Three-body integral equations when applied to collisions of protons with hydrogen atoms yield the amplitudes for direct scattering and electron exchange which automatically satisfy two-body and (if solved exactly also) three-body unitarity. As a consequence, differential cross sections (DCS) for both reactions are calculated on equal footing, and the total exchange cross section (TCS) results from the corresponding DCS. This is to be contrasted with standard models used in atomic physics which either are of the (Distorted-Wave) Born type (to be applied at higher energies), or solve the Schrodinger equation by expanding the wave function in a basis (close-coupling method which has problems especially for rearrangement processes). Hence, at higher energies in general separate models have to be developed to describe the DCS for either direct scattering or electron exchange, and frequently also for the TCS. For instance, the most sophisticated traditional models, which provide a very good reproduction of the TCS data, are the continuum distorted wave [1] and the boundary corrected first Born model [2]. Both, however, fail to describe differential cross sections which represent a much more stringent test. On the other hand, three-body integral equations suffer from the principal defect that their kernels are not compact when particles with charges of different sign are involved, as it happens in applications to atomic reactions (references can be traced from [3]). The consequence is that naive application of standard solution methods of integral equations theory would not be justified. Additional, practical difficulties arise from the complicated singularity structure of the off-shell twoparticle Coulomb T-matrix which is the basic dynamical ingredient. As is well known, the latter develops nasty singularities in the on-shell limit and, in case of attraction, has in addition an infinite number of poles.