Dear Editor: I read the article by Yennurajalingam et al. (Reference Yennurajalingam, Edwards and Arthur2021a) with interest. The authors developed a nomogram to predict the risk of non-medical opioid use (NMOU) among cancer patients receiving outpatient supportive care. They defined the risk of NMOU (+risk) as Screener and Opioid Assessment for Patients with Pain (SOAPP-14) score of ≥7. By the multivariate logistic regression model, the probability of +risk was estimated. +risk was significantly associated with gender, race, marital status, smoking status, depression, anxiety, financial distress, morphine equivalent daily dose (MEDO), and Cut Down-Annoyed-Guilty-Eye Opener (CAGE-AID) score. They established a practical nomogram to assess the +risk, which can be conducted by routinely collected clinical data, and I present a comment about their study.

The same authors reported the independent predictors for NMOU behavior (Yennurajalingam et al., Reference Yennurajalingam, Arthur and Reddy2021b). A total of 19% patients developed NMOU behavior within a median duration of eight weeks after initial supportive care clinic consultation. Adjusted hazard ratio (95% CI) of single, divorced, SOAPP score of ≥7, MEDD, and Edmonton Symptom Assessment Scale pain level for the presence of NMOU behavior were 1.58 (1.15–2.18), 1.43 (1.01–2.03), 1.35 (1.04–1.74), 1.003 (1.002–1.004), and 1.11 (1.06–1.16), respectively. This paper presented that NMOU behavior could be significantly associated with five factors, and SOAPP score of ≥7 was one of the significant contributors for NMOU behavior. Although the authors established a practical nomogram to assess the +risk, NMOU behavior cannot be estimated adequately by their practical nomogram. I recommend the authors presenting a nomogram for estimating NMOU behavior instead of +risk.