MEMP PhD Thesis Defense: Joe Hakim

Causal Machine Learning for Safety Critical Decisions in Obstetrics, Pharmacovigilance, and Critical Care

Understanding treatment effects of e.g. drugs using data besides randomized controlled trials (RCTs) would greatly broaden the evidence base used for evidence-based medicine yet needs experts to evaluate untestable assumptions and can require large observational datasets. However, done carefully, causal inference has been used in policy decisions to supplement our understanding of the effectiveness of specific drugs, and this thesis defense presentation focuses on a case of 17-hydroxyprogesterone caproate, a drug originally approved for reducing the risk of spontaneous recurrent preterm birth. Originally having received accelerated approval contingent on a follow-up trial, two conflicting RCTs gave rise to a need and opportunity to analyze historical insurance records to perform a target trial analysis of the effectiveness of the drug, which we performed with insurance data. After careful adjustments and trial design to emulate the RCTs' structure, extensive sensitivity analyses and subgroup analyses, we found no evidence of risk reduction. Our report, among several other such reports and argumentation, demonstrated to the FDA a consistency and a preponderance of evidence that they used to withdrawal of the drug from the market. This defense describes the causal inference framework, but also the evidence and debate process of public policy making and its intersection with the incentives of the public and pharmaceutical companies. We will also very briefly touch on other works done, including a language model guardrails platform we developed for pharmacovigilance in industry, and some new techniques in clustering for causal longitudinal data in intensive care unit settings for sepsis.

Thesis Supervisor:
Andrew Beam, Ph.D.
Associate Professor of Epidemiology, Harvard T.H. Chan School of Public Health

Thesis Committee Chair:
Marzyeh Ghassemi, Ph.D.
Associate Professor, Electrical Engineering and Computer Science and Institute for Medical Engineering and Science, MIT 

Thesis Reader:
Jordan Smoller, M.D., Sc.D.
MGH Trustees Endowed Chair in Psychiatric Neuroscience, Professor of Psychiatry, Harvard Medical School
Professor of Epidemiology, Harvard T.H. Chan School of Public Health
________________________________________________________________________________________

Zoom Invitation
Joe Hakim is inviting you to a scheduled Zoom meeting

Topic: Joe Hakim MEMP PhD Thesis Defense
Time: Monday, February 9, 2026, 2:30 PM Eastern Time (US and Canada)

Your participation is important to us: please notify hst [at] mit.edu (hst[at]mit[dot]edu), at least 3 business days in advance, if you require accommodations in order to access this event.

Join Zoom Meeting
https://mit.zoom.us/j/95108109196 

Password: 563563

One tap mobile
+16465588656,,95108109196# US (New York)
+16699006833,,95108109196# US (San Jose)

Meeting ID: 951 0810 9196

US: +1 646 558 8656 or +1 669 900 6833

International Numbers: https://mit.zoom.us/u/adlGFd5YEa

Join by SIP
95108109196 [at] zoomcrc.com (95108109196[at]zoomcrc[dot]com)

Join by Skype for Business
https://mit.zoom.us/skype/95108109196