Photo credit: Regenstrief Institute
Researchers at Regenstrief Institute and Indiana University have further enhanced the performance of Uppstroms, a machine learning application that identifies patients who may need full-service referrals, by integrating additional data sources at the personal and population levels, as well as advanced analytical approaches.
Members of the research team include Regenstrief, IU Fairbanks School of Public Health at IUPUI, IU School of Medicine, and Eskenazi Health.
Uppstroms has been used in nine clinics connected to a safety net hospital in Indianapolis. The algorithm identifies patients in primary care with social risks such as behavioral health or problems with food or housing. This allows doctors to offer these patients referrals to specialized services such as nutritionists, behavioral medicine, or social workers to meet the need before it becomes a crisis.
There is evidence that at least one in four adults, and possibly one in two, has a need that is determined by social determinants of health.
“These all-round services can enhance primary care delivery by addressing socio-economic, behavioral, and financial needs that primary care providers cannot address,” said Suranga Kasthurirathne, Ph.D., first author of the paper, Regenstrief research scientist and assistant professor for Pediatrics at the IU School of Medicine. “To make it more useful in the clinical setting, we’ve integrated a wide range of patient-level data and more detailed population status data to improve the app’s accuracy and reduce false positives.”
Innovations to previous approaches
Additional data added to the algorithm included social determinants of health, insurance, drug history, and behavioral health history at the patient level. This data comes from Eskenazi Health’s electronic health record system and the Indiana Network for Patient Care, which is maintained by the Indiana Health Information Exchange. Social determinants of population-level health, measured in the area of the census area smaller than the area covered by a zip code, were derived from the US Census Bureau, the Marion County Public Health Department, and community health surveys.
The research team evaluated the new decision models and found that they outperformed previous models. The new patient-level data and advanced analytical approaches played a key role in improving precision.
“So much of health impacts happen outside of a doctor’s office,” said senior author Joshua R. Vest, Ph.D., MPH, research scientist at Regenstrief and professor and director of the Center for Health Policy at IU Fairbanks School of Public Health at IUPUI. “Health systems are working to integrate these social determinants of health into the EHR. This study shows the advantage of capturing social factors in the EHR during clinical visits and using them for clinical decisions.”
In addition to the added data elements, the study team adapted the application in a manufacturer-neutral manner so that it can be implemented in any electronic patient record system.
The next steps for the researchers are to develop a way to use unstructured data in the EHR and include it in the algorithm.
“Precision Health Enabled Machine Learning to Determine Circumferential Social Services Needs Using Patient and Population Level Datasets: Algorithm Development and Validation” was published online in JMIR Medical Informatics.
Machine learning models are helping clinicians identify people in need of advanced treatment for depression
Suranga N. Kasthurirathne et al. Precise, Health-Enabled Machine Learning to Identify Comprehensive Social Services Needs Using Data Sets at the Patient and Population Level: Algorithm Development and Validation, JMIR Medical Informatics (2020). DOI: 10.2196 / 16129 Provided by the Regenstrief Institute
Quote: Additional data, advanced analytics improve the performance of the machine learning referral app (2020, October 16) released October 16, 2020 from https://medicalxpress.com/news/2020-10-additional-advanced-analytics -machine-referral.html was obtained
This document is subject to copyright. Except for fair trade for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.