Registered Trial Tests EHR Alert for Missed Cognitive Impairment in Primary Care

A cluster-randomized trial in 38 primary care clinics is testing whether an alert built into the electronic health record can help clinicians catch cognitive impairment they might otherwise miss, with the first published results in on clinician confidence, not on diagnosis rates.

A pragmatic, cluster-randomized trial registered on ClinicalTrials.gov (NCT05723523) has tested whether a clinical decision support system built into the electronic health record changes how often primary care clinicians detect and manage cognitive impairment, according to the trial's published protocol in Contemporary Clinical Trials. The study, run across 38 primary care clinics in the Twin Cities metropolitan area and randomized 1:1 to the tool or usual care, uses a system called CI-CDSS that scans electronic health record data at every primary care visit and flags patients who have an abnormal cognitive assessment on file or who score as high risk on a prediction model estimating the likelihood of developing cognitive impairment within three years. Flagged cases route to the clinician with point-of-care prompts for evaluating, diagnosing and managing mild cognitive impairment and dementia.

The first published results from the trial, in BMC Primary Care, do not address whether the tool actually increased diagnoses. They report on a secondary measure: clinician confidence, gathered by survey before the tool went live and again eight months after. Clinicians exposed to the CI-CDSS and its accompanying training reported more confidence in managing patients with cognitive impairment, the study found, but no corresponding gain in confidence to diagnose it. The trial's primary outcome, whether cognitive impairment is actually documented more often in the electronic record among patients at CI-CDSS clinics compared with usual-care clinics, has not yet been reported; the protocol specifies that outcome is measured up to 18 months after a patient enters the study.

Clinics were enrolled from August 2023 through April 2024, and patient follow-up continued through February 2025. The protocol instructs that clinicians be trained the system is meant to supplement, not replace, clinical judgment, and that they remain free to disregard its prompts at any visit. That built-in override, together with the fact that the headline detection-rate result is still pending, means the trial so far documents a shift in how clinicians say they feel about managing cognitive impairment, not a change in how often it gets diagnosed or what happens to patients afterward.

The trial's co-investigators say results on the primary detection outcome and on healthcare costs are still to come; until they are published, the effect of the tool on actual diagnosis rates in primary care remains unknown. Families and clinics wanting to know whether a similar alert is active in their own electronic record should ask their primary care clinic directly.