A systematic approach to produce robust, comparable, and timely cost-effectiveness estimates for a set of interventions: proof of concept in two low-income countries

Abstract

Under the current paradigm, cost-effectiveness studies provide limited value to policymakers in low-resource settings. Studies appear with substantial delays in the academic literature and are often based on large-scale multi-intervention assessments in settings with drastically different infrastructure, resources, and cultures. Timely and contextual evidence is rarely available. Given recent developments in standardizing the analysis of the global burden of disease (GBD), we believe a similar approach can be applied to the generation of cost-effectiveness estimates. To achieve this, we are developing a systematic protocol and guidelines for conducting cost-effectiveness analyses based on the integration of information. We are applying this approach to two low-income countries – Kenya and Zambia – as a proof of concept.

Methods

We define cost-effectiveness as a combination of five inputs: incremental costing, the current coverage of interventions, the remaining burden of disease that needs to be addressed, efficacy of interventions, and the gap between efficacy and effectiveness, which we label as quality. The first step is to identify a set of interventions based on highest potential impact and strategic priorities of the two countries involved. The list of interventions for Kenya is currently being finalized. To develop cost functions, we will use data collected through the Access, Bottlenecks, Costs, and Equity (ABCE) project that incorporate facility-level efficiency. GBD estimates will be used to determine the burden. We will initially develop first-order approximations of coverage based on available survey data, or encounter data for interventions that are not normally included in demographic health surveys. We will map from efficacy in the units reported in the literature to changes in disability-adjusted life years (DALYs), checking for consistency with GBD assumptions regarding prevalence, case-fatality rates, severity distributions, and disability weights. To account for the impact of provider quality and consumer behavior on the real-world effectiveness of interventions, we are collaborating with Emory University in developing a framework to estimate effectiveness and its determinants.

Findings

Bringing together data on the five inputs will allow us to produce estimates of the cost-effectiveness of the interventions of interest to policymakers in Kenya and Zambia. We aim to produce our first round of estimates in 2015 for a subset of those interventions.

Interpretation

Developing a system that is able to generate timely, evidence-based, setting-specific, and up-to-date estimates of cost-effectiveness for each country will take multiple iterations. Ultimately, the aim is to be able to determine the fraction of each disease that can be averted over a defined period with policies that meet certain threshold definitions of cost per DALY averted, while incorporating uncertainty.

Funding

Disease Control Priorities Network through the Bill & Melinda Gates Foundation.