How well do disease-specific studies and WHO-CHOICE cost estimates align? Example of estimating cost per episode of diarrhoea and respiratory syncytial virus in 128 low-income and middle-income countries

OBJECTIVE: Non-disease-specific WHO-CHOICE (CHOosing Interventions that are Cost-Effective) unit costs are often used in cost and cost-effectiveness studies in the absence of country-specific data. This study aims to compare reported country-specific disease costs and the corresponding WHO-CHOICE estimates, using generically defined 'diarrhoea' (including rotavirus diarrhoea) and pathogen-specific 'respiratory syncytial virus (RSV)' disease in children as examples.

METHODS: We updated systematic reviews for both diseases in low-income (LICs), lower middle-income (LMICs) and upper middle-income (UMICs) countries. Diarrhoeal (including a subanalysis of rotavirus-specific diarrhoea) and RSV-specific outpatient and inpatient costs per episode in children were extracted and compared with WHO-CHOICE estimates in the same countries. All costs were updated to 2022 international dollar values. If a consistent pattern of underestimation or overestimation was identified, we quantified the magnitude of the discrepancy as the ratio of published disease-specific costs and corresponding WHO-CHOICE-based estimates.

RESULTS: Out of 1979 records identified, 23 cost studies were included. Including previous reviews, we retained 31 diarrhoea and 16 RSV studies for comparison. WHO-CHOICE-based direct medical costs were similar for diarrhoeal disease (including rotavirus diarrhoea), but lower for RSV-related disease. We estimated the cost per episode of diarrhoea and RSV in 128 countries. RSV outpatient costs were adjusted by multiplying WHO-CHOICE costs by 6.89 (95% uncertainty interval: 5.58 to 8.58) in LICs and LMICs and 5.87 (4.95 to 6.96) in UMICs; RSV inpatient costs were multiplied by 1.43 (1.01 to 2.01) and 1.36 (0.82 to 2.27), respectively.

CONCLUSION: While informative for economic evaluations, WHO-CHOICE-based cost estimates should be used cautiously. Our analysis shows they aligned well with empirical studies for diarrhoeal disease but underestimated the costs of RSV-related disease. For diseases with few country-specific costing studies, comparing findings of the empirical studies with WHO-CHOICE estimates is crucial before conducting economic evaluations for countries without data. We propose a simple approach for calculating adjustment factors for WHO-CHOICE estimates when empirical data on disease-specific diagnosis and treatment costs are limited.