Institution Speaker

Christopher Boyer, Ph.D.

Dr. Christopher Boyer

Dr. Christopher Boyer is an assistant professor from Department of Quantitative Health Sciences, Cleveland Clinic. He is interested in causal inference, deep learning, and epidemiology applied to infectious diseases with extensive experience running randomized controlled trials of social and public health interventions.

  • Title: Counterfactual prediction: A framework for estimation, validation, and transportability of models under hypothetical interventions

  • Abstract: Counterfactual prediction methods are required when a model will be deployed in a setting where treatment policies differ from the setting where the model was developed, or when a model provides predictions under hypothetical interventions to support decision-making. However, estimating and evaluating counterfactual prediction models is challenging because, unlike traditional (factual) prediction, one does not observe the full set of potential outcomes for all individuals. Here, we discuss how to fit or tailor a model to target a counterfactual estimand, how to assess the model’s performance, and how to perform model and tuning parameter selection when (1) data are available from a randomized trial in a source population that differs from the target population in which the model will be deployed, (2) data are available from an observational study conducted in the target population, or (3) both are available. We provide identifiability and estimation results for building a counterfactual prediction model and for multiple measures of counterfactual model performance including loss-based measures, the area under the receiver operating characteristics curve, and calibration. Importantly, our results allow valid estimates of model performance under counterfactual intervention even if the candidate model is misspecified, permitting a wider array of use cases. We illustrate these methods using simulation and apply them to representative counterfactual prediction tasks in cardiovascular disease.

Andy Ni, Ph.D.

Dr. Andy Ni

Dr. Andy Ni is an assistant professor from Division of Biostatistics, College of Public Health, The Ohio State University. Dr. Ni’s main research interests are in the individualized treatment regimen estimation with survival outcomes. He is also interested in regularized variable selection and causal inference under two-stage sampling designs with survival outcome. He collaborates with faculties from multiple colleges at OSU and Nationwide Children’s Hospital on the design and analysis of various observational studies and clinical trials.

  • Title: Contrast Weighted Learning for Individualized Treatment Rule Estimation

  • Abstract: Precision medicine aims to tailor medical decisions based on patient-specific characteristics. An individualized treatment rule (ITR) assigns an optimal treatment to a patient based on their personal characteristics. An archetypal ITR estimation approach is outcome-weighted learning (OWL) based on a weighted classification framework with clinical outcomes as the weights. Existing OWL methods are susceptible to irregularities of outcome distributions, such as outliers and heavy tails. Moreover, it is unclear how these methods can be used for multivariate survival outcomes that are commonly encountered in cancer and cardiovascular research. In this study, we propose contrast-weighted learning (CWL) for ITR estimation. CWL exploits the flexibility and robustness of contrast functions between pairs of patients to enable robust ITR estimation for a wide range of clinical outcomes. By introducing win status into the contrast function, CWL naturally handles multivariate survival outcomes while incorporating the clinical importance of different events into ITR estimation. We established the theoretical properties of the estimated ITR. We conducted simulations to evaluate the finite sample performance of CWL in comparison to several alternative ITR estimation methods. Finally, we apply the proposed CWL method to three real datasets to demonstrate its real-world application.

Holly Hartman, Ph.D.

Dr. Holly Hartman

Dr. Holly Hartman is an assistant professor from Department of Population & Quantitative Health Sciences, Case Western Reserve University. Dr. Hartman’s research interests are focused on guiding patient care. She brings her strong statistical background to the design and analysis of clinical trials.

  • Title: Sequential, Multiple Assignment, Randomized Trials (SMARTs) with continuous outcomes and tailoring functions

  • Abstract: Multi-stage trials like sequential, multiple assignment, randomized trials (SMARTs) have typically relied on a binary variable to define response which is used in assigning the next stage treatment assignment. Instead, we develop a function of a continuous outcome to assign a probability of staying on the same treatment and then randomly assign the next treatment using a multinomial distribution. First, we develop a new trial design for small sample SMARTs (snSMARTs). The overall goal of the trial is to determine the optimal first stage treatment. We use a function, called the mapping function, to map the first stage outcome to a probability of staying on the same treatment and Bayesian regression methods to analyze data from both stages. Re-randomization based on a mapping function of a continuous outcome allows for snSMARTs to be conducted without requiring a binary outcome. Next, we apply similar concepts to a standard size SMART with continuous outcomes where the goal is to determine the optimal dynamic treatment regimen (DTR). We present a new trial design for SMARTs that use a tailoring function instead of a binary tailoring variable. In this trial design, we simultaneously develop a tailoring variable and estimate the DTR. We perform simulation studies to compare the proposed design with continuous outcomes to standard designs with binary outcomes. The proposed designs results in more efficient treatment effect estimates and similar or better outcomes for trial patients.