We have developed mathematical models to estimate the risk of perioperative
adverse events in patients with pre-existing conditions undergoing day-cas
e surgery. We studied 17 638 consecutive day-case surgical patients in a pr
ospective study. Preoperative, intraoperative and postoperative data were c
ollected. Risk modelling was performed with backward stepwise multiple logi
stic regression and validated on a separate subset of our patients. Eightee
n preexisting conditions were entered into the model. We adjusted for age,
sex, and duration and type of surgery. Seven associations between pre-exist
ing medical conditions and perioperative adverse events were statistically
significant. Hypertension predicted the occurrence of any intraoperative ev
ent and intraoperative cardiovascular events. Obesity predicted intraoperat
ive and postoperative respiratory events, and smoking and asthma predicted
postoperative respiratory events. Castro-oesophageal reflux predicted intub
ation-related events. The presented models of risk estimation were validate
d internally and provided a useful tool for accurate risk estimation.