<p>Determining staffing levels is essential to ensure appropriate capacity for delivering the services required by the population. This task is complex due to uncertainties in both the volume and type of demand. Deterministic approaches that use the average expected demand can lead to significant imbalances between the suggested staffing levels and the actual demand. Therefore, it is necessary to explore innovative methodologies that can account for uncertainty and provide robust staffing level definitions. In this paper, we address the staffing level problem in the healthcare context, focusing on estimating how future patient demand translates into resource usage. To this end, we propose a four-step methodology. Firstly, patients’ needs are modeled as service pathways, which determine not only the sequence but also the specific timing at which a patient’s care requires the allocation of medical resources (human and/or material). Secondly, patient arrivals are simulated, and each patient is randomly assigned a service pathway. Thirdly, the workload for each type of resource during each time period is computed by aggregating the needs of all patients within that period. Finally, a scenario-based method, the Sample Average Approximation (SAA), is used to determine resource levels for each period in the planning horizon. Numerical experiments based on randomly generated instances demonstrate the feasibility of the proposed approach and highlight its potential usefulness in supporting managers in their decision-making processes.</p>