JMIR Publications

ABSTRACT

Background:

The digitization of medical data and advances in interoperability have opened opportunities for research studies to use more comprehensive, longitudinal patient data from multiple sources. As patients often interact with many providers and payers over time, collecting data from these organizations may have critical implications for accuracy and bias in study results. United States policy has promoted exchanging health information among providers, payers, and patients, but less attention has focused on facilitating data collection for research, which presents unique challenges.

Objective:

To identify and evaluate existing and emerging approaches for collecting comprehensive provider and payer data for research in the US, with the twin goals of informing researchers of possible methods and generating evidence to inform policy initiatives. Our focus was on electronic approaches to data aggregation for studies requiring patient consent.

Methods:

We conducted a landscape analysis through subject matter experts (SMEs) interviews. SMEs were selected based on expertise. Using these data sources, we created a list of evaluation criteria; identified existing and emerging approaches; and described benefits and limitations of each approach by applying the evaluation criteria. We interviewed SMEs until we achieved saturation in concepts. Data was limited to the United States and may not represent all relevant perspectives.

Results:

A total of 20 SMEs helped identify 8 distinct approaches: (A) general purpose smartphone app, (B) commercial app, (C) research community app, (D) structured data export, (E) TEFCA Individual Access Service, (F) regional study query, (G) national study query, and (H) aggregated data source. Participant-mediated exchange approaches (A-E) leveraged patients’ right of access. Three approaches leveraged existing data exchange services (E, F, G). To evaluate these approaches, we identified 12 criteria which included perspectives of participants, research teams, and broader stakeholders. Each approach had benefits and limitations; no single approach emerged as superior for all use cases. While currently available approaches for participant-mediated exchange bypass the need for complex governance arrangements, they are limited by participant burden, effort needed by research teams, and data gaps, especially from payers. Some regional health data utilities and aggregated data sources address governance challenges and can provide services such as preparing analytic data sets but are restricted to specific locations and/or coverage of data sources. National networks currently do not allow queries for research and confront challenges establishing trust and enforcing compliance with data sharing requirements among network sites.

Conclusions:

Collecting comprehensive health data from multiple providers and payers in the US is a complex and evolving process. The suitability of an approach may vary based on the specific characteristics and needs of a study. Given the numerous barriers and the lack of a clear dominant method, further exploration and benchmark comparisons of all identified approaches are necessary. Ongoing public policy efforts will likely play an important role in progress.