Cost-Effectiveness of Interventions
Cost-effectiveness ratios of diarrheal disease interventions were calculated by World Bank region in terms of disability-adjusted life years (DALYs) averted for a model population of 1 million, following the standardized guidelines of the Disease Control Priorities Project for economic analyses (see chapter 15). Europe and Central Asia were excluded because data were lacking owing to the low prevalence of disease. Input variables included (a) region-specific diarrhea morbidity rates adapted from Kosek, Bern, and Guerrant (2003); (b) region-specific underlying mortality rates and age structures provided by the Disease Control Priorities Project; (c) median intervention effectiveness rates (that is, percentage of diarrheal morbidity reduction and percentage of diarrheal mortality reduction); and (d) median per capita intervention costs gathered from the literature and from personal communications (table 19.1).
[Table .]
Because approximately 90 percent of all cases in the developing world occur in children under five, the analysis focused on this age group alone. Uniform intervention effectiveness rates were assumed for all regions because region-specific information was not available. Regional variations in cost-effectiveness were due to regional variations in the prevalence of diarrheal disease, in the diarrhea-attributable morbidity and mortality, and in the intervention cost, where region-specific information was available.
Disability-adjusted life years are averted through the avoidance of cotemporaneous disability and mortality attributable to diarrhea. We did not consider long-term developmental and cognitive effects of childhood diarrhea or the external benefits of interventions unrelated to diarrhea (for instance, benefits of measles immunization unrelated to diarrhea or other health benefits of improved public water and sanitation). Therefore, our estimates err on the conservative side.
We explored two general categories of interventions: early interventions that take place within the first year of life—breastfeeding promotion and immunizations for rotavirus (with the prototype Rhesus reassortant tetravalent vaccine), cholera (with live oral vaccine), and measles—and other interventions that treat an entire cohort of children under five simultaneously (improved water and sanitation). For early interventions, cost-effectiveness ratios were calculated by considering the cost of treating all newborns in a single year and the benefits (DALYs averted) from those treatments that occur over the first five years of life. These benefits include avoided mortality that allows individuals to live to the expected life expectancy for the region. Other interventions included ORT and improved water and sanitation infrastructure. Because a single year of these interventions yields only cotemporaneous benefits—because effectively treated individuals do not necessarily live to life expectancy given that they are likely to be reinfected the next year—we calculated cost-effectiveness of a five-year intervention. Analysis of a five-year intervention enabled us to consider the case in which an entire cohort of children age zero to four avoids early childhood diarrheal mortality because of the intervention and receives the benefit of living to life expectancy.
Disability and deaths averted for those benefiting from improved water and sanitation were calculated from only the fraction of the model populations currently without access. For each region, the proportion of rural and urban children age zero to four currently without access to improved water and sanitation was calculated using region-specific information from World Bank Development Indicators (World Bank 2002) for 2000. Infrastructure improvements for rural and urban populations were considered separately because of differences in infrastructure type and cost, although the same effectiveness rates were used for both.
The per child treatment costs and effectiveness rates used are presented in table 19.1. Cost per treatment of ORT varied widely depending on the type and method of ORT implemented. Oral rehydration therapy can be as inexpensive as US$0.02 per child treated—the cost of a home remedy with sugar and salt. However, treatment can become substantially more expensive if commercially manufactured ORS is used or if there are substantial personnel or infrastructure costs (Martinez, Phillips, and Feachem 1993). Finally, our analysis considered only long-run marginal costs (which vary with the number of individuals treated) and did not include fixed costs of initiating a program where none currently exists.
Figure 19.3 shows the cost-effectiveness of all interventions over the first five years of life. Two interventions administered during the first year of life—breastfeeding promotion (US$930 per DALY) and measles immunization (US$981 per DALY)—were the most cost-effective. ORT (US$1,062 per DALY) and water and sanitation in rural areas (US$1,974 per DALY) were the next most cost-effective, but only if they were implemented continuously for five years, thereby allowing an entire cohort of effectively treated children age zero to four to survive past the age at which they are most at risk for diarrheal infection, disability, and mortality. Rotavirus immunization (US$2,478 per DALY), cholera immunization (US$2,945 per DALY), and water and sanitation in urban areas (US$6,396 per DALY) were the least cost-effective.
[Figure
19.3]
Among the early interventions, breastfeeding promotion was less effective than other interventions but also less expensive than rotavirus and measles vaccination (table 19.1). Cholera vaccination was less expensive than breastfeeding promotion, but it was also many times less effective because of the significantly higher prevalence of diarrhea that is not related to cholera—making cholera vaccination the least cost-effective of the early interventions considered. Oral rehydration therapy and water and sanitation interventions were more effective than breastfeeding and vaccination interventions in reducing morbidity and mortality caused by diarrhea, but they were also more expensive. However, our analysis for water and sanitation did not consider the benefits of this intervention other than those related to health, and the high cost-effectiveness ratio is more a limitation of our methodology than of the intervention itself.
The high cost-effectiveness ratio for ORT is attributable to the high variation in reported treatment costs, which may inflate the median cost used in this analysis (table 19.2). Given the range of reported treatment costs (table 19.1), the cost-effectiveness ratio of ORT could be as low as US$4 per DALY or as high as US$2,124 per DALY in low- and middle-income countries. High variation in reported treatment costs results in high variation in cost-effectiveness for the other regions as well. There remains little doubt, however, about the effect of widespread use of ORT on diarrhea morbidity and mortality and about the associated direct and indirect cost savings for treatment and hospitalization.
[Table .]
