About

Alyssa Bilinski, PhD, is the Peterson Family Assistant Professor of Health Policy at Brown University School of Public Health, within the Departments of Health Services, Policy & Practice and Biostatistics. Her research focuses on the intersection of policy modeling and evaluation, developing novel methods to support decision-making and applying these methods to identify interventions that can most efficiently improve population health and well-being. She concentrates on applications in infectious disease and maternal health. Dr. Bilinski has published her work in leading peer-reviewed journals, and her research has been covered by major news outlets. She has collaborated with state, local, and federal public health officials to translate her research into practice and has served as a committee member for the National Academies of Sciences, Engineering, and Medicine. She holds a PhD in Health Policy (Evaluative Science & Statistics) from Harvard University, an AM in Statistics from Harvard, an MSc in Medical Statistics from the London School of Hygiene and Tropical Medicine as a Marshall Scholar, and a BA from Yale College.

Research topics

• Environmental health
• Computer Science
• Medicine
• Virology
• Internal medicine
• Demography
• Political Science
• Psychology
• Artificial Intelligence
• Business
• Internet privacy
• Statistics
• World Wide Web
• Public relations
• Epistemology
• Economics
• Psychiatry
• Econometrics
• Finance
• Geography
• Gerontology

Selected publications

• Oversights in global gynaecological disability measurement

  The Lancet · 2026-01-01

  article1st authorCorresponding
  PublisherDOI

• Nothing to See Here? A Non‐Inferiority Approach to Parallel Trends

  Statistics in Medicine · 2026-02-01 · 103 citations

  preprintOpen access1st authorCorresponding

  Difference-in-differences is a popular method for observational health policy evaluation. It relies on a causal assumption that in the absence of intervention, treatment groups' outcomes would have evolved in parallel to those of comparison groups. Researchers frequently look for parallel trends in the pre-intervention period to bolster confidence in this assumption. The popular "parallel trends test" evaluates a null hypothesis of parallel trends and, failing to find evidence against the null, concludes that the assumption holds. This tightly controls the probability of falsely concluding that trends are not parallel, but may have low power to detect non-parallel trends. When used as a screening step, it can also introduce bias in treatment effect estimates. We propose a non-inferiority/equivalence approach that tightly controls the probability of missing large violations of parallel trends, measured on the scale of the treatment effect. Our framework nests several common use cases, including linear trend tests and event studies. We show that our approach may induce no or minimal bias when used as a screening step under commonly assumed error structures and, absent violations, can offer a higher-power alternative to testing treatment effects in more flexible models. We illustrate our ideas by reconsidering a study of the impact of the Affordable Care Act's dependent coverage provision.

  PublisherOA PDFDOI

• Abortion May Be Controversial—Supporting Children and Families Need Not Be

  JAMA · 2025-02-13

  article1st authorCorresponding
  PublisherDOI

• Sins of Omission: Model-Based Estimates of the Health Effects of Excluding Pregnant Participants From Randomized Controlled Trials

  Annals of Internal Medicine · 2025-04-28 · 7 citations

  review1st authorCorresponding

  BACKGROUND: More than 90 million women in the United States have given birth. Randomized controlled trials (RCTs) of medications almost always exclude pregnant participants. OBJECTIVE: To quantify the health effects of excluding pregnant participants from RCTs. DESIGN: Decision analytic framework applied to case studies of thalidomide, COVID-19 vaccines, and dolutegravir. SETTING: Varied. PARTICIPANTS: Pregnant people and their children. MEASUREMENTS: The authors modeled the ex post facto health effects of RCTs, comparing projected health effects of medication uptake had an RCT been conducted versus historically observed outcomes. They also modeled the a priori health effects that could have been anticipated in trial planning. They converted health effect estimates to monetary value using standard benchmarks. RESULTS: Across case studies, health benefits from conducting RCTs during pregnancy were projected to far exceed expected adverse effects (AEs) from RCTs. For example, had thalidomide been tested in a completed RCT with 200 treated participants, about 33 children would have experienced severe AEs, whereas knowledge from the RCT would have prevented 8000 thalidomide-related birth defects, 99.6% of all thalidomide-related birth defects from 1956 to 1962. Likewise, if RCTs for COVID-19 vaccines had included pregnant participants and if posttrial pregnant uptake were conservatively assumed to mirror that of age- and state-matched nonpregnant women, a projected 20% of COVID-19-related maternal deaths and stillbirths (8% of all maternal deaths and 1% of all stillbirths) in the United States would have been prevented from March to November 2021. Across case studies, the a priori value of RCT data would have exceeded the approximately $100 million cost of phase 1 to 3 RCTs. LIMITATION: Parameter uncertainty. CONCLUSION: Systematic inclusion in RCTs could benefit both pregnant people and their children by both speeding AE detection and increasing uptake of beneficial medications. PRIMARY FUNDING SOURCE: None.

  PublisherDOI

• Estimating Time-Varying Epidemic Severity Rates with Adaptive Deconvolution

  ArXiv.org · 2025-10-17

  preprintOpen access

  Several key metrics in public health convey the probability that a primary event will lead to a more serious secondary event in the future. These "severity rates" can change over the course of an epidemic in response to shifting conditions like new therapeutics, variants, or public health interventions. In practice, time-varying parameters such as the case-fatality rate are typically estimated from aggregate count data. Prior work has demonstrated that commonly-used ratio-based estimators can be highly biased, motivating the development of new methods. In this paper, we develop an adaptive deconvolution approach based on approximating a Poisson-binomial model for secondary events, and we regularize the maximum likelihood solution in this model with a trend filtering penalty to produce smooth but locally adaptive estimates of severity rates over time. This enables us to compute severity rates both retrospectively and in real time. Experiments based on COVID-19 death and hospitalization data, both real and simulated, demonstrate that our deconvolution estimator is generally more accurate than the standard ratio-based methods, and displays reasonable robustness to model misspecification.

  PublisherOA PDFDOI

• Trends in Maternal, Fetal, and Infant Mortality in the US, 2000-2023

  JAMA Pediatrics · 2025-04-28 · 9 citations

  articleOpen access

  Importance: Accurately measuring maternal mortality trends has been challenging due to changes in data collection. This work disambiguates trends from the effects of introducing the pregnancy checkbox on death certificates and also analyzes closely related fetal and infant mortality. Objective: To describe trends in maternal, fetal, and infant deaths since 2000, including the impact of the COVID-19 pandemic. Design, Setting, and Participants: A national, population-level, epidemiological, cross-sectional analysis during 2000 to 2023 was conducted as well as a staggered difference-in-differences analysis on the pregnancy checkbox, using the US Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research (WONDER) database on underlying causes of death in the US to identify maternal, infant, and fetal deaths. Study population was restricted to mothers aged 15 to 44 years for all definitions of maternal mortality. Exposures: Staggered introduction of the pregnancy checkbox on death certificates across different states. Main Outcomes and Measures: Longitudinal study (2000-2023) reporting crude rates per 100 000 population for adjusted maternal mortality and per 1000 population for fetal and infant mortality at the national level and by US Census Bureau-designated main census regions, age groups, and race and ethnicity. Staggered difference-in-differences counterfactuals (1999-2023) on impact of pregnancy checkbox. Results: The introduction of the pregnancy checkbox was associated with 6.78 (95% CI, 1.47-12.09) deaths per 100 000 live births increase in reported maternal mortality, 66% (95% CI, 14%-117%) of the total increase from 2000 to 2019, with a smaller impact on maternal mortality excluding cause unspecified (adjusted maternal death rates). Adjusted maternal death rates remained consistently between 6.75 (95% CI, 5.97-7.61) to 10.24 (95% CI, 9.22-11.34) per 100 000 live births from 2000 until 2021, when it peaked at 18.86 (95% CI, 17.48-20.32); the rate dropped to 10.23 (95% CI, 9.22-11.32) in 2022. The death rates of Native American or Alaska Native women increased the most during the COVID-19 period, almost tripling from 2011 to 2019 (10.70 per 100 000 live births; 95% CI, 7.64-14.57) to the 2020 to 2022 period (27.47 per 100 000 live births; 95% CI, 18.39-39.45). The death rates of non-Hispanic Black women were highest across time-approximately triple the rate of non-Hispanic White women in each time period. Infant death rates per 1000 live births dropped from 6.93 (95% CI, 6.85-7.01) in 2000 to 5.44 (95% CI, 5.36-5.51) in 2020, increasing slightly to 2018 levels in 2021 to 2023. Fetal death rates per 1000 live births decreased from 6.28 (95% CI, 6.16-6.31) in 2005 to 5.53 (95% CI, 5.45-5.60) in 2022. Conclusion and Relevance: Using difference-in-differences analyses, results of this study reveal that the pregnancy checkbox explained much of the observed increase in maternal mortality before the COVID-19 pandemic. Nevertheless, results of this cross-sectional study suggest that, even adjusting for pregnancy checkbox effects, most groups saw increases from 2011 to 2019 to the 2020 to 2022 period, indicating that the COVID-19 pandemic led to worse outcomes. The findings demonstrate the relevance of public health emergencies to maternal health outcomes.

  PublisherOA PDFDOI

• Difference‐in‐Differences for Health Policy and Practice: A Review of Modern Methods

  Statistics in Medicine · 2025-10-01 · 6 citations

  reviewSenior authorCorresponding

  Difference-in-differences (DiD) is a popular observational causal inference method in health policy, employed to evaluate the real-world impact of policies and programs. To estimate treatment effects, DiD relies on a "parallel trends assumption" that treatment and comparison groups would have had parallel trajectories on average in the absence of an intervention. Recent years have seen both growing use of DiD in health policy and medicine and rapid advancements in DiD methods. To support DiD implementation in these fields, this paper reviews and synthesizes best practices and recent innovations. We provide recommendations to practitioners in four areas: (1) assessing causal assumptions; (2) adjusting for covariates and other approaches to relax causal assumptions; (3) accounting for staggered treatment timing; and (4) conducting robust inference, especially when normal-based clustered standard errors are inappropriate. For each, we explain challenges and common pitfalls in traditional DiD and recommend methods to address these. We explore current treatment of these topics through a focused literature review of medical DiD studies.

  PublisherDOI

• Defining and Estimating Outcomes Directly Averted by a Vaccination Program when Rollout Occurs Over Time

  ArXiv.org · 2025-09-05

  preprintOpen access

  During the COVID-19 pandemic, estimating the total deaths averted by vaccination has been of great public health interest. Instead of estimating total deaths averted by vaccination among both vaccinated and unvaccinated individuals, some studies empirically estimated only "directly averted" deaths among vaccinated individuals, typically suggesting that vaccines prevented more deaths overall than directly due to the indirect effect. Here, we define the causal estimand to quantify outcomes "directly averted" by vaccination$\unicode{x2014}$i.e., the impact of vaccination for vaccinated individuals, holding vaccination coverage fixed$\unicode{x2014}$for vaccination at multiple time points, and show that this estimand is a lower bound on the total outcomes averted when the indirect effect is non-negative. We develop an unbiased estimator for the causal estimand in a one-stage randomized controlled trial (RCT) and explore the bias of a popular "hazard difference" estimator frequently used in empirical studies. We show that even in an RCT, the hazard difference estimator is biased if vaccination has a non-null effect, as it fails to incorporate the greater depletion of susceptibles among the unvaccinated individuals. In simulations, the overestimation is small for averted deaths when infection-fatality rate is low, as for many important pathogens. However, the overestimation can be large for averted infections given a high basic reproduction number. Additionally, we define and compare estimand and estimators for avertible outcomes (i.e., outcomes that could have been averted by vaccination, but were not due to failure to vaccinate). Future studies can explore the identifiability of the causal estimand in observational settings.

  PublisherOA PDFDOI

• Difference-in-differences analysis with repeated cross-sectional survey data

  Health Services and Outcomes Research Methodology · 2025-10-31

  articleOpen access

  Difference-in-differences (DiD) approach is one of the most widely used approaches for evaluating policy effects. However, traditional DiD methods may not recover the population-level average treatment effect on the treated (ATT) in the absence of population-level panel data, particularly when the composition of units in the treatment group changes over time. In this work, we address the following two challenges when applying DiD methods with repeated cross-sectional (RCS) survey data: (1) heterogeneous compositions of study samples across different time points, and (2) availability of data for only a sample of the population. We introduce a policy-relevant target estimand and establish its identification conditions. We then propose a new weighting approach that incorporates both estimated propensity scores and given survey weights. We establish the theoretical properties of the proposed method and examine its finite-sample performance through simulations. Finally, we apply our proposed method to a real-world data application, estimating the effect of a beverage tax on adolescent soda consumption in Philadelphia.

  PublisherOA PDFDOI

• Trends in Maternal, Fetal, and Infant Mortality in the US, 2000–2023

  Obstetrical & Gynecological Survey · 2025-10-01

  article

  (Abstracted from JAMA Pediatr 2025;179(7):765–772) Measuring trends in maternal mortality has been challenging to study due to data reporting changes in the United States. One contributing factor was the staggered introduction of a pregnancy checkbox on death certificates across states.

  PublisherDOI

Frequent coauthors

• Joshua A. Salomon

  Stanford University

  34 shared

• Meagan C. Fitzpatrick

  University of Maryland, Baltimore

  17 shared

• Andrea Ciaranello

  Harvard University

  14 shared

• Michelle Wilson

  Brown University

  9 shared

• Joseph W. Hogan

  AMPATH

  9 shared

• Laura C. Chambers

  Brown University

  9 shared

• Taylor Fortnam

  Rhode Island Department of Health

  9 shared

• Roberta DeVito

  Rhode Island Department of Health

  9 shared

Education

• Ph.D., Health Policy (Evaluative Science & Statistics)

  Harvard University

  2021

• M.S., Medical Statistics

  The London School of Hygiene and Tropical Medicine

  2015

• B.A.

  Yale College

  2013