 The study assesses the impact of dimensionality reduction on COVID-19 mortality prediction models using standard machine learning classifiers on a geographically diverse sample of 212 patients with populated entries for all 25 available features. The results show that feature reduction provides strong benefits for performance, improving accuracy and reducing variable sourcing burden at hospital admission with little performance loss. Extreme feature reduction to a single most salient feature demonstrates large standalone explanatory power, highlighting the importance of feature reduction in future model construction and the feasibility of deprioritizing large, hard-to-source and non-essential feature sets in real world settings. This article was authored by Ricardo Doyle.