Using Veradigm’s EHR Dataset to Replicate ROCKET-AF Clinical Trial
The COVID-19 pandemic has created numerous healthcare-related obstacles, including obstacles in the realm of clinical research. In the past year, 80% of non-COVID related clinical trials have been suspended or abandoned completely. Ongoing trials face severe difficulties recruiting patients and are facing new challenges in conducting studies. Prior to the COVID-19 pandemic, clinical trials were already increasing in cost and complexity, creating uncertainty for clinical trial planning and execution.1
However, real-world data (RWD), such as that found in the Veradigm EHR Health Insights dataset, can be used to help overcome these challenges. This white paper presents a robust example of using this EHR dataset to replicate a clinical trial. It describes how the Veradigm EHR dataset was analyzed by the HealthPals CLINT™ analytics platform to successfully replicate the results of ROCKET-AF, a large, landmark, international clinical trial. It also demonstrated how the Veradigm EHR dataset could be used to construct an external control arm (ECA) for an ongoing clinical trial.1, 2
The ROCKET-AF Clinical Trial
ROCKET-AF was a large, international, randomized cardiovascular clinical trial that compared the safety and benefit of rivaroxaban versus warfarin in patients with nonvalvular atrial fibrillation (AF).1 This trial utilized numerous clinical variables, such as age, BMI, hemoglobin, blood pressure, and platelet count, as part of its methodology. Replicating the ROCKET-AF trial with RWD required researchers to capture all these clinical variables. Doing so, however, required working with a large, longitudinal dataset, such as that provided by Veradigm. The Veradigm EHR dataset includes over 150M patients, treated by 250,000 clinicians from 35,000 practices.1, 2
Researchers mapped clinical concepts from the trial to all the data available in the EHR records. They identified 644,877 patients in the Veradigm EHR dataset diagnosed with AF.2 Using the clinical concepts, the researchers could then apply inclusion and exclusion criteria from the ROCKET-AF trial, further narrowing the patient sample to match the clinical trial criteria. Patients were then separated into the control and experimental study arms, based on which treatment drug the patients received, rivaroxaban or warfarin.1
Capturing all the clinical concepts from the ROCKET-AF trial allowed researchers to effectively capture the medical state of each patient. This, in turn, allowed the two study arms to be balanced based on the patients’ medical state to reduce bias associated with prognostic variables in the study.1
Replicating a Clinical Trial with an EHR Dataset
This white paper describes advancements in replicating clinical trials with RWD because of the scale and level of detail that an EHR dataset like Veradigm Health Insights offers.1, 2 Researchers have previously tried to replicate clinical trials using data from insurance claims. However, the RWD provided by the large, longitudinal Veradigm EHR dataset offers multiple advantages over insurance claims data.1 Claims data are collected patient information for billing and reimbursement, not for recording clinical details, thus introducing the following limitations when repurposed for clinical trial replication:
- There is often a significant lag between the time when information appears in a patient’s EHR record and the time it appears in their insurance claim. As a result, it takes longer for data to get in the hands of the researcher.1
- Querying the claims data using only ICD codes yields only a fraction of the clinical depth captured by querying the EHR data, so there is more data accessible for analysis in the EHR dataset.1
- Clinical information that is not directly tied to a billed item may not be recorded in the claims data at all, but will likely be recorded in the EHR data.1
External Control Arms
In addition to replicating the ROCKET-AF trial, this study demonstrated how the Veradigm EHR dataset could be used to construct an external control arm (ECA) for an ongoing clinical trial. An ECA consists of a patient cohort derived from RWD that is external to the control arm in the clinical trial. ECA records are matched to the patients in the experimental arm in a way designed to simulate randomization. This matching process reduces bias by balancing characteristics of patients equally between the two arms.1
An ECA can be used to provide a comparison control arm for the experimental arm in the trial, as well as reduce the required sample size for the study control arm. This can then reduce the cost of the trial associated with recruiting study participants. An ECA could also be used to supplement data provided by the existing control arm. Supplementing data used for regulatory submission could help mitigate the risks associated with seeking regulatory approvals.1
The Power of Veradigm’s Rich, Longitudinal EHR Dataset
This white paper shows that Veradigm’s EHR dataset provides RWD that can be used to replicate the methodology of a clinical trial. In this study, Veradigm Health Insights Ambulatory EHR™ provided significantly larger sample sizes for both the control and experimental arms compared to those in the actual ROCKET-AF trial. They also provided data covering a time period four years greater in length than that reported in the original trial.1, 2
These results show that the EHR dataset could be used to evaluate treatment effect longevity for any drug for which one can track the relevant outcomes for an extended time period. They also illustrate that this type of analysis, using a robust dataset such as Veradigm’s, may generate novel insights beyond those yielded by the original trial.1
This study demonstrates that successful analysis of RWD in Veradigm’s EHR dataset can be used as a supplement or substitute for parts of the traditional clinical trial approach for new therapies or expanded use of existing therapies. When paired with the right technology, a RWD-based approach can help reduce a trial’s duration and, as a result, help reduce its cost. A RWD-based approached can also reduce costs by developing an ECA. In the very least, RWD-based approaches can help guide the design of clinical trials and tailor patient recruitment to help them be more efficient.1
Click below to read the full text of this white paper to learn more.
References:
- HealthPals, Veradigm. First Demonstrated Replication of a Randomized Controlled Trial on Electronic Health Record Data Using the Clinical Intelligence Platform (CLINTTM). Accessed March 3, 2021,
- Dash R, MD PhD. HealthPals’ CLINT Demonstrates First Large-Scale RCT Replication with Veradigm’s EHR Data. *Not yet published. Updated January 11, 2021. Accessed March 3, 2021, https://rajdash.medium.com/healthpals-clint-demonstrates-first-large-scale-rct-replication-with-veradigm-s-ehr-data-b23fc64e84ea.