Post-marketing surveillance of combined oral contraceptives (COCs) for rare complications such as venous thromboembolism (VTE) presents unique challenges. Prospective studies, which are costly and time consuming, have to date been undertaken by only a few contraceptive manufacturers willing to commit to full evaluation of product safety. Often such studies are conducted with the approval of regulatory authorities as a precondition for marketing. Alternatively, independent investigators with access to large databases have conducted retrospective studies to compare the incidence of VTE between new and older products. Such studies, however, run the risk of erroneous conclusions if they cannot ensure comparable risk profiles for users of these different products. Often database studies are unable to access information on important confounders, and medical records may not be available to validate the actual diagnosis of VTE. "Pill scares" generated following publication and media dissemination of worrisome findings, when the conclusions are in doubt and not corroborated by stronger prospective study designs, are frequently damaging to public health. From a review of recent publications on the VTE risk with drospirenone-containing COCs, it can be concluded that the best quality evidence does not support a difference in risk between users of COCs containing drospirenone and those of COCs containing levonorgestrel.