Background: Prescription drug monitoring programs (PDMP) provide medication-related information about high-risk medicines to support clinical decision-making. This study aims to determine how PDMP alerts and other patient, pharmacy and medication-related characteristics influence pharmacists’ decisions to supply opioids.
Methods: Australian pharmacists completed an online randomised controlled factorial experiment. They were randomly allocated six clinical vignettes, describing a pharmacy patient and asked to indicate on a 0-10 scale, how likely they were to supply the opioid prescription within the vignette. Mixed-effects linear regression models were used to explore the association between vignette and pharmacy-related characteristics and the likelihood to supply opioids. Results: 598 pharmacists provided data relating to 3370 vignettes. The high-dose and multiple prescribers (4 or more prescribers in the past 3 months) PDMP alerts were the strongest predictors of reduced likelihood to supply, with a respective 2.8- and 3.8-unit decrease on the likelihood to supply scale (p=0.001). Unemployment (β=-0.42, p<0.001), and comorbidities of Hepatitis C (β=-0.26, p=0.009) and Depression (β=-0.30, p=0.003) were also significantly associated with decreased likelihood to supply opioids. Medication-related characteristics including high opioid dose (40mg twice daily) and a co-prescription of opioids with benzodiazepines saw a reduced likelihood to supply of 0.26- and 0.48-units, respectively. Implications: PDMP alerts were the most influential factor associated with reduced likelihood to supply, while other factors associated with economic disadvantage and stigma were also associated with a reduced likelihood to supply. It is important that PDMP-related unintended consequences observed elsewhere, including medication refusal or rapid tapering, are avoided locally.