Pooled PCR testing of dried blood spots for infant HIV diagnosis is cost efficient and accurate
CITATION: Van Schalkwyk, C., et al. 2019. Pooled PCR testing of dried blood spots for infant HIV diagnosis is cost efficient and accurate. BMC Infectious Diseases, 19:136, doi:10.1186/s12879-019-3767-z.
The original publication is available at https://bmcinfectdis.biomedcentral.com
Publication of this article was funded by the Stellenbosch University Open Access Fund
Background: Access to qualitative HIV PCRs for early infant diagnosis (EID) is restricted in resource-limited settings due to cost. We hypothesised that pooling of dried blood spots (DBS), defined as combining multiple patient samples in a single test with subsequent individual testing of positive pools, would be cost saving while retaining clinical accuracy compared to individual patient testing. Methods: Cost savings: A model was developed to simulate reagent and consumable cost saving of pooled compared to individual sample testing. Daily sample/result data of a public health laboratory in South Africa were used to illustrate outputs from the model. Samples were randomly allocated to pools and the process was repeated 1000 times to measure variation in estimates due to this stochasticity. Clinical accuracy: 1170 patient samples were tested using the Roche CAP/CTM Qual assay in pools of five 50 μl DBS. Negative pools comprised DBS previously tested in single reactions; positive pools included 1 positive sample. Results: Pooling would have saved 64% of laboratory costs in 2015. The model is published as an R-based web tool, into which the user enters sample/positivity estimates and workflow management parameters to obtain cost saving estimates at an optimal pool size. Sensitivity of pooled testing was 98.8% overall; 100% for strongly reactive pools. One pool tested false positive which would not impact clinical specificity as individual patient testing is performed prior to reporting. Conclusions: Pooled PCR testing for EID remains accurate and dramatically reduces costs in settings with moderate to low prevalence rates and sufficient sample numbers.