Economics of Malaria Control Interventions
Goodman, Coleman, and Mills's (2000) study represents the most thorough attempt to compare the cost-effectiveness of a wide range of malaria control interventions. They find that in a very low-income country, the cost-effectiveness range per DALY averted was US$19 to US$85 for ITNs (nets plus insecticide), US$32 to US$58 for residual spraying (two rounds per year), US$3 to US$12 for chemoprophylaxis for children (assuming an existing delivery system), US$4 to US$29 for IPT for pregnant women, and US$1 to US$8 for case-management improvements. Goodman, Coleman, and Mills (2000) find that even though some interventions are relatively cheap, achieving high coverage may require a level of expenditure currently out of reach for many African countries and that overcoming operational barriers to achieving widespread coverage is likely to require substantial assistance from external donors.
Analysis
The following analysis incorporates new knowledge on the effects of interventions and on their costs for a low-income, Sub-Saharan African population living in an area of high, stable transmission. The modeling draws on a wide range of sources on the costs and effects of each intervention, extrapolated across settings and operational conditions. The approach allows for changing cost-effectiveness over time, for example, as resistance to antimalarial drugs or insecticides increases. To address problems of uncertainty in relation to many of the parameters, we used probabilistic sensitivity analysis, which allows for multivariate uncertainty by assigning ranges rather than point estimates to input variables. We assumed that cost and effectiveness input variables follow uniform triangular or normal continuous probability distributions (Mulligan, Morel, and Mills 2005).
We consider the cost-effectiveness of a limited subset of interventions: ITNs, IRS, IPT during pregnancy, and patient management with a change of the first-line drug. We include costs to the provider and the community and incremental out-of-pocket expenses for households, but because of major valuation and measurement problems, we do not include the indirect costs of patients' time to seek care and of the lost productivity resulting from morbidity. We consider only gross costs for all interventions except patient management, given the uncertainty inherent in estimating savings.
Insecticide-treated Nets
We based our analysis of ITNs on the delivery mechanism used in the WHO Special Programme for Research and Training in Tropical Diseases (WHO/TDR) trials, where householders, community health workers, and program staff worked together to treat the nets. In relation to insecticide, we considered permethrin and deltamethrin. Deltamethrin is effective for a year; thus, re-treatment is annual. Permethrin lasts for six months; thus, we assumed two treatments per year if the transmission season is longer than six months. The activities undertaken were the training of staff and community health workers, a campaign to inform the community about the intervention, the procurement and transport of the insecticide and nets, and the initial treatment and the re-treatment of the nets. We calculated cost-effectiveness for each intervention for two scenarios: one whereby nets were distributed to households and the second whereby treatment was arranged for existing nets. We drew estimates of the effectiveness of ITNS from a recent meta-analysis of WHO/TDR-sponsored trials conducted in Sub-Saharan Africa (Lengeler 2004). We adjusted the key parameter and effectiveness estimates to account for the proportion of children sleeping and not sleeping under a recently treated net.
With one treatment per year using deltamethrin, the mean cost per DALY averted was US$11 (90 percent range of US$5 to US$21). With one treatment of permethrin per year the cost-effectiveness ratio (CER) increased slightly to a mean of US$12 (90 percent range of US$6 to US$20). Two treatments of permethrin per year increased the mean CER to US$17 (90 percent range of US$9 to US$31). Even if net coverage is low and nets have to be distributed and treated, the intervention remains an extremely attractive use of resources. Moreover, the model includes health benefits only for children under five. If we included benefits for other household members, the CERs would be lower.
Insecticide Residual Spraying
We considered four insecticides for our analysis of the cost-effectiveness of IRS: DDT; malathion; and two pyrethroids, deltamethrin and lambda-cyhalothrin. We used the results of the Cochrane Review meta-analysis of ITN trials conducted in Africa to approximate the results of spraying on morbidity and mortality and adjusted effectiveness estimates to account for noncompliance (Lengeler 2004).
We found little difference between the CERs for the four insecticides when one round of spraying was done per year: they ranged from US$5 to US$18. With two rounds per year, costs increased, but we assumed that effectiveness remained the same, so all the CERs approximately doubled to US$11 to US$34. These results should be interpreted with caution because of uncertainty in relation to the estimates of effectiveness in children under five. The model was based on one round of spraying per year in areas of seasonal transmission and two rounds per year in areas of high, intense (perennial) transmission. Effectiveness will depend on the length of the transmission seasons and on the insecticide. DDT lasts for six months or more, lambda-cyhalothrin for three to six months, and malathion and deltamethrin for only two to three months (Lengeler 2004; Lengeler and Sharp 2003).
IPT during Pregnancy
We analyzed the cost-effectiveness of IPT assuming that primigravid women are given two or three doses of SP at a prenatal clinic. We analyzed benefits to the child by decreased mortality and benefits to the mother resulting from changes in the incidence of severe anemia. We estimated the effect of IPT on the neonatal mortality rate as a function of increased birthweight based on the Cochrane meta-analysis of malaria prevention in pregnancy (Gulmezoglu and Garner 1998). The model allowed for level of drug resistance, probability of initial attendance at a prenatal clinic, probability of returning for a second visit, probability of returning for a third visit, and compliance with the drug regimen. We estimated both incremental and average costs.
The incremental CER for IPT in pregnancy using SP had a 90 percent range from US$9 to US$21 with a mean of US$13. Average total cost-effectiveness had a mean of US$24 (90 percent range of US$16 to US$35).
Change in First-Line Drug
We analyzed the cost-effectiveness of changing first-line therapies to SP and to ACT using a patient-management model with a decision-tree framework in which a patient presents with uncomplicated malaria at an out-patient facility and progresses to full recovery, recovery with neurological sequelae, or death. Three potential drug policy changes were considered (see table 21.5). For current drug policies we assumed either chloroquine or SP as the first-line drug, with SP or amodiaquine as a second-line drug and quinine as the third-line drug. We then examined the cost-effectiveness of policy switches to either SP or ACT as first-line drug (with amodiaquine as second line and quinine as third line).
[Table .]
Amodiaquine was chosen as the second-line drug in the new drug policy because, like chloroquine and SP, it is relatively cheap. Low compliance, adverse effects, and potential cross-resistance between amodiaquine and chloroquine excluded it from first-line selection. We omitted other potential drugs to limit the scope of analysis.
We used commonly found levels of drug resistance to create likely ranges. When used as second-line treatment, we assumed that a drug faced half the level of resistance than when it was used as a first-line drug. We estimated the growth rate of resistance to each drug based on its expected current location along a sigmoid growth curve. We assessed compliance with each drug based on the length and complexity of each regimen. We gave the widest range (20 to 70 percent) to compliance with ACT, which usually requires a three-day treatment, to account for the different lengths of regimen and variety of formulations.
Figure 21.3 shows that a switch from chloroquine to SP is cost-effective (less than $150 per DALY averted) when chloroquine resistance is above 35 percent. Switching from chloroquine to ACT becomes cost-effective as chloroquine resistance reaches around 37 percent. Switching from SP to ACT becomes cost-effective as SP resistance reaches 12 percent. This low threshold is due to the high growth rate of resistance to SP when it is used as a first-line therapy.
[Figure
21.3]
Results and Interpretation
Table 21.6 presents the average CERs for the interventions reviewed. All interventions can be considered attractive using a cutoff of US$150 per DALY averted. For the childhood preventive interventions, the level of existing infrastructure is a key factor in influencing cost-effectiveness. ITNs are an even better use of resources if net coverage is already high, and IPT is even more cost-effective if prenatal care coverage is good. However, even if levels of infrastructure are poor, these interventions are still attractive based on cost-effectiveness criteria.
[Table .]
Curtis and Mnzava's (2000) review comparing worldwide trials of ITNs and IRS for malaria control suggests that they were of equivalent effectiveness. Lengeler and Sharp (2003 21,) also conclude that choosing between IRS and ITNs is "largely a matter of operational feasibility and availability of local resources, rather than one of malaria epidemiology or cost-effectiveness." DDT is the cheapest insecticide but is seldom used because of concerns about its environmental impact. An updated systematic review of the health effects of IRS as well as a full economic comparison between spraying with DDT and other insecticides and ITNs is urgently required. Such reviews should include the environmental benefits of reducing DDT levels and the costs of alternative interventions.
As effective patient management, IPT, and other approaches toward drug use become more widespread, drug resistance will increase, which will affect cost-effectiveness. In terms of first-line treatment, while a switch from chloroquine to SP is unlikely to be costly, cost-effectiveness depends on the initial level of severe resistance to each drug. A switch from chloroquine to ACT is likely to be costly but more effective, given that resistance to ACT is essentially nonexistent and the growth rate of resistance to ACT is likely to be low (Yeung and others 2004). Given the high growth rate of resistance to SP, switching to ACT becomes cost-effective when SP resistance surpasses 12 percent, a relatively low threshold. A switch from chloroquine to ACT appears to be highly cost-effective at all initial levels of chloroquine resistance above 37 percent. Recent studies indicate the remarkable effectiveness of the ACT artemether-lumefantrine in East African areas of chloroquine, amodiaquine, and SP resistance (Mutabingwa and others 2005; Piola and others 2005); it is expected that the availability of ACTs will increase and the costs will decrease in the near future.
Affordability and Scaling Up
Cost-effectiveness analyses can identify which interventions are the most efficient to implement, but information is also needed on affordability. Some interventions are cheap, such as prevention in pregnancy. Achievement of high coverage with an intervention to prevent childhood malaria (for example, ITNs) has a high total cost, which needs to be borne in part by external funding. For example, full coverage of children (assuming 22 percent of the population is under five years of age) would cost US$2.81 million per 1 million persons of the general population covered (assuming the provision of nets plus two rounds of permethrin). The same coverage with IRS would cost about US$4.01 million using deltamethrin (two rounds).
In addition to the difficulties of financial feasibility, the implementation of ITN interventions poses operational and logistical challenges. A number of strategic approaches are available for national ITN upscaling (RBM 2002). The main ones are social marketing (for example, in Malawi and Kenya), assisted commercial sector development (Senegal, Mali, Tanzania), and totally free distribution (Togo). An important feature of all the approaches is that their cost per net distributed decreases significantly as the scale of the undertaking increases, so it is hard to use available estimates to project cost in national programs.
Independent of the main ITN distribution strategy, a number of countries are implementing additional actions targeted at the main high-risk groups. For example, a national voucher scheme is currently being implemented in Tanzania to provide every pregnant woman in the country with a free ITN, and a net is being distributed to children at the time of measles vaccination in Togo.
Re-treatment of ITNs on a large-scale remains a formidable operational issue, and free distribution of insecticide in the way it is done in Vietnam and China is probably the best way forward. This should occur while waiting the availability of long-lasting insecticidal nets that do not require retreatment.
Curtis and Maxwell (2002) point to the successful experience of Vietnam, where ITNs now protect 10 million people and the public health service provides free insecticide. This approach in Africa would require a substantial commitment by donors and governments given that most African governments spend only around US$4 per capita per year on health (World Bank 2003). Eritrea, with 65 percent ITN coverage, and other African countries are beginning to make great strides in getting ITNs to their populations (Barat 2006; WHO 2005; World Bank 2005).
The move toward ACT poses the most difficult questions in relation to the long-term affordability of malaria control interventions. After scale-up, estimates indicate that the additional annual costs of ACT versus current failing drugs range from US$300 million to US$500 million globally (Arrow, Panosian, and Gellband 2004), with the precise amount depending on the extent to which all fevers are treated. This figure does not include drugs for nonmalarial fevers in endemic areas and the costs for the substantial health system strengthening required to make the most effective use of ACT. These include the costs of improved drug regulation, pharmacovigilance, diagnostics, and implementation of different drug policies for different population groups. The introduction of ACT has permanently changed the economic landscape of malaria control. Innovative funding solutions at the global level are required to ensure that effective drugs are made available to the most vulnerable groups. The Global Fund to Fight AIDS, Tuberculosis, and Malaria is providing the major leadership and resources for securing ACTs and other commodities, and many other agencies and organizations are joining coalitions to combat malaria through control, research, training, funding, and advocacy activities (Breman, Alilio, and Mills 2004; Feacham 2004: http://www.theglobalfund.org/en).
Economic Benefits of Malaria Control
Given the substantial total costs of several malaria control interventions, the case for introducing them can be strengthened by evidence on non-health-related benefits, especially evidence of income gains or prevention of income losses. Three different approaches to measuring such benefits are assessing the direct and indirect costs of malaria, studying the relationship between malaria and the output of agricultural or industrial activities, and exploring the macroeconomic impact of malaria. Only the third approach sheds much light on the benefits of control as opposed to the burden of the disease in the absence of control. This point is important, because in places facing the most severe malaria problems, direct information on the economic benefits of control is largely unavailable.
Direct and Indirect Costs of Malaria
The standard approach has been to view the two key determinants of the economic costs of malaria as the direct costs of expenditure on prevention and treatment and the indirect costs of productive labor time lost because of malaria morbidity and mortality, and to estimate the total economic impact by adding the direct and indirect costs (Mills 1992). Households use a range of preventive measures (mosquito coils, aerosol sprays, bednets, and mosquito repellents) to differing degrees. A review of evidence for Sub-Saharan Africa found that monthly per capita household expenditures ranged from US$0.05 per person in rural Malawi to US$2.10 in urban Cameroon, equivalent to US$0.24 and US$15 per household in 1999 U.S. dollars (Chima, Goodman, and Mills 2003). The costs of treatment for malaria include out-of-pocket expenditures for consultation fees, drugs, transport, and subsistence at a distant health facility. For Sub-Saharan Africa, these costs ranged between US$0.41 and US$3.88 per person, equivalent to between US$1.88 and US$26 per household per year. Household expenditure on treatment is usually highly regressive, consuming a much larger proportion of income in the poorest households.
Computations of public expenditures on malaria prevention and treatment are imprecise because most fall within general health service expenditures. About 20 to 40 percent of outpatient visits in malarious Africa are for fever, and suspected malaria among inpatients ranges from 0.5 to 50.0 percent of admissions. Kirigia and others (1998) found that inpatient treatment for pediatric malaria absorbed 15 percent of the annual recurrent costs of inpatient care in one Kenyan hospital and 9 percent in another. Ettling and Shepard (1991) estimated that Rwanda's Ministry of Health spent 19 percent of its operating budget on malaria treatment. Because of the dominance of out-of-pocket spending by households, public expenditures on malaria generally account for a minority of total malaria expenditure.
The methods used to measure and value time lost and a day's work vary considerably between studies; the average time lost per episode for a sick adult and for an adult caring for sick children ranged from one to five days. The average indirect cost per episode ranged from US$0.68 for children under 10 years of age in Malawi to US$23 per adult episode in Ethiopia. Authors have concluded that aggregate productivity losses can be significant for households and for the economy as a whole. In Malawi, the indirect costs of malaria were equivalent to 2.6 percent of annual household income (Ettling and others 1994), and Russell (2004) reports that malaria generally consumes less than 10 percent of family income. Leighton and Foster (1993) estimate that the total annual value of malaria-related production losses was 2 to 6 percent of gross domestic product in Kenya and 1 to 5 percent in Nigeria. The productivity consequences of mortality have received relatively little attention.
The direct and indirect cost approach involves two methodological problems. First, calculations are based mainly on days of illness and neglect mortality, debility (usually from anemia), and neurological and other long-term sequelae. Second, households' coping strategies are likely to reduce the immediate impact of illness, although in the long term they may impose costs through the sale of assets such as livestock, which jeopardizes a household's asset base.
Relationship between Malaria and Output
With the exception of a study of cotton production in Cote d'Ivoire, where the prevalence of parasitemia above a cutoff density had a major impact on labor efficiency (Audibert, Mathonnat, and Henry 2003), studies have not succeeded in identifying impacts on output. One possibility is that the risk of malaria may discourage the planting of crops that require intensive cultivation, the settlement and cultivation of fertile land, or the development of tourism and industry in suitable locations, but good evidence is lacking.
Macroeconomic Impact of Malaria
Recent empirical cross-country comparisons of economic growth indicate that eliminating malaria would have a strong positive impact on economic development. Gallup and Sachs (2001) use cross-country regression analysis to relate the growth in gross domestic product per capita between 1965 and 1990 to initial income levels, initial human capital stock, policy variables, geographical variables, and a malaria index calculated as the product of the fraction of land area with endemic malaria in 1965 and the fraction of malaria cases that were due to P. falciparum in 1990. Their results suggest that countries with a substantial amount of malaria grew 1.3 percent per year less than countries with little or no malaria between 1965 and 1990 (controlling for other influences on growth) and that a 10 percent reduction in malaria was associated with 0.3 percent higher growth per year. McCarthy, Wolf, and Wu (2000) employ a similar approach to explore the impact of malaria on average per capita growth rates during three five-year periods. They find a significant negative association between malaria and economic growth, although the estimated impact differed sharply across countries. The impact was smaller than that found by Gallup and Sachs (2001), exceeding 0.25 percent per year in only a quarter of the sample countries and averaging 0.55 percent for those in Sub-Saharan Africa.
Benefit-Cost Ratios
Mills and Shillcutt 2004 related the evidence on the macroeconomic benefits of malaria control to information on the costs of reducing malaria to calculate a benefit-cost ratio (Mills and Shillcutt 2004). Depending on the assumptions, the benefit-cost ratio ranged between 1.9 and 4.7 (using a 3 percent discount rate). In terms of economic growth alone malaria control is extremely cost beneficial.
