S. A. Brownlee1, S. G. Pappas2, L. A. Gil1, A. Cobb1,2, P. C. Kuo1,2, A. N. Kothari1,2, G. J. Abood2 1Loyola University Medical Center,One:MAP Division Of Clinical Informatics And Analytics,Maywood, IL, USA 2Loyola University Medical Center,Department Of Surgery,Maywood, IL, USA
Introduction: Pain management is a crucial aspect of cancer care, particularly in patients undergoing surgical tumor resection. An increasing awareness of the potential hazards of opioid use, coupled with the high rate of opioid utilization by cancer patients, necessitates further study of risk factors for opioid-related adverse events in patients undergoing oncologic resection. The objective of this study was to construct a simple decision tree model to predict patients likely to experience an opioid-related adverse event following an oncologic resection.
Methods: The Healthcare Cost and Utilization Project (HCUP) State Inpatient Database (SID) and State Emergency Department Database (SEDD) from the state of California for the years 2006-2011 were linked to define the population of interest. Patients undergoing one of four elective oncologic resection procedures (esophagectomy, lung lobectomy, hepatic resection, and colectomy) were included for study. The primary endpoint was an opioid-related adverse event during the year following surgery. A fast and frugal decision tree was constructed to predict the factors that most contributed to the occurrence of an opioid-related adverse event.
Results: 148 699 patients undergoing one of four oncologic resection procedures in CA during 2006-2011 met inclusion criteria. Of these, 230 (0.2%) experienced an opioid-related adverse event in the year following the procedure. Recursive partitioning analysis of the cohort revealed age, income, length of stay after procedure, and comorbidity index as the most significant predictors of opioid-related adverse event, with age being the strongest predictor. For patients less than 63 years old, income level was the next strongest predictor of an opioid event.
Conclusions: Though rare, opioid-related poisonings and adverse events are serious complications for cancer patients undergoing surgical resection. Through the use of four readily-obtainable patient variables (age, income level, length of stay, and comorbid disease burden), this simple prediction model may provide clinicians with a practical tool to help decrease the frequency of opioid-related adverse events in a particularly vulnerable population.
