Abstract
Hospital type is an indicator for structures and processes of care. The effect of hospital type on hip fracture in-hospital mortality is unknown. We determine whether hip fracture in-hospital mortality differs according to hospital type.
We retrieved records of hip fracture for 167,816 patients aged 65 years and older, who were admitted to a Canadian acute hospital between 2004 and 2012. For each hospital type we measured and compared the cumulative incidence of in-hospital death by in-patient day, accounting for discharge as a competing event.
The cumulative incidence of in-hospital death at in-patient day 30 was lowest for teaching hospital admissions (7.3%) and highest for small community hospital admissions (11.5%). The adjusted odds of in-hospital death were 12% (95% CI 1.06–1.19), 25% (95% CI 1.17–1.34), and 64% (95% CI 1.50–1.79) higher for large, medium, and small community hospital versus teaching hospital admissions. The adjusted odds of nonoperative death were 1.6 times (95% CI 1.42–1.86), and 3.4 times (95% CI 2.96–3.94) higher for medium and small community hospital versus teaching hospital admissions. The adjusted odds of postoperative death were 14% (95% CI 1.07–1.22) and 20% (95% CI 1.10–1.31) higher at large and medium community hospitals versus teaching hospitals. The adjusted odds of postoperative death were largest at small community hospitals but the confidence interval crossed 1 (OR = 1.25, 95% CI 0.92–1.70).
A higher proportion of hip fracture patients die at non-teaching compared to teaching hospitals accounting for length of stay. Higher mortality at small community hospitals may reflect disparities in access to resources and delay to treatment.
