The major limitations of meta-analyses are related to non–linear regressions, restricted coverage, confusing effects due to multi-factors, lack of randomization and mediation, and theory-directed approach that may obscure discrepancies. The geometrical reasoning in epidemiological assessments may help to facilitate measuring confounding “confusions”. Current concept resembles a fractal analysis where the analogue of the stressor is the cause (factor) of the disease (or infectious agent) and the analogue of elasticity index is the confounder. The acquisition of fractal analysis in epidemiological assessment of notifiable diseases in surveillance sample may help to locate and measure confounding effects in multi-dimensional system, through iteration of fractal dimension of data sets of self-similar patterns and self-affine traces toward the initial point where no confounder exists.