Multisite Participant Recruitment using EHR-Derived Cohorts:
Guidelines and Best Practices for Researchers

Multisite participant recruitment using EHR-derived cohorts (hereafter referred to as “EHR recruitment”) requires special planning and consideration due to complexities around patient privacy, data limitations and security, and variance in local administrative practices. These guidelines and best practices are intended to aid researchers in creating a multisite EHR recruitment plan that is respectful to patients, minimizes the risk of loss of patient confidentiality, and helps researchers anticipate and prepare for patient feedback.  This document was developed as a collaborative effort by the UC BRAID Participant Recruitment workgroup, comprised of representatives of University of California campuses at Davis, Irvine, Los Angeles, Riverside, San Diego, and San Francisco, with input from Stanford University. 

Note: For the purposes of this document, “EHR recruitment” is defined as: Recruitment of patients who are identified via a systematic query of the electronic health record. The query generates a list of patients who may meet basic study eligibility criteria; providers may span multiple clinics and disease areas and initial contact is conducted outside of the clinic setting.

​

Characteristics of EHR recruitment:

• PI may be unknown to the patient.

• Lists of potentially eligible patients are derived via an EHR data extraction, typically by a certified EHR programmer. 

• Initial contact is conducted outside of the clinic, typically by letter, email, or a patient portal.  

Characteristics of non-EHR recruitment:

• Data is derived from any source other than the EHR (e.g., clinic administrative data, registries, publicly available mailing lists).

• Initial Patient eligibility is determined by a chart review.

•  Initial recruitment approach is in a clinic setting.

IRB approval for any recruitment method is granted on a case-by-case basis; EHR recruitment may not always be appropriate or practical, depending on the context of the study design, patient population, availability of data, and research infrastructure at recruitment sites. Some patient populations may not be suitable for EHR recruitment, such as when eligibility is based on sensitive or rare diagnoses. In these cases, PIs may wish to establish additional collaborations with clinicians working in the disease area to determine appropriate and effective methods for recruiting these patients. 

​

The full Guidelines and Best Practices can be downloaded here as a pdf. It has also been added for other institutions to the Trial Innovation Network Toolkit here.

Stanford