Coverage and Care Delivery in the Medicaid Program
Paper Session
Saturday, Jan. 4, 2025 8:00 AM - 10:00 AM (PST)
- Chair: M. Kate Bundorf, Duke University
Prior Authorization and Inappropriate Antipsychotic Prescribing to Children on Medicaid
Abstract
An important open question in health economics is the extent to which provider behavior can be affected by regulation. We study this question in the context of antipsychotic prescribing to children on Medicaid. These medications have substantial side effects, including long-term on metabolic health. Yet, although they are recommended for only a few indications, prescribing rates are particularly high for children on Medicaid, creating concerns about inappropriate prescribing and health equity. Starting in 2005, many state Medicaid programs rolled out prior authorization requirements for antipsychotic prescribing to children. This paper evaluates the effects of these policies on provider prescribing behavior. Using IQVIA prescription data from 2006 to 2019 covering the entirety of the United States, we find that prior authorization reduced prescriptions to children under the age of seven by 26 percent in the first four years of the policy’s implementation. We explore heterogeneity by provider specialty, training, and prescribing volumes at baseline. Inappropriate prescribing occurs across specialties but happens more frequently when providers have less specialized training. Data on antidepressant prescribing is also examined to measure potential substitution to or complementarities with unaffected psychotropic drugs.Data on Medicaid prior authorization policy data spanning nearly two decades was hand collected through a combination of archival research and Freedom of Information Act requests. Original source documents including Medicaid formularies, provider notices and bulletins, Drug Utilization Review Board minutes, and Pharmacy and Therapeutics Committee minutes were reviewed in order to track changes in all 50 states from 2005 to 2022. These documents reveal substantial heterogeneity in both the design and the implementation of pharmaceutical prior authorization policies, which may also have an impact on their effectiveness.
Navigating Medicaid: Experimental Evidence on Administrative Frictions and Loss of Coverage at Renewal
Abstract
Many Americans who are eligible for means-tested safety net programs do not receive them. The Medicaid program, which provides health insurance coverage for certain low-income Americans, has a long history of low-take-up and low retention considering the value of the benefit. In 2023, 9 million Medicaid beneficiaries lost their coverage because they did not successfully complete the required administrative processes to redetermine their eligibility. Because ineligible people have little reason to complete these processes, however, the impact of administrative frictions on access to Medicaid remains unclear.We conducted a field experiment to test whether outreach methods designed to reduce administrative frictions increased retention of Medicaid coverage. The experimental population included over 100,000 households in Wisconsin enrolled in fee-for-service Medicaid who must renew or lose their coverage during the period from May 2023 to June 2024. The outreach we tested was designed to reduce administrative frictions by a) reminding the recipient about their upcoming Medicaid renewal deadline, and b) by providing instructions to connect with a chatbot or call a hotline offering free, one-on-one assistance with the renewal process. The implementing partner was Wisconsin’s navigator program, Covering Wisconsin. Experimental arms varied the modality of outreach, content of message, and number of messages; while everyone received either text or postcard outreach, half the sample additionally received a pre-recorded outbound call. We used administrative data to assess whether individuals in this population submitted Medicaid renewal applications and whether they lost Medicaid coverage at the redetermination deadline (for any reason, or for procedural reasons). The pre-analysis plan was posted on the AEA RCT registry (https://www.socialscienceregistry.org/trials/9772).
Data on Medicaid enrollment after the redetermination deadline are available for 76,405 households; complete data will be available before the conference. Being randomized to the outbound call arm significantly decreased loss of Medicaid coverage at the redetermination
Quality of Childhood Preventive Care Delivered by Non-Physicians: Evidence from Medicaid Claims
Abstract
To supplement the strained physician workforce and meet growing healthcare demand, nurse practitioners (NP) and physician assistants (PA) are seeing an ever-growing proportion of patients, particularly in the primary care setting. Overall, the share of primary care visits delivered by NPPs nearly doubled between 2013 and 2019. Given the training and licensing differences between NPs, PAs, and physicians, this growth may have implications for quality of care. However, few studies have explored these implications, and even fewer have focused on settings where quality of care has particularly high returns, such as in pediatric preventive medicine.To address this gap, this paper uses Michigan Medicaid claims data for children born between 2013 and 2020 to estimate the effect of using NPPs for early childhood (ages 0 -2) primary care on the likelihood of compliance with official preventive care periodicity schedules. To overcome the endogeneity in choice of provider, we instrument for the percent of visits with an NPP in early childhood by whether the child saw an NPP at the first well-visit. We provide empirical evidence supporting the relevance and exogeneity of the instrument, and as such it is valid for estimating the effect of receiving more early childhood care from an NPP on preventive care outcomes.
Results indicate that children seeing NPs in early childhood do not exhibit notable differences in preventive care use relative to children seeing physicians. But children seeing PAs have 28% fewer well-visits and are on schedule for vaccinations 57% fewer months than children exclusively seeing physicians. We find that this poor quality of care results in subsequently worse outcomes for children seeing PAs in terms of avoidable emergency department visits. In addition, we study whether the quality of care by NPPs is influenced by the extent of concordance, or demographic match, between provider and patient.
Discussant(s)
Benjamin Chartock
,
Bentley College
Anirban Basu
,
University of Washington
Kosali Simon
,
Indiana University
Rebecca Staiger
,
University of California-Berkeley
JEL Classifications
- I0 - General
- H0 - General