When a decision-making process relies on the information provided by a measurement or simulation result, the right decision demands a good quality result, in other words, a low uncertainty result. In order to establish public policies for environmental noise control, it is essential to identify the impact of each type of noise pollution (e.g. road, aircraft and rail transportation noise) on the population affected. One of the noise impact metrics that can be used is the number of highly noise annoyed people in a region whose estimated value is obtained from the corresponding exposure-response function and noise and population density maps. However, an estimated value of the noise impact metric with high uncertainty makes it difficult to realize the actual severity of the problem and its priority in relation to other public health issues. In this work, a Monte Carlo simulation method is used to assess the uncertainty of a noise impact metric result, namely the number of people highly disturbed by road noise in a city. This article also presents a sensitivity analysis of uncertainty sources that allows quantification of the main uncertainty components, which supports improvements in noise impact metric results.