Determinants of antenatal and delivery care utilization in Tigray region, Ethiopia: A cross-sectional study

Introduction. Despite the international emphasis in the last few years on the need to address the unmet health needs of pregnant women and children, progress in reducing maternal mortality has been slow. This is particularly worrying in sub-Saharan Africa where over 162,000 women still die each year during pregnancy and childbirth, most of them because of the lack of access to skilled delivery attendance and emergency care. With a maternal mortality ratio of 673 per 100,000 live births and 19,000 maternal deaths annually, Ethiopia is a major contributor to the worldwide death toll of mothers. While some studies have looked at different risk factors for antenatal care (ANC) and delivery service utilisation in the country, information coming from community-based studies related to the Health Extension Programme (HEP) in rural areas is limited. This study aims to determine the prevalence of maternal health care utilisation and explore its determinants among rural women aged 15-49 years in Tigray, Ethiopia. Methods. The study was a community-based cross-sectional survey using a structured questionnaire. A cluster sampling technique was used to select women who had given birth at least once in the five years prior to the survey period. Univariable and multivariable logistic regression analyses were carried out to elicit the impact of each factor on ANC and institutional delivery service utilisation. Results: The response rate was 99% (n=1113). The mean age of the participants was 30.4 years. The proportion of women who received ANC for their recent births was 54%; only 46 (4.1%) of women gave birth at a health facility. Factors associated with ANC utilisation were marital status, education, proximity of health facility to the village, and husband’s occupation, while use of institutional delivery was mainly associated with parity, education, having received ANC advice, a history of difficult/prolonged labour, and husbands’ occupation. Conclusions: A relatively acceptable utilisation of ANC services but extremely low institutional delivery was observed. Classical socio-demographic factors were associated with both ANC and institutional delivery attendance. ANC advice can contribute to increase institutional delivery use. Different aspects of HEP need to be strengthened to improve maternal health in Tigray. © 2013 Tsegay et al.; licensee BioMed Central Ltd. Tigray regional state is located in the northern part of the country and has an estimated total population of 4.3 million of which 50.8% are females. Eighty percent of the population are estimated to live in rural areas and the majority of the inhabitants are Christian [27]. The region is divided into seven zones and 47 weredas (districts), of which 35 are rural and 12 are urban. There is one specialised referral hospital as well as five zonal hospitals, seven district hospitals, 208 health centres and more than 600 health posts in the region. Coverage estimations from the Tigray Health Bureau indicate 75% are for ANC, 20% for skilled delivery, 13% for clean and safe deliveries (those attended by HEWs) and 90% for contraceptive use [28]. The study district of Samre-Saharti is located in the northern part of the state of Tigray, 55 km from the capital, Mekelle. The district has 23 tabias (sub-districts) and each tabia has a health post with two HEWs. There is one health centre (HC) in the district’s town which functions as a referral center to the four HCs stationed in the rural areas. Samre-Saharti has an estimated population of 124,499 of which 50.2% are female [26]. Women of reproductive age (15–49 years) constitute approximately 14,375 (23%) of the population and the number of deliveries in 2007 was estimated to be 646 [27]. The study design was a community-based cross-sectional survey. Of the 23 tabias in the district, four were difficult to reach due to floods and were therefore excluded, so 19 were included in the study. A lottery was drawn among the 19 tabias to sort their respective villages (kushets) randomly and then the cluster sampling technique was used to select the study population. The sample size for the study was determined using the single population proportion formula. Assuming the proportion of institutional delivery to be 6%, a 95% level of confidence, a 2% marginal error and a design effect of 2, and the required sample size was 1,115 women. The 30 clusters were chosen by population proportional to size sampling. Households were chosen after a random start at a central place in the village. A pen was spun and the data collectors walked to the edge of the village in the direction that the pen pointed, numbering all households along the way. A random number was chosen to identify one of these households as the starting household for the cluster and collection continued on the right-hand side of this starting house until the required number of individuals had been recruited for the sample. If neither household members nor a woman as per the selection criteria, were present at the time of the survey visit, the next closest household was chosen. A total of 1,115 households from the 19 tabias and 30 selected clusters were visited from August to September 2009. The units of analysis for this study were women (aged 15–49 years) who had given at least one live birth during the five years before the survey. A structured questionnaire was prepared in English, based on an existing tool and translated into the local language (Tigrigna) prior to the start of the fieldwork (Additional file 1). To ensure that the questions were clear and could be understood by both the enumerators and the respondents, the questionnaire was pretested and further refined based on the results. The questionnaire collected information on socio-demographic and obstetric characteristics, use of ANC and place of delivery. Fifteen HEW enumerators who were fluent in the local language and four supervisors with experience in maternal health service provision were selected for data collection. The supervisors were assigned to supervise the data collection process and perform quality checks. Three days of training were given to both the data collectors and the supervisors and were managed by the investigator. The training focused on the quality of the field operation (how to fill in the questionnaire, mock interviews and other practical exercises). Permission to carry out the study was obtained from the Tigray Regional Health Bureau and the Samre-Saharti District Health Office. Each respondent gave informed verbal consent after being told the purpose and procedures of the study. All responses were kept confidential and anonymous. Two response variables were created from questions included in the study questionnaire on ANC and place of delivery. ANC use was defined as whether the mother paid at least one visit to the health post during her pregnancy. Place of delivery was classified as home delivery or institutional delivery, the latter including births that took place at a health centre or at a health post. Despite its importance, PNC was not included in this study since this strategy is under developed in Tigray region. In order to study the influence of explanatory variables on the utilisation of ANC and place of delivery, several predictor variables were selected based on (national and international) literature, national guidelines, field observations and common local practices. The independent variables were categorised as follows. The ages of mothers were grouped as 16–29 years, 30–39 years and 40–49 years. Marital status was classified as married for those who were currently living with their partners, single for those who had never been married, divorced for those currently separated, and widowed for those who had lost their husbands. Respondents’ education was classified as illiterate, grade 1–4, grade 5–8 and grade 9–12 and above. Husbands´ occupation was classified as farmers, and other occupations (which included pay-in-cash jobs such as daily labourers, merchants and governmental employees). Proximity of residence to health facility was defined as the availability or not of a health facility in the village. Parity was grouped as 1–4, 5–7 and 8–11 children. History of obstructed and prolonged labour was defined as whether the mother had reported experience of difficult labour in a previous pregnancy. A question about receiving pregnancy advice or not during ANC visits was also included in the questionnaire. Data were collected, compiled and reviewed by the supervisors and then entered into Epi Info software, coded, cleaned, and finally imported into STATA version 10 software for analysis. Univariable logistic regression was carried out between the selected predictor variables and the outcomes (ANC and institutional delivery service utilisation). Those variables which were significant (i.e. with a p value<0.05) in the univariable logistic regression were selected and retained in the multivariable logistic model. Marital status, education, parity, health facility availability in kushet and husband’s occupation were included for the first outcome variable ANC. In addition to the mentioned variables, ANC attendance, ANC advice and difficult/prolonged labour were included in the adjusted models for the delivery care outcome variable. Both variables ‘Attended ANC’ and ‘ANC advice’ were included in the analysis despite potential collinearity (although correlation coefficient of 0.56 was observed in the analysis) because not all women attending ANC received advice and we wanted to capture this dimension. Odds ratios and their 95% confidence intervals (CIs) were calculated.