Background: Misclassification of AI (American Indian)/AN (Alaskan Native) race is a well-documented problem in surveillance systems such as cancer and death certificates and it is unknown to what extent miscoding of race occurs in current state and national trauma registries. The goal of this study was to use linked Indian health registration records to evaluate the completeness and accuracy of race reporting in the Washington State Trauma Registry, characterize injury patterns and mechanisms, and understand risk factors for adverse outcomes among Washington’s AI/AN population.
Methods: We performed a probabilistic linkage of Indian health registration records from Indian Health Service (IHS), tribal, and urban clinics in the Pacific Northwest with patient records from the Washington State Trauma Registry, 2005-2009, updating missing or misclassified AI/AN status. Cases were selected using the principle ICD-9-CM codes 800.0-959.9.
Results: Record linkage increased ascertainment of AI/AN trauma cases by 69.4%, from 1,600 pre-linkage to 2,711 post-linkage. Among the 90% of matched records with recorded race information, 50.7% were misclassified as another race in the state data system. AI/AN trauma patients tended to be younger (mean age =35.1 years vs. 45.9 years for whites) and were more likely to be male (66.6% vs. 62.5% for whites). Age-adjusted rates were almost twice as high among the AI/AN population (455.2 per 100,000 vs. 244.37 per 100,000 for whites). Differences were seen in transport times from scene to hospital (AI/AN mean=39.1 minutes vs. 34.2 minutes for whites), mechanism of injury (31.7% of AI/AN injuries were due to motor vehicle traffic collision vs. 24.5% for whites) and intent (homicide/assault accounted for 17.5% of AI/AN cases vs. 5.0% for whites). In bivariate analysis AI/AN patients were slightly less likely to die than whites (OR = 0.78, 95% CI: [0.62 – 0.98]). AI/AN race was not a significant predictor of death after controlling for ISS, age, sex, intent of injury, ICU admission, and transport time.
Conclusions: To our knowledge, this is the first report based on linked data between a state trauma registry and Indian health databases. Although some notable differences in trauma patterns were identified, AI/ANs in Washington State do not appear to experience excess mortality compared with their white counterparts. This project demonstrated how data linkage can increase the quality of injury data for the AI/AN population and highlights opportunities for further analysis to better understand risk factors for injury and ultimately improve outcomes for AI/AN trauma patients.