Improving palliative care with deep learning

By Anand Avati, Kenneth Jung, S. Harman, L. Downing, Andrew Y. Ng, Nigam H. Shah. IEEE International Conference on Bioinformatics and Biomedicine, 2017

410 citations recorded by the source corpus.

BackgroundAccess to palliative care is a key quality metric which most healthcare organizations strive to improve. The primary challenges to increasing palliative care access are a combination of physicians over-estimating patient prognoses, and a shortage of palliative staff in general. This, in combination with treatment inertia can result in a mismatch between patient wishes, and their actual care towards the end of life.MethodsIn this work, we address this problem, with Institutional Review Board approval, using machine learning and Electronic Health Record (EHR) data of patients. We train a Deep Neural Network model on the EHR data of patients from previous years, to predict mortality of patients within the next 3-12 month period. This prediction is used as a proxy decision for identifying patients who could benefit from palliative care.ResultsThe EHR data of all admitted patients are evaluated every night by this algorithm, and the palliative care team is automatically notified of the list of patients with a positive prediction. In addition, we present a novel technique for decision interpretation, using which we provide explanations for the model’s predictions.ConclusionThe automatic screening and notification saves the palliative care team the burden of time consuming chart reviews of all patients, and allows them to take a proactive approach in reaching out to such patients rather then relying on referrals from the treating physicians.

View Paper: https://www.semanticscholar.org/paper/597de1eed79608bcceee1bba0bafd2ac48f0d26a

How researchers see this paper

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

Citations in the corpus

Hopular: Modern Hopfield Networks for Tabular Data (2022, methodology): “Since Deep Learning methods have already been successfully applied to larger tabular datasets (Avati et al., 2018; Simm et al., 2018; Zhang et al., 2019b; Mayr et al., 2018) we want to know whether Hopular is competitive on small tabular datasets.”
Detecting semantic anomalies (2019, background): “In situations where positive examples are not only much rarer, but also of primary interest for detection, AUROC scores are a poor reﬂection of detection performance; precision is more relevant than the false positive rate [23–25].”
NGBoost: Natural Gradient Boosting for Probabilistic Prediction (2019, background): “The thin gray lines are the upper and lower quantiles covering 95% of the prediction distribution. on structured medical records of the patient (Avati et al., 2018)) are important examples.”
Deep Learning in Pharmacogenomics: From Gene Regulation to Patient Stratification (2018, background): “In addition, recent studies distinguish between interpreting a model and interpreting its decision, arguing that although interpreting a complex model is hard, often users only want an explanation for the prediction made by the model for a given example [42].”
Scalable and accurate deep learning with electronic health records (2018, unknown): “However, there are other possible techniques for interpretability of deep learning models [4, 92], and further research is needed regarding both the cognitive impact of this approach and its clinical utility.”
Axes of a revolution: challenges and promises of big data in healthcare (2020, background): “…subgroups of patients who are most likely to deteriorate or develop a certain complications of the disease can enable targeting of these patients and the use of strategies such as more frequent follow-up schedule, changes in medication regime or a shift from traditional care to palliative care 92 .”
Survey on categorical data for neural networks (2020, unknown): “We copy a table from [55] to give the reader a clear idea of how Avati et al. segment EHR data into windows and slices.”
Artificial intelligence in retina. (2018, methodology): “Such records were successfully used to predict the need for palliative care, in-hospital mortality and unplanned read-mission (Avati et al., 2017; Rajkomar et al., 2018).”
TabLLM: Few-shot Classification of Tabular Data with Large Language Models (2022, unknown): “We predicted the end-of-life (EoL) of all patients older than 70 years, which can be used to inform care in a palliative setting (Avati et al., 2018).”
Explainable Prediction of Medical Codes from Clinical Text (2018, background): “ICD codes have a variety of uses, ranging from billing to predictive modeling of patient state (Choi et al., 2016; Ranganath et al., 2015; Denny et al., 2010; Avati et al., 2017).”