Harold Varmus (NCI/PLoS): Changing the Way We Publish
Harold Varmus: "Changing the Way We Publish" I'm Harold Varmus. I work as a cancer biologist. I've worked in various academic institutions and the government. And for the last several years, I've been concerned about the way that scientists present their work to each other and to the public, through publication. Publishing reports of our scientific work is one of the most important aspects of a scientists' career. It is the way in which we let others know about the work we've done, the discoveries we've made. It influences chances of getting a job, keeping a job, getting promoted. It influences our change of being evaluated for scientific societies and for prizes. It's the way in which we contribute our knowledge to the next generation. Still, despite the importance of this aspect of our work and the power of the internet to change the way we publish our work, the scientific community has generally stuck to a publication system that hasn't changed very much since it was invented about 350 years ago with the first publication by the Royal Society of London and others. I first began to think about this, about the failures of our publication system one morning at breakfast in December of 1999, when as then the director of the NIH, I was having a late breakfast with a close colleague, Pat Brown, from Stanford. Pat had recently learned that the physicists of the world had united to present their work in the form of e-prints on a website called "ArXiv" and their work would then be exchanged easily and freely among anyone who wanted to comment and would then get published in the traditional physics journals. And while that was not, to our mind, an ideal model for those of us who do work in biology and medicine, it made us think about the way in which the internet could be used more fruitfully to promote the value and dissemination of scientific work. I then thought more about this, and wrote an essay called "e-Biomed", in which I propose that the NIH act as a catalyst for ensuring that scientific work is made much more available through the creation of a repository, a library, that could be seen by anybody.
And that lead to the development of, in conjunction with David Lipman at the National Library of Medicine, of a full scale, full text library called "PubMed Central". Many of you are probably familiar with PubMed, which represents the authors, journal name, and abstract, of almost anything that is published in our fields, but for those who want to see the full-text of an article and the data, that required a subscription to the journal, or use of a library. Now for many scientists, seeing that kind of information is not difficult because their institutions use their library fees to publish such information and make it available. But for people who are working in less affluent institutions, who are working in other parts of the world where access to subscriptions are not as available, for members of the public who are interested in scientific work, disease advocacy groups, high school students, journalists, general members of the public who have an interest in science, even though the work is supported largely by the government, and not-for-profit foundations it's frequently very difficult to see the work and to help evaluate it and learn what scientists are doing. So we proposed at the time that we'd form a group called the "Public Library of Science" that would advocate for the creation of large public digital libraries that would host all of the world's work on biology and medicine. Well we declared that interest and the Library of Medicine was helping us but not very many journals were willing to participate because they were afraid that by contributing their articles to this public digital library, even several months or a year after publication, that they would lose the subscriptions that sustain the journals, and clearly journals can't work without money. So we came up with another approach to this question, which entailed the use of a model for publishing called "open-access" publishing.
In traditional publishing, when an article is submitted to a journal, and accepted, the author transfers copyright to the publisher, the publisher charges all readers for subscriptions to the journal and access to it, and that is how the cost of reviewing and editing and printing and so forth are covered. In the new model, the open-access model, the journal would allow the author to retain their own copyright, that the articles would be submitted and published immediately online with access to all and the cost of doing that would be covered by a fee paid by the authors, which means in practice, paid by those who sponsored the scientific work, usually government funding agencies or not-for-profit agencies, and then everybody would have access to, and the ability to use that work as long as there continued to be acknowledgement of the source of the information. The Public Library of Science became, in a word, a publishing house as well as an organ of advocacy. And over the next 7 years, from 2003, when we first began to publish PLoS Biology to the present, when we have published at least seven journals, two highly prestigious journals called PLoS Biology and PLoS Medicine, several so called specialty or community journals in Microbiology, Computational sciences, and other fields, and a very large-scale journal called PLoS One, that publishes in eventually all fields. We have, first of all, become a major publisher in the biomedical sciences. We have established the credibility of this new business model and cultural model for publishing, open-access publishing, and we have become financially self-sustaining. Indeed, in this year, 2010, we will be making a modest profit that will allow us to invest in some new adventures. In view of this 7 year period of dramatic growth and increasing financial stability we have begun to think about what can be done beyond developing the model of open-access publishing to increase the utility of the way in which scientists work through the internet with new information generated by our communities.
In particular, we're thinking about ways in which to make publishing faster, right now as everybody who has experience in publishing in traditional journals know, the publishing process is slowed by a very, some would argue excessively, stringent review process that gives to 2 or 3 people the power to limit the ability of someone to publish their work by requesting more and more additional experiments, that impede the ability of scientists to record their work in a journal. We think that science should progress more rapidly, and one way to do that is to provide a faster means of review that is less stringent initially but a method that transfers responsibility for evaluation to the scientific community at-large once a paper has appeared on the web, and that can be done by making post-publication comments online, and exactly how we do this is still a matter under discussion. The second thing we are interested in is the nature of the scientific article. Traditionally, articles have been organized with abstracts, introductions, result sections, materials and methods, and discussion, all in a narrative sequence, but the electronic publishing changes the way one can envision a paper. One can think about a layered approach in which a summary of the paper can be dug into by the reader to go deeper and deeper on topics of interest and individual components of papers can be seen as individual items of research, results, and data sets. So we are interested in investigated the way in which one can disaggregate a paper. Third, we're interested in the general culture of publishing, and the way in which publishing is viewed by the scientific and academic community. As someone who has run institutions, I know that in the promotion and appointment process, it's frequently the case that an argument for appointment and promotion is based, not on the content of work, but instead on the number of papers published in journals that are viewed as the most revered by scientists.
This seems to me an undermining of the traditional mode of evaluation by scientists, which should be carried out by an engagement in the nature of discoveries, the kind of work, the quality of presentation, something that can only be achieved by having scientists look directly at that work and evaluate it not by the impact factor of a journal in which the work appeared, but instead by looking closely at the work itself. This is a manner of curation, and appreciation of the scientific record that we need to return to. These cultural issues are particularly important for students and post-docs who are trained to be scientists, because I frequently hear that young people are interested in publishing in a way that is beneficial to the greater scientific community. That means publishing in a way that provides access for everybody. And yet there is a reluctance that is passed on by those who hold power in the scientific community, to consider publishing in journals other than Cell, Nature, and Science, the so-called CNS triumvirate that seem to have the greatest clout because of their high impact factor. It's important that we move away from that. We move away from it by saying, I'm going to publish my papers in places that have the greatest impact on the society of science. I'm going to make a declaration of the importance of disseminating my work to the widest possible audience, and I'm going to upset some of the traditional standards by which people are evaluated. In my own institution, we insist that the candidates for promotion to medium level and senior positions, say here are my 5 most important discoveries, and the members of the review committee need to read those papers, not just say that they were published in journals of a certain stature. Being aware of these cultural icons and being willing to challenge them is a very important aspect of life in sciences and I urge all of you who are in positions of training to think about the cultural manifestations of what we do, and to challenge those that seem wrong to you.
And that lead to the development of, in conjunction with David Lipman at the National Library of Medicine, of a full scale, full text library called "PubMed Central". Many of you are probably familiar with PubMed, which represents the authors, journal name, and abstract, of almost anything that is published in our fields, but for those who want to see the full-text of an article and the data, that required a subscription to the journal, or use of a library. Now for many scientists, seeing that kind of information is not difficult because their institutions use their library fees to publish such information and make it available. But for people who are working in less affluent institutions, who are working in other parts of the world where access to subscriptions are not as available, for members of the public who are interested in scientific work, disease advocacy groups, high school students, journalists, general members of the public who have an interest in science, even though the work is supported largely by the government, and not-for-profit foundations it's frequently very difficult to see the work and to help evaluate it and learn what scientists are doing. So we proposed at the time that we'd form a group called the "Public Library of Science" that would advocate for the creation of large public digital libraries that would host all of the world's work on biology and medicine. Well we declared that interest and the Library of Medicine was helping us but not very many journals were willing to participate because they were afraid that by contributing their articles to this public digital library, even several months or a year after publication, that they would lose the subscriptions that sustain the journals, and clearly journals can't work without money. So we came up with another approach to this question, which entailed the use of a model for publishing called "open-access" publishing.
In traditional publishing, when an article is submitted to a journal, and accepted, the author transfers copyright to the publisher, the publisher charges all readers for subscriptions to the journal and access to it, and that is how the cost of reviewing and editing and printing and so forth are covered. In the new model, the open-access model, the journal would allow the author to retain their own copyright, that the articles would be submitted and published immediately online with access to all and the cost of doing that would be covered by a fee paid by the authors, which means in practice, paid by those who sponsored the scientific work, usually government funding agencies or not-for-profit agencies, and then everybody would have access to, and the ability to use that work as long as there continued to be acknowledgement of the source of the information. The Public Library of Science became, in a word, a publishing house as well as an organ of advocacy. And over the next 7 years, from 2003, when we first began to publish PLoS Biology to the present, when we have published at least seven journals, two highly prestigious journals called PLoS Biology and PLoS Medicine, several so called specialty or community journals in Microbiology, Computational sciences, and other fields, and a very large-scale journal called PLoS One, that publishes in eventually all fields. We have, first of all, become a major publisher in the biomedical sciences. We have established the credibility of this new business model and cultural model for publishing, open-access publishing, and we have become financially self-sustaining. Indeed, in this year, 2010, we will be making a modest profit that will allow us to invest in some new adventures. In view of this 7 year period of dramatic growth and increasing financial stability we have begun to think about what can be done beyond developing the model of open-access publishing to increase the utility of the way in which scientists work through the internet with new information generated by our communities.
In particular, we're thinking about ways in which to make publishing faster, right now as everybody who has experience in publishing in traditional journals know, the publishing process is slowed by a very, some would argue excessively, stringent review process that gives to 2 or 3 people the power to limit the ability of someone to publish their work by requesting more and more additional experiments, that impede the ability of scientists to record their work in a journal. We think that science should progress more rapidly, and one way to do that is to provide a faster means of review that is less stringent initially but a method that transfers responsibility for evaluation to the scientific community at-large once a paper has appeared on the web, and that can be done by making post-publication comments online, and exactly how we do this is still a matter under discussion. The second thing we are interested in is the nature of the scientific article. Traditionally, articles have been organized with abstracts, introductions, result sections, materials and methods, and discussion, all in a narrative sequence, but the electronic publishing changes the way one can envision a paper. One can think about a layered approach in which a summary of the paper can be dug into by the reader to go deeper and deeper on topics of interest and individual components of papers can be seen as individual items of research, results, and data sets. So we are interested in investigated the way in which one can disaggregate a paper. Third, we're interested in the general culture of publishing, and the way in which publishing is viewed by the scientific and academic community. As someone who has run institutions, I know that in the promotion and appointment process, it's frequently the case that an argument for appointment and promotion is based, not on the content of work, but instead on the number of papers published in journals that are viewed as the most revered by scientists.
This seems to me an undermining of the traditional mode of evaluation by scientists, which should be carried out by an engagement in the nature of discoveries, the kind of work, the quality of presentation, something that can only be achieved by having scientists look directly at that work and evaluate it not by the impact factor of a journal in which the work appeared, but instead by looking closely at the work itself. This is a manner of curation, and appreciation of the scientific record that we need to return to. These cultural issues are particularly important for students and post-docs who are trained to be scientists, because I frequently hear that young people are interested in publishing in a way that is beneficial to the greater scientific community. That means publishing in a way that provides access for everybody. And yet there is a reluctance that is passed on by those who hold power in the scientific community, to consider publishing in journals other than Cell, Nature, and Science, the so-called CNS triumvirate that seem to have the greatest clout because of their high impact factor. It's important that we move away from that. We move away from it by saying, I'm going to publish my papers in places that have the greatest impact on the society of science. I'm going to make a declaration of the importance of disseminating my work to the widest possible audience, and I'm going to upset some of the traditional standards by which people are evaluated. In my own institution, we insist that the candidates for promotion to medium level and senior positions, say here are my 5 most important discoveries, and the members of the review committee need to read those papers, not just say that they were published in journals of a certain stature. Being aware of these cultural icons and being willing to challenge them is a very important aspect of life in sciences and I urge all of you who are in positions of training to think about the cultural manifestations of what we do, and to challenge those that seem wrong to you.