How to find the right source of real-world data: Claims vs. EHR data

With so many types of real-world data available, it can be difficult to select the source that will meet your needs—that is, the data that will help you accurately and efficiently answer your research questions.

For behavioral health research, a field in which forward momentum has been inhibited by a lack of standardized measurements and biomarkers, choosing the right source of real-world data is especially critical in order to gain new insights into these complex conditions.

Retrospective studies are often conducted using real-world data, including:

• Claims data, or data from provided healthcare services that are reimbursed by health insurance companies. Claims data are often used for retrospective research on healthcare utilization and other questions.
• Data from the electronic health record (EHR), or information that is routinely collected as part of patients’ clinical care. Our NeuroBlu Database contains EHR-derived real-world data representing over 1.4 million patients.

To help you decide which real-world data you should use for your project, we break down below some of the major differences between claims and EHR data.

Representativeness

Pro of claims data:

• Nationally representative

• Certain individuals are not represented in claims, such as the uninsured (estimated at about 8 percent of the U.S. population as of 2022) or those who pay for healthcare out of pocket.

• Because data is captured in routine clinical care, the data are likely to be reflective of all patients who receive care.

• Certain individuals, such as those not receiving treatment, are not represented within EHRs.
• Although national representativeness is more difficult to achieve for EHR data, it is a goal we are working toward. Our NeuroBlu Database currently represents over 30 geographically diverse treatment centers, with goals to incorporate data from every state by the end of the year.

Other data features

Pro of claims data:

• Because claims data are recorded each time a claim is submitted for reimbursement, data are likely to be more complete than many EHR-derived sources, which rely on providers to accurately and completely fill in the EHR and often suffer from data missingness.

At Holmusk, we are working to improve the completeness of EHR-derived data by going straight to the source: data capture. With HolCare EHR+, our EHR platform designed specifically for behavioral health, we’ve made it easier for providers to record accurate and complete data while still focusing on patient care.

Pro of EHR data:

• The type of information captured by claims forms is limited and often much less detailed than what is captured by the EHR. Claims data cover information on demographics, diagnoses, and treatments—but EHR data can often provide even more granular information that provides a more comprehensive view of the patient experience.

It’s important to remember that different sources of EHR data will vary in terms of what kinds of information they can deliver. For example, our NeuroBlu Database contains clinical features from the Mental Status Examination (MSE) such as symptoms, stressors, and outcomes—all valuable contextual data that are obtained through our natural language processing techniques and not always available via traditional EHR sources.

It is clear that both claims data and EHR data have their strengths and weaknesses. Different types of real-world data often complement each other by filling in each other’s gaps—but if you’re faced with selecting one source of real-world data, that choice should be made carefully depending on the questions you’re asking and the goals of your project.

Ready to see if the real-world data within the NeuroBlu Database can help to advance your goals? Get in touch today.