This study is designed to evaluate the efficacy of automated electronic alerts in the electronic health record to improve rates of best practices in the treatment of patients with hyperlipidemia who present in the setting of outpatient primary care and family medicine practices within the Yale New Haven Health System.

Cardiovascular disease (CVD) remains the leading cause of death in most developed countries, including the United States. A key risk factor for CVD is an elevation in low-density lipoprotein cholesterol, or LDL-C. Based on favorable results from large clinical trials, guidelines support the use of high intensity statin therapy for the reduction of LDL-C in very high risk patients and addition of non-statin therapies including ezetimibe and PCSK9 inhibitors if the LDL-C remains above 70 mg/dL. However, implementation of these guidelines is poor and there remains substantial underuse of evidence-based lipid lowering therapies across healthcare systems.

Real-time alerting to important clinical conditions, when fired in an appropriate and timely manner, can have positive impacts on patient outcomes through increased physician awareness and adherence to best practices. To this end, this study has been designed to evaluate whether automated electronic alerts built into the Epic electronic health record can improve the management of hyperlipidemia among patients of outpatient internal medicine and cardiology practices within the Yale New Haven Health System who are at very high risk for future atherosclerotic CVD (ASCVD) events.

One hundred physicians will undergo cluster randomization to either the alert group or the control (usual care) group. Upon opening the order entry screen of an eligible patient with hyperlipidemia, the alert group will receive an informational alert that informs the provider that the patient has hyperlipidemia and is at very high risk for future ASCVD events and provides ACC/AHA-guideline-directed actionable items, including a link to a hyperlipidemia "order set" to include both diagnostic and therapeutic options.

The primary outcome will be the proportion of patients who have intensification of their lipid lowering therapy (increase in statin dose or addition of ezetimibe or addition of a PCSK9 inhibitor) at 90-days. The secondary outcomes will be achieved LDL-C at 6-months and proportion of patients with LDL-C levels less than 70 mg/dL and less than 55 mg/dL. In an optional extension phase, long-term exploratory outcomes looking at the rate of major cardiovascular cardiac events (MACE), defined as hospitalization for myocardial infarction, stroke, unstable angina, or coronary or peripheral artery revascularization will be evaluated at 2 years.

Inclusion Criteria for Patient Subjects:

• Adults 18 years of age or greater
• Currently being seen as an outpatient of Internal Medicine or Cardiology within the Yale New Haven Health System
• At very high risk for ASCVD, defined by a history of a major ASCVD event, and most recent lipid profile with LDL-C greater than 70 mg/dL, with or without other high-risk features (diabetes, CKD, age >65 years)

Exclusion Criteria for Patient Subjects:

• Heart transplant recipient
• Left ventricular assist device recipient
• Hospital inpatient status
• Pregnancy
• Have opted out of clinical research in MyChart

Inclusion Criteria for Provider Subjects:

• Practicing at an outpatient cardiology or internal medicine practice with the Yale New Haven Health System
• High frequency of eligible patients seen, based on retrospective review of outpatient records

Deidentified data underlying results for publication will be shared upon publication.

After publication of results; indefinitely

There were 1130 and 1370 patient subjects enrolled in the alert vs. usual care group, respectively. There were 49 and 47 providers enrolled in the alert vs. usual care group, respectively. The following baseline characteristics are described only for the patient subjects as primary and secondary outcomes were collected only on patient subjects in this study.

6 months

Adverse events are reported to the IRB via written report within 5 calendar days of the Principal Investigator becoming aware of the event.

Adverse events reporting is only described for patient subjects as the exposure of interest would be expected to affect this study population. Adverse events were not monitored or assessed for provider subjects.

All study sites were in a single academic health system, and the alert was developed within this site's EHR, reducing generalizability. The study relied on accurately documented ASCVD ICD-10 codes. Contamination between providers at shared sites was possible, biasing the outcome toward the null. Finally, many providers chose to defer the alert, suggesting responsibility of LLT intensification belonged to primary care providers or cardiologists.