Selecting malaria interventions: A top-down approach
Introduction
Selecting malaria intervention strategies is a complex problem, but one that is highly motivated by the disease's great burden on economic development and human morbidity and mortality. Most studies take a bottom-up approach in addressing the problem. They concentrate on a single subproblem in great detail, such as modeling the economic burden of malaria, or studying the cost-effectiveness of a single intervention. In this study, we take a top-down approach, by combining the subproblems into a single large-scale geographic optimization for selecting malaria interventions.
At a high level, our optimization takes as input (1) data on available intervention strategies (2) ecological and demographic data and (3) an intervention budget constraint. As output, the optimization delivers the location of the supply distribution centers and a geographic plan of intervention delivery that minimizes the disease's impact, subject to the budget constraint. The optimization can be used to construct an efficient frontier, depicting the achievable morbidity or mortality at each budget value.
Geographic epidemiological intervention models have been studied in the past. Most of these models consider a small number of intervention strategies (typically less than 5), e.g., contact tracing and quarantine, and then simulate their effect across the entire geographic region [1], [2], [3], [4]. Large-scale intervention optimization, where a combinatorially large set of intervention strategies is considered, is an active area of research. Recently, large-scale intervention optimization has addressed the distribution of a limited stockpile of vaccines to halt pandemic influenza [5], however that study is not geographically targeted. Little work exists in the area of geographically targeted, large-scale intervention optimization [6]; despite explicit calls for the computation of geographically targeted malaria interventions [7]. This paper's main contribution is a novel large-scale, geographically targeted malaria intervention decision support tool, derived using a top-down approach. The output of the optimization model is a fine-grained geographic intervention plan, specifying a set of actions for each geographic sub-region.
While our top-down optimization approach is able to produce detailed geographic intervention plans, it should be regarded simply as a decision support tool. The complexity and main contribution of a top-down approach comes in combining subproblems to produce the final intervention decision. The drawback of a top-down approach is that each subproblem may be considered in less detail than a typical bottom-up study dedicated solely to an individual subproblem. Our optimization attempts to capture as many aspects of the intervention selection problem as is both (1) computationally tractable and (2) reasonably quantifiable. No model or optimization can capture all the salient factors for a problem as complex as malaria, in which even cultural factors may be definitive in determining the effectiveness of an intervention strategy. Expert knowledge is critical to decision-making; the role of decision support tools is to make the task of decision makers easier by providing quantitative data on the effects of various strategies.
We structure the rest of the paper as follows. First, in Section 2 we outline the structure of our top-down approach. In the Section 3, we cover in detail the execution of the approach for our case study of malaria in Nigeria. Finally, in Section 4, we draw some conclusions.
Section snippets
Materials and methods
In this section we describe our top-down approach in detail. In Section 2.1, we present the stages, i.e., the building blocks of our approach. In Section 2.2, we focus on the mathematical methodology at the core of the optimization.
Results: a case study of malaria in Nigeria
We illustrate the top-down approach by considering malaria intervention in Nigeria. In each of the following subsections, we describe the stages of the approach as applied to this case study.
Discussion
We propose a top-down approach for suggesting disease intervention strategies across a given geographic region. A top-down approach combines models for many subproblems into one. In contrast, most studies take a bottom-up approach, where a detailed model for a single subproblem is developed. The benefit of a top-down approach is that it results in a decision support tool. The drawback of a top-down approach is that to maintain tractability, often, the models for each of the individual
Acknowledgments
This work has been supported by the National Science Foundation through grants CMMI-0653916 and CMMI-0800676 and the Defense Threat Reduction Agency through grant HDTRA1-08-1-0029.
References (66)
- et al.
Optimal policies for controlled Markov chains with a constraint
J Math Anal Appl
(1985) Naturally acquired immunity to Plasmodium falciparum malaria in Africa
Acta Trop
(2005)- et al.
The epidemiology of severe malaria in an area of low transmission in Thailand
Trans R Soc Trop Med Hyg
(1997) - et al.
The economic impact of malaria in Africa: a critical review of the evidence
Health Policy
(2003) - et al.
Annual Plasmodium falciparum entomological inoculation rates (EIR) across Africa: literature survey, internet access and review
Trans R Soc Trop Med Hyg
(2000) - et al.
Taking malaria transmission out of the bottle: implications of mosquito dispersal for vector-control interventions
Lancet Infect Dis
(2003) Cost-of-illness analysis: what room in health economics?
Health Policy
(2006)- et al.
Modeling the worldwide spread of pandemic influenza: baseline case and containment interventions
PLoS Med
(2007) - et al.
Strategies for containing an emerging influenza pandemic in Southeast Asia
Nature
(2005) - et al.
Containing bioterrorist smallpox
Science
(2002)
Smallpox transmission and control: spatial dynamics in Great Britain
PNAS
Optimizing influenza vaccine distribution
Science
Optimizing tactics for use of the U.S. antiviral strategic national stockpile for pandemic influenza
PLoS ONE
Spatial targeting of interventions against malaria
Bull World Health Organ
Malaria in Africa: vector species niche models and relative risk maps
PLoS ONE
An anatomy of risk
Spatial analysis in epidemiology
Biogeography of diseases: a framework for analysis
Naturwissenschaften
Combinatorial design of a stochastic Markov decision process
Markov decision processes: discrete stochastic dynamic programming
Constrained Markov decision models with weighted discounted rewards
Math Oper Res
Randomized and past dependent policies for Markov decision processes with finite action set
Oper Res
Survey of linear programs for standard and nonstandard Markovian control problems. Part I: theory
Math Methods Oper Res
Very high resolution interpolated climate surfaces for global land areas
Int J Climatol
Estimated global resources needed to attain international malaria control goals
Bull World Health Organ
Indicators of life-threatening malaria in African children
N Engl J Med
Cited by (5)
Literature review: The vaccine supply chain
2018, European Journal of Operational ResearchCitation Excerpt :The allocation of vaccines has been studied for a broad range of diseases, which is also apparent from the papers that apply cost-effectiveness analyses to analyze vaccine allocations. Already in the OR/OM community there are studies on hepatitis B (Hutton et al., 2011), HIV (Edwards et al., 1998; Rauner, 2002), malaria (Dimitrov et al., 2013; Parker, 1983), polio (Thompson, Tebbens, Pallansch, Wassilak, & Cochi, 2015), rabies (Frerichs & Prawda, 1975), and tuberculosis (Denysiuk et al., 2015; Reveller et al., 1969). In most cases, the goal of a vaccination program is to contain an outbreak.
A hierarchical compromise model for the joint optimization of recovery operations and distribution of emergency goods in Humanitarian Logistics
2014, Computers and Operations ResearchCitation Excerpt :Our future research plans include the use of RecHADS to devise recovery operations in Jipijapa, a canton of Ecuador frequently affected by heavy rains and landslides, in collaboration with an international NGO. This may require extending RecHADS to capture other complex features arising from the necessities of the case considered, such as identifying the best locations for distribution centers [8]. Also, we are planning to take a step forward and develop a stochastic model based on RecHADS for the protection of the transportation infrastructure prior to disasters [17,16,18–20,27].
Optimal Covid-19 vaccine stations location and allocation strategies
2023, BenchmarkingThe endangerment of Mendriq's traditional knowledge in Malaysia
2019, International Journal of Sustainable SocietyMulti-year optimization of malaria intervention: A mathematical model
2016, Malaria Journal