Encouraging Team Science

The research paradigm is shifting. Once dominated by single investigators leading small groups of subordinate researchers on independent research projects, the research landscape is now dotted with interdisciplinary research teams made up of members who are often peers working together on an independent research project. Attributing proper recognition—which is key for tenure and promotion decisions—to team scientists on grant applications and awards, however, does not come easily under the traditional single principal investigator (PI) model.

Federal research agencies are recognizing this limitation. The White House Office of Science & Technology Policy (OSTP) and the Office of Management & Budget (OMB) are developing plans to put in place a mechanism that allows for multiple PIs to be recognized on a single grant. Plans for a multiple-PI model were announced in a memorandum released last January by OSTP (C&EN, Jan. 24, page 24). Since that time, OSTP, in collaboration with OMB, has continued to seek input from the research community on the details of implementing such a model.

As research becomes more interdisciplinary and federal agencies encourage more of this type of work, scientists are coming together—often with different expertise or tools—and designing projects jointly, explains Geoffrey Grant, staff director for the Research Business Models Subcommittee, which is the lead group on this issue at OSTP. And whether they are in the same institution or different ones, the scientists want to share in the credit, he says.

According to Grant, the Research Business Models Subcommittee—part of the National Science & Technology Councils Committee on Science—has been studying this issue for several years. On the basis of the subcommittees work, it became clear that developing a way to deal with multiple PIs on federal grants is an important issue in the research community.

The resulting memorandum from OSTP to the federal agencies encourages them to implement a multiple-PI policy. Since the release of that directive, Grant notes, work has focused on implementation. To that end, OSTP and OMB issued a request for information about how best to deal with multiple PIs. Grant notes that approximately 35 comments were received from a mixture of individuals and associations. The office is still reviewing the comments, and, Grant says, it expects to issue another announcement next year on the elements of the implementation at all of the agencies.

As OSTP continues to work to facilitate the process and remove barriers to implementation of a multiple-PI model across the federal system, the federal agencies are developing individual internal policies for dealing with the new model that take into account their unique needs. One agency that has been proactive in this area is the National Institutes of Health.

According to NIH Deputy Director for Extramural Research Norka Ruiz Bravo, developing a model to recognize multiple PIs is important for team-led research. She notes that at a 2003 Bioengineering Consortium meeting held at NIH, the inability to recognize more than one PI was seen as a barrier to team science and, as such, the group asked NIH to look into ways to improve the situation.

Echoing the meetings sentiment, Ruiz Bravo says developing a multiple-PI model is important for NIH because it will facilitate team science, particularly for those investigators who are either new or reluctant to get into team science because of issues related to tenure and promotion. Allowing for multiple PIs presents some challenges—including regulatory issues, process issues, and cultural issues, she says, noting that none of the barriers is insurmountable.

To help NIH develop policies specific to its needs, the agency released a request for information coincident with the OSTP request. Ruiz Bravo tells C&EN that 745 comments were received. Of these, 547 were from faculty members and 87 from institutional administrators. The bulk of the comments were very supportive of the concept of recognizing multiple PIs, she says.

One issue that both Ruiz Bravo and OSTPs Grant note as an area of concern is nomenclature. For example, according to the OSTP model, one researcher must still be designated to serve as the point of contact, termed the contact PI.

We can trip over words, Grant explains. It is important to have a contact person, but the terminology shouldnt suggest a difference in scientific stature, he says. We dont want to infer that that person is also the most senior one on the team.

This concern was addressed in the written comments from the outside community submitted to OSTP and to NIH. The proposed contact PI appears to be a clerk rather than a leader; someone appointed to facilitate communications between the research team and the funding agency and to ensure that all reports are submitted in a timely manner, wrote Anthony P. DeCrappeo, president of the Council on Government Relations (COGR), an organization of research universities. DeCrappeo pointed out that the tasks and role of contact PI as defined may likely devolve to the most junior person on a team.

In fact, the COGR comment proposes that recognizing multiple PIs would be best served by maintaining a PI and allowing co-PIs. We believe that the proposed approach is flawed, as it fails to fully recognize and acknowledge the value of scientific leadership—a principal investigator responsible for the intellectual direction and outcomes of the research, DeCrappeo wrote.

Concern for diluting the meaning of a PI was also expressed in written comments by the Federation of American Societies for Experimental Biology (FASEB), which supports the new model.

Not every coinvestigator is a PI, pointed out FASEB President Bruce R. Bistrian. It is possible that overuse of the multiple-PI designation will decrease the utility of the title, an important part of the reward system for career researchers, he wrote, adding some suggestions to be included in the PI definition to avoid diluting its meaning.

Ruiz Bravo says NIH is still working on the issue of defining not only a PI but also who—the institution or the agency—decides which investigators are designated as PIs. In addition, she notes that NIH is working through options for issues such as budget apportionment, database changes needed to accommodate multiple PIs, and continued outreach to institutions.

The OSTP plan will be released sometime next year, according to Grant. At NIH, Ruiz Bravo says its multiple-PI model will be available in some form in May 2006.

Its a wonderful thing in concept, and we are all there, Ruiz Bravo says. Its just going to take some time to work out all of the kinks.