Accessibility of Early Infant Diagnostic Services by Under-5 Years and HIV Exposed Children in Muheza District, North-East Tanzania

Introduction: Early infant diagnosis (EID) of Human Immunodeficiency Virus (HIV) provides an opportunity for follow up of HIV exposed children for early detection of infection and timely access to antiretroviral treatment. We assessed predictors for accessing HIV diagnostic services among under-five children exposed to HIV infection in Muheza district, Tanzania. Methods: A cross sectional facility-based study among mother/guardian-child pairs of HIV exposed children was conducted from June 2015 to June 2016. Using a structured questionnaire, we collected information on HIV status, socio-demographic characteristics and other relevant data. Multiple regression analyses were used to investigate associations of potential predictors of accessing EID services. Results: A total of 576 children with their respective mothers/guardians were recruited. Of the 576 mothers/guardians, 549 (95.3%) were the biological mothers with a median age of 34 years (inter-quartile range: 30–38 years). The median age of the 576 children was 15 months (inter- quartile range: 8.5–38.0 months). A total of 251 (43.6%) children were born to mothers with unknown HIV status at conception. Only 329 (57.1%) children accessed EID between 4 and 6 weeks of age. Children born to mothers with unknown HIV status at conception (AOR = 0.6, 95% CI 0.4–0.8) and those with ages 13–59 months (AOR = 0.4, 95% CI 0.2–0.6) were the significant predictors of missed opportunity to access EID. Children living with the head of household with at least a high education level had higher chances of accessing EID (AOR = 1.8, 95% CI 1.1–3.3). Their chances of accessing EID services was three-fold higher among mothers/guardians with good knowledge of HIV infection prevention of mother to child transmission (AOR = 3.2, 95% CI 2.0–5.2) than those with poor knowledge. Mothers/guardians living in rural areas had poorer knowledge of HIV infection prevention of mother to child transmission (AOR = 0.6, 95% CI 0.4–0.9) than those living in urban areas. Conclusion: Accessibility of EID services among children below 5 years exposed to HIV infection in Muheza is low. These findings stress the need for continued HIV education and outreach services, particularly in rural areas in order to improve maternal and child health. The study was conducted in Muheza district north-eastern Tanzania (4°, 45′S; 39°00′E). The district has a total of 46 health facilities (one district hospital, 4 health centers and 41 dispensaries); 42 of which provide RCH services, 36 offer PMTCT services and 28 offer EID services (33). The prevalence of HIV in the district as reported in 2013 was 3.9% (34). The HIV prevalence among pregnant women in Muheza district in 2015 was 4.0% higher than that of its neighboring districts as detailed in the Supplementary Material (35). Muheza district was selected for this study because it is among the leading districts in Tanga Region with high HIV prevalence among pregnant women (35). The study was conducted in 18 health facilities (one district hospital, three health centers and 14 dispensaries). This was a cross-sectional study using a multi-stage sampling approach conducted during the period June 2015 to June 2016. In the first step, we aimed to select a total of 18 out of 46 health facilities in the district as detailed in the Supplementary Material. The health facilities defined the primary sampling units (PSUs)/clusters. A list of 46 PSUs was obtained from Muheza District Council and the facilities were numbered from one to 46 according to their geographical location. Then the sampling interval was obtained and 18 health facilities were randomly selected from the list. During the selection process, health facilities that were selected from the list and were found to have no EID services were removed and the next on the list was included. The process was repeated until all 18 facilities that were required for the study were included. The study population included all selected mothers/guardians with children below 5 years born to HIV positive mothers who were not breast feeding for ≥6 weeks. The sample size was calculated based on the formula that accounted for simple random sampling and the design effect which accounted for between and within cluster variation (36). We assumed the highest exposure of infection risk and a response rate of 90% and thus the estimated sample size was 836 children. Four nurse counselors were trained on how to conduct interviews using a pre-tested questionnaire. Most of the questions were closed with an open option, and some were open-ended. Information on HIV status, socio-demographic characteristics and other relevant family and health status were collected. Socio-demographic information included age, sex, religion, residence, income, education, marital status, occupation, and size of household. Further information collected included knowledge on HIV transmission and prevention, time period when the child’s first HIV test ought to be performed, distance to the nearest health facility (which was defined by time taken in minutes to reach the facility on foot) and maternal information which included whether the index pregnancy was planned or not, as well as HIV status prior to conception. Mothers/guardians’ satisfaction on services at health facility was assessed based on five key questions. The questions were mainly addressing the mother/guardian’s perceptions regarding health care services such as HIV testing, treatment, advice, nutritional and social support received at the health facility. Each question had five options of pre-coded responses with a neutral point being neither satisfied nor dissatisfied, or neither poor nor good. Moreover, the general quality of all health services was assessed on how low or high they were satisfied with the overall quality of health care services provided at a particular health facility. This included services as a whole, reception by health workers, comfortability with the environment of care, freedom to interact with health personnel (asking questions for clarification), family planning, adequate medicines and diagnostic services. Mother/guardian’s willingness to receive the HIV testing and diagnosis for their children was also assessed. Mother/guardian’s knowledge and beliefs on HIV prevention and transmission including MTCT of HIV was also assessed based on four key questions. These questions were mainly addressing general HIV transmission, timing of post exposure prophylaxis to HIV exposed infant, factors affecting HIV transmission and prevention on the risk of MTCT. Mothers/guardians whose children received an HIV test late (i.e., testing at ≥7 weeks of age) were asked to state the reasons for the delay. Multiple responses regarding the factors associated with missed opportunity to receive the first HIV test for their children at the age between 4 and 6 weeks were assessed. During recruitment, more data of the child and her mother were extracted from registers, child card and hospital case files to supplement the collected primary data. The HIV DNA PCR testing was used to confirm HIV infection in all HIV exposed children less than18 months of age according to national guidelines. DBS samples were routinely collected from HIV exposed children at the first encounter to the health facility. For infants, using a heel prick, five circles were filled with blood on a specific Whatman filter paper card and air dried for ≥4 h on a drying rack. After drying, the cards were placed in a gas-impermeable zip locked bag with desiccant sachets and stored in a safe location. All DBS samples from peripheral health centers and dispensaries were transported to the district hospital and then were mailed to the KCMC Clinical Laboratory in Moshi, Tanzania. In the laboratory, the DBS were tested for HIV DNA using version two of the COBAS AmpliPrep/COBAS TaqMan 48 system. One DBS circle was used to run a DNA-PCR test and if positive, a second circle was analyzed to confirm the first result. A test result was considered positive if both PCR assays were positive. The Rapid HIV antibody test was done in all HIV exposed children aged more than 18 months using Determine®/Bioline® rapid tests for the first test. If the first test was positive, the second test was performed using Uni-Gold® rapid test. HIV diagnosis was confirmed based on concordance of the results of these two HIV rapid tests. Furthermore, the sample turnaround time (TAT) which was defined as time between DBS sample collection and return of HIV DNA PCR results to the facility, for the first HIV DNA PCR test was assessed. To facilitate data entry, responses provided in open-ended questions were re-coded into themes which were developed to respond to study objectives. Data were double entered in Epi Data database version 3.1 (http://www.epidata.dk/) and then transferred to STATA version 13 statistical package (Stata Corp, College Station, Texas, USA). Continuous variables were described using median and inter quartile range (IQR). Categorical variables were described using frequencies and percentages. The accessibility to first HIV test by the child was categorized into a binary variable (dependent variable), “yes” as obtained or “no” as not obtained the test at the age of 4–6 weeks. A composite variable on knowledge and beliefs on HIV prevention and MTCT of the mother/guardian was generated based on the four key questions as described above. The knowledge level was categorized into a binary variable “good” or “poor.” Guardians’ perceptions to healthcare services based on five options of pre-coded responses with a neutral point being neither satisfied nor dissatisfied, or neither poor nor good were described by using percentages. Binary logistic regression was done and all factors with p-values of ≤ 0.2 including priori factors were considered for multiple variable logistic regression analysis. Multiple logistic regression analyses were used to examine the associations between various factors (including child and maternal, guardian, head of household characteristics) and service accessibility by the child at the age of 4–6 weeks. A manual backward stepwise selection was employed by removing non-significant variables (one at a time starting with those with the highest p-value). Goodness of fit of the final model was tested using likelihood ratio test. The final model consisted of variables that were significant at p-value of ≤ 0.05 including those with epidemiological importance. Adjusted odds ratios with their corresponding 95% confidence intervals were estimated and presented. Three individual multiple regression models were fitted for (1) the child and maternal variables, (2) guardian variables, and (3) the head of household variables. Subsequently, two combined models were fitted: (1) a model that contains the three set variables; and (2), a model that includes only significant variables which was used in the final interpretation. In addition, a separate model to assess the mother/guardian’s knowledge of HIV (including MTCT and PMTCT) was also fitted. In this study, “accessibility of first HIV test by the child at the age of 4–6 weeks” was defined as the time when the child was obtaining HIV test at the age of 4–6 weeks. Ethical approval was obtained from the Medical Research Coordinating Committee of the National Institute for Medical Research in Tanzania (Ref: NIMR/HQ/R.8a/Vol. IX/1978) and University of Zambia Biomedical Research Ethics Committee (Ref: 001-01-15). Permission to conduct this study was given by Muheza District Council Authority. Written consent was obtained from each guardian before recruitment. Counseling of the mothers/guardians of HIV-exposed children before and after the HIV test was performed in accordance with standard testing and counseling guidelines. In this study the HIV status for all HIV exposed children was determined for effective appropriate care and management. All confirmed HIV positive children were referred to CTC for initiation of ART in accordance with Tanzania guidelines on the management of HIV and AIDS.