Let the reader beware - Letters | The Star Online

"The body of medical evidence is not as safe and strong as it should be, due to researcher dishonesty or incompetence. This is well known to research physicians but perhaps not to others.

Medical data can refer to both image data and numeric data. Dishonest researchers can fabricate or falsify either or both. We are increasingly being made aware of this problem by fraud detection software. However, this software can only raise the probability of fraud. Confirmation is still by manual data audit. Unfortunately, auditing medical research data is not yet a widespread practice, hence the extent of data fraud is unknown.

Researcher incompetence gives rise to the problem of pseudo-scientific research. Here, the problems are that the study design or statistical analysis is not the correct one to answer the research question posed. This may mislead the busy attending physician and ultimately harm patients."