T-03: Diabetes Defender – Changing At-Risk Population Behavior Using Analytics
Chief Medical Officer
Allscripts Analytics United States
Development of an innovative approach to identify and manage diabetic populations is needed. Studies estimate that diabetes affects 8% of the U.S. population, but confirming this in undiagnosed and high risk patients remains challenging.
We identified pre-diabetics in the USA from more than 3000 Allscripts client sites and included de-identified longitudinal records on 40 million patients nationally using the AMA-CDC retrospective algorithm. 3.48 million patients met inclusion criteria by the algorithm.
Of the approximately 40 million patients analyzed, 3.48 million were pre-diabetic patients. Patients who met inclusion criteria demonstrated significant conversion of pre-diabetes to diabetes mellitus (near 80%) in tracked by time-series analysis of HbA1c’s over the three-year period. The burden of pre-diabetes was demonstrated geographically at the national, state and county level with real-time analysis. Given the complexity of the disease, provision of tailored insights (emphasis on diet, exercise, medications, optimize geography) to empower and incentivize individual patients to assume responsibility for preventative care in the real world is needed.
Once we have defined at-risk patients, we can begin to look at specific complex social and environmental risk factors that impact diabetes – that aren’t necessarily in the health data. For example, counseling a patient on diet modification is unlikely to help if your patient only has access to fast-food delivery because they can’t walk, or resides in a ‘food desert’ where fresh fruits and vegetables are unavailable. With targeted, informed interventions, patients can get treatment tailored to and relevant to their environments, which may include dietary advice, exercise programs or medication availability. This information must not only be delivered to the point of care for health providers, but also directly to the patient in a relevant format.
We are using historical insight from 40 million records in real populations to better understand pathogenesis of disease among different groups of people. If we can show positive patient results from thousands of other diabetic patients with similar BMI, or race or age, and other sociobehavioral drivers of health, we can empower people with understanding and responsibility for their own health.
Future goals are to create and refine predictive models to understand progression to DM, successfully identify at-risk populations for enrollment into DPP’s (Diabetes Prevention Programs) and other personalized treatment plans, quantify cost savings, and identify additional areas for potential intervention.
In this session, Dr. Paruk will also discuss the analytics findings from clinical (identifying and treating pre-diabetic populations) and financial data supporting the need for proactive management of this population, and discuss the need for incorporation of environment data to better enable successful health outcomes. With targeted, informed interventions, patients can get treatment tailored to and most relevant to their environments, which may include dietary advice, exercise programs or medication availability.