# Integrating Acute and Ambulatory Care with Post-Discharge Monitoring and Machine Learning

> **NIH NIH R01** · STANFORD UNIVERSITY · 2024 · $726,452

## Abstract

ABSTRACT
Over 20% of patients discharged from the Emergency Department (ED) have unplanned revisits within 30
days, often due to preventable causes. Upon revisit, the ED physician lacks vital data on the timeline of events
and physiologic changes leading to the patient’s return, which can lead to delayed diagnosis and over-testing.
Continuous monitoring of vital signs and activity can produce detailed information about a patient’s condition
and stability, both in the hospital and after discharge. It is not known, however, which patients benefit most from
post-discharge monitoring (PDM), which monitoring signals and strategies best predict quality of life and ED
revisit risk for specific patient populations, and how PDM data can be made diagnostically useful when a
patient returns to the hospital. To address these gaps, we aim to produce a framework for the integration of
PDM and acute care to improve our understanding of ED patient trajectories, both after discharge and upon
revisit. Specifically, we hypothesize that integrating hospital data and PDM can improve the predictability of ED
revisits, identify potential targets for post-discharge interventions, and improve diagnosis and disposition of ED
revisits that cannot be prevented. We will enroll a clinically and demographically diverse cohort of ED patients
at high risk of revisit within 30 days, and configure noninvasive wearable monitors with an accompanying
smartphone app to continuously track activity and physiology after discharge. We will develop interpretable
deep learning models to predict revisits and changes in health-related quality of life, and characterize, for
specific patient populations, the monitoring signals and measurement frequencies most relevant to predicting
revisits and quality of life, and the prediction horizons in which preventive interventions could be delivered.
Finally, we will combine in-hospital and PDM data to develop and evaluate an intervisit report for the ED
physician treating a returning patient, summarizing the relevant trends in patient physiology, activity, and
health-related quality of life between visits, and including a large language model-derived interpretation of the
antecedents of the return visit. Better understanding how and for which patient populations PDM can predict
ED revisits and quality of life can improve the integration of acute and ambulatory care, identify new clinical
use cases for existing monitoring technologies, and inform the design and timing of preventive interventions.
Analyzing intervisit trajectories can reveal the antecedents of acute presentations, and improve diagnosis and
disposition upon ED revisit.

## Key facts

- **NIH application ID:** 10850216
- **Project number:** 1R01HL172794-01
- **Recipient organization:** STANFORD UNIVERSITY
- **Principal Investigator:** David Andrew Kim
- **Activity code:** R01 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2024
- **Award amount:** $726,452
- **Award type:** 1
- **Project period:** 2024-06-01 → 2029-03-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10850216

## Citation

> US National Institutes of Health, RePORTER application 10850216, Integrating Acute and Ambulatory Care with Post-Discharge Monitoring and Machine Learning (1R01HL172794-01). Retrieved via AI Analytics 2026-05-26 from https://api.ai-analytics.org/grant/nih/10850216. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*