4 Neutral trials: If the difference between interventions is not statistically significant, is there truly no difference?

Neutral trials (i.e. trials without a statistically significant difference in the primary outcome) should not all be interpreted equally, as they will differ in the degree in which they rule out a difference between interventions. It is important to be able to recognize that “no statistically significant difference” does not mean “no difference”.

Neutral trials are also sometimes referred to as “negative trials”.

Checklist Question

Does the confidence interval (CI) exclude a clinically important difference?

Does the confidence interval exclude a clinically important difference?

If the 95% CI is wide enough to include a clinically important difference, it remains possible that a future trial with greater precision, or a meta-analysis of multiple trials, may find a clinically meaningful difference.

E.g. Authors of a trial (Nguyen-Khac E et al.) evaluating the effect of adding N-acetylcysteine to prednisone in 180 patients with acute alcoholic hepatitis concluded that mortality was not reduced with the combination vs. steroid alone. However, these were the results at 6 months:
Table 9. N-acetylcysteine plus prednisone vs. prednisone alone for patients with acute alcoholic hepatitis on the outcome of mortality.
Outcome Combination Steroid Alone Absolute risk difference (95% CI)
Mortality 27% 38% -11% (-22% to +5%) 

The uncertainty of the estimated reduction represented by the 95% CI means that the trial could not exclude the possibility of an absolute reduction in mortality with combination therapy as high as 22%.

 

definition

License

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

NERDCAT Copyright © 2022 by Ricky Turgeon is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book