The analysis presented above has been based on the best estimates available for each of the key variables used in the economic model. In order to test the robustness of the results, sensitivity analysis has been conducted on a number of the variables affecting the outcomes. These are:
- Halving the rate of effect of NSPs on HIV. This analysis seeks to address the issue of the extent to which NSPs contribute to the reduction in HIV as opposed to other concomitant activities (see Section 3.1.7).
- Quartering the effect of NSPs on HIV. This analysis further extends the examination of reduced NSP effects on HIV.
- Doubling the level of investment in NSPs over the ten years. This analysis examines the result of increasing the expenditure on NSPs without any increase in effect on HIV. By so doing, it takes into account the potential contribution of the commercial pharmacy market.
- Halving the annual treatment costs for HIV. This analysis considers the results of possible future reductions in the costs of HIV treatment.
The analysis indicates that the outcomes previously presented are most sensitive to the impact of NSPs on HIV incidence. This is to be expected because of the nature of the estimation technique employed, which uses the logit scale as its base. Consequently, halving the rate of effect of NSPs on HIV incidence has a proportionally greater effect on the number of cases avoided over time. Nevertheless, even at the most conservative estimate of effect (one-quarter of the original effect estimate) the return on investment on both government expenditure and total expenditure on NSPs is positive. This also holds true for variations to the other input variables in the model. Of some interest is the fact that even when the annual costs of treatment for HIV are halved, NSPs continue to meet the required investment criteria. The sensitivity analysis indicates that the results presented are robust, and that the return on investment from NSPs is positive in all other tested scenarios.
Table 4.8.1 Net Present Value, 1991 ($million, year 2000 prices) of investment in NSPs for HIV – Sensitivity Analysis
|Lifetime costs of treatment||Government expenditure||Total expenditure|
|Half NSP effect on HIV|
|Quarter NSP effect on HIV|
|Double NSP investment|
|Half HIV annual treatment costs|