Background: When comparing hospitals on outcome indicators, proper adjustment for case mix (a combination of patient and disease characteristics) is indispensable. This study examines the need for case mix adjustment in evaluating hospital outcomes for non-small cell lung cancer surgery.
Methods: Data from the Dutch Lung Cancer Audit for Surgery were used to validate factors associated with postoperative 30-day mortality and complicated course with multivariable logistic regression models. Between-hospital variation in case mix was studied by calculating medians and interquartile ranges for separate factors on the hospital level and the "expected" outcomes per hospital as a composite measure.
Results: A total of 8,040 patients, distributed over 51 Dutch hospitals, were included for analysis. Mean observed postoperative mortality and complicated course were 2.2% and 13.6%, respectively. Age, American Society of Anesthesiologists classification, Eastern Cooperative Oncology Group performance score, lung function, extent of resection, tumor stage, and postoperative histopathologic findings were individual significant predictors for both outcomes of postoperative mortality and complicated course. A considerable variation of these case mix factors among hospital populations was observed, with the expected mortality and complicated course per hospital ranging from 1.4% to 3.2% and from 11.5% to 17.1%, respectively.
Conclusions: The between-hospital variation in case mix of patients undergoing surgical treatment for non-small cell lung cancer emphasizes the importance of proper adjustment when comparing hospitals on outcome indicators.
Copyright © 2018 The Society of Thoracic Surgeons. Published by Elsevier Inc. All rights reserved.