Correction to “Geographic Variation in the Prescription of Schedule II Opioid Analgesics among Outpatients in the United States”

Abstract

In this issue of HSR, we report findings from our analysis of geographic variations in the prescription of Schedule II opioid analgesics in the United States (Curtis et al. 2006). We write now to inform HSR readers of inaccuracies in the report. It has come to our attention that the “total claims” field in the county-level data set we received was mislabeled. Instead of total claims, the field reflected the number of subjects in each county according to an alternate set of inclusion and exclusion criteria. Believing that the total claims field was accurately labeled, we reported a 12-fold variation in claim rates for Schedule II oral opioid analgesics at the state level, and variation of a similar magnitude in claim rates for controlled-release oxycodone. We also reported that the presence of a state-based prescription monitoring program was associated with lower claim rates at the county level. We now know that the denominator used in those calculations was not claims, but rather a measure of the number of subjects. Because of a corporate acquisition since the time of the analysis, the raw data file is no longer available. As a consequence, it is not possible to recalculate the claim rates as claims per 1000 total claims. Nevertheless, using the known subject population as the denominator, we recalculated the claim rates and updated Table 2 and Figure 1 accordingly. We also corrected the title and row headings of Table 3 to reflect that the denominator of the dependent variable is subjects, not claims. The recalculated data support our original conclusion that considerable variation exists in the prescription of oral opioid analgesics, including controlled-release oxycodone. However, because the denominator refers to subjects rather than claims, we cannot eliminate the possibility that the geographic variations we observed reflect geographic variations in overall prescription drug use. In other words, if claims were the denominator, geographic variations in the use of oral opioid analgesics could be distinguished from geographic variations in the use of total prescription drugs. However, with subjects as the denominator, we cannot account for the variations due to different intensity of overall prescription drug use by region. Figure 1 (Revised): Claim Rates by State for All Opioid Analgesics and Controlled-Release Oxycodone. The figure also displays the year in which state-based Schedule II prescriptionmonitoring programs were first enacted in the states that have such a program. Table 2 (Revised). Claims for Opioid Analgesics by Location* Table 3 (Revised). Multivariable Regression Models for County-Level Opioid Analgesic Prescription Claim Rates per 1000 Subjects*