Effects of laparoscopy on surgical discharge practice patterns

Abstract

Objectives: The length of stay (LOS) after surgical procedures often varies regionally and reflects local idiosyncratic practice patterns. We sought to elicit the influence of local pressures (in the form of the technologic imperative exerted by laparoscopy) on discharge physician practice style as measured by LOS.

Methods: Patients undergoing prostatectomy, nephrectomy, and hysterectomy in 2003 were identified from the Nationwide Inpatient Sample using the International Classification of Diseases, Ninth Revision, codes. Within each procedure, the laparoscopy volume was measured and sorted into tertiles. Patients remaining in the hospital longer than the 90th percentile after each surgery were excluded to minimize the influence of imperative indications on LOS. For each procedure, generalized linear models were used to measure the relationship between the hospital laparoscopy volume and log-transformed LOS outcomes among patients undergoing conventional surgery.

Results: In 2003, 2%, 19%, and 16% of patients underwent laparoscopic prostatectomy, nephrectomy, and hysterectomy, respectively. A hospital's laparoscopy and conventional volumes were moderately to highly correlated for hysterectomy (r = 0.65) and nephrectomy (r = 0.58) and less so for prostatectomy (r = 0.24). Among patients undergoing conventional prostatectomy and nephrectomy, those treated at high-volume laparoscopy hospitals had 0.41 and 0.30 fewer hospital days, respectively, compared with those treated at hospitals at which only conventional surgery was performed (both P <0.05). This effect was partially explained by a hospital's open procedure volume (both P = 0.07). Similar relationships were not evident for hysterectomy.

Conclusions: These data highlight that patients treated by conventional methods at high-volume laparoscopy centers have shorter hospital stays and suggest that physician practices might susceptible to external influences such as the technologic imperative.