The recent publication of FDA Guidelines regarding the use of electronic health record (EHR) data in clinical trials heralds an exciting evolution in the design and execution of such studies. Now clinical notes, physician orders, radiology, laboratory, and pharmacy records from routine care — even records from foreign clinical sites — can be used in product submissions. These real-time, real-world data can be combined, aggregated, and analyzed, providing clinical investigators access to unprecedented mid-trial insights, and facilitating post-trial follow-up to assess long-term safety and efficacy. Further, by simplifying follow-up of a large patient base, the use of EHRs can help capture data in prophylaxis studies when the outcome of interest rarely occurs, and help inform regulatory decision making in the approval of new indications for approved drugs.
Yet, EHRs remain outside the purview of the FDA and its regulated entities, so for such data to be ultimately appropriate for FDA submission, data quality must be thoroughly verifiable. To take full advantage of the expanded trove of data found in EHRs, it is important that sponsors not only select a data-agnostic electronic data capture (EDC) platform but also establish clear-cut parameters to ensure data integrity.
To assist in this process, we offer the following best practices for use when incorporating EHR data into clinical trials.
- Use ONC Health IT Certified EHR sources when possible. First, since the FDA recommends using ONC-certified sources, doing so will simplify compliance. Better, since ONC certification requirements help ensure interoperable data sharing and data security, such certification likely means that related trial data can be analyzed and that it will remain HIPPA-compliant. Of course, data from foreign studies will not be from ONC-certified systems. For those, ensure that other safeguards listed here are firmly in place, from creating audit trails to limiting access to electronic records.
PRO TIP: Maintain a list of EHR systems used by each investigation site; include manufacturer, model number, and version number. If a system is decertified during the course of a study, determine the cause, and whether it compromises the integrity of the relevant study data.
- Choose an EHR that is targeting the right data. EHRs were not developed to capture data for clinical trials — so it is important for a sponsor to ensure that the EHR in question is collecting the data most pertinent to their trial. For instance, the sponsor of an oncology trial should avoid linking to a primary care EHR; it is not designed to capture the information needed for the trial. But EHRs used by the Cancer Centers of America and other such institutions were designed specifically for oncology data, so they gather such information in an appropriate, analyzable way.
- Choose an EDC platform that can collect complex variables and flag data quality issues. In any clinical study, data quality is crucial. With the right parameters in place, an EDC platform can select the correct information even from a more open-ended EHR. Moreover, it can flag any data in the EHR that may be not perfectly precise. For instance, a patient’s weight may be recorded as 150 pounds one day, 142 the next. For most purposes that doesn’t matter — but in a clinical trial, it could matter a lot. The right EDC will issue a query, so the discrepancy can be resolved. Further, because the FDA demands a clear audit trail (see point 7 below), the optimal platform will not only flag those inconsistencies but record every aspect of the input, from the initial record to the discovery of a potential issue, to any corrective activity.
- Protect patient privacy. Patient consent forms must note the extent to which records can and will remain strictly confidential, and precisely who might gain access to the patient’s records — including physicians, investigators, sponsors, and even the FDA. To provide an accurate picture for the patient of all the parties who might access their data, sponsors and investigators should have a detailed understanding of data flow and data visibility within all systems. Of course, circumstances are always important: Even when privacy cannot be absolutely guaranteed, many patients will be happy for their information to be used in the quest to cure a disease; however, few will want to simply enable the monetary gain of a pharmaceutical or biotech company.
- Double-check that a blinded study will remain blinded. To ensure that a blinded study remain valid, sponsors should confirm that neither the EHR nor the EDC platform has the potential to unblind the treatment allocation — and if either one does have that potential, sponsors should put the appropriate controls in place to prevent an unblinding.
- Limit EHR access to authorized users. Prior to the study initiation, each electronic data element should be associated with a data originator. Once data are in the system, the clinical investigator and other pre-designated key personnel should be the only individuals authorized to modify or correct data.
- Create audit trails with identifiable authors. Complete audit trails track all changes to data, including precisely who made any changes, the date and time the change is made, and the reason for the change. Critically, for FDA approval, the system must retain each set of data, both original and modified, so that the FDA has a complete record. Some systems (such as the Remarque Quality Management System) automatically create such records, providing a built-in audit trail. The clinical investigator should always review and electronically sign the record for each study participant before data are submitted or archived.
- Retain records for as long as required. The FDA requires access to all records pertaining to a clinical trial — both paper and electronic — for two years after a marketing application is approved (or two years after an investigation is discontinued). If the necessary records are not available — for instance, for a foreign study — it is possible that the FDA won’t accept the study data.
The bottom line: Employ effective quality checks for EHR data
The ability to harness EHR data dramatically expands the boundaries of clinical trials, creating the potential to conduct broader and deeper studies more economically. Yet sponsors must remember that EHRs were designed to capture data that tracks ongoing patient treatment or data with which to bill insurance companies — not data with which to substantiate clinical trial results. Consequently, they must ensure that effective quality checks are in place so that when study results are presented, the data are FDA-compliant. To learn more about how Remarque enables you to leverage EHR data in your own clinical trials, request a personalized demo today.