Adequate diet during pregnancy has positive effects on the mother and pregnancy outcome. Assessment of diet quality during pregnancy is particularly important in areas where household food security is suboptimal, to enable appropriate targeting and intervention. This study assessed diet quality and identified predicting factors among pregnant women in northern Ghana. A cross-sectional study involving 403 pregnant women was conducted in May 2018. Pregnant women attending antenatal care clinics (ANC) were selected using simple random sampling technique. We assessed socio-demographic characteristics, 24-h recall and household food security. The minimum dietary diversity for women (MDD-W) was used as a proxy measure for diet quality based on Food and Agricultural Organization (FAO) guidelines. Logistic regression models were fitted to determine the predictors of diet quality. The mean dietary diversity score (DDS) of 10 food groups was 4.4 ± 1.1 (95% CI: 4.3–4.5). Logistic regression showed that women of high educational level (adjusted odds ratio [AOR] = 2.42; 95% confidence interval [CI] [1.21–4.84]; P = 0.01), women of high household wealth index (AOR = 1.78; 95% CI [1.14–2.77]; P = 0.01], none/mild household hunger (AOR = 2.71; 95% CI [1.26–5.82]; P = 0.01), medium household size (6–15 members) (AOR = 1.66; 95% CI [1.04–2.66]; P = 0.03) and women of gestational age 20–35 weeks (AOR = 1.89; 95% CI [1.05–3.40]; P = 0.03) were more likely to have quality diets after adjusting for potential confounding variables. Diet quality among pregnant women was low and was predicted by educational level, household wealth, gestational age and food security. Women education and improvements in household food security could impact diets of pregnant women in northern Ghana.
The study was conducted in the Savelugu‐Nantong district in the northern region of Ghana; it has a single farming season, a potential for livestock and crop production (maize, millet, groundnut, cowpea, soybeans, yam, guinea corn and cassava) with a total population of 139,283 (Gss & Macro, 2009). A cross‐sectional design was conducted in May 2018. The study population comprised pregnant women (aged 15–49 years) utilizing antenatal care in health facilities. Pregnant women at all stages of pregnancy were included in order to have a representative measure of diet quality among pregnant women. We selected four health facilities for data collection (Savelugu hospital, Nantong clinic, Savelugu Reproductive and Child Health Centre [RCH] and Mbia Laboratory services.). The health facilities were selected using simple random sampling technique from a list of all health facilities in the district using Excel‐generated random numbers. We used the probability proportional to size (PPS) methodology to select 403 study participants from the four health facilities. Lists of pregnant women attending antenatal care clinics (ANC) in the four facilities were compiled to form the sampling frame. Using the formula for estimating sample size of single proportions (Cochran, 2007) with a 95% confidence interval (CI), an assumed prevalence of 50% of pregnant women meeting minimum dietary diversity, a margin of error of 0.05 and a 5% contingency, the sample size was 403. The dependent variable was diet quality measured using a 24‐h dietary diversity score (DDS). The independent variables included the socio‐demographic characteristics of pregnant women such as age, highest education completed, occupation, religion, marital status, household size (grouped into 1–5, 6–15 and above 15 household members who live together and eat from a common cooking pot), household wealth and gestational age (weeks of pregnancy recorded from pregnant women’s antenatal cards), household food security and nutrition knowledge. A brief description of the key study variables is as follows. Household wealth index was used as a proxy indicator for socio‐economic status (SES) of households. This was based on an earlier concept by Garenne and Hohmann‐Garenne (2003) and Saaka et al. (2017), using absolute measurement of household items such as housing quality, water availability and type of toilet facilities, and ownership of durable goods and livestock. Households were given scores depending on their possession of household assets. Household wealth index was the summation of individual scores of each household items; the scores ranged from 0 to 18. Households with wealth scores less than or equal to 13 were considered as having low wealth index score and households with wealth scores greater than 14 as high wealth index. A dichotomous variable had favourable model fit characteristics compared with the usual quantile system, which would have increased the number of parameters estimated for household wealth. Diet quality was measured using DDS because it reflects intake of nutrient adequacy (Mirmiran et al., 2006). The Food and Agricultural Organization defines dietary diversity as the number of food groups an individual eats over the past 24 h (Kennedy et al., 2011). Dietary diversity of the pregnant women was assessed using the Food and Agricultural Organization (FAO) women dietary diversity questionnaire to determine women dietary diversity score (WDDS) (FAO, 2016). Emphasis was on dietary diversity; hence, food frequency or amount of food taken was not considered. The food groups that were included in the questionnaire were as follows: (1) grains, white roots and tubers and plantain (maize, rice, yam and wheat or foods made from these, e.g., bread, TZ, porridge and banku); (2) Vitamin A‐rich dark green leafy vegetables (Amaranth, okra leaves, baobab leaves, ayoyo, alefu and bra); (3) other Vitamin‐A rich vegetables (tomatoes, carrots, orange flesh potatoes, ripe mangoes, dawadawa pulp and watermelon); (4) other vegetables (garden eggs, okra fruits, garlic, green and red pepper and onion); (5) other fruits (orange, lemon, guava, ebony fruits, blackberry and banana); (6) flesh meat (beef, lamb, chicken and any organ meat, e.g., liver and offals); (7) eggs (chicken and guinea fowl); (8) nuts and seeds (groundnuts, cashew nuts, bungu [sesame] and melon seeds); (9) pulses (beans, peas and lentils); (10) milk and milk products (milk, cheese and yogurt). Respondents were given a score of one for consuming food from a specific food group and 0 for otherwise, this was used to calculate the total women dietary diversity score (WDDS) by summing scores from all foods eaten. The maximum women dietary diversity score (WDDS) was 10 for consuming food from all food groups and 0 for not consuming food from any food group at all. Respondents who had dietary diversity score of 5 and above were considered to have met the women dietary diversity (MDD‐W), because dietary diversity was used as a proxy for measuring diet quality, respondents who met their diet quality were considered to have quality diets. The knowledge questions were adapted from the guidelines for assessing nutrition‐related knowledge, attitudes and practices (KAP) manual published by FAO (Macías & Glasauer, 2014). The questionnaires were field tested and validated in several countries. The focus of our assessment was on basic maternal nutrition that is usually taught in nutrition education sessions to pregnant women. The content included food groups and their functions, balanced diet, nutrition during pregnancy and lactation, causes and prevention of anaemia, iron absorption enhancers and inhibitors and foods sources of key nutrients (iron and Vitamin A). Maternal nutrition knowledge was measured on a continuous scoring scale. The assessment was based on nine questions on knowledge of nutrient rich foods required during pregnancy and on anaemia. A score of 1 was assigned when a correct answer was given otherwise 0 to give a maximum score of 9. Overall nutritional knowledge was rated by calculating the total of all the valid responses pregnant women made. The overall composite nutrition knowledge index ranged from a minimum of 2 to a maximum of 9. The median score was 6.0. The scores were divided into two categories: pregnant women whose overall nutrition knowledge score was less than the median score were considered to have low knowledge and those with scores of 6 and above as high knowledge. Measurement of mid‐upper arm circumference (MUAC) was used to assess nutritional status, as this measurement is independent of pregnancy and therefore it is an effective indicator for a woman’s nutritional status throughout the reproductive life. MUAC was measured following WHO standard procedure (WHO, 1995). MUAC measurement was used to classify pregnant women as normal or malnourished; pregnant women with MUAC less 23 cm were considered malnourished, and those with MUAC greater than or equal to 23 cm were considered to be normal (Tang et al., 2016). Household food security was measured using a modified FANTA household food insecurity and access scale (HFIAS) (Coates et al., 2007). The HFIAS has previously been modified to simply reflect household hunger, the household hunger scale (HHS), using three questions and three sets of responses (Deitchler et al., 2011; Saaka et al., 2017; Silventoinen, 2003). The three responses (never, rarely or sometimes and often) were scored and used to classify food insecurity/hunger. Never was assigned a score of 0; rarely or sometimes had a score of 1; and often was assigned a score of 2. Therefore, 0 was the minimum score, and 6 was the maximum score. Scores of 0–1 indicate little to no household hunger, 2–3 indicate moderate household hunger, and 4–6 indicate severe household hunger (Deitchler et al., 2011). In regression analysis, we combined none and mild household hunger and those with medium or and severe as the second to reduce the possibility of data sparsity arising from empty cells in the fully adjusted model. Questionnaire was translated to Dagbanli and pretested on pregnant women attending antenatal care at health facilities. The questionnaires were administered by four trained undergraduate nutrition officers who attended a 3‐day training on administering the questionnaire and anthropometric assessment (MUAC). Data were collected at designated places at various health facilities to ensure privacy and were then checked on site for completeness on daily basis; incomplete questionnaires were identified, and respondents were followed up to complete missing data. Data analysis was performed using SPSS version 21 (IBM Inc.). Means and standard deviations were used for continuous data. Frequencies and percentages were used for categorical data. Bivariate analyses were performed using chi‐square/Fisher exact test as preliminary analysis to identify predictors of women’s dietary diversity. Variables showing evidence of association with the outcome variable from the preliminary analysis were included for multivariate logistic regression analysis. We used the variance inflation factor (VIF) to assess collinearity among variables in the regression model, variables did not show evidence of significant collinearity (VIF < 5) (Ringle et al., 2015). Independent factors such as gestational age, educational level, household wealth index and household size were accounted for in the multivariate logistic regression modelling. Ethical clearance for the study was obtained from the School of Allied Health Sciences, University for Development Studies, Tamale, Ghana. Informed consent was obtained from each participant before they enrolled on the study.
N/A