Health systems face nonfinancial constraints that can influence the opportunity cost of interventions. Empirical methods to explore their impact, however, are underdeveloped. We develop a conceptual framework for defining health system constraints and empirical estimation methods that rely on routine data. We then present an empirical approach for incorporating nonfinancial constraints in cost-effectiveness models of health benefit packages for the health sector.
We illustrate the application of this approach through a case study of defining a package of services for tuberculosis case-finding in South Africa. An economic model combining transmission model outputs with unit costs was developed to examine the cost-effectiveness of alternative screening and diagnostic algorithms. Constraints were operationalized as restrictions on achievable coverage based on: (1) financial resources; (2) human resources; and (3) policy constraints around diagnostics purchasing. Cost-effectiveness of the interventions was assessed under one “unconstrained” and several “constrained” scenarios. For the unconstrained scenario, incremental cost-effectiveness ratios were estimated with and without the costs of “relaxing” constraints.
We find substantial differences in incremental cost-effectiveness ratios across scenarios, leading to variations in the decision rules for prioritizing interventions. In constrained scenarios, the limiting factor for most interventions was not financial, but rather the availability of human resources.
We find that optimal prioritization among different tuberculosis control strategies in South Africa is influenced by whether and how constraints are taken into consideration. We thus demonstrate both the importance and feasibility of considering nonfinancial constraints in health sector resource allocation models.