Dunn, David Tyre; (1997) Statistical methods for assessing the risk and timing of vertical transmission of human immunodeficiency virus. Doctoral thesis (Ph.D.), University College London (United Kingdom).

Text Statistical_methods_for_assess.pdf Download (9MB)

Abstract

Various methods have been used to estimate the HIV vertical transmission rate and the paediatric AIDS incubation period. These are reviewed, and the standard method of analysis, which ignores children of indeterminate infection status, is shown to be biased. A new method of estimation that is appropriate for prospective studies conducted in non-breastfeeding populations is described. The method, based on an EM algorithm, takes into account clinical, virological, and immunological data and is more efficient than previous approaches. The method was applied to data from the European Collaborative Study and revealed evidence of temporal changes in the transmission rate and the AIDS incubation period. An individual patient meta-analysis of polymerase chain reaction (PCR) data on neonates subsequently shown to be HIV-infected was conducted. The main objectives were to estimate the age-specific sensitivity of the assay and the relative contributions of intrauterine and intrapartum transmission. Distribution-free and parametric approaches were used for the analysis of the data, which were interval-censored. The sensitivity of PCR was shown to be higher than previously thought. Approximately one-third of vertically-acquired HIV infection could be attributable to intrauterine transmission. In a retrospective cohort study conducted in the State of Sao Paulo, Brazil, there was considerable variation in breastfeeding practice. Models, with structures reflecting the series of potential exposures to the virus (intrauterine, intrapartum, breastfeeding), were developed to estimate the risk of breastfeeding according to duration. Although breastfeeding per se was a strong risk factor for transmission, duration of breastfeeding was not. The extent to which this finding could be explained by variability in maternal infectivity was investigated.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item