[Surgical risk assessment using multivariate analysis].
INTRODUCTION: Over and above the risk of a disease itself, surgical treatment implies additional risk involving possible temporary or permanent deterioration in a patient's health. Most studies of anesthetic-surgical risk are based on mortality or serious morbidity (severe, permanent deterioration). They may therefore be of limited applicability in hospitals where such outcomes are rare due to the types of patients treated. In order to compare and classify patients as well as improve outcome in such health care settings other parameters must be evaluated. PATIENTS AND METHODS: Indicators presaging adverse situations related directly to surgery in our postoperative intensive care units involve blood transfusion and perioperative events. By applying multivariate (regression) analysis we can determine the factors that condition these unfavorable situations. RESULTS: In the setting we studied, the main determinants of unfavorable events after surgery are patient health as shown by ASA classification, duration of operation and type of surgery. CONCLUSIONS: 1) Peripheral surgery in patients in good health is very safe. 2) Easily obtained parameters can be used to obtain a fairly efficient prediction of risk. 3) Single-setting risk studies are easy to carry out and useful for providing information about a specific hospital center.