Policy ForumLife Sciences Careers

A new data effort to inform career choices in biomedicine

See allHide authors and affiliations

Science  15 Dec 2017:
Vol. 358, Issue 6369, pp. 1388-1389
DOI: 10.1126/science.aar4638

The biomedical research enterprise finds itself in a moment of intense self-reflection, with science leaders, professional organizations, and funders all working to enhance their support for the next generation of biomedical scientists. One focus of their attention has been the lack of robust and publicly available information on education and training outcomes. In the absence of such information, students are prevented from making informed choices about their pre- and postdoctoral training activities, and universities from preparing trainees for a full range of careers. Today, we presidents and chancellors of nine U.S. research universities and one research institute are announcing a new initiative, the Coalition for Next Generation Life Science (1), that responds to these challenges by adopting a series of transparency enhancing efforts, the first of which is to begin reporting data in early 2018. We have agreed to start with the biomedical research arena because of the considerable attention that workforce issues in this domain have received, but the logic of our initiative extends to other scholarly disciplines.

The current tensions in the biomedical workforce are well-documented and familiar: In the United States, real levels of federal research funding declined by nearly 20% from 2003 to 2016; there has been a steady rise in the age at which the first R01 grant from the National Institutes of Health (NIH) is received for major independent research; and hypercompetition for academic positions has intensified such that only about 10% of biomedical trainees in the United States will find themselves in tenure-track positions at U.S. institutions 5 years after completing a Ph.D., although vastly more enter training in search of those positions (24). Elsewhere around the world, similar concerns are surfacing. In the United Kingdom, for example, only 3.5% of science Ph.D.'s will obtain a permanent scientific research position in a university, and national studies are voicing concern about whether the training and research system is meeting the needs of early career researchers (5, 6).


One factor that has contributed inevitably to these trends is the lack of comprehensive data collected and disseminated to prospective Ph.D. candidates on education and training outcomes in the life sciences. Although this affects a number of issues in the biomedical enterprise, it plays perhaps the most immediate role in the career expectations of biomedical doctoral students and the mismatch between the supply of biomedical doctoral students seeking academic appointments and the available tenure-track positions in the United States. This mismatch in turn has contributed to a marked increase in the number of trainees spending years in postdoctoral fellowships, with few opportunities for advancement into academic positions.

To be certain, many of these trainees enter the market for academic jobs with full knowledge of the scarcity of faculty openings. But studies show that a great many students are entering their doctoral and postdoctoral experiences without having given full consideration to the range of important research opportunities inside and outside the academy, including in industry, entrepreneurship, government, and science communication roles (7). The majority of trainees will eventually choose to pursue those careers, but only after having made irreversible investments in what is often more than a decade in training for academic jobs that do not exist (8). And at least some of this training activity may be unnecessary for their eventual career choices (9).

For these reasons, numerous reports and articles over the past two decades have called for research institutions to make meaningful data on career outcomes available to their trainees (1013). Their argument in favor of disclosure can be simply stated: Open data will allow students and postdoctoral fellows to understand fully the range of likely outcomes of their eventual training and career choices. It will help universities to better target their programs to actual career outcomes, ensuring that they are providing students the education and skills needed for the full diversity of career opportunities inside and outside academia. And it can help to hold universities and other research institutions to account for their success in training and placing graduate students. Each of these measures is directed at the cardinal goal of making advanced training in the life sciences more efficient and humane.

However, only a handful of research institutions have responded to these calls. The University of Michigan offers a rare example of an especially comprehensive platform. More recently, organizations such as the Broadening Experiences in Scientific Training (BEST) Consortium, the Council of Graduate Schools, Rescuing Biomedical Research, the National Research Council, and the Association of American Universities have acted to highlight transparency as a foundational principle, develop best practices for how institutions could collect these data, and even jump-start data collection efforts at certain universities (1417).

These are laudable and important steps. However, the publication of data is still the exception rather than the rule. Further, those universities that are posting data online publish different categories of information using a variety of reporting formats, which impedes the scaling of this information nationally. Many present only a snapshot of information rather than regular or continuous reporting that would permit trend analysis. And there are few, if any, efforts under way to collect data for postdoctoral fellows, despite the substantial length of time that many trainees spend in this setting. We are still falling far short of a norm of transparency that has been the call of countless reports.

How do we explain the reticence to act? Cost, of course, is one factor. The additional expense that institutions will need to incur in collecting, analyzing, and disseminating the data cannot be dismissed, especially in an environment of contracting NIH funding and cutbacks in state funding to higher education. There is also a natural fear on the part of institutions that they will suffer a competitive disadvantage in recruitment if they move to adopt unvarnished transparency measures and their peers do not. Institutions may worry that a frank depiction of poor completion rates or placement outcomes by some programs will penalize some of their programs unfairly against other institutions that do not disclose their data at all. And surely there are political considerations as well; as is often the case in such a complex system, the status quo serves some quite well.

We recognize that the arguments in favor of transparency are sound, and the time has come for institutions to develop credible mechanisms that are responsive to these concerns—hence our commitment to this new coalition in which we pledge to implement a number of concrete steps that will improve the collection and publication of information on our life sciences Ph.D. students and postdoctoral trainees.

The coalition seeks to build on the efforts of the above-mentioned organizations and institutions but to overcome hurdles to progress in several ways. We will hold ourselves accountable through a commitment to act against a set of agreed-upon milestones. Through conversations among peers, we are committed to working through open questions and obstacles with a goal of agreeing on common standards. And over time, we hope to establish a useful precedent that will promote easier and replicable modes of collection and publication, as well as drive down the costs and lower the barriers to this work for other institutions.

The main work of our coalition will center on the collection and publication online of several categories of data for trainees in the life sciences (see the text box). We will start by releasing the first two categories of data in February 2018, with a series of cascading milestones for the others through the ensuing year and a half. In addition to the publication of data, each participating institution has agreed to work broadly to provide meaningful career exploration and placement support for a broad array of potential career paths, improve mentorship at both the doctoral and postdoctoral stages, and increase and improve recruitment and retention aimed at diversifying the life sciences workforce.

Our hope is that these commitments will lead to a number of salutary effects. Public data on Ph.D. programs, training periods, and career outcomes will enable prospective Ph.D. students and trainees to make better decisions about their training and careers at a much earlier point than at present. Beyond that, at the individual university or institute level, open data will facilitate more responsible system stewardship. These data will, we predict, encourage institutions to make evidence-based reforms that enhance the effectiveness of their training programs, in areas such as curricula, career development programming, and faculty mentoring.

Further, programs that are under-performing relative to peers—through, for instance, high rates of attrition, unusually long education or training periods, or chronically poor placement records—will be identifiable and will provoke stakeholder demands for timely reform. Finally, there is still relatively little reliable national data on the biomedical workforce, and as more institutions participate in the effort, it will become easier to discern stress points across the biomedical research enterprise that demand national policy interventions.

In these efforts, we recall U.S. Supreme Court Justice Louis Brandeis's oft-repeated statement that “sunlight is said to be the best of disinfectants; electric light the most efficient policeman.” We do not claim that increased transparency and accountability will be a panacea for all of the challenges facing the next generation of investigators in the U.S. biomedical research system, but we believe that they do reflect an essential step in acknowledging our shared obligation to protect and enhance the vibrancy, humanity and fairness of this system. We hope that in due course other institutions will join our efforts.

References and Notes

  1. Acknowledgments: The efforts discussed in this paper are catalyzed by philanthropic investment from the Gordon and Betty Moore Foundation and an anonymous gift. These efforts are being managed by E. Watkins at University of California, San Francisco, and P. Espenshade at Johns Hopkins University.


Navigate This Article