I am comparing two methods of muscle strength assessment. I have the means of both measurements together with the mean difference (bias), standard deviation, 95% confidence interval and lower/upper agreement levels (of the difference). Is there an equation to calculate the percentage of error of the second method compared to the first?
The formula in the (AACN Adv Crit Care. 2008 Apr-Jun;19(2):223-34.
Design, analysis, and interpretation of method-comparison studies.
Hanneman) (highlighted) states that I should take the confidence limits (range between 95% CI) divided by the mean of the established (first) method. Is it correct or are there other more suitable equations for my case?
Thank you
Mohamed
Hi Mohamed - you might find this helpful:
http://www.ncbi.nlm.nih.gov/pubmed/12858311
Applying the right statistics: analyses of measurement studies. 2003 Bland JM, Altman DG.
Also, it might make more sense to talk about disagreement rather than error unless one of your measures is a criterion.
P
http://www.ncbi.nlm.nih.gov/pubmed/12858311
Dear Mohamed,
I suggest you read the fallowing papers attached.
Regards
Dear Mohamed,
I see that your data are continuous, since you have means and standard deviation, so the papers already cited by Paul and Eurico are good resources for Bland and Altman 95% limits of agreement and typical error.
A caveat for applying limits of agreement is that the differences between the 2 methods of measurement should follow an approximate normal distribution and that the mean and standard deviation of the differences should be constant at different points in the measurement scale, i.e. that the mean (Sd) difference in strength in your case should be the same for people with low, moderate and high strength. The 2003 Bland and Altman paper offers solutions if these data assumptions do not hold.
Kind regards
Catherine
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology