Improved Adherence to Lipid Screening in the Pediatric Cardiology Clinic: A Quality Improvement Project

Abstract

Introduction:

Lipid screening identifies at-risk patients to facilitate cardiovascular risk reduction. National pediatric guidelines recommend universal lipid screening between 9–11 and 17–21 years of age. We aimed to improve adherence to lipid screening for all age-appropriate outpatient pediatric cardiology visits from a baseline of 35% to 90% between November 2021 and July 2023.

Methods:

All outpatient visits for patients 9–11 and 17–21 years were included. Chart review and an Epic electronic health record report identified patients screened, lipid test results, and need for further testing. A P-chart was generated. After establishing a baseline for 8 weeks, interventions, including an Epic dot phrase, group and individual feedback, and Epic best practice alert (BPA), were incorporated via plan-do-study-act cycles. Balancing measures included anonymous provider surveys on visit length and experience.

Results:

More than 1,700 patient visits were included. At baseline, 35% of all age-appropriate patients were screened. The Epic dot phrase prompted a positive shift with a new mean of 59% screened. Another change occurred after the BPA alert, with an increase in screening to 84%. Lipid screening prompted by this initiative found that 38% of those with testing results in Epic had abnormal results requiring follow-up. Providers did not report a significant change in visit length due to screening.

Conclusions:

Quality improvement interventions improved adherence to universal lipid screening guidelines. The Epic dot phrase and BPA facilitated positive shifts. These simple interventions can be spread to other practices to improve adherence to lipid screening and other guidelines.

INTRODUCTION

Problem Description, Available Knowledge, and Rationale

Atherosclerotic cardiovascular disease (ASCVD) is the leading cause of mortality both in the United States and globally.¹ Although ASCVD rarely manifests before adulthood, atherosclerosis can begin in early childhood, especially in those with risk factors for disease, including hypertension, familial hyperlipidemia, and obesity.² Early histological signs of coronary artery disease have been demonstrated in over half of the right coronary arteries evaluated in teenagers as young as 15 years of age.³ Lipid screening can identify at-risk pediatric and adult patients to facilitate cardiovascular risk reduction. These at-risk individuals require extensive counseling on diet and lifestyle modifications, serial monitoring, and medication management. The linear relationship between lipid levels and rates of cardiovascular events has been established. Early intervention to manage hypercholesterolemia decreases the risk of future cardiovascular complications.⁴

Despite the known risks of early coronary artery disease in youth, most teenagers were not screened for hyperlipidemia even in the early 21st century. A study including multiple health systems reported screening rates as low as 8.9% in 9- to 11-year-old and 24.3% in 17- to 19-year-old patients.⁵ In 2011, the National Heart, Lung, and Blood Institute, with the endorsement of the American Academy of Pediatrics, established recommendations for universal lipid screening between 9–11 and 17–21 years, age ranges at which lipid levels have been shown to peak during childhood.^6–9 These guidelines also recommend lipid screening between 2 and 8 years of age in patients with a family history of early cardiovascular disease, parents with dyslipidemia or other risk factors for poor cardiac health, or high-risk conditions such as Kawasaki disease or cardiac transplant.⁷ Although this has been a national recommendation for over a decade, studies in the years following the publication of these guidelines have demonstrated persistently low lipid screening rates ranging from 3.5% to 27% in patients 9–11 years of age seen in pediatric clinics.^8,10,11 Although general practitioners typically perform lipid screening, we sought to address the gap in adherence to universal screening guidelines in our subspecialty practice because a higher proportion of patients with ASCVD risk factors, such as hypertension and obesity, are referred to pediatric cardiology clinic.

Specific Aims

We aimed to improve adherence to lipid screening for all age-appropriate outpatient visits within our pediatric cardiology clinic from a baseline of 35% to 90% during 20 months from November 2021 to July 2023.

METHODS

Context

Atlantic Health System is a large community hospital network that cares for pediatric and adult patients. The pediatric cardiology division consists of 7 physicians who attend 4 outpatient clinic sites across northern and central New Jersey. Patients are referred for various preventative/general cardiology concerns, structural cardiac issues, and cardiomyopathies. The clinic does not include patients who have undergone heart transplantation. Clinical documentation is in a single electronic health record, Epic (Epic Systems Corporation, Verona, Wis.).

Project Development and Measures

To ensure that all age-appropriate visits were included, the Atlantic Health System’s Epic information technology team flagged all new and established patients seen in our clinics within the targeted age ranges. In the year before the study started, 720 patients within the age ranges were seen, with an average of 20 age-appropriate patients seen per week. Based on this, we determined that 8 weeks provided an appropriate timeframe to assess baseline adherence to lipid screening. The 7 pediatric cardiologists in the division comprised the quality improvement (QI) team. A key driver diagram was developed based on physician feedback (Fig. 1).

Fig. 1.

Key driver diagram demonstrating SMART (specific, measured, achievable, relevant, and time) aim, key drivers, and interventions. EMR, electronic medical record.

Our primary measure was the percent adherence to lipid screening for patients 9–11 and 17–21 years of age. Lipid screening determined whether the patient had undergone bloodwork to check a lipid panel, typically including total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), triglycerides, and very LDL. Samples were venous blood draws that could be performed fasting or nonfasting at the physician’s discretion. Point-of-care lipid testing was not available for the clinic. Abnormal lipid values were defined as per the age-specific values established by the American Academy of Pediatrics and National Heart, Lung, and Blood Institute with total cholesterol ≥200 mg/dL, LDL ≥130 mg/dL, non-HDL cholesterol ≥145 mg/dL, HDL ≤40 md/dL, triglycerides ≥100 mg/dL for patients 9 years of age and ≥130 mg/dL for patients older than 9 years of age.⁷

Chart review and an Epic report identified whether patients were screened, test results, and the need for further testing. Each patient screened had four possible results: normal, abnormal, future testing by a primary doctor recommended, or lipid panel ordered on the day of visit. The patient was considered to have had lipid screening performed if the physician documented this discussion in their note. The goal was to have 90% of all age-appropriate patients screened. For patients with multiple visits during the study period, only the most recent visit was included in the analysis to ensure that each patient was only accounted for once. Balancing measures were then established by anonymous provider surveys to determine whether lipid screening impacted visit length or their visit experience rated on a 5-point Likert scale. The anonymous survey was administered to all physicians involved in the QI project.

Improvement Activities

Based on physician feedback, we incorporated our SMART (specific, measured, achievable, relevant, and time) aim with plan-do-study-act cycles. Our plan-do-study-act cycles included physician education regarding lipid screening guidelines, creating a dot phrase for implementation into the note, group feedback, individual feedback, and the creation of an Epic best practice alert (BPA). Interventions were selected based on actionable activities that were easily integrated into the clinical workflow with the assistance of the information technology team.

Lipid Screening Dot Phrase

The first intervention was creating a lipid screening dot phrase. A dot phrase is a statement that can be inserted into the electronic health record documentation by multiple practitioners. In conjunction with the Epic team, a dot phrase was created for all pediatric cardiologists to include in their outpatient notes. Previously, some physicians used a personal dot phrase to document lipid screening, but these were replaced with a more concise phrase that autopopulated the results of the patient’s last lipid panel performed within Atlantic Health System. The dot phrase included an Epic Smart List from which the physicians could select the result of the lipid screening (normal, abnormal, future testing recommended, or lipid panel ordered). Physicians were encouraged to incorporate this dot phrase into their note templates. Physician adherence to the use of the lipid screening dot phrase was a process measure tracked with the use of an Epic-generated report.

Reviewing QI Project

The next intervention was reviewing the QI project aims and age ranges for patient screening with all the pediatric cardiologists. This was performed at one of the weekly physician meetings, in which all physicians were in attendance. Individual discussions with the physicians following the meeting ensured physicians’ understanding of the project aims and screening guidelines.

Group and Individual Feedback

P-charts were reviewed with all physicians at the weekly physician meeting. Individual feedback was provided via a funnel chart demonstrating each of the pediatric cardiologists’ progress throughout the study. The cardiologists were each represented by randomly assigned alphabetical letters to maintain anonymity.

Best Practice Alert

A final intervention was to create a BPA within Epic. In conjunction with the Epic team, a banner alert was created that appeared whenever a chart was opened for an encounter for a pediatric cardiology visit for a patient within the targeted age range. The BPA would indicate that the patient met the age criteria for lipid screening and would list the most recent lipid panel results if available in Epic. The alert continued to seem until acknowledged by the provider.

STUDYING THE INTERVENTIONS

Analysis

Statistical process control charts were used to analyze the data. P-charts were generated for the percentage of patients screened, with each data point consisting of 2 weeks of patient visits to capture multiple visits for each provider. There was a P-chart for all patients and separate charts for 9–11 and 17–21 years old. Special cause variation was defined as an external change to the system with 8 or more consecutive points above or below the centerline.¹²

Ethical Considerations

Per institutional policy, QI projects do not constitute human subjects research, and institutional review board approval and informed consent documentation were not required. Pediatric cardiologists participated in the project voluntarily. Standards for Quality Improvement Reporting Excellence 2.0 guidelines were used to report the findings.

RESULTS

More than 1,700 patients of 9–11 and 17–21 years of age were seen for 20 months between November 2021 and July 2023. Fifty-six percent of these were established patients seen for a follow-up visit. At baseline, 35% of all age-appropriate patients were screened without a significant difference between the younger and older age groups. Following the implementation of the lipid screening dot phrase, a positive shift occurred with 8 data points consecutively above the baseline. The lipid screening dot phrase was used for 89% of the visits in 3 months following the initiation of the dot phrase. This prompted a shift in the mean lipid screening adherence to 59% by June 2022. Lipid screening dot phrase use remained at 90% at the end of the study period.

Individual and group feedback was incorporated without significant improvement. Implementing the BPA prompted a subsequent positive shift in screening to 84% (Fig. 2). For the 9–11 age group, the baseline was 35% of all patients screened, with a positive shift occurring to a mean of 61% after the lipid screening dot phrase was introduced. The BPA alert prompted another shift to a mean adherence of 83% (Fig. 3). For the 17–21 age group, the baseline was 33% of patients screened, with a positive shift occurring to 59% after introducing the lipid screening dot phrase. Another positive shift to 85% occurred after introducing the BPA alert (Fig. 4). Overall lipid screening adherence was comparable between the age groups (P = 0.6) (Figs. 3, 4). Funnel charts were used to assess each pediatric cardiologist’s performance. The funnel chart demonstrated that most pediatric cardiologists improved the percentage of patients screened following individual and group feedback and that all 7 demonstrated further improvement after implementing the BPA (Fig. 5).

Fig. 2.

P-chart demonstrating lipid screening adherence rates after QI interventions.

Fig. 3.

P-chart demonstrating lipid screening adherence rates over time in 9- to 11-year-olds.

Fig. 4.

P-chart demonstrating lipid screening adherence rates over time in 17- to 21-year-olds.

Fig. 5.

Lipid screening adherence rates by individual physicians.

Of the 1,035 patients who underwent lipid screening during this study, 55% (564) did not previously have a lipid panel obtained at the time of their visit. Among these patients, 43% (241) had a lipid panel ordered during their visit, of which 40% (105) performed the lipid testing, with results available in Epic. Thirty-eight percent (43) of these patients with available results had abnormal test results that required follow-up (Table 1). Of the 43 patients with abnormal lipid screening, the most common abnormality was isolated hypertriglyceridemia. The second most common abnormality was a combination of high total cholesterol, triglycerides, and LDL levels (Table 1).

Table 1.

Breakdown of Abnormal Lipid Results in 40 Patients

Abnormal Results	No. Patients
Triglycerides	9
Triglycerides, total cholesterol, and LDL	8
Triglycerides and total cholesterol	5
Total cholesterol	4
Total cholesterol and LDL	3
Total cholesterol, triglycerides, LDL, and low HDL	2
LDL and low HDL	2
Low HDL	2
Triglycerides and low HDL	1
Triglycerides and vLDL	1
Total cholesterol, triglycerides, and vLDL	1
Total cholesterol, LDL, and low HDL	1
LDL and low HDL	1

vLDL, very low-density lipoprotein.

As a balancing measure, providers were given anonymous surveys to determine whether lipid screening impacted visit length or their visit experience. All providers reported that documenting lipid screening never or rarely significantly increased the duration of their patient encounters. The majority felt that lipid screening improved their visit experience most of the time (72%), sometimes (14%), or rarely (14%).

DISCUSSION

Summary and Interpretation

Our QI project has demonstrated improvement in the number of patients screened for lipid disorders, with the goal of nearly 90% of patients being screened. Prior QI studies have focused on improving lipid screening in 9- to 11-year-old patients, primarily in general pediatric practices.^13,14 One of these studies assessed lipid screening and performance of lipid testing for 2 years in a general pediatric academic hospital-based clinic. The interventions included provider education, chart review feedback, electronic health record cues prompting orders, and in-office phlebotomy. Despite the variety of interventions, this study was only able to improve lipid screening adherence from 32% to 44%.¹⁴ Pradhan et al¹⁵ described a lipid screening quality initiative conducted in a pediatric cardiology clinic using electronic reminders and informational sessions, which prompted rapid improvement in lipid screening over a few months, but sustained improvement was not reported. Our study is unique because it uses the electronic health record to autopopulate lipid testing results and generates alerts prompting physician screening. These higher reliability interventions allowed for more sustained improvement in screening. Recent lipid screening adherence has remained at an average of 84% for 9 months, including 6 months following project completion. Our study demonstrated a higher rate of abnormal lipid panel results with 38% of those who obtained lipid testing as a result of the screening compared with a rate of 20% reported previously.¹³ This higher rate could be related to more cardiovascular risk factors in patients referred to pediatric cardiology practice.

This study was performed in community-based hospital subspecialty clinics, including satellite clinics. Using Epic to generate the lipid screening report minimized the need for manual chart review. These technical improvements are generalizable to other electronic health systems with the support of an institutions’ information technology staff to generate automated reporting and advisories or alerts available to providers. Tracking individual physicians’ progress may also help, particularly in larger practices. These simple interventions can be replicated by other practices to improve adherence to lipid screening. The next phase of the lipid screening intervention will be to spread these improvement techniques, including BPA alert and lipid screening dot phrase to the family medicine and pediatric practices within the hospital network.

Study limitations include the inability to review or autopopulate lipid testing performed outside of our hospital system and variability in physician documentation of the screening performed. A fair number of patients who had lipid testing performed at our institution had abnormal results flagged by Epic. This may be partly due to the low percentage of patients who completed their laboratory tests. Only 40% of patients with lipid testing ordered during their visit underwent laboratory draws during the study timeframe. The limited sample size may introduce bias into our results and partially explain the high percentage of abnormal results.

Additionally, lipid samples were a mix of fasting and nonfasting tests and could have impacted the rates of abnormal lipid results. Although many of these patients do not require lipid-lowering medication, lifestyle modifications early on can help reduce cardiovascular disease risk in the future. Interventions offered to patients with abnormal lipid screens were individualized based on the type and degree of abnormalities and included dietary/lifestyle modifications, consultation with nutrition and/or endocrinology if other comorbidities were present, genetic testing for familial hyperlipidemia, and medication management for persistent dyslipidemia. Future initiatives can focus on outcome measures, such as potential improvement in dyslipidemia in patients with abnormal lipid panels.

CONCLUSIONS

QI methodology was implemented in local community-based pediatric cardiology clinics to improve adherence to lipid screening. The electronic health record was used to develop high reliability, automated reminders that can be implemented in other local practices to identify patients at risk for early cardiovascular disease.

ACKNOWLEDGMENT

The authors acknowledge Dr. Jeffrey Dayton’s assistance with this project.