Elsevier

Social Science & Medicine

Volume 86, June 2013, Pages 26-34
Social Science & Medicine

Does antenatal care matter in the use of skilled birth attendance in rural Africa: A multi-country analysis

https://doi.org/10.1016/j.socscimed.2013.02.047Get rights and content

Abstract

While the importance of antenatal care for maternal and child health continues to be debated, several researchers have documented its impact on intermediate variables affecting survival such as birth weight. These studies have also highlighted the problems of causality that are typically not taken into account when estimating the effects of antenatal care on skilled birth attendance. In this study, we revisit this relation in the rural areas of four countries: Ghana, Kenya, Uganda and Tanzania. Using a structural equation modeling approach that corrects for endogeneity, in all four countries we find that the usual simpler probit (or logit) models tend to underestimate the direct effect of antenatal care on skilled birth attendance. Furthermore, in two of the countries, this estimated effect is mediated by the range of services offered to women during antenatal care. These results suggest that governments and NGOs should place more importance on the role of antenatal care providers and on the services they offer, in efforts to promote skilled birth attendance.

Highlights

► The effect of antenatal care on skilled birth attendance is biased downward in all four countries if endogeneity is not taken into account. ► This effect is mediated by the quality of antenatal care service in Kenya and Uganda but not in Ghana and Tanzania. ► Women's level of education influence both the quality of antenatal care received and the delivery with professional. ► Distance to health facility has negative effect on skilled birth attendance in all countries.

Introduction

The importance of antenatal care (ANC) for the prevention of maternal and infant mortality in developing countries is an ongoing debate (Villar et al., 2001). Beyond its role of detecting malformation problems and other risk factors, antenatal care can also be a means of educating women on the advantages of giving birth in medically-controlled conditions (de Bernis, Sherratt, AbouZahr, & Van Lerberghe, 2003). In areas where skilled birth attendance (SBA) remains uncommon, this second role is far from negligible in importance, as the timely use of qualified personnel reduces the risk of death for both the mother and newborn (de Bernis et al., 2003; Gabrysch & Campbell, 2009; Say & Raine, 2007). By increasing the use of formal health services at delivery, antenatal care can have an indirect influence on the survival of mothers and children.

The effect of antenatal care on women's decision to seek skilled birth attendance has received less attention by researchers than its impact on maternal and child health outcomes (Guilkey, Popkin, Akin, & Wong, 1989; Jewell & Rous, 2009). The few studies examining this association have revealed a positive effect of the frequency (de Allegri et al., 2011; Gage, 2007; Guliani, Sepehri, & Serieux, 2012; Stephenson, Baschieri, Clements, Hennink, & Madise, 2006) and the content/quality (Barber, 2006; Bloom, Lippeveld, & Wypij, 1999; Nikiema, Beninguisse, & Haggerty, 2009; Rockers, Wilson, Mbaruku, & Kruk, 2009) of antenatal visits on births attended by qualified practitioners or occurring in a health center. Three pathways may explain the relationship between antenatal care and skilled birth attendance: through the quality of services provided and the information given to women (Akin, Guilkey, & Denton, 1995; Nikiema et al., 2009); by increasing their familiarity with medical personal and thus reducing the “psychological costs” related to seeking their services (Barber, 2006); and by creating or reinforcing habits to make use of this care (Luszczynska & Schwarzer, 2005; Zerai & Tsui, 2001).

Studies on the topic must deal with difficult measurement issues. The number of consultations – by far the most common indicator used to measure antenatal care – does not differentiate women in terms of differences in the quality or range of services received (Bloom et al., 1999) or of their specific motivations, such as perceiving pregnancy complications that underlie their frequency of visits (Ram & Singh, 2005). By defining a composite indicator relating the frequency of ANC care to its content, Bloom et al. (1999) have shown that urban Uttar Pradesh women in the highest quartile of this indicator, whose use is deemed “adequate”, are on average four times more likely to deliver in the presence of trained staff than women in the lowest quartile. With composite indicators, however, it is not possible to separate out which part of the effect on recourse to skilled birth attendance is due to the frequency rather than to the quality of antenatal care (Barber, 2006; Rockers et al., 2009). Nor can they throw light on the relationship linking the frequency of care visits to the quality of the services (Rani, Bonu, & Harvey, 2008).

Aside from questions of measurement, a major problem with previous studies on the relation between antenatal care and medically assisted childbirth is their failure to take into account possible endogeneity biases (Frick & Lantz, 1996; Rockers et al., 2009). There are several reasons to believe that the decisions to seek antenatal care and qualified help at delivery (or to give birth in a health center) are interrelated. First, various characteristics of women or their households such as schooling attainment or income can explain why women may opt for both types of care. For instance, Nikiema et al. (2009) argue that women wishing to give birth in a health center are also those who make the most use of ANC services. In this case, unless we are able to include all the characteristics influencing the use of both of these services, there will be a simultaneity bias in the estimated effect of antenatal care on skilled birth attendance (Cramer, 1995; Joyce, 1994). Analyzing the influence of antenatal care on birth weight, Joyce (1994) showed that women who receive adequate care are different from other women with regard to certain unobservable factors which, if not taken into account, cause the effects of antenatal care to be underestimated. As a consequence, women are likely to differ in terms of unobservable factors associated not only with the use of ANC care, but also with the quality of services received and their likelihood to use skilled personal for delivery (Nikiema et al., 2009; Rockers et al., 2009).

Concern over possible health problems may also affect both the demand for antenatal care and the likelihood of a skilled birth attendance, acting to bias the estimated impact of ANC on SBA in single equation models (Jewell & Rous, 2009). Thus, women experiencing complications or worried about health problems may seek both antenatal care and skilled attendance at birth more frequently than other women (Bloom, Wypij, & Das Gupta, 2001). However, expecting complications may, at times, have the opposite effect: deterring women from seeking formal care at delivery through fear of a caesarean section (Carter, 2010) or to avoid the direct and indirect costs of interrupting their normal activities. Thus, the direction of the bias cannot be known in advance, as it will depend on the dominant effect.

The antenatal care/skilled birth attendance association may also be partly explained by the existence of contextual factors that affect both phenomena simultaneously. This is the case when certain socio-cultural norms discourage women from using maternity services of any kind – antenatal care or skilled assistance at birth (Beninguisse, Nikiema, Fournier, & Haddad, 2005; Sepehri, Sarma, Simpson, & Moshiri, 2008; Stephenson et al., 2006). For instance, Beninguisse et al. (2005) describe cultural practices that restrict access to antenatal consultations during the first months of pregnancy in some contexts in Cameroon, where women have to hide their pregnancy to avoid attracting the attention of “evil spirits.” These norms also place a high value on natural, “non-medicalised” childbirth. In contrast, most awareness campaigns promote both antenatal care and skilled birth attendance (Guilkey & Hutchinson, 2011).

Qualitative research has shed some light on the endogenous nature of the antenatal care variable. Women may make antenatal visits to verify how their pregnancy is progressing and, once reassured, they may not necessarily seek skilled help for the birth itself (Amooti-Kaguna & Nuwaha, 2000). A variety of reasons may underlie these decisions, and the researcher is not always aware of them. Some women may be seeking a kind of assurance vis-à-vis their community so as to avoid personal blame should their pregnancy end badly (Carter, 2010). In communities that prize childbirth without medical assistance, women may use these services to make sure that the pregnancy is progressing without complications. In addition, antenatal visits can be planned ahead of time (for example, coinciding with a trip to a local market), something that is more difficult to do for childbirth. In cases where health care is not easily accessible, a woman may opt for antenatal care so as to reduce the need for SBA. Finally, antenatal care can act as a guarantee of a hospital birth should the need arise, as women will be registered in the health system (Myer & Harrison, 2003).

While these methodological issues have been recognized by many studies, to our knowledge no attempt has been made to address them through the use of appropriate methods applied to cross-sectional data. This study aims to re-examine this relationship by using methods that correct for the effects of unobserved factors that may bias the estimated effect of ANC care on skilled birth attendance. In this paper, we will address both the methodological and substantive limitations of existing studies. From a methodological standpoint, we test for unobserved heterogeneity in the relationship between ANC and SBA using a recursive biprobit model (Babalola & Kincaid, 2009), and we compare the results with those of a simple probit model. A full structural equation modeling (SEM) approach is then used to understand how the services received during ANC visits may explain the relation between ANC care and SBA, again after accounting for the effects of possible endogeneity. By comparing the results, we can assess their robustness and more accurately the true causal effects of ANC care on SBA in a non-experimental design setting.

Section snippets

Data and selected countries

The study is based on publicly accessible large DHS data sets that passed all ethical reviews and which are carefully designed to ensure the complete confidentiality of respondents.

Data are from four Demographic and Health Surveys (DHS). The DHS apply multi-stage stratified probability based sampling to provide nationally and rural/urban representative samples of women of reproductive age. These surveys used standardized questionnaires that allow for a comparison of results across countries,

Results

Table 1 presents the descriptive statistics for the analysis sample. The proportion of women in rural areas who had four or more antenatal visits ranges from 46% in Uganda to 63% in Ghana. Skilled birth attendance was consistently lower, ranging from 32% in Ghana to 51% in Tanzania.

Discussion

This study aimed to analyze the impact of antenatal care on skilled birth attendance in rural Africa, while attempting to correct for problems of endogeneity in the use of antenatal care. We estimated a structural equations model that separates the effect of the number of visits from that of the content of antenatal care, and delineates the determinants of the content of antenatal care and the decision to seek skilled attendance at birth in the rural areas of four sub-Saharan African countries.

Acknowledgments

The helpful comments from the reviewers are greatly appreciated. An early version of this paper is presented as poster at PAA 2012 and as communication at the Canadian Population Society 2012 Annual Meeting.

References (46)

  • G. Zanconato et al.

    Antenatal care in developing countries: the need for a tailored model

    Seminars in Fetal & Neonatal Medicine

    (2006)
  • Adjiwanou, V. Contexte de genre, autonomie de la femme et utilisation des services de santé maternelle en milieu rural...
  • K. Allendorf

    Going nuclear? Family structure and young women's health in India, 1992–2006

    Demography

    (2012)
  • S. Babalola et al.

    New methods for estimating the impact of health communication programs

    Communication Methods and Measures

    (2009)
  • S.L. Barber

    Does the quality of prenatal care matter in promoting skilled institutional delivery? A study in rural Mexico

    Maternal and Child Health Journal

    (2006)
  • G. Beninguisse et al.

    L'accessibilité culturelle: une exigence de la qualité des services et soins obstétricaux en Afrique

    African Population Studies

    (2005)
  • L. de Bernis et al.

    Skilled attendants for pregnancy, childbirth and postnatal care

    British Medical Bulletin

    (2003)
  • S. Bloom et al.

    Does antenatal care make a difference to safe delivery? A study in urban Uttar Pradesh, India

    Health Policy and Planning

    (1999)
  • S. Bloom et al.

    Dimensions of women's autonomy and the influence on maternal health care utilization in a North Indian City

    Demography

    (2001)
  • J. Bruce

    Fundamental elements of the quality of care: a simple framework

    Studies in Family Planning

    (1990)
  • A. Carter

    Factors that contribute to the low uptake of skilled care during delivery in Malindi, Kenya

    (2010)
  • J.C. Cramer

    Racial and ethnic differences in birthweight: the role of income and financial assistance

    Demography

    (1995)
  • K.D. Frick et al.

    Selection bias in prenatal care utilization: an interdisciplinary framework and review of the literature

    Medical Care Research and Review

    (1996)
  • Cited by (0)

    View full text