Implementing good life with osteoArthritis from Denmark (GLA:D®) in Australian public hospitals. Part 2: Service providers' perceptions

BACKGROUND

Previous literature has reported the successful implementation of the Good Life with osteoArthritis in Denmark (GLA:D®) program into predominantly private practice settings. There may be unique challenges present within the public hospital setting that influence GLA:D® implementation in public health.

OBJECTIVE

Explore the attitudes and experiences of service providers directly involved in implementing GLA:D® in Australian public tertiary hospitals.

DESIGN

Qualitative descriptive study design.

METHOD

Service providers (n = 14) from three participating hospitals took part in semi-structured focus groups at the completion of the 6-month implementation period. Inductive thematic analysis was employed to identify primary domains across all facilities.

RESULTS

Four broad domains were identified. Factors that influenced uptake included GLA:D® being a recognisable, evidence-based product requiring minimal development or adaptation. The fidelity of the GLA:D® Australia program was challenged by referral of patients with multiple/complex medical comorbidities, and patient preference to complete registry data via paper rather than online. Several operational considerations are required when delivering GLA:D® in a public hospital setting, including adequate numbers of GLA:D®-trained staff, additional screening requirements, obtaining appropriate clinical space, and persisting patient barriers to attending the service. GLA:D® provided benefits beyond improvement in pain and function, including social interactivity, high attendance and promotion of long-term self-management, while also maximising service efficiencies.

CONCLUSIONS

Implementing GLA:D® in Australian public hospitals was supported by service providers. Specific operational and administrative factors, including staff training, patient complexity, and registry requirements should be considered when attempting to embed and sustain GLA:D® in large Australian public tertiary hospitals.

Implementing good life with osteoArthritis from Denmark (GLA:D®) in australian public hospitals. Part 1: Feasibility

BACKGROUND

Literature reporting positive outcomes from the Good Life with osteoArthritis in Denmark (GLA:D®) program in Australia mainly involves patients attending private physiotherapy services.

OBJECTIVE

Evaluate the feasibility of implementing GLA:D® in Australian public hospitals.

DESIGN

Implementation study in three metropolitan tertiary public hospitals over six months.

METHOD

Patients aged ≥18 years with knee or hip joint-related problems deemed appropriate for non-surgical care were invited to participate in GLA:D®. Feasibility was evaluated using RE-AIM framework components (Implementation, Effectiveness, Maintenance) using service-level metrics, patient-level data, and program fidelity assessment. Findings of qualitative interviews with service providers are presented in Part 2.

RESULTS

Implementation: 70 patients (69 with knee osteoarthritis) participated (13 cohorts). 55 (79%) patients attended both education sessions, and 49 patients (70%) attended 10-12 exercises sessions. Fidelity was met based on environmental, therapist, participant- and program-related criteria.

EFFECTIVENESS

At 3 months, patients reported lower average pain (visual analogue scale [0-100 mm]: effect size -0.56, 95% CI -0.88 to -0.23) and disability (HOOS/KOOS-12 [100-0]: 0.67, 0.28 to 1.05), and improved quality of life (EQ-5D overall score: 0.46, 0.11 to 0.80). No adverse events were reported. All patients who completed 3-month assessment (n = 52) would recommend GLA:D®. Maintenance: All participating services elected to continue delivering GLA:D® beyond the study.

CONCLUSIONS

Implementing GLA:D® in Australian public hospitals is feasible, safe, and acceptable to patients with knee osteoarthritis. Public hospital patients with knee osteoarthritis reported improvements in pain, disability, and quality of life similar to previous GLA:D® cohorts.

Evaluation of 3D Printed Burr Hole Simulation Models Using 8 Different Materials

OBJECTIVE

3D printing is increasingly used to fabricate three-dimensional neurosurgical simulation models, making training more accessible and economical. 3D printing includes various technologies with different capabilities for reproducing human anatomy. This study evaluated different materials across a broad range of 3D printing technologies to identify the combination that most precisely represents the parietal region of the skull for burr hole simulation.

METHODS

Eight different materials (polyethylene terephthalate glycol, Tough PLA, FibreTuff, White Resin, BoneSTN, SkullSTN, polymide [PA12], glass-filled polyamide [PA12-GF]) across 4 different 3D printing processes (fused filament fabrication, stereolithography, material jetting, selective laser sintering) were produced as skull samples that fit into a larger head model derived from computed tomography imaging. Five neurosurgeons conducted burr holes on each sample while blinded to the details of manufacturing method and cost. Qualities of mechanical drilling, visual appearance, skull exterior, and skull interior (i.e., diploë) and overall opinion were documented, and a final ranking activity was performed along with a semistructured interview.

RESULTS

The study found that 3D printed polyethylene terephthalate glycol (using fused filament fabrication) and White Resin (using stereolithography) were the best models to replicate the skull, surpassing advanced multimaterial samples from a Stratasys J750 Digital Anatomy Printer. The interior (e.g., infill) and exterior structures strongly influenced the overall ranking of samples. All neurosurgeons agreed that practical simulation with 3D printed models can play a vital role in neurosurgical training.

CONCLUSIONS

The study findings reveal that widely accessible desktop 3D printers and materials can play a valuable role in neurosurgical training.

"It's like a nail being driven in the ankle": A qualitative study of individuals' lived experiences to inform a core domain set for ankle osteoarthritis

OBJECTIVES

Explore individuals' lived experience with ankle osteoarthritis and to identify health-related domains for ankle osteoarthritis based on the perspectives of people living with the condition, as an initial step to address the International Foot and Ankle Osteoarthritis Consortium's recommendation to develop a core domain set for ankle osteoarthritis.

METHODS

A qualitative study using semi-structured interviews was conducted. Interviews were undertaken with individuals with symptomatic ankle osteoarthritis who were aged ≥35 years. Interviews were recorded, transcribed verbatim and thematically analysed.

RESULTS

Twenty-three individuals (16 females; mean (range) age 62 (42-80) years) were interviewed. Five themes were identified: pain, often severe, is a central feature of living with ankle osteoarthritis; stiffness and swelling are key symptoms; ankle osteoarthritis induced mobility impairments compromise enjoyment in life; ankle osteoarthritis instability and balance impairments lead to concerns about falling; there are financial implications of living with ankle osteoarthritis. We propose 17 domains based on individuals' experiences.

CONCLUSION

Study findings indicate that individuals with ankle osteoarthritis live with chronic ankle pain, stiffness and swelling which affect their ability to participate in physical, and social activities, maintain an active lifestyle, and work in physical occupations. From the data, we propose 17 domains that are important to people with ankle osteoarthritis. These domains require further evaluation to ascertain their inclusion in a core domain set for ankle osteoarthritis.

L5 mononeuritis, an uncommon cause of foot drop: illustrative case

BACKGROUND

New-onset adult foot drop is commonly encountered in neurosurgical practice and has a broad differential, including radiculopathy, peroneal nerve palsy, demyelinating diseases, and central causes. Etiology is commonly identified with comprehensive history, examination, imaging, and investigations. Despite familiarity with the management of lumbar spondylosis and peroneal nerve compression causes, rare or uncommon presentations of nonsurgical causes are important to consider in order to avoid nonbeneficial surgery.

OBSERVATIONS

The authors report a very uncommon cause of foot drop: new-onset isolated L5 mononeuritis in a 61-year-old nondiabetic male. They provide a review of the etiology and diagnosis of foot drop in neurosurgical practice and detail pitfalls during workup and the strategy for its nonsurgical management.

LESSONS

Uncommon, nonsurgical causes for foot drop, even in the setting of degenerative lumbar spondylosis, should be considered during workup to reduce the likelihood of unnecessary surgical intervention. The authors review strategies for investigation of new-onset adult foot drop and relate these to an uncommon cause, an isolated L5 mononeuritis, and detail its clinical course and response to treatment.

A retrospective cohort study assessing geometrical parameters related to visual deterioration in pituitary macroadenoma

PURPOSE

We aim to identify specific parameters that will allow identification of those patients at greater risk of vision loss, specifically variables such as geometry, volume and size of the pituitary tumour.

METHODOLOGY

Volume measurements were made with both manual (using the formula width × height × length × 0.5) and video size, sub-voxel interpolation using Brainlab software. The shape of the tumour was divided into four categories, ovoid, dumbbell, pyramidal and other. Chiasmal compression was divided into central and lateral compression and the chiasm location was classified based on pre/post and normal fixed chiasm in relation to the tuberculum sella. We used a visual impairment score ranging from (0-100), which was then further subdivided in to four grades: Grade I - 0-25 (mild); Grade II - 26-50 (moderate); Grade III - 51-75 (severe); Grade IV - 76-100 (subtotal or complete).

RESULTS

Univariate analysis of variables related to vision loss found tumour volume, suprasellar growth and lateral chiasmal compression related to pre-operative vision loss. On multivariate regression analysis tumour volume and lateral chiasmal compression remained significant (p < 0.008). In relation to post-operative vision, univariate analysis indicated that variables such as giant macroadenoma, pre-operative visual impairment score and suprasellar growth were significant. Multivariate analysis showed pre-operative visual impairment and suprasellar growth as the only statistically significant variables.

CONCLUSION

Our ROC analysis suggests tumours with suprasellar extension 16 mm (sensitivity 85 %, specificity 67 %) in sagittal plane are more likely to result in moderate to severe grade vision loss (VIS grade ≥ 2).

Expert-Moderated Peer-to-Peer Online Support Group for People With Knee Osteoarthritis: Mixed Methods Randomized Controlled Pilot and Feasibility Study

Background

Osteoarthritis (OA) is a major problem globally. First-line management comprises education and self-management strategies. Online support groups may be a low-cost method of facilitating self-management.

Objective

The aim of this randomized controlled pilot study is to evaluate the feasibility of the study design and implementation of an evidence-informed, expert-moderated, peer-to-peer online support group (My Knee Community) for people with knee OA. The impacts on psychological determinants of self-management, selected self-management behaviors, and health outcomes were secondary investigations.

Methods

This mixed methods study evaluated study feasibility (participant recruitment, retention, and costs), experimental intervention feasibility (acceptability and fidelity to the proposed design, including perceived benefit, satisfaction, and member engagement), psychological determinants (eg, self-efficacy and social support), behavioral measures, health outcomes, and harms. Of a total of 186, 63 (33.9%) participants (41/63, 65% experimental and 22/63, 35% control) with self-reported knee OA were recruited from 186 volunteers. Experimental group participants were provided membership to My Knee Community, which already had existing nonstudy members, and were recommended a web-based education resource (My Joint Pain). The control group received the My Joint Pain website recommendation only. Participants were not blinded to their group allocation or the study interventions. Participant-reported data were collected remotely using web-based questionnaires. A total of 10 experimental group participants also participated in semistructured interviews. The transcribed interview data and all forum posts by the study participants were thematically analyzed.

Results

Study feasibility was supported by acceptable levels of retention; however, there were low levels of engagement with the support group by participants: 15% (6/41) of participants did not log in at all; the median number of times visited was 4 times per participant; only 29% (12/41) of participants posted, and there were relatively low levels of activity overall on the forum. This affected the results for satisfaction (overall mean 5.9/10, SD 2.7) and perceived benefit (17/31, 55%: yes). There were no differences among groups for quantitative outcomes. The themes discussed in the interviews were connections and support, information and advice, and barriers and facilitators. Qualitative data suggest that there is potential for people to derive benefit from connecting with others with knee OA by receiving support and assisting with unmet informational needs.

Conclusions

Although a large-scale study is feasible, the intervention implementation was considered unsatisfactory because of low levels of activity and engagement by members. We recommend that expectations about the support group need to be made clear from the outset. Additionally, the platform design needs to be more engaging and rewarding, and membership should only be offered to people willing to share their personal stories and who are interested in learning from the experiences of others.

Trial Registration

Australian New Zealand Clinical Trials Registry ACTRN12619001230145; http://anzctr.org.au/Trial/Registration/TrialReview.aspx?id=377958

Can radiographic patellofemoral osteoarthritis be diagnosed using clinical assessments?

INTRODUCTION

The aim of this study was to determine whether participant characteristics and clinical assessments could identify radiographic osteoarthritis (OA) in individuals with clinically diagnosed, symptomatic patellofemoral osteoarthritis (PFOA).

METHODS

Participant characteristics and clinical assessments were obtained from 179 individuals aged 50 years and over with clinically diagnosed symptomatic PFOA, who were enrolled in a randomised trial. Anteroposterior, lateral, and skyline X-rays were taken of the symptomatic knee. The presence of radiographic PFOA was defined as "no or early PFOA" (Kellgren and Lawrence [KL] grade ≤1 in the PF compartment) or "definite PFOA" (KL grade ≥2). Diagnostic test statistics were applied to ascertain which participant characteristics and clinical assessments could identify the presence of definite radiographic PFOA.

RESULTS

A total of 118 participants (66%) had definite radiographic PFOA. Univariate analysis identified that older age (>61 years), female sex, higher body mass index (BMI) (>29 kg/m² ), longer pain duration (>2.75 years), higher maximum knee pain during stair ambulation (>47/100 mm), and fewer repeated single step-ups to pain onset (<21) were associated with the presence of definite radiographic PFOA. Multivariate logistic regression indicated that BMI, pain duration, and repeated single step-ups to pain onset were independently associated with radiographic PFOA and identified the presence of definite radiographic PFOA with an overall accuracy of 73%.

CONCLUSION

In individuals over 50 years of age with a clinical diagnosis of PFOA, higher BMI, longer pain duration, and fewer repeated single step-ups to pain onset increased the likelihood of radiographic PFOA. However, overall diagnostic accuracy was modest, suggesting that radiographic PFOA cannot be confidently identified using these tests.

Associations of foot and ankle characteristics with knee symptoms and function in individuals with patellofemoral osteoarthritis

BACKGROUND

Foot and ankle characteristics are associated with patellofemoral pain (PFP) and may also relate to patellofemoral osteoarthritis (PFOA). A greater understanding of these characteristics and PFOA, could help to identify effective targeted treatments.

OBJECTIVES

To determine whether foot and ankle characteristics are associated with knee symptoms and function in individuals with PFOA.

METHODS

For this cross-sectional study we measured weightbearing ankle dorsiflexion range of motion, foot posture (via the Foot Posture Index [FPI]), and midfoot mobility (via the Foot Measurement Platform), and obtained patient-reported outcomes for knee symptoms and function (100 mm visual analogue scales, Anterior Knee Pain Scale [AKPS], Knee injury and Osteoarthritis Outcome Score, repeated single step-ups and double-leg sit-to-stand to knee pain onset). Pearson's r with significance set at p < 0.05 was used to determine the association between foot and ankle charateristics, with knee symptoms and function, adjusting for age.

RESULTS

188 participants (126 [67%] women, mean [SD] age of 59.9 [7.1] years, BMI 29.3 [5.6] kg/m²) with symptomatic PFOA were included in this study. Lower weightbearing ankle dorsiflexion range of motion had a small significant association with higher average knee pain (partial r = - 0.272, p < 0.001) and maximum knee pain during stair ambulation (partial r = - 0.164, p = 0.028), and lower scores on the AKPS (indicative of greater disability; partial r = 0.151, p = 0.042). Higher FPI scores (indicating a more pronated foot posture) and greater midfoot mobility (foot mobility magnitude) were significantly associated with fewer repeated single step-ups (partial r = - 0.181, p = 0.023 and partial r = - 0.197, p = 0.009, respectively) and double-leg sit-to-stands (partial r = - 0.202, p = 0.022 and partial r = - 0.169, p = 0.045, respectively) to knee pain onset, although the magnitude of these relationships was small. The amount of variance in knee pain and disability explained by the foot and ankle characteristics was small (R²-squared 2 to 8%).

CONCLUSIONS

Lower weightbearing ankle dorsiflexion range of motion, a more pronated foot posture, and greater midfoot mobility demonstrated small associations with worse knee pain and greater disability in individuals with PFOA. Given the small magnitude of these relationships, it is unlikely that interventions aimed solely at addressing foot and ankle mobility will have substantial effects on knee symptoms and function in this population.

TRIAL REGISTRATION

The RCT was prospectively registered on 15 March 2017 with the Australia and New Zealand Clinical Trials Registry ( ANZCTRN12617000385347 ).

Design, Delivery, Maintenance, and Outcomes of Peer-to-Peer Online Support Groups for People With Chronic Musculoskeletal Disorders: Systematic Review

Background

Online support groups (OSGs) are one way for people with chronic diseases, their family or friends, and health professionals to communicate, gain information, and provide social support. As the number of peer-to-peer OSGs for chronic musculoskeletal conditions grows, it is important to gain insight into the different designs of groups available, who is accessing them, if and how they may be effective, and what strategies are being used to implement or increase consumer engagement.

Objective

The objectives of this systematic review of people with musculoskeletal conditions were to (1) describe the design features (functions, usage options, moderation, and expert input) of peer-to-peer OSGs, (2) describe the characteristics of the individuals using peer-to-peer OSGs, (3) synthesize the evidence on outcomes of participation, and (4) identify strategies used in the delivery and maintenance of OSGs.

Methods

A search comprising terms related to the population (people with musculoskeletal disorders) and the intervention (peer-to-peer OSGs) was conducted in 6 databases. Results were filtered from 1990 (internet inception) to February 2019. Studies identified in the search were screened according to predefined eligibility criteria using a 2-step process. Quantitative studies were appraised by 2 reviewers using the Risk Of Bias In Non-Randomized Studies of Interventions tool. Qualitative studies were appraised by 2 different reviewers using the Critical Appraisal Skills Programme checklist. Extracted data were synthesized narratively.

Results

We examined 21 studies with low to moderate risk of bias. Of these studies, 13 studies included OSGs hosted on public platforms, 11 studies examined OSGs that were conducted in English, and 6 studies used moderators or peer leaders to facilitate engagement. Studies either reported the number of OSG members (n=1985 across all studies) or the number of posts (range: 223-200,000). The majority of OSG members were females who were not full-time employees and with varied levels of education. There were no randomized controlled trials measuring the efficacy of OSGs. Qualitative and quantitative studies identified empowerment, social support, self-management behavior, and health literacy as primary constructs to measure OSG efficacy. Neutral or marginal improvement was reported in these constructs. Sharing experiences and a greater level of engagement appeared to have an important influence on OSGs efficacy. The extent to which members posted on the website influenced engagement.

Conclusions

Across a diverse range of designs, languages, included features, and delivery platforms, peer-to-peer OSGs for chronic musculoskeletal conditions attract predominantly female participants of all ages and education levels. The level of participation of a member appears to be related to their perceived benefit, health literacy, and empowerment. Future studies are needed to identify which design and maintenance strategies have superior efficacy and whether there are concomitant improvements in health outcomes for people with chronic musculoskeletal conditions resulting from participation in OSGs.

Trial Registration

PROSPERO International Prospective Register of Systematic Reviews CRD42018090326; https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42018090326

The FOOTPATH study: protocol for a multicentre, participant- and assessor-blind, parallel group randomised clinical trial of foot orthoses for patellofemoral osteoarthritis

INTRODUCTION

Patellofemoral (PF) osteoarthritis (OA) is a common and burdensome subgroup of knee OA, with very little evidence for effective treatments. Prefabricated foot orthoses are an affordable and accessible intervention that have been shown to reduce PF pain in younger adults. Similarities between PF pain and PFOA, as well as our pilot work, suggest that foot orthoses may also be an effective intervention for PFOA. The primary objective of this study is to compare the 3 month efficacy of prefabricated foot orthoses and flat shoe inserts in people with PFOA, on knee pain severity.

METHODS AND ANALYSIS

The FOOTPATH Study (FOot OrThoses for PAtellofemoral osteoarTHritis) is a multicentre, randomised, participant- and assessor-blinded superiority trial with two parallel groups, a 3 month observation period (pre-randomisation) and 12 month follow-up. 160 participants with a clinical diagnosis of PFOA will be recruited from three sites in Australia, and randomised to one of two groups (prefabricated foot orthoses or flat shoe inserts). The primary outcome is worst knee pain severity during a self-nominated aggravating activity in the previous week (100 mm visual analogue scale) at 3 months, with a secondary endpoint at 12 months. Secondary outcomes include global rating of change, symptoms, function, health-related quality of life, kinesiophobia, self-efficacy and use of co-interventions for knee pain. Blinded, intention-to-treat analyses of primary and secondary patient-reported outcomes will be performed, as well as economic analyses.

ETHICS AND DISSEMINATION

Ethical approval has been granted by La Trobe University's Human Ethics Committee and The University of Queensland's Medical Research Ethics Committee. Study outcomes will be disseminated via peer-reviewed journals, conference presentations targeting a range of healthcare disciplines and an open access website with clinician resources.

TRIAL REGISTRATION NUMBER

ANZCTRN12617000385347; Pre-results.

The psychological features of patellofemoral pain: a cross-sectional study

Background and aims:

Patellofemoral pain (PFP) is a prevalent and debilitating musculoskeletal condition, considered to have a mechanical aetiology. As such, the physical impairments associated with PFP are well documented and have helped characterise different physical phenotypes. But little is known about the relationship between PFP and psychological well-being. In this study, we aimed to: (1) compare psychological profiles between groups with and without PFP; (2) compare psychological profiles and condition severity between PFP subgroups; and (3) explore relationships between psychological factors and their contribution to disability. We expected to find higher levels of psychological impairment, especially kinesiophobia and catastrophizing in the PFP group. We also expected to identify a sub-group for who worsening levels of disability correspond with worsening psychological well-being.

Methods:

One hundred participants with PFP (72 females, mean±SD age 27±5 years, BMI 25.3±4.8 kg/m2) completed measures of pain, disability, and psychological features (kinesiophobia, catastrophizing, anxiety and depression). Fifty controls, matched by sex, age and activity level (36 females, age 27±5 years, BMI 22.9±4.5 kg/m2) also completed psychological measures. The Knee injury and Osteoarthritis Outcome Score (KOOS) was used to cluster PFP participants (K-means cluster analysis) into more and less severe sub-groups. Differences between the control and PFP groups were analysed using t-tests, analysis of variance, Mann-Whitney U-tests or χ 2 tests as appropriate (p<0.05). Pearson correlations were used to explore relationships between psychological measures. Backward stepwise regression (p out >0.05) evaluated how the psychological factors potentially relate to disability.

Results:

Psychological features did not differ between PFP and pain-free groups. But differences were apparent when the PFP cohort was subgrouped. Compared to controls, the more-severe group had significantly higher levels of depression (MD 1.8, 95% CI 0.8–2.8; p≤0.001) and catastrophizing (MD 5.7, 95% CI 2.4–9; p≤0.001). When compared to less-severe cases, the more-severe group also demonstrated significantly higher levels of kinesiophobia (MD 4.3, 95% CI 2.1–6.5; p≤0.001), depression (MD 1.5 95% CI 0.5–2.6; p=0.01) and catastrophizing (MD 4.9, 95% CI 1–8.8; p=0.01). The weakest relationship between psychological factors was found between kinesiophobia and anxiety (r=0.29; p=0.02). While the strongest relationship existed between depression and anxiety (r=0.52; p≤0.001). Both kinesiophobia (β −0.27, 95% CI −0.265 to −0.274) and depression (β −0.22, 95% CI −0.211 to −0.228) were associated with disability as defined by the KOOS in the regression model (R 2=0.17, p≤0.001).

Conclusions:

Those with more-severe PFP-related disability have higher levels of psychological impairment than less-severe cases. Kinesiophobia seems to stand as an important factor in the experience of PFP, because it was elevated in the PFP group, significantly differed between the PFP sub-groups and contributed to explaining disability. Contrary to our hypothesis, levels of catastrophizing in the PFP group and severe sub-group were low and seemingly not important.

Implications:

These findings draw attention to psychological factors to which clinicians assessing PFP should show vigilance. They also highlight psychological impairments that might be worthwhile targets in optimising PFP management.

Longitudinal recovery following distal radial fractures managed with volar plate fixation

AIMS

To synthesise the literature and perform a meta-analysis detailing the longitudinal recovery in the first two years following a distal radius fracture (DRF) managed with volar plate fixation.

MATERIALS AND METHODS

Three databases were searched to identify relevant articles. Following eligibility screening and quality assessment, data were extracted and outcomes were assimilated at the post-operative time points of interest. A state-of-the-art longitudinal mixed-effects meta-analysis model was employed to analyse the data.

RESULTS

The search identified 5698 articles, of which 46 study reports met the selection criteria. High levels of disability and impairment were reported in the immediate post-operative period with subsequently a rapid initial improvement followed by more gradual improvement for up to one year. The results highlight that the period associated with the greatest physical recovery is in the first three months and suggest that the endpoint of treatment outcomes is best measured at one year post-surgery.

CONCLUSION

Clinically meaningful improvements in outcomes can be expected for 12 months, after which progress plateaus and reaches normal values. This paper adopted a novel approach to meta-analyses in that the research question was of a longitudinal nature, which required a unique method of statistical analysis. Cite this article: Bone Joint J 2017;99-B:1665-76.