The ecology of prescription opioid abuse in the USA: geographic variation in patients' use of multiple prescribers ("doctor shopping").

PURPOSE: This study estimates the prevalence in US counties of opioid patients who use large numbers of prescribers, the amounts of opioids they obtain, and the extent to which their prevalence is predicted by ecological attributes of counties, including general medical exposure to opioids.
METHODS: Finite mixture models were used to estimate the size of an outlier subpopulation of patients with suspiciously large numbers of prescribers (probable doctor shoppers), using a sample of 146 million opioid prescriptions dispensed during 2008. Ordinary least squares regression models of county-level shopper rates included independent variables measuring ecological attributes of counties, including rates of patients prescribed opioids, socioeconomic characteristics of the resident population, supply of physicians, and measures of healthcare service utilization.
RESULTS: The prevalence of shoppers varied widely by county, with rates ranging between 0.6 and 2.5 per 1000 residents. Shopper prevalence was strongly correlated with opioid prescribing for the general population, accounting for 30% of observed county variation in shopper prevalence, after adjusting for physician supply, emergency department visits, in-patient hospital days, poverty rates, percent of county residents living in urban areas, and racial/ethnic composition of resident populations. Approximately 30% of shoppers obtained prescriptions in multiple states.
CONCLUSIONS: The correlation between prevalence of doctor shoppers and opioid patients in a county could indicate either that easy access to legitimate medical treatment raises the risk of abuse or that drug abusers take advantage of greater opportunities in places where access is easy. Approaches to preventing excessive use of different prescribers are discussed.