Collections > Electronic Theses and Dissertations > Optimizing the Public Health Benefit of a Medicaid Controlled Substance Lock-In Program

Medicaid controlled substance lock-in programs (MLIP) have garnered increased attention for their potential role in combating prescription drug abuse. MLIPs are purported to enhance care coordination for high-risk beneficiaries by restricting access to a single prescriber and pharmacy for controlled substance service coverage. However, the MLIP evidence base is largely non-existent. The purpose of this dissertation was to examine two threats to MLIP effectiveness: lock-in circumvention through out-of-pocket purchases and lack of standardized MLIP eligibility criteria. Our aims were to investigate: 1) the effect of enrollment in the North Carolina (NC) MLIP on controlled substance circumvention behavior, 2) prescription-level characteristics of circumvented opioid analgesics, and 3) optimal claims-based measures of high-risk opioid use for assessing MLIP eligibility. We used a retrospective cohort of NC MLIP enrollees and linked NC Medicaid claims and Controlled Substances Reporting System data from 10/1/2009-9/30/2012. Generalized estimating equations estimated the effect of MLIP enrollment and covariates on circumvention behavior (Aim 1). Subjects were 3.6 times more likely to obtain a controlled substance prescription through circumvention after MLIP enrollment. Generalized linear models estimated the association of MLIP enrollment with circumvention of opioids with high-risk prescription-level attributes (Aim 2). Mean prescribed daily opioid dose and the likelihood of a circumvented prescription containing a Schedule II opioid product or a long-acting opioid product did not increase after MLIP enrollment. Medicaid claims data for cross-validation cohorts of opioid users in NC Medicaid were used to test and validate optimal measures of high-risk opioid use (Aim 3). The highest performing dichotomous predictors of overdose were selected using survival receiver operating characteristic models and bivariate Cox model fit. The best measures were ≥5 opioid claims and ≥12 daily milligram morphine equivalents over 60 days, but these exhibited low sensitivity in capturing subjects with overdose. MLIP enrollment induced people to circumvent Medicaid to obtain controlled substance prescriptions, negating the underlying purpose of the MLIP intervention. Also, current MLIP eligibility assessment strategies performed poorly in selecting beneficiaries at highest risk for preventable overdose. Reducing MLIP circumvention and optimizing eligibility criteria to better capture highest risk individuals are necessary to bolster MLIP public health benefit.