Abstract: Neutrino oscillations are the first and only observed phenomena beyond the standard model, and measuring the parameters which govern them is a major area of current research. The textbook theoretical treatment of neutrino oscillations, however, is known to be both approximate and inconsistent. The range of its validity is debated, with some estimates suggesting that neutrino coherence loss could occur in existing experiments, especially if heavier, sterile neutrinos participate in the oscillation. All calculations of these effects to date require an assumption about the initial neutrino state, usually obtained by dimensional analysis, and spanning several orders of magnitude between treatments. In this talk I will show that derivation of neutrino coherence properties without an arbitrary initial-state assumption is possible if the neutrino beam is properly treated as an open quantum system. I will present the first such calculation, where the full oscillation phenomenology of neutrinos produced in accelerator neutrino beams is derived without arbitrary spatial scales at the neutrino production vertex. The resulting range of validity of the typical neutrino oscillation formalism will then be discussed.