The existence of asymptomatic individuals with mild symptoms represents an additional challenge for the control of the coronavirus disease 2019 (COVID-19) pandemic. This challenge puts pressure on public management officials involved in the design of infrastructure for service provision as well as isolation procedures. The challenge is even greater in a country such as Brazil, which has large physical dimensions with the assumption of free movement. Considering this scenario, this work presents a new proposal for estimating health system use for COVID-19 cases, as well as a methodology for using the model in the real world. The estimate was obtained by modifying the dynamic model known as Susceptible, Infected, Removed and Dead (SIRD), including a parameter to model not all cases but only the health system dynamics. The model was tuned from the data available for each state and updated day-by-day, establishing a figure of merit to assess the quality of the model and determining the free parameters that best fit the model to the data. The proposed model and the respective tuned parameters were validated considering the data available for the 26 Brazilian states, demonstrating strong adherence in most cases and allowing the estimation of an epidemic model for the whole of Brazil, which was obtained via the linear combination of the models for each state. In addition to the effective use of the health system, the incidence rate and removal rate were analysed, as was the reproduction rate: baseline R0 and effective Rt. In the specific case of Brazil, the states that make up the federation have autonomy in decision making, which increases the complexity of the analysis of the evolution of the pandemic. With the proposed global model, the method used to tune the parameters and the available results, there was heterogeneity in the dynamics observed for each state, which is compatible with some characteristics of the real-world scenario.