Early treatment for infants cheaper than delaying treatment, CHER study shows

Gesine Meyer-Rath. ©IAS/Steve Forest/Workers' Photos
This article is more than 14 years old. Click here for more recent articles on this topic

Starting antiretroviral treatment early in HIV-infected infants, at a median of seven weeks of age, resulted in cost savings of 80%, compared to deferring treatment until a median age of 29 weeks, reported Gesine Meyer-Rath today at a late-breaker session of the Eighteenth International AIDS Conference in Vienna. Her presentation gave the results of a comparative economic analysis of treatment in the first year of life from the Children with HIV Early Antiretroviral Therapy (CHER) trial in South Africa.

While outpatient costs for those infants in a real-life "routine" care setting were three to four times lower than in the two other arms (early and deferred) due to lack of antiretroviral therapy, the savings were cancelled out by the higher costs of inpatient care, accounting for 85% of total costs. 

The total cost of care for each child in "routine" care with delayed access to antiretroviral treatment was over twice the total cost of care for children who got early treatment.



Of or relating to children.

CD4 cell percentage

The CD4 cell percentage measures the proportion of all white blood cells that are CD4 cells.

disease progression

The worsening of a disease.


Having no symptoms.

The preliminary results of the ongoing CHER landmark study, conducted in Johannesburg and Cape Town, South Africa, have led the World Health Organization (WHO) to recommend immediate treatment of all HIV-infected children under the age of two regardless of their CD4 count or disease stage.

The study showed that starting antiretroviral treatment in asymptomatic HIV-infected infants with a CD4 percentage of less than 25%,at a median age of seven weeks, reduced the death rate from 16 to 4% in a follow-up period of 32 weeks, compared to those where treatment was delayed – representing a 76% reduction in infant deaths in the immediate-treatment arm.  Those with delayed treatment showed a rapid disease progression and sudden death.

Concerns over the cost of early treatment, as well as the affordability of full coverage with paediatric treatment, appear to be one of the the primary obstacles to implementation. This led the researchers to analyse the full cost of care during the first year of life from a provider perspective.

The researchers chose to compare the costs of this strategy with “routine” care in a real-life clinic setting where eligibility for treatment was based on WHO guidelines in use at the time: stage 3 or 4 clinical disease or CD4 percentages under 20%. In this setting, the mean CD4 percentage was 15% when starting antiretroviral treatment.

Data on use of outpatient and inpatient resources data were collected for each of the three arms for the first 12 months of life: the CHER trial comprised both 373 infants randomised to receive either early (arms 2/3; 284 in total) or deferred (arm 1; totalling 89) antiretroviral treatment, and 143 infants who started antiretroviral treatment at the Empilweni Clinic in Johannesburg between 2005 and 2007 (the "routine" arm).

Mean ages at the start of antiretroviral treatment were 10, 20 and 27 weeks for the early, deferred and “routine” arms, respectively.

"Patient resource use" included the number of clinic consultations, antiretroviral drugs given out, laboratory tests and days spent as an inpatient. Clinic and hospital accounts provided staff costs inclusive of benefits, equipment, supplies and overheads. The South African government medical department provided the unit cost of drugs and the government laboratory service provided costs of laboratory tests. All cost data were based on 2009 figures.

The infants were followed from the start of treatment for ten months, nine months and three months in the early, deferred and “routine” arms, respectively.  

The mean cost per child on deferred treatment was US$2432 (95% CI: 1982 to 2889,) whereas on early treatment it was US$1349 (95% CI: 1244 to 1464) – a significant cost saving of 45%.

The mean inpatient costs in the early treatment arm represented 26% of total costs or US$346 whereas inpatient costs in the deferred arm amounted to 51% of the total costs or US$1237.  

The mean cost per child in “routine” care was US$2,908 (95%, CI: 2273 to 3743), and in this arm inpatient costs represented a staggering 85% (US$2523) of the total costs.

The cost differences are mostly due to the number of hospitalisations. The mean number of days as inpatients was:

  • Early: two days for each child (with a maximum of 68 days).
  • Deferred: seven days for each child (with a maximum of 84 days).
  • Routine: 13 days for each child (with a maximum of 121 days).

The level of hospitalisation reflects the severity of illness of each child.  

Laboratory costs and staff and overheads did not differ significantly between the early and deferred arms. 

The researchers then looked at what the budget implications would be for South Africa, based on the National ART Cost Model (NACM)/Budget Review of the National Treasury.

Assuming early paediatric treatment at 90% coverage:

  • For the fiscal year 2010/11, 103,000 children would be on ART at a total cost of US$67 million – representing 6% of the cost of the national ART programme and 1% of the public health budget.
  • For the fiscal year 2011/12, 162,000 children would be on ART at a total cost of US$104 million – representing 6% of the cost of the national ART programme and 1% of the public health budget.
  • For the fiscal year 2012/13, 202,000 children would be on ART at a total cost of  US$133 million – representing 6% of the cost of the national ART programme and 1% of the public health budget.

Dr Meyer-Rath noted that the cost of the paediatric antiretroviral treatment programme will always be overshadowed by the cost of the adult programme, irrespective of eligibility criteria.

Dr Meyer-Rath noted a number of limitations including:

  • Daily inpatient care costs were based on an average cost across all wards, and so represent a conservative estimate.

  • Costs of screening all HIV-exposed children were not included. This would amount to an additional US$300 for each child.

  • The treatment of those children in the early-treatment arm differed from treatment in practice and included better follow-up and being on a lopinavir/ritonavir-containing regimen.

  • All arms consisted of children who had survived to start antiretroviral treatment.

  • The difference in costs depended on children being taken to the hospital for admission.

Early treatment during the first year of life not only improves infant survival but provides significant cost savings, notably in terms of resource use within the hospital or clinic setting, than either deferred or “routine” treatment, the research group concluded.

Further information

View abstract and slides from this session on the official conference website.


Meyer-Rath G et al. The cost of early vs. deferred paediatric antiretroviral treatment in South Africa – a comparative analysis of the first year of the CHER trial. Eighteenth International AIDS Conference, Vienna, late breaker abstract THLBB103, 2010.