November 2012

A. Background:

1. What is the National Institutes of Health (NIH) GWAS Policy?

This is a data sharing policy implemented in 2008 for NIH-funded Genome-Wide Association studies, or GWAS.  (See GWAS Policy). The policy establishes a NIH Database of Genotypes and Phenotypes (dbGAP), which has two components.

• An open access portion that will be freely available to the public and will include:
  • The protocol;
  • Questionnaires;
  • Variables measured; and,
  • Other supporting documentation
• A controlled access portion that will only be available to researchers who have been approved by an NIH Data Access Committee (DAC). This portion will include coded phenotype, exposure, genotype and pedigree data, and summary statistics.

2.  What is a Genome Wide Association Study?

• A GWAS study is defined in the NIH policy as “any study of genetic variation across the entire human genome that is designed to identify genetic associations with observable traits or the presence of a disease or condition.” 
• A GWAS study is typically one in which 100,000 or more SNP markers are tested in individual DNA samples, to produce a “high density” genomic profile.

3. When did the policy go into effect?

 The GWAS Policy went into effect for competing funding applications and new funding proposals submitted to NIH on or after January 25, 2008.

4. To whom does it apply?

The policy applies to investigators who receive any NIH funding for genome-wide analysis of specimens. Of note, unlike the general NIH data sharing policy that has a threshold of $500,000, the GWAS policy has no threshold; therefore, any NIH funding for GWAS triggers requirements under the policy.

B. Requirements for IRB Review:

1. Is the project covered by the GWAS policy?

Yes, if:

• The PI will obtain NIH funds to conduct GWAS studies through a new application or continuing funding application submitted after 1/25/2008; or
• The PI has agreed to voluntarily submit genotype or phenotype data to the NIH GWAS registry.

2. What documents must the PI submit for IRB review and certification?

• The PI must submit an application in eIRB (or amend an existing application) which indicates a request for GWAS submission (currently eIRB section 11 – Sample Size and Source, Q. 12.0).
• The PI must upload into the eIRB application a form which details the following:
  • A description of all data fields (genotype and phenotype) being submitted to the GWAS repository;
  • A description of the method(s) to be used to code data for transmission to the NIH;
  • A description of how the code key(s) will be maintained by the PI;
  • A written confirmation by the PI that the code keys will never be shared with the NIH;
  • A description of any limitations on use of data or samples (e.g., subjects signed consent forms stipulating use of their data or specimens only in particular fields of research);
  • A statement as to whether whole genome or whole exome sequencing will be performed on any samples being submitted; and,
  • The name of the applicable Project Officer at NIH.
• If the submission to the IRB is an amendment to an existing study and will involve consent forms previously approved and signed by JHM subjects without GWAS-specific language, the PI must provide the IRB with a copy of each version of the consent form.  The form(s) must be uploaded in the eIRB application.  To facilitate review and approval, the PI may wish to provide an opinion to the IRB concerning whether the language in any previous version of the consent form(s) would prohibit a submission of subject data to NIH.

C. IRB Review Process:

1. For any submission of phenotype data and genotype data to an NIH GWAS repository, the JHM IRBs must determine the following:

• All the information to be submitted meets the HIPAA standard for de-identification;
• GWAS studies of these data do not pose any risks (e.g., stigmatization) to particular populations or groups; and,
• Specimens and data were or will be collected in compliance with 45 CFR 46 and state law;

2. For the submission of data or specimens from studies previously approved by the JHM IRBs (in addition to the items in number 1, above), the IRBs must determine that existing or past versions of the consent forms are not inconsistent with sharing genotype or phenotype information with the NIH GWAS repository.  

• In reviewing this requirement, the IRB should consider the following questions:
  • Do any of the consent forms contain statements such as “your data/specimens will not be shared” or “will only be seen by the research team?”
  • Were any of the specimen donors minors?
  • Did the original consent forms limit future use to specific projects or disease states, or to non-commercial research?
  • Were any of the data (e.g., medical record information) collected under a waiver of consent?

(See section D, below for information on how the IRB may address some of these findings).

3. For the submission of prospectively collected specimens or data (in addition to the items in number 1, above) the consent form in the new protocol application must:

• Describe the GWAS repository and the risks associated with this research, and
• Describe the likelihood that re-identification might become possible in the future.

(Please note that language sufficiently addressing these items may be found in the most current version of the JHM IRB’s consent form template).

D. What if the IRB determines that the original consent forms are not consistent with the requirements for sharing, that no research consent forms were signed, or the risk of re-identification seems too high?

1. If the research consent was not consistent with GWAS sharing requirements or there are no research consent forms (e.g., samples were obtained under a surgical consent or waiver of consent) subjects must be contacted and consented before the PI may submit their data to the NIH GWAS repository.  If appropriate, the IRB may approve an amendment to allow the PI to re-contact subjects to seek their re-consent, or, in limited circumstances and when consistent with regulations, the IRB may waive consent to allow submission of data. 

2. If the consent forms impose use restrictions that are inconsistent with broad data sharing, the IRB may impose use restrictions upon the data submitted to the NIH, (e.g., a restriction to specify that the data may be disclosed by NIH only for the study of a particular disease or only for non-commercial research).

3. To reduce the risk of re-identification, the IRB may limit the types or fields of phenotype data that the PI may submit to the GWAS repository.

E. Certification Letter

When the IRB has completed its review and made all necessary determinations to allow for sharing, JHU will issue a certification letter as requested.  This letter will be sent to the PI, who must then send it to NIH to satisfy the requirements of the GWAS policy.  Certification letters will be reviewed by IRB Counsel and signed by the Vice Dean for Clinical Investigation.

For studies in which Hopkins serves as the lead or coordinating site in a multi-institutional effort, the JHM IRBs and the Institution will not make certifications for data or specimens to be submitted to an NIH GWAS repository from cooperating sites.  It is incumbent upon Hopkins investigators who serve as lead or coordinating PIs to maintain documentation from each individual site’s respective IRB that a certification has been made.