The four main conceptual paradigms of basic biomedical research are that the explication of mechanisms is a primary goal of science; that scientific data must be "hard"; that formulation of hypotheses and counter-hypotheses is a principal creative challenge; and that experiments are the main mechanism for using scientific methods to conduct investigations. Although suitable for the spectacular accomplishments of biomedical research, these paradigms do not offer a satisfactory foundation for the additional basic science needed in the managerial decisions of patient care. For these decisions, the primary scientific goal is usually prediction rather than explication alone; the crucial data often involve "soft" information about clinical and personal phenomena; and the main creative challenge is to develop an improved methodology for getting adequate data and making unbiased comparisons. Although the experiments conducted as randomized clinical trials have been scientifically helpful, the improved methods will have to include evidence obtained in the often unplanned "experiments" of ordinary clinical practice.