Study design, area, and period
A community-based unmatched case-control study was conducted in Adama district, central Ethiopia, from March 1 to April 30, 2023. Adama district is bordered by the Arsi zone to the south, the Lome district to the west, the Boset district to the east, and the Amhara Regional State to the north. The district is comprised of 4 urban and 36 rural kebeles, situated approximately 100 km from Addis Ababa, the capital of Ethiopia. According to the regional health office, the estimated total population is 229,237 (115,567 males and 113,673 females). Among this population, women of reproductive age account for 49,614, and under-five children account for 36,083.
Study population, and eligibility criteria
The study population consisted of all women of reproductive age who had experienced at least two consecutive births in the past five years. Cases were defined as those with a birth spacing of less than 33 months, while women of reproductive age with a birth spacing between 33 and 59 months (including 33 months and up to 59 months) were considered controls. The study included women of reproductive age who had given birth within the last 5 years. Women who were seriously ill, unable to communicate, or mentally ill during the data collection period were excluded from the study.
Sample size, and sampling procedure
The sample size was calculated using Epi info version 7.2 statistical software, employing an unmatched case-control formula. This was done under the assumptions of 80% power, a 95% confidence interval (CI), and a 1:1 case-to-control ratio. Five determinants of suboptimal birth spacing practices identified in previous studies, such as residence, maternal education, wealth index, status of the index child, and postnatal care use after the previous birth, were used as determinant variables(21–23). Under the assumptions, the variable 'postnatal care use after the previous birth' yielded the largest sample size, which was 344. After accounting for a design effect of 1.5 and a 10% non-response rate, the sample size consequently became 568. As a result, a total of 284 cases and 284 controls were included in this study.
A multistage sampling approach was employed to select study participants. Initially, the district was categorized into rural and urban kebeles. Subsequently, 10 rural and 2 urban kebeles were randomly chosen from a total of 4 urban and 36 rural kebeles in the district. In the first phase, a preliminary survey was conducted to compile a list of all eligible women of reproductive age who had given birth from all health posts' family folders to identify the cases and controls. During this survey, each woman of reproductive age, having had two or more successive births and residing in the same household, was individually registered. Following this, a proportional allocation was implemented to determine the required sample size from each kebele. Finally, cases and controls were selected from the respective study population using a simple random sampling technique, with the household list serving as the sampling frame.
Study variables
Dependent variable
Suboptimal birth spacing (Yes/No)
Independent variables
Socio-demographic: maternal age, maternal educational level, husband educational level, residence, religion, ethnicity, occupation of husband, maternal occupation, family size, wealth index.
Maternal, obstetrics, and reproductive health services: parity, age at first birth, postnatal care, antenatal care, knowledge of optimal birth spacing, contraceptive use, place of previous birth, and planned pregnancy.
Child and child-related characteristics: sex of the index child, survival status of the index child, breastfeeding duration.
Operational definitions
Optimal birth spacing: The period between pregnancies that allows the mother time to recuperate from pregnancy, labor, and lactation, with inter-birth intervals ranging from 33 to 59 months(1).
Suboptimal/ short birth spacing: Inter-birth intervals of less than 33 months(1).
Knowledge: Knowledge of optimal birthing spacing was assessed using twelve multiple-choice questions. The total knowledge score was dichotomized into inadequate and adequate categories, with a score greater than 60% considered adequate(18,24,25).
Index child: The child born immediately before a subsequent child(25,26).
Data collection procedure and quality control
Data were collected using a pretested, structured, interviewer-administered questionnaire. The questionnaires were adapted from various relevant literature with necessary modifications tailored to the specific context of the study. A team of 8 trained data collectors and 2 supervisors was enlisted for the data collection process. Throughout this phase, continuous supervision of data collectors took place, and regular meetings were convened among the data collectors, supervisors, and investigators. Additional visits were conducted for participants who were unavailable during the initial visit. The collected data were reviewed and checked for completeness before data entry.
The household's wealth status was assessed using the equity tool, incorporating asset variables such as electricity, electric appliances, refrigerator, television, radio, etc. Principal component analysis was employed for the analysis, with each wealth variable categorized as 0 (no) or 1 (yes) prior to the analysis. The suitability of the data for principal component analysis was confirmed through checks of both Kaiser-Meyer-Oklin (KMO) and Bartlett tests. The wealth status was subsequently categorized into five groups, ranked from the poorest to the wealthiest quintile. Further analysis categorized participants into the first, second, third, fourth, and fifth quintile groups, which were then transformed into three categories representing lower, middle, and higher wealth status.
Data processing and analysis
Following the coding and inputting of data into Epi-Info version 7.2, the data were exported to the Statistical Package for Social Sciences (SPSS) Version 26 for cleaning and analysis. Descriptive statistics were employed to present key characteristics of the study population. The association between independent variables and suboptimal birth spacing practice was modeled using binary logistic regression analysis. In the bivariable logistic regression model, a significance level of 0.25 was set as a threshold to select variables for multivariable logistic regression analysis, aiming to control confounding effects. The existence of multicollinearity among explanatory variables was explored using the variance inflation factor along with standard error. The multivariable logistic regression utilized adjusted odds ratios (AOR) with a 95% confidence interval (CI) to identify factors independently associated with the suboptimal birth spacing practice. The model was fitted using the standard model-building approach. Hosmer and Lemeshow's goodness-of-fit test was used to assess the model's fitness In the final model, variables with a p-value less than 0.05 were deemed statistically significant.