Objectives: Addiction to prescription narcotics is a global issue, and detecting individuals with narcotic use disorder (NUD) at an early stage can help prevent narcotics misuse and abuse. We developed a novel index for the early detection of NUD based on an analysis of real-world prescription patterns in a large hospital.

Methods: We analyzed the narcotic prescriptions of 221,887 patients, prescribed by 8,737 doctors from July 2000 to June 2018. To facilitate the early detection of patients at risk of developing NUD after a prolonged period of narcotic use, we developed a weighted morphine equivalent daily dose (wt-MEDD) score. This score was based on the number of prescription dates where the actual MEDD exceeded the intended MEDD. We compared the performance of the wt-MEDD scoring system in identifying patients diagnosed with NUD by doctors against other high-risk NUD indices. These indices included the MEDD scoring system, the number of days on prescribed narcotics, the frequency and duration of prescriptions, narcotics prescriptions from multiple doctors, and the number of early narcotic refills.

Results: A wt-MEDD score cut-off value of 10.5 successfully identified all outliers and diagnosed patients with NUD with 100% sensitivity and 99.6% specificity. This score demonstrated the highest sensitivity and specificity for detecting NUD compared to all other indexes. The predictive performance was further improved by combining the wt-MEDD score with other high-risk NUD indexes.

Conclusion: We developed a novel index, the wt-MEDD score, which showed excellent performance in the early detection of NUD.