[Skip to Content]
[Skip to Content Landing]
Citations 0
April 6, 2011

Treatment of Chronic Prostatitis/Chronic Pelvic Pain Syndrome—Reply

JAMA. 2011;305(13):1298-1299. doi:10.1001/jama.2011.400

In Reply: Dr Jackson and colleagues are concerned that our meta-analysis alternated between fixed and random effects for pooling. The majority of outcomes in our review were heterogeneous, so we did use a random-effects model. The exceptions to this were pain and voiding scores between phytotherapy and placebo groups and treatment responsiveness between anti-inflammatory drugs and placebo where mild heterogeneity was found (ie, I2 = 0%, 0%, and 24.4%, respectively). Applying fixed- or random-effects models for these first 2 pooled effects obviously yields identical results. For the last outcome, the pooled risk ratio for treatment responsiveness was 1.8 (95% confidence interval [CI], 1.2-2.6) for the random-effects model compared with 1.8 (95% CI, 1.3-2.4) for the fixed-effects model. As Jackson et al point out, the tests for heterogeneity are weak, and that is why we set the threshold P value for this test at .10 rather than .05 and why we also calculated I2 statistics.1,2

First Page Preview View Large
First page PDF preview
First page PDF preview