The burden of traumatic injury among older adults is underappreciated, as it is the predominant cause of death among children and younger adults and therefore often thought to be primarily a disease of the young. 1 However, traumatic injury in older adults is increasing in prevalence as the US population ages. 2 Older patients present unique care challenges because they have decreased physiologic reserve, have more comorbid conditions, and are at risk of specific complications such as delirium. 3 Our objective was to quantify occurrence of and spending on traumatic injury in comparison with other common medical diseases resulting in acute hospitalization among Medicare beneficiaries.
Methods | We evaluated Centers for Medicare & MedicaidServices Parts A and B claims data for inpatient hospitalizations among adults aged 65 years or older with fee-forservice Medicare coverage from 2008 to 2014 (the most recent data available to our research group) using a nationally representative 20% sample. We computed all index hospitalization, readmission, and postacute care (skilled nursing facility, long-term acute care facility, inpatient rehabilitation, and outpatient Part B) payments within 90 days of discharge. Payments were categorized according to primary indication for hospitalization based on the first International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code listed for each episode. Traumatic injury was defined as a first-position ICD-9-CM code from 800 to 959.9, excluding late effects, superficial injuries, or foreign bodies entering through orifice. 4 Additional ICD-9-CM diagnosis groupings were created for congestive heart failure, pneumonia, stroke, and acute myocardial infarction. These diagnosis groupings were chosen because they represent common acute diseases among Medicare beneficiaries.Analyses were performed using SAS version 9.4 (SAS Institute Inc). Proportions of inpatient hospitalizations and Medicare payments were compared between disease categories using 2-sided z tests or t tests. 95% Wald confidence intervals were constructed. Statistical significance was set at P < .005 to account for multiple comparisons. Payment data are presented in real dollars (not inflation adjusted). The Michigan Medicine Institutional Review Board approved this study with a waiver of informed consent.