Associations between implementation of Project Lazarus and opioid analgesic dispensing and buprenorphine utilization in North Carolina, 2009-2014

Background: Project Lazarus (PL) is a seven-strategy, community-coalition-based intervention designed to reduce opioid overdose and dependence. The seven strategies include: community education, provider education, hospital emergency department policy change, diversion control, support programs for patients with pain, naloxone policies, and addiction treatment expansion. PL was originally developed in Wilkes County, NC. It was made available to all counties in North Carolina starting in March 2013 with funding of up to $34,400 per county per year. We examined the association between PL implementation and 1) overall dispensing rate of opioid analgesics, and 2) utilization of buprenorphine. Buprenorphine is often used in connection with medication assisted treatment (MAT) for opioid dependence.

Methods: Observational interrupted time series analysis of 100 counties over 2009-2014 (n = 7200 county-months) in North Carolina. The intervention period was March 2013-December 2014. 74 of 100 counties implemented the intervention. Exposure data sources comprised process surveys, training records, Prescription Drug Monitoring Program (PDMP) data, and methadone treatment program quality data. Outcomes were PDMP-derived counts of opioid prescriptions and buprenorphine patients. Incidence Rate Ratios were estimated with adjusted GEE Poisson regression models of all seven PL strategies.

Results: In adjusted models, diversion control efforts were positively associated with increased dispensing of opioid analgesics (IRR: 1.06; 95% CI: 1.03, 1.09). None of the other PL strategies were associated with reduced prescribing of opioid analgesics. Support programs for patients with pain were associated with a non-significant decrease in buprenorphine utilization (IRR: 0.93; 95% CI: 0.85, 1.02), but addiction treatment expansion efforts were associated with no change in buprenorphine utilization (IRR: 0.98; 95% CI: 0.91, 1.06).

Conclusions: Implementation of PL strategies did not appreciably reduce opioid dispensing and did not increase buprenorphine utilization. These results are consistent with previous findings of limited impact of PL strategies on overdose morbidity and mortality. Future studies should analyze the uptake of MAT using a more expansive view of institutional barriers, treating community coalition activity around MAT as an effect modifier.