Construction and validation of medication deviation in cancer pain patients with oral opioid formulations during hospital-to-home transition

Objective The transition from hospital to home is a high-risk period for medication errors, particularly in patients receiving opioids. We constructed and validated a Medication Deviation Risk Prediction Model (MDRP) in cancer pain patients during hospital-to-home transition. Methods The medication deviation assessment table was constructed to determine whether there was a medication deviation in the MDRP modeling group. Univariate analysis and logistic regression were used to analyze influencing factors. The model’s goodness of predictive effect was tested using Hosmer-Lemeshow (H-L) and Receiver Operating Characteristic (ROC) curves. External validation was performed using the same methods, and developed a simple risk scoring scale. Results In the modelling group, 33.33% (51/153) had medication deviation, while 66.67% (102/153) had no medication deviation. BPI score, number of comorbidities, presence of long-term caregivers, medication adherence, and presence of anxiety/depression were the five independent influencing factors in the construction of MDRP (P < 0.05). The H-L test yielded p = 0.402, and the area under the ROC curves (AUC) was 0.875, with sensitivity at 0.765 and specificity at 0.882. The validation group results were consistent with the modeling group. a simple risk scoring scale was developed with a total score of 6, a cutoff value of 4, and an AUC of 0.886. The predictive accuracy of the scoring scale was 86.90%. Conclusion The MDRP of cancer pain patients had high sensitivity and specificity. The simple risk scoring scale was convenient and practical in clinical practice.