Table 4 Table of five example patients, their contributory predictor variables with their model-derived predicted risk of consuming an opioid in the first week after surgical discharge, alongside the prediction classification at a 50% threshold, the real-world clinician decision and real-world patient consumption outcome
Age | Gender | Smoking status | Procedure | BUPA category | Urgency | Discharge day OME used on the day of discharge | Prior 6-month opioid use | Total OME consumed after discharge | Predicted risk of consuming opioids | Binary prediction | Clinician decision | Patient consumption |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
62 | Female | No | Knee arthroplasty | MAJOR | Elective | 65 | Yes | 334.6 | 0.906 | Yes | Yes | Yes |
21 | Male | No | Appendicectomy | INTER | Emergency | 25 | No | 80 | 0.514 | Yes | Yes | Yes |
24 | Male | No | Appendicectomy | INTER | Elective | 0 | No | 0 | 0.024 | No | No | No |
77 | Male | No | Inguinal hernia repair | MINOR | Elective | 40 | No | 26 | 0.2 | No | Yes | Yes |
58 | Female | No | Colorectal resection | MAJOR | Elective | 0.4 | No | 0 | 0.078 | No | No | No |
