Routine immunization status of nomadic children aged five years and below in Volta Region, Ghana in the post-COVID-19 pandemic era: a cross-sectional study | BMC Public Health

Study design

We conducted a community-based analytical cross-sectional study in the Adaklu and Akatsi North Districts from July to October 2022.

Study area and population

The study was conducted in the Adaklu and Akatsi North Districts of the Volta Region of Ghana, among nomadic children aged 5 years and below. The two districts are among the eighteen (18) Municipalities and Districts in the Volta Region of Ghana and have nomadic communities widely dispersed in almost all the sub-districts. Currently, there are seventeen (17) health facilities in Adaklu District (eleven Community-based Health Planning and Services (CHPS) compounds, four health centres and two mission facilities. Akatsi North District has a total of twelve facilities, comprising ten CHPS compounds and two health centres. There is no hospital in the two districts in the region. The main occupations of the populace in the two districts are farming, trading and cattle rearing by the nomads. Both districts have nomadic populations, and it is against this background that the two districts were selected for the study. The study population consisted of nomadic children aged five years and below residing in the Adaklu and Akatsi North Districts and their primary caregivers.

Figure 1 illustrates the study site map in the Volta Region of Ghana.

Sample size determination

The sample size for this study was estimated with the Cochran’s formula [n= (Z² Pq)/d²] [15]. The parameters in the formula include;

n = sample size to be determined.

Z = Reliability of coefficient corresponding to a 95% confidence interval (1.96).

P = proportion of fully immunized children in Ghana was 89.5% [16]. This estimate was used because the study district had similar demographic and health service characteristics with the current study area.

q = the acceptable deviation from the assumed proportion (1-0.895).

d = the margin of error around p estimated as 0.05 in this study.

Adjusting for an anticipated 5% non-response rate, a minimum sample size of 157 primary caregivers with nomadic children under five years was recruited for the study.

Sampling

Adaklu and Akatsi North Districts each consist of five sub-districts, encompassing a total of 104 and 140 communities, respectively. Of these, there are nine nomadic communities in Adaklu and eight in Akatsi North. The initial listing of nomadic communities was purposive based on official and field-level expert identification.

To ensure a systematic and representative sampling process, a multistage sampling technique was employed. This approach resulted in the selection of four sub-districts, twenty communities, and 157 participants from the two districts. The steps taken to achieve this are outlined below:

In Stage 1, sub-districts within each district were randomly selected using the lottery method. A sampling frame of all sub-districts with nomadic populations in each district was created. The names of these sub-districts were written on pieces of paper, folded, and placed in a container. A health professional then randomly selected sub-districts from the container without replacement. Two sub-districts were chosen from each district, resulting in a total of four sub-districts.

Stage 2: A sampling frame of all nomadic communities within the selected sub-districts was created. Five communities were randomly selected from each of the two selected sub-districts in both districts using the lottery method. The names of the communities were written on pieces of paper, folded, and placed in a container. A health professional then randomly selected communities from the container without replacement. Five communities were chosen from each sub-district, resulting in a total of twenty communities. Proportionate sampling ensured each selected community had an equal representative sample size.

Next, in stage 3, in each selected community, data collectors identified a central point such as a market, community center, school, or chief’s palace that served as the geographical center of the community. At this central location, a bottle was spun on the ground to randomly determine a direction in which household selection would begin. Following the direction indicated by the bottle tip, all houses from the central point to the edge of the community along that path were counted and assigned unique sequential numbers. These numbers were then written on pieces of paper, folded, and placed in a container. One number was randomly drawn to identify the first house for data collection.

After the initial household was selected, subsequent households were visited consecutively along that same direction until the required number of eligible participants for the community was obtained. If a selected house had more than one household with eligible participants, one was randomly selected using simple random techniques. If a household did not meet the inclusion criteria or declined participation, the next household was approached.

Finally, in stage 4, the participants were chosen. We recruited a mother or caregiver in each selected household who met the inclusion criteria and consented to participate. In cases where multiple households existed within a house, with eligible children, a simple random selection was done to recruit the participant.

Data collection instruments and procedures

Two trained research assistants who were fluent in the local dialect (Ewe) of the study participants, in addition to the principal investigator, collected data on sociodemographic characteristics, health system, and community-related factors using a questionnaire through interviews. The questionnaire was developed based on insights from the literature reviewed for this study (see Additional File 1). The questionnaire underwent expert consultation through discussions with the district EPI Officers at both study sites to enhance its content validity and ensure alignment with the study objectives. Furthermore, it was subsequently pilot tested among caregivers of children under five years old in a site not included in the main study to identify ambiguous or inconsistent questions before data collection. Before conducting the interviews, each primary caregiver involved in the study provided consent to participate. During the interviews, primary caregivers were asked to present the child’s health record card for review on immunization status. The health record card was used to determine the child’s immunization status. To avoid recall bias, we focused on primary caregivers with health record cards for their children.

Operational definitions

Based on previous studies [17, 18], we defined immunization status and primary caregivers as follows:

Fully immunized

For children aged 12–23 months: receiving one dose of BCG, at least three doses of OPV, three doses of pentavalent vaccine, three doses of PCV, two doses of rotavirus vaccine, and one dose of measles vaccines all administered before the child’s first birthday.

For children older than 23 months: receiving one dose of BCG, at least three doses of OPV, three doses of rotavirus vaccines, one dose of measles vaccine, plus a second dose of the measles-rubella (M-R) vaccine and one dose of the meningitis A vaccine.

Partially immunized

Partially immunized describes a child who has not received one or more of the prescribed vaccine doses based on their age, necessary for safeguarding against vaccine-preventable diseases.

Not immunized

Not immunized refers to a child who has not received any of the prescribed vaccine doses, which are meant to protect against vaccine-preventable diseases.

Primary caregivers

A primary caregiver refers to someone who has the main responsibility of caring for the child on a day-to-day basis.

Data processing and analysis

The data collected was compiled and entered using Epi-Data software version 4.0 and exported into STATA 17.0 for analysis. Data cleaning and validation were done to ensure data quality before analysis. Descriptive statistics, such as frequencies and proportions, were performed on categorical variables, while means and standard deviations were used for quantitative variables and presented in tables and charts. A binary logistic regression model was used to determine the association between the full immunization status of nomadic children (aged 12 months and above) and independent variables. Variables with p < 0.2 in the univariate regression were included in the multivariate binary logistic regression. A p < 0.05 was considered statistically significant in the final model. For post-estimation, we conducted the Hosmer-Lemeshow goodness-of-fit test to assess the model fit. The test indicated that the model fit the data adequately (p-value > 0.05), suggesting no significant difference between observed and predicted values.