Background and purpose: The technologic imperative has prompted the adoption of complex laparoscopic techniques by physicians with various degrees of skill. We sought to measure the impact of both case mix and physician practice (perioperative process/risk factors) on length of stay (LOS)-a common benchmark- after laparoscopic surgery.
Patients and methods: We identified 911 patients undergoing laparoscopic retroperitoneal surgery between 1996 and 2004, who comprise our study population. Patients remaining in the hospital >5 days-the 90th percentile for the sample-were classified as having a prolonged LOS. Adjusted models were developed to determine the independent association of case mix and process measures with a prolonged LOS. The likelihood ratio test was used to discern the improvement of fit of the process model compared with the case-mix model.
Results: Among factors related to case mix and structure of care, increasing age (odds ratio [OR] 1.1; 95% CI 1.0, 1.2), less surgeon experience (OR 6.1; 95% CI 2.1, 17.2), male gender (OR 2.1; 95% CI 1.2, 4.0), and American Society of Anesthesiologists score of 3 or 4 (OR 7.2; 95% CI 2.2, 23.3) were independently associated with a prolonged LOS. The need for a transfusion (OR 9.4; 95% CI 33.9, 23.2), the development of a postoperative complication (OR 4.6; 95% CI 2.2, 9.5), and longer operative time (OR 1.5; 95% CI 1.3, 1.8) explained additional variation in prolonged LOS outcomes when considering perioperative process/risk factors in the model. Perioperative factors significantly improved the fit of the model (chi (2) statistic 101.8; p < 0.0001).
Conclusions: Significant variation in outcomes is explained by factors describing aspects of surgical expertise. Variability in the surgical skill set is likely greatest during the laparoscopic learning curve, which raises a quality-of-care concern during the initial implementation of the technique. Policies attempting to smooth the laparoscopic learning curve, such as mentoring and skill measurement prior to credentialing, could improve the quality of care.