A matched-cohort study of HIV-infected and HIV-uninfected mothers and their infants 18-20 months after their estimated delivery date was conducted between August and December 2009.
This study was conducted in Zomba district (population 670,000) in southern Malawi where 80% of inhabitants are rural [8
]. Antenatal surveillance data from sentinel health centers show that HIV prevalence is high, ranging from 12% to 34% [4
]. Public health services are provided at one central hospital, one mission hospital and 33 health centers, run by the Zomba District Health Office (DHO). Clinical services and medications are provided by the MOH without user fees. Since 2005, the Zomba DHO in partnership with Dignitas International, have implemented PMTCT services. By 2007, 22 ANC sites in Zomba District provided PMTCT services. Zomba District data indicate that 96% of all women attend at least one ANC visit and that 56% of deliveries are performed by skilled birth attendants in health institutions [9
20 out of 22 sites in the district where PMTCT services were provided for at least 18 months were included in this study. 2 sites, Zomba Central Hospital and St. Luke's Mission Hospital, were excluded as women who gave birth at these district level referral hospitals may not live in proximity and therefore would be more difficult to trace, and they may also be a different population than those who attend health centres. All HIV-infected mothers who attended one of the 20 rural public facilities that provide antenatal services, and had an estimated delivery date between March 1st and May 31st 2008, were identified through antenatal, delivery and postnatal facility registers. By reviewing all available registers, we attempted to capture the entire cohort of HIV-infected pregnant women in the district. For every HIV-infected mother, the next registered HIV-uninfected mother was identified as a control. 5 other health centers were excluded during this process since few (<10) HIV positive women were identified.
Community sensitization meetings were conducted prior to data collection to inform community leaders and health facility teams about the study. In addition, study advertisement posters were displayed at health centers and in the villages. Health Surveillance Assistants (HSAs) helped trace participants and asked the identified mother-child pairs to come to the health centre for interviews. Since equal numbers of HIV-infected and HIV-uninfected mothers were recruited there was no association between study participation and HIV status. In cases where the mother had died, the child's primary caregiver was interviewed. If the mother or the primary caregiver did not come for interview, the interviewer, accompanied by the HSA, visited the participant at home for the interview. In cases where study participants were reported to have moved or died, village headmen were consulted and village registers reviewed for verification.
Data collection was conducted by 6 trained female interviewers through semi-structured interviews using a standardized questionnaire.The questionnaire administered to the mothers included questions regarding socio-economic demographics, parity, disclosure of HIV status, partner testing, uptake of PMTCT services from HTC at ANC to uptake of sd-NVP and HAART, and feeding options utilized. Information concerning prior HIV-testing and ARVs was verified through personal health passports, a small personal paper booklet that provides a record of clinical encounters. Responses on uptake of sd-NVP or HAART were verified with available information in health passports and registers. When information differed or was missing (e.g. when mother had died) the information from the registers was included.
Mothers and/or children with negative or unknown HIV status were offered point of care HIV rapid testing. All testing was done by trained counselors as per the Malawian MOH guidelines. Mothers and/or children found HIV-infected were referred to the nearest clinic which could provide appropriate ART services, including ongoing care and assessment for ART eligibility.
Data and statistical analysis were conducted using STATA 9.1 (StataCorp, Texas, USA) and SPSS 17.0 (SPSS, Inc., Chicago, IL, USA). Comparisons between groups were made using t-tests and exact tests. Relative risks (RR) as a ratio of two proportions with a confidence level of 95% were derived from two-by-two contingency tables. Reported RR's are unadjusted, unless stated otherwise and all variables were considered significant with a p < 0.05. HIV-free survival by 18-20 months was calculated as: total number of cohort minus number of transmissions, deaths and not tested by 18-20 months divided by total number of cohort minus number not tested.
This study received ethical approval from the National Health Science Research Committee in Malawi. Written informed consent was given by all participants.