P-values and significance: The null hypothesis that they are not related is correct.

False signals and therapeutic optimism have led medicine down many a wrong pathway. Apart from the unnecessary costs of care and redundant research expenditure and efforts this has caused, therapies which were eventually discontinued may have come at a staggering cost of lives lost. Still most statistical information in the medical literature is presented with its p-values and little else. The use of p<0.05 was proposed as an arbitrary threshold for defining a statistically significant difference. Strong signals manifest even with the use of small sample sizes and are highly replicable. However, weak signals, which are increasingly the focus of modern research, may be obscured by the presence of other powerful variables in the dataset are more likely to decay in subsequent studies due to multiple causes. It is time to set higher standards for evidence in medical research. Data that should always be presented in conjunction with the p-value are the Confidence Intervals, which illustrate the uncertainty inherent to the results, and the Fragility Index, which reflects result robustness. Multiple RCTs should be the standard for implementing change. Ideally these studies should consistently demonstrate p-values <0.005, study and control groups with well separated 95% CIs and high fragility indices.