| Summary: | Abstract Background Numerous studies have reported the economic burden of childhood diarrhea in low- and middle-income countries (LMICs). Yet, empirical data on the cost of diarrheal illness is sparse, particularly in LMICs. In this study we review the existing literature on the cost of childhood diarrhea in LMICs and generate comparable estimates of cost of diarrhea across 137 LMICs. Methods The systematic literature review included all articles reporting cost estimates of diarrhea illness and treatment from LMICs published between January 2006 and July 2018. To generate country-specific costs, we used service delivery unit costs from the World Health Organization’s Choosing Interventions that are Cost-Effective (WHO–CHOICE database). Non-medical costs were calculated using the ratio between direct medical and direct non-medical costs, derived from the literature review. Indirect costs (lost wages to caregivers) were calculated by multiplying the average GDP per capita per day by the average number of days lost to illness identified from the literature. All cost estimates are reported in 2015 USD. We also generated estimates using the IHME’s service delivery unit costs to explore input sensitivity on modelled cost estimates. Results We identified 25 articles with 64 data points on either direct or indirect cost of diarrhoeal illness in children aged < 5 years in 20 LMICs. Of the 64 data points, 17 were on the cost of outpatient care, 28 were on the cost of inpatient care, and 19 were unspecified. The average cost of illness was US$36.56 (median $15.73; range $4.30 – $145.47) per outpatient episode and $159.90 (median $85.85; range $41.01 – $538.33) per inpatient episode. Direct medical costs accounted for 79% (83% for inpatient and 74% for outpatient) of the total direct costs. Our modelled estimates, across all 137 countries, averaged (weighted) $52.16 (median $47.56; range $8.81 – $201.91) per outpatient episode and $216.36 (median $177.20; range $23.77 –$1225.36) per inpatient episode. In the 12 countries with primary data, there was reasonable agreement between our modelled estimates and the reported data (Pearson’s correlation coefficient = .75). Conclusion Our modelled estimates generally correspond to estimates observed in the literature, with a few exceptions. These estimates can serve as useful inputs for planning and prioritizing appropriate health interventions for childhood diarrheal diseases in LMICs in the absence of empirical data.
|
|---|
