Client’s Challenge
Optimizing the Medical Resident Experience
Our client needed to evaluate medical residents effectively and objectively across all involved sites of care while customizing the learning experience for each resident.
Case Studies
Building Machine Learning Models for Medical Residence Program
Optimizing Student Evaluations Through Customized Data Science Tools
Our Solution
Developing Machine Learning Models
After analyzing existing resident data, InfoWorks data scientists developed machine learning models for three significant areas of data that affect the outcome of the resident training programs.
Value Realized
Customized and Objective Resident Program
Using the machine learning models, our client will be able to improve the Residence program, gain knowledge to tailor trainings to individual resident needs, and add objectivity to resident/fellow assessments.
Project Details
Our client, a Fortune 100 healthcare organization, wanted to more effectively and objectively evaluate students in their residency program across all involved sites of care. In addition, they wanted the ability to tailor trainings to individual residents, helping them to meet their goals and advance in their specialized field more efficiently.
The InfoWorks data science team analyzed existing resident data, identified relevant data assets, redesigned and retrained their machine learning models using resident-associated EHR data, and developed machine learning models for three significant areas of data that impact the outcome of the resident training programs.
With these machine learning models, our client is able to offer individualized training programs, continuously gain knowledge on how to tailor trainings to meet individual resident needs, and add objectivity to resident/fellow assessments.
To ensure they were successful, we provided coaching on how to utilize data science for the medical residence program, which tools fit into their existing framework, and how to utilize the tools for increased impact.
The retrained machine learning models resulted in an increase in the average model accuracy rate from 53% to 90%. Our client was able to more effectively and objectively evaluate students and offer tailored training solutions.
