Adjustment for reporting bias in network meta-analysis of antidepressant trials

Abstract

BackgroundNetwork meta-analysis (NMA), a generalization of conventional MA, allows for assessing the relative effectiveness of multiple interventions. Reporting bias is a major threat to the validity of MA and NMA. Numerous methods are available to assess the robustness of MA results to reporting bias. We aimed to extend such methods to NMA.MethodsWe introduced 2 adjustment models for Bayesian NMA. First, we extended a meta-regression model that allows the effect size to depend on its standard error. Second, we used a selection model that estimates the propensity of trial results being published and in which trials with lower propensity are weighted up in the NMA model. Both models rely on the assumption that biases are exchangeable across the network. We applied the models to 2 networks of placebo-controlled trials of 12 antidepressants, with 74 trials in the US Food and Drug Administration (FDA) database but only 51 with published results. NMA and adjustment models were used to estimate the effects of the 12 drugs relative to placebo, the 66 effect sizes for all possible pair-wise comparisons between drugs, probabilities of being the best drug and ranking of drugs. We compared the results from the 2 adjustment models applied to published data and NMAs of published data and NMAs of FDA data, considered as representing the totality of the data.ResultsBoth adjustment models showed reduced estimated effects for the 12 drugs relative to the placebo as compared with NMA of published data. Pair-wise effect sizes between drugs, probabilities of being the best drug and ranking of drugs were modified. Estimated drug effects relative to the placebo from both adjustment models were corrected (i.e., similar to those from NMA of FDA data) for some drugs but not others, which resulted in differences in pair-wise effect sizes between drugs and ranking.ConclusionsIn this case study, adjustment models showed that NMA of published data was not robust to reporting bias and provided estimates closer to that of NMA of FDA data, although not optimal. The validity of such methods depends on the number of trials in the network and the assumption that conventional MAs in the network share a common mean bias mechanism.

Effectiveness and Tolerability of Anticoagulants for Thromboprophylaxis after Major Joint Surgery: a Network Meta-Analysis

“…To address global inconsistency from all possible sources, we used a design-by-treatment interaction model when adjusting results for entire publication bias [26]. Comparison-adjusted funnel plots were used to determine whether small-study effects were present in our analysis [27]. …”

Comparative effectiveness of pharmacologic and nonpharmacologic interventions for orthodontic pain relief at peak pain intensity: A Bayesian network meta-analysis

“…35,36 Therefore, we extracted the pairwise estimates from the network meta-analysis in a way that would reflect that in trials comparing active and inactive treatments (eg, placebo, control group), it could be reasonably assumed that the active treatment would be favored by a smallstudy bias. 35 A similar approach was used to extract pairwise estimates to draw the comparison-adjusted funnel plot to assess small-study effects. The comparisonadjusted funnel plot was drawn using the "netfunnel" command 37 in Stata software (version 13; StataCorp, College Station, Tex).…”

“…34 Small-study effects were assessed using the metaregression-based approach recently introduced to Bayesian network meta-analyses. 20,35 This approach is based on the assumption that biases are exchangeable across the network; ie, biases, if present, operate in a similar way in trials across the network. 35,36 Therefore, we extracted the pairwise estimates from the network meta-analysis in a way that would reflect that in trials comparing active and inactive treatments (eg, placebo, control group), it could be reasonably assumed that the active treatment would be favored by a smallstudy bias.…”

“…20,35 This approach is based on the assumption that biases are exchangeable across the network; ie, biases, if present, operate in a similar way in trials across the network. 35,36 Therefore, we extracted the pairwise estimates from the network meta-analysis in a way that would reflect that in trials comparing active and inactive treatments (eg, placebo, control group), it could be reasonably assumed that the active treatment would be favored by a smallstudy bias. 35 A similar approach was used to extract pairwise estimates to draw the comparison-adjusted funnel plot to assess small-study effects.…”

Intravesical treatment for interstitial cystitis/painful bladder syndrome: a network meta-analysis

“…To rank the treatments for an outcome, surface under the cumulative ranking (SUCRA) probabilities was used [24]. A Bcomparison-adjusted^funnel plot was used to assess the presence of small-study effects in the network metaanalysis [25]. Two-tailed p < 0.05 was considered statistically significant.…”