The Effects of Custom-made Foot Orthoses on Foot Pain, Foot Function, Gait Function and Freeliving Walking Activities in Psoriatic Arthritis (PsA) Patients: a Pre-experimental Trial

Introduction: Foot involvement is a major concern in psoriatic arthritis (PsA) as it can lead to severe levels of foot pain and disability as well as reduced mobility and quality of life. Previous studies have shown moderate ecacy in reducing foot pain and disability in rheumatoid arthritis patients with the use of custom-made foot orthoses (CFO). However, evidence on the ecacy of CFO in PsA patients is lacking. Objectives: Explore the effects of CFO on foot function, foot and lower limb pain, gait function, and freeliving walking activities (FWA) in PsA patients. Methods: A Pre-experimental study including 20 PsA patients (mean age: 54.10 ± 9.06 y and disease duration: 11.53 ± 10.22 y), was conducted. All the participants received and wore CFO for a 7-week period. Foot and lower limb pain and foot function were measured before and after the intervention using the numerical rating scale (NRS) and the foot function index (FFI). Gait function was assessed from gait spatiotemporal parameters (STPs) extracted during a 10-meter walk test with an gait analysis system (Mobility Lab). Freeliving walking activities (step count, freeliving cadence, time spent in different ambulatory physical activities (APA)) were recorded over 7 days using accelerometer data collected from an instrumented sock worn during waking hours. Results: PsA patients reported severe baseline levels of foot pain (54.46 ± 14.58 %) and disability (46.65 ± 16.14%) on the FFI. Statistically and clinically signicant improvements with large effect sizes (Cohen’s effect size > 1, p<0.005) in foot pain and foot function were observed after 7 weeks of CFO use. A signicant correlation (r=-0.64, p<0.01) between CFO wear time after the adaption period and foot function on the FFI at 7 weeks was observed. However, no signicant changes were demonstrated for gait STP nor for free-living walking activities after 7 weeks of CFO use. Conclusion: Results support the clinical and biomechanical plausibility of using CFO with PsA patients to reduce pain and improve foot function. Larger and controlled studies are needed to conrm these ndings and a multidisciplinary approach including the prescription of exercise therapy and physiotherapy in combination with CFO could be relevant to improve STP and promote APA in PsA patients. APA: ambulatory physical activity; BMI: body mass index, CRP: c reactive protein; CFO: Custom foot othoses; DMARD: disease-modifying anti-rheumatic drugs; FFI: foot function index; FWA: Freeliving walking activities; GCT: gait cycle time; MCID: minimal clinically important difference; NRC: numerical rating scale; PA: physical activity; PsA: psoriatic arthritis; STPs: spatiotemporal parameters; 10MWT: 10-meter walk test.

meet the recommendations of the international guidelines for PA [18,19,20]. Inherently, gait and walking are closely intertwined. For instance, a reduced gait speed is a signi cant predictor of decreased daily walking activity and it could therefore have major daily life and PA impacts [21,22]. The impact of PA and walking on health outcomes is well established [23,24]. It has also been reported that PA may be bene cial on disease activity, muscle strength, fatigue, pain, and quality of life in PsA patients [25].
However, a recent systematic review demonstrated low PA levels in PsA patients, which could be detrimental on PsA associated comorbidities (i.e. cardiovascular health) and on the disease per se [25].
Importantly, a close relationship between foot pain and gait function defeciencies has been demonstrated [28,29]. This suggests that interventions aiming at improving foot pain and function could eventually improve gait function and consequently promote walking activity and overall PA in PsA patients.
PsA management is mainly based on pharmacological treatments including disease-modifying antirheumatic drugs (DMARDs), Janus Kinase inhibitors (JAKi) and biological therapies which have been proven e cient on symptoms control and disease activity reduction [30][31][32]. Remarkably, in ammation and foot pain as well as the related disability can still be observed in a large proportion of patients on pharmacological therapy [15,17] [33][34][35]. From a physiopathological perspective, mechanical stress and trauma have been suggested as joints and soft tissues in ammation triggers in PsA [36][37][38]. This could partly explain the high prevalence and persistence of foot and ankle problems, considering the weight bearing function and the continuous mechanical stress applied to the foot during daily living activities [39]. In this context, experts stress out the need for targeted therapies addressing biomechanical abnormalities in the foot to reduce mechanically triggered in ammation and pain in PsA patients with persistent foot problems.
Among such therapies, custom foot orthoses (CFO) are medical devices designed spec cally to correct the foot position, realign the lower limb articulations, redistribute pressures, and o oad painful structures [40]. CFO are common practice in patients with foot problems associated to rheumatic diseases. In patients with rheumatoid arthritis (RA), a disease characterized by high prevalence and severity of foot problems similar to those reported in PsA, the use of CFO showed positive effects on foot pain and disability [41][42][43][44][45][46]. However, to our knowledge, there are no records in the literature on the e cacy of CFO in PsA patients. In addition, despite the similarities of foot problems, RA and PsA are different entities with different clinical presentations making direct extension of results from RA to PsA inadequate [47]. Therefore, the primary objective of this study is to explore the effects of a 7-week CFO use period on foot function in PsA patients. The secondary objectives are to explore the effects of CFO on foot pain, the relationship between the CFO wearing time and foot pain and function, and the effects of CFO on gait function (STP) and free-living walking activities (FWA) in PsA patients.

Study design and participants
A pre-experimental study with pre-test/post-test design was conducted. Twenty-two PsA patients were recruited from the rheumatology out-patient clinics at the Université de Sherbrooke Hotel Dieu University Hospital (CHUS). Inclusion criteria were: between 20 and 70 years of age, a con rmed diagnosis of PsA by a trained rheumatologist, moderate to severe and recurrent foot pain and stable medication over three months preceding the recruitment. Patients with diabetes, neurological disease or any musculoskeletal disease that could impact the normal gait pattern, who received an intra-articular corticosteroid injection or any conservative foot treatment such as foot orthoses/footwear intervention within the past three months were excluded. The study was approved by the CIUSSS de l'Estrie-CHUS Ethics Board, and all the participants gave their informed consent prior to data collection. Data collection procedure Demographic and baseline characteristics were assessed on a rst data collection session (T1) at Sherbrooke's Research Center on Aging. Foot and lower limb pain, global pain, and foot function were then measured using self-reported questionnaires. Afterwards, participants were tted with inertial measurement units (IMUs) and performed a 10 meters gait test at a self-selected speed. Participants were then examined by a podiatrist who performed foot casts and designed the CFO. At the CFO delivery visit (T2), the participants' foot and lower limb pain, and foot function were evaluated for a second time. This second pre-intervention measurement was planned considering the T1 -T2 delay as well as the uctuating nature of PsA patients' pain. CFO were then dispensed and adjusted if necessary, to improve t and comfort. The participants were then directed to wear the CFO for a 7-week period. They were provided with a diary to record daily CFO wearing time, weekly foot pain intensity and any changes in medication during the intervention period. At a nal follow up visit at the end of the 7-week period (T7), foot and lower limb pain and foot function were re-assessed and the instrumented 10 meters gait test was repeated. Free walking activities (FWA) were recorded over seven consecutive days before the CFO rst use, and after the seven-week intervention period using an instrumented sock (inertial sensors module positioned at the ankle and embedded in a sock) worn daily during waking hours by the participants.

Intervention
Functional foot orthoses were custom-made for each participant. The CFO were designed by the same experienced podiatrist (GBC) based on a detailed clinical and biomechanical assessment. Foot scans were obtained with the Occipital Structure Sensor 3D scanner (Occipital Inc), the patient in a prone position while the subtalar joint was held in a neutral position. The obtained scan was edited and smoothed using the MSoft software (Techmed3d Inc, Levis, QC, Canada). The prescriptions/corrections were de ned by the podiatrist before manufacturing. Three-quarter length CFO were 3D printed in a rigid material (Nylon) using a MultiJet Fusion 3D printer (HP, Palo Alto, CA, USA). The thickness of the CFO ranged between 2 and 3.4 mm depending on the participants' medial arch weight and height. Most of the orthoses included a medial arch support, a heel stabilizer and a metatarsal pad. The degrees and types of corrections added to the CFO were determined speci cally for each participant based on the clinical examination and were adjusted at the CFO delivery according to each patient's comfort and tolerance.
The participants were taught to wear the CFO progressively during the rst two weeks to allow for the lower limbs' muscles and structures to adjust to the CFO, and to wear the orthoses for the next 5 weeks, 7 days a week as often as they could. To document adherence to the CFO, all the participants completed a diary to report the daily wearing time in hours. In addition, the participants were advised to wear the CFO with adapted shoes after the general characteristics (e.g. heel height, malleability of the sole etc.) of such shoes were explained.

Outcomes measures
Sociodemographic, baseline characteristics and control variables: Sex, age, body mass index, professional occupation, and previous history of foot injuries were assessed using a self-reported questionnaire. Disease duration and C-reactive protein levels (CRP) at baseline were obtained from the patients' medical records. Foot type, foot deformities and foot pain sites were obtained from the podiatrist clinical examination records. Changes in medication that occurred during the CFO intervention period were recorded by the participant in a diary.
Foot function: Foot function was measured using the Foot function index (FFI). The FFI is a valid, reliable and responsive self-reported questionnaire widely used in studies of foot and ankle disorders previously used to evaluate CFO effects on foot function [48,49]. The FFI is composed of 23 items divided into three subscales measuring foot pain (FFI-P), foot disability (FFI-D) and foot-related activity limitation (FFI-AL).
Each item of the FFI is recorded on a 0 to 10 numeric rating scale (NRS) allowing for subscale and total scores calculation. A sub-scale score is obtained by totaling the sub-scale items' scores and dividing the total by the sub-scale maximum possible total [48]. The total score is obtained by adding all subscales nal percentages and dividing by the total number of subscales. The values reported are presented as percentages ranging between 0 and 100 %, with higher values indicating greater pain, disability, and activity limitation. The minimal clinically important difference (MCID) for the FFI total score was found to be 7 points in patients with plantar fasciitis [50].
Foot pain: The intensity of foot pain was measured using a zero-to-ten NRS considered a valid, reliable, and responsive tool for pain intensity assessment [51]. The MCID for patients with chronic pain is a 2 points change [52]. Participants were asked to circle the number between 0 and 10 that better matched their average foot pain intensity in the seven days preceding data collection. Foot pain was assessed more in detail (e.g. pain walking with foot orthoses, pain walking with shoes, pain at the end of the day etc.) with the FFI pain sub-scale. To monitor the evolution of weekly foot pain intensity during the intervention period, the patients were asked to record it at the end of each week, in a diary using 0 to 10 NRS. Foot pain and foot function were measured at three time points, at rst data collection (T1), at CFO delivery (T2), and after the 7-week intervention period (T7). As there were no statistical differences in the FFI and NRS scores between T1 and T2, the average of these two time points was used as a baseline measure while the last time point (T7) represented the nal measure. Global and Lower limb pain: Global pain and pain at the knee, hip, and lower back pain, were measured at T1, T2 and T7 using the NRS (0, no pain -10, worst imaginable pain). Similarly to foot pain and foot function, data obtained during the rst two time points was used as baseline.
Gait function: Gait function was assessed using an instrumented gait analysis system. Gait spatiotemporal parameters (STPs) including cadence, gait cycle duration, gait speed, stride length, double support, swing time, foot strike angle, and stride time variability, were recorded using the Mobility Lab system (APDM Wearable Technologies) during 10 meters walk test (10MWT). Mobility Lab is a research grade system proven to accurately and reliably estimate STP [53][54][55]. Mobility Lab uses a set of six OPAL inertial measurement units (IMUs) and a software that allows for an automated and easy extraction of STP. All the participants performed three 10MWT trials over a 14-meters straight walkway with the Mobility lab's IMUs xed with elastics straps on the chest, the lower back, both wrists and feet. The 10MWT trials were performed at the participants' comfortable speed and the average of the three trials was calculated.
Freeliving walking activities (FWA): FWA including step count, freeliving cadence, and time spent in ambulatory physical activity (APA) intensity-based categories were measured using an instrumented sock (Sensoria Inc, Redmond, WA, USA) with an embedded 9 axis IMU positioned at the ankle. The instrumented sock connects automatically, without any manipulation needed from the participants, via Bluetooth to a smartwatch (Apple Watch, series 3) where the raw inertial measures of motion (3D accelerometer) are stored and then transferred for data reduction and processing to extract walking activities speci c outcomes (see Additional le for signal processing and steps identi cation description).
Step count was assessed as the total number of steps per day. Besides, to describe the steps accumulation pattern throughout the day, the number of steps was also assessed for each of the following active events' categories determined based on the number of consecutive steps within each category: 0 to 20 steps, 20 to 60 steps, 60 to 120 steps, 120 to 300 steps, 300 to 600 steps, and more than 600 steps. Free-living cadence included cadence averaged for total wearing time (mean cadence/day). Moreover, as cadence may vary depending on the number of consecutive steps performed, it was also assessed for each of the above active events' categories. APA intensity-based categories were determined using cadence data and included stepping activities (0 to 59 steps/min), slow walking (60 to 79 step/min), medium walking (80 to 99 step/min), moderate intensity APA (MAPA) (100 to 119 step/min) and vigorous APA (VAPA) ( > 120 step/min), as previously de ned [56,57]. The time spent in each of the last three categories was calculated only for more than 60 steps events referred to as purposeful walking.
Free-living walking activities were assessed over a 7 consecutive days period before and after the intervention during which, the participants were instructed to wear the instrumented sock during waking hours while they perform daily activities. Only valid days de ned as days with at least 8 hours of recordings, and only valid participants with a minimum of one valid day were included in the analyses.
Adherence to the instrumented sock use was tested by the system itself and by asking the participants to record the daily wearing time in a diary. To avoid in uencing the participant's physical activity, they were only told that the study would assess the effects of CFO on foot pain, foot function and gait STP.

Sample size and statistical analysis
Given the exploratory nature of this study, the sample size was calculated assuming a large effect size (0.8) for the pre-post differences in the FFI total score. Using a two tailed paired T-test, a signi cance level α of 5%, a power of 90%, and assuming a 15% dropout rate, a total sample size of 22 participants was required. The Shapiro-Wilk test was used to examine data distribution and parametric and non parametric statictical tests were used depending on data distribution. Paired t-tests (or the Wilcoxon signed-rank test) were used to assess the differences in foot pain, foot function, global and lower limb pain between baseline and T7 and Cohen's effect size was calculated to quantify the magnitude of these differences. A one-way repeated measures ANOVA was conducted to determine whether there was a statistically signi cant difference in weekly recorded foot pain intensity and CFO wearing time over the 7-week intervention period. Spearman correlation coe cients were calculated to assess the relationships between the CFO wearing time, foot pain and foot function at T7. As a two-week adaptation period was required, correlation analyses were performed using the average CFO wearing time per week calculated for the last 5 weeks of the intervention period. Correlation coe cients were considered weak, moderate, and strong for values between 0.1 and 0.3, 0.3 and 0.5, and > 0.5, respectively [58]. The pre-post differences in gait function and freeliving walking activities were assessed using the Wilcoxon signedrank test. The signi cance level was set at 0.05. The statistical analyses were performed using SPSS Version 26 (IBM statistics Corporation, Armonk, NY).

Results
Demographic and baseline clinical characteristics: Of twenty-two PsA patients initially recruited for the study, one patient was excluded after podiatry examination because the CFO was considered not clinically appropriate. Another patient was excluded because of an acute outbreak during follow up. The remaining 20 participants were included in the analyses. Demographic and baseline clinical characteristics are presented in Table 1. The study sample was composed of 5 males and 15 females with a mean age of 54.10 ± 9.06 years, a mean BMI of 29.3 ± 4.5 kg/m 2 and a mean disease duration of 11.53 ± 10.22 years. The CRP baseline levels were within the normative ranges for 75 % of the patients. Two participants recorded changes in medication during the intervention period ( Table 1). The rst participant started taking NSAIDs at the end of the fourth week for generalized body pain, and the second received a new biological therapy starting at week 4.
Foot function, global and lower limb pain: Figure 1 presents the FFI (1a) and NRS (1b) scores at baseline and after 7 weeks of using the CFO. All the FFI sub-scores and total score decreased at T7 indicating a signi cant improvement in foot pain ( Seventeen participants recorded the weekly foot pain intensity and CFO wearing time using the provided diaries. The CFO daily wearing time and the evolution of foot pain are presented in Figure 2a and 2b. As it can be observed, there was a tendency for foot pain to decrease during the rst 4 weeks, a plateau was observed after this period. The same tendency was also observed for wear time which increased progressively during the rst 4 weeks and remained almost stable afterwards.  (Figure 3).
Gait function: Spatiotemporal gait parameters measured before and after the intervention period are presented in Table 2. Results show slight but signi cant increase and decrease in double support time and swing time, respectively, at T7. However, for all the remaining STP including cadence, gait cycle duration, gait speed, stride length, and foot strike angle, there were no signi cant differences at T7 compared to baseline. Similarly, no signi cant changes were reported for effects of CFO on stride time variability.  Freeliving walking activities (FWA): Freeliving walking activities are presented in Figures 4, 5 and 6. Out of the 20 participants, four were excluded from the FWA analyses because they didn't accumulate at least one valid day of recordings (i.e., recording time > 8 hours) at baseline and T7. Consequently, the results presented in this section include data from 16 participants. On average, the patients accumulated a total of 10.79 ± 1.18 hours and 9.99 ± 1.02 hours of recorded time at baseline and T7, respectively, with no signi cant differences between baseline and T7 (p=0.09). The average active time was 3.9 ± 1.7 hours at baseline and 3.5 ± 1.6 hours at T7 (p=0.08). The average number of steps per day was 6460 ± 2818 step/day at baseline, and 5836 ± 3079 step/day at the end of the intervention period with no signi cant differences between the two time points (p=0.18) (Figure 4. a). The pattern of accumulating these steps is presented in Figure 4. b. As it can be observed, the total number of steps was accumulated in short active events composed of 20 to 300 consecutive steps (Figure 4. b). The daily average freeliving cadence was 96.68 ± 5.85 steps/min and 96.03 ± 5.58 steps/min before and after the intervention, respectively (p=0.79) (Figure 5. a). Results show that cadence increases considerably for active events of more than 20 steps compared to those of less than 20 steps (Figure 5.

Discussion
The primary purpose of this study was to explore the effects on foot function of wearing CFO for a 7week period in PsA patients. As a secondary objective we sought to assess the effects of CFO on foot pain, gait function (instrumented gait STP analysis) and FWA, and to assess the relationship between the CFO wearing time, foot pain and function in PsA patients with foot involvement.
Foot pain and foot function: The main ndings showed a signi cant improvement in foot pain (NRS and FFI) and foot function (FFI) after the use of CFO for a 7-week intervention period. The FFI total score decreased by 22.30 % and a large effect size (d= 1.25) was reported. Interestingly, this improvement was above the MCID reported for the FFI total score meaning that the improvement was clinically signi cant and it was observed after only 7 weeks of wearing time.
Research in rheumatoid arthritis (RA) patients has shown some e cacy of CFO on foot pain which is consistent with our ndings [41,42,43]. However, con icting results regarding the of CFO effects on foot function have been reported. Indeed, while some studies endorsed the effectiveness of CFO on foot function [44,46], others reported no signi cant effects for this outcome [59][60][61]. Several factors could explain such discrepancies. First, the type of orthoses can differ between the studies according to the targeted therapeutical aim (e.g. palliative vs functional). Moreover, foot casting and manufacturing materials and techniques can signi cantly vary from one study to another depending on the latest advances in technologies and on the national practice guidelines making between-studies comparison di cult. In our study, we used functional CFO in all the participants. The materials and techniques of foot casting and manufacturing were standardized for all the participants. However, the degrees and types of corrections were speci c for each participant's needs and tolerance. The CFO were also adjusted at delivery and during the following two weeks if needed, to increase comfort and acceptance. This could explain the positive effects on pain and foot function observed in this study. Furthermore, there was no consensus regarding the intervention durations and the control conditions in the previous studies. Some trials used placebo orthoses while others used shoes only as a control condition. As the former are usually made from thick and cushioned materials they cannot be considered as having no therapeutic effects, thereby attenuating the CFO effects. In contrast, using shoes-only as control could overestimate the effect of CFO [62].
Importantly, CFO wearing time was not documented in all the previous studies which can compromise their conclusions as a lack of effectiveness could partially be related to insu cient use of the orthoses. In our study, a su cient wearing time (8hours/day) was reported and a strong correlation between the CFO wearing time and the FFI total score was demonstrated. These ndings are similar to those reported in the few studies in RA patients that did report wearing time [44,46,60]. This observation highlights the relevance of monitoring this variable in future studies and the importance of continuous and regular wearing of the CFO.
Also of note, the intervention characteristics, sampling groups and individual characteristics such as age, sex distribution, disease duration, and baseline levels of foot pain and disability are also important considerations as they may in uence the effectiveness of CFO. For instance, data from a previous study [63] showed that shorter disease duration, younger age and higher baseline values of foot pain and disability predicted outcomes after CFO use. In our study, the participants had a mean age of 54.10 ± 9.06, a median disease duration of 6 years, a high prevalence of foot pain and severe levels of disability at baseline. It is therefore not unexpected that the CFO was helpful relieving pain and improving perceived foot function. In addition to foot pain and disability, the ndings showed a signi cant decrease in global and hip pain, with large and medium effect sizes reported for these variables, respectively. Knee and lower back pain also improved but the differences did not reach the signi cance level. Although there is evidence that CFO could have bene cial effects on lower limb pain [64-67], further studies using speci c tools to assess knee, hip and lower back pain in details, are needed to con rm the bene cial effects of CFO on these variables.
Gait function: Gait STPs are indicators of functional capacity, overall health and autonomy, which makes them the mostrelevant and most often measured biomechanical parameters for gait analysis in healthy and pathological populations [13,73]. We noticed a lack of scienti c evidence regarding the effects of CFO on gait STPs in patients with arthritic foot diseases. For instance, studies in RA patients, showed only a minimal or no effect of CFO for increasing gait speed, cadence and stride length [43,70,71] and some other studies involving RA patients showed that wearing CFO allowed for an improvement of gait stability [46,72].
Our ndings showed that gait STPs were altered compared to the normative values reported for healthy adults [73,74]. However, there were no signi cant changes in STP and gait stability after 7 weeks of CFO use with the exception for a negligible effect on double support and swing time. These results could be potentially attributed to short-term nature of our intervention period and the long disease duration reported in this study. In fact, gait impairments demonstrated in PsA patients may be acquired over years and as the disease progresses, the patients can develop antalgic walking strategies that would necessitate longer intervention periods to improve. Aditionnaly, the patients could have developed fearavoidance beliefs which have been reported as signi cant predictors of poorer response to CFO in other patient populations [75]. Another possible explanation for the lack of STP response could be that PsA is responsive to CFO treatment in earlier stages of the disease. Therefore, long-term studies including patients with early disease would be more relevant to properly assess the effects of CFO on gait STP.
Additionally, although the 10MWT is reliable and responsive in patients with major gait impairments secondary to neurological disease such as stroke and Parkinson [76], it might have not been responsive in this study because of the relatively milder gait impairment related to PsA and the short duration of the intervention and follow up period [77]. Longer time/distance tests should also be considered when assessing gait in PsA patients as gait impairements reported in this population could be attributed to commun features of in ammatory arthritis patients such as muscle weakness, reduced range of motion, and deformities [78, 79,80] and may be more evident during sustained walking activities.
Freeliving walking activities: The bene ts of physical activity on health are undeniable. Results from a recent systematic review [25] suggest that physical activity may have positive effects on disease activity, muscle strength, fatigue, pain, quality of life and cardiovascular risk in PsA patients. Therefore, promoting physical activity in this population should be a health care priority. Among physical activities, walking is a practical, accessible and low skill mode of physical activity that also plays a major role in occupational and social activities in everyday living [81]. This study has limitations that should be mentioned. First, there was no control group to minimize the placebo effect and to control for the natural evolution of the disease. However, as this is the rst study to assess the effects of CFO in patients with RP, the results are just exploratory but encouraging for performing RCTs. Second, the follow-up duration may have been too short to show changes in outcomes such as gait parameters and FWA that may necessitate longer time to improve. Moreover, the small sample size and the high proportion of female participants which is not representative of the gender distribution usually seen in PsA, would compromise the generalizability of our ndings.

Conclusion
The use of CFO for 7-week period improved foot pain and related disability in PsA patients. This suggests that CFO may be used in conjunction with pharmacological therapy to relief pain and improve perceived foot function. However, the results should be con rmed in larger and controlled studies. The effects of CFO on gait function and freeliving walking activities were not obvious in this study due to the small sample size, short follow-up period and environmental factors that have not been measured. Future RCTs with lager sample sizes and longer intervention durations are needed for any conclusions to be drawn.

Declarations
Ethics approval and consent to participate The study was approved by the the Institutional Review Board of the CIUSSS de l'Estrie-CHUS (2019-3182), and all the participants gave their informed consent to participate in the study.

Consent for publication
Not applicable Availability of data and materials The dataset used and analyzed during the current study is available from walha.roua@usherbrooke.ca on reasonable request.

Competing interests
The authors declare no con ict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Funding
This research was funded by the Canadian MSK Rehab Research Network and supported through a donation from Janssen Pharmaceutical compagnies to Fondation pour la recherche et l'enseignement en orthopédie de Sherbrooke.
Authors' contributions RW and PB conceived the study. RW collected and analyzed the data and wrote the rst full draft of the manuscript. PB assisted with the analysis of data. PB, GBC, PD and NG all reviewed the manuscript drafts. All authors read and approved the nal manuscript.  Correlation between weekly CFO wearing time averaged for the last 5 weeks of the intervention period and the FFI subscales and the total score recorded at T7. Step count. (a) Daily averaged steps and (b) number of steps the different step-based active episodes categories. Figure 5