U.S. Reports

Disruption Of Healthcare With Machine Learning Is The Future Report

Issue link: https://hub.orionhealth.com/i/951768

Contents of this Issue

Navigation

Page 0 of 1

Disruption Of Healthcare With Machine Learning Is The Future, How Will This Curb Avoidable Readmission Rates and Reduce Healthcare Costs? Amadeus Intelligence curve) of 0.62 (where 0.5 is random chance). Published evidence now demonstrates that AUC scores of 0.75 to 0.85 are now achievable through the use of machine learning tools. There are two key themes related to improving accuracy that emerge from the literature: • Having a large data set for the Machine Learning Model to train on; the more patient records and the larger the medical history, the better the model can train itself. • The second is the breadth of data types available to the model; combined clinical data from acute and community providers, claims data, patient reported outcomes, social and demographic data can all be used to improve the predictive power of a Machine Learning model. Machine learning has the ability to provide clinical and financial insights that are driving healthcare toward personalized treatments and maximizing the cost effectiveness of interventions. Current Challenges for Healthcare Organizations Healthcare costs in the United States are spriralling upwards, with an aging population and a growing number of people with multiple chronic illnesses. Healthcare organizations are having to provide care for a large number of patients who require more intense and costly interventions. One of the reasons cost is increasing for healthcare organizations, is the high hospital readmission rates. Healthcare organizations are looking for ways to slow down or stop preventable readmissions. A useful tool would be the ability to predict which patients require hospital level care, and when, so that strategies can be put into place. Preventable readmissions is an area where machine learning can be used to augment current approaches and significantly reduce the cost and waste in healthcare. Hospital admissions are a major focus for payers and providers, 17.6% of hospitalizations in the United States result in a re- hospitalization within 30 days with an estimated 76% of those re-hospitalizations being potentially avoidable. The financial risk of high hospital readmission rates has been shifting to providers through financial penalties and withheld payments. Medicare estimate that they will withhold $564m in payments over the 2018 financial year. While there are many effective programs and interventions to reduce hospitalization rates—such as intensive transition management and patient support—come with a relatively high price tag. Current readmission prediction tools, such as the LACE Index for Readmission - Length of stay (days), Acute (emergent) admission, Charlson Comorbidity Index and Number of ED visits within 6 months (LACE), are relatively poor at predicting patient readmission. If health systems can't target these high cost, highly effective readmission interventions at the right patients, the interventions are unlikely to save a health system any money. The Machine Learning Approach Current models that predict readmission risk such as LACE have relatively poor predictive performance with AUC (area under the Why is Machine Learning Important? Orion Health is leading ground-breaking research in machine learning, exploring meaningful ways to minimize waste, reduce operating costs and help clinicians make more accurate decisions at the point of care. Significant amounts of data exist that will support better decision making, drawing on information from entire populations to treat and manage a person's health. The healthcare sector is being transformed by the ability to record massive amounts of information about patients and their environments. Machine learning provides a new way to find patterns and reason about data, which enables healthcare professionals to move closer to personalized medicine. There are many possibilities for how machine learning can be used in healthcare, and all of them depend on having sufficient data and permission to use it.

Articles in this issue

view archives of U.S. Reports - Disruption Of Healthcare With Machine Learning Is The Future Report