Our findings indicate that a full-scale RCT of contoured foot orthoses versus flat insoles for adolescents with PFP would not be feasible with the protocol used in this study, based on our a priori criteria to inform feasibility. Although our observed recruitment rate (1.2 participants per week) exceeded the pre-specified criterion of 1 participant per week, parameters for minimum adherence with shoe insert wear and drop-out rate were not met. Of the 15 participants (42%) who provided adherence data, only 7 (47%) met minimum adherence of 2 hours per day, 5 days per week. At 3 months, 28% of participants had dropped out (11/36), of which 4 were related to the COVID-19 pandemic. There was one minor, transient adverse event reported. Participants perceived both interventions to be credible, and results of the Credibility and Expectancy Questionnaire demonstrated success of blinding. The method of patient-reported outcome collection was feasible and acceptable in this group.
Our recruitment strategy involved recruiting adolescent volunteers from the community via advertisements, as well as contacting adolescents with PFP from a pre-existing database. Combined, this strategy led to a recruitment rate of 1.2 participants per week over 32 weeks. However, we acknowledge that a future full-scale RCT may not have access to such a database, and thus is likely to rely solely on recruitment from the community. If the 6 participants who were recruited from the pre-existing database are not considered, our recruitment rate for the remaining 30 participants was 0.94 participants per week. This does not meet our a priori criterion (18). Therefore, a full-scale trial in this population may need to factor in a longer recruitment period to meet sample size targets, or be run across multiple sites or use a broader recruitment strategy (e.g., through schools) to meet a recruitment rate of 1 participant per week. Of the advertising methods used, social media was most successful, accounting for 90% of participants recruited from advertising (27/30). Future studies recruiting adolescents with PFP should consider using social media as their primary advertising method.
Our observed percentage of total volunteers entering the study (12.9%) is in line with previous PFP RCTs. Rathleff et al (13) reported a 16.7% inclusion rate of adolescents with knee pain at screening, who were recruited through schools. Collins et al (11) included 11.7% of adults volunteers with PFP, who were recruited through community advertising. The most common reasons for excluding volunteers were that they were already wearing foot orthoses (n=70), had sustained a previous lower limb injury (n=29), or had anterior knee pain from another source (e.g. patellar tendinopathy, Osgood-Schlatter disease) (n=18). Interestingly, 97 out of 260 (37%) adolescents who responded to advertisements and completed the initial online screening form did not respond to follow-up email or telephone communications. Therefore, it is not possible to know how many of these volunteers were potentially eligible for physical screening. Our observed rate of non-response is substantially higher than the proportion of adults with PFP who did not respond to follow-up communications after volunteering for a previous study (182/1530, 11.9%) (11). Although the reasons for this are unclear, adolescents are likely to have different considerations and preferences for communicating or engaging with researchers and barriers to participation, which need to be explored further.
We evaluated two aspects of adherence – logbook completion and time wearing the allocated inserts. 42% (15/36) of participants returned their logbooks, suggesting that data from the logbooks may not be representative of the whole group. Because the logbooks were the primary source of data relating to insert wear time, this also suggests that our insert adherence data may not be representative. 38% (3/8) of participants in the contoured foot orthoses group reported wear time that met our predefined criteria for minimum adherence, compared to 57% (4/7) of participants from the flat insole group. This is substantially lower than a previous study in older adults with patellofemoral osteoarthritis (mean age 60 years), where logbooks were completed by 69% of participants and insert wear time was, on average, 37 hours per week (37). Our findings are likely to overestimate insert wear, as participants who completed and returned their logbooks are also more likely to be adherent with other aspects of the study, such as wearing their allocated insert. Further research is needed to determine the most effective methods of facilitating and monitoring adherence in this population. For example, app- or web-based methods may be more acceptable to adolescents with PFP due to the high use of electronic devices in this demographic (38). Notwithstanding this, the low adherence rates that we observed may indicate that shoe inserts are not a viable intervention in this population. This requires further exploration to determine the barriers and facilitators to wearing shoe inserts, as well as footwear preferences of adolescents with PFP. Alternatively, there may be other factors that affected adherence with wearing the shoe inserts, such as the COVID-19 pandemic. Participants were in lockdown (confined to their homes) during the study period and were undertaking home schooling. This likely resulted in participants substantially reducing their shoe wear time, compared to normal school attendance and sport participation.
We set an a priori drop-out rate of ≤20% to inform feasibility (18). Overall, 28% (10/36) of participants were lost to follow up at 3 months. 31% (11/36) of participants did not attend any physiotherapy appointments, and therefore did not receive their allocated intervention. It is important to explore the reasons why these adolescents dropped out or did not attend physiotherapy. One consideration is the impact of the COVID-19 pandemic. Four participants were due to schedule their initial physiotherapy appointments when COVID-19 restrictions were coming into effect in Queensland. If these participants are not considered, then the total loss to follow-up within the study is 17% (6/36), which is within our pre-specified feasibility criterion, and the number of participants not attending any physiotherapy appointments is 7 (19%). However, it is likely that there were other barriers to our participants attending physiotherapy appointments, such as relying on parents for transport and the geographic location of the clinics involved in our trial. These need to be addressed if a full-scale RCT is to be feasible.
There were minimal differences between groups on the Credibility and Expectancy Questionnaire. At baseline, prior to randomization, median scores for the two groups were identical or within 1 response category. After receiving the allocated intervention, the group who received flat insoles demonstrated lower scores on all items of the Credibility and Expectancy Questionnaire, although median scores for the two groups were within 1.5 categories. The similarity in scores between groups after receiving the intervention, and between baseline and follow up for item 6 (By the end of your therapy, how much improvement in your symptoms do you really feel will occur?) indicates success of blinding. No other studies have assessed credibility and expectancy outcomes in adolescents with PFP. However, success of blinding was demonstrated in a study investigating lower limb injury risk in military recruits randomized to receive contoured prefabricated foot orthoses or flat insoles (39). This suggests that flat insoles are an appropriate comparator for foot orthoses, including for adolescents with PFP.
Our findings also indicate that adolescents with PFP have minimal adverse events associated with wearing shoe inserts. Only one participant (who received flat insoles) reported minor discomfort and skin redness due to rubbing of the insert. This resolved in two days, and the participant was able to continue wearing the intervention. Similar to previous studies involving shoe inserts for knee pain (11, 16, 40), we found that contoured foot orthoses, as well as flat shoe inserts, were tolerated well by participants.
There are two key limitations of this study. Firstly, there was a higher drop-out rate observed in the flat insole group at 3 months (42% in flat insole vs. 12% in foot orthoses group). However, we observed similar outcomes on the Credibility and Expectancy Questionnaire after participants had received their allocated shoe inserts. This suggests that there may be other factors affecting retention in this group. Secondly, only 42% of participants returned the logbook, resulting in incomplete data for adherence with the shoe inserts, use of co-interventions and pain relief, activities undertaken, and adverse reactions. Thus, it is not clear whether these data are representative of the entire cohort.