OHRS Alert: dbGaP Default GSR Setting May Pose Risk to Sensitive Populations

October 03, 2019

If you have deposited genomic data into dbGaP, you received an email from the NIH with the subject line: ACTION NEEDED: Update to NIH Access Procedures for Genomic Summary Results. The email requested action by May 1, 2019 regarding NIH’s new plan to share genomic summary results (GSR) in an open-access fashion within dbGaP.

This new policy allows unrestricted access by default to GSR which are sometimes called “aggregate genomic data” or “summary data.” For most studies in dbGaP this means that GSR represent a summary of the information generated from hundreds or thousands of research participants. This may pose a risk of identification for sensitive populations.

NIH defines sensitive populations as including, but not limited to: potentially vulnerable populations (e.g., small sample sizes, isolated or identified geographic regions, Native Americans/Alaska Natives or other indigenous populations, rare disease communities) or potentially stigmatizing traits. Some examples of cohorts that may include sensitive populations are the following: HIV, some pediatric, rare diseases. When writing your grants, identifying these rare populations within your Resource Sharing Plan or Genomic Data Sharing Plan is a requirement.

The default for GSR becoming public does not put the individual-level data at risk; however, in cases where the disease under analysis may be defined as sensitive, limiting the general public to this summary data may be in the best interest of these participants. This determination is left up to the investigator and should be discussed with OHRS.

NIH believes designating data as 'sensitive' with respect to summary statistics will only be necessary in very rare circumstances.

For your datasets in dbGAP:

  • If you determine that genomic summary results from this data should not be considered sensitive, then no action is needed. The summary data became available in the open access repository after 5/1/2019.

  • If you determine that the genomic summary results from this data should be considered sensitive, as outlined above, and should not be available in an open (unrestricted) access repository, please request a new Institutional Certification with OHRS and alert the NIH of this determination.

Please don’t hesitate to contact OHRS if you have any questions about the NIH plan or your available choices.