Research Activities, December 2012: Acute Care/Hospitalization: A few preoperative factors can be used to predict inpatient deaths following surgery at resource-limited hospitals

Research Activities, December 2012: Acute Care/Hospitalization: A few preoperative factors can be used to predict inpatient deaths following surgery at resource-limited hospitals

A few preoperative factors can be used to predict inpatient deaths following surgery at resource-limited hospitals

Various organizations have developed programs to define, measure, and evaluate the quality of surgical care delivery. However, most of these models are based on hospitals in urban areas. As a result, many suburban and rural hospitals are overlooked. Therefore, there is a need to develop a risk-adjusted tool to measure surgical outcomes in these smaller institutions, particularly those that are resource-limited. Researchers have now come up with a model that uses fewer than six preoperative variables, making it easier for these hospitals to conduct their own surgical outcome evaluations.
For the development of this new model, the researchers used patient data collected on more than 130 variables by the National Surgical Quality Improvement Program. Within this group of variables are 66 preoperative ones designed to predict inpatient mortality. The researchers used a six-step process that added each variable sequentially to predict inpatient death. Models developed with the highest receiver operating charge (ROC) values, that is, best sensitivity and specificity, were then subjected to validation from a rural U.S. hospital.
A 3-variable model reached higher than a 90 percent ROC value at predicting inpatient mortality. The three variables were the American Society of Anesthesiologists physical status classification, functional status at the time of surgery, and patient age. Adding a fourth variable only increased value to 91 percent; a 6-variable model neared 92 percent. Very little additional gain was realized by adding more variables to the 3-variable model.
This new model can be easily implemented in resource-poor settings, including hospitals located in low- and middle-income countries. The researchers point out that while collecting outcomes on in-hospital mortality is important, future models should consider other factors such as surgical complications, morbidity, and disability-adjusted life-years. The study was supported in part by the Agency for Healthcare Research and Quality (HS19913).
See "Brief tool to measure risk-adjusted surgical outcomes in resource-limited hospitals," by Jamie E. Anderson, M.P.H., Randi Lassiter, B.S., Stephen W. Bickler, M.D., and others in the May 21, 2012, Archives of Surgery [Epub ahead of print].

— KB