Validation of a prediction model for post-discharge nausea and vomiting after general anaesthesia in a cohort of Swedish ambulatory surgery patients.

Abstract

In ambulatory surgery, post-discharge nausea and vomiting (PDNV) has been identified as a significant problem occurring in more than one-third of patients. To validate a simplified PDNV score in a Swedish population. Prospective observational study. Two county hospitals in Sweden: Sundsvall from June 2012 to May 2013 and Sunderbyn from January to October 2014. Adult patients undergoing ambulatory surgery under general anaesthesia. Postoperative outcomes with a focus on nausea and vomiting were collected at 2, 4, and 6 h after surgery and on the first three postoperative days. The simplified PDNV score, calculated before discharge, included the factors: female sex, age less than 50 years, history of postoperative nausea and vomiting, postoperative nausea and opioids given postoperatively. The prediction performance of the simplified PDNV score was evaluated in terms of discrimination (area under receiver-operating characteristics curve) and calibration plots and was compared with that of the original development study. A total of 559 patients were asked to participate, of which 431 were included in the final study cohort. The overall risk of postoperative nausea and vomiting and PDNV were 18.8 [95% confidence interval (CI), 15.4-22.8]% and 28.1 (95% CI, 24.0-32.5)%, respectively. The discrimination capacity of the simplified PDNV score in our study was similar to that of the original dataset [area under the curve 0.693 (95% CI, 0.638-0.748) vs. 0.706 (0.681-0.731), absolute difference 0.013]. The slope of the calibration curve was 0.893, with a constant of 0.021 (R-square 0.884). In a Swedish cohort of patients, the simplified PDNV score performs well in discriminating between patients who will experience post-discharge nausea and/or vomiting after ambulatory surgery. Our results indicate that the simplified PDNV score is as valid in other cohorts as it was in the original development cohort.