News Release

The good and bad of medical research

Peer-Reviewed Publication

The Lancet_DELETED

The difficult area of using observational methods in medical research is explored in a Comment by The Lancet's Editor Dr Richard Horton, published early Online and in an upcoming edition of The Lancet. The Comment coincides with the publication of a report by the working party of The Academy of Medical Sciences, on which Dr Horton sat.

He says: "In most areas of science, non-experimental evidence does not command a great deal of public attention. In health, it frequently becomes a matter of urgent public concern. From the safety of vaccination to the risks of passive smoking, the validity of non-experimental research has often been a flashpoint of political as well as scientific debate. Yet observational research underpins a vast amount of clinical and public health knowledge that is fundamental to daily medical practice and to policymaking. High-profile reversals of a few flawed or over-interpreted studies should not be allowed to damage the overall credibility of non-experimental methods in medicine."

The Comment refers to the recommendations by the Academy of Medical Science on the environmental cause of disease, following the inquiry they initiated in 2006, to investigate various areas surrounding this issue -- including the strengths, limitations, and potential of non-experimental methods for the identification of environmental causes of disease; and lessons from successful and less successful examples of non-experimental research.

Dr Horton points out there are many good examples of how non-experimental research studies have produced reliable causal inferences -- smoking and lung cancer; lipids and coronary artery disease; perinatal studies in HIV infection; folic acid and neural tube defects; and thalidomide and teratogenicity. However, he adds: "There have also been some notable errors and failures -- over the measles, mumps and rubella vaccine; hormone replacement therapy and coronary artery disease; the safety of calcium-channel blockers; caffeine in pregnancy; and vitamin supplements and mortality. The common lesson from the misleading claims that arose from these examples is that small pilot studies with no or inadequate controls, perhaps driven by individuals who are seeking to advance one particular and contentious point-of-view, should be viewed with great caution."

He continues by reinforcing the point that all results of scientific research, experimental or non-experimental -- are provisional, and that no one study should be relied to the change behaviour or policy - it should first be replicated across several contrasting populations. Dr Horton says: "The safest course of action for the policymaker is always to be guided by the totality of available evidence. Sometimes the risk or hazard associated with not acting may be greater than the risk of acting. Before the results of randomised clinical trials became available, many experts argued that this was the case with male circumcision in the prevention of HIV transmission. It is certainly the case for child abuse."

The Comment draws attention to the evidence of a decline in the scientific expertise of the UK's civil service, stating that "All senior government officials should have at least a basic understanding of scientific methods and peer review." And Dr Horton warns that even when non-experimental research follows best scientific practice, communication of that research can be a major obstacle. This task is often considered the burden of science and health journalists, but is a responsibility shared by editors of medical journals. He discusses the conflicting responsibilities faced by editors -- to ensure preservation and strengthening of research literature, the risks of balancing newness with trueness, and the need to be vigilant when considering potentially controversial findings based on non-experimental methods. A recent Science investigation of the South Korean cloning fraud suggested high-risk papers (which could trigger controversy) should trigger more intensive-peer review.

Dr Horton praises the Academy of Medical Sciences report's guidance for clinicians, policymakers and funders, saying: "These proposals are supplementing and strengthening existing recommendations about the reporting and interpretation of non-experimental research."

He concludes: "An overarching goal for everyone who generates and uses research should be to embrace the value of scientific evidence at all levels of public policymaking. Each of us has a role to play in building a strong and sustainable evaluative culture in society, We need to think more carefully and creatively about how we discharge that role."

###

This paper associated with this release can be found at the link below:
http://eurekalert.org/jrnls/lance/medicalresearchfinal.pdf


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.