Susan Murphy

Raaz Dwivedi, Katherine Tian, Sabina Tomkins et al. · harvard, mit

We introduce and analyze an improved variant of nearest neighbors (NN) for estimation with missing data in latent factor models. We consider a matrix completion problem with missing data, where the $(i, t)$-th entry, when observed, is given by its mean $f(u_i, v_t)$ plus mean-zero noise for an unknown function $f$ and latent factors $u_i$ and $v_t$. Prior NN strategies, like unit-unit NN, for estimating the mean $f(u_i, v_t)$ relies on existence of other rows $j$ with $u_j \approx u_i$. Similarly, time-time NN strategy relies on existence of columns $t'$ with $v_{t'} \approx v_t$. These strategies provide poor performance respectively when similar rows or similar columns are not available. Our estimate is doubly robust to this deficit in two ways: (1) As long as there exist either good row or good column neighbors, our estimate provides a consistent estimate. (2) Furthermore, if both good row and good column neighbors exist, it provides a (near-)quadratic improvement in the non-asymptotic error and admits a significantly narrower asymptotic confidence interval when compared to both unit-unit or time-time NN.

Susobhan Ghosh, Raphael Kim, Prasidh Chhabria et al. · harvard, mit

There is a growing interest in using reinforcement learning (RL) to personalize sequences of treatments in digital health to support users in adopting healthier behaviors. Such sequential decision-making problems involve decisions about when to treat and how to treat based on the user's context (e.g., prior activity level, location, etc.). Online RL is a promising data-driven approach for this problem as it learns based on each user's historical responses and uses that knowledge to personalize these decisions. However, to decide whether the RL algorithm should be included in an ``optimized'' intervention for real-world deployment, we must assess the data evidence indicating that the RL algorithm is actually personalizing the treatments to its users. Due to the stochasticity in the RL algorithm, one may get a false impression that it is learning in certain states and using this learning to provide specific treatments. We use a working definition of personalization and introduce a resampling-based methodology for investigating whether the personalization exhibited by the RL algorithm is an artifact of the RL algorithm stochasticity. We illustrate our methodology with a case study by analyzing the data from a physical activity clinical trial called HeartSteps, which included the use of an online RL algorithm. We demonstrate how our approach enhances data-driven truth-in-advertising of algorithm personalization both across all users as well as within specific users in the study.

Eura Shin, Siddharth Swaroop, Weiwei Pan et al.

Mobile health (mHealth) technologies empower patients to adopt/maintain healthy behaviors in their daily lives, by providing interventions (e.g. push notifications) tailored to the user's needs. In these settings, without intervention, human decision making may be impaired (e.g. valuing near term pleasure over own long term goals). In this work, we formalize this relationship with a framework in which the user optimizes a (potentially impaired) Markov Decision Process (MDP) and the mHealth agent intervenes on the user's MDP parameters. We show that different types of impairments imply different types of optimal intervention. We also provide analytical and empirical explorations of these differences.

Susobhan Ghosh, Yongyi Guo, Pei-Yao Hung et al.

The escalating prevalence of cannabis use poses a significant public health challenge globally. In the U.S., cannabis use is more prevalent among emerging adults (EAs) (ages 18-25) than any other age group, with legalization in the multiple states contributing to a public perception that cannabis is less risky than in prior decades. To address this growing concern, we developed MiWaves, a reinforcement learning (RL) algorithm designed to optimize the delivery of personalized intervention prompts to reduce cannabis use among EAs. MiWaves leverages domain expertise and prior data to tailor the likelihood of delivery of intervention messages. This paper presents a comprehensive overview of the algorithm's design, including key decisions and experimental outcomes. The finalized MiWaves RL algorithm was deployed in a clinical trial from March to May 2024.

Shuangning Li, Lluis Salvat Niell, Sung Won Choi et al.

Mobile health aims to enhance health outcomes by delivering interventions to individuals as they go about their daily life. The involvement of care partners and social support networks often proves crucial in helping individuals managing burdensome medical conditions. This presents opportunities in mobile health to design interventions that target the dyadic relationship -- the relationship between a target person and their care partner -- with the aim of enhancing social support. In this paper, we develop dyadic RL, an online reinforcement learning algorithm designed to personalize intervention delivery based on contextual factors and past responses of a target person and their care partner. Here, multiple sets of interventions impact the dyad across multiple time intervals. The developed dyadic RL is Bayesian and hierarchical. We formally introduce the problem setup, develop dyadic RL and establish a regret bound. We demonstrate dyadic RL's empirical performance through simulation studies on both toy scenarios and on a realistic test bed constructed from data collected in a mobile health study.

Sorawit Saengkyongam, Niklas Pfister, Predrag Klasnja et al.

Policy learning is an important component of many real-world learning systems. A major challenge in policy learning is how to adapt efficiently to unseen environments or tasks. Recently, it has been suggested to exploit invariant conditional distributions to learn models that generalize better to unseen environments. However, assuming invariance of entire conditional distributions (which we call full invariance) may be too strong of an assumption in practice. In this paper, we introduce a relaxation of full invariance called effect-invariance (e-invariance for short) and prove that it is sufficient, under suitable assumptions, for zero-shot policy generalization. We also discuss an extension that exploits e-invariance when we have a small sample from the test environment, enabling few-shot policy generalization. Our work does not assume an underlying causal graph or that the data are generated by a structural causal model; instead, we develop testing procedures to test e-invariance directly from data. We present empirical results using simulated data and a mobile health intervention dataset to demonstrate the effectiveness of our approach.

Yongyi Guo, Ziping Xu, Susan Murphy

We consider the contextual bandit problem where at each time, the agent only has access to a noisy version of the context and the error variance (or an estimator of this variance). This setting is motivated by a wide range of applications where the true context for decision-making is unobserved, and only a prediction of the context by a potentially complex machine learning algorithm is available. When the context error is non-vanishing, classical bandit algorithms fail to achieve sublinear regret. We propose the first online algorithm in this setting with sublinear regret guarantees under mild conditions. The key idea is to extend the measurement error model in classical statistics to the online decision-making setting, which is nontrivial due to the policy being dependent on the noisy context observations. We further demonstrate the benefits of the proposed approach in simulation environments based on synthetic and real digital intervention datasets.

Susobhan Ghosh, Yongyi Guo, Pei-Yao Hung et al.

The escalating prevalence of cannabis use, and associated cannabis-use disorder (CUD), poses a significant public health challenge globally. With a notably wide treatment gap, especially among emerging adults (EAs; ages 18-25), addressing cannabis use and CUD remains a pivotal objective within the 2030 United Nations Agenda for Sustainable Development Goals (SDG). In this work, we develop an online reinforcement learning (RL) algorithm called reBandit which will be utilized in a mobile health study to deliver personalized mobile health interventions aimed at reducing cannabis use among EAs. reBandit utilizes random effects and informative Bayesian priors to learn quickly and efficiently in noisy mobile health environments. Moreover, reBandit employs Empirical Bayes and optimization techniques to autonomously update its hyper-parameters online. To evaluate the performance of our algorithm, we construct a simulation testbed using data from a prior study, and compare against commonly used algorithms in mobile health studies. We show that reBandit performs equally well or better than all the baseline algorithms, and the performance gap widens as population heterogeneity increases in the simulation environment, proving its adeptness to adapt to diverse population of study participants.

Xueqing Liu, Kyra Gan, Esmaeil Keyvanshokooh et al.

Motivated by applications in digital health, this work studies the novel problem of online uniform sampling (OUS), where the goal is to distribute a sampling budget uniformly across unknown decision times. In the OUS problem, the algorithm is given a budget $b$ and a time horizon $T$, and an adversary then chooses a value $τ^* \in [b,T]$, which is revealed to the algorithm online. At each decision time $i \in [τ^*]$, the algorithm must determine a sampling probability that maximizes the budget spent throughout the horizon, respecting budget constraint $b$, while achieving as uniform a distribution as possible over $τ^*$. We present the first randomized algorithm designed for this problem and subsequently extend it to incorporate learning augmentation. We provide worst-case approximation guarantees for both algorithms, and illustrate the utility of the algorithms through both synthetic experiments and a real-world case study involving the HeartSteps mobile application. Our numerical results show strong empirical average performance of our proposed randomized algorithms against previously proposed heuristic solutions.

Ziping Xu, Hinal Jajal, Sung Won Choi et al.

Medication adherence is critical for the recovery of adolescents and young adults (AYAs) who have undergone hematopoietic cell transplantation (HCT). However, maintaining adherence is challenging for AYAs after hospital discharge, who experience both individual (e.g. physical and emotional symptoms) and interpersonal barriers (e.g., relational difficulties with their care partner, who is often involved in medication management). To optimize the effectiveness of a three-component digital intervention targeting both members of the dyad as well as their relationship, we propose a novel Multi-Agent Reinforcement Learning (MARL) approach to personalize the delivery of interventions. By incorporating the domain knowledge, the MARL framework, where each agent is responsible for the delivery of one intervention component, allows for faster learning compared with a flattened agent. Evaluation using a dyadic simulator environment, based on real clinical data, shows a significant improvement in medication adherence (approximately 3%) compared to purely random intervention delivery. The effectiveness of this approach will be further evaluated in an upcoming trial.

Kyla Chasalow, Skyler Wu, Susan Murphy

Missing data in online reinforcement learning (RL) poses challenges compared to missing data in standard tabular data or in offline policy learning. The need to impute and act at each time step means that imputation cannot be put off until enough data exist to produce stable imputation models. It also means future data collection and learning depend on previous imputations. This paper proposes fully online imputation ensembles. We find that maintaining multiple imputation pathways may help balance the need to capture uncertainty under missingness and the need for efficiency in online settings. We consider multiple approaches for incorporating these pathways into learning and action selection. Using a Grid World experiment with various types of missingness, we provide preliminary evidence that multiple imputation pathways may be a useful framework for constructing simple and efficient online missing data RL methods.

Eura Nofshin, Siddharth Swaroop, Weiwei Pan et al.

Many important behavior changes are frictionful; they require individuals to expend effort over a long period with little immediate gratification. Here, an artificial intelligence (AI) agent can provide personalized interventions to help individuals stick to their goals. In these settings, the AI agent must personalize rapidly (before the individual disengages) and interpretably, to help us understand the behavioral interventions. In this paper, we introduce Behavior Model Reinforcement Learning (BMRL), a framework in which an AI agent intervenes on the parameters of a Markov Decision Process (MDP) belonging to a boundedly rational human agent. Our formulation of the human decision-maker as a planning agent allows us to attribute undesirable human policies (ones that do not lead to the goal) to their maladapted MDP parameters, such as an extremely low discount factor. Furthermore, we propose a class of tractable human models that captures fundamental behaviors in frictionful tasks. Introducing a notion of MDP equivalence specific to BMRL, we theoretically and empirically show that AI planning with our human models can lead to helpful policies on a wide range of more complex, ground-truth humans.

Kyra Gan, Esmaeil Keyvanshokooh, Xueqing Liu et al.

Contextual bandit algorithms are commonly used in digital health to recommend personalized treatments. However, to ensure the effectiveness of the treatments, patients are often requested to take actions that have no immediate benefit to them, which we refer to as pro-treatment actions. In practice, clinicians have a limited budget to encourage patients to take these actions and collect additional information. We introduce a novel optimization and learning algorithm to address this problem. This algorithm effectively combines the strengths of two algorithmic approaches in a seamless manner, including 1) an online primal-dual algorithm for deciding the optimal timing to reach out to patients, and 2) a contextual bandit learning algorithm to deliver personalized treatment to the patient. We prove that this algorithm admits a sub-linear regret bound. We illustrate the usefulness of this algorithm on both synthetic and real-world data.

Raaz Dwivedi, Katherine Tian, Sabina Tomkins et al.

We consider after-study statistical inference for sequentially designed experiments wherein multiple units are assigned treatments for multiple time points using treatment policies that adapt over time. Our goal is to provide inference guarantees for the counterfactual mean at the smallest possible scale -- mean outcome under different treatments for each unit and each time -- with minimal assumptions on the adaptive treatment policy. Without any structural assumptions on the counterfactual means, this challenging task is infeasible due to more unknowns than observed data points. To make progress, we introduce a latent factor model over the counterfactual means that serves as a non-parametric generalization of the non-linear mixed effects model and the bilinear latent factor model considered in prior works. For estimation, we use a non-parametric method, namely a variant of nearest neighbors, and establish a non-asymptotic high probability error bound for the counterfactual mean for each unit and each time. Under regularity conditions, this bound leads to asymptotically valid confidence intervals for the counterfactual mean as the number of units and time points grows to $\infty$ together at suitable rates. We illustrate our theory via several simulations and a case study involving data from a mobile health clinical trial HeartSteps.

Eura Shin, Pedja Klasnja, Susan Murphy et al.

Motivated by the need for efficient and personalized learning in mobile health, we investigate the problem of online kernel selection for Gaussian Process regression in the multi-task setting. We propose a novel generative process on the kernel composition for this purpose. Our method demonstrates that trajectories of kernel evolutions can be transferred between users to improve learning and that the kernels themselves are meaningful for an mHealth prediction goal.

Marianne Menictas, Sabina Tomkins, Susan Murphy

Users can be supported to adopt healthy behaviors, such as regular physical activity, via relevant and timely suggestions on their mobile devices. Recently, reinforcement learning algorithms have been found to be effective for learning the optimal context under which to provide suggestions. However, these algorithms are not necessarily designed for the constraints posed by mobile health (mHealth) settings, that they be efficient, domain-informed and computationally affordable. We propose an algorithm for providing physical activity suggestions in mHealth settings. Using domain-science, we formulate a contextual bandit algorithm which makes use of a linear mixed effects model. We then introduce a procedure to efficiently perform hyper-parameter updating, using far less computational resources than competing approaches. Not only is our approach computationally efficient, it is also easily implemented with closed form matrix algebraic updates and we show improvements over state of the art approaches both in speed and accuracy of up to 99% and 56% respectively.

Sabina Tomkins, Peng Liao, Predrag Klasnja et al.

In mobile health (mHealth) smart devices deliver behavioral treatments repeatedly over time to a user with the goal of helping the user adopt and maintain healthy behaviors. Reinforcement learning appears ideal for learning how to optimally make these sequential treatment decisions. However, significant challenges must be overcome before reinforcement learning can be effectively deployed in a mobile healthcare setting. In this work we are concerned with the following challenges: 1) individuals who are in the same context can exhibit differential response to treatments 2) only a limited amount of data is available for learning on any one individual, and 3) non-stationary responses to treatment. To address these challenges we generalize Thompson-Sampling bandit algorithms to develop IntelligentPooling. IntelligentPooling learns personalized treatment policies thus addressing challenge one. To address the second challenge, IntelligentPooling updates each user's degree of personalization while making use of available data on other users to speed up learning. Lastly, IntelligentPooling allows responsivity to vary as a function of a user's time since beginning treatment, thus addressing challenge three. We show that IntelligentPooling achieves an average of 26% lower regret than state-of-the-art. We demonstrate the promise of this approach and its ability to learn from even a small group of users in a live clinical trial.

Peng Liao, Zhengling Qi, Runzhe Wan et al.

We consider the batch (off-line) policy learning problem in the infinite horizon Markov Decision Process. Motivated by mobile health applications, we focus on learning a policy that maximizes the long-term average reward. We propose a doubly robust estimator for the average reward and show that it achieves semiparametric efficiency. Further we develop an optimization algorithm to compute the optimal policy in a parameterized stochastic policy class. The performance of the estimated policy is measured by the difference between the optimal average reward in the policy class and the average reward of the estimated policy and we establish a finite-sample regret guarantee. The performance of the method is illustrated by simulation studies and an analysis of a mobile health study promoting physical activity.

Jiayu Yao, Emma Brunskill, Weiwei Pan et al.

Contextual bandits often provide simple and effective personalization in decision making problems, making them popular tools to deliver personalized interventions in mobile health as well as other health applications. However, when bandits are deployed in the context of a scientific study -- e.g. a clinical trial to test if a mobile health intervention is effective -- the aim is not only to personalize for an individual, but also to determine, with sufficient statistical power, whether or not the system's intervention is effective. It is essential to assess the effectiveness of the intervention before broader deployment for better resource allocation. The two objectives are often deployed under different model assumptions, making it hard to determine how achieving the personalization and statistical power affect each other. In this work, we develop general meta-algorithms to modify existing algorithms such that sufficient power is guaranteed while still improving each user's well-being. We also demonstrate that our meta-algorithms are robust to various model mis-specifications possibly appearing in statistical studies, thus providing a valuable tool to study designers.

Mashfiqui Rabbi, Meredith Philyaw-Kotov, Jinseok Li et al.

Mobile health (mHealth) applications are a powerful medium for providing behavioral interventions, and systematic reviews suggest that theory-based interventions are more effective. However, how exactly theoretical concepts should be translated into features of technological interventions is often not clear. There is a gulf between the abstract nature of psychological theory and the concreteness of the designs needed to build health technologies. In this paper, we use SARA, a mobile app we developed to support substance-use research among adolescents and young adults, as a case study of a process of translating behavioral theory into mHealth intervention design. SARA was designed to increase adherence to daily self-report in longitudinal epidemiological studies. To achieve this goal, we implemented a number of constructs from the operant conditioning theory. We describe our design process and discuss how we operationalized theoretical constructs in the light of design constraints, user feedback, and empirical data from four formative studies.

Sabina Tomkins, Peng Liao, Predrag Klasnja et al.

In mobile health (mHealth), reinforcement learning algorithms that adapt to one's context without learning personalized policies might fail to distinguish between the needs of individuals. Yet the high amount of noise due to the in situ delivery of mHealth interventions can cripple the ability of an algorithm to learn when given access to only a single user's data, making personalization challenging. We present IntelligentPooling, which learns personalized policies via an adaptive, principled use of other users' data. We show that IntelligentPooling achieves an average of 26% lower regret than state-of-the-art across all generative models. Additionally, we inspect the behavior of this approach in a live clinical trial, demonstrating its ability to learn from even a small group of users.

Peng Liao, Predrag Klasnja, Susan Murphy

Due to the recent advancements in wearables and sensing technology, health scientists are increasingly developing mobile health (mHealth) interventions. In mHealth interventions, mobile devices are used to deliver treatment to individuals as they go about their daily lives. These treatments are generally designed to impact a near time, proximal outcome such as stress or physical activity. The mHealth intervention policies, often called just-in-time adaptive interventions, are decision rules that map an individual's current state (e.g., individual's past behaviors as well as current observations of time, location, social activity, stress and urges to smoke) to a particular treatment at each of many time points. The vast majority of current mHealth interventions deploy expert-derived policies. In this paper, we provide an approach for conducting inference about the performance of one or more such policies using historical data collected under a possibly different policy. Our measure of performance is the average of proximal outcomes over a long time period should the particular mHealth policy be followed. We provide an estimator as well as confidence intervals. This work is motivated by HeartSteps, an mHealth physical activity intervention.

Peng Liao, Kristjan Greenewald, Predrag Klasnja et al.

With the recent evolution of mobile health technologies, health scientists are increasingly interested in developing just-in-time adaptive interventions (JITAIs), typically delivered via notification on mobile device and designed to help the user prevent negative health outcomes and promote the adoption and maintenance of healthy behaviors. A JITAI involves a sequence of decision rules (i.e., treatment policy) that takes the user's current context as input and specifies whether and what type of an intervention should be provided at the moment. In this paper, we develop a Reinforcement Learning (RL) algorithm that continuously learns and improves the treatment policy embedded in the JITAI as the data is being collected from the user. This work is motivated by our collaboration on designing the RL algorithm in HeartSteps V2 based on data from HeartSteps V1. HeartSteps is a physical activity mobile health application. The RL algorithm developed in this paper is being used in HeartSteps V2 to decide, five times per day, whether to deliver a context-tailored activity suggestion.

Sabina Tomkins, Predrag Klasnja, Susan Murphy

In many mobile health interventions, treatments should only be delivered in a particular context, for example when a user is currently stressed, walking or sedentary. Even in an optimal context, concerns about user burden can restrict which treatments are sent. To diffuse the treatment delivery over times when a user is in a desired context, it is critical to predict the future number of times the context will occur. The focus of this paper is on whether personalization can improve predictions in these settings. Though the variance between individuals' behavioral patterns suggest that personalization should be useful, the amount of individual-level data limits its capabilities. Thus, we investigate several methods which pool data across users to overcome these deficiencies and find that pooling lowers the overall error rate relative to both personalized and batch approaches.