Misrepresentation of research data is rife in medical testing, a report has claimed.
A team of British and French scientists found that more than half of 72 investigations exaggerated the beneficial effect of medicines.
One example saw a study conclude that a cancer detection system worked even though the claim was not borne out by statistics, according to Dr Isabelle Boutron, who worked on the project.
She added that not all cases were quite so glaring but said: “Some of it was quite shocking.”
Other examples of data manipulation were more subtle, the study added, explaining that some spin was missed by experts. Typical practices employed here included downplaying risk and removing placebo data from the analysis.
The investigation comes shortly after earlier research suggested that findings were often “spun” when money was involved, for example when a drug maker funds a study of its own product. In such cases, favourable conclusions frequently contradicted the actual results.
Boutron said the findings would make her more suspicious when she reads academic articles but added: “I’m sure most of the authors weren’t cheating, they were just very enthusiastic about their results.”
Copyright Press Association 2010
European Medicines agency