Scope of the Study
The study focused on agricultural households in low-resource regions across Nigeria. Located in West Africa, Nigeria, officially known as the Federal Republic of Nigeria, lies between the Sahel region to the north and the Gulf of Guinea to the south, within the Atlantic Ocean. Covering an area of 923,769 square kilometers (356,669 square miles) and with a population exceeding 230 million, Nigeria is bordered by Niger, Chad, Cameroon, and Benin. The country operates as a federal republic and is divided into 36 states, plus the Federal Capital Territory, where the capital city, Abuja, is situated. Geographically, Nigeria spans latitudes 4° to 14° north and longitudes 2° to 15° east. It is organized into six geopolitical zones: North West, North East, North Central, South West, South East, and South South.
Source and Type of Data
To achieve the objectives of the study, data from the 2018 National Demographic Health Survey (DHS) for Nigeria was utilized. The DHS dataset encompasses a sample ranging from 5,000 to 30,000 households nationwide. This dataset not only provides national estimates but also includes data on key indicators across regions, the six geopolitical zones, states, and the Federal Capital Territory (FCT). Questionnaires were distributed to collect information from households, including men and women, which was then analyzed to extract crucial national indicators. The dataset proved adequate for addressing the primary research questions due to its comprehensive nature and adaptability for various research and assessment purposes. It contains extensive demographic, nutritional, and health information.
For the survey, a representative sample of 42,000 households was initially selected. Ultimately, data was collected from 40,427 households, 41,821 women, and 13,311 men, including 23,403 rural households. However, following data cleaning and sorting, the number of rural households was reduced to 6,514. Given that this study focuses on mothers with children under five in low-resource agricultural areas, the final sample of 6,514 rural households was utilized. Further, these rural households were categorized into agricultural (2,751) and non-agricultural (3,763) households based on the employment status of the household head. Consequently, the study specifically analyzed data from 2,751 women with children under five in low-resource agricultural areas in rural Nigeria.
Analytical Techniques
Multiple Correspondence Analysis (MCA)
Multiple Correspondence Analysis (MCA) was employed to consolidate various indicators of maternal healthcare-seeking behavior into a comprehensive index. In this analysis, a lower weighted average across all indicators signifies a lower level of maternal healthcare-seeking behavior, whereas a higher weighted average reflects a higher level of maternal healthcare-seeking behavior. The formula used for this calculation is provided below.
Where;
Z is the contingency table of the data
P is the indicator matrix of the rows
B is the Burt matrix
Q is the indicator matrix of the columns
The equation for MCA can be interpreted as follows the contingency table is decomposed into a product of three matrices. The first matrix, P, is a matrix of row scores. The second matrix, B, is a matrix of column scores. The third matrix, Q, is a matrix of row-column correlations. The row scores and column scores can be used to visualize the data in a biplot. The biplot is a scatterplot that shows the rows and columns of the contingency table. The rows and columns are connected by lines that show the correlations between them. MCA is an effective method for examining relationships between categorical variables. It helps uncover underlying patterns within the data and enables visualization through a biplot, allowing for a clearer understanding of data structures. Table 1 shows the six indicators of maternal healthcare-seeking behaviour that were adopted for this study as earlier used by [12] to develop the Maternal Healthcare-seeking Behaviour Index (MHSB).
Table 1
Maternal healthcare-seeking behaviour indicators /variables
S/N | Indicators | Definition of indicators | Modalities |
1 | Pre-natal care | Women who received care before pregnancy i.e. family planning, fertility tests, etc. | Yes, No |
2 | Timing of first ante-natal | Pregnant women are first-trimester antenatal starters and those who are not | Yes, No |
3 | Number of ante-natal visits during pregnancy | Total number of antenatal visits to the health facility before delivery | ≥ 4 = 1, \(\:<\) 4 = 0 |
4 | Place of delivery | Women in pregnancy who gave birth at medical facilities and those who delivered in other homes. | Yes, No |
5 | Assisted by a skilled attendant | Pregnant women who were assisted by a skilled birth attendant during delivery and those assisted by others. | Yes, No |
6 | Post-natal care | Women who received care immediately after delivery till 6 weeks after from a trained professional and those who do not | Yes, No |
Descriptive Statistics
Descriptive Statistics were used to profile the immunization status of under-five children. This includes frequencies, tables, and percentages. The various vaccines used by the children were used to describe immunization status of the children as shown in Table 3. Children were described as partially immunized (when a dose of any of the vaccines is skipped), fully immunized (when all the doses of the vaccines are received), and unimmunized (when all the doses of the vaccines are not received/when oral polio vaccine only is received). This approach follows [13].
Table 3
Routine Immunization in Nigeria according to the National Program on Immunization (NPI)
Vaccines | Schedule | Fully Immunized | Unimmunized | Partially immunized |
Bacillius Calmette Guerin (BCG) Hepatitis B OPV 0 | At birth | | | |
PENTA (DPT, Hep B and Hib) (3 doses) | 6,10,14th weeks | | | |
Oral Polio Vaccine (3 doses) | 6,10,14th weeks | | | |
Oral Polio inactive | 14th week | | | |
Measles | 9 months | | | |
*Partially immunized is when a dose of any of the vaccines is skipped.
*Fully immunized is when all the doses of the vaccines are received.
*Unimmunized is when all the doses of the vaccines are not received/when oral polio inactive vaccine only is received.
Extended Ordered Probit Regression Model
The Extended Ordered Probit Regression Model was employed to examine the impact of maternal healthcare utilization on child immunization status. This model is appropriate as it handles endogenous covariates, nonrandom treatment assignment, endogenous sample selection, and can be applied to panel or grouped data. The model is formulated as follows:
$$\:yi=vh\:\:\:iff\:\:\:{k}_{h-1}<{X}_{i}\:\beta\:+\:\:{Ԑ}_{i}\le\:\:{K}_{h}\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(2\right)\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:$$
The limits on the unobserved \(\:{{\epsilon\:}}_{\text{i}}\) based on the observed values of\(\:\:{y}_{i}\) and\(\:\:{X}_{i}\)
\(\:{l}_{1i}={c}_{i(h-1)}\:\:\:\:\:\:\:\:\) if \(\:{y}_{i}=\) \(\:{v}_{h}\:\) (3)
\(\:{u}_{1i}=\) \(\:{c}_{ih}\:\) if \(\:{y}_{i}=\) \(\:{v}_{h}\) (4)
the log-likelihood is written as
1nL\(\:={\sum\:}_{I=1}^{N}{w}_{i}ln{ɸ}_{i}\)(\(\:{l}_{1i}\), \(\:{v}_{1i}\),1) (5)
The conditional probabilities of success can be written For h = 1,……, H as;
Pr (\(\:{y}_{i}\)=\(\:{v}_{h}\)/\(\:{X}_{i}\)) = \(\:{ɸ}_{i}\)(\(\:{c}_{i(h-1)}\),\(\:\:{c}_{ih}\:\), 1) (6)
Y=
Y= \(\:{K}_{h}\:+{X}_{1}+{X}_{2}\:+{X}_{3}+{X}_{4}\dots\:\dots\:\dots\:.\:\:{X}_{N}+£\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(7\right)\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\)
Y1 = = Child Immunization (0-unimmunized, 1- partially immunized, 2- fully immunized)
Child’s Characteristics
X1= Age of the child ( In months)
X2 =Child’s sex. (1 if male and 0, otherwise)
X3 =Child’s birth weight (in kg)
X4 =Child’s age square (in months)
X5 = Total children (Numbers)
X6 = Birth order (Number)
Mother’s Characteristics
Y0= Maternal health care utilisation index
X7 = Mother’s educational status (1 if formal, 0= otherwise).
X8 = Mother’s age at first birth (years)
X9 = Mother’s occupation (0= Agriculture part–time, 1= agriculture full-time.)
X10 = Media exposure (1=exposed, 0=otherwise)
Household’s Characteristics
X11 = Household size (persons)
X12 = Family wealth index
X13 = Sex of household head. (1 if male, 0 if otherwise)
X14 = Husband’s educational status (1= formal, 0= no formal)
N15 = Husband’s age (years)
Region of Residence
X16 = 1 if Northcentral, 0 if otherwise
X17= 1 if Northeast, 0 if otherwise
X18=1 if Northwest, 0 if otherwise
X19 =1 if Southeast, 0 if otherwise
X20 =1 if South-south, 0 if otherwise
X21 =1 if Southwest, 0 if otherwise