Costs and Cost-Effectiveness of Interventions
The published information on cost-effectiveness of interventions for the four diseases is incomplete, and this section provides some new data on the cost per DALY averted that is not available in the literature.
Chagas Disease
For the Southern Cone countries, investment in the control of Chagas disease since 1991 has been about US$320 million, well within the original estimates of US$190 million to US$350 million (Schofield and Dias 1991). Initial predictions of cost-effectiveness suggested an internal rate of return (IRR) for the initiative of about 14 percent, but point studies during the course of the interventions suggest actual IRRs of about 30 percent for Brazil (Akhavan 1998) and more than 60 percent for the province of Salta, Argentina (Basombrio and others 1998).
Brazil had an estimated 2 million infected individuals in 1995. Annual follow-up of the 1.6 million asymptomatic cases would have cost US$98 million. Diagnosis of megasyndromes for 6 percent of infected individuals at an average cost of US$141 each would add US$16.9 million, and corrective surgery for 3 percent of the latter would add US$60 million. Cardiac pacemakers for 5 percent of infected individuals at US$3,000 each would add another US$30 million, so that the partial direct costs for medical attention in Brazil in 1995-96 would have been US$205 million.
In Argentina, the costs for medical attention in 1992 were US$435 for acute cases, US$122 per patient for asymptomatic cases, US$336 for moderate cardiopathy, and US$1,135 for severe cardiopathy. Given that Argentina had 2 million infected individuals, and assuming that 85 percent of them would have been asymptomatic,9 percent would have had mild cardiopathy, 4 percent would have had moderate cardiopathy, and 2 percent would have had severe cardiopathy, then total expenditures for medical attention would have amounted to US$457 million.
In Chile, aggregate treatment costs were estimated at US$37 million in 1991. New estimates in 1997 using the government payment schedule and an estimate of 142,000 people infected, including 26,545 with myocardiopathy of which 9,652 were severe cases, resulted in an estimated cost of US$14 million to US$19 million. In Uruguay, annual costs for treatment were estimated at US$15 million for 1996.
These treatment costs are significantly higher than the costs of vector control, which for 1996 were US$13 million in Argentina, US$28 million in Brazil, US$650,000 in Chile, and US$4,000 in Uruguay. Akhavan's (1998) study in Brazil estimates a cost of US$260 per DALY prevented, and Robles's (1997) study in Bolivia indicated a cost of US$362 to prevent one year of life lost.
Data on the cost-effectiveness of treating those infected are sparse; however, Robles's (1997) study in Bolivia estimates the costs of treating infected children under the age of five, coming up with a cost of US$3,009 per death averted or about US$100 per DALY averted.
Lymphatic Filariasis
The most widely used interventions for LF control are MDA, vector control, and administration of diethylcarbamazine-fortified salt. We estimate the cost-effectiveness of those strategies in terms of DALYs averted from studies in India on the costs and effectiveness of control and for different scenarios for the minimum duration of control required to achieve sustained interruption of transmission. These scenarios assume that all three strategies are implemented in areas with similar levels of endemicity of Culex-transmitted Bancroftian filariasis and that the available cost data for India apply.
We consider three scenarios (table 22.2). Elim1 is an optimistic elimination scenario under which sustained interruption of transmission is achieved after a relatively short period of intervention (six annual rounds of MDA, 10 years of vector control, and 2 years of diethylcarbamazine-fortified salt). Elim2 is a conservative elimination scenario under which sustained interruption is achieved only after a longer period of intervention (10 years of MDA, 15 years of vector control, and 4 years of diethylcarbamazine-fortified salt). Control is a scenario under which transmission is brought to low levels but not interrupted and where control efforts will have to continue.
[Table .]
Because of the slow dynamics of filariasis transmission and disease, the prevalence of the chronic disease manifestations (lymphedema and hydrocele) on which the DALY estimates are based will not fully reflect the effect of control for many years. We have therefore tried to predict the trend in chronic disease over a 30-year period. Recent findings from a longitudinal study (Ramaiah and others 2003) of the effect of MDA in Pondicherry, India, showed that the prevalence of hydrocele and lymphedema had declined by 58 percent after seven annual treatment rounds with diethylcarbamazine. We assumed that from the seventh year of intervention, any further reduction in disease prevalence was attributable exclusively to reduced incidence as a result of reduced transmission, and that 30 years after the initiation of the intervention, the prevalence of disease would have fallen by 90 percent. We assumed that the effect of diethylcarbamazine-fortified salt was similar to that of MDA, whereas for vector control we assumed that prevalence would decline with a delay of seven years.
The predicted costs per DALY averted (table 22.3) indicate that MDA and diethylcarbamazine-fortified salt are extremely cost-effective. Elimination with MDA costs about US$4 to US$8 per DALY averted, and even if transmission were not interrupted and MDA would have to be continued for 30 years (control scenario), the cost would be still only be around US$29 per DALY averted. Diethylcarbamazine-fortified salt would be the cheapest intervention, but governments rarely favor it, and compliance can be difficult to ensure. Vector control is at least 10 times more expensive in terms of DALYs averted, but it offers additional benefits in terms of malaria and dengue control and significant relief from mosquito nuisance.
[Table .]
The effect of MDA on hydrocele and lymphedema is not yet well established and the results of the Indian trial on which the previous calculations are based may be too optimistic. However, even under much less favorable assumptions that the prevalence of hydrocele and lymphedema declines by 20 percent after 7 years of MDA and by 75 percent after 30 years, the estimated cost per DALY averted would be only 50 percent higher than those given in table 22.3, and the interventions would still be very cost-effective.
The prevention of chronic disease also has direct economic benefits (Ramaiah and Das 2004). The cost of preventing one case of chronic disease through six rounds of MDA in India has been estimated at US$8.41. The economic benefits include savings of 58.24 working days per year per case, yielding wages of US$39.39 and treatment costs of US$1.44. On average, chronic patients lose 11 years of productive life; thus, the average economic benefits total US$449.13 per chronic case averted. This figure gives a benefit-cost ratio of 52.6, perhaps one of the highest for any disease control program.
Onchocerciasis
Investment in onchocerciasis control has included about US$570 million provided by donors to the OCP during 1975-2002, US$140 million provided and earmarked for APOC for 1996-2010, and US$10 million for OEPA for 1991-2003. The African onchocerciasis control programs are considered highly cost-effective. No cost-benefit analysis has yet been published for OEPA.
The OCP has been highly successful. More than 40 million people in the program's 11 countries are now considered free from infection and eye lesions, more than 1.5 million people are no longer infected, and more than 200,000 cases of blindness have been prevented. Sixteen million children born since the program began are free of onchocerciasis. The socioeconomic effect has also been dramatic: 25 million hectares of fertile land in the river valleys were made available for resettlement and agriculture. A cost-benefit analysis of the OCP has estimated the net present value for the OCP over a 39-year project horizon from 1974 to 2002 as US$485 million (Kim and Benton 1995). This figure corresponds to an IRR of 20 percent, resulting mainly from increased labor because of prevention of blindness (25 percent of benefits) and increased land use (75 percent of benefits).
A similar cost-benefit analysis for APOC also considered benefits in terms of additional labor resulting from blindness prevention (Benton 1998). It did not consider land use because depopulation of river valleys is rarely seen in APOC countries, where the forest type of onchocerciasis predominates. Nevertheless, the estimated IRR for APOC remained almost as high as that for the OCP (17 percent), because the cost is lower but the number of people served is far greater.
The estimated rates of return for the OCP and APOC did not include the effects of control on onchocercal skin disease. Hence, these rates underestimate the benefits, because troublesome itching accounts for more than 50 percent of the DALYs attributable to onchocerciasis. The cost of ivermectin, which is donated by Merck, was not included in our analyses.
To estimate the approximate cost per DALY averted, we considered the burden of disease and treatment with ivermectin in APOC countries. Using the latest epidemiological mapping data, we estimate that, in 1995, 34.6 million people were infected in APOC countries and that 1.86 million DALYs were lost. Currently more than 44 percent of those infected are covered by community-directed treatment with ivermectin, and expectations are that treatment will be expanded to cover most of the remainder before the end of APOC in 2010. Information from areas where ivermectin treatment has been in effect for more than 15 years shows that the prevalence and intensity of onchocerciasis infection have fallen to low levels (Borsboom and others 2003), and computer simulations predict that the disease could not become a public health problem again for at least another 10 to 20 years if treatment were halted (Remme, Alley, and Plaisier 1995). We therefore estimate that 15 years of ivermectin treatment at 65 percent coverage will prevent at least 25 years of onchocercal disease. If we assume that 70 percent of endemic communities will ultimately be covered by community-directed treatment with ivermectin and that 80 percent of those communities will maintain annual treatment at 65 percent coverage for at least 15 years, at least 26 million DALYs would be prevented over a 25-year period.
The predicted cost of community-directed treatment with ivermectin in APOC countries is US$145 million by the international donor community plus US$64 million by ministries of health and collaborating nongovernmental organizations, giving a total of US$209 million. Therefore we estimate that the cost of community-directed treatment is approximately US$7 per DALY averted.
The ultimate cost-benefit of onchocerciasis control will depend on how long effective control programs will need to be maintained to keep the disease under control. National governments and ministries of health should plan to invest in ivermectin distribution and in surveillance activities for the foreseeable future. Thus, a case must continually be made with national decision makers that if investments do not continue, recrudescence of infection is likely. One strategy for sustaining national investment is to show that ivermectin distribution systems can be made polyfunctional. Treatment programs based on MDA for intestinal parasites, schistosomiasis, and LF and on vitamin A distribution can be integrated with ivermectin distribution programs and thereby further improve cost-benefit ratios. The use of community-directed treatment with ivermectin is also envisaged as a way of strengthening peripheral and district health systems (Homeida and others 2002).
Leprosy
Costs associated with leprosy control include case detection, treatment, prevention of disability, and rehabilitation. We calculate the incremental health service cost to arrive at the average cost of curing a patient with leprosy. Our estimates are based on the limited published cost data available, program expenditure data, and expert opinion, although costs are likely to differ substantially by country.
As case-detection rates decrease, the average cost of detecting one case increases. The previous edition of this volume estimated a cost of US$2 per case detected based on a case-detection rate of about 300 per 100,000; however, case-detection rates are now considerably lower in most countries (Dharmshaktu and others 1999; Ganapati and others 2001; Smith 1999). Many leprosy control programs now rely on voluntary case finding supported by information, education, and communication activities to raise or maintain people's awareness of the early signs and symptoms of leprosy. We estimate the cost of this approach to be about US$1 per case detected. Nevertheless, if active methods are still used in areas where case-detection rates are low, the cost of case detection may be as high as US$108.
The costs of diagnosing and treating leprosy have fallen in the past decade, and diagnosis by clinical examination only is now recommended. We therefore exclude the cost of skin smears. In addition, a shortening of the treatment regimen has lowered drug costs to about US$12 for a multibacillary case and US$1 for a paucibacillary case. Globally, almost 40 percent of leprosy cases are classified as multibacillary cases, with the remaining 60 percent being paucibacillary cases. Thus, we estimate the average drug cost as US$5.40 per case.
The cost of treatment, however, is more than the cost of drugs alone. WHO guidelines recommend that a multibacillary case receive supervised treatment for 12 months and that a paucibacillary patient receive treatment for 6 months. Using cost data from Ethiopia and Pakistan, we estimate these treatment costs at US$20 to US$30 in low-income countries. Data from studies of tuberculosis interventions show that community-supervised treatment may reduce costs by up to 50 percent (Khan and others 2002), and this approach is being advocated as part of "flexible MDT delivery" (ILA 2002) and "accompanied MDT" (WHO 2002b). Reducing the nondrug costs of treating leprosy to about US$10 to US$20 per patient may, therefore, be possible. We thus estimate the costs of treating a case of leprosy with MDT to be between US$15.40 and US$35.40 per case, depending on the strategy used.
About 10 to 20 percent of new leprosy cases are likely to suffer a reaction during or after MDT. We estimate treatment of those reactions to cost US$25 per patient. Of these patients, 1 percent may develop severe complications requiring hospitalization, at an estimated cost of US$480 per patient. In addition, 10 percent of new cases will develop neural or secondary impairments and may require footwear and education about wound management. We estimate the lifetime cost of protective footwear at US$300 per patient (Seboka, Saunderson, and Currie 1998) and education at US$10 per patient. In 1 percent of cases, reconstructive surgery may be required at about US$455 per patient. We therefore estimate the average incremental cost of interventions for prevention of disability to be US$44.15 per new case of disability. Because about 3 percent of new patients will need rehabilitation, we estimate the average cost at under US$1 for each new case of leprosy detected (Jagannathan and others 1993). However, a backlog of old cases exists. Although data in this area are weak, up to a third of the 4 million people living with leprosy globally (2 million with grade 1 disability and 2 million with grade 2 disability) could require rehabilitation.
Few data are available on the program costs associated with leprosy. A review of expenditure in Asia found that up to 40 percent of the total costs could be classified as programmatic costs, although this amount may now be less because leprosy programs have increasingly been integrated into general health services. Data from Indonesia demonstrate that program costs can be reduced by up to 35 percent by integrating them with tuberculosis programs (Plag 1995). We therefore estimate the average cost of finding, treating, and preventing disabilities and rehabilitating a new case of leprosy at US$76 to US$264.
In practice, many leprosy programs will also be providing disability prevention and rehabilitation interventions to a large backlog of patients, so the average cost per new case will be higher than here. Programs that face a high proportion of multibacillary cases and cases presenting with high levels of disability are also likely to have higher costs.
Assuming a cure rate of around 85 percent, we estimate the costs of curing one patient of leprosy to be about US$93 per new case. Using data from India (25 percent of those with leprosy will self-cure, an average age of onset of 27, a disability weighting of 0.152, and a life expectancy at age 25 to 29 of 44.75), we estimate the cost per DALY of detecting and treating a new case of leprosy to be US$38.
In addition, assuming a 90 percent success rate, we calculate a cost per DALY of US$7 for patients needing treatment for reactions and ulcers, US$75 for those needing footwear and self-care education, and US$110 for those needing reconstructive surgery. These estimates provide only a broad indication because data on the effectiveness of these interventions are scarce, and the application of the disability weight of 0.152 to all interventions may overestimate their benefits.
Data on the economic effect of leprosy at the national level are not available. However, leprosy affects those who are economically active, with a peak in incidence at 10 to 20 years of age and again at 30 to 50 years of age. Studies of the impact of leprosy on productivity show that deformity from leprosy can reduce the probability of obtaining employment and can reduce household income and expenditure on food (Diffey and others 2000; Kopparty 1995). In addition, leprosy can have a significant social impact because participation in the community may be restricted. This impact continues well beyond the actual treatment period because leprosy-related impairments have a tendency to get worse over time even after the infection has been arrested.
Summary
Available information indicates that interventions for the four diseases are highly cost-effective and that the benefit-cost ratio of control is high (table 22.3).
