Association of Medicaid Managed Care Drug Carve Outs With Hepatitis C Virus Prescription Use

Key Points Question Did the use of direct-acting antiviral hepatitis C (HCV) medications change after these medications were carved out from Medicaid managed care organization (MCO) coverage and financed fee-for-service in 4 state Medicaid programs? Findings In this cross-sectional study, carve outs were associated with a mean quarterly increase of 22.1 HCV prescription fills per 100 000 Medicaid enrollees compared with synthetic control states, translating to a relative increase of 86%. Meaning This study suggests that carve outs of HCV medications from Medicaid MCO coverage may increase access to these medications, potentially improving the health of Medicaid enrollees with HCV and reducing the economic burden of untreated HCV on the US health care system.

The table displays the percentage that each state contributed to the synthetic control. Numbers may not sum to 100 due to rounding.
--indicate a state received no weight in that synthetic control model eMethods

Sample and Measurement of Outcome Variables
The pre-period for our analysis ran from January 1, 2015 to each states' respective effective quarter for MCO medication carve-outs, and our post-period was the two years following the carve-out effective quarter (see Appendix II for effective dates and post-periods for each state). We included states in our analysis if they had at least one year of pre-period data following a HCV medication carve-out, provided criteria for HCV medication access publicly, and did not implement subscription-based payment models during the study period. Justifications for the exclusion of specific states are discussed in the main text. Over final "donor pool" included the following: Alabama, Arizona, Arkansas, California, Colorado, Connecticut, District of Columbia, Florida, Georgia, Hawaii, Idaho, Kansas, Kentucky, Maine, Maryland, Massachusetts, Minnesota, Mississippi, Missouri, Nevada, New Jersey, New Mexico, New York, North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, Tennessee, Texas, Utah, Vermont, Virginia, and Wisconsin.

Unadjusted Comparisons
Unadjusted changes in study outcomes were identified by comparing the four calendar quarters preceding carve-outs with the following eight quarters. Significance levels and confidence intervals were calculated using two-sided t-tests.

Synthetic Control Comparisons
The basic idea of synthetic control methods is to use a weighted average of each state's donor pool, with the weights chosen so that pre-trends in outcomes were as similar as possible between the treated state(s) and synthetic control. 1 We followed the approach outlined in Robbins, Saunders, and Kilmer (2016) and implemented in package 'microsynth' for R Statistical Software version 4.0.2. 2 We estimated separate synthetic controls models for each treated state and study outcome.
Taylor series linearization (TSL) was then used to calculate estimates and 95% confidence intervals for the effects of SBPM implementation during the following year. This approach accounts for the complex weighting of control states used to create the synthetic control. TSL uses a linear function of the observed data to approximate the estimator, and the variance estimation formulae for a linear estimator can then be applied to the linear approximation. In general, this leads to a statistical consistent estimator of the variance. 3 The TSL method tells us whether any differences in outcome between the treated states and their synthetic controls are statistically significant.
We also conducted a series of permutation tests to determine placebo effect sizes. This procedure occurs in three steps. First, we subset the donor pool of control states for a given treated state and outcome. We then iteratively reassign the "treatment" to each control state. Weights are then calculated to match the placebo treatment to a new synthetic control, and a placebo effect, sampling distribution and associated pvalue are generated. Weights were selected to minimize not only outcome pre-trends, but also differences in state liver damage and sobriety restrictions during the pre-period. The number of permutation tests varied by state; individual permutation tests that could not achieve a suitable match during the pre-period were dropped when calculating results. This was defined as a mean squared error (MSE) between treated state and synthetic control of greater than 1. Any MSE cutoff is by definition arbitrary, thus we also tested thresholds of 0.5 and 1.5 and achieved highly similar results. We then use a two-sided t-test to determine whether the observed effect in the treated state is likely to have occurred by chance, given the distribution of placebo effect estimates. 4 eFigure 1. Trends of Hepatitis C mortality per 100,000 residents in treated states and nationally.
Notes. Dotted lines indicate the quarter carve-outs occurred in treated states.