Integrating clinical research into electronic health record workflows to support a learning health system

Abstract

Objective

Integrating clinical research into routine clinical care workflows within electronic health record systems (EHRs) can be challenging, expensive, and labor-intensive. This case study presents a large-scale clinical research project conducted entirely within a commercial EHR during the COVID-19 pandemic.

Case Report

The UCSD and UCSDH COVID-19 NeutraliZing Antibody Project (ZAP) aimed to evaluate antibody levels to SARS-CoV-2 virus in a large population at an academic medical center and examine the association between antibody levels and subsequent infection diagnosis.

Results

The project rapidly and successfully enrolled and consented over 2000 participants, integrating the research trial with standing COVID-19 testing operations, staff, lab, and mobile applications. EHR-integration increased enrollment, ease of scheduling, survey distribution, and return of research results at a low cost by utilizing existing resources.

Conclusion

The case study highlights the potential benefits of EHR-integrated clinical research, expanding their reach across multiple health systems and facilitating rapid learning during a global health crisis.

Introduction

Integrating clinical research into routine clinical care workflows within the same electronic health system tends to be challenging, as well as expensive and labor intensive.¹ Historically, clinical and research endeavors fundamentally differ in how encounters are managed, how data is collected and managed, and the degree of procedural rigor, including decisional flexibility at the point of care and revenue capture. In the past, this has led to the natural evolving development of separate clinical and research information systems, even though using a single system would most benefit the participant and the healthcare delivery system conducting the research. However, as healthcare embraces continuous improvement through learning healthcare systems (LHSs), clinical research and clinical care activities are being conducted in parallel to a greater extent and, in many cases, overlap. Now with a shared goal—learning from clinical interventions—health systems’ electronic health records (EHRs) are increasingly being utilized for research purposes, primarily for data aggregation, information exchange, and large clinical trials,^1–4 though processes remain limited and often fragmented.⁵

The global COVID-19 pandemic created unprecedented challenges to health systems at large, and has forced creativity and flexibility around both learning and work environments. There is an immense amount of unknown information, and yet the speed at which new information is generated is rapid, in many instances outstripping our ability to analyze and interpret it. To adapt to this environment, utilizing a highly reliable LHS’s resources, particularly informatics resources, efficiently and effectively becomes vital.^6–11 It is here where integration and alignment of research trial and clinical activities through the EHR are not only novel and invaluable, but also necessary.

The Agency for Healthcare Research and Quality (AHRQ) defines LHSs as those in which internal data and experience are systematically integrated with external evidence to facilitate data-driven intervention.^6,12 A key characteristic of LHSs is the use of the “Deming Cycle,” a derivation of the scientific method.¹³ Contemporary EHRs are evolving to better support this cycle in LHSs. We posited that research that aligns with evidence-based evaluations commonly done in LHS could be conducted at scale with speed using a modern commercial EHR.

We present our experience leveraging an existing commercial EHR to conduct a large-scale clinical research project quickly during the COVID-19 pandemic. The aim of this specific project was to investigate population-level immunological response to COVID-19 by integrating clinical data on diagnostic testing and vaccination and research-based quantification of antibodies against multiple SARS-CoV-2 variants. Speed was of the essence during the pandemic, as it was important to understand the level of immunity in the population from both viral infection and vaccination, as well as the utility of the antibody assays utilized for the study. The primary goal of this case report is to share this novel experience of integrating research participant recruitment and follow-up within existing routine clinical workflows in the EHR.

Methods

The University of California San Diego (UCSD) and University of California San Diego Health (UCSDH) COVID-19 NeutraliZing Antibody Project (ZAP) is an ongoing project that began in January 2022 aiming to evaluate the immune response to SARS-CoV-2 virus in the population of a large academic institution, and examine whether antibody levels (from either vaccination or infection) are associated with subsequent diagnosis of SARS-CoV-2 infection. The project is open to all staff and students at UCSD and UCSDH, who were regularly testing for SARS-CoV-2 infection from October 2020 to July 2022 according to institutional and state guidelines. This project represented an LHS activity; however, quality improvement (QI) activities can sometimes involve risk (minimal in this case) and the protocol was submitted and approved by the Institutional Review Board. In addition, this report aligns with the Consolidation Framework for Implementation Research (CFIR).

Given the societal emphasis on social distancing during respective COVID-19 surges, recruitment for this project needed to be conducted primarily electronically—specific methods included emails that contained links and QR (Quick Response) codes to the scheduling platform, UCSD mobile application announcements, and flyers with QR codes that were posted in clinical and educational sites where staff and students had in-person activities. Participants self-scheduled research visits linked to their existing medical records using the patient-facing portal for the UCSDH EHR (Epic MyChart, Verona, WI). As a result of the EHR integration of the scheduling platform, participants had the ability to complete an “eCheck-in” process in Epic MyChart prior to their study visits, which included a study questionnaire and electronic informed study consent¹⁴ and HIPAA (Health Insurance Portability and Accountability Act) authorization (eConsent) in addition to relevant ambulatory items as necessary. This eCheck-in process could be completed remotely from any computer or mobile device with internet access.

Operationally, the in-person fingerstick blood spot testing was primarily integrated into existing vaccination and/or COVID-19 testing sites at UCSD and UCSDH. At the testing sites, participants utilize their UCSD mobile application scanner to capture barcodes on the blood spot collection cards (the same method used for on-site COVID-19 staff self-testing), which sends linked barcode/participant identification pairings to the lab performing the serology and neutralizing antibody testing. A secure laboratory information system (LIS) accessions and tracks the study samples, and intakes and stores serology and neutralizing antibody results, which are then returned securely to participants back in the patient-facing EHR, along with a link to the ZAP project website, which contains additional information for contextualizing individual results.

Participants are sent follow-up study questionnaires through the EHR patient portal 30 and 90 days after their initial study visit, which can be completed on the MyChart mobile or desktop applications. Participants are also invited for return visits to track results longitudinally. Given the project’s full integration with the EHR, individual and group reports are able to be pulled directly from the EHR’s data warehouse and analyzed using the research team’s analytic programs of choice (Figure 1).

Figure 1.

Alignment of clinical care and research trials through the electronic health record (EHR) at a learning healthcare system.

Results

By the end of 2022, over 2500 UCSD and UCSDH staff and students had participated in the project. There were 2523 initial visits and 652 repeat visits (participants were allowed to present for multiple research visits). Figure 2 shows cumulative consent and sample collections over time; we note that intensive recruitment was focused on the first 2 months of the study period. Participant demographics are outlined in Table 1. More than 90% of those who enrolled and consented for the study completed at least one visit, including completion of the pre-visit questionnaire and submission of a fingerstick blood sample. Follow-up survey response rate was 70.1% at 30-day post initial visit, and 48.5% at 90-day post initial visit.

Figure 2.

ZAP enrollment over time. Solid line representing participants who consented to the study, and dotted line representing participants who consented and from whom samples were collected and recorded.

Table 1.

ZAP participant demographics.

Participant information (consented N = 2727)
Demographics
Age (years)
Mean (SD)	38.3	12.8
Median (Min, Max)	37	18 80
Sex (n, %)
Male	867	31.8%
Female	1852	67.9%
Unknown	8	0.3%
Ethnicity (n, %)
Hispanic	473	17.4%
Non-Hispanic	2095	76.8%
Unknown	159	5.8%
Race (n, %)
White	1503	55.1%
Black	65	2.4%
Asian	703	25.8%
Native American/Alaska Native	17	0.6%
Native Hawaiian/Pacific Islander	32	1.2%
Other/Mixed	283	10.4%
Unknown	124	4.5%
Prior exposure status
Vaccinated, full (n, %)a	2678	98.2%
Study visits
Initial visitsb	2523	92.5%
Repeat visits (total)	652	–
Unique follow-up visitsc	601	22.0%
Survey rates
Baseline	2583	94.7%
30 days	1912	70.1%
90 days	1322	48.5%

^aAn additional 8 participants were partially vaccinated (1 dose of either Pfizer or Moderna).

^bNumber sampled; does not include participants who arrived for a visit but did not scan a sample.

^cNumber of participants with at least 1 follow-up sample.

Discussion

This project is one of the first published reports describing a clinical research project that is integrated into an existing commercial EHR as a single centralized study platform and conducted at a quaternary healthcare system. In addition to commercial EHR integration, the research project also leveraged existing COVID-19 testing and vaccination operations, including clinical and laboratory staff, the UCSD mobile application, and a clinical LIS. This utilization of existing resources, including EHR integration, allowed for increased enrollment during a global pandemic, at low cost. During this process, many valuable lessons were learned.

The implementation of electronic health record systems for clinical operations, though met with some criticisms, has been overall beneficial for record-keeping and searchability, health information exchange,^15,16 patient safety,^17,18 and patient engagement.¹⁹ Research systems, including traditional clinical trial management systems (CTMSs), were initially developed separately because of their many workflow differences, data rigor differences and different goals at the time. More recently however, given the benefits of the EHR as outlined, it seems logical that integrating clinical research projects using available EHRs might reap these same benefits as a result of increased interoperability.

The Veterans Affairs Healthcare System (VA) through the Massachusetts Veterans Epidemiology Research and Information Center (MAVERIC) has conducted several integrated “Point of Care Clinical Trials” (POCCT) using the EHR.^20–22 In these trials, clinicians are prompted in real time during clinical care patient encounters by a clinical decision support system to consider enrolling their patient.²⁰ Unlike the MAVERIC POCCT trials, which utilized a bespoke EHR developed for and restricted to the VA Healthcare System, ZAP used a commercial vendor, which has the valuable potential for expansion across multiple health systems.

ZAP was conducted in a unique environment at an increased speed necessary to understand the COVID-19 pandemic as it progressed. As an established “rapid learning” healthcare system,⁶ UCSDH was able to support this kind of speed, particularly in this instance related to an infectious disease that infects both healthcare and non-healthcare personnel the same. As a result, the research team demonstrated that a large-scale research trial can be performed using existing information ecosystems, without negatively impacting or interfering with other clinical processes.

It is important to note that the complete functionality does not yet exist in this vendor’s EHR to support a fully interventional randomized clinical trial (RCT) involving possible blinding of treatment. However, the system supports several research-focused functions, such as research specific data collection in electronic case report forms, documenting adverse events/severe adverse events (AE/SAE), and external trial monitoring, albeit these are nascent in their functional evolution. This case study highlights the potential benefits of EHR-integrated clinical research for a subset of research trials or projects that can be embedded into standard clinical workflows, facilitating rapid learning during a global health crisis.²³

The sample size of trial participants in other clinical trials conducted to correlate COVID-19 antibody assays with vaccine protection ranges from 338 to 1575, at the highest.^24–28 The ZAP project was able to accrue higher levels of participants at relatively low cost as the result of EHR-integration. The ease of study enrollment using the electronic platform reduced friction substantially. The primary cost of patient acquisition was labor, though this was mitigated by integrating with existing self-testing workflows. Therefore, ZAP highlights the demonstration that one can use Office of the National Coordinator for Health Information Technology (ONC)-certified (commercial) EHRs to very rapidly conduct research in which the cohort, the workflows, and the intervention are optimally aligned with the capabilities of existing commercial EHRs.

This high and rapid enrollment was also feasible in part because the study targeted a student and employee population, including both healthcare and non-healthcare personnel. Importantly, at the study institution, both the employee health and student health systems are EHR-integrated with 100% enrollment in the EHR patient-facing portal.^29,30 We view this as a unique cohort of clinical research participants—relatively motivated and healthy people who participate in educational and/or employment activities every day, are highly knowledgeable and engaged with their healthcare and the UCSDH system, are generally comfortable with technology and mobile applications, underwent regular COVID-19 testing during the pandemic, and were mostly (98.2%) fully vaccinated. Although the study population did not represent the full patient population of the study institution, it was an ideal population for proof-of-concept of an EHR-integrated research project, and to understand the relative contributions of COVID-19 vaccination and infection on immune response.

An LHS is not just a place to render care. A highly reliable LHS is a forward-thinking healthcare system,^6,31 learning from not only our patients but from ourselves (institutional employees and students). This project is one example of this, and led to important insights into the implementation of a fully EHR-integrated clinical research project that utilizes existing institutional operations.

Though many successes were identified from this project, there were also several obstacles faced by the both the investigators and participants. While end-to-end integration of electronic systems from recruitment and enrollment, through sample and data collection, to reporting of results, streamlined record-keeping and standardization, parts of the process, such as eCheck-in, were surprisingly difficult for some participants to complete efficiently. Additionally, there were ethical issues to consider given the participant population, including how to establish an LHS by learning from its own staff and students while preserving and protecting personnel autonomy. It is important to note that individual results were released to the participant only, and were not made available to supervisors or professors. Lastly, this project was conducted at a single institution with strong existing digital health infrastructure (though with multiple campuses), which may impact generalizability of the study results.

Our vision for the future of learning workflow integration into the EHR is bright. The EHR can not only house and make relatively available large amounts of clinical data for observational research, but it can also significantly lower the oftentimes prohibitive costs of prospective clinical research trials. One group has previously described a “green button function” that leverages aggregate EHR data for real-time personalized comparative effectiveness informational support at the point of care.³² With this approach, a successful learning health care system could enhance clinical patient care and research to help guide providers and patients toward informed health decisions at the bedside in the present and in the future.

Author contributions

Nicole H. Goldhaber (Conception, Design, Data acquisition/interpretation, Drafting, Final approval), Marni B. Jacobs (Conception, Design, Data acquisition/analysis/interpretation, Review, Final approval), Louise C. Laurent (Conception, Design, Data interpretation, Review, Final approval), Rob Knight (Conception, Design, Review, Final approval), Wenhong Zhu (Design, Data analysis, Review, Final approval), Dean Pham (Design, Data acquisition, Review, Final approval), Allen Tran (Design, Data acquisition, Review, Final approval), Sandip P. Patel (Conception, Design, Review, Final approval), Michael Hogarth (Conception, Design, Data interpretation, Review, Final approval), and Christopher A. Longhurs (Conception, Design, Data interpretation, Review, Final approval)

Funding

This work was supported internally by UCSD Chancellor Khosla and UCSDH CEO Maysent.

Conflicts of interest

The authors have no competing interests to declare.

Data availability

The data underlying this article may be shared on reasonable request to the corresponding author.