AEGiS-14IAC: Projecting performance measures and costs of alternative HIV rapid test strategies.

14th International AIDS Conference

Barcelona, Spain - July 7-12, 2002

Projecting performance measures and costs of alternative HIV rapid test strategies.

Int Conf AIDS 2002 Jul 7-12; 14:(abstract no. B10167)

Shields JM, Mwinga AZ, Nelson DB
CDC/GAP - Zambia, Lusaka, Zambia

BACKGROUND: Increasing numbers of HIV Rapid Tests present testing protocol optimization challenges to researchers and program specialists. The number of possible protocols exceeds even the combinations of the available tests. The many calculations required to evaluate alternatives mitigate against making analytically based choices. Field evaluation should precede programmatic implementation, but a decision support tool may be used to short list possible protocols for in-country evaluation.

METHODS: An Excel spreadsheet was prepared to facilitate short listing alternative protocols by providing protocol templates and a means of selecting tests from a user modifiable list to model a specific protocol. The user specifies conditions for the selected tests, inputs local performance parameters (or selects best, mid, or worst case published sensitivity and specificities), and stipulates estimated prevalence and number of persons to be tested.

RESULTS: The spreadsheet performs Bayesian and other required statistical calculations to produce EXPECTED net protocol sensitivity and specificity, false negative and false positive counts, number of HIV tests of each type required for the protocol, total cost of tests used, and a measure of the protocol testing bias. During a discussion of the merits of introducing electronic readers at a planning meeting in Zambia, using the tool demonstrated no significant expected marginal benefit for cost, given prior local performance.

CONCLUSIONS: This decision support tool has been useful in ad hoc modeling of selected rapid test strategies and quickly comparing their projected performance. It is flexible in that intermediate calculations and formulae are available for review and adaptation to special cases, and most researchers and program specialists can use spreadsheets. We expect the tool will be useful for those planning to implement new or evaluate existing HIV testing strategies.

Keywords: AEGIS, Costs and Cost Analysis, AIDS Serodiagnosis, Mass Screening, Sensitivity and Specificity, Weights and Measures, Laboratory Techniques and Procedures, Reproducibility of Results, Blotting, Western, Zambia, economics

020707
B10167