Background In Ethiopia, the uptake of antenatal care services has been low. Moreover, there is less frequent and late attendance of antenatal care among women who attend. Using the Anderson-Newman model of health care utilization, this study identified factors that either facilitate or impede antenatal care utilization in Kersa district, Eastern Ethiopia. Method A community-based cross-sectional study was conducted. A total of 1294 eligible women participated in the study. Data were collected using face to face interviews with a pre-tested structured questionnaire administered with a digital survey tool. Data were collected in a house to house survey of eligible women in the community. Bivariate and multivariate logistic regression analyses were used to examine the predisposing, enabling and need factors associated with antenatal care utilization. Result Out of the 1294 respondents, 53.6% (95% CI: 50.8%, 56.3%) attended antenatal care at least once during their last pregnancy. Only 15.3% attended four or more antenatal care visits and just 32.6% attended prior to the 12th week of gestation. Educational status, previous use of antenatal care and best friend’s use of maternal care were significant predisposing factors associated with at least one antenatal care visit. Type of kebele, wealth index and husband’s attitude towards antenatal care were significant enabling factors associated with at least one antenatal care consultation. Health Extension Workers providing home visits, perceived importance of ANC and awareness of pregnancy complications were significant need factor associated with at least one antenatal care consultation. Husband’s attitude towards ANC, head of the household, awareness of pregnancy complications, and history of abortion were predictors of attending four or more antenatal care visits. Conclusion More than half of the women attended at least one antenatal care visit. A sizable proportion of women had infrequent and delayed antenatal care. Intervention efforts to improve antenatal care utilization should involve the following: improving women’s educational achievement, peer education programs to mobilize and support women, programs to change husbands’ attitudes, ameliorate the quality of antenatal care, increasing the Health Extension Worker’s home visits program, and increasing the awareness of pregnancy complications.
A community-based cross-sectional study was conducted in Kersa district, Eastern Hararghe zone of Oromia regional state, in Eastern Ethiopia from June to August 2017. The district capital is Kersa town, which is located 486 kilometres from the capital city, Addis Ababa. According to the population projection for Ethiopia published by the Central Statistical Agency in 2014, the district has an estimated total population of 205,628. The population is predominantly rural (92%). The district has 38 kebeles of which three are urban and 35 are rural [29, 30]. The kebele is the lowest administrative unit in Ethiopia consisting of around 1000 households, or an approximate population of 3000 to 5000 [31]. The district has 24 Health and Demographic Surveillance System (HDSS) kebeles and 14 non-HDSS kebeles. HDSS is a platform to regularly follow the health and demographic characteristics of a community residing in a distinct geographic area. It monitors new health threats, tracks the change in population number through fertility and migration rates, and measures the effect of interventions on communities [32, 33]. All kebeles have access to non-asphalt roads, though the terrain is mountainous in the majority of the rural kebeles. The district has seven health centres, thirty four health posts and eight private pharmacies at different locations within the district. In each kebele, there are two Health Extension Workers (HEWs) providing health promotion activities. According to the information from the district health office, the health coverage (physical accessibility of health facilities) of the district is more than 80% [34]. The study population was all reproductive-aged women in the Kersa district who gave birth in the three years prior to the survey, regardless of the birth outcome. Women who had lived in the district for more than six months, and delivered their most recent baby after 28 weeks of gestation were included. Women who did not volunteer to participate in the study, were critically ill and physically or mentally disabled during the data collection period, were excluded from the study. The sample size (n = 1320) was primarily determined for a broad study on the level of maternal health service uptake and associated factors in Kersa district, Eastern Ethiopia. A total of ten kebeles (seven from areas where HDSS is conducted and three from non-HDSS areas) were included in the survey. Households with eligible women were identified using the HEW health management information system registration log sheet and the number of eligible women in each included kebele was determined. The total sample size of the study was proportionally allocated to each kebele. Respondents were selected through systematic random sampling techniques and invited to take part in the survey. When two or more eligible women were found within the selected household, one was selected by the lottery method and invited to do the interview. At least one ANC attendance: Women who have attended at least one ANC check-up during their most recent pregnancy as reported by the participant. Four or more ANC attendance: Women who attended four or more ANC visits during their last pregnancy as reported by the participant. The predictor variables were conceptualized based on the ANBM of health care utilization and grouped into three set of factors: predisposing, enabling and need factors as shown in Table 1. To obtain the data, we used house-to-house interviews with eligible reproductive-aged women using a structured questionnaire. The study tool for the survey was adapted from pertinent literature. The study tool was first prepared in English and subsequently translated into the local language (Oromiffa) to collect the data. The tool was re-translated back to English to check for consistency. A template of the study tool was prepared using an online survey tool (Survey Gizmo) and downloaded onto iPads for offline data collection. The data was collected by resident HDSS data collectors who have extensive experience in conducting interviews in both urban and rural kebeles using the iPads. The principal author (GT) and a supervisor closely monitored the overall data collection process. The study tool was pre-tested on 65 women living in a neighbouring district. All required revisions were made to the study tool based on the pre-test. Experienced HDSS data collectors and a supervisor were recruited and deployed for the data collection. A two-day intensive training course was provided to the data collectors and the supervisor about the aim of the study and sampling procedures; data collectors also performed simulated exercises on how to interview respondents. To ensure correct inclusion of the participants, the student researcher made the random selection of the kebeles, was responsible for the proportional allocation of the samples, and carried out the random selection of the interviewees. We used iPads for data collection to avoid missing or incomplete responses. The supervisor cross-checked the completed responses on the iPads by repeating the interviews with 10% of the respondents to check for correct completion of valid responses. The responses were uploaded into the online survey tool on a daily basis and the lead author double-checked for any inconsistencies and gave feedback to the interviewers on daily basis. The data were directly exported from the digital survey tool into SPSS software version 22 for analysis. Before commencing data analysis, appropriate transformations were made on the variables and missing values were also managed as necessary. Descriptive statistics and appropriate measure of central tendencies were used to summarize the key variables. Variables with missing data due to the skipping nature of the question (husband education and living in a model family) were managed by re-coding the system missing value into an existing relevant category of the variable. Categories of some variables with negligible frequency, such as, “don’t know” or categories not relevant for a particular inferential test due to small cell value were recoded into another related category. Before fitting the full multivariate model, all of the variables were considered for the multicollinearity diagnostics and all showed no multi-collinearity with a variance inflation factor of less than five. Bivariate logistic regression analysis was conducted to examine the association between the predictor and outcome variables using the Crude Odds Ratio (COR) at a 95% confidence interval (CI). Factors that were significant with a p-value of less than 0.05 were retained for further consideration in three blocks of the multivariate logistic regression model. The three block models were built in such a way that, model 1 contained only the predisposing factors, model 2 included predisposing and enabling factors, and model 3 (the final full model) considered all three factors simultaneously (S1 and S2 Tables). In the final multivariate model, factors with a p-value of less than 0.05 were declared statistically significant [35]. Two separate multivariate models were fitted for primiparous and multiparous women for the “at least one ANC” outcome variable. The factors that showed statistically significant association in the multivariate logistic regression analysis were mapped into the three domains: predisposing, enabling and need factors to streamline the analysis using the ANBM. The study was conducted after securing ethical approval from the Institutional Health Research Ethics Committee (IHREC) of the College of Health and Medical Sciences, Haramaya University, Ethiopia with approval number (IHRERC/129/2017) and the Human Research Ethics Committee (HREC) of the University of Newcastle, Australia with approval number (H-2016-0403). Informed verbal consent was obtained from each respondent before commencing interviews. The informed verbal consent procedure had been approved by both Ethics Committees. Participants read or listened to the Information Statement provided by the interviewer and it was expected that they understood its contents. Then, if they decided to participate, they informed the interviewer that they were happy to take part in the research, and this was taken as participant’s informed verbal consent. For respondents aged 15–18 years, we obtained informed consent from themselves as they were married, had had at least one birth, assumed social responsibility and are considered to be mature minors. The Ethics Committees had also approved the minor consent procedure for women between 15 and 18 years. The confidentiality of the respondents was ensured by avoiding personal identification details in the study tool.