DeepCare: A Deep Dynamic Memory Model for Predictive Medicine

By Trang Pham, Truyen Tran, Dinh Q. Phung, Svetha Venkatesh. 2016

304 citations recorded by the source corpus.

Personalized predictive medicine necessitates modeling of patient illness and care processes, which inherently have long-term temporal dependencies. Healthcare observations, recorded in electronic medical records, are episodic and irregular in time. We introduce DeepCare, a deep dynamic neural network that reads medical records and predicts future medical outcomes. At the data level, DeepCare models patient health state trajectories with explicit memory of illness. Built on Long Short-Term Memory LSTM, DeepCare introduces time parameterizations to handle irregular timing by moderating the forgetting and consolidation of illness memory. DeepCare also incorporates medical interventions that change the course of illness and shape future medical risk. Moving upi¾źto the health state level, historical and present health states are then aggregated through multiscale temporal pooling, before passing through a neural network that estimates future outcomes. We demonstrate the efficacy of DeepCare for disease progression modeling and readmission prediction in diabetes, a chronic disease with large economic burden. The results show improved modeling and risk prediction accuracy.

Subfields: Artificial Intelligence, Health Information Management, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Health Information Management

In Signal Processing

Citations in the corpus

Deep Reinforcement Learning for Dynamic Treatment Regimes on Medical Registry Data (2017, background): “Some pioneering works [20], [21] are using recurrent neural networks (RNNs) to model the sequence of hospital admission, diagnosis and treatments for broad disease categories with large electronic health record (EHR) datasets.”
Interpretable Representation Learning for Healthcare via Capturing Disease Progression through Time (2018, background): “[22] modifies the forget gate of the standard LSTM unit to account for time irregularity of admissions.”
Recurrent Neural Networks for Multivariate Time Series with Missing Values (2016, neutral): “ods, we average the soft-labels of several classiﬁers and treat it as the ensemble prediction. Recently RNN models have been explored for modeling diseases and patient diagnosis in health care domain [14, 5, 19] using doctor notes but are not readily applicable for comparison in our time series clas”
Patient Subtyping via Time-Aware LSTM Networks (2017, methodology): “First approach, denoted by MF1-LSTM, multiplies the output of the forget gate by ( ∆ t ) such as f t = д ( ∆ t ) ∗ f t . whereas MF2-LSTM utilizes a parametric time weight such as 3 when ∆ t is measured in days similar to [24].”
Deep learning for healthcare: review, opportunities and challenges (2018, unknown): “RNNs with long short-term memory (LSTM) hidden units, pooling and word embedding were used in DeepCare [58], an end-to-end deep dynamic network that infers current illness states and predicts future medical outcomes.”
Deep EHR: A Survey of Recent Advances in Deep Learning Techniques for Electronic Health Record (EHR) Analysis (2017, methodology): “Restricted Boltzmann Machine (BM) Popular RNN variants include the long short-term memory (LSTM) and gated recurrent unit (GRU) models, both referred to as gated RNNs.”
The rise of artificial intelligence in healthcare applications (2020, background): “The long short-term memory of the system models the illness trajectory and healthcare processes of users via a time-stamped sequence of events and in this way allows capturing long-term dependencies [41].”
BEHRT: Transformer for Electronic Health Records (2019, methodology): “s (e.g., key diagnoses such as diabetes can stay a risk factor over a person’s life, even decades after their ﬁrst occurrence; certain surgeries may prohibit certain future interventions). Pham et al [15] introduced a Long Short-Term Memory (LSTM) architecture with attention mechanism, called DeepCa”