Readmission prediction using deep learning on electronic health records

• Unscheduled 30-day readmissions are a hallmark of Congestive Heart Failure (CHF) patients that pose significant health risks and escalate care cost. In order to reduce readmissions and curb the cost of care, it is important to initiate targeted intervention programs for patients at risk of readmission. This requires identifying high-risk patients at the time of discharge from hospital. Here, using real data from over 7500 CHF patients hospitalized between 2012 and 2016 in Sweden, we built and tested a deep learning framework to predict 30-day unscheduled readmission. We present a cost-sensitive formulation of Long Short-Term Memory (LSTM) neural network using expert features and contextual embedding of clinical concepts. This study targets key elements of an Electronic Health Record (EHR) driven prediction model in a single framework: using both expert and machine derived features, incorporating sequential patterns and addressing the class imbalance problem. We evaluate the contribution of each element towards prediction performance (ROC-AUC, F1-measure) and cost-savings. We show that the model with all key elements achieves higher discrimination ability (AUC: 0.77; F1: 0.51; Cost: 22% of maximum possible savings) outperforming the reduced models in at least two evaluation metrics. Additionally, we present a simple financial analysis to estimate annual savings if targeted interventions are offered to high risk patients. © 2019 The Authors

Ämnesord

MEDICIN OCH HÄLSOVETENSKAP  -- Medicinska och farmaceutiska grundvetenskaper -- Cell- och molekylärbiologi (hsv//swe)

MEDICAL AND HEALTH SCIENCES  -- Basic Medicine -- Cell and Molecular Biology (hsv//eng)

MEDICIN OCH HÄLSOVETENSKAP  -- Hälsovetenskap -- Hälso- och sjukvårdsorganisation, hälsopolitik och hälsoekonomi (hsv//swe)

MEDICAL AND HEALTH SCIENCES  -- Health Sciences -- Health Care Service and Management, Health Policy and Services and Health Economy (hsv//eng)

Nyckelord

Contextual embeddings

Electronic health records

Long short-term memory networks

Readmission prediction

Brain

Cost reduction

E-learning

Embeddings

Forecasting

Health risks

Long short-term memory

Patient treatment

Records management

Risk assessment

Risk perception

Class imbalance problems

Congestive heart failures

Discrimination ability

Electronic health record

Intervention programs

Prediction performance

Sequential patterns

Short term memory

Deep learning

Electronic health records

Publikations- och innehållstyp

ref (ämneskategori)

art (ämneskategori)