AI models are often evaluated based on their ability to predict the outcome of interest. However, in many AI for social impact applications, the presence of an intervention that affects the outcome can bias the evaluation. Randomized controlled trials (RCTs) randomly assign interventions, allowing data from the control group to be used for unbiased model evaluation. However, this approach is inefficient because it ignores data from the treatment group. Given the complexity and cost often associated with RCTs, making the most use of the data is essential. Thus, we investigate model evaluation strategies that leverage all data from an RCT. First, we theoretically quantify the estimation bias that arises from naïvely aggregating performance estimates from treatment and control groups and derive the condition under which this bias leads to incorrect model selection. Leveraging these theoretical insights, we propose nuisance parameter weighting (NPW), an unbiased model evaluation approach that reweights data from the treatment group to mimic the distribution of samples that would or would not experience the outcome under no intervention. Using synthetic and real-world datasets, we demonstrate that our proposed evaluation approach consistently yields better model selection than the standard approach, which ignores data from the treatment group, across various intervention effect and sample size settings. Our contribution represents a meaningful step towards more efficient model evaluation in real-world contexts.

Peer Reviewed

Many reinforcement learning (RL) applications have combinatorial action spaces, where each action is a composition of sub-actions. A standard RL approach ignores this inherent factorization structure, resulting in a potential failure to make meaningful inferences about rarely observed sub-action combinations; this is particularly problematic for offline settings, where data may be limited. In this work, we propose a form of linear Q-function decomposition induced by factored action spaces. We study the theoretical properties of our approach, identifying scenarios where it is guaranteed to lead to zero bias when used to approximate the Q-function. Outside the regimes with theoretical guarantees, we show that our approach can still be useful because it leads to better sample efficiency without necessarily sacrificing policy optimality, allowing us to achieve a better bias-variance trade-off. Across several offline RL problems using simulators and real-world datasets motivated by healthcare, we demonstrate that incorporating factored action spaces into value-based RL can result in better-performing policies. Our approach can help an agent make more accurate inferences within underexplored regions of the state-action space when applying RL to observational datasets.

During training, models can exploit spurious correlations as shortcuts, resulting in poor generalization performance when shortcuts do not persist. In this work, assuming access to a representation based on domain knowledge (i.e., known concepts) that is invariant to shortcuts, we aim to learn robust and accurate models from biased training data. In contrast to previous work, we do not rely solely on known concepts, but allow the model to also learn unknown concepts. We propose two approaches for mitigating shortcuts that incorporate domain knowledge, while accounting for potentially important yet unknown concepts. The first approach is two-staged. After fitting a model using known concepts, it accounts for the residual using unknown concepts. While flexible, we show that this approach is vulnerable when shortcuts are correlated with the unknown concepts. This limitation is addressed by our second approach that extends a recently proposed regularization penalty. Applied to two real-world datasets, we demonstrate that both approaches can successfully mitigate shortcut learning.

OBJECTIVE: When patients develop acute respiratory failure (ARF), accurately identifying the underlying etiology is essential for determining the best treatment. However, differentiating between common medical diagnoses can be challenging in clinical practice. Machine learning models could improve medical diagnosis by aiding in the diagnostic evaluation of these patients. MATERIALS AND METHODS: Machine learning models were trained to predict the common causes of ARF (pneumonia, heart failure, and/or chronic obstructive pulmonary disease [COPD]). Models were trained using chest radiographs and clinical data from the electronic health record (EHR) and applied to an internal and external cohort. RESULTS: The internal cohort of 1618 patients included 508 (31%) with pneumonia, 363 (22%) with heart failure, and 137 (8%) with COPD based on physician chart review. A model combining chest radiographs and EHR data outperformed models based on each modality alone. Models had similar or better performance compared to a randomly selected physician reviewer. For pneumonia, the combined model area under the receiver operating characteristic curve (AUROC) was 0.79 (0.77-0.79), image model AUROC was 0.74 (0.72-0.75), and EHR model AUROC was 0.74 (0.70-0.76). For heart failure, combined: 0.83 (0.77-0.84), image: 0.80 (0.71-0.81), and EHR: 0.79 (0.75-0.82). For COPD, combined: AUROC = 0.88 (0.83-0.91), image: 0.83 (0.77-0.89), and EHR: 0.80 (0.76-0.84). In the external cohort, performance was consistent for heart failure and increased for COPD, but declined slightly for pneumonia. CONCLUSIONS: Machine learning models combining chest radiographs and EHR data can accurately differentiate between common causes of ARF. Further work is needed to determine how these models could act as a diagnostic aid to clinicians in clinical settings.

Acute respiratory distress syndrome (ARDS) is a fulminant inflammatory lung injury that develops in patients with critical illnesses, affecting 200,000 patients in the United States annually. However, a recent study suggests that most patients with ARDS are diagnosed late or missed completely and fail to receive life-saving treatments. This is primarily due to the dependency of current diagnosis criteria on chest x-ray, which is not necessarily available at the time of diagnosis. In machine learning, such an information is known as Privileged Information - information that is available at training but not at testing. However, in diagnosing ARDS, privileged information (chest x-rays) are sometimes only available for a portion of the training data. To address this issue, the Learning Using Partially Available Privileged Information (LUPAPI) paradigm is proposed. As there are multiple ways to incorporate partially available privileged information, three models built on classical SVM are described. Another complexity of diagnosing ARDS is the uncertainty in clinical interpretation of chest x-rays. To address this, the LUPAPI framework is then extended to incorporate label uncertainty, resulting in a novel and comprehensive machine learning paradigm - Learning Using Label Uncertainty and Partially Available Privileged Information (LULUPAPI). The proposed frameworks use Electronic Health Record (EHR) data as regular information, chest x-rays as partially available privileged information, and clinicians' confidence levels in ARDS diagnosis as a measure of label uncertainty. Experiments on an ARDS dataset demonstrate that both the LUPAPI and LULUPAPI models outperform SVM, with LULUPAPI performing better than LUPAPI.

Standard reinforcement learning (RL) aims to find an optimal policy that identifies the best action for each state. However, in healthcare settings, many actions may be near-equivalent with respect to the reward (e.g., survival). We consider an alternative objective – learning set-valued policies to capture near-equivalent actions that lead to similar cumulative rewards. We propose a model-free algorithm based on temporal difference learning and a near-greedy heuristic for action selection. We analyze the theoretical properties of the proposed algorithm, providing optimality guarantees and demonstrate our approach on simulated environments and a real clinical task. Empirically, the proposed algorithm exhibits good convergence properties and discovers meaningful near-equivalent actions. Our work provides theoretical, as well as practical, foundations for clinician/human-in-the-loop decision making, in which humans (e.g., clinicians, patients) can incorporate additional knowledge (e.g., side effects, patient preference) when selecting among near-equivalent actions.

OBJECTIVE: In applying machine learning (ML) to electronic health record (EHR) data, many decisions must be made before any ML is applied; such preprocessing requires substantial effort and can be labor-intensive. As the role of ML in health care grows, there is an increasing need for systematic and reproducible preprocessing techniques for EHR data. Thus, we developed FIDDLE (Flexible Data-Driven Pipeline), an open-source framework that streamlines the preprocessing of data extracted from the EHR. MATERIALS AND METHODS: Largely data-driven, FIDDLE systematically transforms structured EHR data into feature vectors, limiting the number of decisions a user must make while incorporating good practices from the literature. To demonstrate its utility and flexibility, we conducted a proof-of-concept experiment in which we applied FIDDLE to 2 publicly available EHR data sets collected from intensive care units: MIMIC-III and the eICU Collaborative Research Database. We trained different ML models to predict 3 clinically important outcomes: in-hospital mortality, acute respiratory failure, and shock. We evaluated models using the area under the receiver operating characteristics curve (AUROC), and compared it to several baselines. RESULTS: Across tasks, FIDDLE extracted 2,528 to 7,403 features from MIMIC-III and eICU, respectively. On all tasks, FIDDLE-based models achieved good discriminative performance, with AUROCs of 0.757-0.886, comparable to the performance of MIMIC-Extract, a preprocessing pipeline designed specifically for MIMIC-III. Furthermore, our results showed that FIDDLE is generalizable across different prediction times, ML algorithms, and data sets, while being relatively robust to different settings of user-defined arguments. CONCLUSIONS: FIDDLE, an open-source preprocessing pipeline, facilitates applying ML to structured EHR data. By accelerating and standardizing labor-intensive preprocessing, FIDDLE can help stimulate progress in building clinically useful ML tools for EHR data.

When training a machine learning algorithm for a supervised-learning task in some clinical applications, uncertainty in the correct labels of some patients may adversely affect the performance of the algorithm. For example, even clinical experts may have less confidence when assigning a medical diagnosis to some patients because of ambiguity in the patient's case or imperfect reliability of the diagnostic criteria. As a result, some cases used in algorithm training may be mislabeled, adversely affecting the algorithm's performance. However, experts may also be able to quantify their diagnostic uncertainty in these cases. We present a robust method implemented with support vector machines (SVM) to account for such clinical diagnostic uncertainty when training an algorithm to detect patients who develop the acute respiratory distress syndrome (ARDS). ARDS is a syndrome of the critically ill that is diagnosed using clinical criteria known to be imperfect. We represent uncertainty in the diagnosis of ARDS as a graded weight of confidence associated with each training label. We also performed a novel time-series sampling method to address the problem of intercorrelation among the longitudinal clinical data from each patient used in model training to limit overfitting. Preliminary results show that we can achieve meaningful improvement in the performance of algorithm to detect patients with ARDS on a hold-out sample, when we compare our method that accounts for the uncertainty of training labels with a conventional SVM algorithm.