Trading off health and financial protection benefits with multiobjective optimization

Abstract

Countries which are introducing a system of Universal health coverage have to make a number of key tradeoffs, of which one is the tradeoff between the level of coverage and the degree to which patients are exposed to potentially catastrophic financial risk. In this study, we first present a way in which decision makers might be supported to focus on in a particular part of the tradeoff curve and ultimately choose an efficient solution. We then introduce some multiobjective optimization models for generating the tradeoff curves given data about potential treatment numbers, costs, and benefits. Using a dataset from Malawi, we demonstrate the approach and suggest a core index metric to make specific observations on the individual treatments. Moreover, as there has been some debate about the best way to measure financial exposure, we also investigate the extent to sensitivity of our results to the precise technical choice of financial exposure metric. Specifically, we consider two metrics, which are the total number of cases protected from catastrophic expenditure and a convex penalty function that penalizes out‐of‐pocket expenditures in an increasingly growing way, respectively.

Balancing health and financial protection in health benefit package design

“…We did not formally account for health disparities (e.g., distribution of health outcomes across socioeconomic groups) or priority to the worse‐off or special disadvantaged groups. We also limited our health outcome to deaths averted omitting DALY estimation (see, e.g., Karsu & Morton, 2021 for the use of DALYs as a health outcome measure). Deaths averted, as a health outcome, was chosen for local policy relevance (e.g., high mortality levels in Ethiopia), simplicity of exposition, and due to paucity of data.…”

“…This work is grounded in a growing literature that calls for or applies mathematical optimization and operations research methods to resource allocation problems in health policy (examples include Crown et al., 2017 ; Epstein et al., 2007 ; Kerr et al., 2015 ; Ochalek et al., 2020 ; Smith & Yip, 2016 ; Verguet, 2013 ; Wilson & Blower, 2006 ; and most recently Karsu & Morton, 2021 ). With the move toward UHC as a core component of the SDGs, revising and updating national essential health benefit packages will become an increasingly important policy milestone in many LMICs (Verguet et al., 2021 ).…”

“…We considered a range of constraint values ( K ) for the second objective that covered the entire feasible funding space to generate the Pareto efficiency frontier; and K was incrementally changed by single units of either deaths averted or CHE cases averted. The technical appendix of Karsu and Morton (Karsu & Morton, 2021 ) provides a detailed explanation for implementation of the epsilon constraint method which can efficiently modify the value of K iteratively (see also Ehrgott, 2005 ). The analysis used R statistical software (version 3.5.0; www.r‐project.org ).…”

“…We illustrate our optimization approach with the case study of selected interventions in a low‐income country context. Recently, Karsu and Morton developed similar methods for trading off health and FRP benefits with multi‐objective optimization in the context of Malawi (Karsu & Morton, 2021 ). The work we present here builds on Karsu and Morton's contribution by focusing on the practical aspects of intervention selection, parameterization requirements and translation of WHO's UHC recommendations in the context of Ethiopia.…”

How to do (or not to do)… health resource allocations using constrained mathematical optimization

“…Both CEA and constrained optimization methods can be used to inform decisions within the context of a more holistic health system review, which may also assess factors such as purchasing/procurement arrangements, supply chain management, the funding landscape, equipment allocation and use and overall patient satisfaction, among others ( Crown et al. , 2018 ; Verguet et al ., 2021 ; Karsu and Morton, 2021 ).…”