Motorcycle Trauma Outcome Registry (MOTOR) Trial Dataset
Description
The present data includes the MOTOR Trial Dataset, Data Dictionary, and Data Analysis Codes. The data for the MOTOR trial were derived from a multi-center cluster randomized trial that aimed to evaluate the impact of rural trauma team development training and its implementation on various metrics, including prehospital time, the interval from referral decision to hospital discharge, and the clinical outcomes associated with motorcycle-related neurological and orthopedic injuries in Uganda (Pan African Clinical Trial Registry: PACTR202308851460352). The study was predicated on the null hypothesis asserting no significant difference in outcomes between intervention hospitals that received rural trauma team training and implementation and control hospitals that did not receive such training. The results indicated a notable reduction in prehospital time and all-cause mortality by more than 50% in the intervention group, while also not demonstrating any deterioration in patient-reported trauma morbidity, as assessed by the Glasgow Coma Scale and Trauma Outcome Measure scores. Therefore, the level II prognostic evidence derived from this trial suggests that locally contextualized, trainee-led rural trauma team development intervention programs are both feasible and effective in enhancing clinical processes and improving patient outcomes within low- and middle-income contexts.
Files
Steps to reproduce
We assessed the normality of distribution and the equality of variance utilizing the Shapiro-Wilk and Levene’s tests, respectively. For the primary outcomes, a two-sample Wilcoxon rank-sum test was employed to compare the median values (interquartile range, IQR) due to the skewed nature of the data. In analyzing the secondary outcomes, an adjusted Chi-square test was utilized to compare the differences in proportions of all-cause 90-day mortality and the proportions of unfavorable Glasgow Outcome Scale (GOS) and Trauma Outcome Measure Scale (TOMS). GOS was stratified into unfavorable (score 1-3) versus favorable outcomes (score 4-5), whereas TOMS were dichotomized into unfavorable (TOMS < Trauma Expectation Factor Score, TEFS) vs. favorable (TOMS ≥ TEFS) The fixed effects variable was the treatment arm (intervention vs. control) as the unit of analysis with the odds ratios and their corresponding 95% CI as a direct estimate of the effect size. Intracluster correlation coefficients (ICC) were computed in the mixed effects restricted maximum likelihood (REML) regression model using the “estat icc” command as the ratio of between cluster variance of outcomes to total variance of outcomes between and within cluster. Analyses were performed in Stata 15.0
Institutions
Categories
Funding
University of Turku
Turku University Hospital
University of California, San Francisco