Opioid abuse risk model uses clinical, administrative and social determinant data
PALO ALTO, Calif. /PRNewswire/ -- Recent data from the CDC show that overdose deaths from opioids have increased by more than five times in the last 20 years. In fact, two-thirds of all drug overdose deaths in 2016 were opioid-related. The CDC estimates the financial cost of opioid abuse at $78 billion annually, with over one-third of that amount going towards increased health care and substance abuse treatment costs.
These statistics and alarming increases in opioid abuse led HBI Solutions to build predictive algorithms on integrated clinical, administrative and social determinant data to help identify patients most at risk for opioid abuse. Powered by data from across the care continuum and HBI's proven data science capabilities, the new model identifies individuals at risk for abusing opioids, assigns a risk score, and exposes the clinical, social and lifestyle factors driving that risk to help clinicians take action.
It couldn't come at a better time, as the push for improved opioid prescribing methods and better population health management has led hospitals and health systems to increase their focus on social determinants of health and improve care for patients at risk for abusing opioids.
Traditionally providers have turned to standard prescription, dosage and monitoring practices. However, non-medical conditions and situations like housing status, income and even zip code can influence opioid abuse risk as much as some medical conditions and are leading the drive towards more patient-centered clinical practices.
HBI's new model helps identify which patients in a certain population should be targeted for intervention and gives providers both clinical and non-clinical factors driving that risk. "We've known for some time that the collection and analysis of social determinant of health data is helpful in the identification of those at risk for any type of health event," said HBI Solutions CEO, Eric Widen.
Always expanding its risk model inventory, HBI decided to add opioid abuse risk to its list of peer-reviewed and validated models in 2017. "In building this and other new models, our machine learning processes have found social determinants to be particularly important to improving the accuracy of the predictions," said Widen.
Officially offered to the market in late 2017, two HBI Solutions clients are putting the model to the test. Hospitals and general medical providers often don't have a full medical and social profile for the patient and may not have time to dig through past visit notes and summaries where risk factors, particularly social determinants, could be hidden. Having this machine learning tool at the point of care, helps clients target the highest risk patients, better allocate resources, address harms of opioid use and hopefully save lives.
According to Widen, "Our data science approach is unique. Our risk models use any and all available data, employ natural language processing to incorporate data from notes and summaries, deliver predictions in real-time, and present risk in context with a patient's full medical profile."
