Background: Loss to follow-up (LTFU) deprives HIV-exposed infants the lifesaving care required and results in exposing HIV free infants to virus requisition risk. We aimed to determine the rate of LTFU, postnatal mother-to-child HIV-transmission (MTCT) and to identify maternal factors associated with LTFU among HIV-exposed infants enrolled at Mbarara Regional Referral Hospital PMTCT clinic. Methods: Study participants were infants born to HIV-positive mothers enrolled in the PMTCT clinic for HIV care at Mbarara Regional Referral Hospital. While access database in the Early Infant Diagnosis (EID) clinic provided data on infants, the open medical record system database at the ISS clinic provided that for mothers. Infants were classified as LTFU if they had not completed their follow-up schedule by 18 months of age. At 18 months, an infant is expected to receive a rapid diagnostic test before being discharged from the PMTCT clinic. Postnatal MTCT of HIV was calculated as a proportion of infants followed and tested from birth to 18 months of age. Logistic regression was used to determine possible associations between mothers’ characteristics and LTFU. In-depth interviews of mothers of LTFU infants and health workers who attend to the HIV-exposed infants were carried out to identify factors not captured in the electronic database. Results: Out of 1624 infants enrolled at the clinic, 533 (33%) were dropped for lack of mother’s clinic identification number, 18 (1.1%) were either dead or transferred out. Out of 1073 infants analysed, 515 (48%) were LTFU by 18 months of age while out of the 558 who completed their follow-up schedule, 20 (3.6%) tested positive for HIV. Young age of mother, far distance to hospital and non-use of family planning were identified as outstanding factors responsible for LTFU. In addition, in-depth interviews revealed facility-level factors such as “waiting time” which would not be found in routine client databases. Conclusion: This study has revealed a high rate of loss to follow up among HIV-exposed infants enrolled at Mbarara Regional Referral hospital PMTCT clinic. Young maternal age, long distance to health facility and failure to use family planning were significantly associated with LTFU. Incorporating family planning services in the ART and PMTCT clinics could reduce loss to follow-up of HIV exposed infants. Young mothers should be targeted with information on the importance of completing the EID follow-up schedule and also, their clinic identification number be gotten at each visit.
This was a mixed study design comprising a descriptive retrospective registry-based cohort study and In-depth Interviews. The retrospective study was carried out on infants born to HIV positive mothers attending the PMTCT/EID clinic at Mbarara Hospital Immune Suppression Syndrome (ISS) clinic. In addition, qualitative interviews of mothers whose babies were lost to follow-up and health workers in the EID/PMTCT clinic were carried out. The study was conducted at the ISS clinic of Mbarara Regional Referral Hospital (MRRH) in Mbarara district. Mbarara district is found in the south-western part of Uganda, and is located 270kms from Kampala city. At this hospital like in all maternal child health (MCH) facilities, all pregnant women attending antenatal care (ANC) clinic receiving counselling, are tested for HIV and are recorded in PMTCT or ANC registers. All HIV positive women are recorded in PMTCT care registers and instantly placed on lifelong ART treatment regardless of CD4 count or gestation period. They are started on ART combination of TDF, 3TC and EFV. Infants born to mothers infected by HIV are documented in the EID register after birth, and followed up till they are 18 months old and within this time, mothers receive counselling on infant feeding along with ARV prophylaxis for PMTCT of HIV. To this effect, infants whose mothers are placed on lifelong ART are given once daily NVP from birth to 6 weeks of age regardless of whether they are exclusively breastfed or given replacement feeding and thereafter receive cotrimoxazole prophylaxis. According to the Uganda HIV infant testing algorithm, an HIV DNA PCR test is carried out at ≤6 weeks of age and cotrimoxazole is then started. In the case where the 1st DNA PCR is negative and the child has been breastfeeding, the 2nd DNA PCR test is carried out 6 weeks after breastfeeding has stopped. When a 2nd PCR turns out to be negative, a rapid HIV test is conducted at 18 months old prior exit of care by child. On the other hand, if the 1st or 2nd PCR is positive, the infant is referred for ART initiation. By June, 2014, the EID/PMTCT clinic at Mbarara Hospital ISS clinic had 3120 infant’s cumulative enrolment since its inception in 2005, and by January 2013, 2072 infants enrolled. The study was delimited to HIV exposed infants who enrolled for care at the Mbarara Hospital PMTCT/EID clinic between January 2010 and January 2013 and whose mothers received care from Mbarara Regional Referral Hospital ISS clinic during the same period. Infants whose mothers’ lacked clinic identification numbers were deemed ineligible. All HIV-exposed infants who had been enrolled from January 2010 to January 2013 and followed up for 18 months and whose mothers received care from ISS clinic were analysed. A sample size of 772 mother-infant pairs was estimated based on assumptions of 80% power; odds ratio for loss to follow up of 2.2 [7] and level of significance of 0.05 (two-sided). This estimate was reached at using a method in observational epidemiology for calculating sample size for unmatched cross-sectional studies, cohort studies and randomized clinical trials [7, 8]. Data on HIV-exposed infants were obtained from the Access database system for the PMTCT/EID clinic whereas, data on mothers were obtained from the electronic Open Medical Record System (Open MRS), the database for the ISS clinic. Data of eligible mother-infant pairs were identified by unique client numbers in the electronic databases and the dataset converted into STATA format (Stata Corporation Inc.). The main outcome variable was loss to follow-up among HIV-exposed infants. An infant was classified loss-to follow-up if he/she did not complete follow up to the point of being discharged and was not declared deceased. Data on independent variables were extracted as recorded in databases and later categorised where necessary as described by Kabakyenga and colleagues (1). For example, age was categorised and coded as 0 “18–23”, 1 “24–29” and 2 “> 30 years”. Fourteen interviews were conducted, ten of them with selected HIV positive mothers whose exposed infants were lost to follow up and four with health workers who attend to these infants. The purpose of the qualitative interviews was to capture factors associated with LTFU that could have been missed out in the electronic database of the ISS clinic such as facility-level client experiences and health worker behaviours. Tracing of mothers of lost babies was done by phone calls after their information had been generated from the data set. Only mothers who had left their phone contacts with the clinic were contacted. Interviews were audio recorded, transcribed word verbatim and translated approximately from the local dialect to English language for analysis. Quantitative data analysis was done using Stata version 11.0 (Stata Corp, College Station, TX, USA). Frequency counts and percentages were obtained to describe socio-demographics and other categorical variables. The median and inter-quartile ranges were used to describe quantitative variables with skewed distributions such as CD4 count. In the bivariate analysis, independent variables were cross tabulated with the outcome variable to determine possible associations. Odds Ratios and their 95% confidence intervals were calculated. Independent variables with p-value ≤ 0.2 in the bivariate analysis were entered into a multivariable logistic regression model to adjust for confounding. Statistical significance of variables in the final model was assessed based on a p-value threshold of ≤0.05. In analysing the qualitative data, English transcripts were read in between the lines to identify codes and themes. Identification of codes was guided by the conceptual framework and study objectives with room for emergent themes from the data. Coding and analysis were done manually using a cut and paste approach where segments from the transcripts were copied and assigned to the generated codes.
N/A