Understanding P-Values
A p-value is a measure used in statistics to determine the strength of evidence against a null hypothesis (which usually states that there is no effect). A smaller p-value indicates stronger evidence against the null hypothesis.
🟢 State the number: It’s important to report the exact p-value (like 0.004) rather than just saying it's less than 0.05. This is because a p-value of 0.04 and one of 0.0001 are not equivalent in terms of the evidence they provide.
🟢 Not Dichotomous: P-values aren’t just 'significant' or 'not significant'. They provide a continuum of evidence against the null hypothesis.
Examples:
🔵 Paradigm-HF Trial (Sac/valsartan vs Enalapril): With a hazard ratio (HR) of 0.8 and a p-value of 0.0000004, this indicates extremely strong evidence against the null hypothesis. The chance of this result occurring by chance is less than 1 in a million, which strongly suggests that Sac/valsartan is superior to enalapril.
🔵 IMPROVE-It Trial (Ezetimibe vs. Placebo): An odds ratio (OR) of 0.94 with a p-value of 0.016 provides some evidence of a modest treatment benefit, but it's not as overwhelming as in the Paradigm-HF trial.
🔵 P-Value of 0.06: This doesn’t definitively mean there's no treatment effect. It suggests that there's insufficient evidence to declare an effect. The treatment may have a modest effect, or it may have no effect – the data is inconclusive.
🔵 P-Value of 0.41 (Astronaut Trial): With a p-value this high, it's reasonable to conclude that there's no evidence of a treatment effect.
P-values are not black and white. They offer varying degrees of evidence. Always report the actual p-value, and understand that the smaller the p-value, the stronger the evidence against the null hypothesis. Avoid oversimplifying by categorising trials as "positive" or "negative" based solely on the p-value of the primary outcome.
