The case for using flexible healthcare capacity constraints to optimise pandemic control strategies

During the COVID-19 pandemic, a key challenge was designing control strategies that balance the benefits and costs of public health and social measures while preventing healthcare systems becoming overwhelmed. This was often implemented by epidemiological modellers as a binding (hard) constraint based on the maximum number of hospital beds that could be occupied at any given time. However, experience from the pandemic demonstrated that healthcare capacity depends not only on the number of beds available, but also on staffing and the availability of other resources. We argue that defining healthcare capacity using a single number (beds available) does not adequately capture pressures in healthcare settings as high occupancy is sustained. We therefore introduce a framework for implementing flexible (soft) constraints on healthcare capacity by allowing the cost of control strategies to depend continuously on intensive care unit (ICU) occupancy. We illustrate pandemic scenarios where using a soft constraint is better to capture long-term pressures on the healthcare system. Additionally, we highlight that explicitly accounting for uncertainty is essential to choose a robust control strategy. For optimal control studies to provide the best evidence to policy for future pandemics, they must carefully consider how their healthcare capacity constraints are implemented.