Three-factor tool could be all you need to assess surgery risk

By Sarah Guy, medwireNews Reporter

A simple tool involving less than six variables to measure surgical outcomes could be all that is needed to predict inpatient mortality, show US study results.

Area under the receiver operating characteristic curve (AUC) scores were "substantially" high, at above 90% for a tool involving only patients' American Society of Anesthesiologists (ASA) class, their functional status at the time of surgery, and their age, report the researchers.

The finding could benefit resource-poor hospitals that are unable to collect the variables required by other multi-variable tools such as the National Surgical Quality Improvement Program (NSQIP), which contains 130 variables and includes 30-day patient follow-up data.

"Allowing hospitals in resource-limited countries to participate in surgical quality improvement efforts through the development of a simplified tool to measure surgical outcomes is the next critical step to improving surgical outcomes globally," say Jamie Anderson and colleagues from the University of California, San Diego.

The team designed the tool using a list of all preoperative variables listed in the NSQIP database and data for 631,449 patients treated between 2005 and 2009, adding additional variables sequentially and testing its ability to discriminate between surgical survivors and nonsurvivors.

The single variable with the greatest discriminatory ability was the ASA physical status classification, report Anderson et al, with an AUC score of 0.85, where 1.00 denotes perfect discrimination.

The next most discriminative variables were albumin measures, functional status, age, sepsis status, and preoperative hematocrit.

However, the model with the highest AUC score, at 0.94, included just three of these - age, ASA classification and functional status. Indeed, adding further variables decreased the discriminative ability of the model, to 0.91 with four variables, and 0.92 with five.

Even including the full 66 preoperative variables available only gave an AUC score of 0.91, remark the authors in the Archives of Surgery.

"The data presented allow for a wide range of possible risk adjustment models, allowing surgical systems to choose the most appropriate model given their unique resources," they say.

Future risk-adjustment models would benefit from considering surgical complications and morbidity rather than straight mortality, they conclude.

Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.