Although deep learning models trained on electronic health records (EHR) data have shown state-of-the-art performance in many predictive clinical tasks, the discovery of adversarial examples (i.e., input data that are engineered to cause misclassification) has exposed vulnerabilities with lab and imaging data. We specifically consider adversarial examples with longitudinal EHR data, an area that has not been previously examined because of the challenges with temporal high-dimensional and sparse features. We propose Longitudinal AdVersarial Attack (, a saliency score based adversarial example using a method that requires a minimal number of perturbations and that automatically minimizes the likelihood of detection. Features are selected and modified by jointly modeling a saliency map and attention mechanism. Experimental results with longitudinal EHR data show that an substantially reduce model performance for attention-based target models (from AUPR = 0.5 to AUPR = 0.08).

This paper presents a new method, which we call SUSTain, that extends real-valued matrix and tensor factorizations to data where values are integers. Such data are common when the values correspond to event counts or ordinal measures. The conventional approach is to treat integer data as real, and then apply real-valued factorizations. However, doing so fails to preserve important characteristics of the original data, thereby making it hard to interpret the results. Instead, our approach extracts factor values from integer datasets as scores that are constrained to take values from a small integer set. These scores are easy to interpret: a score of zero indicates no feature contribution and higher scores indicate distinct levels of feature importance. At its core, SUSTain relies on: a) a problem partitioning into integer-constrained subproblems, so that they can be optimally solved in an efficient manner; and b) organizing the order of the subproblems' solution, to promote reuse of shared intermediate results. We propose two variants, SUSTain_M and SUSTain_T, to handle both matrix and tensor inputs, respectively. We evaluate SUSTain against several state-of-the-art baselines on both synthetic and real Electronic Health Record (EHR) datasets. Comparing to those baselines, SUSTain shows either significantly better fit or orders of magnitude speedups that achieve a comparable fit (up to 425× faster). We apply SUSTain to EHR datasets to extract patient phenotypes (i.e., clinically meaningful patient clusters). Furthermore, 87% of them were validated as clinically meaningful phenotypes related to heart failure by a cardiologist.

Deep learning models exhibit state-of-the-art performance for many predictive healthcare tasks using electronic health records (EHR) data, but these models typically require training data volume that exceeds the capacity of most healthcare systems. External resources such as medical ontologies are used to bridge the data volume constraint, but this approach is often not directly applicable or useful because of inconsistencies with terminology. To solve the data insufficiency challenge, we leverage the inherent multilevel structure of EHR data and, in particular, the encoded relationships among medical codes. We propose Multilevel Medical Embedding (MiME) which learns the multilevel embedding of EHR data while jointly performing auxiliary prediction tasks that rely on this inherent EHR structure without the need for external labels. We conducted two prediction tasks, heart failure prediction and sequential disease prediction, where MiME outperformed baseline methods in diverse evaluation settings. In particular, MiME consistently outperformed all baselines when predicting heart failure on datasets of different volumes, especially demonstrating the greatest performance improvement (15% relative gain in PR-AUC over the best baseline) on the smallest dataset, demonstrating its ability to effectively model the multilevel structure of EHR data.

Clustering, the process of grouping together similar items into distinct partitions, is a common type of unsupervised machine learning that can be useful for summarizing and aggregating complex multi-dimensional data. However, data can be clustered in many ways, and there exist a large body of algorithms designed to reveal different patterns. While having access to a wide variety of algorithms is helpful, in practice, it is quite difficult for data scientists to choose and parameterize algorithms to get the clustering results relevant for their dataset and analytical tasks. To alleviate this problem, we built Clustervision, a visual analytics tool that helps ensure data scientists find the right clustering among the large amount of techniques and parameters available. Our system clusters data using a variety of clustering techniques and parameters and then ranks clustering results utilizing five quality metrics. In addition, users can guide the system to produce more relevant results by providing task-relevant constraints on the data. Our visual user interface allows users to find high quality clustering results, explore the clusters using several coordinated visualization techniques, and select the cluster result that best suits their task. We demonstrate this novel approach using a case study with a team of researchers in the medical domain and showcase that our system empowers users to choose an effective representation of their complex data.

accuracy, data needs, interpretability) of GRAM to various methods including the recurrent neural network (RNN) in two sequential diagnoses prediction tasks and one heart failure prediction task. Compared to the basic RNN, GRAM achieved 10% higher accuracy for predicting diseases rarely observed in the training data and 3% improved area under the ROC curve for predicting heart failure using an order of magnitude less training data. Additionally, unlike other methods, the medical concept representations learned by GRAM are well aligned with the medical ontology. Finally, GRAM exhibits intuitive attention behaviors by adaptively generalizing to higher level concepts when facing data insufficiency at the lower level concepts.

Managing patients with complex multimorbidity has long been recognized as a difficult problem due to complex disease and medication dependencies and the potential risk of adverse drug interactions. Existing work either uses complicated rule-based protocols which are hard to implement and maintain, or simple statistical models that treat each disease independently, which may lead to sub-optimal or even harmful drug combinations. In this work, we propose the LEAP (LEArn to Prescribe) algorithm to decompose the treatment recommendation into a sequential decision-making process while automatically determining the appropriate number of medications. A recurrent decoder is used to model label dependencies and content-based attention is used to capture label instance mapping. We further leverage reinforcement learning to fine tune the model parameters to ensure accuracy and completeness. We incorporate external clinical knowledge into the design of the reinforcement reward to effectively prevent generating unfavorable drug combinations. Both quantitative experiments and qualitative case studies are conducted on two real world electronic health record datasets to verify the effectiveness of our solution. On both datasets, LEAP significantly outperforms baselines by up to 10-30% in terms of mean Jaccard coefficient and removes 99.8% adverse drug interactions in the recommended treatment sets.

Objective: We explored whether use of deep learning to model temporal relations among events in electronic health records (EHRs) would improve model performance in predicting initial diagnosis of heart failure (HF) compared to conventional methods that ignore temporality. Materials and Methods: Data were from a health system's EHR on 3884 incident HF cases and 28 903 controls, identified as primary care patients, between May 16, 2000, and May 23, 2013. Recurrent neural network (RNN) models using gated recurrent units (GRUs) were adapted to detect relations among time-stamped events (eg, disease diagnosis, medication orders, procedure orders, etc.) with a 12- to 18-month observation window of cases and controls. Model performance metrics were compared to regularized logistic regression, neural network, support vector machine, and K-nearest neighbor classifier approaches. Results: Using a 12-month observation window, the area under the curve (AUC) for the RNN model was 0.777, compared to AUCs for logistic regression (0.747), multilayer perceptron (MLP) with 1 hidden layer (0.765), support vector machine (SVM) (0.743), and K-nearest neighbor (KNN) (0.730). When using an 18-month observation window, the AUC for the RNN model increased to 0.883 and was significantly higher than the 0.834 AUC for the best of the baseline methods (MLP). Conclusion: Deep learning models adapted to leverage temporal relations appear to improve performance of models for detection of incident heart failure with a short observation window of 12-18 months.

OBJECTIVE: We describe how electronic health record (EHR) audit files can be used to understand how time is spent in primary care (PC). MATERIALS/METHODS: We used audit file data from the Geisinger Clinic to quantify elements of the clinical workflow and to determine how these times vary by patient and encounter factors. We randomly selected audit file records representing 36 437 PC encounters across 26 clinic locations. Audit file data were used to estimate duration and variance of: (1) time in the waiting room, (2) nurse time with the patient, (3) time in the exam room without a nurse or physician, and (4) physician time with the patient. Multivariate modeling was used to test for differences by patient and by encounter features. RESULTS: On average, a PC encounter took 54.6 minutes, with 5 minutes of nurse time, 15.5 minutes of physician time, and the remaining 62% of the time spent waiting to see a clinician or check out. Older age, female sex, and chronic disease were associated with longer wait times and longer time with clinicians. Level of service and numbers of medications, procedures, and lab orders were associated with longer time with clinicians. Late check-in and same-day visits were associated with shorter wait time and clinician time. CONCLUSIONS: This study provides insights on uses of audit file data for workflow analysis during PC encounters. DISCUSSION: Scalable ways to quantify clinical encounter workflow elements may provide the means to develop more efficient approaches to care and improve the patient experience.

Accuracy and interpretability are two dominant features of successful predictive models. Typically, a choice must be made in favor of complex black box models such as recurrent neural networks (RNN) for accuracy versus less accurate but more interpretable traditional models such as logistic regression. This tradeoff poses challenges in medicine where both accuracy and interpretability are important. We addressed this challenge by developing the REverse Time AttentIoN model (RETAIN) for application to Electronic Health Records (EHR) data. RETAIN achieves high accuracy while remaining clinically interpretable and is based on a two-level neural attention model that detects influential past visits and significant clinical variables within those visits (e.g. key diagnoses). RETAIN mimics physician practice by attending the EHR data in a reverse time order so that recent clinical visits are likely to receive higher attention. RETAIN was tested on a large health system EHR dataset with 14 million visits completed by 263K patients over an 8 year period and demonstrated predictive accuracy and computational scalability comparable to state-of-the-art methods such as RNN, and ease of interpretability comparable to traditional models.

Diverse stakeholders--clinicians, researchers, business leaders, policy makers, and the public--have good reason to believe that the effective use of electronic health care records (EHRs) is essential to meaningful advances in health care quality and patient safety. However, several reports have documented the potential of EHRs to contribute to health care system flaws and patient harm. As organizations (including small hospitals and physician practices) with limited resources for care-process transformation, human-factors engineering, software safety, and project management begin to use EHRs, the chance of EHR-associated harm may increase. The authors propose a coordinated set of steps to advance the practice and theory of safe EHR design, implementation, and continuous improvement. These include setting EHR implementation in the context of health care process improvement, building safety into the specification and design of EHRs, safety testing and reporting, and rapid communication of EHR-related safety flaws and incidents.