WORKS IN PROGRESS
MEDICAL MARIJUANA LAWS AND PERINATAL HEALTH
With Angelica Meinhofer, Allison Witman, and Kosali Simon.
Past month marijuana use by pregnant women increased 103% between 2003 and 2017, nearly twice the growth among the general U.S. population. Observational studies suggest that marijuana use in pregnancy is associated with adverse newborn health outcomes including stillbirth, low birth weight, and low gestational age. We study the association between marijuana liberalization policies and newborn health with a difference-in-differences design that exploits cross-state variation in effective dates of medical marijuana laws (MML) and recreational marijuana laws (RML). We find no association between MMLs and neonatal mortality, low gestational age, and low birth weight. In contrast, we find a 3% increase in low birth weight and low gestational age after RML implementation, along with growth in exposure to noxious substances via placenta or breast milk and in neonatal drug withdrawal syndrome. This growth is accompanied by increases in maternal substance use and use disorder, including marijuana use disorder. Our findings should not be interpreted as capturing an association between newborn health and marijuana use, do not reflect all outcomes that could be affected by marijuana liberalization policies, and need not extend to all subpopulations.
MEDICAID APPLICATIONS SPIKE DURING MARKETPLACE OPEN ENROLLMENT: LESSONS FROM COVERED CALIFORNIA.
With Paul Shafer.
Several states already have or are pursuing Medicaid 1115 waivers that would restrict or eliminate retroactive eligibility, which typically offers coverage for any medical expenses incurred for the 90 days prior to application. One of the key arguments for these restrictions has been to encourage people to enroll in Medicaid as soon as they are eligible, making Medicaid enrollment more like private insurance. In states using HealthCare.gov, the open enrollment period during which anyone can enroll is now only 45 days (November 1 to December 15). Individuals cannot enroll for the rest of the year unless they qualify for a special enrollment period and can prove that a qualifying event occurred (e.g., change in income or family size). Some states have extended their open enrollment periods to give individuals a longer time to enroll. However, individuals who are Medicaid eligible can enroll year-round. In this study, we assess whether Marketplace open enrollment influences Medicaid enrollment patterns. In an environment with perfect information, we would expect to see a uniform distribution of Medicaid enrollment throughout the year or at least one that mirrors seasonality in employment. However, many individuals who attempt to enroll in Marketplace coverage during open enrollment may find out they are eligible for Medicaid. A large increase in Medicaid applications during Marketplace open enrollment would indicate two potential information gaps, that 1) people don’t know that they are eligible and 2) people don’t know that they can enroll year-round. We assembled a 3-year time series of statewide monthly Medicaid applications for California (Medi-Cal) from public state and federal reports for July 2016 to June 2019, deriving daily averages to account for variation in number of days in each month. We used a log-linear ordinary least squares model with month and year fixed effects to estimate the effect of Marketplace open enrollment on Medicaid applications. Over one-third of all Medicaid applications during this time period (34.6%) were received during months with Marketplace open enrollment. We find a 32.5% increase (p<0.05) in daily average Medicaid applications during the months with Marketplace open enrollment, accounting for seasonality and year-specific differences. These results have implications both for state-level Medicaid policy and administrative burdens on social service agencies. A lack of uniformity in the distribution of Medicaid applications creates uneven administrative burdens for state social service agencies and unnecessary gaps in coverage when patients may be foregoing care and becoming non-adherent to medication. In California, like many other states, eligibility redetermination takes place 12 months from the approval date, meaning that peak new and renewal application caseloads are occurring during months with prominent holidays and less available staff time. Our findings also create concern about whether there is enough consumer education and outreach to potential enrollees to limit gaps in coverage and associated barriers in access to care.
THE IMPLEMENTATION & SUSTAINMENT FACILITATION (ISF) STRATEGY: COST AND COST-EFFECTIVENESS RESULTS FROM A 39-SITE CLUSTER RANDOMIZED TRIAL INTEGRATING SUBSTANCE USE SERVICES IN AIDS SERVICE ORGANIZATIONS
With Bryan Garner, Colleen Watson, Stephen Tueller, Elizabeth Ball, and Rasika Ramanan.
Background: Substance use among people with HIV is both prevalent and problematic, yet the integration of substance use treatment within HIV service settings is rare. The Substance Abuse Treatment to HIV Care Project was funded to test an organization-focused strategy called Implementation & Sustainment Facilitation (ISF) as an adjunct to the Addiction Technology Transfer Center (ATTC) strategy. This presentation presents the cost and cost-effectiveness results from this cluster randomized implementation experiment.
Methods: Thirty-nine AIDS Service Organizations (ASO) and two brief intervention (BI) staff per ASO (N=78) were randomized to receive (1) the ATTC strategy (ATTC only) or (2) the ATTC strategy plus the ISF strategy (ATTC+ISF). We estimated costs using primary data on the time spent on the ATTC strategy, on the ISF strategy, and implementing BIs. Salary information was collected via staff surveys. We conducted a cost-effectiveness analysis at the staff level on the number of BIs implemented, the overall MIBI quality scores achieved, and total average patient days abstinent.
Results: Per BI staff costs were $3,176 for the ATTC strategy and $5,752 for the ATTC+ISF strategy, resulting in an incremental cost of $2,576. The incremental difference was approximately 4 for BIs delivered and 780 for MIBI quality scores achieved, yielding incremental cost effectiveness ratios of $644 and $3, respectively. Patients receiving a BI in the ATTC+ISF condition averaged 60 more days of abstinence at follow-up per BI staff, yielding an incremental cost effectiveness ratio of $42.
Discussion: The ISF strategy was found to be a cost-effective adjunct to the ATTC’s current state-of-the-art implementation strategy. The current finding is important given that it suggests ISF as a promising strategy to improve the integration of substance use treatment within ASOs.
REVISITING THE AFFORDABLE CARE ACT’S TAX CREDITS AND COST-SHARING SUBSIDIES: A FOUR-YEAR FOLLOW-UP ON THE INCENTIVES TO PURCHASE INDIVIDUAL HEALTH INSURANCE
With David Anderson.
In 2014, the Patient Protection and Affordable Care Act provided a heterogenous set of incentives to purchase health insurance in the individual market. Consumers with incomes at 100% of the federal poverty level (FPL) gain eligibility for substantial premium tax credits and cost-sharing reductions, lose eligibility for cost-sharing reductions at 250% FPL, and lose eligibility for premium tax credits at 400% FPL. Previous work has found coverage effects at the lowest FPL thresholds and highest thresholds. Since 2014, the ACA exchanges have experienced dynamic changes and this study assesses changes in the effects of the Marketplace subsidies on insurance coverage as the ACA Marketplace premiums and regulations have evolved across times. Using the Current Population Survey and American Community survey, we assess whether individual insurance cover changes in response discontinuous changes in the subsidies at three income cutoffs dynamically across time in response to various incentives. Analyses suggest that the response at the 138% FPL cutoff in 2014 has attenuated across time in Medicaid expansion states. There is evidence of increased coverage near 250% FPL in Medicaid expansion states in 2017 and increases in insurance coverage in 2016 and 2017 in non-expansion states at 100% FPL which correspond to increases in the effective value of the tax credits. We also find evidence that downstream out-of-pocket medical expenditures may have decreased.
DO EMPLOYEES REALLY FACE AN EARNINGS TRADEOFF FROM INCREASING EMPLOYER-SPONSORED HEALTH INSURANCE PREMIUMS?
This study re-examines a seminal question in the health and labor economics literature: Is there a tradeoff between pecuniary benefits (i.e., wages) and non-pecuniary benefits (i.e., employer-sponsored health insurance premiums). Several studies in the 1990s and early 2000s presented empirical evidence of a large tradeoff, however, recent evidence for an earnings penalty is mixed. We provide a review of the existing literature and examine the role of sample selection and selection bias among these studies. In an empirical exercise using data from the Current Population Survey, we show that large earnings penalties may be due to the inclusion of part-time or part-year workers and that earnings penalties may be offset by employers increasing employee paid premiums.