Epidemiology and short-term outcomes of injured medicare patients

OBJECTIVES: To describe characteristics and short-term outcomes of Medicare patients hospitalized after injuries in 1999. DESIGN: Analysis of national population-based case series. SETTING: Hospitalized Medicare patients. PARTICIPANTS: All fee-for-service Medicare patients aged 65 and older admitted for the first time in 1999 with principal injury diagnoses (International Classification of Diseases, Ninth Revision, codes 800-904, 910-929, 940-957, 959). MEASUREMENTS: Incidence rates, stratified by anatomic location (hip, other extremity, spine, head, chest, other), sex, and age group (65-74, 75-84, greater than or equal to85). For each category, Charlson comorbidity scores, Abbreviated Injury Scores, hospital length of stay, discharge disposition, hospital mortality, 30-day mortality, and readmissions within 30 days of discharge. RESULTS: A total of 439,605 persons were admitted at least once (crude rate 1,654/100,000). Rates of hospitalization increased with age and were generally higher in women (except head injuries). Comorbidities were more common in men. Hip fractures constituted 46.6% of cases and other extremity injuries another 30.7%. Hospital mortality (3.7% overall) increased with age, was greater in men, and was highest in patients with head injuries. The proportion discharged to skilled nursing facilities (43.8% overall, range 10.0-61.9% by age/sex/anatomic category) also increased with age, was higher in women, and was highest in patients with hip fractures. Slightly more than one-tenth (12.3%) of patients were readmitted within 30 days. Thirty-day mortality was 2.0 times hospital mortality (range 1.2-3.4 by category). CONCLUSION: Most injuries resulting in hospitalization for the Medicare population involve the extremities, but other injuries have higher mortality. Many injured patients are not discharged home but receive additional institutional care. Thirty-day survival is much lower than observed hospital survival. Further studies of injuries using Medicare data are warranted.