Abstract

‍
Achieving a balanced healthcare workforce requires health planners to adjust the supply of health human resources (HHR). Mathematical programming models have been widely used to assist such planning, but the way uncertainty is usually considered in these models entails methodological and practical issues and often disregards radical yet plausible changes to the future. This study proposes a new socio-technical methodology to factor in uncertainty over the future within mathematical programming modelling. The methodological approach makes use of foresight and scenario planning concepts to build tailor-made scenarios and scenario fit input parameters, which are then used within mathematical programming models. Health stakeholders and experts are engaged in the scenario building process. Causal map modelling and morphological analysis are adopted to digest stakeholders and experts’ information about the future and give origin to contrasting and meaningful scenarios describing plausible future. These scenarios are then adjusted and validated by stakeholders and experts, who then elicit their best quantitative estimates for coherent combinations of input parameters for the mathematical programming model under each scenario. These sets of parameters for each scenario are then fed to the mathematical programming model to obtain optimal solutions that can be interpreted in light of the meaning of the scenario. The proposed methodology has been applied to a case study involving HHR planning in Portugal, but its scope far extends HHR planning, being especially suited for addressing strategic and policy planning problems that are sensitive to input parameters.

‍

Highlights

‍
• We develop a comprehensive methodology that addresses uncertainty in mathematical programming models through foresight and scenario-building.

• This methodology is socio-technical, being especially suited for strategic and policy problems.

• We illustrate this methodology by applying it to the planning of Human Health Resources in Portugal.

• The main outputs of the methodology are then used to model uncertain parameters in a MILP optimization model that sets optimal medical vacancies.

• We show how this methodology is capable of capturing different sensitivities from relevant stakeholders and experts, going far beyond traditional approaches such as robust or stochastic programming.

‍

Read the research paper here