Deficiency of Peer Review - Muir Gray, NHS National Knowledge Service
"The future is here; it is just not generally distributed - William Gibson"
Muir is talking about the issue of sharing data usefully - data may be known but for reasons of privacy and trust it may not always readily available at your fingertips. Thus he will be talking about the quality of knowledge (from a medical perspective).
"most articles in clinical journals are not appropriatte for direct application by individual clinicians" - various refs etc. will be available later.
"Most reviews published in peer reviewed journals or funded by industry have serious methodological flaws that limit their alue to guide decisions" - A.R. Jaded et al (2004)
Muir notes that most journals answer questions raised by researchers, not always the useful date or areas that you would wish.
The way evidence is presented can have a huge impact on the apparent effectiveness of a given drug/technique etc. Researchers often duplicate trials and like positive research - this can lead to misleading results and medical efficacy.
Peer reviewing is the target here *but* it is not the only problem. Planning, funding, research topic selection, positive bias etc. cause issues at all stages. Huge positive publication bias as many drug companies, researchers etc don't want to publish negative results - "we need a journal of negative results!"
People frequently do not report their findings and what they were funded for. They do not report all hypotheses, generally it is the negative findings that are not reported. Industry follow this trend perhaps even more enthusiastically. Publishing constraints also support the bias with 91% of 1994/5 contrained from publication: only 9% of protocols reported.
NB: this presentation is heavy on quotes which are flashing by quickly so I would recommend seeing the slides directly if/when they are made available.
Solutions:
The same principles are being applied to other types of research - a key issue is the perception of absolute vs. relative risk which researchers are not well aware of at present and publishers must therefore have a role in addressing this.
Finally we plan to reduce the carbon footprint of the NHS and that means reducing the carbon footprint of its supply chain ("campaign for greener healthcare"). Hopefully entirely digital publication usage by 2011 - digital has its own carbon footprint but it is better. Muir's organisation is knowledgeintoaction.org. This is a little off-topic but a personal concern of Muir's own charity. He would like us all to walk everywhere please (great for 1-to-1 meetings apparently). Also he wants to set up a knowledge service using internet radio. And logical (online) bookshops for healthcare professionals.
[NB: Q&A for this and other sessions to come later. ]
Muir is talking about the issue of sharing data usefully - data may be known but for reasons of privacy and trust it may not always readily available at your fingertips. Thus he will be talking about the quality of knowledge (from a medical perspective).
"most articles in clinical journals are not appropriatte for direct application by individual clinicians" - various refs etc. will be available later.
"Most reviews published in peer reviewed journals or funded by industry have serious methodological flaws that limit their alue to guide decisions" - A.R. Jaded et al (2004)
Muir notes that most journals answer questions raised by researchers, not always the useful date or areas that you would wish.
The way evidence is presented can have a huge impact on the apparent effectiveness of a given drug/technique etc. Researchers often duplicate trials and like positive research - this can lead to misleading results and medical efficacy.
Peer reviewing is the target here *but* it is not the only problem. Planning, funding, research topic selection, positive bias etc. cause issues at all stages. Huge positive publication bias as many drug companies, researchers etc don't want to publish negative results - "we need a journal of negative results!"
People frequently do not report their findings and what they were funded for. They do not report all hypotheses, generally it is the negative findings that are not reported. Industry follow this trend perhaps even more enthusiastically. Publishing constraints also support the bias with 91% of 1994/5 contrained from publication: only 9% of protocols reported.
NB: this presentation is heavy on quotes which are flashing by quickly so I would recommend seeing the slides directly if/when they are made available.
Solutions:
- Good Systematic Reviews
- Better peer reviewing and editing - what do we mean by peer review? It has many meanings and so how do we define and work out what we want it to achieve.
- Improve standards of reporting and include caveats to make clear that the reader must beware.
The same principles are being applied to other types of research - a key issue is the perception of absolute vs. relative risk which researchers are not well aware of at present and publishers must therefore have a role in addressing this.
Finally we plan to reduce the carbon footprint of the NHS and that means reducing the carbon footprint of its supply chain ("campaign for greener healthcare"). Hopefully entirely digital publication usage by 2011 - digital has its own carbon footprint but it is better. Muir's organisation is knowledgeintoaction.org. This is a little off-topic but a personal concern of Muir's own charity. He would like us all to walk everywhere please (great for 1-to-1 meetings apparently). Also he wants to set up a knowledge service using internet radio. And logical (online) bookshops for healthcare professionals.
[NB: Q&A for this and other sessions to come later. ]
Labels: green issues, health, muir gray, NHS
posted by Nicola Osborne
at 4/07/2008 12:21:00 pm
2 comments
Subscribe to this blog