Richard Smith: The dangers of textbooks

richard_smith_2014Recently published research has found that about half of the authors of Harrison’s Principles of Internal Medicine, which is nothing short of a Bible in medicine, have received payments from 17 pharmaceutical and device companies and about a quarter hold patents. Some of these payments were considerable: authors in total had received more than $11m over four years, and one author had been paid $560 021 over a 17 month period, 90% of it for consulting. But none of these conflicts of interest were declared in the textbook. The research found that other textbooks had authors with financial conflicts of interest that were not declared, confirming earlier research. There is simply no tradition of publishing conflicts of interest in textbooks, as there is now, at long last, for articles in journals. Guidelines are beginning to follow with publishing conflicts of interest, and surely textbooks will follow eventually. But how much does it matter that textbooks don’t publish the conflicts of interest of authors?

It’s not surprising that authors of textbooks have financial conflicts of interest. Most authors of articles in journals have financial conflicts of interest, and the authors of textbooks are the same authors. What’s more, authors of a textbook like Harrison’s are by definition KOLs (pronounced to rhyme with trolls), key opinion leaders, who are valued by the pharmaceutical industry as big game hunters value tigers. A supportive word from a KOL can be worth millions, and even if they aren’t supportive they might hold back from negative comments if paid enough.

Do the financial conflicts of interest lead to biased information in textbooks? We don’t have direct evidence that they do, but we should expect that they will–because we have evidence that they do in journal articles and guidelines. Why would textbooks be different?

But I marvel that textbooks matter at all. It takes years to assemble a textbook, and much of the information they contain will be out of date by the time they are published. There is also something terribly old fashioned about the whole nature of textbooks, wisdom brought down from on high as Moses brought down from the mountain the 10 commandments. Doctors now have access online to information sources like BMJ Best Practice, UpToDate, and the Cochrane Library and increasingly decision support aids that are regularly updated. Going to look up something in a textbook is a bit like accessing the Dead Sea Scrolls for guidance.

Plus, although we don’t know how exactly conflicts of evidence corrupt textbooks, we do know that textbooks are far from evidence-based. Iain Chalmers, founder of the Cochrane Collaboration, told a House of Lords Committee that the Oxford Textbook of Medicine had continued to publish wrong advice years after it had been proven incorrect by a research study. This led to The Sunday Times running the headline “Hundreds killed by doctors relying on outdated manuals.” A famous study by Tom Chalmers and others showed that textbooks continued to publish conflicting, often wrong, advice years after it could have been known that thrombolytic agents could reduce deaths from heart attacks. 

Although often out of date and inaccurate, a successful textbook is a gold mine. Harrison’s, now in its 19th edition, costs £142, and you can also buy an accompanying book for self-assessment costing £60. It was first published in 1950 and has been translated into 14 languages. I couldn’t find data on how many copies have been sold, but it must be well over a million and might be closer to 10 million. The printing costs will be relatively small, and the authors, both conflicted and unconflicted, will not be paid much (their real payment will be the glory of being a prophet in such a sacred text), so the profits must be huge. The 20th edition will be published in August 2018. There is huge vested interest in the continuation of such gold mines, but my advice is stay away from textbooks whether or not written by conflicted authors: you could be misled, and your patients could suffer.

Richard Smith was the editor of The BMJ until 2004.

Competing interest: RS receives a pension from the BMA, which owns BMJ, which publishes BMJ Best Practice, and when editor of the BMJ and chief executive of the BMJ Publishing Group he was involved in the creation of Clinical Evidence, a forerunner of BMJ Best Practice. He was also until 2016 a chair of the Cochrane Library Oversight Committee.