Research in PCORnet^®

Abstract

PCORnet^® is a large, complex network that may appear intimidating for new investigators and funders, posing a perceived barrier to expanding the use of this powerful resource for research. In this commentary, I highlight the importance of mentorship, immersion, and didactic training to build the future cadre of investigators leveraging the PCORnet infrastructure and expand the impact of PCORnet^® Studies, while sharing personal experience navigating the opportunities and challenges offered by this network.

THE CHALLENGE

As described in this issue, the PCORnet infrastructure currently supports more than $1 billion in funded research. PCORnet^® Studies are agnostic to therapeutic area or methodological design and range from retrospective cohort studies to randomized clinical trials to implementation science programs. In my experience, scopes and budgets for projects using PCORnet range from modest observational studies to trials of definitive size, from budgets of $500,000 to $20,000,000.

Well-known PCORnet^® Studies, such as ADAPTABLE,^1,2 PREVENTABLE,³ HERO,^4–8 ACTIV-6,^9–12 and RECOVER,¹³ are nationally-scaled pragmatic research programs led by large and experienced teams at dozens of sites and budgets as high as hundreds of millions of dollars. Junior investigators may lack the confidence, skillset, and track record to lead programs of similar scopes, and also may not understand that the PCORnet infrastructure can support projects of any design, scope, funding, and experience level. I offer personal case examples of pathways to navigate the PCORnet infrastructure with specific suggestions for steps early-stage investigators can take to develop collaborations and explore publicly available resources provided by the network.

A CASE EXAMPLE

As a practicing preventative cardiologist, my research interests are primarily related to the patients I see every day, by developing better care options for the patients who are struggling to control many risk factors for heart disease. My work focuses on topics such as obesity, high blood pressure, elevated cholesterol, and the disparities in care and outcomes that disproportionately affect minority populations. I have a longstanding commitment to patient-centered research and meeting the needs of patients with diverse backgrounds.

My journey with PCORnet began in 2016, when my mentor introduced me to a project leveraging the PCORnet infrastructure to understand the current state of obesity in the United States, and to examine how weight change over time affects modifiable risk factors and outcomes. At the time, PCORnet was an emerging Network, and little was known about PCORnet or how to design a study utilizing the resources.

My first project leveraging PCORnet resources was a retrospective cohort study of overweight patients with body mass index ≥25 kg/m².¹⁴ The budget was modest, and the timeline was relatively quick; <$500,000 within 12 months. The study leveraged data from 11 health systems participating in PCORnet to analyze electronic health record data from 882,712 patients, whose weight change (stable, gained, and lost) was assessed over a 2-year period. Our results showed that weight loss changes are not sustained over time, and also, at that time, pharmacotherapies for weight loss were not well-utilized.

Although ours was among the largest cohorts to describe obesity trajectories and outcomes, there were several lessons learned, many of which reflect variation in clinical practice and coding across sites. For example, health systems are highly variable in how they capture simple yet necessary metrics such as body mass index in the electronic health record. The study team experienced significant delays navigating the conversion of differing indices into an analytic dataset. Fortunately, researchers across the PCORnet infrastructure were invested in identifying data source problems and helping to correct those. Despite these challenges, the study team and funder were enthusiastic about our access to this volume of data, and the ability to find patients meeting our criteria.

Key lessons learned from this study were: (1) electronic health record data are imperfect and are not easily interpretable by statisticians accustomed to receiving “clean” datasets for analysis; and (2) it is essential to have both clinical and informatics staff engaged for any study–these dyads will have to work together to obtain and interpret data from sites participating in PCORnet. These types of collaborations are essential to optimizing the value of using PCORnet. I often find that the local clinical expert who is enrolling patients is not well-connected with the local informatics expert who is extracting patient data, and vice versa. Connecting these individuals early in the planning process is crucial in encompassing all available expertise and harnessing the power of PCORnet^®.

It is also valuable for investigators new to using the PCORnet infrastructure to understand the challenges of contracting and related administrative details in a large national network. At the time of my initial study in 2016, conducting a multisite study was extremely challenging. Since then, embracing a learning health system model, the PCORnet^® Network Partners have implemented numerous efficiencies that I believe have streamlined some of these pain points. Examples include the creation of a master data sharing agreement, master contract template, and operating procedures to facilitate a single IRB.¹⁵ The PCORnet website now includes pages dedicated to sharing tools to facilitate study start-up [https://pcornet.org/resources/tools-for-study-start-up/] and also to cataloging resources to share tools and best practices [https://pcornet.org/resources/].

Since this initial PCORnet^® Study over 8 years ago, I have co-created 2 additional observational studies, 1 randomized clinical trial, and 2 randomized implementation science programs powered by PCORnet^® [NCT06471036, NCT06249165]. Each project has been progressively more complex, resourced, and comprehensive.

On the basis of the 6 studies I have conducted in collaboration with PCORnet^® Network Partners, I have identified 3 primary lessons learned. Each of these lessons contributes to a set of associated recommendations for leadership of the Network—as well as research teams new to using the PCORnet infrastructure—to support the ongoing development of PCORnet as a learning health system that creates strong career development opportunities for investigators.

MENTORSHIP

Mentors who are enthusiastic and willing to educate junior investigators about the nuances of the PCORnet infrastructure are critical. I was uniquely positioned in that my mentor was the Principal Investigator for the Coordinating Center for PCORnet^® and many of the research staff who worked on my project were already familiar with the network. Despite this nearly ideal support system, I still found the Network both daunting and confusing. A training program to familiarize investigators with how to navigate the PCORnet infrastructure is important and would have been valuable to me as I initiated my first study using the network. The newly launched PCORnet^® Playbook provides a useful resource for future investigators to understand how to benefit from the unique resources accessible through this network (https://pcornet.org/pcornet-playbook/). The Playbook is a living and expanding resource that will continue to offer insights into the use of this Network, including practical guidance on approaches to patient and community engagement in research, a critical component if an investigator wants to pursue funding for patient-centered research. PCORnet^® Population Insight reports (https://pcornet.org/data/population-insights/) are another resource to help investigators understand the network. The PCORnet^® Population Insight reports offer a high-level overview of the types of data and cohorts that have already been developed using the network. Interested investigators may familiarize themselves with the breadth and depth of the populations and comprehensive clinical conditions that may be studied using PCORnet.

In addition to understanding the nuances of the network and considerations for executing a PCORnet^® Study, researchers may also benefit from recognizing the funding opportunities available to support research leveraging PCORnet, such as the relatively new PCORI Broad Pragmatic Studies—category 3 funding announcement, which requires the use of PCORnet resources (https://www.pcori.org/funding-opportunities/announcement/broad-pragmatic-studies-pcori-funding-announcement-cycle-1-2025).

Mentors are essential to helping new investigators navigate these and other funding opportunities using PCORnet. For example, the PCORI Broad Pragmatic Studies opportunity mentioned above includes an option for investigators to propose coverage of structured mentorship activities for early-career or mid-career investigators, investigators making a career transition to patient-centered comparative effectiveness research (CER), and/or patient and community partners. This program can cover up to 25% effort for mentees and 5% effort for mentors and provides one pathway to structured mentorship using PCORnet.

The PCORnet community has also recently embarked on an immersive and educational series of workshops across the nation to increase the visibility and public awareness of PCORnet. I believe these will be valuable in helping navigate the challenges of building a research portfolio that utilizes this unique resource and can help pair mentors and mentees interested in using PCORnet.

Expect to Be a Trailblazer: Plan to Share Your Lessons Learned. The mantra of see-one, do-one, teach-one somewhat describes my experience in navigating the real-world data resources accessible through the PCORnet infrastructure. The early years of PCORnet did not allow the experience of a “see-one” moment before starting a new project. Because the size of this national network is so large, particularly in the early days, each new project powered by PCORnet was addressed as a “do-one.” For example, in my first project, given the reliance on electronic health record (EHR) data that differed across sites, my first PCORnet^® Study team was unable to determine reasons for weight change and could not attribute weight change to clinical outcomes. As our statistical team learned, electronic health record data are a messy reflection of clinical documentation. Even now, EHR data cannot be expected to be complete nor traditionally analysis-ready, and are subject to a variety of interpretative challenges.

The challenges of real-world data mean it is imperative for clinical investigators and informaticists to work together closely. In those early days, the clinical investigators at the sites were not well-connected with the informatics teams at the sites, creating some confusion and delays for us all. These connections across team members have improved over time as the network has matured. Processes, such as data curation and quality checks have become more common, and as documentation has become much more readily available. Nonetheless, changing the culture of research to work across so many sites requires a genuine commitment to collaboration and team science—and willingness to “pay it forward” by documenting and sharing effective approaches. The PCORnet infrastructure now offers many ways for investigators and research team members to share lessons learned, from modules in the PCORnet^® Playbook, to best practice sharing sessions and webinars, to the PCORnetwork handle on LinkedIn.

TRAINING

To increase the usage of PCORnet, training is essential. Workshops, local presentations, and site visits from those who have utilized the network will be useful in preparing and supporting new investigators who may be interested in partnering with PCORnet^® Network Partners. There are lessons to learn related to community and team building as well as understanding the data. The PCORnet infrastructure is complex and warrants dedicated time to educate the broader research community about the opportunities.

The newly initiated educational workshops and events and the recently launched PCORnet^® Playbook will be instrumental in providing this fundamental level of understanding. I have also recently contributed to the PCORnet^® Best Practice Sharing Sessions to transparently share my lessons learned and advice for future researchers utilizing this incredible resource. Despite the initial learning curve required to use PCORnet, I believe it is worth the effort to tap into this unique national resource that provides access to extensive clinical data as well as clinical and patient partner expertise.

OPPORTUNITY

PCORnet is a complex and large network. It can be intimidating for new researchers to know where to start. Through the last 10 years, I have learned and experienced the exciting benefits of partnering with the PCORnet^® Network Partners. I have been pleased to observe that the PCORnet infrastructure has matured substantially in this time and is now ready to invite more early-stage investigators to use the network.

For investigators looking to get started with the network, I highly recommend starting with the PCORnet^® Front Door team at www.pcornet.org/front-door. The Coordinating Center for PCORnet^® for the network can provide resources to new investigators that can help create connections and collaborations. As my experience and research track record demonstrate, with appropriate mentorship; willingness to experiment and build experience; and pursue training, PCORnet is a valuable resource for everyone – from the most junior to most senior levels.