Research In Action

The use of electronic health records (EHR) in research is growing due to their ability to systematically collect patient and population information in an electronic format. Such databases can be efficiently utilized in larger research studies because they contain diagnostic and health service information on sizeable numbers of individuals, are time efficient and relatively inexpensive to access, and can be linked to other data sets in order to examine relationships between health services and outcomes.

My colleagues and I at Children’s Hospital of Philadelphia (CHOP) recently published a study in the Journal of Attention Disorders that validated an electronic health record (EHR)–based algorithm to classify the ADHD status of pediatric patients. This internal validation study, the first study to estimate sensitivity and specificity of EHR-based ADHD diagnosis codes in classifying ADHD status, is part of a larger retrospective cohort study that aims to examine the association between ADHD and driving outcomes among adolescents. We leveraged the strength of a linked EHR system throughout a large healthcare network and found that the EHR-based algorithm was able to identify ADHD cases accurately and efficiently.

We identified all 15,000 primary care patients of the CHOP Care Network who were born between 1987-1995 and resided in New Jersey. Besides classifying data for diagnostic code, medication, and office visits specifically for ADHD, two trained abstractors performed a manual review of each patient’s entire EHR to search for independent sources of information that either confirmed ADHD or non-ADHD status.

This validated data is currently being used by CIRP@CHOP’s Teen Driver Safety Research team to examine the rates of licensure and risk of crashing among adolescents and young adults with neurodevelopmental differences to help establish the epidemiologic foundation for future translational research.

Useful Considerations

Because the EHR data from the CHOP Care Network serves a broad range of patients from various socioeconomic and racial/ethnic background and includes a variety of providers, we believe that practices, providers, and patients represented in this study likely reflect those in other large health care systems. Before conducting their own studies using EHR data, however, researchers should consider:

• Are the data contained in the EHR accurate? For example, is there variability in how ADHD or other medical conditions are diagnosed?
• Have the data been validated for your chosen exposure and outcome?
• Does the research include patients who were seen in a variety of clinical settings? Provider behaviors and patient characteristics may differ depending on type of setting, such as primary care versus specialty care.

At CIRP we are using EHR data as the foundation for our applied research in several areas of injury prevention, including traffic injury, concussion, and post-injury care and recovery. The amount of EHR data readily available for research purposes is rapidly growing and likely to substantially increase in the future.

Because EHRs offer quick access to data for the dual purpose of clinical and research efforts and the ability to feed back into one another, they are concurrently impacting best practices and helping to advance public health, including teen driver safety.