One of PCORI’s goals is to improve the methods that researchers use for patient-centered outcomes research. PCORI funds methods projects like this one to better understand and advance the use of research methods that improve the strength and quality of comparative effectiveness research.
This research project is in progress. PCORI will post the research findings on this page within 90 days after the results are final.
What is the research about?
Researchers can use data from electronic health records, or EHRs, to identify patients who are at risk for certain diseases. Diagnostic codes are one way that EHRs record data on patients’ health problems. But these codes are not always correct. Combining different types of information from EHRs, such as prescriptions and lab test results, may be more accurate than using diagnostic codes alone.
In this study, the research team is developing methods to combine different types of information from a patients’ EHR to estimate latent phenotypes. A latent phenotype is the researcher’s best guess at a patient’s characteristics, such as the presence of a health problem, based on all available information.
Who can this research help?
This research may help researchers obtain accurate patient data from EHRs to help them identify patients to include in research studies and to help measure the effects of treatments.
What is the research team doing?
First, the research team is developing statistical methods to estimate latent phenotypes from EHR data. Second, the research team is developing ways to include this information on latent phenotypes in analyses to improve analyses that look at health outcomes. The team is also creating software for using the new methods. With computer simulation, the team is comparing the new methods to existing EHR-based methods. The team is testing the new methods using several different ways of handling missing data.
Third, the team is using the new methods to estimate a latent phenotype for type 2 diabetes in children. Using data from eight children’s hospitals, the team is looking to see how this phenotype is associated with body mass index in early life. Clinicians, patients, and parents are working with the research team to choose which health conditions to study.
Research methods at a glance
|Approach||Statistical methods development, software development, simulation studies|
- Douglas Lunsford (parent), Nationwide Children’s Hospital
- Drew Wittman
Other Stakeholder Partners
- Nikki Highfield, Cincinnati Children’s Hospital
- Rohan Henry, Nationwide Children’s Hospital
- Robert Siegel, Cincinnati Children’s Hospital
- Monika Chaudhari, Nationwide Children’s Hospital