A significant body of research has shown a relationship between nurse staffing (in particular, skill-mix: the proportion of Registered Nurses [RNs]) and both morbidity and mortality. This relationship is typically investigated by measuring the incidence of Nursing Sensitive Outcomes (NSOs) under different skill-mix levels. Yet whilst the evidence suggests that richer skill-mix is associated with a lower incidence of NSOs, recent Australian policy reforms have proposed the replacement of Registered Nurses with less qualified staff. The present study sought to examine the relationship between staffing, skill-mix, and incidence of NSOs at two hospitals in one Australian state. The study sought to determine the rate of occurrence of several NSOs, the relationship of skill-mix to that rate, and the number of patients affected per annum. It was found that the current rate of NSOs across wards ranged from 0.17% to 1.05%, and that there was an inverse relationship between the proportion of hours worked by RNs and NSO rates: an increase of 10% in the proportion of hours worked by RNs was linked to a decrease in NSO rates by between 11% and 45%. It was estimated that increasing the RN staffing percentage by 10% would mean 160 fewer adverse outcomes for patients per year across these two hospitals. Importantly, increases in nursing hours overall (without increases in skill-mix) had no significant effect on patient outcomes. These findings challenge current policy recommendations, which propose increasing the number of unregistered staff without increasing skill-mix.