Many studies and analyes we do require statistics to back up. However, many times the results run out to be statistically significant, yet when interpret it in the clinical context, it is too small to be significant.
This is rather frustrating. Any solution for solving this?
Is those very small interval unit scale more easily affected?
Can we change our analytics method to cope with the results?
Beyond statistical significance: clinical interpretation of rehabilitation research literature.
Int J Sports Phys Ther. 2014;9(5):726–736.
The clinical data if well presented and analysed with appropriate statistical tool should be taken for decision.
Could you clarify what is too small to be clinically insignificant? is it the level of a certain biomarker?
E.g. I measured the Hounsfield unit (HU) of an implant on Computed tomography.
There is statistically significant difference between different groups, yet the absolute difference is only 3-4 HU, which on clinical interpretation is not significant.
A result may be statistically significant but if it doesn't produce a change in the patients that is significant to them then it is clinically insignificant.
One has to have a large number of examples in a study to ensure truism.
Statistically significant but clinically not significant
The following articles clarify this issue in detail with examples:
1. Article Statistical and clinical significance, and how to use confid...
2. Article Clinical significance versus statistical significance
3. Article Beyond statistical significance: Clinical interpretation of ...
4. Article Statistical significance versus clinical relevance
Critical appraisal of any research in healthcare is important to determine external validity (i.e. clinical significance beyond the setting of the study). One aspect is the magnitude of the effect size. Another is the quirks of the healthcare setting in which you work. Sadly the majority of clinical research is flawed in one way or another so external validity is limited.
- Badges
- Science topic
- Similar topics
- Medicine