Jadad criteria allocate a point each for randomization, double-blind design, and description of dropouts. If randomization and double-blind concealment are assured, an additional 2 points are added. If randomization or double-blind concealment is not assured, a point is deducted for each. A trial with a score of 3 or more is regarded as high quality. Data from trials with scores of 3 or more were grouped and analysed separately from those scoring less than 3.