Heart valve disease and weakened heart muscle (left ventricular systolic dysfunction) are common in older adults and often go undetected until serious complications such as heart failure develop. Detection currently depends on a clinician hearing a murmur and then ordering an echocardiogram, which is easily missed or delayed.
This study tests whether a brief, artificial intelligence (AI)-guided handheld heart ultrasound - point-of-care ultrasound, or POCUS - performed by trained clinic staff during a routine visit identifies these conditions earlier than usual care.
Primary care and geriatrics providers, rather than individual patients, are assigned by chance to one of two groups. Patients seen by providers in the AI-ultrasound group are offered a POCUS scan and a one-time blood test at their regular visit, and are referred for a confirmatory echocardiogram if the scan is abnormal. Patients seen by providers in the usual care group receive standard clinic care. Researchers will compare how often previously undiagnosed structural heart disease is newly identified in each group.
Physician or advanced practice provider (physician assistant or nurse practitioner) practicing in primary care or geriatrics at a participating institution who independently assesses patients
Patient panel comprising >50% adults aged 65 years or older
Sees patients in ambulatory clinic at least 1 day per week on average
Agrees to participate, to undergo randomization, and to permit inclusion of all eligible patients from the panel
Serving in a locum tenens or other temporary capacity
Unwilling or unable to nominate an AI-POCUS champion from the clinic
Echocardiogram performed within the past 5 years
At least moderate aortic or mitral valve disease, or left ventricular ejection fraction 50% or less, documented on any prior echocardiogram