Regionalization of the surgical care of children: A risk-adjusted comparison of hospital surgical outcomes by geographic areas

Background There is an ongoing debate among pediatric surgeons regarding the need or lack thereof to centralize the surgical care of children to high-volume children's centers. Risk-adjusted comparisons of hospitals performing pediatric surgery are needed. Methods Admissions from 2006 to 2010 from two national administrative databases were analyzed. Only nontrauma pediatric patients undergoing a noncardiac surgical procedure were included. Risk-adjustment was performed with a validated International Classification of Diseases, 9th Revision code-based tool. Hospitals were grouped into metropolitan regions using the first three digits of their zip code. Poorly performing outlier hospitals were defined by an odds ratio >1 and P value <.05 for mortality compared with the center with the greatest pediatric operative volume in that same region. Results Information was obtained from 415,546 pediatric surgical admissions, and 173 hospitals in 55 regions were compared. A total of 18 poor performing hospitals (adjusted odds ratio, range 1.91-35.95) in 15 regions were identified. Mortality in poor performers ranged from 1.11% to 10.19% whereas that in the high-volume reference centers was 0.37-2.41%. A subset analysis in patients <1 year of age showed 37 poor performers in 46 regions. Median number of surgical admissions was 345 (interquartile range 152-907) for nonoutlier and 240 (interquartile range 135-566) for outlier centers (P =.30). Conclusion The present analysis is a novel risk-adjusted assessment of the performance of hospitals delivering pediatric surgical care. By identifying the existence of multiple poor performing outlier hospitals, this study provides valuable data for discussion as health care delivery systems continue to debate optimal resource distribution and regionalization of the surgical care of children.