University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Cost-effectiveness of routine rapid human immunodeficiency virus antibody testing before DNA-PCR testing for early diagnosis of infants in resource-limited settings
Global Health Sciences Literature Digest
Published December 13, 2010
Journal Article

Menzies NA, Homsy J, Chang Pitter JY, et al. Cost-effectiveness of routine rapid human immunodeficiency virus antibody testing before DNA-PCR testing for early diagnosis of infants in resource-limited settings. Pediatr Infect Dis J. 2009 Sep;28(9):819-25.

In Context

Infants born to HIV-positive mothers benefit from early and regular testing so that they can be treated early if they become infected. In resource-poor settings, however, widespread infant HIV testing is logistically and financially difficult. Specific tests, such as polymerase chain reaction (PCR) for HIV-DNA, are usually too expensive, require transportation from the field to centralized laboratories, or both.(1, 2, 3, 4) Antibody-based rapid HIV tests (RHTs) are used widely for adults.5 In infants, however, these tests also detect maternal antibodies and therefore can be falsely positive.(4, 5)


To determine whether it is cost-effective to use rapid HIV antibody tests to screen out antibody-negative infants prior to DNA-PCR testing; to evaluate the sensitivity and specificity of rapid tests in infants.


Tororo District Hospital in eastern Uganda and Mulago National Referral Hospital, Kampala.

Study Design

Modeling of cost-effectiveness; a cross-sectional cohort study comparing rapid HIV and DNA-PCR testing.


HIV-exposed infants and children 1.5 to 18 months of age and their mothers/caregivers


Mother-infant pairs were enrolled in a pilot HIV-testing program in 2005-2006. Blood samples from the infant were tested with a single RHT (Determine or Stat-Pak); a dried blood-spot (DBS) sample also was sent for HIV-1 DNA-PCR testing at the reference laboratory. An HIV-negative diagnosis was only provisional for infants still breastfeeding or who had been weaned less than six weeks earlier. Additional retesting was required after all breastfeeding had been stopped. Sensitivity and specificity of RHT were calculated using DNA-PCR results as the standard. Costs were based on time-motion studies; capital equipment costs were included (3% discount rate). A decision-analytic model (model #1) was used to compare conventional testing (PCR-DNA tests for all infants) versus a "modified algorithm" in which infants older than age three months were screened with an RHT, and only infants with positive results had DBSs sent for PCR testing. DNA-PCR testing was used for all infants three months of age or younger. The model included costs, HIV prevalence, RHT performance, and client loss to follow-up. Outputs were a) percentage of infected infants correctly diagnosed and caregivers informed; b) average cost per infant tested; and c) average cost per HIV-infected infant identified. Analyses were stratified by infant age, HIV prevalence, and symptoms.

In model #2, the performance of the conventional and the modified algorithms were compared for a hypothetical cohort of HIV-exposed infants at a program level. Different scenarios were constructed based on different levels of compliance with testing, breastfeeding, and follow-up.


In the cohort study, data were available for 788 HIV-exposed infants. The HIV prevalence was 26% overall, increasing steadily from 16% in the youngest group (1.5-3 months) to 48% among those >12-18mos of age. The sensitivity of RHT was 93%-94%; specificity was 37%-28% and improved with infant age. Supplies were the primary costs, particularly the PCR Amplicor test kit costs (nearly $20). The percentage of infected infants correctly diagnosed was slightly better for the conventional algorithm (94.3%) compared to the modified algorithm (87.8%). The modified algorithm showed lower costs for infants younger than age three months, reducing costs by 16%-68%. These cost savings were greater in asymptomatic infants, where a higher proportion of uninfected infants could be screened out by RHT. The modified algorithm also showed reductions in the number of clinic visits, the number of PCR tests needed, and wait time for results. In the conventional algorithm, 94.4% of infants were diagnosed and caretakers informed of results, whereas for the modified algorithm, it was 94.9% among infants aged 1.5-3 months, but 99.8% in those aged 9-12 months. The average cost per infant tested was $23.47 using the conventional algorithm. Model #2 projected effectiveness, costs, and cost-effectiveness algorithms with results calculated under perfect, intermediate and poor compliance scenarios; intermediate compliance was defined as 10%-11% loss to follow-up, and 30% of infants continuing to breastfeed until 18 months; poor compliance was 9%-20% loss to follow-up, and 53% continuing to breastfeed until 18 months. Incremental cost per infant correctly diagnosed and informed of results (comparing the conventional and the modified algorithm) ranged from $539 with low compliance to $7165 with perfect compliance. These costs were reduced if retesting at nine and 18 months was instituted-the cost per infected infant diagnosed and informed of result was $23.83. Sensitivity analysis indicated that costs per infected infant diagnosed and informed were most sensitive to changes in DNA-PCR costs, breastfeeding behavior, and initial HIV prevalence.


This study shows that using RHT before DNA-PCR is almost as effective as using only a single DNA-PCR test, but is less costly and more cost-effective. Cost-effectiveness was further improved by retesting uninfected but exposed infants at nine and 18 months of age.

Quality Rating

This was a high quality cost-effectiveness study and included sensitivity analyses.

Programmatic Implications

Using RHT to identify infants who are negative for HIV antibodies and older than three months could be a less costly approach to early infant diagnosis. Only infants who are HIV positive on the rapid antibody test would need further PCR testing for confirmation because the rapid testing algorithm is likely to identify false-positive infants. This method, however, is likely to miss up to 6% of HIV-infected infants when they are younger than three months of age. Regardless of method, testing must be repeated among infants who continue to breastfeed.


  1. Lambert JS, Harris DR, Stiehm ER, et al. Performance characteristics of HIV-1 culture and HIV-1 DNA and RNA amplification assays for early diagnosis of perinatal HIV-1 infection. JAIDS. 2003;34:512-519.
  2. Sherman GG, Cooper PA, Coovadia AH, et al. Polymerase chain reaction for diagnosis of human immunodeficiency virus infection in infancy in low resource settings. Pediatr Infect Dis J. 2005;24:993-997.
  3. Sherman GG, Stevens G, Jones SA, et al. Dried blood spots improve access to HIV diagnosis and care for infants in low-resource settings. JAIDS. 2005;38:615- 617.
  4. Bremer JW, Lew JF, Cooper E, et al. Diagnosis of infection with human immunodeficiency virus type 1 by a DNA polymerase chain reaction assay among infants enrolled in the Women and Infants' Transmission Study. J Pediatr. 1996;129:198 -207.
  5. Downing RG, Otten RA, Marum E, et al. Optimizing the delivery of HIV counseling and testing services: the Uganda experience using rapid HIV antibody test algorithms. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;18:384 -388.