Please enable JavaScript.
Coggle requires JavaScript to display documents.
Continuous Outcome Data, Compare by taking difference in means/ Summary…
Continuous Outcome Data
-
EXAMPLE
- Null H - the difference between the mean SBP for the new treatment and the current treatment is equal to zero
Alt H - the difference between the mean SBP for the new treatment and the current treatment is not equal to zero
- Primary outcome is continuous data (SBP) = mean of two groups
SS = 0, evidence to suggest the effect of both treatments are the same
- Estimate SS = mean difference
-
- SE used to calculate the CI (4.95%)
-
- Test the NH producing a p-value
-
6.Interpretation
On average , participants randomised to the new treatment had a statistically significant lower SBP of 4.1mmHg (95% CI: 1.21,6.99) compare to the participants randomised to the current treatment (p-value = 0.034)
-
-
Clinical Significance
95% CI
(-1,1) difference of 2, HIGH precision, VERY clinically relevant - almost no difference in SBP
(-25,25) difference 50, LOW precision & LITTLE clinical relevance
(-2.5,-0.5) difference of 2, HIGH precision, LITTLE clinical relevance - small difference
(-65.5,-1.5) difference of 60, LOW precision & LITTLE clinical relevance - difference may be small or large
(-25,-15) difference of 10, MODERATE precision, REASONABLE clinical relevance
Linear Regression
For continuous outcome (example of stratification variables used for randomisation), this can be done by linear regression. Obtaining an estimate for adjusted difference in means with 95% CI & p-value
-
-