SUPER-LIPID: Non-visit based automated orders sent to central pharmacy increased appropriate statin prescriptions

By Lucas Marinacci on

Key Points

  • In two simultaneous pragmatic trials, an asynchronous, non-visit based, automated order for statins placed to a centralized pharmacy significantly increased both overall statin and appropriate dose statin prescriptions, while an interruptive visit based EHR notification strategy had an overall smaller effect.

Despite their effectiveness at reducing cardiovascular risk, the majority of patients who have indications for a statin are undertreated.  Pharmacists may be able to fill the gaps in evidenced-based statin prescription.  On November 12, 2023 the results of the “SUPER LIPID Program: Two Randomized Controlled Trials of Nudges to Encourage Referrals to Centralized Pharmacy Services for Evidence-Based Statin Initiation in High-Risk Patients” were presented at AHA Scientific Sessions 2023.   The purpose of this project was to test two different strategies for integrating centralized pharmacy services into statin prescription.


This project described two parallel clinical trials.  The first was a stepped wedge design that tested visit-based pop-op notification to refer eligible patients to the centralized pharmacy for statin initiation versus usual care.  The second was a cluster randomized trial of non-visit based, automatically placed orders for referral to centralized pharmacy for statin initiation versus usual care. Patients who were assigned a primary care providers at participating clinics in Pennsylvania and who were not prescribed a moderate- or high-intensity statin despite an indication were eligible to participate and were identified via electronic health record (EHR)-based algorithms.  Patients with documented statin allergy or intolerance, advanced chronic kidney disease, or prescribed a PCSK9-inhibitor were excluded.  Once a pharmacists received a referral, the pharmacist reviewed the chart, confirmed that it was appropriate for a statin prescriptions, and ensured the patient did not previously decline a statin when offered in the past two years.  They then called the patient and after shared decision making if patient agreed, prescribed the statin as well as any indicated follow up labs.  


For the first stepped-wedge trial of the visit-based intervention, 16 primary care providers were randomly assigned to two groups.  Both groups of providers performed usual care for 3 months.  Then for months 4-6, only group 1 got the pop-up during patient encounter for pharmacist referral. Finally for months 7-9 both groups received the intervention.   There were 970 patients exposed to group 1 and 672 patients exposed to group 2; there were significantly more woman in the second group.  The odds of any statin prescription were significantly higher for those exposed to the intervention (OR 1.43, 95% CI 1.02, 2.00), but there was no significant difference of statin prescription at appropriate dose.   Of note, this trial did not reach its enrollment goal.


For the cluster RCT of the non-visit-based intervention, 10 primary care practices were randomly assigned to usual care or the intervention.  The primary outcome at 6 months was percentage of eligible participants prescribed a statin.  Both arms had 975 patients, with an average age of ~64.  There was some imbalance based on age, gender, race, baseline LDL, smoking, and comorbid hypertension.  Significantly more patients in the intervention arm were prescribed a statin compared to the usual care arm (31.6% vs 15.2%, OR 2.22, 95% CI 1.47-3.33).  Unlike for the stepped-wedge trial, significantly more patients were prescribed an appropriate dose statin in the intervention arm (24.8% vs. 7.7%, OR 6.79, 95% CI 4.00, 11.53). Both trials had limitations related to randomization at the physician or practice level, namely there was some baseline imbalance between the groups.  These trials were separate and therefore there was no direct comparison between the visit and non-visitbased strategies (although the including criteria, pharmacy protocol, and endpoints were the same).  


Dr Alexander Fanaroff of the University of Pennsylvania concluded: “A centralized asynchronous model could be an effective adjunct to visit-based clinical interactions in increasing statin prescribing for high-risk patients.”