In the United States and around the world, people are living longer lives. As the population ages, so does the number of older adults who may experience declines in memory, attention, reasoning, or other thinking skills. Some of these changes in cognition can be treated and reversed if caught early. Others can be slowed down and hopefully one day prevented. Unfortunately, people with cognitive decline or very mild dementia often are not recognized until late in the disease course when treatments are less effective. As the first health care professional most people reach out to about medical concerns, primary care providers play a critical role in detecting cognitive decline early. While many primary care providers conduct cognitive screening at Medicare Annual Wellness Visits and when patients voice concerns, 9 out of 10 would like more information about who to screen, which assessment tool to use, and what to say if screening is positive. Deciding who to screen with a brief cognitive assessment tool is a key part of the process because not everyone needs to be screened, and primary care providers already face time pressures to address the obvious and immediate concerns of their patients. The long-term goal of this project is to develop a risk assessment and cognitive screening tool that requires minimal time and effort from primary care providers or their staff and is sensitive to cognitive decline in older adults from diverse educational and racial/ethnic backgrounds. The tool will be integrated into electronic health record systems to make it easy for primary care providers and patients to see results. The specific aims of the first phase of the project are to modify an existing dementia risk screening index to identify older adults who are at high-risk for cognitive impairment, develop a brief cognitive assessment tool using tasks that are easy for older adults to perform yet are sensitive to cognitive decline, confirm their utility in 150 people with varying levels of cognitive abilities that have already been well defined, and test ways to integrate findings into the electronic health record. The specific aims for the second phase are to further test the effectiveness of the newly developed tool in 250 older adults receiving care in a primary care clinic, to find out from primary care providers using the tool how much they liked it and if it was useful and easy to use, and to integrate findings into multiple electronic health record systems. Findings from this project will fill a gap in the existing toolkit of primary care providers and will make screening for cognitive decline quick, easy, and effective.

Inclusion Criteria:

• Cognitively normal must have no cognitive concerns, be independent in all activities of daily living, be medically stable and willing to participate in the study, and fluent in English
• Mild cognitive impairment must have self or informant concern of cognitive decline, impairment in one or more cognitive domains on objective tests when possible, preservation of independence in activities of daily living, be medically stable and willing to participate in the study, and fluent in English
• Mild dementia must exhibit cognitive and/or behavioral symptoms that represent a decline from prior levels of functioning, interfere with the ability to function at work or in usual activities, are not explained by delirium or a major psychiatric disorder, and are evident in at least two cognitive/behavioral domains

Exclusion Criteria:

• Cognitively normal and mild cognitive impairment cannot have confounding medical or psychiatric conditions that could impact cognition, dementia of any etiology, or active substance abuse that could impact cognition or substance use disorder not in remission for at least 1 year.
• Dementia cannot have psychiatric conditions that could impact cognition or active substance abuse that could impact cognition or substance use disorder not in remission for at least 1 year.

Older adults

All research resources (e.g., methods, tools, materials) and other information developed in this project will be made readily available to the scientific community for non-profit research purposes.

As per the proposal guidelines, we will publish all of our code and scripts for statistical analysis and machine learning development on a publically available GitHub repository. We will publish or provide complete recipes for all aspects of data collection and analysis once all references to patient information have been removed. We are committed to sharing our data, models, and software in a practical, user-friendly way; we strongly believe this is required to enable this research to reach its maximum potential. More broadly, the Dell Medical School at the University of Texas at Austin is committed to improving the quality and delivery of care through research discoveries that change clinical practice.

One year after study completion