Journal of Free Speech Law: "The NIH's Genomic Data Sharing Policy and the First Amendment," by Prof. Adam Candeub

An article from the Information as Medicine symposium.

Here's the Abstract; the full article is here:

Genome-wide association studies (GWAS) use DNA statistical analy­ses to examine the relationship between genotypic differences and phenotypic traits. Revolutionizing genetics, these studies have discovered more than 50,000 associations of genome-wide significance between genetic variants and common diseases and traits. GWAS also have transformed the study of physical anthropology, establishing the relatedness of modern and proto-humans and other primates as well as modern humans' ancient migration patterns.

The NIH's 2014 Genomic Data Sharing Policy (GDSP) governs collecting, storing, and accessing the databases upon which most GWAS research in this country relies. Many data repositories refuse access to those who pursue what the NIH categorizes as "stigmatizing" or "sensitive" research.

The GDSP does not comply with the Administrative Procedure Act (APA). The policy's "sensitive" and "stigmatizing" standard lacks any statutory basis and is perforce arbitrary and capricious.

And even assuming that the GDSP is consistent with administrative law, the policy is best viewed as a condition to obtain a government benefit or as a viewpoint-based restriction of generally available government information. So characterized, the GDSP violates the First Amendment.

Comments

[Longtobefree]

Big deal. It violates the first amendment.
Like so much of the federal government.
Wake me when the first bureaucrat goes up against the wall.

[John F. Carr]

It appears to me that some of the restriction is imposed by private actors receiving government funding.

Let's say I get a grant to collect genetic data. I sign a contract that allows NIH to hold the data and requires NIH to prohibit access to researchers who might denigrate Ashkenazi Jews. Perhaps NIH should not have approved the contract, but it did and I relied on NIH's promise when collecting the data. Can a third party come along and demand my data set to prove that Hitler was right? NIH has the technical ability to provide my data but doing so would exceed the rights I granted.

[Longtobefree]

Sorry, old son; you have no rights.

[River]

Constitutional law generally trumps contract law. If the government has the constitution saying it must allow access to the data, and a contract saying it can't, it must follow the constitution. As with all other contracts, if you are harmed by the breach, you can potentially sue for breach of contract. And since the First Amendment predates your contract, you were on notice when you signed it that the NIH's ability to comply would be constrained by the First Amendment.

[ReaderY]

But there is no constitutional right to access to non-government data, data the government doesn’t own. And if the grant of ownership to the government was a partial one, the government doesn’t own the data outside the scope of the grant. It’s not the government’s except when used for the purposes or other limitations specified in the grant. And scopes of grants to the government are determined by contract.

Let’s put look at it this way. A developer grants an easement to the government to permit inspection and collection of water use data in the homes by qualified professionals, subject to notice etc. I want to install video cameras in all the showers in the subdivision to inspect the occupants’ showering (using the water) - which is clearly data on water use by a straight dictionary definition - as part of my voyeur sex video. Can I do that?

Your argument says I most certainly can. After all, government generally has to grant access to its property to the public for first amendment purposes, and was expressly granted a right to inspect. Since the first amendment trumps contracts, comtractual limitations on government right to inspect and hence on its ability to give away that right to the public are completely ignored. Data is data. Inspection is inspection. Any concept of a PARTIAL or LIMITED right of use is contractual in nature and hence trumped by the First Amendment. And of course government can’t discriminate against me because it disapproves of voyeur videos created by installing hidden cameras in private home showerz, while approving of other kinds and uses of “water use” data. That’s obvious content-based discrimination.

It’s not like the showers were private or the residents could possibly have had any expectation of privacy. It’s very straightforward. ANY grant to the government for ANY purpose makes it wholly the government’s. Since limitations are contractual in nature, they are ignored. So since the residents bought homes subject to a grant of a right to the government to collect data on water use, contractual limitations on that data are invalidated by the First Amendment, and anyone can enter and collect data, for any purpose, whenever they damn well feel like it.

[River]

The disanalogy between your hypothetical and this situation is that here, the NIH actually has the data already. In your hypothetical, the government does not have the data. All the government has is an easement which allows it, under one dubious interpretation, to collect the data. To make your hypothetical work, the government would have to have already installed the cameras itself, recorded the videos, and have the videos sitting on its hard drives somewhere. So lets suppose all that has happened, and now your hypothetical voyeur asserts a First Amendment right to access the videos. Does the First Amendment actually cover videos of naked people taken without their consent? I would think not. Would any sane judge construe a right to collect data on water usage as consent to record videos of people showering? Again, I would think not. So in your hypothetical, the First Amendment claim fails.

More generally, applying the concept of ownership to information is always problematic. In this case, there is possession of the data, which the NIH already has. And John F. Carr has asked us to consider the hypothetical that there is a contract, to which the NIH is a party, which says that it cannot allow the data to be used for certain purposes. You can describe this as "partial ownership" if you want to, but it doesn't change the fact that what there is is possession and contract, and the First Amendment trumps a contract.

Here's the concern on the other side. If, by signing a contract with a private entity, the government could gain a legal excuse to violate the First Amendment, don't you think they'd be doing it all the time, to the point that the First Amendment would become utterly useless? I do.

[ReaderY]

Let’s do a closer analolgy. A key is simply information. analog information, about how to get into a house. Suppose the government has the key - actual information in its pisession. Must it give away that information?

Now suppose instead of an analog key there is a digital lock and the government has a digital key. On the one hand. You might feel much more comfortable calling this “information,” whoch you say has no ownership. It looks and transmits just like any other string of bits. On the other hand, it serves exactly the same function as an analog key - it enables you to get into somebody’s house.

I don’t see why the First Amendment can permit people to have property rights to their analog keys, to the point where making a copy without getting permission can be made a crime, without letting them have the same rights to digital keys. And I see no meaningful difference between access codes and other digital information people don’t want made public.

[Roger S]

Rules like these make it difficult to find genetic causes for high blood pressure, and a lot of other legitimate medical research.

[John F. Carr]

There was an idea floating around that the harsh conditions on slave ships selected for people with genes that cause high blood pressure. Slaves with normal blood pressure or normal salt balance were less likely to survive. I don't know if people still think that.

[Armchair]

This feels like a bridge too far. Especially the stretch to arbitrary and capricious from sensitive and stimatizing....

Let's start with this. This is a government database. The government has many databases. Some are freely accessible. Some are highly restricted. And some are in between. I don't think any reasonable individual would believe that all government databases should be freely accessible to any individual.

The GWAS database would fall in the middle. An application process is needed to access the database. The information is potentially sensitive. While very useful for medical research, it also has the potential to be misused in potentially severe ways. In an especially severe case, a researcher might want access the database to help design a virus that is selective against certain ethnicities. Or a different researcher may somehow develop an algorithm that can somehow take the genomes of the individuals and use that to identify individuals. Neither something the NIH wants to help along. So, a certain amount of caution is called for. The NIH has given itself a legal out in those cases with the "stigmatizing or sensitive" part. And each application would be evaluated on a case by case basis.

It's hard to see how an evaluation on a case by case basis is "arbitrary and capricious"...especially in the absence of any case examples. Furthermore, let's assume the authors succeed in a first amendment challenge. The most likely response by the NIH would be to further, more heavily restrict access to the database. If they're required to give it to the public, no matter what causes it might be abused for...instead, it would "only" be used by NIH researchers, and everyone else would be shut out. It would seem to be counterproductive that way.

[ReaderY]

Is this a government database? If data was given to the government subject to use limitations, the governemnt does not own the data except for purposes within those limitations.

Property is a bundle of rights. Concepts of divided and limited rights occur all over the place in the law. A grant of an easement on land to the government for a particular purpose doesn’t make it “government land” whose uses somehow cannot be constitutionally restricted. Outside the scope of the grant, it just ISN’T the government’s land.

Nothing in the First Amendment prohibits ordinary contract and property principles, including limited grants to the government subject to restrictions of various kinds, from being applied to data.

[MaverickNH]

What they mean is, you can’t have access to the data unless you demonstrate that your work furthers DEI and that you’re employing diverse people to do the work, etc. And if you’re asking any questions about race, ethnicity, gender that might not support that it’s all nurture and none nature, you can’t have the data. That, and if we give you the data and you stumble on any findings that suggest it’s not 100% nurture (including negative nurture forms like privilege, supremacy, etc.) - bury it or you’ll never publish again.

[John F. Carr]

A new NSF rule also puts political restrictions on research:

Effective 20 May, any scientist doing research “that may impact the resources or interests of a federally recognized American Indian or Alaska Native Tribal Nation” must obtain “prior written approval” from that nation before NSF will make an award.

https://www.science.org/content/article/new-nsf-rule-requires-tribal-approval-research-affecting-their-interests

The rule will also cut down on redundant climate change research. Alaskan locals are suffering "research fatigue" due to too many teams visiting to to ask "how do you feel about seas rising to flood your village?"