More than half of the 6.5 million Americans living with Alzheimer’s disease and related dementias (ADRD) do not have a diagnosis, which can lead to poor quality care for these patients and inadequate reimbursement for costs of care for health plans. This Small Business Innovation Research (SBIR) Direct Phase II grant will enable us to test, refine, and bring to market a unique algorithm that will use electronic health data to support earlier identification of people living with undiagnosed ADRD.

The primary goals of this Small Business Innovation Research (SBIR) Direct Phase II project are to test, refine, and bring to market a novel algorithm that uses electronic health data (including electronic health records [EHR] and insurance claims) to improve detection and diagnosis of Alzheimer’s disease and related dementia (ADRD). More than half of the 6.5 million Americans currently living with ADRD are undiagnosed. Lack of a diagnosis can result in delayed access to treatment, poor care management, unnecessary hospitalizations and emergency department (ED) visits, difficulty managing other chronic conditions, inability to participate in planning for future care, risky behaviors such as driving, and delayed access to caregiver support. Studies also have shown that healthcare costs for people with ADRD are three times higher than those without ADRD. For this reason, health plans are reimbursed by Medicare at a higher rate for patients diagnosed with ADRD. Multiple principal investigator (PI) Dr. Deborah Barnes and colleagues have developed and validated a novel algorithm called eRADAR (electronic health record [EHR] Risk of Alzheimer’s and Dementia Assessment Rule) that uses EHR data to identify patients at risk of having undiagnosed dementia with high accuracy. They have externally validated eRADAR in two healthcare systems, including confirming that the algorithm works well over time and in different race and ethnic groups. A limitation of eRADAR, however, is that it was designed as a static tool, and it does not learn from or adapt to new data streams. Many organizations that are interested in improving dementia detection, such as Medicare Advantage plans, have access to insurance claims data rather than EHR data. As part of Phase I proof-of-concept work for this grant, we created RADAR (Risk of Alzheimer’s and Dementia AlgoRithm), which uses Medicare claims data rather than EHR data to calculate a risk score for undiagnosed dementia that is similar to eRADAR. The primary goals of this SBIR Direct Phase II are to: 1) Retrospectively assess the accuracy of RADAR to detect undiagnosed dementia using historical Medicare claims data; 2) Develop an adaptive process using machine learning techniques to identify optimal RADAR predictor sets based on data inputs; and 3) Determine the optimal strategy for sharing RADAR’s output with clients interested in improving dementia detection and diagnosis. Together Senior Health’s mission is to serve people living with memory loss, cognitive decline and ADRD by providing evidence-based, online programming. Our current products include Moving TogetherTM–a mind-body group movement program for people with mild ADRD and care partners—and Brain Health Together, which combines Moving Together with brain health coaching to address modifiable risk factors. Our current clients include Medicare Advantage plans, home health agencies, and ADRD care centers. Adding RADAR to our product line will enable us to better support our business clients while also expanding access to our services and improving quality of life for people living with ADRD and those who care for them.