In a clinical trial what amount of decrease in HbA1c would you consider to be clinically significant?
Ideally a drop to 6 or less but if that is not possible, a statistically significant difference of or beyond 2 standard deviation will be useful. Long-term hyperglycemia related systemic derangement is positively associated with the degree of hyperglycemia of which temporal profile of HbA1c is a good indicator.
From a clinical perspective, I would say if the interpretation of the HbA1c drops to normal levels from high/borderline... meaning you would analyze the data using logistic regression to see if the "trial" had a clinically meaningful effect. At the lab I work at, 5.7% is the diagnostic cutoff for healthy individuals; however, the diagnostic cutoff at different labs/sites can vary depending on the exact testing methodology used.
you can compare their hemoglobin with other healthy groups when you find the difference is significant it can be taken
For research, having a control group and a statistically significant difference between intervention group and control group would be sufficient enough.
It would be clinically significant, if the post intervention HbA1C is close to 6.5 (or 7%), irrespective of the baseline data. In research however, itt mainly depends on the baseline HbA1C of the subjects... i.e. it depends whether the baseline HbA1C is 10% or 8%
Dear colleagues,
Thank you for your valuable suggestions and contributions, which I really appreciate. Professor Niessen, thanks for your guidance as usual in times of need.The preliminary analysis from our mobile phone clinical trial intervention in Bangladesh shows:
In the SMS group, using ITT analysis, HbA1c declined from 8.12±1.67 % at baseline to 7.38±1.05 % at 6 months (p 0.001). A similar reduction occurred in the control group (8.49±1.94% to 8.07±1.51% (p 0.064). The model-adjusted intervention difference of -0.741 (95% CI -1.10%, -0.38%) indicated that mobile phone SMS intervention was superior to standard-of-care alone in reducing HbA1c among patients with type 2 diabetes on oral medication but not statistically significant (p 0.272 ). Similar findings were observed in the complete-case analyses, where the difference in mean changes in the SMS group between baseline and 6 months were -0.79 (-1.17, -0.42) (p 0.000). However, the difference-in-difference calculated as an effect of the SMS intervention was only -0.32 (95% CI -0.89%, 0.25%) (p 0.272).
I would appreciate any further comments and suggestions.
Best regards,
Sharif
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Peptides