First of all it must be clarified that a lattice is a collection of geometrical point all characterized by a same surrounding (that is, each lattice point has the same number of nearest neighbour lattice points arranged with the same point distances and angles).
Hence, widespread definitions such as “lattice distortion”, “lattice dynamics”, “lattice energy” and so on are misleading, but unfortunately commonly used; in this case the term “lattice” should be substituted by “structure”. A crystal structure is obtained by the convolution of a lattice with a basis (i.e. an atom of a group of atoms).
Having clarified this, structural distortion can be evaluated by diffraction. Distortion can be dynamic or static. Pair distribution function analysis provides a better estimate of the Debye-Waller parameters, whereas the widths of the peaks give the bond distance distribution of the considered atomic pair. But this last information characterizes the local scale, not the structural parameters.
Diffraction line (peak) broadening can be used for evaluating the structural strain, that is the fluctuation of the inter-planar distances along the different crystallographic directions. Simplifying, the components of the diffraction line profile have a $\tan(2\theta)$ dependence on the scattering angle $\theta$ by means of which strain along the different directions can be valuated.
The qualitative analysis of the strain distribution is rather easy from high quality diffraction data, but an accurate quantitative evaluation is not. This is not a predictive method (neither is PDF analysis), but gives you qualitative information about the structural distortion.
