Evidence Bias.pdf

The Italian Group for Evidence-Based Medicine (GIMBE) organized a workshop on clinical decision-making and efficacy trials. The event focused on the integrity of research, highlighting issues such as publication bias, scientific fraud, ethics, and conflicts of interest.

Publication bias occurs when trials with negative results are less likely to be published than those with positive results, leading to an overestimation of therapeutic effects. This bias can affect up to 50% of meta-analyses, with 20% of cases being highly probable and 10% of cases resulting in modified outcomes.

Scientific fraud includes fabrication, falsification, plagiarism, and failure to obtain informed consent. Examples of fraud include inventing data, distorting research results, and publishing reviews that appear impartial but are actually written by industry "ghost writers."

Consequences of fraud include overestimation of therapeutic efficacy, promotion of professional careers, and harm to patients. Ethical issues arise when placebo-controlled trials are used to demonstrate the superiority of new drugs, and when the "game of doses" is played to favor the new drug.

Conflicts of interest can also influence research, as seen in the debate over calcium-channel antagonists. A study found that authors with financial ties to the pharmaceutical industry were more likely to support the use of these drugs.

In conclusion, the integrity of research is compromised by publication bias, scientific fraud, ethical issues, and conflicts of interest, which can have serious consequences for patients and public health.