Introduction

Available evidence shows that countries in sub-Saharan African (SSA), Nigeria, inclusive are lagging in Antenatal care (ANC) coverage. A WHO survey on the coverage of maternity care showed that while almost all pregnant women in developed countries attend ANC; coverage remains low in SSA [1] due to several factors which include poor health-seeking behaviour, social and cultural barriers, lack of access to services and poverty [2]. Globally, notable progress has been made in ANC coverage rate, which reflected in a 70% increase between 1990 and 2013 [3-5]. However, SSA has recorded slow progress, and for Nigeria, ANC utilization rate has consistently remained below the global average [6]. In 2013, the global ANC utilization rate (at least one visit) was 81%, and for SSA, it was 75%, while the rate for Nigeria ranged between 61% and 66% [7-8]. Nigeria's poor performance in terms of the number of women that utilized ANC is more glaring when we compare Nigeria's rate with neighboring African countries: Ghana [78.2%], Sierra Leone [76.0%], Liberia [78.1%] and Lesotho [70.4%] [9].

It is, therefore, no source of surprise that SSA accounts for a disproportionate burden of maternal mortality. Available evidence shows that about 585 000 women die yearly from pregnancy-related complications, and out of this, SSA accounts for 66% [10], Nigeria constitutes about 2% of global population yet accounts for over 14% of global maternal mortality burden [8, 11-12]. It is believed that high maternal mortality in SSA is due to the underutilization of maternal care services [13]. In SSA, the deplorable state of maternal health outcomes is because a good number of women do not use modern healthcare services. Consistent medical checkup has been recommended as an intervention strategy to reduce the risk of both maternal and neonatal mortality [14]. NDHS [7] reported that among the women surveyed who delivered within the last five years of the survey, 17% did not have access to ANC, 30-37% did not undergo any medical tests or examination, 27% were not administered tetanus toxoid (TT) injection and 40% were never informed of the danger signs of complications [15].

From the available evidence, maternal death is higher among women who do not attend ANC and those who deliver without skilled birth attendants (SBAs) [16-17]. Nigeria's women's attitude towards ANC has been reported to be unimpressive. Available evidence shows that while several Nigerian women do not attend ANC, those who attend do not make early visits and do not make up to a minimum of four ANC visits, which is the minimum recommended number. A good number of women who utilize ANC still do not deliver their babies in health facilities. The WHO [18] recommends ready access by women to evidence-based maternal and perinatal healthcare (antenatal, Intrapartum and postnatal care) as an essential intervention for improving maternal and perinatal outcomes in low resource settings. However, many women in Nigeria do not keep to this recommendation.

The benefit of ANC is well documented in the literature. Pregnant women are advised to attend ANC to track the progress of their pregnancy and put in place necessary measures to avert complications in case of potential risks. Evidence has shown that women who start ANC within the first three months of their pregnancies more often record better pregnancy outcomes than women who either start late or refrain from using care [19, 20]. ANC guarantees optimal health for both mothers and newborn babies. Appropriate care during pregnancy is essential for the health of the mother and the development of the unborn baby, promotion of healthy behaviours and parenting skills, and establishes a link between the women and the health system [6, 21]. While utilization of ANC is beneficial, also important is the gestational age of the pregnancy when a woman begins to attend ANC and the number of times she attends. According to Arthur [21], to maximize the benefits from ANC, a woman must begin the service in the first three months of pregnancy and maintain a minimum of four visits. WHO [22] recommends a minimum of four visits for pregnant women before delivery if the woman cannot make the recommended number of visits by physicians. Villar et al. [23] revealed that a maximum of four visits suffice for uncomplicated pregnancies, and more visits may be recommended for complicated pregnancies. Concerning early timing of ANC visits, the WHO [22] has recommended that pregnant women should attend ANC in the first trimester of pregnancy, but the most recent survey revealed that in Nigeria only 14.6% of rural women began ANC within the first four months of pregnancy and for urban women 23.1% started at the same time [7].

A significant disparity exists in maternal care utilization between rural and urban women; poor and the rich and women living in Southern and Northern geopolitical zones. Several health facilities are located in the urban part of the country where rural women may not be able to get to due to transportation problems and poor road networks [24]. Available evidence shows that while 86% of urban women utilized ANC from skilled ANC providers (doctor, nurse/midwife, or auxiliary nurse), only 46.5% of rural women received ANC from such sources [12]. Of the women surveyed in the last NDHS [7], while 74.5% of urban women made at least four ANC visits, it is only 38.2% of rural women that made at least four ANC visits [7]

PHC has been recommended as the health system strategy to improve access of women to SBAs and ANC, particularly in low-resource settings such as Nigeria [17].In 2006, the Nigerian Presidential Committee in a report titled, "Accelerating the attainment of MDG -5" reported that poverty and the attendant's inability to pay for healthcare services were the major barriers hindering women from accessing safe maternal care [25]. To reduce financial barriers to maternal and childcare services, the Federal Ministry of Health (FMoH) in 2007 recommended that all states (the second tier of government) should implement the policy of free maternal and child care services. By December 2009, the National Primary Healthcare Development Agency (NPHCDA) initiated the Midwife Services Scheme (MSS), a programme designed to redress supply bottlenecks in the provision of maternal healthcare services in the rural part of Nigeria. By the end of 2010, over 50% of Nigerian States have implemented the policy of free care services to mothers and newborn babies [12, 17]. In 2007 the Nigerian government launched the Integrated Maternal and Child Health (IMCH) policy, which was designed to provide a continuum of care to women from the preconception period to the immediate postnatal period [17]. IMCH aims at addressing barriers to maternal care utilization to the vulnerable section of society. To increase access to maternal care in the rural part of the country, the FMoH adopted PHC as the implementation strategy of the policy. Despite this, evidence has shown the underutilization of PHCs by rural Nigerian women [26], and consequently, a yawning disparity in maternal health outcomes exists between both regions. Why do rural women still utilize ANC far less than their urban counterparts despite the innovative approach of PHCs, designed to bring care to the doorstep of rural women?

Predictors of ANC utilization have been the subject of several studies in Nigeria, though a good number of the studies did not disaggregate ANC into its primary health component. There remain gaps in the study considering the dearth of research studies on predictors of ANC in Primary Healthcare Centers (PHCs). In Nigeria, there are only a few studies that focused on maternal care utilization in PHCs. For instance, in a descriptive study, Ejembi et al. [26] investigated the determinants of ANC and family planning utilization in PHCs.However, the study was a mere descriptive study. Also, Egbewale & Odu [27] investigated people's perception of PHCs and also determinants of utilization of healthcare services in PHCs but did not focus on maternal care utilization. Okonofua et al. [12] utilized both descriptive and analytical techniques to investigate the predictors of both ANC and SBAs in twenty communities in Edo State. However, this study did not consider the early timing of ANC visits and a minimum of four ANC visits; this constitutes the gap that this present study filled. This is the first study that assessed predictors of ANC in PHCs in Ughelli North LGA, Delta State Southern Nigeria. The motivation of this study originated in this context, and it is poised to fill the gaps in both normative and empirical literature. It is believed that the findings from this study will provide evidence-based information on the socioeconomic predictors of maternal care utilization in PHCs focusing on early timing of ANC visits and a minimum of four ANC visits in eight randomly selected rural communities in Delta State, Southern Nigeria.

Conceptual Framework

This study is built on the conceptual framework of Andersen and Newman [28]. This framework was initially developed in the 1960s and was principally designed to understand the predictors of healthcare utilization in the United States of American (USA). However, the framework has been utilized in research outside the USA, as reported by a systematic review of studies that utilized the model between 1998 and 2011[29]. According to this model, health service utilization is a form of individual characteristics that are influenced by the health system and societal factors. The societal determinants include technology and societal norms such as healthcare financing, and it affects the individual determinants directly and through health system factors [12]. Health system determinants include resources- volume and distribution of labour and capital for health care, organization- patient's access to the healthcare system and structure - what happens after entry into the health system. The individual characteristics that influence healthcare utilization behaviour are categorized into three: predisposing factors, enabling factors and need-based characteristics [6]. Predisposing factors are the ability to secure health services and it includes demographic characteristics such as age, sex, marital status, past illness, social structure (education, race, occupation, family size, ethnicity, religion, and residential mobility) and beliefs (values about health and illness, attitudes towards health services/ providers, knowledge about diseases). Past studies show that individual characteristics are significant predictors of healthcare utilization. For instance, women that are married have been reported to utilize maternal care services better than their unmarried counterparts. The past illness factors show that women experience with pregnancy-related complications, birth order and past experiences with the health system can influence their maternal care behaviour [28, 30].

The enabling factors are the means or resources require to access healthcare services. Enabling factors include family resources (income, wealth, health insurance, and other sources of third-party payment, type of regular source of payment and accessibility to them) and community characteristics (ratio of health personnel to patients, health facility density ratio, rural-urban location). It, therefore, follows that women's access to maternal care services depends in part on the availability of services and access to resources to pay for services.

The need-based characteristics include the perception of the need for health services, whether individual, social or clinically-evaluated perception of need [31]. Individual perception of illness in terms of his or her vulnerability to illness, consequences of illness and benefits from the utilization of services will influence his/her behaviour in utilizing care. For instance, a woman who once suffered obstetric complications and who foresees that possible complications may result in either death or disability may take a measure to protect herself. According to Andersen & Newman [28], these factors represent the most immediate determinants of health service utilization. The need-based component suggests that the utilization of maternal health services can be dictated by a woman's perception of the relative benefits of modern health services versus traditional methods of care. Added to this is a woman's understanding of child-birth related complications and her desire to deliver safely and obtain a healthy baby.

Materials and Methods

Study Communities

This study is a community-based cross-sectional study design that employed quantitative data collection method to investigate the predictors of early timing of ANC visits and a minimum of four ANC visits in PHCs in eight randomly selected rural communities in Delta State: Unukpo, Agbarha-otor, Saniko, Oguname, Uneni, Evwereni, Umusu, and Ekrerhavwe. These eight communities are located in Ughelli North local government council. Ughelli North is one of the LGAs in Delta State. It lies between 90 45 'N and 80 43'E with a landmass of 818 square km.

Sample Size Calculation

The data for the study was obtained from 900 women residing in eight randomly selected communities in Ughelli North L.G.A. The sample size was selected using Cochrane's [32] sample size formula. This method uses the risk the researcher is willing to take which is commonly called the margin of error at a particular alpha level, which in the present study is adopted as 0.05, and 50% assumed maternal care utilization in PHCs, which was the rate used by a previous Nigerian study [12]. A sample size of 900 was selected using a standard formula for sample size as follows: , which is used for large and finite populations [32].

Where:

n = sample size, Z=Standard Score corresponding to 95% confidence interval that is,1.96, d = proportion of sample Error permitted in a given situation(that is, 5%), Ρ = The prevalent rate of maternal care utilization from PHCs (that is,50% ), q = 1-ρ or percentage of failure rate =1-50%. Hence, the sample size can be worked out below:

,

n

A multi-stage sampling technique was used in selecting respondents for the study. Data collection involved three stages. In the first stage, a purposive sampling technique was utilized in selecting four political wards out of the eleven political wards that make up Ughelli North LGA. The selected wards were: Agbarha-otor ward, Evwereni Ward, Orogun ward I, and Agbarho ward II.

In the second stage, a simple random sampling technique was used in selecting two communities from each political ward. We observed that some communities have PHCs while others do not, hence in each political ward, one community with PHC and one without PHC was selected. For Agbarha- otor ward, Agbarha-otor community was selected for having PHC while Saniko selected for having no PHC. For Orogun ward, Ovara Unukpo was selected for having PHC while Umusu selected for having no PHC. For Agbarho ward II, Oguname was selected without PHC, and Ekrerhavwe selected for having PHC. For Evwereni ward, Evwereni community was selected for having PHC while Uneni selected without PHC.

In the third stage, household survey was conducted. Through the aid of lead contact persons, the research team made a tour around the communities, counting the number of houses in major roads and streets, and identifying houses with at least one woman that met the inclusion criteria. The eligibility criteria were the age 15-49 years, ever married, currently pregnant, or have had a birth in the past five years preceding the survey and must be currently residing in the communities. A second tour was made round the communities on separate days. This time it was only houses marked with qualified women that were visited. Before the visitation, community town criers were made to announce early in the morning that women within the reproductive ages should wait at their various homes during a specified period to see the research team. In houses where more than one eligible woman was on the ground at the time of the survey, simple balloting was used to select one out of them

The research plan proposed an equal number of 113 women to be selected from each community because there is no accurate population data to enable proportional sampling. However, some communities have fewer women who met the eligibility criteria, and fewer than 113 were surveyed in these communities. Hence, the number of women surveyed in communities with larger population size was increased to get the required study sample size of 900.

The analysis of data showed that all the communities were fairly covered in the selection of participants. However, Ekrerhavwe has more participants 132 [14.7%], followed by Agbarha-otor 125 [13.9%]; Oguname 120 [13.3%]; Uneni 117 [13.0%], Evwereni 114 [12.7%], Umusu 101 [11.2%]; Ovara Unukpo 104 [11.6%] and Saniko 87 [9.7%] have the least numbers of participants [analysis not presented in Table].

Variables and Measures

Two outcome indicators were utilized in this study. They were early timing in ANC utilization in PHCs and a minimum of four ANC visits in PHCs. Early timing in ANC in PHCs is a binary variable, and it was coded 1 for ANC utilization in PHCs within the first three months of pregnancy and 0 otherwise. A minimum of four ANC visits in PHCs is binary variable coded as 1 for ANC visit of at least four visits in PHCs and 0 otherwise. Drawing from the Andersen and Newman [28] model and past studies [12], thirteen independent variables were utilized for the study. The independent variables laced under the various determinants as espoused by Andersen and Newman [28] are predisposing factors [maternal age, maternal education, religion, marital status, woman's level of autonomy and perception of quality of care in the nearest health center]; enabling factors [income and employment status and time involved in traveling to the nearest health center], need-based characteristics [The number of children ever given birth to, pregnancy-related complications and birth preparedness].

Maternal age is a continuous variable and measures in years of the participants, maternal education is categorical and coded into no formal education, primary education, secondary education, and higher education. Marital status is classified into married, living together, single, separated, and widow. The number of children ever given birth to ascertains the number of children the woman has ever had. It is categorized into 0-2, 3-4 and ? 5. Religion is categorized into Catholic, other Christians, Islam, African traditional religion, and others. Employment status is binary and categorized into working and not working. Monthly income is categorized in to the following groups: < N5, 000, 5,000-9,999; 10,000-14,999; 15,000-99,999. Time engage in walking to the nearest health centers (in minutes) considered three categories of < 30 minutes' walk, between 30 and 59 minutes' walk and at least 60 minutes' walk.

Perception of quality of care render in the nearest health centers assessed respondents' valuation of care rendered in the nearest PHCs. We considered the following options: poor, good and excellent. Being classified as one prepared for birth entailed undertaking one of the following: saving money for delivery, had a transport arrangement, made an arrangement with a blood donor, and identified where the baby was born. History of pregnancy-related complications was categorized into yes for respondents who once suffered pregnancy-related complications and no otherwise. Exposure to media index was generated by asking participants questions on the number of times they watch television (TV) and listening to the radio. Three categories of responses were provided to participants. The response options are everyday as opportunity allows, at least once a week and not all. The option of everyday as opportunity allows is scored 2 marks; at least once a week scored 1 mark, and 0 was attached to the option of not at all. The responses were aggregated to generate five categories of exposure to media index: no exposure [0%], low exposure [11%-30%]; moderate exposure [31%- 50%]; high exposure [51%- 69%] and very high exposure [70%-100%]. We generated index for woman's autonomy by probing in to seven core areas of household decision-making: Ownership of land /house, participation in woman's healthcare decision, participation in household's large purchase decision, participation on household's decision on daily minor purchase, participation on household's decision on choice of food to cook and freedom to visit friends and relatives. Ownership of land or house was included in the study because women in the study area were allowed to own land or house if they wish. The response options provided were respondent alone, the husband, respondent and husband and others. We assigned 1 mark to the response respondent alone, 0.5 marks to the husband alone, and zero to every other response that precludes the respondents. Using principal component analysis, five-category index of autonomy was generated: no autonomy (0%), low autonomy [7%-30%], moderate autonomy [31%-50%], high autonomy [51%- 69%] and very high autonomy [70%-100%]. Women's participation in household decision-making in areas of healthcare, minor daily purchase, major purchase, freedom to visit friends and relatives, and ownership of land/ house have been used by several studies as a proxy for women's autonomy[12].

Data Collection Procedure

The women were interviewed to elicit information from them on the pattern of utilization of ANC in PHCs in the course of last pregnancy using a pre-tested adapted questionnaire. The questionnaire was administered through a face-to-face interview and was fielded in either Pidgin English or English language. Trained field Research Assistants with one-day training assisted in administering the questionnaires. The questionnaire was structured into five sections. Section 1 focused on respondent's background, Section 2 on husband's characteristics and respondent's participation in household decisions, section 3 reproductive history, section 4 on Antenatal care, Intrapartum and postnatal care experiences during last pregnancy and section 5 on reasons for use and no use of PHCs for various maternal care needs. The responses generated from these sections established the foundation upon which the study was situated in its socio-cultural context. Data collection lasted for three months. During the fieldwork, interviewers were closely monitored by field supervisors who performed spot checks and possibly re-interviewed where appropriate.

Data Analysis

The distribution of the respondents' socioeconomic characteristics, together with their attitude towards maternal care (early timing of ANC visits and a minimum of four ANC visits) was expressed as simple percentages and frequencies. Bivariate analysis was conducted to examine whether a significant association existed between the two outcomes indicators and the various socioeconomic factors, using the Pearson's chi-square (X²) test at 5% level of significance. The binary logistic regression was utilized to examine the various socioeconomic factors influencing the two outcome indicators while adjusting for the effect of other predictors. The model has advantages over others [multiple regressions] because it ensures prediction of the probability of choice within ranges (1, 0), easier, and more convenient to compute since it is based on the cumulative ordered logistic probability function.

The variables included in the logistic regression model were those variables that were either significant at 1%, 5%, or 10% level of significance or conceptually important drawing from previous studies on maternal care utilization in Nigeria. The results of the logistic regression were presented as odds ratio (OR) with a 95% confidence interval for the pooled observation and both sets of communities. Some variables that were significant at the bivariate level were not included because of problems of collinearity.

Some of the independent variables were re-coded both for the chi-square test and binomial logistic regression analysis because of few numbers of cases in some categories. Variables that were re-coded include maternal age, religion, maternal education, and marital status. Maternal age was initially a continuous variable, but they were categorized for frequency distribution. Religion was re-coded into three categories: Catholic Christian, other Christian and others. This is owing to a few numbers of persons in the categories of pagan and African Traditional Religion. Marital status was also recoded because of the few numbers of cases in the categories: separated, widowed, and single; they were lumped together as others; hence, marital status was re-categorized into married, living together, and others. Only a few numbers of participants and their spouses have post-secondary education; hence, secondary education and post-secondary education were lumped together as at least secondary education, so education was re-coded into no formal education, primary education, and higher education. To remove women without previous birth experience, the number of children ever given birth to was recoded into 1-2, 3-4, and ? 5.

For the two indicators (a minimum of four ANC visits and early timing of ANC visits), attempts were made to estimate separate models for both sets of communities based on whether the variable "communities" was significant at 5%. The variable "communities " was binary, hence coded 1 for respondents drawn from communities with PHCs and 2 for those from communities without PHCs. In cases where the variable "communities" was significant, it was possible to deduce that different sets of predictors influence the particular outcome, hence separate analyses for that particular indicators were made for both sets of communities. However, where the variable communities were not significant, only an estimated model for the pooled observation was made.

Both the early timing of ANC visits and a minimum of four ANC visits were bounded for women who reported they utilized PHCs as the place of ANC providers. Of the 819 women who reported recent birth in the last five years preceding the survey, 435 of them utilized ANC in PHCs. These were the women use for the analysis of the determinants of early timing of ANC visits and a minimum of four ANC visits in PHCs. The analyses were bounded for users of PHC for ANC because it is only users that can use timely and adequately.

Ethical Consideration

This study is part of a study designed to assess barriers to maternal care utilization in PHCs in the rural part of Delta State, Southern Nigeria. Ethical clearance and Approval to conduct the study was obtained from the University of Benin Ethics Review Committee with protocol number ADM/E22/A/VOL.VII/14689. Permission to undertake the survey was sought from the leaders of the various communities (that is, Ovies) and chiefs of the study area where the survey was conducted. Also, consent was obtained from the heads of individual households where participants were drawn from. Finally, informed consent was obtained from participants. Participants were informed of the purpose of the study and were assured of confidentiality. They were educated that they had the choice to discontinue the study if they had wish without consequences. Finally, participants were made to sign a consent form, which showed that they understood very well what was explained to them and that they also gave their consent to partake in the survey.No names or specific contact information was obtained from participants as they were identified by unique numbers.

Results

Socio-demographic characteristics of the respondents

The demographic characteristics of the women are presented in Table 1 above. For simplicity and easy comprehension, we present summary statistics for the pooled observation and for both sets of communities that are for communities with PHCs and those without PHCs. For the pooled observation, the average maternal age is 34.4 years with a standard deviation (SD) of 7.2 years. Participants drawn from the set of communities with PHCs have higher average maternal age than those from the sets of communities without PHCs [34.9 years vs. 33.9 years]. The majority of the participants belong to the age group 35-39 [29.9 %]. For the set of communities with PHCs, 32.7% [n= 155] belong to the age group 35-39, while for the set of communities without PHCs, 26.8% [n = 114] belong to the same age group. A large proportion of respondents reported primary educational qualification [54.4%]. Other Christians constitute the vast majority [49.1%], with more participants reporting other Christians in the set of communities without PHCs compared to the set of communities with PHCs [50.8% vs. 47.6%]. Approximately 95% of the women were engaged in economic activities outside their homes. Analysis of detailed of employment revealed that they were largely engaged in the informal sector. Approximately 38% [n = 383] earned less than N5, 000. In the set of communities with PHCs, the majority of the women earned less than N5 000 [52.5%], while for the set of communities without PHCs majority earned between N (5,000-9,999) [38.4%].

Analysis of media exposure showed that the majority of women had high media exposure [40.9%]. More women had high media exposure among the sets of communities without PHCs compared to those with PHCs [45.4% vs. 36.8%]. The majority of women are married [65.2%]. Slightly above 30% were in a consensual union (32.7%). More women were in informal union among the set of communities with PHCs than the set of communities without PHCs [42.9% vs. 21.2%]. The majority of the women live in homes where they have to walk for at least 60 minutes to the nearest health center [57.7%]. A higher proportion of women will need to walk for at least 60 minutes among the set of communities without PHCs than those with PHCs [72.5% vs. 54.9%]. While 65.6% of the women rated the quality of care in the nearest health center to be poor; 6.3% rated good, and 28.1% rated excellent. More women rated poor among the set of communities without PHCs than among the set of communities with PHCs [74.8% vs. 57.3%]. Approximately 68% of the women reported no previous pregnancy-related complications. More women reported they once suffered pregnancy-related complications among the set of women without PHCs than among the set of women with PHCs [32.2% vs. 31.4%]. A higher proportion of women did not prepare for delivery [67.8%]. Analysis of woman's level of autonomy shows that the majority of the respondents reported low autonomy (43.6%). More women reported low autonomy among the set of communities with PHCs compared to the set of communities without PHCs [54.5% vs. 31.8%].

Examination of reproductive characteristics revealed that the majority of the respondents reported that they have given birth to at least five children [41.2%]. Approximately, 27.0% [n = 244] of the respondents were currently pregnant. More women reported current pregnancy among the sets of communities with PHCs compared to the set of communities without PHCs [32.2% vs. 21.4%]. Approximately 75% [n = 182] of currently pregnant women utilized ANC and approximately 69% utilized care in PHCs.

As touching delivery care, 819 women had a birth in the last five years preceding the survey. Approximately 77% [n = 627] of women with recent birth went for ANC checkup in the course of pregnancy, and 435 [69.4%] went for ANC checkup in PHCs.

Factors Associated with Early Timing of ANC Visits in PHCs

The Chi-square test, which examined the association between socio-demographic factors and early timing of ANC visits in PHCs is presented in Table 2. The early timing of ANC in PHCs was significantly associated with maternal age, maternal education, religion, exposure to media, time involved in walking to the nearest health centers, perception of the quality of care rendered in the nearest health center, pregnancy-related complications and birth preparedness [ρ < 0.05]. A higher proportion of women within the age bracket 16-19 [50%] compared to those within any other age group made early ANC visits in PHCs. Again, 40.5 % of respondents who reported formal/ primary education made early ANC visits in PHCs compared to 24.3% of those who reported secondary/higher education. An examination of religion revealed that 51.1% of women who reported other religions (traditional African worshippers, pagan) made early ANC visits in PHCs. Majority of the women who reported low media exposure [48.9%] made early ANC visits in PHCs when compared to any other level of media exposure. While 63.1% of those who lived within 30 minutes' walk to the nearest PHCs made early visit in PHCs; 16.0% of those who live within 30-59 minutes' walk made early visit in PHCs and 17.7% of those who lived within at least 60 minutes' walk made early ANC visits in PHC. A higher proportion of women who rated the quality of care in nearest PHCs as excellent made early ANC in PHCs [54.9%] compared to women who rated poor [28.8%] and good [35.5%].