About

Turgay Ayer is the Virginia C. and Joseph C. Mello Chair & Professor at the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Institute of Technology. The provided page primarily lists his students and their academic achievements, with no specific details about his research focus, background, or key contributions. Therefore, there is no additional biographical information available on this page.

Research topics

• Political Science
• Computer Science
• Medicine
• Artificial Intelligence
• Economics
• Geography
• Econometrics
• Business
• Machine Learning
• Engineering
• Environmental health
• Actuarial science
• Operations research
• World Wide Web
• Surgery
• Virology
• Demography
• Meteorology
• Statistics
• Biology
• Internal medicine
• Data science
• Psychiatry
• Mathematics

Selected publications

• Sa2016 COST-EFFECTIVENESS OF PHARMACOLOGIC THERAPIES FOR METABOLIC DYSFUNCTION-ASSOCIATED STEATOHEPATITIS WITH F2-F3 FIBROSIS IN THE UNITED STATES

  Gastroenterology · 2026-05-01

  article
  PublisherDOI

• Transfer Learning for Meta-analysis Under Covariate Shift

  ArXiv.org · 2026-04-03

  articleOpen accessSenior author

  Randomized controlled trials often do not represent the populations where decisions are made, and covariate shift across studies can invalidate standard IPD meta-analysis and transport estimators. We propose a placebo-anchored transport framework that treats source-trial outcomes as abundant proxy signals and target-trial placebo outcomes as scarce, high-fidelity gold labels to calibrate baseline risk. A low-complexity (sparse) correction anchors proxy outcome models to the target population, and the anchored models are embedded in a cross-fitted doubly robust learner, yielding a Neyman-orthogonal, target-site doubly robust estimator for patient-level heterogeneous treatment effects when target treated outcomes are available. We distinguish two regimes: in connected targets (with a treated arm), the method yields target-identified effect estimates; in disconnected targets (placebo-only), it reduces to a principled screen--then--transport procedure under explicit working-model transport assumptions. Experiments on synthetic data and a semi-synthetic IHDP benchmark evaluate pointwise CATE accuracy, ATE error, ranking quality for targeting, decision-theoretic policy regret, and calibration. Across connected settings, the proposed method is best or near-best and improves substantially over proxy-only, target-only, and transport baselines at small target sample sizes; in disconnected settings, it retains strong ranking performance for targeting while pointwise accuracy depends on the strength of the working transport condition.

  PublisherOA PDF

• Sa2016 COST-EFFECTIVENESS OF PHARMACOLOGIC THERAPIES FOR METABOLIC DYSFUNCTION-ASSOCIATED STEATOHEPATITIS WITH F2-F3 FIBROSIS IN THE UNITED STATES

  Gastrointestinal Endoscopy · 2026-05-01

  article
  PublisherDOI

• Capacity allocation in a school-based asthma care model for pediatric patients

  IISE Transactions on Healthcare Systems Engineering · 2026-05-15

  articleCorresponding
  PublisherDOI

• Transfer Learning for Meta-analysis Under Covariate Shift

  arXiv (Cornell University) · 2026-04-03

  preprintOpen accessSenior author

  Randomized controlled trials often do not represent the populations where decisions are made, and covariate shift across studies can invalidate standard IPD meta-analysis and transport estimators. We propose a placebo-anchored transport framework that treats source-trial outcomes as abundant proxy signals and target-trial placebo outcomes as scarce, high-fidelity gold labels to calibrate baseline risk. A low-complexity (sparse) correction anchors proxy outcome models to the target population, and the anchored models are embedded in a cross-fitted doubly robust learner, yielding a Neyman-orthogonal, target-site doubly robust estimator for patient-level heterogeneous treatment effects when target treated outcomes are available. We distinguish two regimes: in connected targets (with a treated arm), the method yields target-identified effect estimates; in disconnected targets (placebo-only), it reduces to a principled screen--then--transport procedure under explicit working-model transport assumptions. Experiments on synthetic data and a semi-synthetic IHDP benchmark evaluate pointwise CATE accuracy, ATE error, ranking quality for targeting, decision-theoretic policy regret, and calibration. Across connected settings, the proposed method is best or near-best and improves substantially over proxy-only, target-only, and transport baselines at small target sample sizes; in disconnected settings, it retains strong ranking performance for targeting while pointwise accuracy depends on the strength of the working transport condition.

  PublisherDOI

• EPH162 Modeling the Long-Term Bone Health Impact of TDF/FTC PrEP in Adolescent MSM

  Value in Health · 2025-12-01

  article1st authorCorresponding
  PublisherDOI

• 1056-P: Cost-Effectiveness of the WHO Diabetes Compact Targets for Diabetes Control in India—A Microsimulation Modelling Study

  Diabetes · 2025-06-13

  article

  Introduction and Objective: Diabetes affects over 100 million adults in India, but fewer than one in 10 are achieving optimal glycemic (HbA1c < 8.0%), blood pressure (BP; <140/90 mmHg), and LDL cholesterol (<100mg/dl or statin use) control. The WHO Diabetes Compact recommends target achievement for these metrics among people with diabetes to achieve broad population benefits. We estimated the health and economic impacts of achieving these targets for India to inform future health system planning. Methods: We developed a micro-simulation model to estimate the cost-effectiveness of attaining three Compact targets among those with diagnosed diabetes: 80% glycemic control, 80% BP control, and 60% statin use (for those >40 years). We simulated 20-year rates of complications (cardiovascular events, retinopathy, neuropathy, nephropathy) and mortality using data from the Longitudinal Aging Study in India (n = 9,678; 52.4% female). Clinical intervention effectiveness was based on national treatment guidelines and parameters from prior models. We evaluated health system costs in 2023 INT$ and INR₹, health status in disability adjusted life years (DALYs), and the incremental cost-effectiveness ratio (ICER) as cost per DALY averted. Results: Over 20 years achievement of three WHO Compact targets was associated with reductions of 22.8% in cases of retinopathy, 13.9% in nephropathy, 15.6% in foot ulcers, 26.1% in total myocardial infarctions, 30.0% in strokes, 28.0% in cardiovascular-related deaths. Total health system costs were lower in the optimal scenario, resulting in a mean cost savings of INT$630.20 (₹12,730.00) per patient, with an average 0.104 DALYs averted. Achieving these three Compact targets was cost saving, with an ICER of INT$-6059.60 (₹-122,403.85) per DALY averted. Conclusion: Attainment of WHO Diabetes Compact targets is expected to be cost saving for the Indian healthcare system and deliver improved health outcomes and reduced healthcare costs. Disclosure E.L. Kocher: None. A.A. Abdeen: None. Y. Long: None. A. Baker: None. A. Sharma: None. T. Ayer: Consultant; GlaxoSmithKline plc, Gilead Sciences, Inc. J. Julien: None. M.K. Ali: Advisory Panel; Eli Lilly and Company. N. Tandon: None. J. Varghese: None. G. Garcia: None.

  PublisherDOI

• MSR155 Assessing the Effectiveness of Large Language Models in Automating Systematic Literature Reviews: Findings from Recent Studies

  Value in Health · 2025-07-01

  article
  PublisherDOI

• Causal Clustering for Conditional Average Treatment Effects Estimation and Subgroup Discovery

  ArXiv.org · 2025-09-06

  preprintOpen access

  Estimating heterogeneous treatment effects is critical in domains such as personalized medicine, resource allocation, and policy evaluation. A central challenge lies in identifying subpopulations that respond differently to interventions, thereby enabling more targeted and effective decision-making. While clustering methods are well-studied in unsupervised learning, their integration with causal inference remains limited. We propose a novel framework that clusters individuals based on estimated treatment effects using a learned kernel derived from causal forests, revealing latent subgroup structures. Our approach consists of two main steps. First, we estimate debiased Conditional Average Treatment Effects (CATEs) using orthogonalized learners via the Robinson decomposition, yielding a kernel matrix that encodes sample-level similarities in treatment responsiveness. Second, we apply kernelized clustering to this matrix to uncover distinct, treatment-sensitive subpopulations and compute cluster-level average CATEs. We present this kernelized clustering step as a form of regularization within the residual-on-residual regression framework. Through extensive experiments on semi-synthetic and real-world datasets, supported by ablation studies and exploratory analyses, we demonstrate the effectiveness of our method in capturing meaningful treatment effect heterogeneity.

  PublisherOA PDFDOI

• MSR30 Evaluating Generative AI in Replicating Health Economic Models: A Case Study on Ulcerative Colitis

  Value in Health · 2025-07-01

  article
  PublisherDOI

Recent grants

• CAREER: Optimal Management of Chronic Diseases Caused by Infections

  NSF · $500k · 2015–2021

• SCH: EXP: Smart Adaptive Adherence-Enhancing Intervention Strategies for Breast Cancer Prevention

  NSF · $290k · 2017–2020

• SCH: INT: Collaborative Research: Smart Intervention Strategies for Hepatitis C Elimination

  NSF · $430k · 2017–2023

Frequent coauthors

• Jagpreet Chhatwal

  Massachusetts General Hospital

  271 shared

• Neehar D. Parikh

  University of Michigan–Ann Arbor

  88 shared

• A. Burak Ozbay

  Exact Sciences (United States)

  87 shared

• Benjamin Haaland

  University of Utah

  82 shared

• Kyle Porter

  81 shared

• Carol Kirshner

  74 shared

• Shubham Chakankar

  The University of Texas Southwestern Medical Center

  72 shared

• Qiushi Chen

  Pennsylvania State University

  65 shared

Labs

• Turgay Ayer LabPI

Awards & honors

• Career Award, National Science Foundation (2015)
• Seth Bonder Foundation Research Award (2012)
• First Prize, INFORMS Doing Good with Good OR Competition (20…
• Second Prize, INFORMS MSOM Society Student Paper Competition…
• Finalist in INFORMS Decision Analysis Society Student Paper…