Many pregnant women in Africa who access professional antenatal care do not receive all the WHO-recommended components of care. Using Demographic and Health Survey (DHS) data from Kenya, Malawi and Nigeria, this study assesses the relationship of education level with the quality of antenatal care received and highlights how the number of antenatal visits mediates this relationship. The results show that a large proportion of the effect of education level on quality of care is direct, while only a small portion is mediated through the number of antenatal visits. Efforts to improve pregnancy outcomes for under-privileged women should focus on removing structural barriers to access, strengthening the technical and interpersonal skills of providers, and addressing providers’ biases and discriminatory practices towards these women. Such efforts should also seek to empower underprivileged women to insist on quality antenatal care by explaining what to expect during an antenatal visit.
Antenatal care, education, quality, mediation analysis, Africa
Recent estimates of maternal mortality indicate that, globally, about 358,000 women died from problems related to pregnancy and childbirth in 2008.1 Poor maternal health contributes significantly to poor clinical outcomes for the infant. About half of newborn deaths occur during the first 24 hours of life, due to infections, asphyxia and other complications occurring during childbirth and the post-partum period.2 Most of the women experiencing complications during pregnancy, childbirth, and the post-partum period reside in developing countries where access to quality maternal health care is limited. Sub-Saharan Africa presents a disproportionately high burden of maternal morbidity and mortality. About 60% of all maternal deaths in 2008 were estimated to have occurred in sub-Saharan Africa.1 Complications related to pregnancy and childbirth are major causes of death for women of reproductive age in this region. The major causes of maternal morbidity and mortality (prolonged bleeding, infection, high blood pressure, unsafe abortions, and obstructed labor) can safely be managed with early detection and access to the right type of care and intervention. Many women [End Page 161] do not access the care necessary to prevent and treat maternal health complications because of multiple individual, household, normative, health system and policy factors.
The Millennium Development Goal number 5 (MDG 5) focuses on improving maternal health.3 One component of this goal seeks to achieve a 75% reduction in maternal mortality ratio (MMR) between 1990 and 2015. There has been significant progress worldwide towards this target; the largest gains in this area have been in Southeast Asia and Northern Africa.3 Nonetheless, some African countries experienced increased MMR between 1990 and 2008, including Botswana, Zimbabwe, South Africa, Swaziland, and Lesotho. The second component of the MDG 5 goal was to achieve universal access to reproductive health care by 2015. One of the main indicators for this goal is antenatal care. Available data suggest that although most women are attending at least one health care visit during pregnancy, increasing from 64% in 1990 to 81% in 2008, not enough women are receiving the recommended quality of antenatal care. For example, about half of women in developing countries received the recommended four antenatal visits in 2009 although this number represents a significant increase from 35% in 1990.3 Antenatal care provides an opportunity to check health indicators and a forum to educate women on healthy pregnancy outcomes and complications that may arise. The World Health Organization recommends that, as part of routine antenatal care, all pregnant women should receive monitoring of pregnancy progress, appropriate tests to detect pregnancy-related complications, tetanus vaccination, anemia prevention and control, and information about home care, nutrition, family planning and healthy lifestyles.4 There is evidence that many women who received antenatal care services from a skilled provider do not receive the recommended components of care. Studies have shown that even in African settings where the majority of women attend antenatal clinics, most do not receive the information necessary for appropriate pregnancy care, including danger signs in pregnancy, and what to do in case of complications.5,6
Education delivers multiple market and non-market benefits. A woman’s education has been shown to be associated with lower fertility and improvements in her earning potential, productivity, health, and longevity, as well as to offer benefits for her family and community.7–14 Education has been consistently linked with pregnancy and maternal outcomes for women. Evidence from many studies suggest that, compared with women having more education, women with less education may be at an increased risk for preterm birth, small-for-gestational-age birth, still-birth, neonatal and post-neonatal death, and maternal death.15–18 Many studies have found that better educated women are more likely than their peers with little or no education to use skilled antenatal care.19–22
Access to antenatal care, however, only paints part of the picture. What happens when women eventually access antenatal care? Does education play a role in the components of antenatal care received once a woman makes it to the health facility? A few studies in Africa and elsewhere have looked at the role of women’s education in maternal care and found that women with lower levels of education are far less likely to receive the recommended care during pregnancy or attend a health care facility for delivery.23–27 However, these previous studies have generally looked at a single component of care. In this paper, we focus on women who reportedly received skilled care during their most recent pregnancy, that is, who reported that they saw a health professional (doctor, [End Page 162] nurse, or nurse-midwife) during an antenatal visit. We examined the relationship between education level and the quality of antenatal care, defined in this paper using a combination of key components performed during antenatal care visits. Education may also be associated with the total number of antenatal care visits a woman makes during a pregnancy. The number of antenatal visits defines the opportunity for receiving quality antenatal care. This study examines the mediating role of number of antenatal visits in the relationship between a woman’s education level and the quality of antenatal care received. We used data from three African countries to examine commonalities and differences across settings. We hypothesize that education level is associated with quality of care directly and indirectly through its association with the number of antenatal care visits (see Figure 1). We acknowledge the oversimplification inherent in the analytic framework presented in Figure 1. Indeed, it is obvious that there are factors operating at the health system and wider societal levels that may prevent even the most educated women from accessing quality care. For example, a woman’s education cannot reasonably be expected to compensate for health facility understaffing, the lack of provider skills, stock-out of necessary supplies or the absence of required equipment. Nonetheless, a mediation model, such as the one employed in this paper, allows us a better understanding of the complex relationship between education and quality of care.
The data analyzed in this paper derived from Demographic and Health Surveys (DHS) conducted in three African countries: Kenya (2008–09), Malawi (2010), and Nigeria (2008) by ICF Macro and its country partners. These three countries were selected to represent three different regions of Africa: East, South, and West. The surveys followed a similar sampling design with eligible women selected through a multi-stage procedure. A detailed description of the sampling design for each survey is provided in the respective study report, accessible from the Measure DHS website (http://www.measuredhs.com/countries). The survey sampling design allowed representativeness at the national or sub-national levels. Survey teams used a structured questionnaire to interview eligible respondents who consented to participate in the survey. [End Page 163]
The analyses focused on the most recent birth among women of reproductive age who had a birth during the five years preceding the survey and who reportedly saw a skilled provider during the antenatal period.
The outcome variable that we analyzed in this paper is the quality of antenatal care received during the most recent pregnancy. We defined quality as a combination of the following components of antenatal care: whether the woman’s blood pressure was measured during an antenatal visit (BPtaken), whether a sample of her urine was taken for analysis (urinanalyzed), whether a sample of her blood was taken for analysis (bloodanalyzed), whether the woman was told about possible complications during pregnancy (complication_info), and whether she bought or was given iron tablets during an antenatal visit (given_boughtiron). A description of how these variables were measured is provided on
The key independent variable assessed in this paper is the woman’s education level, defined as a binary variable distinguishing between women with primary or no formal education (i.e., having completed between 0 and 6 years of schooling in Nigeria, between 0 and 7 years in Kenya, and between 0 and 8 years in Malawi), and post-primary education (i.e., having completed at least 7 years of schooling in Nigeria, at least 8 years in Kenya, and at least 9 years in Malawi). This classification may result in differently classifying women with the same number of years of education across the three study countries. Nonetheless, we prefer to use this approach rather than standardize across countries using an arbitrary number of years of schooling since it is plausible to assume that a person’s perceived status depends on their level of education rather than on the number of years of schooling they have completed. In the estimated models, we examined the mediating role of the number of antenatal visits or the extent to which the effects of education on quality of care are channeled through this variable. The estimated models controlled for six covariates: age, age-squared, place of antenatal care (hospital versus lower-level facilities), number of children ever born, household wealth status and urban residence.
We started with bivariate analysis looking at the following relationships between: (1) the number of antenatal visits and education level; (2) the number of antenatal visits and quality of care; and, (3) education level and quality of antenatal care. Subsequently, we performed mediation analysis to assess the direct and indirect effects of education level on quality of care. In the estimated mediation models, the number of antenatal visits was modeled as a mediator variable while age, age-squared, place of antenatal care, number of children ever born, household wealth status and urban residence were covariates. In its most basic form, mediation analysis as originally proposed by Baron and Kenny involves four steps.28 In the first step, the relationship between the independent variable (X), and a dependent variable (Y) is estimated. In the second step, the association between the independent variable and mediating variable (M) is obtained. The third step involves estimating the relationship between M and Y. In the fourth step, the mediating effect is obtained as the product of the coefficient of [End Page 164]
. Description of Dependent and Independent Variables
|Variable||Description and measurement|
|Quality of skilled care||Derived from five components of care variables:
BPtaken: Takes on a value of 1 if the woman reported that her blood pressure was measured and 0 otherwise;
Urinanalyzed: Takes on a value of 1 if the woman reported that her urine sample was taken for analysis during an antenatal visit and 0 otherwise;
Bloodanalyzed: Takes on a value of 1 if the woman reported that her blood sample was taken for analysis during an antenatal visit and 0 otherwise;
Complication_info: Takes on a value of 1 if the woman reported that she received information about possible pregnancy-related complications during an antenatal visit and 0 otherwise; and,
Given_boughtiron: Takes on a value of 1 if the woman reported that she bought or was given iron tablets during an antenatal visit and 0 otherwise;
|The quality of care indicator is a binary variable, obtained by combining the five variables above, and splitting the resulting score into two categories: score below 5 (low quality) and a score of 5 (high quality).|
|Education||Measured as the highest level of education attained by the woman. The variable is defined as binary, distinguishing between women who had secondary or tertiary level (i.e., at least 7 years of schooling in Nigeria, at least 8 years in Kenya, and at least 9 years in Malawi; value = 1) and those with primary or no education (i.e., between 0 and 6 years of schooling in Nigeria, between 0 and 7 years in Kenya, and between 0 and 8 years in Malawi; value = 0).|
|Mediator Variable (Number of antenatal visits)||Measured as number of visits for antenatal care during the pregnancy of the most recent birth The variable was not normally distributed. Therefore, in the regression models, we used a version of this variable that was log-transformed using the lnskew0 command in Stata.|
|Current age||Age of the woman at the time of the survey; continuous variable measured in single years|
|Current age squared||The square of current age. This variable was introduced to test the hypothesis that the relationship between number of antenatal visits and age is not linear but curvilinear.|
|Type of health facility used||A binary variable that distinguishes between using a hospital for any antenatal visit and using a health center or a lower level facility.|
|Number of children ever born||Continuous variable measured as the number of sons and daughters to which the woman had ever given birth.|
|Type of place of residence||A binary variable measured as residence in an urban (value = 1) or a rural location (value = 0).|
|Wealth||The survey data included an asset-based measure of wealth that divides the respondents into five groups representing the poorest, poor, medium, rich and richest quintiles|
[End Page 166]
the relationship of X to Y and that of M to Y. The limitations of such a causal steps-approach have been widely discussed.29 The models were estimated using the medeff command in Stata.30 Medeff is simulation-based and uses a parametric algorithm that builds on the seminal work by Baron and Kenny and extends the basic approach to binary dependent variables.28 We performed the analyses on all women who met the selection criteria in each country. In addition, to assess whether the effects of education level on quality of care depend on wealth status, we estimated separate models for rich (fourth and fifth wealth quintiles) and non-rich (lower three quintiles) women.
This study involves secondary analysis of publicly available DHS data and there was no need for the author to obtain IRB approval to conduct the study. Nonetheless, for the original data collection, ICF Macro and its partners obtained local IRB approval from the Scientific and Ethical Review Committee of Kenya Medical Research Institute (KEMRI) in Kenya, Malawi Health Sciences Research Committee in Malawi, and the National Health Research Ethics Committee in Nigeria.
Table 2 shows the number of women interviewed, the number that had given birth during the past five years, and the number that were eligible for the analysis in each country. The proportion of women with a recent birth who saw a skilled professional during pregnancy was particularly low in Nigeria compared with the two other study countries (Table 1).
Socio-demographic characteristics of the study participants differ across countries (Table 2). For example, on average, Kenyan and Nigerian women were better educated than their peers from Malawi. In addition, Nigerian women were older and more likely to reside in an urban area than women from Kenya or Malawi. There are also significant variations in the proportion of women that received all the essential components of antenatal care during pregnancy. Almost half of the women who saw a skilled provider in Nigeria received all the five essential components of care, compared with about one-fifth [End Page 167] in Malawi and one quarter in Kenya (Table 2). Most pregnant women in the three countries reportedly had their blood pressure measured (over 80%) or had a blood test (over 75%). In contrast, the proportion that reportedly had urine analysis was very low in Malawi (27.9%) and only moderate in Kenya (68.2%). Similarly, many pregnant women in Kenya (56.4%) and in Nigeria (35.0%) were reportedly not told about possible pregnancy-related complications. In each study country, post-primary education significantly increases the odds that a woman will receive each of the components of antenatal care (data not shown).
The results of the bivariate logistic regression models presented on Table 3 show that in each country, post-primary education is significantly positively associated with increased number of antenatal visits. Note that because the number of [End Page 168]
antenatal visits was log-transformed, interpretation of the coefficients is not very simple; the coefficients are interpreted in terms of percent change in the mean score.31 In Kenya for example, compared with no education or primary education, post-primary education significantly increases the mean number of antenatal visits by 18.2% (calculated as [e(0.167) – 1] multiplied by 100; p < .001). Similarly, in Malawi and Nigeria, post-primary education is associated with a 6.1% (p < .001) and 37.2% (p < .001) increase in the mean score for the number of antenatal visits, respectively. The bivariate results further indicate that the number of antenatal visits is associated with significant increases in the odds of receiving high quality care in the three countries. Substantively, the coefficient indicates that a 2.71-fold (Napier’s constant, e) increase in the number of antenatal visits is associated with more than a three-fold increase in the odds of receiving quality care in Kenya, 1.6-fold increase in Malawi and 2.6-fold increase in Nigeria. Finally, in the three countries, post-primary education is associated with greater odds of receiving high quality care compared with primary or no education. In sum, the bivariate analyses indicate that the data meet the three basic conditions for mediation analysis in the three countries.28 The independent variable (education level) is related to both the mediator (number of antenatal visits) and the dependent variable (quality of antenatal [End Page 169] care). In addition, the mediator is related to the dependent variable. In the next section, we estimate the amount of the effect of education level on the quality of care that is mediated through the number of antenatal visits.
The results for the multivariate mediation analysis are presented on Table 4. We present the findings by country in the following paragraphs.
In Kenya, overall, the association of education level with the number of antenatal visits was positive and very significant. After controlling for the relevant correlates, post-primary education was associated with a 10% increase in the mean number of antenatal visits. In addition, education level was associated with about a 45% increase in the odds of receiving high quality antenatal care. The total effect of post-primary education on the quality of antenatal care received was 0.088, less than one fifth of which was mediated through the number of antenatal visits. Although the magnitude of mediated effect was small, it was nonetheless significant. The patterns of effects differ between the rich and the non-rich. Specifically, the relationship between education level and the number of antenatal visits was strongly significant for rich women but weak for their non-rich peers. In contrast, it appears that wealth does not moderate the association of education level with the quality of antenatal care. In other words, wealth tends to accentuate the positive effect of education level on the number of antenatal visits but does not change the relationship between education level and quality of care. The mediated effect was insignificant for non-rich women whereas it was highly significant for rich women.
The results for Malawi show that, overall, education level is predictive of both the number of antenatal visits and the quality of antenatal care received. For example, compared with lower level of education, post-primary education is associated with a 25% increase in the odds of receiving high quality antenatal care. The number of antenatal visits also significantly predicts the quality of care received such that a 2.7-fold increase in the number of antenatal visits is associated with a 56% increase in the odds of receiving quality antenatal care. Overall, only a very small, albeit significant, proportion of the total effect of education level on the quality of care received was mediated through the number of antenatal visits. There are noticeable differences between the rich and the non-rich regarding the role of education level. For the non-rich, education level did not significantly predict the number of antenatal visits or the quality of care received; the reverse is the case for rich women. In other words, for non-rich women, education level does not play any significant role, direct or indirect, in the quality of antenatal care received. For both rich and non-rich women, the number of antenatal visits significantly and positively predicts the quality of antenatal care received.
Overall, post-primary education was significantly and positively associated with the number of antenatal visits and with receiving better quality antenatal care. Compared with a lower level of education, post-primary education is associated with a 24.2% increase in the mean number of antenatal visits. Similarly, the odds of receiving high quality antenatal care were 34% higher among women with post-primary education than among their peers with primary education or less. In addition, the number of antenatal visits positively and significantly predicts the quality of antenatal care. The data show that a 2.7-fold increase in the number of antenatal visits is associated with about a two-fold increase in the odds of receiving high-quality antenatal care. A [End Page 170]
[End Page 172]
noticeable proportion of the total effects of education level on quality of antenatal care was due to the mediating role of the number of antenatal visits. As in the other two study countries, there were noticeable differences between rich and non-rich women, especially with respect to the mediating role of the number of antenatal visits. Wealth appears to attenuate the mediating role of the number of antenatal visits in the relationship between education level and the quality of antenatal care received. Indeed, whereas for rich women, only 12.6% of the total effect of education level on the quality of care was mediated through the number of antenatal visits, the comparative proportion was 25.2% for non-rich women.
This study focused on women who gave birth during the five years preceding the survey and who reportedly accessed skilled antenatal care during their most recent pregnancy in three African countries: Kenya, Malawi, and Nigeria. The study shows that many women attending antenatal care do not receive all the essential components of care. The findings further show considerable educational inequalities in the quality of antenatal care received. Using data from three countries is a key strength of this study, making it possible to see how the relationships vary across settings. In each of the three study countries, the total effect of education level on quality of care is significant. In Kenya and Malawi, only a fraction of the total effect of education level on the quality of care received is mediated through the number of antenatal visits. In Nigeria, the proportion of total effect mediated is more substantial. The data show that education level has an effect on the quality of care above and beyond its association with the number of antenatal visits after controlling for the place of care. In other words, even when the woman accesses the right type of care in a regular manner, her education level still makes a difference in the quality of care that she receives. This finding supports the inverse care law and is consistent with evidence from other studies.32–34 The less educated women are the ones most in need of quality preventive maternal health services. They are more likely to experience complications during pregnancy and childbirth; more likely to be malnourished, to live in areas with limited access to emergency obstetric care, and to be less knowledgeable about infection prevention; they are also more likely to have preexisting health conditions that may lead to complications during pregnancy. For example, a cross-sectional study conducted in Ghana found that signs and symptoms of preeclampsia were significantly higher among women with no formal education.35
The strong direct effect of education level on the quality of care does persist after controlling for household wealth, indicating that the relationship is not attributable to the association of wealth with education. A possible explanation of the observed inequality may be that educated women are more knowledgeable about the procedures to expect during an antenatal visit and more likely to request such procedures. Little education, in particular, can create a social distance between the patient and the provider, thereby making effective communication problematic. Another possible explanation of the observed educational inequalities may have to do with the way women with low levels of education are treated in health facilities. It is possible that service providers [End Page 173] tend to discriminate against women with little education by not providing them with comprehensive information about pregnancy care, not performing all the required tests, and not offering preventive medications (such as iron tablets). Low levels of utilization of maternal health services among underprivileged women have been linked to perceived stigma and discrimination in the health care setting.36,37 Evidence from other settings and for other health issues suggest that service providers tend to display active discrimination and prejudice against disadvantaged groups, including women with little education.38,39 Other studies have linked patient’s education with satisfaction with the quality of client-provider’s interactions in family planning clinics,40 doctor-patient encounters,41 quality of information received during consultations,42 active patient participation,41,43 and level of involvement in decisions regarding care and treatment.44 Future studies should include measures of discrimination and stigma in antenatal care settings to assess the extent of active discrimination against women with low levels of education and determine how this discrimination affects the quality of care received.
Wealth appears to moderate the indirect effect of education level on the quality of care. However, the pattern of moderation effect is not similar across countries. In Kenya, wealth appears to accentuate the positive effect of post-primary education on the number of antenatal visits such that the relationship of education level with the number of antenatal visits was very strong for rich women but weak for their non-rich peers. We can look at the way in which health care system is organized and financed in Kenya for a partial explanation of this pattern of effects. Irrespective of wealth status, the most important source of antenatal care is the public sector. Nonetheless, Kenya has one of the best-developed private health sectors in Africa, contributing over 40% of health services in the country.45 In 2010, of the 6,641 health facilities in the country, 3,541 are in the private sector (comprising 1,425 for-profit and 2,116 non-profit or faith-based facilities). Unlike most other African countries, the private sector in Kenya is very accessible to the poor even in the rural areas.46 Women in the poorest quintiles are as likely to use a private sector health facility for antenatal care as women in the richer quintiles. Nonetheless, the majority of the poor who use private sector facilities use a facility run by a faith-based organization (FBO) whereas the majority of the rich who use a private facilities use a for-profit facility.46 In general, the quality of antenatal care is highest in FBO-run facilities. For example in 2010, the proportion of facilities with all essential supplies for basic antenatal care was 12% in government facilities, 10% in NGO-run facilities, 38% in for-profit private facilities and 68% in FBO-run facilities.45 Taken together, these data indicate that proportionally more poor than non-poor women use facilities with potentially high-quality ANC services, thereby leaving less room for variableness by educational or other socio-demographic differences. Moreover, the mission of FBO facilities is to cater to the health care needs of the poor and underprivileged; it is reasonable to assume that discrimination against women with little education will be less in such facilities compared with facilities that target a broader segment of the population. It is therefore surprising that education level has a significant direct effect on quality of care among the poor in Kenya.
In Malawi, wealth seems to accentuate the effects of education level on both the number of antenatal visits and the quality of care received. Indeed, the data did not reveal any significant effects of education level among the poor on either of these two [End Page 174] outcomes. This finding suggests that inability to pay for services is a strong factor overriding the privileges derived from post-primary education and further hindering access to quality services among the poor. Although antenatal services are generally free in Malawi as in other African countries, shadow fees (sometimes charged to cover supplies) and transportation costs may prevent many non-rich women from making the required number of antenatal visits and receiving quality care, irrespective of their education level.
In Nigeria, wealth appears to depress the positive effect of post-primary education on the number of antenatal visits. The combination of poverty and little education is particularly detrimental for the number of antenatal visits. This finding makes intuitive sense. Indeed, it is possible that rich women, irrespective of their educational level, have better access to the media and other sources of health information and are thereby more knowledgeable about ideal care seeking behaviors. It is also possible that household chores and other domestic responsibilities constitute less of a barrier to accessing care for rich women than for non-rich women.
The findings from this study have significant programmatic and policy implications, especially in connection with the achievement of the MDG 5. Although the study shows that education level has strong direct effects on the quality of antenatal care, the findings should not be interpreted as suggesting that achieving MDG 5 is contingent upon increasing educational opportunities for women. The findings highlight the need to intensify efforts to reach less educated women with quality antenatal care. Obviously, improving the overall quality of antenatal services is essential. Increasing the availability of quality of antenatal care in all types of health facilities, especially in those that are most likely to be visited by women with low level of education is essential. Efforts to remove structural barriers to quality antenatal care for poor and lowly educated women should be intensified. Service providers should understand the rights of women of all backgrounds to quality antenatal care and do their best to provide the best possible quality care to women irrespective of their background. The findings from this study may suggest the presence of active discrimination against women with low level of education; therefore, service providers also must understand the effects that their personal biases may have on disadvantaged groups who are in greatest need of quality antenatal service. In particular, attention should be paid to making underprivileged women feel welcome and comfortable in the health facility. Relevant efforts should not just be supply-focused. Women with low level of education should be empowered to make a demand on health services for quality. Programs that target women with low level of education with relevant information about what they should expect during antenatal visits and that coach them to demand such quality services are relevant.
This study uses data from the most recent DHS in the study country; however, it is important to note that there are some on-going efforts to promote access to quality antenatal care in the study countries. For example, in Nigeria, the federal government and some state governments are working with local and international partners to implement programs that provide monetary and other incentives to women who obtain at least four antenatal visits. In addition, the government is intensifying efforts to recruit and retain qualified midwives and relevant health workers.47,48 The Emergency Human Resources Program (EHRP), focusing on improved staffing of maternal and other [End Page 175] health services, construction of maternity waiting houses, and increased use of Health Surveillance Assistants for home visits are some of the strategies that the government of Malawi is implementing to improve maternal health outcomes.49 In Kenya, the rights-based, community health approach that the Ministry of Public Health and Sanitation adopted in 2007 as part of its National Heath Sector Strategic Plan II (NHSSP II) has been shown to be effective in increasing antenatal attendance and other reproductive health outcomes.50
This study has some limitations that we highlight below. First, the use of cross-sectional data makes claims about causality impossible since both the dependent and the independent variables were measured during the same time period. However, the strong association found between education level and quality of services suggests that the policy and programmatic implications of the results should be taken seriously. Second, the survey data were self-reported and thereby subject to memory lapse, social desirability, and other biases. Focusing on the most recent birth may have helped to minimize the bias due to memory lapse. A related limitation is that it is plausible that women with low education were less likely to understand and thereby recall some of the specific procedures connected with the components of antenatal care that we examined in this paper. Nonetheless, because the relevant questions focused on descriptions of the procedures rather than using specific nomenclatures, it is reasonable to assume that this bias is minimized. Additionally, the way quality of care was operationalized in this study might have introduced some biases into the results. Due to limitations inherent in the survey data analyzed, some essential WHO-recommended interventions during antenatal visits were not included in our indicator of quality of antenatal care. For example, although the indicator included whether or not blood sample was taken, the WHO recommendations were specific about the tests should be conducted with the blood sample, including tests for haemoglobin, HIV, syphilis, and rhesus group. Similarly, although we included whether or not urine sample was taken in the indicator of quality, the WHO recommends testing the urine for bacteriuria and proteinuria. Regrettably, the survey data did not include such details. Finally, the use of education level rather than years of schooling to compare outcomes in countries with different educational systems may have introduced a bias into the results and affected the comparability of the role of education across countries. Nonetheless, the fact that the association of education with the outcome variables is similar across countries makes this problem less of an issue.
Stella Babalola is Associate Professor and teaches Health Communication in the Department of Health, Behavior and Society Johns Hopkins University. She is also Senior Research Advisor at the Center for Communications Programs of the same university. Dr. Babalola has a wealth of experience in international health, teaching, communication and research. During the last five years, her research has been largely in the areas of HIV risk reduction and women’s reproductive health.
1. World Health Organization. Trends in maternal mortality: 1990 to 2008. Estimates developed by WHO, UNICEF, UNFPA and the World Bank. Geneva, Switzerland: World Health Organization, 2010.
2. The Partnership for Maternal, Newborn & Child Health. A global review of the key interventions related to Reproductive, Maternal, Newborn and Child Health (RMNCH). Geneva, Switzerland: The Partnership for Maternal, Newborn & Child Health, 2011.
3. United Nations. The millennium development goals report 2011. New York, NY: United Nations, 2011 [End Page 176]
4. World Health Organization. Pregnancy, childbirth, postpartum and newborn care: a guide for essential practice (2nd Edition). Geneva, Switzerland: World Health Organization, 2006
5. Anya SE, Hydara A, Jaiteh LE. Antenatal care in the Gambia: missed opportunity for information, education and communication. BMC Pregnancy Childbirth. 2008 Mar 7;8:9. http://dx.doi.org/10.1186/1471-2393-8-9; PMid:18325122PMCid:PMC2322944
6. Pembe AB, Carlstedt A, Urassa DP, et al. Quality of antenatal care in rural Tanzania: counselling on pregnancy danger signs. BMC Pregnancy Childbirth. 2010 Jul 1;10:35. http://dx.doi.org/10.1186/1471-2393-10-35; PMid:20594341 PMCid:PMC2907301
8. Hanushek EA. Schooling, gender equity, and economic outcomes. In: Girls’ education in the21st century: gender equality, empowerment, and economic growth. Edited by Tembon M, Fort L, eds. Washington, DC: The International Bank for Reconstruction and Development, 2008.
9. Grossman M. Education and nonmarket outcomes. In: Handbook of the economics of education (Volume 1). Hanushek E, Welch F, eds. Amsterdam, Netherlands: Elsevier, 2006.
10. Wolfe B, Zukevas S. Nonmarket outcomes of schooling. Madison, WI: Institute for Research, 1995.
11. Acemoglu D, Angrist J. How large are the social returns to education? evidence from compulsory attendance laws. Cambridge, MA: National Bureau for Economic Research, 2000.
12. Watson C. Addressing the MDGs and targets for education and gender: comments on selected aspects linked to the ICPD programme of action. New York, NY: United Nations International Children’s Emergency Fund, 2005.
13. Simler KR, Mukherjee S, Dava GL, et al. Rebuilding after war: micro-level determinants of poverty reduction in Mozambique. Washington, DC: International Food Policy Research Institute, 2004. PMCid:PMC1747058
15. Mumbare S, Maindarkar G, Darade R, et al. Maternal risk factors associated with term low birth weight neonates: a matched-pair case control study. Indian Pediatr. 2012 Jan;49(1):25–8. Epub 2011 May 30.
16. Luo ZC, Wilkins R, Kramer MS, et al. Effect of neighbourhood income and maternal education on birth outcomes: a population-based study. CMAJ. 2006 May 9;174(10):1415–20. http://dx.doi.org/10.1503/cmaj.051096; PMid:16682708 PMCid:PMC1455422
17. van den Broek NR, White SA, Ntonya C, et al. Reproductive health in rural Malawi: a population-based survey. BJOG. 2003 Oct;110(10):902–8. PMid:14550359
18. Karlsen S, Say L, Souza JP, et al. The relationship between maternal education and mortality among women giving birth in health care institutions: analysis of the cross sectional WHO Global Survey on Maternal and Perinatal Health. BMC Public Health. 2011 Jul 29;11:606. http://dx.doi.org/10.1186/1471-2458-11-606; PMid:21801399 PMCid:PMC3162526
19. Elo IT. Utilisation of maternal health-care services in Peru: the role of women’s education. Health Transit Rev. 1992 Apr;2(1):49–69. PMid:10148665
20. Celik Y, Hotchkiss DR. The socioeconomic determinants of maternal health care utilization [End Page 177] in Turkey. Soc Sci Med. 2000 Jun;50(12):1797–806. http://dx.doi.org/10.1016/S0277-9536(99)00418-9
21. Anwar I, Sami M, Akhtar N, et al. Inequity in maternal health-care services: evidence from home-based skilled-birth-attendance programmes in Bangladesh. Bull World Health Organ. 2008 Apr;86:252–9. http://dx.doi.org/10.2471/BLT.07.042754; PMid:18438513 PMCid:PMC2647426
22. Babalola S, Fatusi A. Determinants of use of maternal health services in Nigeria—looking beyond individual and household factors. BMC Pregnancy Childbirth. 2009 Sep 15;9:43. http://dx.doi.org/10.1186/1471-2393-9-43; PMid:19754941 PMCid:PMC2754433
23. Agha S, Carton TW. Determinants of institutional delivery in rural Jhang, Pakistan. Int J Equity Health. 2011 Jul 30;10:31. http://dx.doi.org/10.1186/1475-9276-10-31; PMid:21801437 PMCid:PMC3159141
24. Edmonds JK, Hruschka D, Bernard HR, et al. Women’s social networks and birth attendant decisions: application of the Network-Episode Model. Soc Sci Med. 2012 Feb; 74(3):452–9. Epub 2011 Dec 7. http://dx.doi.org/10.1016/j.socscimed.2011.10.032; PMid:22196965 PMCid:PMC3265634
25. Ahmed S, Creanga AA, Gillespie DG, et al. Economic status, education and empowerment: implications for maternal health service utilization in developing countries. PLoS One. 2010 Jun 23;5(6):e11190. http://dx.doi.org/10.1371/journal.pone.0011190; PMid:20585646 PMCid:PMC2890410
26. Habibov NN. On the socio-economic determinants of antenatal care utilization in Azerbaijan: evidence and policy implications for reforms. Health Econ Policy Law. 2011 Apr;6(2):175–203. Epub 2010 Jul 2. http://dx.doi.org/10.1017/S1744133110000174; PMid:20598213
27. Rani M, Bonu S, Harvey S. Differentials in the quality of antenatal care in India. Int J Qual Health C. 2008 Feb;20(1):62–71. Epub 2007 Nov 17. http://dx.doi.org/10.1093/intqhc/mzm052; PMid:18024998
28. Baron RM, Kenny DA. The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986 Dec;51(6):1173–82. http://dx.doi.org/10.1037/0022-35188.8.131.523; PMid:3806354
29. MacKinnon DP, Fairchild AJ, Fritz MS. Mediation analysis. Annu Rev Psychol. 2007;58: 593-614. http://dx.doi.org/10.1146/annurev.psych.58.110405.085542; PMid:16968208 PMCid:PMC2819368
30. Hicks R, Tingley D. Causal mediation analysis. Stata J. 2011;11(4):609-15.
31. VittinghoffE, Glidden DV, Shiboski SC, et al. Regression methods in biostatistics: linear, logistic, survival, and repeated measures models. New York, NY: Springer, 2005.
32. Hart TJ. The inverse care law. Lancet. 1971 Feb 27;297(7696):40512.
34. Mercer SW, Watt GC. The inverse care law: clinical primary care encounters in deprived and affluent areas of Scotland. Ann Fam Med. 2007 Nov–Dec;5(6):503–10. http://dx.doi.org/10.1370/afm.778; PMid:18025487 PMCid:PMC2094031
35. Dinglas C, Lardner D, Homchaudhari A, et al. Relationship of reported clinical features of pre-eclampsia and postpartum haemorrhage to demographic and other variables. West Afr J Med. 2011 Mar–Apr;30(2):84–8. PMid:21984453 [End Page 178]
36. Adamson PC, Krupp K, Niranjankumar B, et al. Are marginalized women being left behind? a population-based study of institutional deliveries in Karnataka, India. BMC Public Health. 2012 Jan 12;12:30. http://dx.doi.org/10.1186/1471-2458-12-30; PMid:22240002 PMCid:PMC3269389
37. Napravnik S, Royce R, Walter E, et al. HIV-1 infected women and prenatal care utilization: barriers and facilitators. AIDS Patient Care STDS. 2000 Aug;14(8):411–20. http://dx.doi.org/10.1089/108729100416623; PMid:10977970
38. Barber SL, Bertozzi SM, Gertler PJ. Variations in prenatal care quality for the rural poor in Mexico. Health Affair (Millwood). 2007 May–Jun;26(3):w310–23. Epub 2007 Mar 27. http://dx.doi.org/10.1377/hlthaff.26.3.w310; PMid:17389636
39. Burgess DJ, Fu SS, van Ryn M. Why do providers contribute to disparities and what can be done about it? J Gen Intern Med. 2004 Nov;19(11):1154–9. http://dx.doi.org/10.1111/j.1525-1497.2004.30227.x; PMid:15566446 PMCid:PMC1494785
41. Peck BM. Age-related differences in doctor-patient interaction and patient satisfaction. Curr Gerontol Geriatr Res. 2011;2011:137492. Epub 2011 Oct 5.
42. Hall JA, Roter DL, Katz NR. Meta-analysis of correlates of provider behavior in medical encounters. Med Care. 1988 Jul;26(7):657–75. http://dx.doi.org/10.1097/00005650-198807000-00002; PMid:3292851
43. Street R. Information giving in medical consultations: the influence of patients’ communicative styles and personal characteristics. Soc Sci Med. 1991;32(5):541–8. http://dx.doi.org/10.1016/0277-9536(91)90288-N
44. Smith SK, Dixon A, Trevena L, et al. Exploring patient involvement in healthcare decision making across education and functional health literacy groups. Soc Sci Med. 2009 Dec; 69(12):1805–12. Epub 2009 Oct 19. http://dx.doi.org/10.1016/j.socscimed.2009.09.056; PMid:19846245
45. National Coordinating Agency for Population and Development, Kenya Ministry of Medical Services, Kenya Ministry of Public Health and Sanitation, et al. Kenya service provision assessment survey 2010. Nairobi, Kenya: National Coordinating Agency for Population and Development, 2011.
47. Cooke J, Tahir R. Maternal health in Nigeria: with leadership, progress is possible: a report of the Center for Strategic and International Studies (CSIS) Global Health Policy Center. Washington, DC: Center for Strategic and International Studies, 2013.
48. Doctor HV, Findley SE, Ager A, et al. Using community-based research to shape the design and delivery of maternal health services in Northern Nigeria. Reprod Health Matters. 2012 Jun;20(39):104–12. http://dx.doi.org/10.1016/S0968-8080(12)39615-8
49. Ergo A, Shah N, Rashidi T et al. Malawi case study: how health system strengthening efforts have affected maternal health. Washington, DC: Maternal and Child Health Integrated Program, 2011.
50. Wangalwa G, Cudjoe B, Wamalwa D, et al. Effectiveness of Kenya’s Community Health Strategy in delivering community-based maternal and newborn health care in Busia County, Kenya: non-randomized pre-test post test study. Pan Afr Med J. 2012;13 Suppl 1:12. Epub 2012 Dec 26; PMid:23467438 PMCid:PMC3587017 [End Page 179]