But our latest innovation will do something truly unique to align with the industry’s growing focus on personalized health.
And if we play our cards right, it could also result in a nice reward from CMS.
Currently in the testing phase, we expect to unveil a solution by year end that will identify key factors that influence hospital readmission for diabetes and predict the likelihood of diabetic patient readmission. With preliminary results showing the solution being more than 90 percent accurate, we are very enthusiastic about its future success.
Why is this such a big deal?
According to a recent American Diabetes Association report, the estimated cost of diagnosed diabetes in 2017 was $327 billion, including $237 billion in direct medical costs, with hospital inpatient care accounting for 30 percent of the total medical cost. For the cost categories analyzed in the report, care for people with diagnosed diabetes accounts for 1 in 4 health care dollars in the U.S., and more than half of that expenditure is directly attributable to diabetes.
Additionally, hospital readmission is considered as a metric for quality of care by the Centers for Medicare and Medicare Services (CMS) and under the Hospital Readmission Reduction Program, CMS penalizes hospitals for high rates of readmission.
Assuming the solution works as we envision, it will expand the scope of applying artificial intelligence to help prevent readmissions for other conditions like cancer, COPD, mental illness and substance abuse.
Last month CMS announced The CMS AI health outcomes challenge, an industry competition promoting innovation in the healthcare analytics space to demonstrate how AI tools can be used to predict such things as unplanned hospital and skilled nursing facility admissions. With our solution aligning perfectly with the competition and $1.65 million in total prize funds up for grabs, count us in!
In terms of our R&D efforts, we have been consulting with top global diabetes specialists and running data through AI models to predict which patients will end up in the hospital, find cohorts of diabetic patients not getting the right drug (or could do better if the drug were changed) and building data models along those lines. After exhaustive rounds of tests to get our models in line with targets, we have reached an inflection point in the development of a product that combines data with artificial intelligence.
Relying on data from across the care continuum (clinical, claims, pharmacy, labs) the solution combines machine learning with massive amounts of data, including longitudinal patient records, previous hospitalizations for patients or a patient cohort, family history, social determinants of health, care coordinator feedback, patient-supplied updates on glucose, exercise and diet, and other personal and population factors. It will be able to predict when diabetic patients are likely to be readmitted to the hospital should they not address the change in health detected by the solution.
A simple alert to a patient that their current blood levels combined with other factors put them in the same statistical range where they have been previously hospitalized, also triggers a call to the primary care provider.
The goal is for diabetic patients to be notified that they are nearing a person healthcare crisis, and if immediate action isn’t taken they will likely be heading to the hospital soon.
We look forward to providing an update on our progress!