IMPORTANCE: Detection of Alzheimer disease and related dementias (ADRD) is a challenge in primary care settings. The Quick Dementia Rating System (QDRS) and the Passive Digital Marker (PDM) were developed as 2 scalable tools for early detection of ADRD. The QDRS is a patient-reported outcome measure, while the PDM is a machine learning algorithm that uses electronic health record (EHR) data. Both can be embedded in the EHR for ease of use.
OBJECTIVE: To evaluate the effect of the combined approach of QDRS plus PDM in improving the annual rate of new documented ADRD diagnoses in primary care practices.
DESIGN, SETTING, AND PARTICIPANTS: This randomized clinical trial was performed between July 2, 2022, and July 1, 2024. Nine federally qualified primary health care clinics in Indianapolis, Indiana, were randomized to usual care (no routine screening for ADRD), PDM only, or QDRS plus PDM. Participants were treated as per the randomization arm. Participants included adults 65 years and older, without a diagnosis of mild cognitive impairment, dementia, or severe mental illness. Data analysis was based on intention to treat and was performed between November 1, 2024, and August 20, 2025.
INTERVENTION: Randomization of clinic to usual care, PDM only, or QDRS plus PDM.
MAIN OUTCOMES AND MEASURES: The primary outcome was 12-month cumulative incidence of ADRD diagnoses; the secondary outcome was any ADRD diagnostic workup, such as laboratory tests, neuropsychological testing, or brain imaging.
RESULTS: The study included 5325 patients (mean [SD] age, 71.1 [5.9] years), 3312 (62.2%) of whom were female. Compared with the usual care clinics (12-month incidence, 213 of 1724 [12.4%]), the odds of an incident ADRD diagnosis were higher in the clinics randomized to QDRS plus PDM (12-month incidence, 355 of 2301 [15.4%]; adjusted odds ratio [AOR], 1.31; 95% CI, 1.05-1.64) but not the clinics randomized to PDM only (12-month incidence, 134 of 1300 [10.3%]; AOR, 0.84; 95% CI, 0.63-1.11]). Compared with the usual care clinics (12-month incidence, 500 of 1724 [29.0%]), the odds of ADRD diagnostic assessments were higher in the QDRS plus PDM clinics (12-month incidence, 844 of 2301 [36.7%]; AOR, 1.41; 95% CI, 1.12-1.77) but not the PDM clinics (12-month incidence, 362 of 1300 [27.8%]; AOR, 0.94; 95% CI, 0.72-1.22).
CONCLUSIONS AND RELEVANCE: This randomized clinical trial found that the combined approach was effective at scale for the early detection of ADRD in primary care settings. This is an important feature in busy primary care settings that can benefit both the health care system and patients.
TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT05231954.
| Discipline Area | Score |
|---|---|
| Physician |  |
Physician rater
Not clear that the machine learning algorithm (PDM) adds any value to patient-reported data (QDRS). Isn't it possible that if a patient reports dementia symptoms, it will on its own lead to the doctor diagnosing dementia?
Physician rater
For diagnosing dementia, discipline matters. This fascinating manuscript compares more rigorous questioning on the dimensions of cognitive impairment and incorporating electronic data abstraction into assessing patients. The benefits in terms of patient safety and outcomes of detecting early cognitive decline has yet to be determined.
, McMaster University