Automated Risk Calculator Promises Easily Implemented Prompt for Cognitive Screening

As new treatments and strategies to slow the progression of Alzheimer’s disease and dementia become more available, patients and providers face a new challenge: diagnosing cognitive decline early enough so that interventions may still help.

For many patients with Alzheimer’s disease, the window for successful intervention may close before cognitive symptoms are noticed. In response, a multidisciplinary team at Cleveland Clinic has developed a low-cost, easy-to-use automated tool that may help identify older adults in the earliest stages of Alzheimer’s disease and dementia, when the conditions are often still unrecognized.

The tool, published in Alzheimer’s & Dementia (2025;17[2]:e70136), is a cognitive risk calculator built from routine electronic health record (EHR) data. It integrates into the charting system, analyzing patient data during visits and alerting physicians when further screening should be conducted. Led by Darlene Floden, PhD, a neuropsychologist with Cleveland Clinic’s Center for Neurological Restoration, the automated tool aims to make early detection easier, more accurate and more accessible.

The calculator’s development at a glance

Cognitive problems often go undiagnosed. Almost all risk calculators for cognitive decline and dementia are trained on existing diagnoses from the medical record. Dr. Floden says as many as 50% of those records may contain a misdiagnosis or missing diagnosis for dementia. To avoid carrying that type of flawed data forward, she and Cleveland Clinic colleagues in the Center for Geriatric Medicine and the Department of Internal Medicine set out to use objective cognitive testing to establish cognitive status and thereby develop a cognitive risk model.

They prospectively recruited 337 older patients (aged 60 or above) from five primary care and geriatrics clinics in Northeast Ohio. All underwent full neuropsychological assessment, with 109 individuals (32%) being diagnosed with early cognitive impairment.

Biostatisticians then analyzed five years of EHR data from all patients’ annual visits to identify patterns that might predict cognitive decline.

“We wanted to assess whether there’s a way to accurately identify cognitive risk simply by using the wealth of information collected in a typical primary care record,” Dr. Floden explains. “This would make it easy to integrate into the electronic health record.”

The researchers developed a model that estimates dementia risk based on seven common variables associated with cognitive decline that are typically available in the EHR: age, race, pulse, systolic blood pressure, use of nonsteroidal anti-inflammatory drugs, history of a mood disorder and family history of neurologic disease.

When applied to the 337-patient primary care sample, the model successfully differentiated patients with cognitive impairment from those without cognitive impairment, with a concordance statistic of 0.72, which is comparable to more complex tools that require specialized biomarker testing.

“The strengths of our approach are that it’s built on objective cognitive testing and it establishes risk solely on the basis of data readily available in the primary care record,” notes Dr. Floden. “In these ways our model overcomes limitations of earlier passive risk calculators.”

Next steps in development

Cleveland Clinic is now piloting the tool in primary care. When a patient’s profile suggests elevated cognitive risk, a flag appears in their EHR. The program doesn’t trigger an automatic referral or diagnosis, but it prompts the provider to consider whether a cognitive screen is appropriate.

“It’s not meant to be the final word,” Dr. Floden says. “It’s a nudge — a scalable, low-cost way to help busy providers know where to look more closely.”

Early results are promising. Patients with flagged records are being screened for cognitive issues at higher rates than those without. Dr. Floden and her team are also collecting physician feedback about how easy the tool is to use in everyday clinical practice.

“Primary care visits are packed, so we wanted to make this as easy and painless as possible,” Dr. Floden says. “It’s meant to help physicians focus their time where it matters, and it reminds patients that cognitive health is worth discussing, even before symptoms appear.”

The above piloting efforts are part of an ongoing implementation trial that is formally evaluating the risk model’s utility and effectiveness. Future research plans include external validation in an additional sample of primary care patients who complete objective clinical assessment.