January 5, 2016

Precision medicine needs a population approach


Dr. Eric B. Larson says a new JAMA study of arrhythmia shows the power of research with big, diverse, real-world populations such as Group Health

The year 2003 was a milestone in human genetics. An international consortium, including scientists here in Seattle, finished sequencing the human genome. That accomplishment created a database of all the information in our chromosomes’ DNA—which determines characteristics ranging from hair color to risk of disease. 

As researchers and physicians, we were thrilled about the human genome sequence. It created the promise of “personalized” or “precision” medicine: tailoring treatments based on a person’s traits, including individual genetics, instead of choosing therapies based on what we know about the entire population of humans. Now, a new paper in JAMA (Journal of the American Medical Association) shows the power of applying big, broad, population-based analysis—like we often do at Group Health Research Institute—to resolving the complexities of human genetics.

A search for arrhythmia genes

We know that some variants of human genes directly cause disease: examples are cystic fibrosis, sickle cell anemia, and Huntington disease. The causes of other conditions are more complicated. For example, certain specific variants of the genes SCN5A and KCNH2 appear to cause rare heart arrhythmias, which can be lethal. Other versions of these genes might also increase risk of arrhythmia but we don't know if this is true. The JAMA study from eMERGE (Electronic Medical Records and Genomics), a National Institutes of Health-supported network that uses genome databases and electronic medical records for health research, looked for additional variants of SCN5A and KCNH2 that might identify people at high risk for arrhythmias.

Group Health Research Institute, in partnership with the University of Washington, is a founding eMERGE member, so Group Health volunteers and others from around the country contributed DNA samples to the study. New variants of the two genes were identified in the study population. Electronic medical records were used to find out if these variants were linked to heart rhythm problems. And this is where it got complicated. Many people with a variant in one of the genes, including some followed past age 90, never had arrhythmia. This evidence showed that the newly identified gene variants were not good indicators of arrhythmia. The predictability of the genes was not as great as we expected. 

The power of big, diverse populations

Is this result a blow to personalized medicine and its promise of more precise diagnosis and treatment? I don’t think so. Instead, this study shows the value of research using large, diverse, real-world populations. By studying the additional SCN5A and KCNH2 gene variants in a broad population, we could see that some variants appeared in people with arrhythmia, but also in people without the condition. The JAMA study shows the importance of working with an unselected group of patients rather than focusing too narrowly on people with a specific condition.

We use this approach frequently in our studies because we are fortunate to have access to information from Group Health members, who are generally very open to participating in research. By taking this approach, we made a somewhat unexpected discovery in this eMERGE project: The best way to advance precision medicine that focuses on the individual may be to study a big population.


by Eric B. Larson, MD, MPH, Executive Director, Group Health Research Institute; Vice President for Research, Group Health