Approved Research

Developing and validating an Electronic Health Record-based Frailty Index Using Machine Learning

Lay summary

Frailty is a condition that affects older adults and can lead to bad health results, especially after surgery. It is hard to measure frailty in large groups of people and even before surgery, because it takes a lot of time, equipment, and space. To make it easier, we created a method that uses electronic health records (EHR) to check for frailty. In the first part of this study, we will use a computer technique to predict a person's frailty using their EHR from UK Biobank. We will create models that use information from EHR to predict if someone is frail and see how this connects to common health problems.

In the second part of the study, we will test a model we created before that predicts frailty before surgery. We used artificial intelligence techniques and data from the University of Florida to create a frailty index based on EHR that can predict if someone is frail before surgery. We will test if our model works well with data from UK Biobank participants and if it can predict how people do after surgery. If it works, the frailty index could be used by doctors to see if a patient is frail before surgery, helping them plan the surgery and care afterward.