How many sites should an orthopedic trauma prospective multicenter trial have? A marginal analysis of the Major Extremity Trauma Research Consortium completed trials.

BACKGROUND: Multicenter trials in orthopedic trauma are costly, yet crucial to advance the science behind clinical care. The number of sites is a key cost determinant. Each site has a fixed overhead cost, so more sites cost more to the study. However, more sites can reduce total costs by shortening the study duration. We propose to determine the optimal number of sites based on known costs and predictable site enrollment. METHODS: This retrospective marginal analysis utilized administrative and financial data from 12 trials completed by the Major Extremity Trauma Research Consortium. The studies varied in size, design, and clinical focus. Enrollment across the studies ranged from 1054 to 33 patients. Design ranged from an observational study with light data collection to a placebo-controlled, double-blinded, randomized controlled trial. Initial modeling identified the optimal number of sites for each study and sensitivity analyses determined the sensitivity of the model to variation in fixed overhead costs. RESULTS: No study was optimized in terms of the number of participating sites. Excess sites ranged from 2 to 39. Excess costs associated with extra sites ranged from $17K to $330K with a median excess cost of $96K. Excess costs were, on average, 7% of the total study budget. Sensitivity analyses demonstrated that studies with higher overhead costs require more sites to complete the study as quickly as possible. CONCLUSIONS: Our data support that this model may be used by clinical researchers to achieve future study goals in a more cost-effective manner. TRIAL REGISTRATION: Please see Table 1 for individual trial registration numbers and dates of registration.

Minimal Clinically Important Difference (MCID) for the Short Musculoskeletal Function Assessment (SMFA) in Severe Lower Extremity Trauma: Pooled Data from 7 Multicenter, Prospective Clinical Trials.

BACKGROUND: The Short Musculoskeletal Function Assessment (SMFA) is a well validated, widely used patient-reported outcome (PRO) measure for orthopaedic patients. Despite its widespread use and acceptance, this measure does not have an agreed upon minimal clinically important difference (MCID). The purpose of the present study was to create distributional MCIDs with use of a large cohort of research participants with severe lower extremity fractures. METHODS: Three distributional approaches were used to calculate MCIDs for the Dysfunction and Bother Indices of the SMFA as well as all its domains: (1) half of the standard deviation (one-half SD), (2) twice the standard error of measurement (2SEM), and (3) minimal detectable change (MDC). In addition to evaluating by patient characteristics and the timing of assessment, we reviewed these calculations across several injury groups likely to affect functional outcomes. RESULTS: A total of 4,298 SMFA assessments were collected from 3,185 patients who had undergone surgical treatment of traumatic injuries of the lower extremity at 60 Level-I trauma centers across 7 multicenter, prospective clinical studies. Depending on the statistical approach used, the MCID associated with the overall sample ranged from 7.7 to 10.7 for the SMFA Dysfunction Index and from 11.0 to 16.8 for the SMFA Bother Index. For the Dysfunction Index, the variability across the scores was small (<5%) within the sex and age subgroups but was modest (12% to 18%) across subgroups related to assessment timing. CONCLUSIONS: A defensible MCID can be found between 7 and 11 points for the Dysfunction Index and between 11 and 17 points for the Bother Index. The precise choice of MCID may depend on the preferred statistical approach and the population under study. While differences exist between MCID values based on the calculation method, values were consistent across the categories of the various subgroups presented. LEVEL OF EVIDENCE: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.