For the past half century or so, academic publishers have been making ?by getting the world¡¯s best minds to give them copyrights for research that was often sponsored by public agencies. Next, without compensation, highly accomplished scientists voluntarily review and edit the articles. Then, the for-profit publishers turn around and sell the research to the universities and scientists that gave them the product and labour for free.

Now, in the internet-era pursuit of open access, production and distribution costs are a mere fraction of what they were in the days of print journals. Yet publishers are charging researchers thousands of dollars to publish papers.

None of this is news to any Times ÖйúAƬ reader. But I believe there are three reasons why academia¡¯s relationships with for-profit publishers must finally be severed.

First, the peer review system is broken. In the old days, the most accomplished experts usually agreed to evaluate papers. Now, editors report sending 15 or more requests to find two warm bodies to offer an opinion.

ÖйúAƬ

Second,?academics often can¡¯t afford those high open access fees ¨C especially faculty outside the sciences, the wealthier institutions and the developed world. This makes it more likely that journals will fill their pages with papers by authors who have money, as opposed to authors who have good ideas. Pay to play is simply the wrong model for academia.

Third, publishers have resisted repeated attempts to make their contracts with universities more transparent. A 2014 showed that the University of Michigan, Ann Arbor paid Elsevier $2.16 million (?1.77 million) for the exact same package of journals sold to the University of Wisconsin, Madison for $1.22 million. Yale, with about 12,500 students, paid Springer $711,564 for the same package that the University of Texas, Austin, with more than 50,000 students, purchased for $481,932.

ÖйúAƬ

Scientific publications need to get back to their original goal of distributing the best scientific information to the largest audience at the lowest cost. To replace the expensive, dysfunctional system, we need a national or global digital library that will edit and post peer-reviewed scientific papers.

This will require multi-institutional consortia and a substantial expansion of university libraries and professional librarians. Oversight will also be needed ¨C and could be provided by a distinguished non-profit entity, similar to the national academies. Subcommittees representing the interests of each academic discipline should define the content of the new electronic publications and appoint their editors and editorial boards.

Peer review will be similar to current practice, but databases of reviewer participation could help incentivise it. Faculty, researchers and members of scholarly societies could be expected to complete a fixed number of reviews each year as evidence of their worthiness for promotion or advancement. Those who always refuse might be excluded from submitting their own articles.

Editors could assess the quality of reviews and offer feedback to the reviewers¡¯ institutions. Not only would this incentivise better reviews, it would also help reviewers improve. Attractive bound paper journals will become relics of the past, but in their place will be downloadable pdf documents that are curated, peer-reviewed, and organised using modern ontological systems.

ÖйúAƬ

A peer-controlled system might also eliminate unnecessary burdens, such as idiosyncratic formatting conventions. As readers, we don¡¯t care about the exact style. Yet the average Canadian researcher currently 52 hours per year reformatting papers, at a cost of $1,908. And that does not include the half day that it often requires to navigate the crazymaking, illogical article submission portals.

Over the past decade, UK universities spent?more than ?1 billion on academic journals, and US institutions spent many times more. These funds could be reallocated to support the digital library. Eliminating costs for marketing, reducing duplication, and centralising administration will result in further savings. In theory, libraries could achieve full open access and still reduce their expenditures ¨C although another option would be to charge individuals and non-academic institutions a small fee for subscriptions or individual publications.??

Even if the system didn¡¯t save universities a penny, it would result in higher-quality review, wider dissemination of research, and reduced pressure for authors to pay open access fees out of their own pockets.

These ideas are preliminary and, certainly, there will be challenges. Some institutions will assume a disproportionate share of the expenses, for instance ¨C even if that might be justified on the basis of their greater usage of the library. A further problem is that if all research is completely open access, there will be little motivation for institutions to continue contributing resources. You might argue that scholarship should be considered a public good, so taxpayers should fund the digital library. But would politicians (or the public) agree?

ÖйúAƬ

Not everyone will consider the trade-offs I propose to be worth it, and I offer these suggestions only to start the conversation. But let¡¯s have that discussion soon. With a little ingenuity, we have the opportunity to reinvigorate the review process, reduce the financial burden on authors and universities, and get quality research in front of many more readers.

Robert M. Kaplan is a faculty member at Stanford University¡¯s Clinical Excellence Research Center. He is former associate director of the National Institutes of Health and former chief science officer for the US Agency for Healthcare Research and Quality. He is former editor-in-chief of two journals: Annals of Behavioral Medicine and Health Psychology.

ÖйúAƬ