Evaluating an Artificial Intelligence-Based Diagnostic Support Tool for Older Adults in Primary Care

Advancing Diagnostic Excellence For Older Adults Through Collective Intelligence And Imitation Learning

Older adults commonly experience diagnostic errors that may lead to direct harms and increased healthcare costs. Older adults are especially at risk because of higher rates of comorbidity burden, medical complexity, frailty, and cognitive impairment. An artificial intelligence (AI) clinical decision support system (CDSS) offer a promising approach to promote diagnostic excellence for older adults.

The purpose of this study is to assess the acceptability and feasibility of a new AI CDSS for older adults in primary care. The goal of this AI CDSS is to provide diagnostic support during primary care visits (i.e., help make timely and accurate diagnoses) and support communication amongst patients, doctors, and caregivers about the patient's health.

In this study, participants will use the AI CDSS in a primary care visit and review its suggestions for diagnoses and tests. Afterwards, they will complete a feedback survey and interview where they share their thoughts about and experience using the AI CDSS.

Inclusion Criteria:

Clinicians

1. Work in the University of Pennsylvania Health System
2. Work in a primary care setting (i.e., Internal Medicine, Geriatric Medicine, Family Medicine, and Penn Primary Care)
3. Actively treat adult patients who are 65 years old or older

Patients

1. 65 years of age or older
2. Have an upcoming encounter with a participating primary care clinician
3. Indicate a new or worsening health concern that they wish to discuss at the upcoming encounter

Caregivers

1. 18 years old or older
2. Accompanying the participating patient during the encounter with the AI CDSS
3. Identified by the patient as a caregiver

Exclusion Criteria:

An individual who meets any of the following criteria will be excluded from participation in this study:

1. Under the age of 18 years old
2. Unable to provide informed consent in the opinion of the investigator
3. Has a "Research Do Not Contact" status in the electronic health record