Dementia risk prediction: Zero-minute assessment at less than a dollar

“Identification of dementia risk is important for appropriate care management and planning,” said study senior author Malaz Boustani, MD, MPH, of the Regenstrief Institute and IU School of Medicine. “We wanted to solve the problem of early detection of people who are likely to develop dementia with a solution that is both scalable and cost-effective for the healthcare system.

“To do this, we use existing information—passive data—that is already in the patient’s medical notes for what we call a zero-minute assessment at a cost of less than a dollar. Decision-driven content selection methodology is used to develop individualized dementia risk prediction or to show evidence of mild cognitive impairment.” .”

This technique uses machine learning to select a phrase or subset of sentences from medical notes in a patient’s electronic health record (EHR) written by a doctor, nurse, social worker, or other provider that relates to a target outcome over a defined period of time. observation period. Medical notes are narratives in the EHR that describe the patient’s health in free-text format.

Information selected to be extracted from medical notes to estimate dementia risk may include clinician comments, patient comments, blood pressure or cholesterol values ​​over time, observations of a family member’s mental state, or medication history, including prescription and over-the-counter medications. medicines as well as “natural” remedies and supplements.

Predicting dementia risk helps the patient, family, and healthcare providers access resources such as support groups and the Centers for Medicare and Medicaid GUIDE model program, which supports keeping individuals in their homes longer. This may also encourage the clinician to prescribe medications that are commonly used by older adults but are known to adversely affect the brain and to talk to the patient about over-the-counter medications that have similar properties. Knowing the risk of dementia may prompt physicians to consider new FDA-approved amyloid-lowering therapies that alter the course of Alzheimer’s disease.

“Our methodology combines both supervised and unsupervised machine learning to extract dementia-related phrases from the large amount of medical notes available for each patient,” said study co-author Zina Ben Miled, PhD, MS, of the Regenstrief Institute. affiliated scientist and a former Purdue University faculty member in Indianapolis. “In addition to improving predictive accuracy, this allows the healthcare provider to quickly confirm cognitive impairment by reviewing the specific text used to guide risk assessment based on our language model.”

“Regenstrief Institute and Indiana University researchers have been pioneers in demonstrating the utility of electronic health records since the early 1970s. Given the tremendous effort that both clinicians and patients expend to capture EHR data, the goal should be to achieve maximum clinical value. Regenstrief and IU School of Medicine “These data can be leveraged even beyond their central role in medical care,” said study co-author Paul Dexter, MD, from “In the future, machine learning methods can be used to identify patients at high risk of dementia.” This study provides an excellent and innovative example of the clinical value that can be derived from EHRs. Early identification of dementia will become increasingly vital, especially as new treatments become available. developed.”

While patients and caregivers are the ultimate beneficiaries of the use of the new technique, providing zero-minute assessment at a cost of less than a dollar has a clear advantage for primary care clinicians who are overburdened and often lack the time and training needed to implement specialized cognitive management. tests.

A 5-year clinical trial of the risk prediction tools that the study authors are conducting in Indianapolis and Miami is in its final year. Lessons learned from this trial will enable them to advance the use of the dementia risk prediction framework in primary care practices. The researchers plan future studies on combining medical notes with other information contained in electronic health records and environmental data.