Selection on Observables and Unobservables: Level I Trauma Center Effects on Return to Work Outcomes

Trauma accounts for twelve percent of Emergency Room visits in America. Level-I Trauma Centers provide care to persons with severe injury. Their specialization raises the possibility that they provide better care to patients there admitted than an equivalently injured person in a non-trauma hospital. MacKenzie et al (2006), using matching methods, showed that the risk of death is significantly lower when care is provided in a Level-I Trauma Center than in a non-trauma center. This dissertation explores whether for persons who were mainly working prior to their injury, was treatment at a trauma center more likely to result in their returning to work within 3 and within 12 months after injury. Outcomes are compared among patients treated in 18 hospitals with a Level-I Trauma Center and 51 non-trauma center hospitals located in 12 states. Patients 18 to 64 years old with a moderate-to-severe injury and who were mainly working prior to injury were eligible. We use linear probability models and propensity-score weighting to adjust for observable differences between patients treated at trauma centers and those treated at non-trauma centers. We also use Instrumental Variables methods to adjust for unobservable differences between patients. Covariates include severity of injury, labor market conditions, and individual socio-economic characteristics. Data comes from the National Study on the Costs and Outcomes of Trauma (NSCOT) conducted between 2001 and 2002. Using the percentage of people covered by helicopter transportation services at the state level as an instrument for the probability to be transported to a Level-I Trauma, we find that treatment at these centers improve the probability of returning to work within 3 months after injury by ten to twelve percentage points. These results are robust to sensitivity tests on specification and to sample variation