Health-service costs for the treatment of multidrug-resistant/rifampicin-resistant tuberculosis in children in South Africa: application of a real-world dataset

Background

Children with multidrug-resistant (MDR) / rifampicin-resistant (RR)-tuberculosis (TB) are an important but neglected group in cost-effectiveness research. Digital health information systems enable new approaches to health-service cost analysis. The Provincial Health Data Centre (PHDC) in the Western Cape, South Africa, collates disparate health system data including hospital inpatient and outpatient data, medication, laboratory tests, and primary healthcare utilisation.

Methods

Cost analysis was conducted using anonymised, integrated PHDC data for children with MDR/RR-TB between 1 January 2018 and 31 December 2021. Health-service costs were assessed by patient and disease characteristics including age, sex, drug susceptibility type, site of disease, and HIV status.

Results

There was significant cost variation across the n=271 children in the data sample (median US$7,576 (IQR $2,725 - $22,986)). The distribution of total per patient costs fitted a gamma distribution (mean US$13,435, α = 0.93, β = 14,496). Regression analysis indicates age, disease site, living with HIV, and treatment duration had significant impact on costs, whereas the impact of resistance profile and sex was not significant.

Conclusion

Treatment for MDR/RR-TB in children remains costly for health systems. Utilising routinely collected, real-world data from an established health information system enables accurate and representative insights to overall costs and major cost drivers. Costs were highly skewed, with a small proportion of patients incurring very high costs. This cost analysis can assist in decision-making and programme development at local and international levels and as an input to secondary analysis.

Key points

• Health-service costs of treating children who have drug resistant tuberculosis can be high, but the cost vary widely among patients

• Key factors that affect costs are age, the site of the tuberculosis disease, whether the child has HIV, and the length of time that a child receives treatment.

• Treating children in hospital is a major component of costs, indicating that programs or treatments that can reduce hospitalisation are likely to reduce costs.