Prenatal hepatitis B screening and associated factors in a high prevalence district of Lira, northern Uganda: A community based cross sectional study

Background: Chronic hepatitis B Virus (HBV) infection affects 80-100 million people in sub-Saharan Africa and accounts for an estimated 650,000 deaths annually. The prevalence of active hepatitis B virus infection among women aged 15-64 in mid-Northern Uganda is about 5%. Lira district is among the high prevalence areas where government embarked on mass HBV screening since 2015 as a gateway for access to prevention, treatment services, and an effective response to the hepatitis B epidemic. The current proportion of pregnant women screened and the factors associated with prenatal HBVscreening in Lira are not known despite the fact that women contribute largely to both vertical and horizontal transmission of HBV. This study aimed at determining the proportion of pregnant women screened for HBV and factors associated with prenatal HBV screening in Lira district. Methods: This was a community based cross sectional study conducted among 423 pregnant women in the sub counties of Aromo and Agweng in Lira district. Data were collected using open data kit and analysed using STATA version 14. The outcome variable was prenatal HBV screening while predictor variables were community, individual and health facility factors associated with HBV screening. Multivariable logistic regression was used to determine factors associated with prenatal HBV screening. Results: Thirty five women (8.3%) had been screened for HBV during the current pregnancy. Factors associated with prenatal HBV screening in Lira included perceived risk (Adjusted Odds Ratio (AOR) 3.78, 95% CI 1.01-6.14), respondent’s age (AOR = 3.98, 95% CI 1.39-5.09), husband/partner’s education (AOR = 3.34, 95% CI 1.10-5.12) and past failure to access to HBV screening services at government health facilities (AOR = 6.44, 95% CI 2.10-8.02). Conclusion: The level of HBV screening among pregnant women in Lira was low and is mainly associated with perceived risk, age, access to HBV screening services and spousal education level. More effort is needed in creating mass awareness on the need and importance of HBV screening most especially among pregnant women. We carried out the study in Agweng and Aromo sub counties, Lira district. These sub-counties have the highest rates of maternal and child mortality in Lira district, estimated at over 400 per 100,000 live births and 85 per 1000 live births respectively [12]. Lira District is located in Lango sub-region in Northern Uganda and is bordered by the districts of Pader and Otuke in the North and North East, Alebtong in the East, Dokolo in the South and Apac in the West. It is 375kms from Kampala via Karuma- Kamdini. The total population in Lira district is about 403,100. Most of the inhabitants are subsistence peasant farmers. This was a community-based cross sectional study nested in a cluster-randomized trial that considered the same study population (pregnant women). The cluster randomized trial was on the effectiveness of an integrated intervention consisting of pregnancy buddies, mobile phone messages, and mama kits in increasing facility-based births. Sample size estimation was done using the Kish Leslie formula for cross sectional studies [13] and a 10% adjustment for non-response was made to come up with 423 respondents. Where N = sample size estimate of pregnant women. P = assumed true population prevalence of Hepatitis B screening services (50%). Zα = Standard normal deviate at 95% confidence interval corresponding to 1.96. δ = Absolute error between the estimated and true population prevalence of Hepatitis B screening, (5%) at 95% CI. Consecutive sampling method was used whereby every woman known to be pregnant within every village in the two sub-counties was approached and those who were eligible and consented to participate in the study were included until a sufficient sample size was accrued. Though not a probability sampling method, it allows one to select all the accessible population in an area during the study period. This method is recommended for RCTs including the one in which our study was nested. The outcome variable was self reported prenatal hepatitis B screening. A pregnant woman was included if she self-reported to have been screened for Hepatitis B since conception of the current pregnancy. Independent variables were community, individual and health facility factors that affected prenatal Hepatitis B screening. Community factors included cultural beliefs and practices, stigma, community mobilisation and sensitisation. Individual factors included; formal education level, gender, age, marital status, attitude towards the services, knowledge and awareness about hepatitis B infection, HBV transmission, HBV screening perceived risk and complications. Health facility factors included; health worker attitude, availability of skilled health workers, convenience of obtaining care, cost, and distance. A pregnant woman was eligible for the study if she was in the last trimester of her pregnancy (i.e., based on self-reported information using dates for the last normal menstrual period) and was a resident in one of the two sub counties. Pregnant women with psychiatric ailments that prevented them from providing an informed consent were excluded. We employed quantitative data collection methods. Data were collected electronically using interviewer administered structured, standardized, pre-coded and pre-tested questionnaires. The questionnaires were prepared in English and translated to Lango and then back-translated to English to ensure consistency of the tool. Data were uploaded into ODK software on android mobile phone devices that were configured to have an instant check for validity and could not allow certain types of erroneous responses to be entered. Range and consistency checks were also incorporated in the data collection system to ensure completeness. The original study recruited pregnancy monitors in every village of the study area. These were elected by the community in a public meeting. Their role was to identify all pregnant women in the area, and inform the study team. In order to ensure that all pregnant women had been enrolled, the study also employed village health team leaders to conduct a census of all pregnant women in the study area. At the end of each field day, data would be uploaded into a secure database that was encrypted and password protected to preserve participant confidentiality. Research assistants that fluently spoke Lango were recruited from the study area and trained on electronic data collection. Data were cleaned using MS EXCEL and exported to STATA version 14 for analysis. Frequencies and percentages were obtained for all categorical variables and means (Standard Deviation) and medians (Inter Quartile Range) were generated for continuous variables. The number of pregnant women that had been screened for HBV during the current pregnancy was expressed as a proportion of the total sample size. Bivariable logistic regression was used to examine the crude association between prenatal HBV screening and the predictor variables. Variables that had p < 0.25 [14] at bivariable analysis, those backed by literature were considered in a stepwise and logical way for the multivariable models aimed at determining independent factors associated with the outcome. The most parsimonious model was tested for goodness of fit using the Archer-Lemeshow goodness-of-fit test (l fit test). Considering the null hypothesis that the model was a good fit, the model was found to have a p > 0.05 using the 95% level of significance, which implied that it was a good fit.