A community-based cross-sectional study among pregnant mothers was conducted on 500 study participants from September 11 to February 2020. The target population was all pregnant women in East Borena Zone. The source populations for the study were all pregnant mothers in Liban District whereas the accessible population was all pregnant women within reproductive age groups living in 5 randomly selected rural and 1 urban Kebeles. In addition to this, all pregnant women who lived in the district for more than six months were included if she was volunteer to participate, and no problem to do so.
4.4. Sample size determination and Sampling procedure
The sample size was estimated using single population proportion formula, considering the 46.5% of prevalence of under-nutrition among pregnant women from a study done in Jimma Town(28)Other parameters considered are a 5% margin of error, 95% CI, 10%
= X 0.465(1- 0.535) = 382
Where n = Sample size Z α ⁄2 = Z value corresponding to a 95% level of significance = 1.96
p = expected proportion of practices of mothers on nutrition during pregnancy = 50% =0.5
d = absolute precision (5%).
Therefore, from the above sample size is: 382
n = 384 and none response rate = 10% which is 38, n = 382 + 38 = 420
Sample size determination for specific objective two
Insert Table 1
Table 1
sample size calculation for second specific objective
Variables | Magnitude | Power CI level | AOR | Sample size | Reference |
Exposed | Non exposed |
Average monthly income of HH | 15.83% | 4.8% | 80%, 95% | 8.72 | 73 | (12) |
Decision making autonomy of pregnant women | 11.09% | 1.81% | 80%, 95% | 2.7 | 160 | (26) |
Work load on women | 28.8% | 6.3% | 80%, 95% | 13.6 | 46 | (26) |
Practice frequent hand washing habits | 14.20% | 3.02% | 80%, 95% | 6.55 | 68 | (12) |
Educational status | 2.91 | 0.77 | 80%, 95% | 1.50 | 131 | (11,59) |
Sampling Procedures:
First, kebeles in the district were stratified into urban and rural areas (kebele is the lowest governmental administrative structure in Ethiopia). The sample size was proportionally allocated for each stratum and then, representative pregnant women were randomly selected..A random sampling technique was utilized to select 10 Kebeles out of 35 total Kebeles. Finally, 420 samples were allocated proportionally to each selected Kebeles based on their total number of pregnant mothers. The calculated sample size of 500 were proportionally allocated to randomly select 5 health post out of 20 in the Liban district and two health center out of five based on the number of clients attending antenatal care (4) at health post and health center. Then every seven pregnant women, as registered, were included in the study at each antenatal care unit till the desired sample size is achieved (14).
Insert Fig. 2.
The dependent variable of this study was the nutritional status of pregnant women. Independent variables were Socio-demographic characteristics of the pregnant women like age, marital status, education, religion, ethnicity, residence area, family size, income, women's decision-making autonomy, Intra-house holds violence and polygamy. Reproductive, medical, and behavioral characteristics of the study participants like age at first marriage/ pregnancy, trimester of pregnancy, pregnancy intention, gravidity, parity, abortion, inter-pregnancy interval, and recent illness in the past 15 days and substances abuses are independent variables considered in this study. Others are health care and environmental characteristics such as accessibility to health care, prenatal dietary advice, antenatal care follow-up, drinking water source, and latrine possession. Dietary characteristics of the study participants such as Minimum Dietary Diversity of Women, household food insecurity, improved dietary feeding, skipping meals/snack,s and eating an additional meal are also independent variables included in the study.
Data collection tools and procedure:
During data collection, face-to-face-interview, observation, anthropometric measurements, and standard checklists were used to collect data from pregnant women after the interviewers explained the purpose of the study and obtained the participant’s verbal consent to participate in the study. In this study, MDDW was measured by the FAO-2016 standard checklist developed for this purpose which is recommended for 24 h dietary recalls. Household food insecurity was measured by the FANTA-2007 standard tool that has nine questions with each comprising 3 responses; 27-score-based HFIAS scale. Undernutrition was measured by MUAC (in cm) on their left arms at the midpoint between tip of the shoulder (olecranon process) and tip of the elbow (acromion process) and insertion type of MUAC tape was Benonelastic and non-stretchable to take the value with correct tension (not too loose/tight) with nearest 0.1 cm reading. Age, age at pregnancy, and inter-birth time were approximated to local memorable events. The participants were nutritionally accessed via 24-hour recall. Additionally, anthropometric assessment MUAC measurement was involved. A structured questionnaire was developed and adopted from Mini-EDHS 2019, the food frequency questionnaire, and WHO standard.That all the variables to be assessed were incorporated(8).
Data quality control all study instruments were translated into local languages ( by native speakers and then back-translated to English by two other competent persons. Six interviewers and two supervisors were recruited for the survey and were trained on the overall data collection process. All the six data collectors are BSC midwives and the supervisors are senior public health experts with master of public health degrees who are competent in local languages. Completeness and consistency of data were assured through direct and daily supervision by the supervisors and principal investigators. Interviewers re-administered the questionnaire to the respondent under supervision by the supervisor.
To ensure the quality of data, training of data collectors and supervisors was undertaken questionnaires will also be translated into the local language to facilitate understanding of the respondents. In addition to written documentation of responses from study participants, tape recordings were done after obtaining verbal consent to ensure that all feedback are captured for analysis.
Data collectors and supervisors were selected based on their educational background (particularly those who have received training on essential nutrition actions), work position, and experience of data collection. Supervisors and data collectors were trained on the objective and methodologies and data collection techniques of the study. Daily discussions and check-ups of data completeness were made with supervisors and the principal investigator. The data cleaning and entry were conducted exclusively by the principal investigator. The questionnaire was pre-tested among 5% of the total sample size to assess its clarity, length, completeness, and consistency. After the pre-test was conducted adjustments were done according to enhance the reliability and validity of the tool. The structured questionnaire was then rephrased in light of the responses. Test-retest reliability was established by examining the consistency of pre-test responses using and the three main components of the test-retest method are as follows Test-retest reliability of the research instrument was established during pretesting. Pretesting was done on two occasions but on the same respondents, on Monday and Friday: assume there is no change in the underlying condition (or trait you were trying to measure) between test 1 and test 2. And finally, compute the correlation between the two separate measurements and if test 1 and test 2 have become consistent, the questionnaire was considered reliable
The collected data were checked for incompleteness and inconsistency. Data were entered into Epi-Info version 7.2 software and then, exported to SPSS version-21 for analysis. Prior to running for analysis, data were cleaned, composite indexes were computed and recorded aver missing values, and extreme values were identified and trimmed. Descriptive statistics were used to describe the sample accordingly. Bivariate logistic regression was carried out to see the association of each independent variable with acute under-nutrition and those with p- values below 0.25 remained in the final models (multivariate logistic regressions). Odds Ratios (OR) were generated for each variable and the independence of any association was controlled by entering all variables into the model using the backward step-wise method. The magnitude of the association between the independent variables in relation to acute under-nutrition was measured using adjusted odds ratios (AOR) and 95% confidence interval (CI) and P-values below 0.05 were considered statistically significant. Descriptive statistics were used to show Socio-demographic characteristics and the prevalence of nutritional practices.
Logistic regression analysis was used to identify the association between factors and the nutritional status of pregnant mothers and multivariate logistic regression were performed to determine independent predictors of the nutritional status of pregnant mothers. A p-value < 0.05 was considered statistically significant. VIF and tolerance tastes were checked for the presence of multidisciplinary among the independent variables. Step-wise model building strategy with p-value = 0.05 was applied to identify independent predictors of nutritional practice and the Hosmer-Lemeshow Test of Goodness-of-Fit was used to test how well the model explains the data. Adjusted Odds Ratios and their 95% Confidence Intervals were reported. Additionally; tables and figures were used to present the findings.