2. Total cost implications of providing anti-retroviral therapy

The unit costs presented in 1. can be used to estimate what the annual per patient cost of providing ARVs might be. While the drug costs are fairly uniform among countries and will only differ according to freight/customs charges, the costs of the non-drug components are likely to vary substantially - as the costs for a day in hospital shown above indicate. Since most of the non-drug cost data are from the USA, this means that cost estimates cannot be very precise. This problem is aggravated by the fact that approaches to case management are not well defined for many countries - for example the frequency and type of laboratory tests which would be appropriate, how often monitoring visits would be required, etc. This makes costing difficult, since the required resource inputs cannot be clearly identified.

Moreover, a portion - perhaps even a majority - of patients will not be able to tolerate the drugs and will suffer adverse effects (Hay et al, 1988), or will decline to take them at all (Alcorn 1995). This means that the average annual cost of ARV therapy per HIV infected person may be either more or less than the average annual cost for a patient who tolerates the therapy without any problems. It will be more if the costs of patients suffering adverse effects in a given year exceeds the cost of those patients continuing with therapy for that year; it will be less if the cost of adverse effects is less than the costs which would be incurred were those patients suffering from adverse effects to continue therapy.

Due to these difficulties, a simplified cost analysis is given here. The annual cost for a patient receiving ARV is assumed to be the cost of the drugs, the cost of four outpatient visits (with a range from 1 to 120), the cost of four complete blood cell counts (with a range of 2 to 21), four CD4 counts (with a range from 30 to 157), four viral load tests (a frequency suggested in the minutes of the February 26th 1997 meeting of the "HIV physicians forum for the North West Region in the UK") at US$163 each, and four chemistry panels (with a range from 12 to 35). The range in costs for the final three components reflects the range shown in the Table 3 cost data. The range in outpatient costs uses the highest figure for the USA shown in Table 3 as an upper estimate, but US$1 is used as a lower estimate since the day in hospital cost for Malawi suggests that in very poor countries this may be a realistic figure. Counselling costs and HIV test costs are ignored since these are one-off costs and are minor in comparison with other cost components. The cost impact of adverse effects is also ignored since there are so few data which can provide guidance. However, it is worth noting that in the USA it has been suggested that adverse effects will substantially increase ARV therapy costs (Bozette et al, 1994). It has also been suggested that outpatient costs for HIV-infected people are higher than those for non-infected patients (McDermott et al, 1991), and both the minimum and maximum estimates used here may therefore underestimate the cost of an OPD visit for patients receiving ARV which would be incurred in practice.

Using the above assumptions, the per patient annual cost for ARV therapy would range from

US$ 3 570 to US$ 4 722 for AZT therapy, and from US$ 8 776 to US$ 13 902 for triple-combination therapy (excluding a regimen which includes ritonavir). It is noteworthy that the drugs constitute between 58% and 77% of total per patient annual costs with AZT therapy, and between 86% and 91% of total per patient annual costs with triple combination therapy. This illustrates that if drug costs can be reduced, as has already happened with AZT, and as as been demonstrated to be possible in Uruguay, a substantial impact will be made on the affordability of therapy. Partnerships among international agencies, individual governments and drug companies may also have potential to have a large impact on the cost of drugs. However, at present, at market prices, even the cost of AZT therapy is higher than average incomes in most countries.

To estimate the total annual cost which would be incurred if ARV therapy were to be provided in different geographic regions, these figures are then used in combination with 1996 data concerning the number of people with HIV infection but not AIDS, and the number of people with AIDS (taken from figures quoted in Mann and Tarantola, Chapter 1, 1996). Two scenarios are presented. In the first, everyone receives ARV therapy - though unrealistic, it is useful in providing an upper total cost estimate; in the second, 50% or those eligible are assumed to receive therapy. The results are shown in Table 4 below. When considering these figures, it is worth bearing in mind that it has been calculated that implementation of six major prevention strategies in developing countries would in total cost between US$ 1.5 and US$ 2.9 billion (Broomberg and Schopper, 1996).