UC San Diego

BACKGROUND : The San Diego trauma population is aging over time and older trauma patients represent a vulnerable population at risk for mortality and severe morbidity. Reductions in validity of the prognostic metrics may be improved by evaluating chronic diseases, but there is minimal evidence to support this. Few studies have evaluated mortality after trauma center discharge, and none have evaluated mortality in the context of discharge as a competing risk. METHODS : Three studies were performed to evaluate chronic conditions on trauma-related outcomes as part of historical cohort study on blunt- injured patients age ≥55 years admitted between 01/2006-12 /2012. The first evaluated chronic conditions on in- hospital mortality (with a competing event) and hospital length of stay (HLOS). The second evaluated chronic conditions and mortality after trauma center discharge. In the third study, a chronic disease-based prognostic model for trauma-related mortality was constructed and compared to other leading metrics. RESULTS : There were 4653 unique patients who met all criteria for inclusion. Of 40 conditions, 23 were associated with in-hospital mortality. HLOS was associated with 32 of 40 conditions. Of 4442 survivors, 938 died within two years of discharge. Patients discharged to care facilities showed worse survival within the first 30 days (early-term mortality). Injury-related variables were associated only with early- term mortality, while chronic conditions were associated with two-year mortality after discharge. The model development procedure identified twelve conditions for inclusion. Model tests showed moderate performance for mortality and ours was superior versus other chronic disease metrics. Validation showed moderate performance of our model that did not significantly differ from other chronic disease metrics. Injury metrics, overall, were poor at predicting mortality. CONCLUSION : Chronic conditions vary in their relationship with mortality by time. In older trauma populations, chronic conditions show stronger associations with mortality than injury-related measures. Discharges to care facilities are valid competing events to in-hospital mortality and should be accounted for when assessing the quality of trauma care. Our developed model performed well with twelve variables compared to others with far more. Results can be used to inform patients and care providers of the risks outside the realm of injury