Value of Survey Information for Geographical Targeting Interventions

Subnational Estimates Working Group of the HIV Modelling Consortium. District-level Map of Kenya -Forthcoming

    
This project, which is led by University of York, is the second element of a two-phased work package investigating how program planners may use data or methods that describe subnational heterogeneity of HIV epidemiology when developing their response to the epidemic. The first element focused on the development of optimal methodologies for determining the spatial heterogeneity based on current data, while this project aims to identify the value of collecting more precise estimates of epidemiological data on HIV so as to reduce decision uncertainty and errors.


Background

A uniform national policy for HIV interventions within a country may not lead to the most efficient use of resources and recognising and identifying the heterogeneity between areas could result in improved health outcomes. Previous work by Anderson et al. in Kenya has demonstrated how geographic and epidemiological data can be used to identify foci of HIV transmission within countries (Anderson et al.). If resources were then prioritized to these foci (i.e. increased stratification of populations) large efficiency gains could be made (i.e. more health gain could be generated from the same budget). The work raised the issue that a uniform national policy for HIV interventions within a country may not lead to the most efficient use of resources and that potentially large gains in health outcomes could be realised by recognising and identifying the heterogeneity in HIV epidemics between areas.

This work raises two core issues in the prioritization of resources for HIV programs. Firstly, there is likely to be value in targeting different populations with different programs (i.e. stratifying care) to increase overall health gain (e.g. where populations are treated differently according to their locality). Secondly, there may be value in generating surveillance evidence on the epidemiology of HIV, whether area specific or national level evidence, to reduce decision uncertainty. This project will aim to build upon the earlier work to identify the value of increasing the stratification of populations to increase overall health gain; this is often termed the value of heterogeneity (Espinoza, 2012). Further it will also consider the value of collecting more precise estimates of epidemiological data on HIV so as to reduce decision uncertainty and therefore to reduce the cost of decision errors (Claxton, 1999).

Approach

The Secretariat, in conjunction with collaborators from University of York will produce a conceptual framework outlining how to identify the value of increased stratification of care and the value of collecting additional epidemiological data to reduce decision uncertainty. This will be approached in three stages:

  • Firstly, the value of allowing for greater stratification in programming decisions across treatment, prevention and all program components based on current evidence. 
  • Secondly, to consider the value of obtaining additional HIV related surveillance data conditional on the level of stratification of programming decisions across treatment, prevention and all program components. 
  • Finally to consider different levels of stratification together with associated values of collecting additional epidemiological surveillance data as a means to demonstrate how the value of information depends upon the feasible level of stratification. This can be termed the dynamic value of stratification. This will combine the first and second stages outlined previously. 

In all of these scenarios the impacts of different levels of ‘transaction costs’, those costs associated with changing the current configuration of HIV services, including those related to fully or partially irreversible configurations of services, will be considered. Geographic variation in costs will also be incorporated, where data are available.

References

Anderson, S., Hallet, T., Cremin, I. & Borquez, A. Year. Using information on epidemic heterogeneities in resource allocation. In.
Claxton, K. 1999. The irrelevance of inference: a decision-making approach to the stochastic evaluation of health care technologies. Journal of Health Economics, 18, 342-364.
Espinoza, M. Year. The value of heterogeneity for cost-effectiveness subgroup analaysis. In: ISPOR, 2012 Berlin.