Statistical Fragility of Venous Thromboembolism Prophylaxis Following Total Joint Arthroplasty

Background: Statistical fragility is a quantitative measure of the robustness of the statistical conclusions drawn in a study. Although statistical fragility has been comprehensively evaluated in the arthroplasty literature, the statistical fragility of large-scale randomized trials evaluating venous thromboembolism (VTE) prophylaxis has not been evaluated. The purpose of this study was to determine the utility of applying the fragility index (FI) and the fragility quotient (FQ) analysis to randomized controlled trials (RCTs) evaluating VTE prophylaxis following total joint arthroplasty. Methods: A systematic review was performed by searching multiple databases to identify RCTs that evaluated VTE prophylaxis following total joint arthroplasty from 2000 to 2020. The FI was determined by manipulating each reported dichotomous outcome event until a reversal of significance was appreciated with 2 x 2 contingency tables. The associated FQ was determined by dividing the FI by the sample size.Results: Thirty-two RCTs were ultimately included for analysis. The overall FI incorporating all 32 RCTs was only 7 (interquartile range 3-9), suggesting that the reversal of only 7 events is required to change study significance. The associated FQ was determined to be 0.01. Of the RCTs that reported lost-to-followup data, the majority of studies had lost-to-follow-up numbers greater than 7.Conclusions: Our findings suggest that RCTs evaluating VTE prophylaxis following total hip arthroplasty and total knee arthroplasty may lack statistical stability as few outcome events are required to reverse the significance of outcomes. Future randomized trials should consider reporting FI and FQ along with the P value analysis to provide better context to the integrity of statistical stability.(c) 2023 The Authors. Published by Elsevier Inc. on behalf of The American Association of Hip and Knee Surgeons. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).