DRC has a surface area of 2.345.000 km² and ≈60 million inhabitants, for a density of 25 inhabitants/km². Administratively, DRC is subdivided into 11 provinces, and HAT is endemic in 9 of them (Programme National de Lutte contre la Trypanosomiase Humaine Africaine [PNLTHA], unpub. data). Since 1990, the country has been devastated by political turmoil and civil war (1996–1997 and 1998–2003). The health status of the population has deteriorated because of progressive breakdown of health infrastructures, disease outbreaks, and the reemergence of endemic diseases such as tuberculosis and HAT. The emergence of HIV/AIDS has added to this catalog of health disasters.

HAT control in DRC is organized by a national program, PNLTHA. This program divides HAT-endemic areas into 7 regions, each under the responsibility of a regional coordinator (Figure 1). These regions do not coincide with the administrative divisions (provinces).

The main control strategy of PNLTHA is to actively screen the population at risk by specialized mobile teams (15), who refer patients with confirmed cases to regular health services and specialized centers for treatment. Screening was based on the palpation of cervical glands until 1996, when a serologic screening test (card agglutination test for trypanosomiasis [CATT]) was added to the algorithm (16). Each mobile team screens ≈40,000 persons/year. A considerable amount of passive case finding takes place as well, for example, when the regular health service staff diagnose HAT in a patient who arrives for a consultation. PNLTHA's control strategies also include vector control.

We used the PNLTHA epidemiologic surveillance database that included all HAT cases detected by mobile teams and regular health services since 1926. For 1993–2003, we examined the monthly reports compiled by the regional PNLTHA coordinators, with the exception of those from Maniema-Katanga and the Province Orientale because they were incomplete and fragmentary as a consequence of the ongoing war.

We distinguish 2 discrete periods for the analysis of international aid. From 1993 to 1997, only humanitarian aid budgets were allocated to HAT control, typically lasting for a maximum of 6 months. Because of the political turmoil at that time in DRC, international aid for HAT was given as "indirect aid," i.e., donors would give cash grants to 3 nongovernmental organizations (NGOs)—Fonds Médical Tropical (FOMETRO), Medische Missie Samenwerking (MEMISA), and Médecins sans Frontières (MSF)—and rely on them for implementation. MEMISA and MSF would supplement this indirect aid with funds they had privately raised. Between 1998 and 2003, the Congolese government again benefitted from long-term international aid programs, and the Belgian Technical Cooperation (BTC) launched its own technical assistance program for HAT control. The same NGOs continued to play a major role in implementation, as well as partly funding, these control activities. For the period under study, WHO funds were donated directly to PNLTHA, while those from the European Union were given to FOMETRO.

We obtained financial data on budgets and expenditure for HAT control directly from the various donor agencies and cross-checked data with all the implementing agencies (PNLTHA, the 3 NGOs, and the BTC engaged in HAT control in DRC during the period under study) (Table 1). Only funds allocated to HAT control were incorporated in our study; we excluded funds earmarked for research. To avoid duplication, we categorized financial resources by donating and not by implementing agency. Over the entire study period, the Congolese government only allocated funds for personnel costs, and those were included in our computations at an average salary of US $12.50 per month per person. All international aid was donated in cash directly to the NGOs or BTC.

Expenditure in Belgian francs was converted into US dollars, according to the exchange rate that applied at time of expenditure. Expenditure in euros was converted at a fixed rate of 40.3399 Belgian francs = 1 euro. The exchange rate between the US dollar and euro was the average exchange rate per year based on Federal Reserve Statistical Release (available at http://www.federalreserve.gov/releases/H10/Hist/). All current dollars were converted to constant 1998 dollars by using the US Office of Labor Consumer price Index. All data were stored and analyzed in an Excel database (Microsoft Corp., Redmond, WA, USA).

PNLTHA defines a new HAT patient as a person whose condition has, for the first time, been diagnosed by parasitologic examination as sleeping sickness. Relapse cases are thus not included in this study. The HAT detection rate is the number of newly detected cases, expressed as a proportion of the screened population. We distinguish the active detection rate (ADR), in which data are collected through active case finding, from the overall detection rate, which also includes cases detected at health facilities. The coverage rate of the population is the proportion of the population tested (through active or passive case finding) divided by the population at risk for HAT. The participation rate applies only to active case finding and is defined as the number of persons screened by the mobile teams divided by the target population. The proportion of treated patients is the number of persons who received HAT treatment divided by the number of persons detected with HAT.

We structured our evaluation of the HAT control program in the form of input, process, output, and outcome analyses (17) and according to the method of Bouchet et al. (18) (Table 2). Input represents the human and financial resources invested in the program. Drug availability, measured as the number of occasions that the stock ran out during the period under study, was also considered as an input. Process indicators are not reported in this study because they are relevant only to the daily management of the program. Program output was measured through an analysis of coverage of the population at risk, the participation rate in screening, the number of detected HAT cases, the proportion of patients with detected HAT patients who received treatment, and the proportion of patients with treated cases that have been followed-up correctly. We report the annual HAT detection rate, both nationally and for each region, as indicators of program outcome.

Figure 2 represents the evolution in the annual number of newly detected HAT cases in DRC from 1926 to 2003. Between 1960 and 1989, the figure shows an increasing trend with 2 small peaks in 1970 and 1986. This trend was interrupted in 1990 and 1991, which coincides with the sudden arrest of control activities in 1990. When humanitarian aid was launched in 1993, the annual number of detected cases increased markedly. The peak was reached in 1998, when the control program detected 26,318 new HAT cases in a screened population of 1,472,674 persons.

After 1998, a marked decline occurred in the number of HAT cases detected, which was not due to an overall decrease in screening activities; the number of operational mobile teams and number of screened persons continued to increase over that period (Figure 3). The overall HAT detection rate, based on active case finding, declined from 1.1% in 1994 to 0.3% in 2002. However, this overall decline in detected HAT cases masks differences between regions. Figure 4 shows the evolution of the number of HAT cases and active detection rate, by region, from 1993 until 2002.

In 1990, 25 mobile teams covered the population at risk. With the reduction in external financial support, 10 teams remained operational from 1991 to 1993. This number slowly increased to 33 teams in 1998 and to 46 teams in 2002. These increases were concentrated in the regions of Bandundu, Equateur-Nord, and Kasai, where the number of teams rose to 13, 13, and 7, respectively, which accounts for 33 mobile teams of 46. In 1993, PNLTHA staff was 250. This number increased progressively to 580 in 2001 and remained stable at 580 in 2003.

From 1993 to 1997, total annual expenditure amounted to US$1,302,646, while in the period 1998–2003, total annual expenditure doubled to US$2,786,916. Table 1 shows the amount and the origin of the financial resources. The budget breakdown was as follows: functioning costs (fuel, vehicle maintenance, supervision, training, stationery, etc.) (32.0%), personnel time (26.4%), HAT drugs (24.3%), laboratory reagents (7.3%), and other material and equipment (10%). The average expenditure per HAT case detected and treated is shown in Table 3.

From 1993 to 2001, implementing agencies spent, on average, 24% of their budgets in purchasing trypanocidal drugs (range 12%–44%). Though the Sanofi-Aventis/Bayer donation program was established in 2001, in practice, implementing agencies could continue working throughout 2002 with existing drug stocks. For 2003, a detailed analysis of amount of donated drugs versus purchased drugs consumed was not possible, and we therefore ignored the in-kind drug donation in Table 1. The full effect of the donation will only become clear after 2004, although can be estimated by its monetary value (Table 4). PNLTHA records the number of patients who have been treated by drug regimen, and these data allowed us to estimate the quantity of the trypanocidal drugs that were required, as well as the total cost of those drugs, calculated according to the preferential price, which was valid until 2001. We estimated that the total drug cost per year in DRC, for treating ≈14,000 HAT patients/year, corresponded to ≈US$600,000/year. This value is consistent with the previous estimate that 24% of budgets are reportedly used for drugs. Notably, since the public-private partnership was established in 2001, supplies of trypanocides never ran short, whereas this problem was a matter of continuous concern before.

The population at risk in the DRC has been estimated at 12,600,000 persons (PNLTHA, unpub. data). Screening and treatment of the at-risk population is estimated to have risen from 6% in 1993 to 19% in 2003. We observed notable differences between regions. Equateur-Nord had a coverage rate of >50%, while in the other regions the rate ranged from 10% to 20%. Figure 3 shows that the number of persons screened each year almost tripled from 1993 to 2003.

The participation rate of the population in active case finding was almost 96% in 1998. By 2002, the rate had fallen to 78%. The proportion of new patients who received treatment was ≈95% throughout the whole country but varied from region to region, from 89% to 100%. From 2001 to 2003, a total of 44,247 patients were treated with pentamidine (31.5%), suramin (5.8%), the combination pentamidine-suramin (0.2%), melarsoprol (55.2%), eflornithine (0.85%), nifurtimox (2.6%), and the combination melarsoprol-nifurtimox (3.5%) (because nifurtimox is not registered for use against HAT in DRC, it was given on a compassionate basis when no other drugs were available or when melarsoprol treatment failed).