Pediatric lower extremity fracture treatments: a statistical review of randomized controlled trials

Abstract

Introduction

Pediatric lower extremity fractures present with significant clinical challenges and largely rely on randomized controlled trials (RCTs) for intervention evaluation. The statistical robustness of these trials is seldom evaluated, but when examined reveals issues such as small sample sizes and underpowered results. Therefore, this study aimed to test the statistical robustness of RCTs evaluating pediatric lower extremity fracture interventions using fragility statistics.

Methods

PubMed, Embase, and MEDLINE were systematically searched for recent RCTs (2000–2023) assessing outcomes for pediatric patients with lower extremity fractures. The fragility index (FI), or number of event reversals required to alter statistical significance, was calculated for all dichotomous outcomes. The fragility quotient (FQ) was then determined by dividing the FI by the study sample size.

Results

After screening, 14 studies were included for analysis. Across 83 total dichotomous outcomes, the median FI was 5 (IQR 3–6) with an associated median FQ of 0.070 (IQR 0.033–0.107), suggesting that just event reversal in 5 patients, or 7.0 % of the study population, would alter significance for 50 % of outcomes. 15 outcomes were statistically significant (FQ = 0.029), and 68 outcomes were non-significant (FQ = 0.083). Outcomes were grouped into three categories, including fracture reduction or union (28 outcomes), functional improvement and patient satisfaction (26 outcomes), and adverse events (29 outcomes). Category FQs were 0.061, 0.069, and 0.076, respectively. 47 outcomes were extracted from studies published prior to 2018 (FQ = 0.086), while the 36 outcomes published in 2018 or later resulted in a median FQ of 0.048. Three fracture types were assessed, including tibia (37 outcomes), femur (39 outcomes), and ankle (7 outcomes) fractures, with FQs of 0.075, 0.060, and 0.050, respectively. Across the 59 outcomes from studies with operative interventions, the FQ was 0.075, whereas the 24 outcomes from studies with non-operative interventions reported a FQ of 0.060.

Conclusion

The efficacy of treatments in pediatric lower extremity fractures from RCTs are slightly fragile, particularly among significant outcomes. Over time, these studies have become less robust. Larger RCTs that combine the reporting of p-values with FI and FQ metrics may provide more robust evidence for guiding effective treatment strategies in pediatric lower extremity fractures.