Search for: All records

Controlled experiments are widely applied in many areas such as clinical trials or user behavior studies in IT companies. Recently, it is popular to study experimental design problems to facilitate personalized decision making. In this paper, we investigate the problem of optimal design of multiple treatment allocation for personalized decision making in the presence of observational covariates associated with experimental units (often, patients or users). We assume that the response of a subject assigned to a treatment follows a linear model which includes the interaction between covariates and treatments to facilitate precision decision making. We define the optimal objective as the maximum variance of estimated personalized treatment effects over different treatments and different covariates values. The optimal design is obtained by minimizing this objective. Under a semi-definite program reformulation of the original optimization problem, we use a YALMIP and MOSEK based optimization solver to provide the optimal design. Numerical studies are provided to assess the quality of the optimal design. 

A key and challenging step toward personalized/precision medicine is the ability to redesign dose-finding clinical trials. This work studies a problem of fully response-adaptive Bayesian design of phase II dose-finding clinical trials with patient information, where the decision maker seeks to identify the right dose for each patient type (often defined as an effective target dose for each group of patients) by minimizing the expected (over patient types) variance of the right dose. We formulate this problem by a stochastic dynamic program and exploit a few properties of this class of learning problems. Because the optimal solution is intractable, we propose an approximate policy by an adaptation of a one-step look-ahead framework. We show the optimality of the proposed policy for a setting with homogeneous patients and two doses and find its asymptotic rate of sampling. We adapt a number of commonly applied allocation policies in dose-finding clinical trials, such as posterior adaptive sampling, and test their performance against our proposed policy via extensive simulations with synthetic and real data. Our numerical analyses provide insights regarding the connection between the structure of the dose-response curve for each patient type and the performance of allocation policies. This paper provides a practical framework for the Food and Drug Administration and pharmaceutical companies to transition from the current phase II procedures to the era of personalized dose-finding clinical trials. Funding: This research is supported by the National Science Foundation [Grant 1651912]. Supplemental Material: The online appendices are available at https://doi.org/10.1287/serv.2022.0306 . 

Identifying the right dose is one of the most important decisions in drug development. Adaptive designs are promoted to conduct dose-finding clinical trials as they are more efficient and ethical compared with static designs. However, current techniques in response-adaptive designs for dose allocation are complex and need significant computational effort, which is a major impediment for implementation in practice. This study proposes a Bayesian nonparametric framework for estimating the dose-response curve, which uses a piecewise linear approximation to the curve by consecutively connecting the expected mean response at each dose. Our extensive numerical results reveal that a first-order Bayesian nonparametric model with a known correlation structure in prior for the expected mean response performs competitively when compared with the standard approach and other more complex models in terms of several relevant metrics and enjoys computational efficiency. Furthermore, structural properties for the optimal learning problem, which seeks to minimize the variance of the target dose, are established under this simple model. Summary of Contribution: In this work, we propose a methodology to derive efficient patient allocation rules in response-adaptive dose-finding clinical trials, where computational issues are the main concern. We show that our methodologies are competitive with the state-of-the-art methodology in terms of solution quality, are significantly more computationally efficient, and are more robust in terms of the shape of the dose-response curve, among other parameter changes. This research fits in “the intersection of computing and operations research” as it adapts operations research techniques to produce computationally attractive solutions to patient allocation problems in dose-finding clinical trials. 

In this work, we study the optimal design of two-armed clinical trials to maximize the accuracy of parameter estimation in a statistical model, where the interaction between patient covariates and treatment are explicitly incorporated to enable precision medication decisions. Such a modeling extension leads to significant complexities for the produced optimization problems because they include optimization over design and covariates concurrently. We take a min-max optimization model and minimize (over design) the maximum (over population) variance of the estimated interaction effect between treatment and patient covariates. This results in a min-max bilevel mixed integer nonlinear programming problem, which is notably challenging to solve. To address this challenge, we introduce a surrogate optimization model by approximating the objective function, for which we propose two solution approaches. The first approach provides an exact solution based on reformulation and decomposition techniques. In the second approach, we provide a lower bound for the inner optimization problem and solve the outer optimization problem over the lower bound. We test our proposed algorithms with synthetic and real-world data sets and compare them with standard (re)randomization methods. Our numerical analysis suggests that the proposed approaches provide higher-quality solutions in terms of the variance of estimators and probability of correct selection. We also show the value of covariate information in precision medicine clinical trials by comparing our proposed approaches to an alternative optimal design approach that does not consider the interaction terms between covariates and treatment. Summary of Contribution: Precision medicine is the future of healthcare where treatment is prescribed based on each patient information. Designing precision medicine clinical trials, which are the cornerstone of precision medicine, is extremely challenging because sample size is limited and patient information may be multidimensional. This work proposes a novel approach to optimally estimate the treatment effect for each patient type in a two-armed clinical trial by reducing the largest variance of personalized treatment effect. We use several statistical and optimization techniques to produce efficient solution methodologies. Results have the potential to save countless lives by transforming the design and implementation of future clinical trials to ensure the right treatments for the right patients. Doing so will reduce patient risks and reduce costs in the healthcare system. 

The primary objective of this paper is to develop computationally efficient methods for optimal stopping of an adaptive Phase II dose-finding clinical trial, where the decision maker may terminate the trial for efficacy or abandon it as a result of futility. We develop two solution methods and compare them in terms of computational time and several performance metrics such as the probability of correct stopping decision. One proposed method is an application of the one-step look-ahead policy to this problem. The second proposal builds a diffusion approximation to the state variable in the continuous regime and approximates the trial’s stopping time by optimal stopping of a diffusion process. The secondary objective of the paper is to compare these methods on different dose-response curves, particularly when the true dose-response curve has no significant advantage over a placebo. Our results, which include a real clinical trial case study, show that look-ahead policies perform poorly in terms of the probability of correct decision in this setting, whereas our diffusion approximation method provides robust solutions. 

We developed a model to compare the impacts of different lifestyle interventions among prediabetes individuals and to identify the optimal age groups for such interventions. A stochastic simulation was developed to replicate the prediabetes and diabetes trends (1997–2010) in the U.S. adult population. We then simulated the population-wide impacts of three lifestyle diabetes prevention programs, i.e., the Diabetes Prevention Program (DPP), DPP-YMCA, and the Healthy Living Partnerships to Prevent Diabetes (HELP-PD), over a course of 10, 15 and 30 years. Our model replicated the temporal trends of diabetes in the U.S. adult population. Compared to no intervention, the diabetes incidence declined 0.3 per 1,000 by DPP, 0.2 by DPP-YMCA, and 0.4 by HELP-PD over the 15-year period. Our simulations identified HELP-PD as the most cost-effective intervention, which achieved the highest 10-year savings of $38 billion for those aged 25–65, assuming all eligible individuals participate in the intervention and considering intervention achievement rates. Our model simulates the diabetes trends in the U.S. population based on individual-level longitudinal data. However, it may be used to identify the optimal intervention for different subgroups in defined populations.

 

We study the classical ranking and selection problem, where the ultimate goal is to find the unknown best alternative in terms of the probability of correct selection or expected opportunity cost. However, this paper adopts an alternative sampling approach to achieve this goal, where sampling decisions are made with the objective of maximizing information about the unknown best alternative, or equivalently, minimizing its Shannon entropy. This adaptive learning is formulated via a Bayesian stochastic dynamic programming problem, by which several properties of the learning problem are presented, including the monotonicity of the optimal value function in an information‐seeking setting. Since the state space of the stochastic dynamic program is unbounded in the Gaussian setting, a one‐step look‐ahead approach is used to develop a policy. The proposed policy seeks to maximize the one‐step information gain about the unknown best alternative, and therefore, it is called information gradient (IG). It is also proved that the IG policy is consistent, that is, as the sampling budget grows to infinity, the IG policy finds the true best alternative almost surely. Later, a computationally efficient estimate of the proposed policy, called approximated information gradient (AIG), is introduced and in the numerical experiments its performance is tested against recent benchmarks alongside several sensitivity analyses. Results show that AIG performs competitively against other algorithms from the literature.