Clinical InvestigationsGeneralizability of an EHR-network dataset to the United States for cardiovascular disease conditions: Comparison of Cerner real world data with the national inpatient sample
Section snippets
Cerner real world data
Health systems in the United States who use Cerner's EHR or Cerner's population health management platform are eligible to participate in CRWD. Systems who opt-in contribute EHR data from both Cerner and, if necessary, non-Cerner EHRs from their health system. Data are mapped to a Cerner-specific common data model, fully de-identified, and made available to researchers for analyses on a cloud-based data science platform.
We used data from CRWD to identify all adult (age >18) hospitalizations for
Results
Initially, 86 health systems were included contributing data to CRWD, representing 6,612,117 hospitalizations (range 4-987,711, median 35,729, interquartile range 3,308-85,630) between January 1, 2017 and December 31, 2018. Among these, 33 health systems were excluded for either not identifying final diagnoses for the hospitalization, or being hospitalized in a system with 30 or fewer patients with MIs during the study period (eg, children's hospitals). Characteristics of health systems that
Discussion
Our analysis overall found that EHR data from, CRWD, an aggregate EHR based network, was closely matched to data found in the NIS. In particular, we found that demographics, procedure rates, and hospitalization outcomes were similar for patients hospitalized for MI, CHF, and stroke. Overall, these analyses support the representativeness and generalizability of CRWD as a source to evaluate national-level patterns of care and outcomes for cardiovascular disease.
The NIS has long been used for
Funding
This project was supported by grant funding to Duke University and UT Southwestern Medical Center from Janssen. The sponsor had no role in the development of the analytic plan, data analysis, data interpretation, or decision to publish.
Disclosures
NP Shah has received research funding from Amgen, Janssen, and Novartis, and honoraria and consulting fees from Amgen, Novartis, and Esperion. AM Navar and ED Peterson receive research funding to their institution for research from Bristol Myers Squibb outside the scope of this research project, consulting fees for research consulting from Bristol Myers Squibb, Boehringer Ingelheim, and Janssen Pharmaceuticals, and consulting fees for research advising from Cerner outside of this project. The
References (12)
- et al.
STS adult cardiac surgery database: 2021 update on outcomes, quality, and research
Ann Thorac Surg
(2021) - et al.
Risk adjustment for in-hospital mortality of contemporary patients with acute myocardial infarction: The Acute Coronary Treatment and Intervention Outcomes Network (ACTION) Registry®–Get With The Guidelines (GWTG)TM acute myocardial infarction mortality mo
Am Heart J
(2011) - et al.
Challenges of using medical insurance claims data for utilization analysis
Am J Med Qual
(2006) - et al.
Strategies for building delivery science in an integrated health care system
J Gen Intern Med
(2019) - EPIC Systems. Published 2022. https://www.epic.com/about. Accessed October 13,...
- Oracle Cerner. Published 2022. https://www.cerner.com/. Accessed October 13,...