Bulletin of the World Health Organization

Global cost of child survival: estimates from country-level validation

Liselore van Ekdom, Karin Stenberg, Robert W Scherpbier, Louis W Niessen & on behalf of the ad hoc Study Group for Child Health Cost Validation

Volume 89, Number 4, April 2011, 267-277

Table 1. Stepwise validation models and data updates applied in the revision of 2007 World Health Organization (WHO) global estimates for child survival interventions, 2010–2015

Modela and description Reference year for price data Country-validated assumptions Reference year for population, incidence and intervention coverage data Inputs updated Expected effect on overall costs, all else being the same Finding
Original 2007 WHO
Original price tag 2004 No Population, 2002;
incidence and coverage, 2004
_ _ _
Validation (V) and update (U) analysis (n = 26)
Model V: original price tag estimates updated with country inputs 2004 Yes Population, 2002;
incidence and coverage, 2004
Country inputs Unknown (depending on higher/lower country validation of ingredients) Costs increased by 53% from original
Model U1: as per model V, with population updates 2004 Yes Population, 2008 Population, update from 2002 to 2008 projections (medium variant) Expect higher costs, as population estimates have increased on average since the 2002 projections for the 75 countries Costs for V2 decreased by 3% (vs V)
Model U2: as per model V, with incidence updates 2004 Yes Incidence15 of severe malnutrition Incidence, new formulas available for estimating incidence from prevalence Expect higher costs, as incidence estimates will increase Costs increased 15% (vs V)
Model U3: As per model V, with coverage updates 2004 Yes Coverage, Countdown 2008b Intervention coverage Expect lower additional costs (than original) because current coverage has most likely increased Coverage resulted in 5% higher costs (vs V) but effect not significant when combined with demographic data (U4)
Model U4: V and U1–U3 combined (population, incidence and coverage updates) 2004 Yes Population, 2008; incidence15 and coverage, Countdown 2008b Examine combined effect of updating population, incidence and coverage Unknown _
Sensitivity (S) analysis (n = 26)
Model S1: Model U4 rerun with alternative scale-up strategy (linear) 2004 Yes Population, 2008; incidence15 and coverage, Countdown 2008b Linear scale-up Unknown Costs decreased slightly due to cost drivers in the sample, e.g. China, Egypt and India, now with a slower scale-up trajectory than in the original analysis
Model S2: Model U4 rerun with alternative population projection, high variant 2004 Yes Population, 2008; incidence15 and coverage, Countdown 2008b Population, based on UN 2008 projections, high variant Expect higher additional costs Overall costs increased by 2% (vs U4); patient costs increased by 4%
Model S3: Model U4 rerun with alternative population projection, low variant 2004 Yes Population, 2008; incidence15 and coverage, Countdown 2008b Population, based on UN 2008 projections, low variant Expect lower additional costs Overall costs decreased by 3% (vs U4); patient costs decreased by 6%
Updating (U) analysis of the global price tag
Model U5 (n = 26): Model U4 rerun with updated 2005 WHO-CHOICE prices 2005 Yes Population, 2008; incidence15 and coverage, Countdown 2008b WHO-CHOICE prices Expect higher costs since WHO-CHOICE price update gives higher price estimates (due to changes in technology mix over time) Overall costs increased by 16% (vs U4)
Extrapolation
To 75 countries: results from Model U5 extrapolated 2005 Yes Population, 2008; incidence15 and coverage, Countdown 2008b No additional changes in assumptions: extrapolation from 26 to 75 countries _ _
Combination
Combinationc: estimates combined with costs for immunization, malaria and PMTCT of HIV 2005 NA NA Costs taken from recent publications on HIV/AIDS, malaria and immunization _ _

AIDS, acquired immunodeficiency syndrome; CAH, Department of Child and Adolescent Health and Development (WHO); HIV, human immunodeficiency virus; NA, not applicable; PMTCT, prevention of mother-to-child transmission; UN, United Nations.

a The CAH model includes patient-level intervention cost components: breastfeeding counselling, improvement of complementary feeding, severe malnutrition management, pneumonia management, diarrhoea management, antibiotic treatment for dysentery, measles complications, community-based case management, neonatal infections, vitamin A supplementation and regular deworming. In addition, it includes the following programme-level cost components: community health workers, supervision, training, monitoring and evaluation; information, education and communication; advocacy; laws, policy and regulation; infrastructure; technical assistance; general management.

b Countdown coverage data were only available for three interventions (management of diarrhoea, management of pneumonia and vitamin A supplementation) and for 23 of the 26 selected countries.

c Refers to a combination with other models that include immunization costs; general management and infrastructure for immunization; PMTCT of HIV programmes; general management of PMTCT programmes; costs of malaria interventions in children under five and general management of such malaria programmes.