Those involved in current electronic health record implementations or preparing for "meaningful use" reporting may find it difficult to think beyond these challenging goals. Yet a medical practice or hospital with an EHR is in a unique position to utilize this resource for research and not just quality reporting. EHRs offer a source of rich clinical data for use in research. Several examples are emerging from the secondary use of EHR data.
EHRs can be the basis of disease registries. Traditional registries require time-consuming manual data abstraction, and it often takes many months to obtain data. EHRs make data readily available to study the characteristics of a large cohort with a specific condition and provide a unique opportunity to advance medical science.
A recently developed cohort of chronic kidney disease that emerged from a registry of 57,000 patients can serve as a model for studying conditions seeing growing incidence levels.
But smaller registries of rare diseases can be developed as well. The opportunity to collaborate with other medical centers using EHR-based registries will provide larger patient cohorts and more opportunities for studying subgroups within a population.
This type of data can enable comparative effectiveness research as well. For instance, with EHR systems that include data on large groups of patients, it is possible to compare treatment outcomes in two or more groups on different drugs. The potential to perform clinical effectiveness research without the daunting task of large clinical trial recruitment provides an incredible advantage. A recent study using EHR data which compared four oral agents for diabetes demonstrated the ability to predict a six-year risk of patients with Type 2 diabetes based on the class of diabetes drug they were prescribed.
The EHR provides a platform for clinical trial recruitment as well. Using inclusion and exclusion criteria, clinical trial alerts -- which pop up like a clinical decision support alert -- can be created to prompt the physician to contact the research nurse. Studies have demonstrated a significant increase in clinical trial recruitment using these alerts. In addition, EHR data can be used to evaluate the feasibility of studies by tapping into these large cohorts to refine inclusion criteria. By choosing inclusion criteria with a large enough potential sample size for recruitment, the potential for adequate subject recruitment is enhanced.
Recently, it has been recommended to use EHRs to record adverse events for post-marketing evaluation of drugs and devices. This could provide valuable information for early identification of side effects that might not have been identified during clinical trials. A recent study demonstrated that integration of adverse event reporting into the EHR workflow made it more likely for physicians to report those events (217 events were reported in 16 months).
With the current concerns about privacy and security, greater efforts toward monitoring secondary use of EHR data need to be implemented.
Specifically, these large data sets should be de-identified to the greatest extent possible without undermining research goals. In some cases, limited data sets, which include some protected health information, will be required for research and should be managed in a secure environment.
Some academic medical centers include consent for research, including the use of EHR data, in their admission consent forms. Institutional Review Boards need to be educated on this new area of research to provide appropriate controls without inhibiting legitimate research.
There are limitations to research using EHR data. First, adoption of EHRs has been gradual, although there are some signs of recent acceleration.
Some EHRs are overly dependent on text data requiring the more complex task of natural language processing for analysis. Clinical research requires discrete data elements in a data set. If the EHR provides this up front, data extraction for research is enabled.
Other limitations include the issue of missing data, which can occur during busy clinical care. Meaningful use reporting might help address some of these issues as the same type of discrete data and data accuracy is required.
With attention to privacy concerns and collecting accurate data at the point of care, clinical research from the EHR has a promising future. These use cases are just beginning to be defined and research reports from these records is just beginning to appear in journals.
While this branch of research will not completely replace clinical trials, it will make major contributions in comparative effectiveness research and research recruitment.
MORE ON THE WEB
- "Development and Validation of an Electronic Health Record–Based Chronic Kidney Disease Registry" (Navaneethan et al., Clinical Journal of the American Society of Nephrology, 11/2010).
- "Using Computerized Registries in Chronic Disease Care" (California HealthCare Foundation, 2/2004).
- "Effect of a Clinical Trial Alert System on Physician Participation in Trial Recruitment" (Embi et al., Archives of Internal Medicine, 10/2005).
- "Predicting 6-Year Mortality Risk in Patients With Type 2 Diabetes" (Wells et al., Diabetes Care, 12/2008).
- "Secondary Use of Electronic Health Record Data: Spontaneous Triggered Adverse Drug Event Reporting" (Linder et al., Pharmacoepidemiology and Drug Safety, 10/2010).
AHRQ, Registries for Evaluating Patient Outcomes: A User’s Guide: 2nd Edition, 2010