Article type
Year
Abstract
Background: Evidence from placebo-control trials comparing the least effective antipsychotics is distorted by small study effects and publication bias (PB). This raises concerns about the validity of the results from network meta-analysis (NMA). Biased evidence in placebo-control comparisons may affect the treatment effects and the ranking of the antispychotics as placebo-controlled trials account for one-third of the information in the entire network.
Objectives: To assess the potential impact of small study effects (SSE) and PB on the recently estimated relative effectiveness and ranking of pharmacological treatments for schizophrenia. To suggest statistical methods for adjusting results in the presence of SSE or PB.
Methods: We used a recently published network of 167 studies involving 36,871 patients and comparing the effectiveness of 15 antipsychotics and placebo. We used novel visual and statistical methods to explore whether smaller studies are associated with larger treatment effects and if the probability of publication is associated with the magnitude of effect. We conducted a NMA of the published evidence as our primary analysis and employed a sensitivity analysis considering low, moderate and severe selection bias (that correspond to the number of unpublished trials) with an aim to evaluate robustness of point estimates and ranking. We explored whether placebo-controlled and head-to-head trials are associated with different levels of PB.
Results: We found that small placebo-controlled trials exaggerated slightly the efficacy of antipsychotics and PB was not unlikely in the evidence based on placebo-controlled trials; however, ranking of antipsychotics remained robust.
Conclusions: The total evidence comprises many head-to-head trials that do not appear to be prone to SSE or PB and indirect evidence ‘washes-out’ some of the bias in the placebo-controlled trials.
Objectives: To assess the potential impact of small study effects (SSE) and PB on the recently estimated relative effectiveness and ranking of pharmacological treatments for schizophrenia. To suggest statistical methods for adjusting results in the presence of SSE or PB.
Methods: We used a recently published network of 167 studies involving 36,871 patients and comparing the effectiveness of 15 antipsychotics and placebo. We used novel visual and statistical methods to explore whether smaller studies are associated with larger treatment effects and if the probability of publication is associated with the magnitude of effect. We conducted a NMA of the published evidence as our primary analysis and employed a sensitivity analysis considering low, moderate and severe selection bias (that correspond to the number of unpublished trials) with an aim to evaluate robustness of point estimates and ranking. We explored whether placebo-controlled and head-to-head trials are associated with different levels of PB.
Results: We found that small placebo-controlled trials exaggerated slightly the efficacy of antipsychotics and PB was not unlikely in the evidence based on placebo-controlled trials; however, ranking of antipsychotics remained robust.
Conclusions: The total evidence comprises many head-to-head trials that do not appear to be prone to SSE or PB and indirect evidence ‘washes-out’ some of the bias in the placebo-controlled trials.