• Sustainability
  • DE&I
  • Pandemic
  • Finance
  • Legal
  • Technology
  • Regulatory
  • Global
  • Pricing
  • Strategy
  • R&D/Clinical Trials
  • Opinion
  • Executive Roundtable
  • Sales & Marketing
  • Executive Profiles
  • Leadership
  • Market Access
  • Patient Engagement
  • Supply Chain
  • Industry Trends

Comparing Registry and Electronic Health Record Data for Real-World Evidence Generation: Heart Failure as a Case Study


Contrasting the view of heart failure patients as seen from the perspectives of two large but distinct sources of real-world data: the Veradigm Cardiology Registry and Practice Fusion

Mac Bonafede
VP of Research Consulting

Mac Bonafede
VP of Research Consulting

Heart failure is becoming an increasingly heavy health burden worldwide, affecting an estimated 23 million people. Classification of heart failure patients depends primarily on the degree of left ventricular systolic dysfunction, which is measured via echocardiography. Once heart failure patients have been identified, many researchers examine the relationship between heart failure and other factors, such as obesity, as measured by body mass index (BMI).

As a result, studying and generating real-world evidence (RWE) related to patients with heart failure requires data not commonly found in administrative claims data. A comparison of two real-world data sources and their differences in terms of collection and content as well as the implications of these differences helps identify key patient populations for assisting in understanding heart failure management and potential recruitment for observational research or clinical trials.

The two data sources—the Veradigm Cardiology Registry (formerly PINNACLE Registry), cardiology’s largest outpatient quality improvement registry, and Practice Fusion, a cloud-based ambulatory electronic health record (HER)—contain data points, collected between 2015 – 2020, needed for studies of heart failure, such as left ventricular ejection fraction (LVEF) and BMI.


The PINNACLE Registry was founded in 2008 by the American College of Cardiology (ACC), who now operates it partnership with its owner Veradigm under the Veradigm Cardiology Registry name. The registry captures data on coronary artery disease, hypertension, heart failure, and atrial fibrillation. The data captured for this heart failure case comparison includes data from 4,000 participating sites and 16.5 million patients. Sites that sign up to participate in the registry transmit their patients’ EHR data into the registry in a HIPAA-compliant, secure manner. From there, the data from each site is mapped to assess 24 cardiovascular-related quality payment program measures. The data includes basic demographic information; presence of cardiovascular conditions and comorbidities; cardiac events such as myocardial infarction, percutaneous coronary intervention, hemorrhage, and other medical events with dates; and exams, labs, and medications.

Registry data is refreshed quarterly, and sites are provided with the 24 quality payment program measure reports. This allows each site to compare its performance against its peers and see how it is doing over time. Sites are also able to report directly to their quality payment program through the registry.

The top specialties represented in the registry are cardiology, family medicine, and internal medicine. More small- and medium-sized practices are represented than larger practices.


In contrast, the Practice Fusion dataset provides data from a large, cloud-based EHR that includes both primary care providers and specialists in all 50 states. Unlike the Veradigm Cardiology Registry, most of these sites are small primary care practices. From 2021 - 2022, Practice Fusion included over 48 million patients, which corresponds to more than 123,000 providers in over 62,000 practices. Also, unlike the registry, the Practice Fusion Research Database is refreshed weekly, making it extremely timely. Its data includes patient demographics; provider specialties, geography, practice links; visits, vitals, encounter events, problem lists; medical histories; diagnoses; prescription(s); lab(s); insurance information; and last data refresh data.

Practice Fusion, which is also owned and operated by Veradigm, has HIPAA-compliant, secure access to all the EHR data for Practice Fusion patients, including physician notes and other free text.


Veradigm’s analysis revealed that, in structured fields, the Practice Fusion data contained fewer than 5,000 LVEF values. However, after its data enrichment services including using NaturalLanguage Processing (NLP) and other data mining techniques, they were able to extract approximately 3.5 million additional data points from semi-structured text.


The Veradigm Cardiology Registry had a substantially higher proportion of patients with readily available LVEF data—41.3% had available LVEF between 2016 and 2019—than the Practice Fusion EHR, in which approximately 1 in 20 patients diagnosed with heart failure had useable LVEF data during this period (TABLE 1).

This difference makes sense when considering that LVEF data was primarily extracted from physician notes in the Practice Fusion EHR using NLP. For this case study, a valid BMI measurement was required so it could be used as another way to categorize patients and their cardiovascular risk. This requirement caused a relative drop of 15% in patients available for analysis for both data sources. The two cohorts examined were similar in demographic characteristics (TABLE 2), with a few exceptions:

  • The Practice Fusion EHR patients appeared to be slightly older; however, the age of the Veradigm Cardiology Registry patients needed to be artificially truncated at age 80 to fit within de-identification requirements.
  • This makes the two cohorts more similar in age than shown in TABLE 2.
    Fewer than 50% of patients were female, which is surprising as the heightened inclusion of males in heart failure clinical research is well-documented.
  • Race and ethnicity were not required fields for both data source, but the Practice Fusion EHR did capture a larger relative proportion of non-white patients. This feature has important implications for using these data sources in clinical trials or for prospective observational research.

Related Content