DECISION SUPPORT FOR THE DEVELOPING WORLD: A PORTABLE HEALTH POLICY MODEL FOR PNEUMOCOCCAL CONJUGATE VACCINATION

PURPOSE: Pneumococcal Conjugate Vaccine (PCV) has the potential to save lives in low income countries. While global initiatives provide financial assistance for the introduction of PCV, policy makers require analytic assistance in weighing the remaining costs against the potential savings and health benefits. We have developed a portable health policy analysis tool that provides decision support for the introduction of PCV, that is particularly well-suited to these resource-poor settings.

METHODS: A comprehensive decision model for the direct and indirect effects of PCV was constructed and refined through a series of expert panels. Experts in pneumococcal disease, global health and economics reached consensus on the structure of the model, the best available evidence, and the analytic requirements for making policy decisions. Data sets were constructed for each country using estimates from recent global analyses of disease burden and epidemiology. Country-specific evidence was pooled from disparate sources, including unit costs from World Health Organization's CHOICE project and vaccine-related costs from the GAVI Alliance (GAVI). The model was then embedded in a general purpose decision support framework consisting of a series of dynamic web pages that provide complete documentation of all input parameters to the model and a mechanism to allow policy makers to enter new values within reasonable limits. The interface and evaluable model are completely self-contained in this set of web pages so that they can be distributed through media such as CD-ROM for use in places where internet access is unavailable, and is operable on any standards-compliant web browser.

RESULTS: The model has been fully populated with sufficient data to demonstrate the cost effectiveness of PCV for the 72 GAVI-eligible countries and partially populated with data to assist policy makers in the remaining 121 UN member states. The interface allows policy makers to provide additional data to further explore both the healthcare and societal perspectives.

CONCLUSIONS: This decision support tool allows geographically dispersed analysts, with no special training in decision analysis, to use a web browser to apply the decision model to their unique situation. This demonstrates a low-cost method for providing distributed decision support to users in resource-poor settings, using a flexible platform-independent framework that provides comprehensive documentation of an underlying decision model.