Active vs passive best-practice alerts to improve guideline-directed lipid-lowering therapy intensification in patients with recent myocardial infarction and elevated LDL-C

Background Patients with a recent myocardial infarction (MI) often do not receive lipid-lowering therapy (LLT) according to guideline recommendations, which leave them at very-high risk of another cardiac event. Purpose We evaluated the impact of changing an electronic health record (EHR) best practice from a passive to an active alert for patients with a recent MI who were not receiving guideline-recommended LLT according to the 2018 AHA/ACC/Multisociety cholesterol guideline. Methods From 2/01/18, our practice group (96 cardiologists, 35 locations) implemented an EHR passive alert that would trigger for patients with a recent MI (<1 year from index visit ) if the LDL-C ≥70mg/dL or there was no LDL-C value in the EHR within 6 months of the index visit. Physicians could dismiss the alert without action. From 8/01/20 to 02/01/22 the EHR alert was changed to an active "hard-stop" alert. Both alerts recommended to either intensify LLT or order an LDL-C test; however, physicians could not dismiss the active alert without action. Patients were followed for up to 1 year. Results Data for a total of 1,320 patients were assessed (passive alert cohort, n=733; active alert cohort, n=587). Baseline characteristics were similar between groups, with mostly males (66% and 61%, respectively) with median ages of 67 (IQR 58-76) and 65 (IQR 57-76) years. During the passive alert period, statin use increased from 59% pre-index to a max of 86% throughout follow-up (Fig 1A). High-intensity statins increased from 39% to 67%. Few passive alert patients received ezetimibe or PCSK9i pre-index or during follow-up. For the period when the active alert was implemented (Fig 1B), statin and statin high-intensity use were already high pre-index (79% and 70% of patients, respectively), with little change during follow-up. In contrast to the passive alert period, the patterns for ezetimibe and PCSK9i improved during follow-up with the active alert; ezetimibe use increased from 12% to a max of 31% of patients and PCSK9i use from 2% to a max of 6% (adjusted odds ratios: 4.6 and 3.8, respectively ). Post-index LDL-C follow-up was less common in the passive group (50% at 6 months, 72% at 12 months versus 60% and 80%, respectively, in the active group (P<0.01). Goal attainment of LDL-C of <70 mg/dL was 39% at 6 months and 55% at 12 months in the passive group (40% and 54%, respectively, in the active group, P=0.6). Conclusions The passive alert improved LLT intensification by primarily increasing the use of statin and high-intensity statin therapy. Changing to active alert improved LDL-C monitoring and led to further improvement by increasing prescribing of non-statin therapies in a population of high-risk patients already on maximally tolerated statin therapy who were not at their LDL-C goal. Despite these improvements, continued efforts are needed to encourage guideline-directed LLT intensification in patients with a recent MI and are at risk of another cardiac event.