Clinical predictive models (CPMs) are crucial for forecasting patient outcomes using available electronic health record (EHR) data. Traditional time-to-event (TTE) models, like the Cox proportional hazards model, assume that hazard ratios remain constant over time, which may not hold in many clinical settings. In this study, we introduce a Discrete Time Neural Network (DTNN) to address these limitations by modeling time-varying predictor importance. The DTNN combines the flexibility of classification models with the advantage of handling time-to-event data, providing a single model fit across multiple time horizons. Using data from patients with end-stage kidney disease (ESKD) undergoing hemodialysis, we demonstrate that the DTNN can flexibly adjust for different risk factors across short-term and long-term mortality predictions. The model was evaluated using cumulative-dynamic area under the receiver operating characteristic (CD-AUROC) to compare patients who remained event-free up to a given time t with those who experienced the event before t. Our results show that the DTNN outperforms traditional TTE models and provides robust predictions across varying time intervals, making it an appealing choice for clinical settings where time-varying predictor importance is essential.

Transformer-based models have achieved stateof-the-art performance in document classification but struggle with long-text processing due to the quadratic computational complexity in the self-attention module.Existing solutions, such as sparse attention, hierarchical models, and key sentence extraction, partially address the issue but still fall short when the input sequence is exceptionally lengthy.To address this challenge, we propose IRIS (Interpretable Retrieval-Augmented Classification for long Interspersed Document Sequences), a novel, lightweight framework that utilizes retrieval to efficiently classify long documents while enhancing interpretability.IRIS segments documents into chunks, stores their embeddings in a vector database, and retrieves those most relevant to a given task using learnable query vectors.A linear attention mechanism then aggregates the retrieved embeddings for classification, allowing the model to process arbitrarily long documents without increasing computational cost and remaining trainable on a single GPU.Our experiments across six datasets show that IRIS achieves comparable performance to baseline models on standard benchmarks, and excels in three clinical note disease risk prediction tasks where documents are extremely long and key information is sparse.Furthermore, IRIS provides global interpretability by revealing a clear summary of key risk factors identified by the model.These findings highlight the potential of IRIS as an efficient and interpretable solution for long-document classification, particularly in healthcare applications where both performance and explainability are crucial.1

Recently developed survival analysis methods improve upon existing approaches by predicting the probability of event occurrence in each of a number pre-specified (discrete) time intervals. By avoiding placing strong parametric assumptions on the event density, this approach tends to improve prediction performance, particularly when data are plentiful. However, in clinical settings with limited available data, it is often preferable to judiciously partition the event time space into a limited number of intervals well suited to the prediction task at hand. In this work, we develop Adaptive Discretization for Event PredicTion (ADEPT) to learn from data a set of cut points defining such a partition. We show that in two simulated datasets, we are able to recover intervals that match the underlying generative model. We then demonstrate improved prediction performance on three real-world observational datasets, including a large, newly harmonized stroke risk prediction dataset. Finally, we argue that our approach facilitates clinical decision-making by suggesting time intervals that are most appropriate for each task, in the sense that they facilitate more accurate risk prediction.

Deep neural networks (DNNs) lack the precise semantics and definitive probabilistic interpretation of probabilistic graphical models (PGMs). In this paper, we propose an innovative solution by constructing infinite tree-structured PGMs that correspond exactly to neural networks. Our research reveals that DNNs, during forward propagation, indeed perform approximations of PGM inference that are precise in this alternative PGM structure. Not only does our research complement existing studies that describe neural networks as kernel machines or infinite-sized Gaussian processes, it also elucidates a more direct approximation that DNNs make to exact inference in PGMs. Potential benefits include improved pedagogy and interpretation of DNNs, and algorithms that can merge the strengths of PGMs and DNNs.

Learning to predict rare medical events is difficult due to the inherent lack of signal in highly imbalanced datasets. Yet, oftentimes we also have access to surrogate or related outcomes that we believe share etiology or underlying risk factors with the event of interest. In this work, we propose the use of two variants of a well-known approach, regularized multi-label learning (MLL), that we hypothesize are uniquely suited to leverage this similarity and improve model performance in rare event settings. Whereas most analyses of MLL emphasize improved performance across all event types, our analyses quantify benefits to rare event prediction offered by our approach when a more common, related event is available to enhance learning. We begin by deriving asymptotic properties and providing theoretical insight into the convergence rates of our proposed estimators. We then provide simulation results highlighting how characteristics of the data generating process, including the event similarity and event rate, affect our proposed models’ performance. We conclude by showing real-world benefit of our approach in two clinical settings: prediction of rare cardiovascular morbidities in the setting of preeclampsia; and early prediction of autism from the electronic health record.

The mixture cure model allows failure probability to be estimated separately from failure timing in settings wherein failure never occurs in a subset of the population. In this paper, we draw on insights from representation learning and causal inference to develop a neural network based mixture cure model that is free of distributional assumptions, yielding improved prediction of failure timing, yet still effectively disentangles information about failure timing from information about failure probability. Our approach also mitigates effects of selection biases in the observation of failure and censoring times on estimation of the failure density and censoring density, respectively. Results suggest this approach could be applied to distinguish factors predicting failure occurrence versus timing and mitigate biases in real-world observational datasets.

Vivek Subramanian, Matthew Engelhard, Sam Berchuck, Liqun Chen, Ricardo Henao, Lawrence Carin. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

Gait impairment in multiple sclerosis (MS) can result from muscle weakness, physical fatigue, lack of coordination, and other symptoms. Walking speed, as measured by a number of clinician-administered walking tests, is the primary measure of gait impairment used by clinical researchers, but inertial gait features from body-worn sensors have been proven to add clinical value. This paper seeks to understand and differentiate the physiological significance of four such features with proven value in MS to facilitate adoption by clinical researchers and incorporation in gait monitoring and analysis systems. In addition, this information can be used to select features that might be appropriate in other forms of disability. Two of the four features are computed using the dynamic time warping (DTW) algorithm: The "DTW Score" is based on the usual DTW distance, and the "Warp Score" is based on the warping length. The third feature, based on kernel density estimation (KDE), is the "KDE Peak" value. Finally, the "Causality Index" is based on the phase slope index between inertial signals from different body parts. Relationships between these measures and the aforementioned gait-related symptoms are determined by applying factor analysis to three common, clinical walking outcomes, then correlating the inertial measures as well as walking speed to each extracted factor. Statistically significant differences in correlation coefficients to the three extracted clinical factors support their distinct physiological meaning and suggest they may have complimentary roles in the analysis of MS-related walking disability.

Multiple sclerosis (MS) interrupts communication between the brain and other parts of the body causing functional deterioration. Gait impairment is a common finding in MS, one caused by several neurological symptoms. We perform an event-specific analysis to study the variable impact of MS on gait components. Our results show that the mid-swing to heel strike (HS) phase of a gait cycle is the most indicative of motor problems. We apply the Hilbert-Huang transform to inertial gait data, corresponding to this phase, to extract the spectral features and study their relationships with the patient-reported outcomes. A number of strong and statistically significant dependencies were found, many having to do with activities of daily living and MS walking scale, leading to the conclusion that the disturbance in mid-swing to HS is specific to deterioration in physical functions. Spearman correlations coefficients and adjusted R2obtained using stepwise linear regression models are reported. We conclude that event-specific gait features can be used to quantify the precise impact of MS symptoms on gait phases and identify markers of balance, stability, or fall risk, etc. We believe that this information supplements on-going MS research and could be used to develop personalized disease-modifying therapies and exercises.

Gait impairment in Multiple Sclerosis (MS) can result from imbalance, physical fatigue, weakness, and other symptoms. Walking speed is the primary measure of gait impairment used by clinical researchers, but inertial gait features from body-worn sensors have been proven to add clinical value. This paper seeks to understand the physiologic significance of two such features in MS. Both features are computed using the dynamic time warping (DTW) algorithm: the “DTW Score” is based on the usual DTW distance, and the “Warp Score” is based on the warping length. Using linear regression and stepwise regression models, the relationship between these features and several gait-related MS symptoms is analyzed. Results show that the DTW Score and Warp Score have distinct physiologic significance in MS compared to walking speed, and these features may also be useful for walking assessment in a wide range of clinical contexts.