The kinetic Monte Carlo (KMC) method provides a versatile tool to investigate the mechanisms underlying photocurrent generation in nanostructured organic solar cells. Currently available algorithms can already support the development of more cost-efficient photovoltaic devices, but so far no attempt has been made to test the validity of some fundamental model assumptions and their impact on the simulation result. A meaningful example is given by the treatment of the electrostatic interactions. In most KMC models, electrostatic interactions are approximated by means of cutoff based potentials, irrespective of the long-range nature of the Coulomb interaction. In this paper, the reliability of such approximation is tested against the exact Ewald sum. The results under short-circuit and flat-band conditions show that use of cutoff-based potentials tends to underestimate real device performance, in terms of internal quantum efficiency and current density. Together with this important finding, we formalize other methodological aspects which have been scarcely discussed in the literature.