We propose to analyse the development of a disease by sanogenesis curves which model the result of interacting exitatory and inhibitory factors on a disease. Assuming that high-dimensional data describing the disease course are driven by a latent complex-valued Gaussian process with Markovian structure, we can identify the sanogenesis curve as the real part of the covariance function of the latent process. By applying techniques of stochastic process theory and partially inverse functions theory this finding allows to estimate the model parameters. In addition, the sanogensis curve also suggests a new model for survival times, where failures (deads) are only observed during critical time periods (crises) defined by the sanogenesis curve. We illustrate our approach by analyzing two real data sets from medicine.