The Statistical Fragility of Platelet-Rich Plasma as Treatment for Plantar Fasciitis: A Systematic Review and Meta-Analysis

CATEGORY: Basic Sciences/Biologics; Hindfoot; Sports INTRODUCTION/PURPOSE: Plantar fasciitis (PF) is estimated to account for 8% of all running-related injuries, affecting nearly 1 out of every 10 adults aged 50 and older. Recent randomized controlled trials (RCTs) suggest platelet-rich plasma (PRP) injection may be an effective non-operative treatment option for chronic, refractory PF. However, RCTs in general have come under scrutiny due to a frequent lack of reproducibility, which is largely attributable to shortcomings of the commonly used p<0.05 threshold for significance. Fragility indices may be an appropriate tool to gauge the clinical importance of RCT findings in these situations. In this study, we utilize the continuous fragility index (CFI) and continuous fragility quotient (CFQ) to determine the statistical robustness of data from RCTs evaluating PRP for PF. METHODS: RCTs comparing outcomes after PRP injection versus other conservative treatment in chronic PF patients were evaluated. Representative datasets were generated for each reported outcome event using summary statistics. The CFI was determined by manipulating each dataset until reversal of significance (a=0.05) was achieved. The corresponding CFQ was calculated by dividing the CFI by the sample size. RESULTS: Of 259 studies screened, 27 studies (59 outcome events) were included in this analysis. The median CFI for all events was 9 (CFQ=0.177), suggesting that altering the treatment of only 9 patients (or 17.7 out of 100) would be sufficient to reverse trial significance (Table 1). Of the 36 outcome events reporting lost to follow-up data, 10 events (27.8%) lost >=9 patients. Outcome events that were originally reported as significant (p<0.05) were considerably more fragile (CFI=5; CFQ=0.122) than events that were reported as nonsignificant (CFI=10; CFQ=0.179). CONCLUSION: Fragility indices are a useful adjunct to p-values and provide an assessment of how easily statistical significance can be overturned. While there is some preliminary evidence supporting the clinical efficacy of PRP, our fragility analysis suggests that RCT findings may be underpowered in some cases. Over 27% of outcome events may have experienced a reversal of significance if the studies had simply maintained follow-up. Given the importance of RCTs in clinical decision-making, fragility indices should be reported alongside p-values to indicate the strength of statistical findings.