Why I published “the albumin paper”: confession of a buccaneering editor Richard Smith Editor, BMJ October 2001.

Why I published “the albumin paper”: confession

of a buccaneering editor

Richard Smith

Editor, BMJOctober 2001

Hypothesis One• Editors are shadowy, wayward pictures who

prefer the dark to the light and are happiest consorting with “les belle de nuits”

• They love sensation, any sensation

• Nothing gives them more pleasure than to upset solid, upstanding people like intensivists

• The “albumin paper” provided a momentary fix to feed these dubious pleasures

Hypothesis two

• The paper asked an important question• The methods were “good enough”• ASIDE: the invention of the “good enough

mother” may be one of the greatest inventions of the 20th century; concepts of the “good enough editor” or “good enough intensivist” follow

• The paper was suitably tentative, even if some of the subsequent comments were not

What I want to talk about

• The new world of evidence based practice

• My version of the albumin story

• Intensive care in an evidence based world

Is evidence based practice a radical

change?• Combines with other drivers of change• “Consumerism”; the resourceful patient• The arrival of the internet• The desire of owners to manage more the clinical

process• Growing gap between what could be done and what

can be afforded

•

Has EBP changed the world?• Source of knowledge

is expert opinion

• Clinical skills are seen as semimystical

• Research is marginal to practice

• Source of knowledge is systematic review of evidence

• Clinical skills can be audited and managed

• Research and evidence go together

Has EBP changed the world?

• Analysis of research is haphazard

• Not important to gather new evidence from patients routinely

• Analysis of research is systematic

• Patients should be included in trials wherever possible


• Only lip service is paid to keeping up to date and learning new skills

• Most medical care is assumed to be beneficial

• Essential to keep learning new skills

• Widespread recognition that the balance between doing good and harm is fine


• Clinical performance is not systematically audited

• Managers have little involvement in clinical proceses

• Clinical performance is regularly reviewed and managed

• Managers are involved in clinical processes


• Organisational model is hierarchical

• Doctor patient relationship is essentially master/pupil

• Organisational model is much more democratic, based on ability to use evidence

• Patient partnership is the norm


• Patients do not have easy access to the knowledge base of doctors

• The doctor is smartest

• Patients have as much access to the evidence base of medicine as doctors

• Often the patient is smarter

The albumin story: my view

• Albumin has been used for 50 years to treat the critically ill

• The “theory” behind the treatment was based on hypoalbuminaemia being associated with higher mortality (undoubtedly true), oedema, and “low serum oncotic pressure”


• The theory was: “X being low is bad: giving X will be good”: Is this simpleminded?

• Like surgical “theory”: something in the body is bad: if we cut it out things will be better: result radical mastectomy; hemicorporectomy


• “Respectable” intensivists had doubts about the effectiveness of albumin

• “Currently, the widespread use of albumin has more to do with word association and the treatment of items that are marked on a pathology form with an asterisk than with scientific medical management.” Neil Soni, BMJ, 1995


• There was big worldwide variation in the use of albumin: generally, Commonwealth countries used it a lot; Americans used it much less

• The story is complicated (as always) by commercial factors: albumin is expensive, and many peoples’ jobs depend on it


• Enter some honest Cochraneites/EBMers with no particular axe to grind, no money to make, no reputation to lose

The albumin story: my view• Experience--especially experimentally

collected data--trumps theory• The thinking behind the renaissance:

surely intensivists are not medievalists• Surely intensivists don’t want to be

associated with the chicanery of management consultancy: “It may(not) work in practice, but will it (not) work in theory?”


• The Cochraneites follow their usual method

• They pose a question, systematically search for all relevant studies, set some quality criteria, perhaps combine the data statistically, and see what the data say

• A crucial observation is that the data are poor: the big, randomised, double blind study that should have been done has not been done


• The data suggested--to their surprise but fairly consistently--that albumin kills more people than it saves

• They write up the study with a suitably tentative conclusion

The albumin story: my view• “There is no evidence that albumin

administration reduces mortality in critically ill patients with hypovolaemia, burns, or hypoalbuminaemia and a strong suggestion that it may increase mortality. These data suggest that use of human albumin in critically ill patients should be urgently reviewed and that it should not be used outside the context of rigorously conducted randomised controlled trials.”


• The study is submitted to the BMJ• We decline to fast track the study• The peer review of the study is even more

extensive than usual, generating pages of comments and revisions

• The clinical reviewer is against publication--partly because of unhappiness with the whole methodology (“garbage in, garbage out”--but this is all the evidence we have)

An aside: problems with peer review• No evidence of effectiveness• “Ineffective”: doesn’t detect errors• A lottery• A black box• Slow• Expensive• Biased• Easily abused• Can’t detect fraud

An aside on peer review

• “The benefit of peer review probably comes not from sorting out what to reject and what to publish but rather from improving what is eventually published.”

Who makes the final decision at the BMJ?

• Two practising doctors (mostly physicians) with extensive experience of peer review

• One or possibly two editors

• A statistician

• Everybody reads every word

• A majority vote carries the day

• The buck stops with the editor (me)


• A heavily revised paper is published

• An editorial written by an intensivist is generally supportive

• A scientific commentary provides modern pathophysiological explanations of why albumin might make things worse rather than better

• A TWIB overdoes it: “Albumin administration increases mortality in critically ill patients”

All hellbreaksloose


• 30 rapid responses

• “For the Editorial Board of the BMJ to sanction a headline-grabbing press release on this paper is nothing short of scaremongering, and further justifies my decision three years ago to resign my membership of the BMA.”

• Keith Judkins, intensivist

The albumin story: my view• A reasonable test is to ask what I would want for

myself, as a patient, or for someone I cared for. In brief, I would attempt to sue anyone who gave me an albumin infusion. And, as for any attempt to secure my informed consent to take part in a randomised trial (or my assent on behalf of someone I cared for who was unable to give informed consent) - forget it!

• [Sir, wow] Iain Chalmers, head of the UK Cochrane Centre

The aftermath

• A moderate editorial argues that “rather than fulminating we seek to answer the questions raised”

• FDA advises that the results deserve serious attention

• A trial is proposed

• Use of albumin drops

The aftermath

• Authors and editor get gently roasted at British intensive care meeting

• Intensivists seem to argue that “A question like whether albumen works is not useful. What matters is the whether the individual intensivist can compensate for the individual patient’s seriously dreranged physiology”

• Editor says this is exactly the argument used by psychoanalysts

Two yearsOn all issweetnessAnd light

Intensivists and evidence

• Albumen

• Low dose dopamine to prevent renal failure

• Pulmonary artery catheters

• Ranitidine to prevent GI bleeding

• Various antesepsis regimens

Why the problem?• RCTs are especially hard to do in intensive

care: urgency, very sick patients, multiple pathology, each patient is unique, consent

• It must be hard to do nothing

• “Good surgeons know how to operate. Better surgeons know when to operate. The best surgeons know when not to operate.” True as well for intensivists?

Conclusion• The “albumen story” has posed

important questions that are now being answered

• It’s prompted understanding (and misunderstanding--they always go together) of EBM among intensivists

• We all got a little carried away

• Cue music: Je ne regret rien

Why I published “the albumin paper”: confession of a buccaneering editor Richard Smith Editor, BMJ October 2001.

Documents

new evidence

aboutthe new world of

intensiviststhe albumin

smarterthe albumin story

clinical processeshas

worldis evidence

normhas ebp

togetherhas ebp