Costs and Cost-Effectiveness of Interventions
Dengue
Few studies are available on the cost-effectiveness of vector control for reducing dengue transmission. One of the difficulties is that the level of vector population control needed to reduce transmission is influenced by the human population's past exposure to the circulating virus serotype. A direct relationship is apparent between seroprevalence rates and levels of vector abundance needed for epidemic transmission. Thus, the paradox is that, as herd immunity declines over time in response to effective vector control, progressively lower vector densities can maintain the same level of transmission.
Modeling of the dynamics of dengue transmission is helping to improve understanding of the interrelationships between virus, vector, and host (Ferguson, Donnelly, and Anderson 1999; Focks and others 1995; Newton and Reiter 1992; Shepard 2001), but the absence of epidemiologically defined target levels for vector control has hindered calculations of cost-effectiveness. According to Shepard and others (2004), average annual costs for dengue vector control per 1,000 population were US$15 in 1998 in Indonesia, US$81 in 1994 and US$188 in 1998 in Thailand, US$240 in 2002 in Malaysia, and US$2,400 in 2000 in Singapore. In 1997, spending on dengue control in 14 Latin American countries ranged from US$20 to US$3,560 per 1,000 population, with a median of US$260. For 17 Caribbean islands in 1990, the corresponding expenditures ranged from US$140 to US$8,490, with a median of US$1,340 (Nathan 1993). By contrast, McConnell and Gubler's (2003) study in Puerto Rico concludes that larval control programs that achieve a 50 percent reduction in dengue transmission and cost less than US$2.50 per person would be cost-effective in that setting. From research based on analytical models (Shepard 2001) and primary data from Singapore, we estimate the cost of using environmental management for control at US$3,139 per DALY averted and the cost of using insecticides at US$1,992 per DALY averted.
Dengue case management depends on the severity of the illness. Despite the lack of information about cost-effective interventions to treat dengue cases, Shepard (2001) estimates an average cost of US$587 per DALY averted by appropriate case management. Were a dengue vaccine to become available, the Shepard model estimates that immunization would cost US$3,040 per DALY averted.
Leishmaniasis
Case Finding and Treatment
For leishmaniasis, diagnosis represents a small proportion of the cost of case finding and treatment, with diagnostic tests becoming available at approximately US$1.50 for the dipstick, US$3.00 for the direct agglutination test using freeze-dried antigen, and US$1.50 for the urine latex agglutination text. These tests can be used in the field. A study in Nepal (Pokhrel 1999) comparing outreach case detection using serology (the dipstick) with parasitological diagnosis at health centers (bone marrow aspirate) concluded that the median cost per VL case detected was US$25 in the outreach program, compared with US$145 at health centers (of which more than 50 percent was due to absence from work). Treatment costs increased these figures to US$131 and US$200 per patient, respectively.
In India, an examination of the costs of drugs and hospitalization and of the evolution of the disease under treatment (cure, relapse, failure, intolerance) indicated that the final cost of successful treatment depends largely on the basic drug cost, which averaged US$86 per patient successfully treated with miltefosine (using reduced pricing because of the large number of patients), US$467 for treatment with amphotericin B, and US$1,613 for treatment with AmBisome. Given current estimates of about 100,000 cases of VL each year in the state of Bihar, India, the estimated total cost of treatment using miltefosine as a first-line drug and amphotericin B as a second-line drug would be about US$11 million, or approximately US$110 per patient (personal communication with P. Olliaro and S. Sundar on treatment options for kalaazar [visceral leishmaniasis], 2003). By contrast, analysis of humanitarian relief interventions by Medecins sans Frontieres-Holland that combined case finding with treatment after a VL epidemic in southern Sudan indicated total costs of US$394 per patient, or an average cost of US$595 per life saved (Griekspoor, Sondorp, and Vos 1999). Thus, the average cost per DALY averted was US$18.40.
Vector Control
Vector control is rarely carried out as a specific approach to leishmaniasis control, and cost-effectiveness estimates are not available. In general, domestic and peridomestic sandfly vectors are more susceptible to indoor residual spraying than are other domestic vectors, such as anopheline mosquitoes or triatomine bugs, so that transient suppression of sandfly populations is seen as an additional benefit of malaria or Chagas disease vector control in areas where these vectors coincide. However, insecticide-treated bednets, which are becoming widely deployed against malaria transmission, may also become cost-effective for reducing leishmaniasis in areas of domestic transmission. In Yenice, Turkey, the use of impregnated bednets reduced the incidence of CL from 1.90 percent to 0.04 percent between 2000 and 2001 (Alten and others 2003).
African Trypanosomiasis
Case Finding and Treatment
WHO (1986, 1998) has analyzed the costs of T.b. gambiense control by means of case finding and treatment based on practice in Cote d'Ivoire and Uganda. This work and other studies indicate that, at current prices, the cost of active detection using the card agglutination trypanosomiasis test with parasitological confirmation varies around US$1 per person screened or slightly more for mobile teams. However, mobile teams are more effective in screening a high proportion of the population and are also more successful in ensuring that a high proportion of patients receive treatment. Unit costs are currently US$0.33 per person for the card agglutination trypanosomiasis test and US$2.20 for the miniature anion exchange centrifugation technique. Less sensitive parasitological techniques, such as examination of lymph node aspirate or blood smears, cost only a few cents but may miss a third to half of patients.
By contrast, treatment is expensive despite the availability of free drugs. Treatment of early-stage disease incurs costs of more than US$100 per person, rising to more than US$250 for late-stage treatment with melarsoprol and about US$700 with eflornithine (WHO 1998). The long hospitalization period is a major component of costs during the second stage, although work undertaken by Burri and others (2000) on a shorter melarsoprol regimen offers opportunities for reducing these costs.
Despite the costs and the risk of complications, treating sleeping sickness patients in the second stage of the disease is cost-effective. In Uganda, costs were less than US$10 per DALY averted for melarsoprol treatment and less than US$20 per person for eflornithine (Politi and others 1995). Similarly, in southern Sudan, the cost per DALY averted ranged from US$4 to US$22 (Trowbridge and others 2001). Shaw and Cattand (2001) considered the costs of case finding and treatment for T.b. gambiense infection for three delivery options and a wide range of prevalences. Given the limited information available on DALYs gained or on the effect on transmission of reducing the size of the human reservoir, they estimate that under different scenarios the costs per DALY averted tend to converge. For most assumptions, the cost per DALY averted fell below a US$25 threshold at prevalences of 0.5 to 1.0 percent but rose sharply at low prevalences, which explains the reluctance of control programs to invest in screening operations when prevalence is less than 0.2 percent. With better quantitative understanding of the effects of screening and removing patients from the reservoir in preventing future epidemics, investigators could demonstrate that even at low prevalences screening for sleeping sickness is highly cost-effective.
Vector Control
Several countries have undertaken community-based programs to trap tsetse flies, notably Cote d'Ivoire, where costs came to US$2.30 per person protected per year (Laveissiere and others 1994), and Uganda, which achieved a cost of US$0.50 per person protected per year (Lancien and Obayi 1993). Vector control costs have been studied in more detail in the context of livestock disease (Maudlin, Holmes, and Miles 2004). These costs vary according to the technique used and the environmental context, often ignoring overheads for organizing and planning. With that caveat in mind, the figures per square kilometer cited for local tsetse-fly eradication range from about US$250 to US$550 at current prices for aerial spraying (based on experience in Somalia, South Africa, Zambia, and Zimbabwe); US$250 to US$400 per square kilometer for ground spraying; and US$200 to US$400 per square kilometer for targets. However, the cost of traps and targets falls to US$25 to US$60 per square kilometer for control or suppression operations alone. Projects treating cattle with insecticides have been implemented at costs of US$50 to US$60 per square kilometer. Use of the sterile insect technique is much more expensive because it relies on prior suppression of fly populations using another technique. The overall costs of the experimental eradication of G. austeni from Zanzibar using the sterile insect and other techniques were about US$3,000 per square kilometer, although the International Atomic Energy Agency (IAEA) envisages that the cost of the sterile insect technique component could be reduced to less than US$800 per square kilometer as the technology is developed and applied on a sufficiently large scale, (Dr. Udo Feldmann, IAEA, Vienna, personal communication).
