The fifth Millennium Development Goal outlines the international commitment to measurably reduce maternal mortality by the year 2015 [1]. Antenatal care (ANC) is a critical strategy in reducing maternal mortality as it facilitates the identification and mitigation of risk factors early in pregnancy [2]. Timely and frequent use of ANC enables delivery of essential services, including malaria treatment, immunizations, and health counselling [3–6].

In particular, ANC clinics act as a key entry-point for implementing nutrition and health educational interventions that promote preventive health behaviours to improve maternal and neonatal health through better knowledge, attitudes and practices. Studies conducted in Turkey, Pakistan and Laos provide evidence to support the role of ANC in improving health knowledge, attitudes and practices (KAP) among women who utilize the service [7]. In the implementation and evaluation of a community-based ANC education programme in Istanbul [8], women in the ANC education group were reported to be more likely to initiate breastfeeding within the first two hours after delivery, bring infants for check-up within seven days after birth and to implement family planning measures at three months after birth, compared to the control group. Similarly, in a cross-sectional survey in Islamabad, Pakistan, Alam and colleagues showed that women attending ANC clinics were more likely to recognize signs of a difficult pregnancy, to realize the importance of eating a healthy diet, and to indicate tetanus immunization uptake, compared to their non-attending counterparts [9]. And in rural Laos, women who had received ANC were more likely to utilize health services at delivery and had a greater mean knowledge score regarding obstetric care compared to the women who had not received any antenatal care [10].

In Kenya, the World Health Organization (WHO) focused antenatal care guidelines, which recommend at least four antenatal visits for women with low-risk pregnancy and provide evidence-based content for each visit, were adopted and implemented in 2001 [11]. Despite a 92% national average for at least one ANC visit in 2008, the maternal mortality rate in Kenya remains at a high of 488 maternal deaths per 100,000 live births [12]. While this disparity may reflect challenges in feasibility and acceptability of the ANC package due to limited resources for materials and supplies, lack of training for service providers, and lack of clear policy direction [11], it may also be an indication of gaps in service provision at the ANC clinics, particularly in educational components. As Nikiema et al. showed in a cross-country analysis of Demographic Health Survey data from 19 sub-Saharan countries, healthcare providers do not routinely provide women with information as part of ANC or fail to provide information in a way that is remembered by women [13], which further impedes the efforts to improve health and nutrition KAP among women.

Among the few studies that assess the impact of ANC services on maternal KAP, few data emerge from sub-Saharan Africa. In a cross-sectional survey of women in rural Kenya, van Eijk et al. found that 11% of the women, who attended ANC clinics prior to delivery, were receptive to the health information provided during their visits [14]. In rural Nigeria women identified midwives and nurses as the major source of health information on maternal and child health during focus group discussions [15]. Yet, the comparative assessment of whether the KAP of women improved due to ANC utilization was not conducted in these studies. Recognizing the disparities in knowledge, attitudes, and practices of pregnant women who attend ANC clinics versus those who do not has important implications for assessing the impact of education interventions provided through ANC services.

The objective of this study, therefore, was to compare nutrition and health knowledge, attitudes, and practices of pregnant women who attended versus did not attend ANC clinics in rural Kenya. Specifically, we aimed to better understand the impact of ANC attendance on nutrition knowledge, health and healthcare knowledge, attitudes towards preventative health practices, and dietary diversity of pregnant women, taking into account other relevant covariates.

Agriculture as a primary activity and selling agricultural products were significantly greater among women who sought ANC, compared to women who did not attend ANC clinics (15.3% vs. 12.5%). Women who did not attend ANC clinics were more likely to be involved in agriculture as their primary activity but did not sell agricultural products (62.7% vs. 52.1%), were more likely to undertake casual labour (45.7% vs. 34.2%) and were more likely to be self-employed (27.2% vs. 20.1%) compared to ANC clinic attendees. The proportion of women involved in two or more income generating activities was greater among women who did not attend ANC clinics compared to ANC seeking women (21.7% vs. 13.6%). Conversely, a greater proportion of women who attended ANC clinics were either not involved in income generating activities or were involved in only one activity. In addition, household head education above primary school, wealth index and frequent radio use were significantly greater for ANC clinic attendees compared to non-attending women. Women who attended ANC clinics were less likely to be in the lowest wealth index quartile (21.3% versus 28.5%) and were more likely to listen to the radio every day or at least once a week compared to women who had not attended ANC clinic.

At the community level, there were no statistically significant differences between women who attended and did not attend ANC clinics. Village size ranged from 38 to 296 households and the number of pregnant women participating from villages ranged from two to 25 women; 72.5% of the women lived in villages which had previously held health meetings organized by The AIDS, Population and Health Integrated Assistance (APHIA) program; village committees for health and nutrition existed for 69.8% of the women; and health interventions had previously occurred in villages for 50.8% of the women.

This study provides key insights into the impact of ANC attendance on the knowledge, attitudes and practices of pregnant women in rural Kenya. Among women who had attended ANC clinic at least once, health knowledge score and proportion who received malaria or anthelminthic treatment were significantly greater compared to their non-attending counterparts. Similar findings were observed in rural Pakistan and Laos where ANC clinic attendees demonstrated greater knowledge of obstetric complications than non-attending women [9, 10]. Further, the positive association observed between the number of ANC clinic visits and higher health knowledge score, after controlling for maternal socio-demographic characteristics, is encouraging. This parallels the positive trend observed in the cross-country analysis of 19 sub-Saharan countries, where compared to women who had attended ANC clinics once, the likelihood of receiving advice was almost twice as much for women who attended ANC clinics three times, and closer to three times as much for women who visited ANC clinics at least five times [13]. This suggests that health and healthcare education provided at the ANC clinics may confer substantive improvements in health knowledge among pregnant women in rural Kenya.

Nonetheless, greater efforts to improve the quality and uptake of educational interventions implemented through ANC services are necessary in sub-Saharan Africa. The lack of difference in nutrition knowledge and dietary diversity outcomes by ANC clinic attendance observed in this study may be a result of homogenization of the maternal KAP due to village nutrition committees in the communities surveyed, limited uptake of information offered at ANC clinics, and/or due to a lack of effective nutrition counselling during ANC visits contributing to minimal nutrition knowledge gain [13]. Sub-optimal nutrition counselling has previously been observed in an acceptability and sustainability analysis of the focused ANC package in Kenya where only one third of the women attending ANC clinics for the first time received nutritional counselling [11]. In rural Uganda, ineffective organization of educational sessions recommended as part of the ANC service provision was noted as a major impediment to increasing maternal knowledge [27], and in The Gambia, Anya et al. showed that despite high rates of ANC attendance, there was limited benefit observed for women who sought ANC services [28]. Moreover, only 14% of the women surveyed in a previous cross-sectional survey conducted in Western Kenya reportedly attended a health talk at the ANC clinic and the health topics covered during ANC visits were noted to be limited in scope [14]. Similarly, in rural Tanzania, delivery of health information, including nutrition counselling, at the ANC clinics was noted to be among the least likely components of ANC to be effectively implemented [29]. These data indicate persistent gaps in effective counselling, specifically in nutrition topics, provided through ANC clinics.

Not surprising then, the average household dietary diversity score for women in this survey was less than five, a level previously demonstrated to correlate with micronutrient inadequacy [19]. Additionally, the reported lack of adequate food consumption during pregnancy, although not unique to the Kenyan context, is concerning. In a cross-sectional survey in rural Bangladesh, only one-third of women surveyed reported increasing their food intake during pregnancy [30], and in a focus-group analysis in Bali, women reported restricting the amount of food consumed during pregnancy to prevent having a large infant and a difficult delivery [31]. Such practices fail to meet the increased nutritional demands necessary to maintain optimal health during pregnancy [32] and undermine public health efforts targeted to improve maternal and peri-natal health.

In addition to ANC attendance, maternal education level was positively associated with health and nutrition knowledge, health attitudes, and dietary diversity. Higher maternal education level has been previously correlated with greater health knowledge scores among women in Laos, China and Turkey [8, 10, 33]. Gestational age, however, was negatively associated with health knowledge score in this population suggesting that a delay in seeking ANC may contribute to less health knowledge as the opportunities to receive advice and education are reduced. Radio use was also noted to be greater among ANC attending women and demonstrated positive associations with dietary diversity score, indicating that radio-broadcasted public health messages may be an effective method for providing educational programs to improve healthcare knowledge in this population. Media interaction, in the form of radio messages regarding health during pregnancy, has been reported to increase awareness among Guatemalan mothers and supplement their clinic-based education [34].

An interesting finding of this study was the positive association between income generating activities and health knowledge score. Although our results parallel those from China [12], where employed women were noted to have higher health knowledge compared to unemployed women, higher involvement in income generating activities was negatively associated with ANC attendance. In this study, a lower proportion of women involved in two income generating activities attended ANC clinics compared to women involved in only one activity or less. This finding provides evidence for the maternal “double-burden”, characterized by managing the increasing demands of labour and economic productivity with care-taking responsibilities [35]. Women who were involved in two income generating activities may have limited time to seek preventative ANC in this population.

Our interpretation regarding the temporality of the observed associations is limited due to the cross-sectional nature of the survey. In addition, although women were notified in advance of the survey date and transportation was provided to those who had difficulties reaching the survey site, inviting women to a centralized survey location in the village may have introduced selection bias and may limit the generalizability of these results. Also, although maternal parity is an important factor in antenatal care use, it was not possible to control for parity in adjusted analysis as this data was not directly collected in the survey. Women’s autonomy, gender relationships and social networks have been previously found to lend to differences in beliefs and preferences relating to maternal healthcare seeking behaviours [36], but were not explored in this analysis given the lack of heterogeneity in the data. Furthermore, although the use of a 24-hour recall to construct the DDS does not account for within-person variance, the DDS is considered to be a valid proxy indicator to reflect nutrition adequacy and is often employed in low resource settings [37]. Future studies should explore the factors associated with maternal health care knowledge and behaviours and utilize qualitative research to better understand the nuances in maternal KAP due to contextual variables.

This study revealed that health knowledge and receiving malaria and anthelmintic practices alone were higher among women who attended ANC clinics compared to those who did not. Attending ANC clinics confers some benefit, but not in nutrition education, dietary diversity or improved food consumption practices. Thus, a greater emphasis on program implementation of health and nutrition counselling is required to ensure effective uptake of health and nutrition knowledge, and improved dietary and food consumption practices. Our findings emphasize the need for programs that aim to address the gaps in education delivery and counselling at the ANC clinics. Maternal knowledge, attitudes, and practices are important evaluation indicators to assess the impact of educational intervention offered through ANC clinics. Future emphasis on multi-level qualitative and quantitative research in this context may facilitate a greater understanding of determinants of health and nutrition knowledge, attitudes and practices among pregnant women.