OA at TACD IP

I previously liveblogged the sessions from Patents, Copyrights and Knowledge Governance: The Next Four Years, hosted by Trans Atlantic Consumer Dialogue (Washington, DC, January 12-13, 2009). Now I want to go back through the notes (which are cursory and, to others, potentially cryptic) and draw out one particular area of discussion, that of open access to research publications and data. (Note that this summary is based on my notes, memory, and impression, so I apologize for any errors.)

Two panels (at least) mentioned it explicitly: the panels on Openness and on Innovation and Access for Medical Technologies. I’ll try to summarize what was notable:

Openness: Tim Hubbard of the Wellcome Trust:

• Some funders are starting to require that their grantees develop data access policies.
• There is an effort to develop a “PubMed Commons”, building on top of PubMed Central, adding e.g. features for comments and tags.
• Recommendation: Policies for OA to publications and data should become standard for research funders.
• There are some problems remaining:
  • OA to publications: name control (e.g. recognizing that “Doe J Q” is the same person as “John Q. Doe”), permission problems (e.g. barriers to data mining)
  • OA to data: cultural attitudes (e.g. how to give credit/attribution, adjusting to new competitive environment — people interpreting your data before you can), practical problems (e.g. format standardization)
• The focus of the presentation was on the conflict between openness (e.g. OA to genetic data) and privacy. Even supposedlu de-identified data can be re-identified, posing a challenge to OA to data on human subjects. He suggested a solution where researchers could have free access to the complete data (with full granularity), but without handing it out in raw form: researchers could load programs onto a trusted broker’s computer, which would execute the algorithm and provide anonymized results. This would not be a perfect protection of privacy — there is still the possibility of leaks and ways to game the system — but it would make it easy enough for legitimate researchers and hard enough for malicious uses so as to discourage most of them.

Comments:

• How many funders actually have open data requirements for their grantees? This seems like a good topic for the Open Access Directory: collecting information on funder policies, as well as on projects and institutions with open data policies for their intramural researchers. This is such a nascent space that there’s a big role for sharing best practices here.
• I hadn’t previously heard of PubMed Commons — I’d certainly like to hear more about it.
• It was striking how much he presented OA to publications as a done deal — even the problems he mentioned had to do with leveraging OA for extra value, rather than the problem of getting OA to begin with. (This is especially striking since Wellcome’s own mandate after 8 months still had less than 30% compliance. I haven’t seen more recent numbers, but Wellcome announced they would start audits to increase the compliance rate.)
• Privacy is an important question when it comes to sharing data on human subjects. Unfortunately, human subjects is one of the areas where data is most critical to share, e.g. to better understand epidemics and more quickly develop responses. Hubbard’s proposal is an interesting one, and it’ll be interesting to see if anyone pilots such a model. At the same time, I think a dedicated person probably could still game such a system where motivation is high, e.g. seeking private data of a celebrity or public figure. Aside from the general damage of such breaches, they set back the research enterprise itself because people are less willing to share data with researchers. So I think there’s room for an additional layer of security, where researchers have to apply for access (which can be revoked if the system is misused). This will better give the public a sense that their data isn’t just “out there”, but there are legitimate filters to keep out bad actors. Inevitably there still will be breaches and leaks, but it’ll seem like more is being done to prevent them. There will also need to be legal and cultural elements to this too (as Hubbard also suggested), such as an effective framework for redress when private data is abused. It’s also important that the authorization system not evolve into something exclusionary: any legitimate research by a trustworthy investigator should be promptly granted access. Unfortunately, the structure is then already in place for charging subscriptions; particularly where databases are expensive to generate and maintain, there may be a temptation to do so.

Openness: Heather Joseph of SPARC:

• There is a growing but limited awareness among policymakers of the benefits of openness. We need policies to remove both access and permission barriers to research. So far most uptake of OA policies has been in the biomedical arena. There’s pushback from opponents even there.
• There’s a recognition that data is valuable and needs to be appropriately managed. But we need to move from mere “data management policies” toward open access to data.
• So far, the signs from the Obama administration are promising. The administration espouses openness (in general) and some key appointments are strong.
• Recommendation: Make OA policies throughout the government, e.g. through executive order or regulation, but not as part of IP policymaking.

Comments:

• I think this is the first time I’ve heard the suggestion that the U.S. adopt OA policies for its publicly-funded research via executive order or regulation. There’s some sense to this, I think: it has to do with the internal management of the government and doesn’t run afoul of any legislation. I see two challenges, though. First, other funding agencies don’t have the pre-existing cyberinfrastructure into which grantees could deposit their articles; funding for that has to come from somewhere (e.g. through the appropriations process). Second, as we saw with the “Fair Copyright in Research Works Act”, there’s some opposition to these policies in Congress: some members might take issue with executive policymaking — but I think that’s a fight we could win, especially with the administration on our side.

Openness: Questions:

• Bruce Perens: What about other IP on publicly-funded research, e.g. Bayh-Dole? Joseph: We need to draw a distinction — we’re focusing on access to the literature and data.
• Jamie Love: If there was an access to knowledge treaty, what specific issues would you want to include? Hubbard: OA data and literature; mandated protection of private data; greater openness to economic data; standards for government data. Joseph: OA literature and data.
• Jonathan Band: Some countries hold copyright on government works (Crown copyright — unlike the U.S.) — this needs to change.

Comments:

• There’s a little surge of discussion lately about patents (and the management thereof) on publicly-funded research, e.g. owing to developments in India and South Africa. It’s an issue I’ve been involved with before, and I’m sympathetic toward efforts to improve the current system. There are some common principles, but they’re also very different questions. Specificity, so we can understand well what we’re discussing, is important in policy, and especially in these areas, which are often so challenging to understand. So: OA can advance without affecting the Bayh-Dole system; the Bayh-Dole system doesn’t necessarily do harm to OA.

Innovation and Access for Medical Technologies: Anthony So of Duke University:

• Recommendation: Public access to full clinical trial protocols and regulatory submissions.

Innovation and Access for Medical Technologies: Judit Rius of Knowledge Ecology International:

• Recommendation: Innovation inducement prizes should include an “open source dividend” to encourage open sharing of knowledge, data, material, and technology
• Recommendation: Governments should support the World Health Organization’s Global Strategy and Plan of Action on Public Health, Innovation and Intellectual Property (as adopted in World Health Assembly resolution 61.21), including open access to scientific articles and data from publicly-funded research

Comments:

• The FDA Amendments Act requires the registration of clinical trials and reporting of some basic results on the OA site ClinicalTrials.gov. But the “basic results” really are pretty basic — there’s room for broader disclosure (and OA) here.
• KEI’s “open source dividend” idea is an interesting one — which may have application outside the area of inducement prizes — and one I’d like to hear more about.
• OA advocates shouldn’t forget about the WHO’s Global Strategy, which includes very favorable language toward OA. It’s worth mentioning when speaking with funders, policymakers, researchers, etc. Unfortunately, the language as adopted in WHA 61.21 only said that states should “strongly encourage” publicly-funded researchers to make their manuscripts OA, retreating from earlier language that states should “require” it. Still, as we saw with the NIH policy — maligned as the earlier voluntary policy was (and rightly so), it laid the groundwork for a mandatory policy (establishing a repository, starting to educate researchers, gathering evidence on the policy’s impact and effectiveness, building a constituency in favor of OA, etc.). We have a document representing global consensus that governments should support OA: we should remind people of that more often.