5A-05 - Assessing the Impact of Florida's Opioid Prescribing Restriction Law on Over-the-Counter Analgesic Sales: Single and Controlled Interrupted Time Series Analyses

Background: Opioid prescribing restriction laws for acute pain have led to significant declines in opioid utilization, though corresponding increases in prescription non-steroidal anti-inflammatory drugs (NSAIDs) have not been observed. It is unknown whether over-the-counter (OTC) non-prescription analgesic products are acting as opioid substitutes. Importantly, drug utilization research on non-prescription medications has been limited by the absence of OTC product data in administrative claims.

Objectives: To compare effect estimates from interrupted time series (ITS) analyses with and without negative internal control outcomes assessing the impact of a state (Florida) opioid prescribing restriction law on OTC analgesic sales.

Methods: We used NielsenIQ Retail Scanner Data from 1,624 Florida stores (food, drug, mass merchandise, and convenience stores) from January 2015 to December 2019. Monthly OTC analgesic sales, adjusted for inflation to 2019 prices, were analyzed combined and by main ingredient (aspirin, ibuprofen, naproxen and acetaminophen). We used generalized linear models accounting for autocorrelation to assess immediate (level change) and ongoing impacts (trend change) of the law’s July 2018 implementation. Controlled ITS analyses were performed using antacid OTC medications as a negative internal control group to further adjust for time-varying confounding. Results from single and controlled ITS analyses were compared to assess potential bias of effect estimates by secular sales trends.

Results: In uncontrolled ITS analyses, significant post-intervention decreasing trend changes were observed for sales of aspirin and naproxen products (-$22,579 per month [95%CI -$30,528, -$14,630] and -$11,990 per month [95%CI -$13,063, -$10,925], respectively), and non-significant decreases for sales of other OTC analgesics. In controlled ITS analyses, we observed a significant post-intervention increase in the trend of acetaminophen sales ($25,907 per month [95%CI $14,495, $37,320]) and immediate post-intervention level change for naproxen ($176,601 [95%CI $62,819, $290,383]).

Conclusions: In contrast to findings from uncontrolled analyses, accounting for secular trends using internal negative control showed a significant increase in the use acetaminophen OTC products after the opioid restriction law. Antacid sales proved to be a suitable negative control group, as they closely mirrored OTC analgesic sales trends before the intervention. This study underscores the importance of selecting appropriate controls in ITS analyses – without adequate controls, results may be distorted and can obfuscate the interpretation of policy effects.