Atherosclerotic Cardiovascular Disease (ASCVD), Coronary Artery Calcification
Conditions
Brief summary
This multi-site study will test whether an opportunistic AI-based CAC screening and notification intervention can improve cholesterol treatment and lower cholesterol levels in adults. The study uses artificial intelligence to detect calcium buildup in heart arteries (coronary artery calcium or CAC) on chest CT scans that patients have already had for other reasons. The study will focus on adults who either have known atherosclerotic cardiovascular disease (ASCVD) or have significant calcium buildup (a CAC score of 100 or higher), and whose cholesterol is not well controlled. It will also evaluate how well this approach can be implemented at scale across multiple health systems. The main questions it aims to answer are: Does notifying patients and their clinicians about incidental CAC increase lipid-lowering therapy(LLT) initiation or intensification? Does the intervention improve Low-Density Lipoprotein(LDL)-cholesterol control and related lipid testing? How does the intervention affect downstream care (e.g., clinic visits, cardiology referrals, and cardiac testing)? Researchers will use an FDA-cleared AI algorithm to quantify CAC on previously performed non-gated chest CT scans and identify eligible participants through the electronic health record. Participants will be randomized to receive CAC notification either right away or after a 6-month delay.
Interventions
OTHERAI-Detected CAC Notification and Care Facilitation
After randomization to the early notification arm, the study team will send a standardized notification message to the participant's affiliated clinician. The study team will send a message to the participant after a brief delay. Each site will determine the timing between the initial message to the clinician and the participant based on stakeholder feedback. The notification is about the AI-CAC identified on the participant's previous chest CT. It will provide an overview of AI-CAC, a personalized image of AI-CAC, and a recommended risk discussion with their clinician. These clinicians will also be notified of the findings. For participants randomized to early notification who do not undergo LLT initiation, intensification, or LDL testing within 2 months, the clinician and participant will receive a second message at that time. The participants in the delayed notification arm will receive a similar notification 6 months later.
Sponsors
Stanford University
Amgen
Baylor Scott and White Health
Kaiser Permanente
Vanderbilt University Medical Center
Duke University
University of California, Los Angeles
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
HEALTH_SERVICES_RESEARCH
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
18 Years to 74 Years
Healthy volunteers
No
Inclusion criteria
* Non-gated chest CT performed within the prior 2 years within the health system * Active health system engagement with an affiliated clinician eligible for notification, defined as ≥1 clinical visit within the prior 2 years AND at least one of the following: 1. Active enrollment in the site's integrated healthcare plan (e.g., Kaiser Permanente Northern California), OR 2. LDL-C measured within the health system within the prior 2 years, OR 3. Outpatient cardiovascular medication prescription within the prior year * Meets one of the following clinical criteria: 1. Clinical ASCVD diagnosis (coronary artery disease, peripheral arterial disease, or ischemic cerebrovascular disease) AND AI-detected CAC \>0 on non-gated chest CT, OR 2. No ASCVD diagnosis AND AI-detected CAC ≥100 * Suboptimal LDL-C control, defined as either: Last LDL-C ≥70 mg/dL in the last 2 years, OR No LDL-C measurement in the last 2 years Note: Site-level variations and additional refinements may occur based on local stakeholder input and patient population identification.
Exclusion criteria
* Dementia * Heart transplant * Active hospice * Other life-limiting illness as determined by the site investigator (e.g., metastatic cancer) * Additional participant-level exclusions as determined by clinical stakeholders at each site
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Lipid-lowering therapy initiation or intensification (LLTI) | 6 months after participant randomization | * Number of new statin prescriptions * Number of statin doses/intensity increase * Number of PSCK9 inhibitor initiation (monoclonal antibodies and siRNA) * Number of Ezetimibe initiations * Number of Bempedoic acid initiations |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Aspirin prescription | 6 months after participant randomization | Number of new Aspirin or other antiplatelet prescription |
| Lipid panel (total cholesterol, HDL-C, LDL-C, triglycerides) ordering | 6 months after participant randomization | Number of Lipid panel ordering |
| Lipoprotein(a) test ordering | 6 months after participant randomization | Number of Lipoprotein(a) test ordering |
| Changes of total cholesterol, HDL-C, LDL-C, triglycerides, Lipoprotein(a) levels | 6 months after participant randomization | This outcome measure the changes of the test levels |
| Healthcare resource use | Baseline through Month 6 | * Number of primary care clinical encounters * Number of cardiology encounters * Number of coronary artery disease-related imaging tests (e.g. noninvasive stress tests, coronary computed tomography angiography (CCTAs), and angiograms) |
Countries
United States
Contacts
PRINCIPAL_INVESTIGATORFatima Rodriguez, MD, MPH
Stanford University
Outcome results
None listed