Background: Interhospital transfer is frequent, and transferred patients can have worse outcomes than direct admissions. We sought to define the incidence of interhospital transfer in general surgery and evaluate its association with surgical outcomes.
Study design: The 2011 American College of Surgeons NSQIP database was used. Transferred patients were compared with urgent, inpatient direct admissions in a series of increasingly complex risk-adjustment models, including multiple regression using modified Poisson and negative binomial models, as well as propensity scores. Primary outcomes were overall complications, mortality, length of stay, and readmission.
Results: Overall, 7% of inpatient general surgery cases were transferred in. Among urgent cases, there were 6,197 transferred patients and 47,267 direct admissions. The most common procedures for direct admissions were appendectomy and cholecystectomy, and transfers had a more complex and broader range of procedures. On unadjusted analysis, transferred patients had a much higher risk for complications (risk ratio [RR] = 1.48; 95% CI, 1.45-1.52) and mortality (RR = 2.70; 95% CI, 2.48-2.94), as well as a longer length of stay (1.74 times longer; 95% CI, 1.69-1.78) and higher risk of readmission (RR = 1.31; 95% CI, 1.20-1.44). In the most sophisticated model, the propensity score match, the difference in outcomes for transferred patients was only modestly higher or equivalent (complications: RR = 1.03; 95% CI, 1.00-1.07; mortality: RR = 0.98; 95% CI, 0.88-1.09; length of stay: 1.08 times longer; 95% CI, 1.04-1.11; readmission: RR = 0.97; 95% CI, 0.88-1.08).
Conclusions: Interhospital transfer is frequent in surgery. Worse outcomes seen in transferred patients are largely due to confounding by patient characteristics rather than any true harm from transfer. Pay-for-performance schemes should adjust for transfer status to avoid unfairly penalizing hospitals that frequently accept transfers