Results: Financing DOTS expansion
NTP budgets, total costs of TB control, and government contributions among HBCs, 2002 and 2003
The NTP budgets of the HBCs for the fiscal year 2003 totalled US$ 430 million, excluding South Africa and Zimbabwe, which provided no data (Table 17). This was lower than the 2003 budget estimate of US$ 481 million for the HBCs reported in Global Tuberculosis Control 2003,5 which included budget figures for Zimbabwe but not for South Africa. The difference is largely due to a change in the budget for the Russian Federation. No data were provided by the Russian Federation MoH in 2002; instead, we used an estimate of US$ 200 million based on recent costing studies.18 For this report, the Russian Federation provided data for 2003 for all expenditures at federal level and for staff expenditures at oblast (regional) level, which totalled US$ 125 million. The oblast data did not include all items funded locally. If these were included, the estimate would probably be similar to that in last year's report.
The total costs of TB control were calculated for the HBCs by adding the 2003 NTP budgets to the costs associated with TB control that were not financed through the NTP (e.g. salaries of health workers and infrastructure costs). The total costs for the HBCs in 2003 were estimated at around US$ 1 billion (Table 17). This is an increase of around US$ 150 million (about 18%) from 2002 expenditures of US$ 834--884 million.
The total expenditures for 2002 were lower than the estimates published in last year's report (US$ 976 million).5 This is primarily because the estimates in last year's report assumed that the number of patients treated by NTPs would be consistent with the progress needed to reach 70% case detection in 2005. In practice, the 2002 notification data show that they treated fewer cases. There was no consistent association between the change in case load between 2002 and 2003 (cases detected by countries in 2002 compared with cases expected in 2003) and the change in costs (2002 expenditures compared with 2003 budgets). All countries that reported 2002 expenditures and 2003 budgets reported an increase in costs in 2003.
The total cost per patient treated in the HBCs in 2003 ranged from US$ 65 in Myanmar to US$ 1419--1824 in the Russian Federation (Table 17). The median total cost per patient was US$ 199 (China). Fourteen of the countries (74% of those providing data) had costs per patient between US$ 125 and US$ 380. Two countries had costs per patient below US$ 100 (India and Myanmar). Three countries stand apart from the rest: in Brazil, the Russian Federation and South Africa, the costs per patient were above US$ 700. Patient care is expensive in these countries partly because the prices of labour and capital are higher, linked to higher GNI (Figure 17). However, costs are also inflated by the heavy reliance on hospital treatment and expensive diagnostic techniques in South Africa and the Russian Federation. In Thailand, the cost per patient was low (US$ 198) relative to GNI (US$ 1980). One explanation is that patients make relatively few visits to clinics during treatment (12 on average).
Between 2002 and 2003, the costs per patient changed little in India and the Philippines (Table 17). This is because, while geographical expansion of DOTS has proceeded rapidly in India, the strategy for implementation has remained the same. The Philippines was already close to full DOTS coverage in 2002.
In other countries, the cost per patient increased markedly between 2002 and 2003 for reasons that differed among countries. Costs have increased in Vietnam because a national prevalence survey was included in the 2003 budget, and because the NTP is expanding to remote areas where detecting and treating cases is more difficult. In Bangladesh, the budget for fiscal year 2003 includes substantial funding for new initiatives such as improvement in the quantity and quality of diagnostic services and training. In Myanmar, the increase is due to large planned investments in vehicles for supervision and in diagnostic equipment. This adds considerably to costs in the year in which these items are bought, though the benefits will be spread over several years.
There will inevitably be delays between investments in TB control and the consequent increase in case detection. For example, where NTPs (e.g. Kenya) have introduced new initiatives to increase case detection -- by involving, for example, the private sector, HIV/AIDS control programmes, or lay members of communities who can recognize TB symptoms and supervise treatment -- the yield in new TB cases will not be immediate. Another possible explanation for increasing per patient costs is that the targets for case detection set by NTPs are not sufficiently ambitious relative to the large increases in their budgets.