Publications

A pipeline for enabling path-specific causal fairness in observational health data

Published in ArXiv [Preprint], 2026

We present a model-agnostic pipeline for training causally fair machine learning models (according to path-specific causal fairness) that combines data-driven analysis of bias with with known contexts of health disparities.

Recommended citation: Kashyap A, Matijevic S, Elhadad N, Kushner SA , Joshi S. A pipeline for enabling path-specific causal fairness in observational health data. ArXiv [Preprint]. 2026 January. doi: https://arxiv.org/abs/2601.09841 https://arxiv.org/abs/2505.16941

FoMoH: A clinically meaningful foundation model evaluation for structured electronic health records

Published in ArXiv [Preprint], 2025

We benchmark foundation models for structured electronic health record data, focusing on clinically relevant tasks and assessing models for discrimination, calibration, and fairness.

Recommended citation: Pang C, Jeanselme V, Choi YS, Jiang X, Jing Z, Kashyap A, Kobayashi Y, Li Y, Pollet F, Natarajan K, Joshi S. FoMoH: A clinically meaningful foundation model evaluation for structured electronic health records. ArXiv [Preprint]. 2025 June. doi: https://arxiv.org/abs/2505.16941 https://arxiv.org/abs/2505.16941

Something Similar: Exploring the Usefulness of an On-the-Spot Meal Recommendation System for Health Goal Attainment

Published in Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, 2025

We conduct a feasibility study to evaluate the utility of similarity-based recommendations in constrained settings.

Recommended citation: Desai PM, Raj A, Albers D, Kashyap A, Mamykina L. 2025. Something Similar: Exploring the Usefulness of an On-the-Spot Meal Recommendation System for Health Goal Attainment. In Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (CHI EA '25). Association for Computing Machinery, New York, NY, USA, Article 490, 1–8. https://doi.org/10.1145/3706599.3719813 https://dl.acm.org/doi/abs/10.1145/3706599.3719813

Investigating racial disparities in drug prescriptions for patients with endometriosis

Published in npj Women's Health, 2025

We assess racial disparities in medication prescription patterns for endometriosis patients across Medicaid administrative claims data, and find larger disparities pre-diagnosis compared to post-diagnosis.

Recommended citation: Kashyap, A., Aziz, M., Sun, T.Y., Lipsky Gorman, S., Opoku-Anane, J., Elhadad, N. Investigating racial disparities in drug prescriptions for patients with endometriosis. npj Womens Health 3, 6 (2025). https://doi.org/10.1038/s44294-025-00053-3 https://www.nature.com/articles/s44294-025-00053-3

The Impact of Evolving Endometriosis Guidelines on Diagnosis and Observational Health Studies

Published in medRxiv [Preprint], 2024

We define five cohorts of endometriosis patients, each based on different diagnostic criteria, and examine population differences between individuals within each phenotype.

Recommended citation: Reyes Nieva H, Kashyap A, Voss EA, Ostropolets A, Anand A, Ketenci M, DeFalco FJ, Choi YS, Li Y, Allen MN, Guang S, Natarajan K, Ryan P, Elhadad N. The Impact of Evolving Endometriosis Guidelines on Diagnosis and Observational Health Studies. medRxiv [Preprint]. 2024 December. doi: 10.1101/2024.12.13.24319010. https://pmc.ncbi.nlm.nih.gov/articles/PMC11702737/

Trade-offs in concentration sensing in dynamic environments

Published in Biophysical Journal, 2024

We model a eukaryotic cell sensing a chemical secreted from bacteria and develop analytical calculations and stochastic simulations of sensing in this environment. We find that cells can have a huge variety of optimal sensing strategies ranging from not time averaging at all to averaging over an arbitrarily long time or having a finite optimal averaging time.

Recommended citation: Kashyap A, Wang W, Camley BA. Trade-offs in concentration sensing in dynamic environments. Biophysical Journal, 123(10). May 2024. 10.1016/j.bpj.2024.03.025 https://www.cell.com/biophysj/fulltext/S0006-3495(24)00205-4

Using machine learning on clinical data to identify unexpected patterns in groups of COVID-19 patients

Published in Nature Scientific Reports, 2023

We leverage JH-CROWN: The COVID Precision Medicine Analytics Platform Registry to identify subgroups of COVID-19 patients who are at high risks for severe disease progression.

Recommended citation: Cowley HP, Robinette MS, Matelsky JK, Xenes D, Kashyap A., Ibrahim NF, Robinson ML, Zeger S, Garibaldi BT, Gray-Roncal W. Using machine learning on clinical data to identify unexpected patterns in groups of covid-19 patients. Scientific Reports, 13(1). February 2023. https://doi.org/10.1038/s41598-022-26294-9 https://www.nature.com/articles/s41598-022-26294-9

Aparajita Kashyap