← All case studies ◆ Health equity · Statistical analysis

Analysis of racial disparities in trauma care and outcomes

A research team analysing National Trauma Data Bank records found racial disparities in US trauma care, with American Indian and Alaska Native patients experiencing more inter-facility transfers and poorer outcomes.

Two hands of different skin tones reaching towards each other

The challenge

A research team investigating healthcare equity found significant racial disparities in US trauma care. Their National Trauma Data Bank analysis showed concerning patterns, especially for American Indian and Alaska Native (AI/AN) patients, who had disproportionately high rates of inter-facility transfers and poorer outcomes than other groups.

What we set out to do

  • Analyse racial disparities in inter-facility transfers for trauma patients
  • Identify key factors affecting withdrawal of life support decisions
  • Evaluate differences in hospital length of stay across demographic groups
  • Assess variations in injury severity and Glasgow Coma Scale scores
  • Develop evidence-based recommendations to improve healthcare equity

Our approach

Data architecture analysis

Extracted patient data from the National Trauma Data Bank, cleaned data to handle missing values, applied statistical methods to address outliers, and used imputation techniques for incomplete records.

Technical implementation

Built linear regression and generalised linear models, used stepwise ANOVA comparisons for model optimisation, ran imputation validation across multiple statistical platforms, and cross-validated results between imputed and unimputed datasets.

Solution delivery

Delivered statistical analysis of racial disparities across several outcomes, assessment of inter-facility transfer patterns, multifactorial analysis of clinical, demographic and institutional influences, and evidence-based recommendations for systemic improvements.

Results and impact

American Indian patients were 96% more likely to undergo inter-facility transfers, while Black and Asian patients had 45% and 39% lower odds of transfer respectively. Life support withdrawal varied by race, with lower odds for non-white patients, and hospital length of stay differed significantly by race and ethnicity. The work revealed interactions between clinical presentation and patient demographics and identified priority areas for intervention and policy.

“American Indian patients in particular showed distinctive patterns suggesting potential systemic biases, with persistent disparities across multiple trauma care metrics shaped by institutional and geographic factors.”

Timeline

  • Initial data extraction and cleaning: 1 week
  • Statistical model development: 1 week
  • Analysis and interpretation: 1 week
  • Manuscript preparation: 1 week
  • Total: about 4 weeks

Looking forward

Developing culturally sensitive end-of-life care protocols, creating improved trauma triage protocols that account for possible assessment bias, and having regional trauma systems review transfer patterns and outcomes through an equity lens.