A meta-analysis of small studies suggests a large benefit from a new haemostatic device, but the funnel plot is markedly asymmetric and larger trials show little effect. What is the most defensible interpretation?
Asymmetry can arise when small negative studies are missing, although heterogeneity and methodological differences are alternative explanations. The result should be explored with sensitivity analyses rather than accepted at face value.
A. The pooled estimate is automatically definitive because meta-analysis is highest-level evidence: Evidence hierarchy does not protect a synthesis from biased or heterogeneous inputs.
B. Funnel-plot asymmetry proves fraud: Asymmetry has several possible causes and is not proof of misconduct.
C. Small-study effects and publication bias may be exaggerating the pooled benefit: Asymmetry can arise when small negative studies are missing, although heterogeneity and methodological differences are alternative explanations. The result should be explored with sensitivity analyses rather than accepted at face value.
D. Discard all large trials: Larger, more precise studies usually deserve substantial weight and may expose small-study effects.
E. Convert the outcome to a surrogate to restore significance: Changing outcomes after seeing results introduces further bias.
If the funnel plot were symmetrical but statistical heterogeneity remained extreme, clinical and methodological sources of heterogeneity would still need investigation.
Independent Clinora educational preparation material; not official JCIE content, an accredited programme, clinical advice, or a substitute for local policy and specialist judgement.
Back to the Section 1 bank