Hospital readmissions carry significant financial costs and are associated with negative patient outcomes. AI could be used to predict the Readmission Risk per patient, aimed at preventing readmissions and improving outcomes for patients.
Business impact: reduce hospital readmission cost, improve treatment efficacy, save more people’s lives and moneyÂ
Understand the Use-case under 5 minutes
Article (6 minutes)
Explains the value of using ML to model patient readmissions. Includes ROI figures from multiple health entities, along with some expected challenges
VisitVideo (4.5 minutes)
Explains the benefit of using Machine Learning vs Manual approaches to calculate Readmission Risk per Patient. Expands on the data sets needed, modelling, and value to providers
WatchGet to know more Business and Technical details about the use-case (15-30 minutes)
More detailed introduction covering business and technical aspects
Video (8.5 minutes)
Using predictive analytics to reduce the risk of an acute hospital readmission. Explains the different models used, provide examples for the data and variables used per each, and share some ROI figures
WatchVideo (7 minutes)
Intuitive step by step explanation for the process of building a Predictive Model that would predict the likelihood of a patient’s readmission to the hospital: data sourcing, feature engineering, model development, and model usage
VisitVideo Demo (6 minutes)
A Demo by ProKarma showing how they can reduce hospital readmissions. Provides a brief about the needed data, solution architecture, modelling work, and system modules
WatchCase studies, Organizational Aspects, Return on Investment examples
Article (9 minutes)
Highlighting the process Mission Health went through when shifting from LACE score to Machine Learning for predicting readmission risk. Including: key business questions asked, assembled team, milestones, and key outcomes
ReadCase Study (6 minutes)
Case studies for hospitals implementing AI for predicting patient readmissions, with ROI up to 25% less readmission rates. read the full peer-reviewed study
ReadCase Study (7 minutes)
University of Washington (UW) Tacoma’s Center for Data Science using machine learning to predict risk-of-readmission factor as well as corresponding actionable guidelines for the patient-provider team
ReadArticle (8 minutes)
Showing the difference between using classical techniques for calculating readmission risk (like LACE score) vs AI. AI provides better accuracy and up to $6.5 Million in cost savings for hospitals
ReadMore details on the technical aspects of the use-case
Video (3 minutes)
Short Demonstration for using IBM Watson to build models Predicting Readmission Risk. The video shows the steps of data ingestion, feature exploration, and model development - most of which are automated by Watson
WatchPaper (16 minutes)
Explaining the technical architecture and models used in Microsoft’s RaaS (Readmission risk as a Service), which is an Azure Service hospitals can use to predict patient readmission
ReadTechnical resources that will help you implement the use-case (notebooks, tutorials..)
Article + Github Repo
Technical Article exploring the problem of Prediction of Hospital Readmission in diabetic inpatients, and explaining how they achieved 94% accuracy. It includes a Github Repo with full code + data
VisitGithub Repo
End to End ML Pipeline including data ingestion, data exploration, feature engineering, model training, model optimization, deployment to production, and real-time inference. Leverages AWS technologies
VisitArticle (50 minutes)
Exploring the methodologies followed and results achieved in 43 major study applying Machine Learning to Predict Hospital Readmission. Includes summery findings and links to the studies
VisitGithub Repo
Using 283,208 clinical notes and discharge summaries on 35,779 patients, various NLP models were assessed in their ability to predict all-cause 30-day readmission for all patients
VisitData Sets you can use to build Demos, POCs, or test Algorithms
The NRD contains data from approximately 18 million discharges each year. Weighted, it estimates roughly 35 million discharges.
10 years (1999-2008) of clinical care at 130 US hospitals and integrated delivery networks. It includes over 50 features representing patient and hospital outcomes. Information was extracted from the database for encounters that satisfied the following criteria
Real hospital stays for patients admitted to a tertiary academic medical center in Boston, MA, USA. MIMIC-IV contains comprehensive information for each patient: lab measurements, medications, vital signs, etc
Off-the-Shelf Products using AI for Predicting Hospital Readmission
Got a Question or a Resource to share with the Community? Please do!
Copyright © 2024 AI Cases. All rights reserved
Session expired
Please log in again. The login page will open in a new tab. After logging in you can close it and return to this page.