ABSTRACT There is observational evidence of a dearth in core-collapse supernova (ccSN) explosions from stars with zero-age main-sequence (ZAMS) mass M0 ≈ 17–30M⊙, referred to as the ‘red supergiant problem’. However, simulations now predict that above 20 M⊙ we should indeed only expect stars within certain pockets of M0 to produce a visible SN explosion. Validating these predictions requires large numbers of ccSNe of different types with measured M0, which is challenging. In this paper, we explore the reliability of using host galaxy emission lines and the H α equivalent width to constrain the age, and thus the M0 of ccSNe progenitors. We use Binary Population and Spectral Synthesis models to infer a stellar population age from MUSE observations of the ionized gas properties and H α EW at the location of eleven ccSNe with reliable M0 measurements. Comparing our results to published M0 values, we find that models that do not consider binary systems yield stellar ages that are systematically too young (thus M0 too large), whereas accounting for binary system interactions typically overpredict the stellar age (thus underpredict M0). Taking into account the effects of photon leakage bring our M0 estimates in much closer agreement with expectations. These results highlight the need for careful modelling of diffuse environments, such as are present in the vicinity of Type II SNe, before ionized emission line spectra can be used as reliable tracers of progenitor stellar age.