Clinical data was extracted from electronic health records from patients treated for hip fractures by the Department of Orthopedic Surgery at NYU Langone Health between October 2014 and February 2020. A total of 1,590 patients aged 55 and older were included in the retrospective review with complete records for patient's sex; body mass index; comorbidities as measured with the Charlson Comorbidity Index (CCI); American Society of Anaesthesiologists Physical Classification score; Glasgow Coma Score; Abbreviated Injury Severity score for the chest, head and neck, and extremities; and pre-injury ambulatory status. Other variables reviewed included post-operative complications such as acute kidney injury, sepsis or septic shock, pneumonia, stroke, and cardiac arrest; mortality (inpatient, 30-day, and 1-year post-discharge); ICU admission; and readmission at 30- and 90-days post-discharge.

The accompanying R code was used to perform propensity matching with baseline and outcome variables using three matching schemes: The Score for Trauma Triage in Geriatric and Middle-Aged (STTGMA), CCI alone, or a combination of sex, age, CCI and BMI. 10% of patients are selected by a pseudorandom number generator and matched with the remaining patients 1:1 by propensity scores generated through logistic regression with one of the aforementioned matching schemes. The code then implements Chi-square, Fisher, or Mann-Whitney U (Wilcoxon Rank Sum) testing with a significance level of 0.05 to compare matched cohorts.