model at a later point to validate and correct model assumptions. The information analyses would help prioritize selection of data to target critical areas of uncertainty within the economic model, and ultimately inform NICE guidance and policy.

Sample size

The minimal clinically important difference (MCID) for the primary outcome measure is 8 to 10,⁸ so we have used a value of 9, the mid-point of this range. To account for 20% unidirectional crossover from the non-surgical arm to the surgical arm, a MCID of 7.2 has been used. If the observed treatment effect was going to be 9, then the observed mean KOOS in one arm would have been x, and in the other arm it would have been x + 9. If there is 20% crossover in one direction, and you analyze by intention-to-treat, then the observed mean KOOS in one arm is now still x, and in the other arm it is now [0.8(x + 9) + 0.2x]= x+7.2. We have assumed the SD for the KOOS will be 15.^8,15-18 To account for and adjusting for baseline KOOS, an assumed correlation between baseline and follow-up measurement of 0.25 has been used. This reduces the SD to 13. Variance2 = (1-rho2) variance1 where variance1 = 152, and rho = 0.5.²⁹ Therefore, new SD=√(168.75) ≈13. The intracluster correlation coefficient of the KOOS among surgeons is likely to be around 0.056.³⁰ Using this value, and assuming there will be approximately five patients per surgeon, gives a design effect of 1.224 (design effect = 1+(0.056*(5-1)) = 1.224). The trial has adjusted for clustering among surgeons through multiplying by the square root of the design effect (1.106) to obtain a SD in the surgical arm of 14.4. In the non-surgical intervention arm of the trial, the number and variety of physiotherapists treating each patient is variable, and clustering by physiotherapist is not appropriate.

However, the components of the non-surgical intervention are likely to be similar with each centre, and vary between centres, so clustering on centre within the non-surgical intervention arm is more appropriate. It is assumed that the intracluster correlation coefficient of the KOOS in the non-surgical intervention arm of each centre is 0.056. If there are ten patients per centre, we get a design effect of 1.504 (design effect = 1+(0.056*(10-1)) = 1.504). Adjustment for clustering in the physiotherapy arm has been done by multiplying the square root of the design effect (1.226) to obtain a SD in the physiotherapy arm of 15.9. Using 90% power, two-sided p = 0.05, with a difference of 7.2, and SDs of 14.4 and 15.9, this gives a sample size of 95 per group. Allowing for 15% loss to follow-up gives 112 per group.

Planned statistical analysis

Primary analysis will compare the KOOS at the 24-month follow-up, between the randomized treatment arms, using a linear mixed model adjusting for surgeon as a random effect. Clustering in the surgical and non-surgical arms will be considered according to the recommended method and as detailed in the samples size calculation.³¹ There are no preplanned subgroup analyses, and no preplanned formal interim analyses. Analysis of secondary outcomes will follow similar methods to those proposed for the primary analysis, where the data are of an appropriate form. A low amount of missing data for the scales (approximately 10% of our primary outcome) is presumed and thus a complete case analysis will be carried out. Primary analysis will be based on intention-to-treat, but a complier average causal effect (CACE) analysis will also be performed as a sensitivity analysis. The statistical aspects of this trial will follow the Standard Operating Procedures of ECTU.

Health economic analysis

Full details of these analyses will be specified in a comprehensive Health Economic Analysis Plan. The following section offers an overview for ease of reference only. Analysis will have two components: 1) a 24-month within trial analysis; and 2) a decision analytic simulation model to account for the impacts of future TKAs and revisions, and their timing relative to retirement and employment potential. To maximize UK policy relevance, health economic analysis will follow NICE reference case recommendations including: NHS and PSS costing perspective for primary analyses; cost-utility approach (results presented in terms of incremental cost-effectiveness ratios (ICERs) on a cost per QALY basis; choice of discount rate; and use of probabilistic sensitivity analysis (PSA).³² Secondary analysis will examine differences between arms in employment status. Healthcare resource use (HCRU) will be combined with standard UK price weights to generate costs,³³ and QALYs will be derived from EQ-5D utility scores using NICE’s preferred scoring algorithm at time of analysis. Within trial cost-utility analysis will be undertaken using generalized linear modelling to account for skew,³⁴ and a recycled predictions approach.³⁵ Change in employment status will be examined using logistic regression. Missing data will be imputed using appropriate techniques depending on degree of missingness, most likely multiple imputation by chained equations (which is considered the gold standard in this area).³⁶ Of note, the most important cost factors (inpatient readmission and outpatient activity) will be obtained from medical records and thus are not anticipated to be prone to missingness. Longer run cost-efficiency in terms of incremental cost per QALY will be examined using decision analytic simulation modelling. This will extrapolate patient outcomes over a lifetime horizon to account for future impacts on the need for TKA and associated revision, and their timing relative to retirement which can have impacts on employment options. We anticipate using Markov modelling, though more sophisticated options will be explored. The model will be parameterized using trial data, complemented with NJR and UKKOR data on longer-term rates of TKA, revision TKA, and associated EQ-5D utility scores, and a Cochrane review.^37,38 Additional, potentially non-systematic, targeted literature searches may be necessary for specific parameters if not available from these sources.

Qualitative analysis – process evaluation

In line with Medical Research Council (MRC) guidance, a mixed-method process of evaluation will be conducted.³⁹ The overall aim is to assess the acceptability and feasibly of the intervention from staff and patients. Specifically, it will address the following research objectives: 1) explore trial eligibility, recruitment and retention rates; 2) acceptability of intervention implementation including trial processes and the collection of routine monitoring data; and 3) patient experience of taking part and the contextual factors that influence this.

The process evaluation design will comprise two parts. Part 1 will be embedded within the pilot study, providing delivery detail and context to support the definitive trial. Data will be collected, via semi-structured telephone/video interviews from delivery staff (n = 12, surgeons/physiotherapists/research nurses) at three sites and patients enrolled in the pilot stage (n = 15 to 20, minimum of four from each site and minimum of four from the control arm) to capture initial understanding/expectations of the trial. They will then be followed up once (via semi-structured telephone interview) during the pilot study (months 3 to 6). Staff interviews will explore experiences of taking part, challenges to recruitment/implementation, and issues related to trial delivery. Patient interviews will explore willingness to be randomized and acceptability of intervention or the non-surgical package if in the comparison group. Findings will be fed back to the Trial Management Group and Trial Steering Committee to improve recruitment/trial delivery for the definitive trial.

Part 2 will extend the process evaluation into the definitive trial. This is important because the pilot trial data captures just one follow-up from three sites, providing limited detail on the research objectives listed above and the factors that influence this. By extending the process evaluation into the definitive trial, the design will move to a prospective, longitudinal, cohort study across six sites. The total sample will be boosted to 20 staff and 30 to 40 patients (approximately seven to ten control and 20 to 30 intervention, spread across the three sites). This will further our understanding of the trial implementation, context (delivery may vary by site, for example), and the mechanisms that influence delivery. It will also provide a richer narrative of the patient journey and stakeholder views, which will inform future rollout and create transferable learning to other disciplines. For example, a key area is to explore movement from the non-surgical to surgical arm of the trial. Being able to examine this over multiple periods will generate a greater depth of understanding of the context in which this occurs and the factors that influence the patient decision to switch. Patients will receive three further telephone interviews. The first at six months and then at 12 and 24 months in line with the clinical follow-up. To prevent sample attrition, patients will receive £15 for each completed interview. Finally, it is important to understand the reasons why patients decided not to take part in the trial and if there is anything that could have been done differently to encourage participation in the future. We will, therefore, conduct ten telephone interviews with these patients, who will also be given a £25 voucher.

Data management

The database and data management plan will be created, validated, and administered by the ECTU, following Standard Operating Procedures. All data will be collected by those trained and delegated via the site delegation log. Follow-up at 12 and 24 months will be collected centrally by ECTU via provision of an online portal or postal/emailed follow-ups. The study will comply with the Data Protection Act 2018, which requires data to be anonymized as soon as it is practical to do so. Data will be entered into a secure online study database provided by ECTU. Paper-based case report forms will be stored on site at ECTU under locked conditions for the duration of the trial; these will be considered source documents for this study. Direct access to source data and documents will be granted to authorized representatives from the sponsor, sites, and regulatory authorities to permit trial-related monitoring, audits, and inspections. The study will be conducted in accordance with the principles of the International Conference on Harmonization Tripartite Guideline for Good Clinical Practice (ICH GCP).

Dissemination

The trial team will inform and engage patients, NHS, and the wider population about our work. In formulating this engagement plan we have taken into consideration stakeholder engagement, format, established opportunities for networking, context, and timing in accordance with NIHR guidelines for dissemination. Dissemination will start immediately after the internal pilot phase is completed and continue throughout the course of the main trial. It will be led in conjunction with constant engagement with our PPI group, who have been closely involved throughout the study development and are part of the trial management group. The investigators will be involved in reviewing drafts of the manuscripts, abstracts, press releases, and any other publications arising from the study. Authorship will be determined in accordance with the International Committee of Medical Journal Editors (ICMJE) guidelines and other contributors will be acknowledged.

Discussion

Approximately 30% to 50% of all patients with knee OA suffer from disease predominantly affecting the medial compartment of the joint.^{15,26,37,40,41} The MOTION trial aims to answer the following research question: for patients aged < 60 years with medial compartment knee OA, what is the relative clinical effectiveness and cost-effectiveness (modelled over a lifetime horizon) of HTO compared with non-surgical management at 12 and 24 months?

A Cochrane review identified 21 randomized studies examining HTO in 1,065 patients.⁴² The mean number of participants in the 21 studies was 52 and the mean age was 60 years (42 to 67). None of these studies compared HTO with non-surgical treatment; 19 studies compared different HTO techniques and two compared HTO and partial knee arthroplasty. HTO reduced pain and improved knee function in patients with knee OA. However, this conclusion was based on within-group comparisons, and not on nonoperative controls. The two studies comparing HTO and partial knee arthroplasty reported no significant differences between the treatments at 7.5 years in terms of treatment failure (risk ratio = 1.32) and pain/functional outcome (precise treatment effect not estimable). However, a probabilistic analysis of systematic review data at ten years following surgery found that HTO was most likely to be cost-effective in patients aged < 60 years and partial knee arthroplasty in patients aged > 60 years.³⁷ Registry data on 3,195 HTOs performed between 1987 to 2008 in Finland demonstrated that 89% did not need a knee arthroplasty within five years and 73% did not need one within ten years.²⁶ Such longer-term survival and satisfactory patient-reported outcomes after HTO have also been reported in a recent systematic review of HTO, but the same review also recognized the lack of high-quality studies investigating HTO.⁴³ Finally, HTO does not compromise the later requirement for knee arthroplasty. Joint registry data from Denmark have shown that, when corrected for age and sex, the ten-year survival of knee arthroplasty is not significantly worse when performed after HTO (n = 1,044) compared to without prior HTO (n = 63,763) – 91% compared to 94% (hazard ratio 1.19, p = 0.09).⁴⁴

The NICE clinical guideline CG177 emphasizes a holistic approach to non-surgical management that includes non-pharmacological treatments, pharmacological treatments, and steroid injections.² A Cochrane review of the effectiveness of exercise alone in knee OA concluded that high-quality evidence indicates that land-based therapeutic exercise provides short-term benefit that is sustained for at least two to six months after cessation of formal treatment in terms of reduced knee pain, and there is moderate-quality evidence showing improvement in physical function among people with knee OA.⁴⁵ Recent systematic reviews from international patient and public organizations dedicated to improving the treatment of OA have also provided substantial evidence to suggest moderate effectiveness of non-surgical treatment for knee OA.^46-48 In a randomized trial published this year, patients with knee OA who underwent physiotherapy had less pain and functional disability at one year than patients who received an intra-articular glucocorticoid injection.⁴⁹ A Danish RCT compared the efficacy of a 12-week non-surgical treatment programme with ‘usual care’ in patients with knee OA.⁵⁰ The treatment programme consisted of individualized progressive neuromuscular exercise, patient education, insoles, dietary advice, and analgesics, while usual care comprised two leaflets with information and advice on knee OA and recommended treatments. Compared with usual care, patients undergoing the treatment programme improved more in the KOOS with an adjusted mean difference of 9.6. In a second RCT, the same research group found that 66% of patients receiving the intervention chose to delay knee arthroplasty for at least two years.⁸ Around 12,796 patients with knee OA have now taken part in a tailored non-surgical management programme of exercise and education in Denmark. This was found to be effective in reducing pain by 13.4 points on a 0 to 100 scale.⁵¹ Finally, data from the lead centre indicate that the non-surgical management of symptomatic knee OA in primary care fails to meet minimum treatment guidelines in the vast majority of patients, and suggests that a tailored approach may be more successful in delaying the need for knee arthroplasty.²³

There is no high-quality evidence comparing non-surgical management with surgery for the younger adult patient with symptomatic medial compartment knee OA who may not be suitable for joint arthroplasty. The findings from the MOTION trial hope to impact the management of this sub-group of patients with knee OA by defining the clinical-effectiveness and cost-effectiveness of the interventions studied in the trial.

Take home message

- For patients aged < 60 years with symptomatic medial compartment knee osteoarthritis (OA), there is uncertainty regarding the most effective management (surgical or non-surgical).

- Surgical realignment with a high tibial osteotomy (HTO) can relieve pain and improve function, while also potentially avoiding or delaying a knee arthroplasty.

- However, HTO has never been compared with non-surgical treatment in this patient group. MOTION (Medial Opening wedge high tIbial osteotomy compared with non-surgical treatment) aims to determine the relative clinical effectiveness and cost-effectiveness of HTO compared with non-surgical management for patients aged < 60 years with medial compartment knee OA.

Data Availability

The datasets generated and analyzed in the current study are not publicly available due to data protection regulations. Access to data is limited to the researchers who have obtained permission for data processing. Further inquiries can be made to the corresponding author.