Objective: To compare the cost and outcomes of patients treated at orthopaedic teaching hospitals (OTHs) with those treated at nonteaching hospitals (NTHs).
Design: Retrospective study.
Setting: The Statewide Planning and Research Cooperative Systems (SPARCS) database, which includes all admissions to New York State hospitals from 2000-2011.
Patients/participants: A total of 165,679 patients with isolated closed hip fracture 65 years of age and older met inclusion criteria. Of them, 57,279 were treated at OTH and 108,400 were treated at NTH.
Intervention: Admission for the management of a hip fracture.
Main outcome measure: Cost, length of stay (LOS), and inpatient mortality.
Results: Univariate analysis shows that mean total hospital costs were higher at OTH ($16,576 ± $17,514) versus NTH ($13,358 ± $11,366) (P < 0.001); LOS was equivalent at OTH (8.0 ± 9.0 days) versus NTH (8.0 ± 7.6 days) (P = 0.904); and mortality was lower in OTH (3.4%) versus NTH (4.0%) (P < 0.001). In the multivariate total cost analysis, in addition to demographic differences, we identified total hospital beds and total ICU beds as significant confounding variables. Interestingly, when controlling for these patient and hospital factors, OTH designation was not a significant predictor of cost. In addition, multivariate analysis found that OTH status decreased LOS by 0.743 days (95% confidence interval: 0.632-0.854, P < 0.001) and mortality by 21% (odds ratio 0.794, 95% confidence interval: 0.733-0.859, P < 0.001), confirming the univariate trends.
Conclusions: While OTH may seem to have higher hospital costs for operative hip fractures on cursory analysis, controlling for patient and hospital factors including hospital bed number negates this effect such that OTH has no additional cost compared with NTH. In addition, OTH status is associated with shorter LOS and lower in-hospital mortality. With the results of this study, health care systems and patients should feel confident that the quality of care at teaching hospitals is no less and potentially better than that at NTH with no added cost.
Level of evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.