[Calculation of radioimmunochemical determinations by "spline approximation" (author's transl)].

A simplified method, based on the "spline approximation", is reported for the calculation of the standard curves of radioimmunochemical determinations. It is possible to manipulate the mathematical function with a pocket calculator, thus making it available for a large number of users. It was shown that, in contrast to the usual procedures, it is possible to achieve optimal quality control in the preparation of the standard curves and in the interpolation of unknown plasma samples. The recaluculation of interpolated values from their own standard curve revealed an error of 4.9% which would normally be an error of interpolation. The new method was compared with two established methods for 8 different radioimmunochemical determinations. The measured values of the standard curve showed a weighting, and there was a resulting quality control of these values, which, according to their statistical evalution, were more accurate than those of the others models (Ekins et al., Yalow et al., (1968), in: Radioisotopes in Medicine: in vitro studies (Hayes, R. L., Goswitz, F.A. & Murphy, B. E. P., eds) USA EC, Oak Ridge) and Rodbard et al. (1971), in: Competitive protein Binding Assys(Odell, W. D. & Danghedy, W. H., eds.) Lipincott, Philadelphia and Toronto). In contrast with these other models, the described method makes no mathematical or kinetic preconditions with respect to the dose-response relationship. To achieve optimal reaction conditions, experimentally determined reaction data are preferable to model theories.