The working principle of lateral flow assays, such as the widely used COVID-19 rapid tests, is based on the capillary-driven liquid transport of a sample fluid to a test line using porous polymeric membranes as the conductive medium. In order to predict this wicking process by simplified analytical models, it is essential to determine an effective capillary radius for the highly porous and open-pored membranes. In this work, a parametric study is performed with selected simplified structures, representing the complex microstructure of the membrane. For this, a phase-field approach with a special wetting boundary condition to describe the meniscus formation and the corresponding mean surface curvature for each structure setup is used. As a main result, an analytical correlation between geometric structure parameters and an effective capillary radius, based on a correction factor, are obtained. The resulting correlation is verified by applying image analysis methods on reconstructed computer tomography scans of two different porous polymeric membranes and thus determining the geometric structure parameters. Subsequently, a macroscale flow model that includes the correlated effective pore size and geometrical capillary radius is applied, and the results are compared with wicking experiments. Based on the derived correction function, it is shown that the analytical prediction of the wicking process in highly porous polymeric membranes is possible without the fitting of experimental wicking data. Furthermore, it can be seen that the estimated effective pore radius of the two membranes is 8 to 10 times higher than their geometric mean pore radii.