Severe Food Allergies and Anxiety: Patient and Family Effects
Abstract
A severe food allergy demands continuous vigilance over an everyday activity and is known to co-occur with anxiety. Using the timing of first diagnosis in U.S. commercial claims, we use staggered difference-in-differences design to show severe food allergies cause anxiety and that this anxiety is not only confined to the patient, but present in the household as well. For a patient, a diagnosis raises the quarterly probability of an anxiety diagnosis by 1.9 percentage points (41.0%) with no pre-trend and continued growth over five years. The effect is specific to food allergies: drug- and venom-induced anaphylaxis, carrying the same trauma and label but no daily vigilance, show no comparable response. Real treatment follows in clinical order, and the burden extends to caregivers, namely mothers.
(Draft available upon request)
Medical Assistance in Dying and Suicide Rates Among Older Adults: Evidence from Canada
with Julian Reif, Jennifer Stewart, and Coady Wing
Abstract
A growing number of countries permit medical professionals to help patients end their lives, yet how these laws affect suicide, defined as deaths from intentional self-harm without medical assistance, remains unresolved. We examine whether Canada’s introduction of Medical Assistance in Dying (MAID) in 2016 affected suicide rates among adults aged 65 and older. Using annual mortality data from Canada and the United States for 2000–2024, we compare changes in suicide rates in Canada with contemporaneous changes in U.S. states without assisted-dying laws. Suicide rates in Canada declined substantially after legalization relative to the U.S. comparison group: by 4.8 deaths per 100,000 for men (24.7%) and 0.9 deaths per 100,000 for women (21.9%). These declines are consistent with partial substitution from suicide to MAID, but are much smaller than the increase in MAID deaths. As more jurisdictions consider assisted-dying laws, the Canadian experience suggests that legalization may partially displace suicide but that most medically assisted deaths do not replace deaths that would otherwise have occurred by suicide.
(Draft available upon request)
with Jaehyun Choi, Seth Freedman, and Coady Wing
Abstract
This book chapter introduces difference-in-differences designs as a core tool for causal inference and explains why standard two-way fixed effects estimators can be misleading under staggered treatment timing or heterogeneous treatment effects. It develops the classic two-group, two-period framework and then extends the discussion to two modern alternatives: stacked difference-in-differences and imputation-based estimators. The chapter emphasizes the importance of separating estimation from aggregation across units and time, providing applied researchers with a framework for implementing credible, interpretable, and robust DID analyses.
The Impact of Women's Health Clinic Closures on STIs
Abstract
Abortion-providing clinics deliver a substantial volume of non-abortion reproductive care, including testing and treatment for sexually transmitted infections (STIs). We ask whether the closure of these clinics affects STI testing, diagnosis, and treatment among the commercially insured. We measure clinic access with a county-by-month panel of driving distances to the nearest abortion facility [Myers, 2026], aggregate it to the metropolitan statistical area, and define a staggered “clinic-closure” treatment as a sustained 25-mile increase in distance above a 2014–2015 baseline. We link this treatment to STI outcomes built from MarketScan commercial claims over 2016–2023 and estimate dynamic effects with the Callaway and Sant’Anna [2021] staggered difference-in-differences estimator. We find no significant effect on STI testing rates. The significant effects fall on diagnosis and treatment and are mixed in direction: syphilis and any-STI treatment rise, while chlamydia treatment and chlamydia diagnosis decline among men. Because the commercially insured face fewer access barriers than the populations most dependent on these clinics, we read these estimates as a lower bound and identify replication in Medicaid claims as the direct next step.
(Draft available upon request)
Measuring Infant Mortality Using Synthetic Survival Curves
with Seth Freedman and Coady Wing
Abstract
This study develops and validates a novel synthetic control framework for estimating the effects of policy interventions on survival functions. Estimating counterfactual survival functions may provide insight into the way that interventions affect outcomes with a duration structure at different follow up times. We study the performance of the method using data on linked infant birth-death data to measure survival over the first year of life for multiple birth cohorts in each US state, and simulating treatment effects from hypothetical interventions. Results from simulated placebo tests and power analysis show that a synthetic survival curve estimators produces stable counterfactual estimates and is able to detect relatively small treatment effects with high statistical power. Power analyses indicate that the method reliably detects survival rate reductions as small as 0.3% with 80% power. These validation results demonstrate the method’s potential in evaluating the effects of changes in abortion policy on infant mortality.
(Draft available upon request)
Past Research/Undergraduate Publication
Exitability and Economic Freedom: Evidence from the U.S.
International Advances in Economic Research (2017, with Josh Hall)