Introduction
Lipoprotein(a) [Lp(a)] is an independent and causal risk factor for atherosclerotic cardiovascular disease (ASCVD) and is implicated in the development of coronary artery disease, peripheral artery disease, and calcific aortic valve stenosis.1–4 Elevated Lp(a) affects approximately 20% of adults, and the prevalence is even higher in people of African ancestry.5 Over 90% of Lp(a) elevation is genetically determined, contributing to the relative stability of Lp(a) levels over a lifetime.6 To improve cardiovascular risk prediction, clinical guidelines and consensus statements increasingly recommend screening for Lp (a) for all adults at least once in a lifetime.6,7
Despite these recommendations, rates of Lp(a) testing remain low. Retrospective analyses estimate testing prevalence at less than 1% of adults.8,9 While the barriers to Lp(a) testing have been described,10 no published quality improvement (QI) efforts have addressed this gap.
A QI team at the University of Pennsylvania aimed to increase Lp(a) testing in a general cardiology fellows’ clinic. Between July 1, 2022, and June 30, 2024, 123 (10.1%) out of 1222 patients had an Lp(a) test order, and 85 (7.0%) had a result. Our goal was to increase the Lp(a) order rate to 20% and the result rate to 15%.
Methods
The interventions took place at an ambulatory cardiovascular clinic, staffed by 35 general cardiology fellows and 21 noninvasive cardiology attendings. The subspecialty preventive cardiology and lipid clinics were excluded.
The electronic health record (EHR) was used to query patients seen in the cardiology fellows’ clinic between July 1, 2022, and May 4, 2025. The primary outcome was the proportion of patients with an Lp(a) order, defined as follows: number of unique patients seen in the fellows’ clinic with an Lp(a) ever ordered / number of unique patients seen in the fellows’ clinic during the same time period. The secondary outcome was the proportion of patients with an Lp(a) result.
Lp(a) orders and results before the study period were included in the primary and secondary outcomes. These were measured bimonthly and reported as cumulative percentages.
Patient demographics, clinical history, and lab data were collected via EHR queries and chart review. Lp(a) risk categories were defined as low (< 30 mg/dL or < 75 nmol/L), elevated (30–49 mg/dL or 75–124 nmol/L), and very high (≥50 mg/dL or ≥125 nmol/L).6 Mass units (mg/dL) were converted to molar units (nmol/L) by using a factor of 2.15. For patients who completed testing at the University of Pennsylvania, the Lp(a) lab sample was sent to ARUP Laboratories (Salt Lake City, UT), which utilizes quantitative immunoturbidimetry.
Some patients had their labs performed at Quest Diagnostics, Labcorp, or at other local clinical lab sites. ASCVD was defined as the presence of coronary artery disease, peripheral artery disease, or ischemic stroke using International Classification of Diseases, 10th Revision (ICD-10) codes and verified for accuracy through randomly selecting and manually reviewing patient charts.
Statistical process control tools were created to assess trends in Lp(a) orders and results.11 Analyses were conducted using Microsoft Excel and Stata v18.
The QI team, led by a cardiology fellow and a lipid specialist, performed a Root-Cause analysis based on clinical experience and a prior institutional survey10 (Supplemental Fig S1). Interventions were implemented through iterative Plan-Do-Study-Act (PDSA) Cycles, with emphasis on education and EHR tools (Table 1).
This project met the criteria for QI by the University of Pennsylvania Institutional Review Board (IRB) and was exempt from further review.
Results
From July 1, 2024, to May 4, 2025, 156 (20.9%) out of the 748 unique patients seen in the fellows’ clinics had an Lp(a) order, and 84 (11.2%) had a result. Lp(a) testing completion rate was 53.9%. The mean age of patients with an Lp(a) order was 53.8 years ( ± 14.6); 51% were male, and 42% identified as Black (Table 2).
Among those tested, mean Lp(a) was 123.1 nmol/L ( ± 116.3), and 58% had levels above the elevated threshold (≥75–124 nmol/L). Very high Lp(a) levels (≥125 nmol/L) were more frequent in Black (65%) vs White patients (24%) (Fig 1A). ASCVD was present in 96 (62%) patients with Lp (a) orders.
Among those with ASCVD and an Lp(a) result, 43% had very high Lp(a). Of the patients with ASCVD and very high Lp(a), 65% were on statins, 26% on high-intensity statins, and < 1% on proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors (Fig 1B).
A control chart showed a sustained increase in Lp(a) orders starting December 2024, with a clear upward trend from October 2024 to January 2025 (Fig 2). A run chart revealed that result rates steadily increased beyond baseline (7%) but fell short of the 15% goal (Supplementary Fig S2). Rates of Lp(a) orders and results were cumulative, beginning July 1, 2024.
The Lp(a) test completion rate decreased from 69.1% to 53.9%.
Discussion
There is a growing consensus among experts in preventive cardiology and lipidology that all adults should have an Lp(a) level tested at least once per lifetime. Lp(a) testing has emerged as part of an armamentarium of tools to assess cardiovascular risk; elevated Lp(a) may prompt clinicians to intensify lipid-lowering therapy for both primary and secondary ASCVD prevention. During this QI initiative, the rates of Lp(a) testing among all patients seen in a cardiology clinic doubled from 10.1% to 20.9%, and the rates of Lp(a) results increased from 7.0% to 11.2%.
The success of this effort was driven by provider education and EHR-based tools. With targeted therapies for Lp(a) on the horizon, it is a critical time to identify patients who may benefit from Lp(a) lowering.15 This QI intervention may serve as a model for other practices interested in increasing Lp(a) testing.
Prior analyses at academic centers found Lp(a) testing rates under 1%, increasing to just under 4% among patients with cardiovascular disease.9,16 Our baseline Lp(a) testing rate was relatively high at 7.0%, which may be attributable to the narrow study population of patients seen by cardiology trainees, who are typically quick to adapt evidence-based interventions. Other institutions have reported lower rates of testing among non-Hispanic Blacks and patients with lower socioeconomic status,16 but testing rates were similar across race and gender in our study. This is of particular importance because patients who identify as Black are more likely to have elevated Lp(a) levels, a finding also seen in our study.5 Inclusive implementation of Lp(a) testing is critical across race and genders, both to ensure health equity and to study ASCVD risk in diverse populations.
Several key barriers to Lp(a) testing emerged during our QI study (Supplementary Fig 1). Lack of familiarity with Lp(a) was common, particularly around indications for testing and clinical application of results. We addressed this through creating and disseminating dot phrases, toolkits, and guideline-based education.
Since Lp(a)-targeted therapies were not yet available, we focused on evidence supporting aggressive LDL-C lowering for individuals with high Lp(a) and preventive cardiology referral for the highest risk patients.17,18 Testing Lp(a) at new patient visits, or when patients are due for standard lipid testing, may be ideal time points for Lp(a) screening. We found that two-thirds of patients with clinical ASCVD are on statin therapy, but utilization of high-intensity statin and non-statin therapies was higher in patients with elevated Lp(a) levels. While there remain opportunities to increase lipid-lowering therapy for all patients with ASCVD, some providers may already be using Lp(a) to guide LDL-cholesterol treatment and ASCVD risk reduction strategies.
The addition of Lp(a) measurement may therefore enhance current risk stratification practices and prompt therapy intensification for both primary and secondary ASCVD prevention. Future efforts should focus on optimizing available therapies, once patients at the highest risk are identified.
Cost was another major concern in the root cause analysis. A prior institutional survey found that over one-third of respondents avoided ordering Lp(a) due to insurance and billing concerns.10 This hesitation was confirmed in our surveys and focus groups. To address this, we provided estimated test costs at local labs, shared billing guidance, and created patient-facing infographics to improve shared decision-making.
Finally, the lack of visibility of Lp(a) in the EHR made it time-consuming for providers to identify patients who had never been tested. The addition of Lp(a) dot phrases and the incorporation of Lp(a) lab results into clinic notes were simple ways to increase visibility of Lp(a) during a clinical encounter. Future interventions may test automatic nudges and opt-out Lp(a) ordering, which are known to have a greater impact on quality and patient outcomes.19
There are several important limitations of this study. First, we used the EHR to screen patients who had ever had an Lp (a) order in our health system. Lp(a) orders and results that occurred before the QI intervention were therefore included in our numerator but are not attributed to our intervention.
Second, our QI study was designed to track rates of Lp(a) testing in a single clinic without comparing rates to other similar clinics. We therefore cannot conclusively attribute the increase in Lp(a) testing to the QI interventions. External factors—such as the development of Lp(a)-targeted therapies and increased provider awareness from outside sources—may have naturally contributed to increased testing.
Third, Lp(a) orders outside of our health system were not accounted for, which could impact both process and outcome measurements. Finally, we did not reach our target of 15% Lp(a) results, and we detected a decreased rate of Lp(a) test completion. Potential reasons for test incompletion may include cost, access to labs, testing performed outside of our health system, and unanswered questions about the test’s utility.
While out of scope for this current project, future work will assess barriers from test order to completion.
Conclusion
This QI initiative successfully increased rates of Lp(a) testing in a cardiology clinic. Though limited in scope, this effort highlights how targeted education and workflow integration can address clinical gaps in cardiovascular risk prevention. Broader implementation of these strategies may improve ASCVD risk assessment and prepare systems for Lp(a)-targeted therapies.
Supplementary Material
Supplementary material associated with this article can be found in the online version at doi:10.1016/j.jacl.2026.01.018.