Now that you know the results are valid and well-reported, are they clinically important?
A very large clinical trial may be able to show small differences between interventions that have little clinical significance. Statistical significance does not always equate to clinical significance.
|Is a treatment effect (therapy) or improvement in sensitivity/specificity (diagnosis), clinically relevant?||A small therapeutic benefit or improvement in sensitivity may be important for a highly fatal disease, but may not be enough to justify a new intervention in a disease with a good prognosis.|
|What level of uncertainty surrounds any results?||Are confidence intervals given? A confidence interval shows the range of results within which the ‘true’ result is likely to fall. If a 95% confidence interval is given, this means that if a study was repeated many times within the same population, the ‘true’ value should fall within these limits 95% of the time.
If confidence intervals are given, do they encompass values that would not be clinically relevant? A larger clinical trial should give a smaller confidence interval, and a more precise estimate of the likely treatment effect.
|Do the risks of a treatment or diagnostic procedure outweigh the potential benefits?||What are the toxicities and side effects of an intervention?
How do these compare with the comparator arm?
Is there potential morbidity or mortality for performing a test that outweighs the benefits of making a diagnosis?