Digital Detection of Dementia Studies (D cubed Studies).

PROJECT SUMMARY/ABSTRACT
Every year Alzheimer’s disease and related dementias (ADRD) adversely affect millions of Americans at a
societal cost of more than $200 million.1 Concurrently, half of Americans living with ADRD never receive a
diagnosis.2-7 Early detection helps those with ADRD and their caregivers better plan and potentially lessen the
burden of lengthy and costly medical care. Current investigational approaches using biomarkers for early
detection are invasive, costly, and sometimes inaccessible to patients. The National Institute on Aging calls for
the development of effective, scalable and low cost approaches for early detection of ADRD (RFA-AG-20-051).
Currently, primary care clinicians provide the majority of care to older adults living with ADRD.1-5 Our
interdisciplinary scientific teams have developed and tested scalable early detection approaches.10, 11 We are
proposing to evaluate an integrated approach embedded in the Annual Wellness Visit (AWV) that leverages
Electronic Health Record systems, machine learning models, and patient reported outcomes to deploy a low-
cost and scalable approach for early detection of ADRD. Our proposed studies will leverage previously
developed machine learning models (Passive Digital Marker) and patient reported outcomes (Quick Dementia
Rating Scale). The design of our proposed studies is predicated on the notion that patient screening is done to
identify a more targeted group of referral for applicable diagnostic and management services. We will conduct
two complementary multi-site studies to evaluate the effectiveness of two scalable approaches for early
detection of ADRD. The first study will be a clinical validation study of the three scalable approaches; the
Passive Digital Marker (PDM) that uses EHR data, the Quick Dementia Rating Scale (QDRS) that uses patient
reported outcomes (PROs) imbedded within the EHR system, and the combination of both (PDM + QDRS).
The second study will be a pragmatic cluster-randomized controlled comparative effectiveness trial of two
screening approaches embedded within the AWV, as compared to the AWV-only process, in increasing the
incidence rate of new ADRD. In the final year of the study, we will share our codes for both the Passive Digital
Marker and the QDRS with Epic headquarters to ensure that these codes are available for any healthcare
system with Epic nationwide.
The high costs of treating Alzheimer’s disease and the costs incurred by patients and caregivers, both tangible
and intangible, are a major threat to public health and the US economy. Developing scalable and low cost
instruments and assessments integrated into EHR data will assist physicians in early detection, more and
better diagnoses, and clinically meaningful care recommendations. Cost effective, scalable, and noninvasive
models are urgently needed to proactively mitigate these costs and prolonged medical care.

Grant Number: 5R01AG069765-05
NIH Institute/Center: NIH
Principal Investigator: MALAZ BOUSTANI