Data analysis
Outcomes were analyzed in matched sets by fitting log-binomial models, frequentist (with cluster robust sandwich estimator of the standard error) and Bayesian with three different priors for the effects of interest: (i) skeptical prior – moderatly informative neutral prior consistent with an a priori hypothesis of no effect, centered at 0 for the Ln(RR) with standard deviation 0.355. It assigns equal probability (50%) for an RR >1.0 and an RR <1.0, with 95% probability between RR=0.50 and RR=2.0; (ii)optimistic prior – moderately informative prior centered at -0.199 for the Ln(RR), with standard deviation 0.4, i.e., 18% relative risk reduction as seen in an up-dated Bayesian meta-analysis of randomized trials of fluvoxamine in this setting2, but with 30% probability of an RR >1.0; (iii) pesimistic prior – weakly informative prior centered at 0.199 for Ln(RR) (reciprocal to the optimistic prior) with a standard deviation of 0.77. Although it suggests harm, it leaves 40% probability of an RR <1.0. We used SAS 9.4 for Windows (SAS Inc, Cary, NC) (proc glimmix ) to fit frequentist and packagerstanarm 16 in R15 to fit Bayesian models. Since the number of people prescribed with fluvoxamine was low and the outcomes of interest were rather infrequent in matched sets, to increase the precision of the A vs. B comparisons, we conducted network meta-analysis of results in matched sets A vs. B, A vs. C and B vs. C. Although derived from the same pools of original patients, matched contrasts were based on different pseudopopulations (by selection and weighting). We performed frequentist (packagenetmeta 17 in R15) and Bayesian (package BUGSnet 18 and packagegemtc 19 in R15, with default priors) network meta-analysis using weighted counts from matched sets, and also using the effect measures (RR) generated in Bayesian analyses with the skeptical prior. We made no multiplicity adjustments of the alpha level: present analysis did not aim at “discoveries”; we intended to simply estimate quantities of interest and see how they related to the RCT data.