Health Services/Policy
Can BRFSS be used to assess the impacts of the Dobbs v. Jackson Women’s Health Organization US Supreme Court decision on healthcare utilization? A preliminary robustness analysis Andrea Molino* Andrea Molino Rachel Winer Vida Henderson Lyndsey Benson Steve Mooney
The June 2022 Dobbs v. Jackson Supreme Court decision created a state-by-state patchwork of abortion policies. Its impact on non-abortion healthcare utilization, such as cervical cancer (CC) screening, is underexplored. When 2024 Behavioral Risk Factor Surveillance System (BRFSS) data become available, we plan to leverage state policy variation and implement a difference-in-differences (DID) design to assess Dobbs’ impact. Concerned that COVID-related healthcare changes may affect our ability to estimate causal effects, we performed preliminary robustness checks on available data.
We used 3 pre-Dobbs BRFSS waves (2018, 2020, 2022). Inclusion criteria were identifying as female, 30-49 years, no hysterectomy or CC history, not currently pregnant, and survey completion pre-Dobbs. State inclusion was determined by post-Dobbs abortion access and state Medicaid expansion stability. Robustness checks included: 1) test of prior trends, 2) placebo test (intervention artificially backdated to 2021), and 3) balancing tests for changes in intervention group composition (age, education, urban/rural).
We included 52,085 participants from 27 states, with 88.3% in 22 abortion accessible and 11.7% in 5 abortion banned states. In 2018, 2020, and 2022, 91.3%, 89.7%, and 58.2%, respectively, were considered CC screening compliant, with 2022 decline likely due to COVID. Prior trends F-statistic for year/group interaction terms was statistically significant (p=0.03). The placebo test did not show that artificial 2021 intervention led to greater screening declines in banned states (OR=0.90, 95%CI: 0.75-1.07). Models assessing group composition did not provide evidence that age (F-stat. p=0.52), education (p=0.11), or urban/rural status (p=0.74) changed over time.
Robustness checks support that DID may yield valid estimates of Dobbs’ impact on CC screening, which will be crucial when the full analysis is performed with 2024 BRFSS data. Prior trends result may reflect large sample size.
