The Effect of Resident Duty Hour Restriction on Trauma Center Outcomes in Teaching Hospitals in the State of Pennsylvania

BACKGROUND: Resident duty hour restriction was instituted to improve patient safety, but actual impact on patient care is unclear. We sought to determine the effect of duty hour restriction on trauma outcomes in Level I trauma centers (TCs; surgery residency programs) versus Level II TCs (those with no surgery residency programs) within the state of Pennsylvania, using noninferiority as our hypothesis testing. METHODS: Outcomes (mortality and length of stay [LOS]) were compared in Level II TCs without surgery residencies (n = 7) with Level I TCs (with surgery residencies; n = 14) PRE80 (2001-2003) and POST80 (2004-2007). The subcategories of critically injured patients, Injury Severity Score (ISS) >15, ISS >25, Trauma and Injury Severity Score (TRISS) ≤ 50, Abbreviated Injury Scale (AIS) head/chest/abdomen score >3, age >65 years, mechanism, and shock, functioned as outcome predictors. RESULTS: There was a decrease in mortality overall PRE80 to POST80 for Level I and II TCs. There was a decrease in mortality in Level I TCs POST80 in ISS >15 (16.5% vs. 14.8%, p = 0.0001), AIS (head) score >3 (20.8% vs. 17.8%, p < 0.0001), age >65 years (12.2% vs. 10.7%, p = 0.0013), and blunt mechanism (5.2% vs. 4.6%, p = 0.0004). LOS was reduced in ISS >15, AIS (head) score >3, age >65 years, and penetrating mechanism in Level I TCs POST80. A similar but more profound decrease was also seen in Level II TCs PRE80 and POST80 (ISS >15, 25; AIS (head) score; shock; blunt mechanism; and TRISS ≤ 50). Testing for inhomogeneity identified less-severely injured patients at Level II TCs POST80 compared with Level I TCs in certain subcategories (ISS >15, 25; AIS (head) score; shock; blunt mechanism; and TRISS ≤ 50) regarding mortality and LOS (TRISS >50%). CONCLUSIONS: Decreases in mortality and LOS during the study periods were likely not related to resident work hour restriction but rather to overall improvement in outcomes seen at Level II (no residents) and Level I (residents) TCs. Resident work hour restrictions had no discernible effect on patient care (noninferiority).