This paper presents a proposal for optimizing the public health service in the state of Parana in terms of the flow of patients within the state's boundaries and the regionalization (division) of the state into new hierarchical configurations for this service. In terms of regionalization, the proposal consists of dividing the state into smaller regions comprising several towns connected to a headquarter city, which would be responsible for decisions affecting the region. With regard to the flow of patients, an algorithm is proposed which organizes the information while simultaneously optimizing the flow. The state was regionalized using a branch-and-price algorithm, which uses a column generating algorithm in each node of a branch-and-bound tree. The technique proposed here to optimize the patient flow proved effective and useful, controlling not only the medical procedures performed in each town but also defining the city to which each patient should be taken, according to the hierarchical division of the state. The branch-and-price algorithm used to optimize the regionalization of the state is very interesting, for it improves the hierarchical division of the state taking into account the number of inhabitants and the number of medical procedures in each city of the state. The results have met the expectations of State Health Services.
P-medians problem; Branch-and-price algorithm; Patient flow; Hierarchical configuration
