Statistics - explanations and formulas

Mean Difference

Measure of continuous data – data before or after an intervention; or difference between 2 groups (ie blood pressure, weight, pain scales, etc)

- Compared to dichotomous data – only 2 points (ie improve or not improve; death vs no death; etc)

A mean difference of 0 implies that the data is equal in both groups. MD >0 implies that the data is increased in the first group. MD <0 implies that the data is less in the first group.

            - from this study; Mean headache intensity (0-10 pain scale – [this is continous type of data]) was also lower at 12 weeks in the melatonin group vs placebo group (4.8 points vs 3.6 points ; Mean Difference, -1.2 points; 95% CI, -1.6 to -0.8) – here, the confidence interval does not cross 0, so this reveals with accuracy and precision, the melatonin group had less pain compared to placebo