What is normal? A critical analysis of the methods quantifying prescription drug use and potential misuse in pharmaceutical claims

Our overall objective is to examine methods of quantifying prescription drug misuse in pharmaceutical claims. We approach this by undertaking a systematic review of the global literature measuring the extent of prescription drug misuse in pharmaceutical claims. Our review highlights four measures, number of prescribers, number of dispensing pharmacies, volume of drug dispensed and number of early refills, are used frequently to define prescription drug misuse. Despite this homogeneity, we found heterogeneity in the thresholds delineating use from misuse and a lack of established or validated benchmarks to accurately measure misuse in pharmaceutical claims. In our empirical work, we focus on prescription opioid analgesics due to the recent and considerable global increase in use and opioid- related harms. We use publically available, routinely collected data to document increases in prescription opioid use and related harms in Australia over 20 years. Over three chapters we explore population norms of prescription drug access in national dispensing claims and examine how access patterns relate to the metrics defining ‘misuse’ identified in our systematic review. We compare prescription drug access in Australia and British Columbia, Canada, for prescription opioids and statins, drug classes with high or no known abuse potential, respectively. We found access norms are remarkably similar across drug classes and healthcare settings. However, extreme access patterns are more common in people dispensed opioids, younger age groups or those receiving income assistance. We then examine opioid access in Australian adults initiating or reinitiating strong opioid treatment. We found the standard metrics defining ‘misuse’, including doctor and pharmacy shopping, are non-specific in that they identify misuse, but are also likely to capture high-need patient groups including individuals with a history of cancer treatment. From a translational perspective, our findings are particularly important as the US Food and Drug Administration recently endorsed using routinely collected data, including pharmaceutical claims, to quantify prescription opioid misuse and measure the effectiveness of interventions aimed to curb the ‘opioid epidemic’. We recommend using these commonly established metrics with caution due to their inability to isolate a population of people misusing opioids