Energy dissipation is highly intermittent in turbulent plasmas, being localized in coherent structures such as current sheets. The statistical analysis of spatial dissipative structures is an effective approach to studying turbulence. In this paper, we generalize this methodology to investigate four-dimensional spatiotemporal structures, i.e., dissipative processes representing sets of interacting coherent structures, which correspond to flares in astrophysical systems. We develop methods for identifying and characterizing these processes, and then perform a statistical analysis of dissipative processes in numerical simulations of driven magnetohydrodynamic turbulence. We find that processes are often highly complex, long-lived, and weakly asymmetric in time. They exhibit robust power-law probability distributions and scaling relations, including a distribution of dissipated energy with power-law index near -1.75, indicating that intense dissipative events dominate the overall energy dissipation. We compare our results with the previously observed statistical properties of solar flares.