Figure 5

Empirical evidence of missing data (data asymmetry). (A) Funnel plotting of the rat (BBB-only) cohort including 40 included experiments (n = 890) reported an overall neurobehavioral efficacy of 18.7% (95% CI 14.4–23.0). (B) Trim and fill analysis added 19 experiments (white circles) indicating missing data. Factoring in missing data points resulted in a reduced effect size of 8.8% (95% CI 2.6–14.9). (C) Contour-enhanced funnel plot adds three shaded areas of statistical significance to the funnel plot. Most of the experiments lie within the grey area (highly significant results p < 0.01) or within the light green area (p < 0.05). Visual inspection for plot symmetry would impute potential missing studies close to the threshold of significance on the left side of the funnel plot or within the grey area of statistical significance⁴⁰.