Since 1979, the NONMEM Project at the University of California at San Francisco has been concerned with the development of data analysis techniques and exportable software for estimating the parameters of nonlinear mixed effects (statistical regression-type) models. These techniques are particularly useful when the data are population pharmacokinetic/pharmacodynamic data, and when there are only a few PK/PD measurements from some individuals sampled from the population, or when the regression design varies considerably between individuals.

However, increasingly, the techniques are also being used advantageously with better-designed experimental type data.