Study design
A community-based cross-sectional study was conducted to assess the prevalence (point-prevalence) of marital dissolution among Hosanna residents.
Study setting
This study was conducted in Hosanna town. Hosanna is the administrative and commercial center of the Hadiya Administrative Zone in the Southern regional state of Ethiopia. It is located 232 km South West of Addis Ababa. As for demographic characteristics, the population of Hosanna has been rapidly growing since its inception. According to the housing and population census of Ethiopia, the population of the Town was 13,467 in 1984, and (31,701) ten years later in 19,944 [31]. In 2007, Hosanna had a population of 69,995 people [31]. According to Hosanna Municipality and the Ethiopian Demographic and Health Survey [31], the town of Hosanna is divided into six Kebele (smaller administrative units in Ethiopia). We used this pre-arranged structure of the town for the current study.
In Ethiopia, only Twenty-seven percent of women aged 15–49 have never married and 11 percent are divorced. Correspondingly, less than one percent of women aged 45–49 have never been married indicating that both marriage and marital dissolution are universal in Ethiopia [32]. Recruitment of participants and data collection were carried out from February 1 to March 30, 2022.
Participants
Individuals who lived in Hosanna town and were able to answer specific marital questions were included in this study. We used marital-specific questions such as the “current marital status,” “the prime reasons for getting married”, “age at marriage”, and “parental marital dissolution history”, to consider people as eligible for this study. These variables were considered because they were associated with marital dissolution, which was the outcome variable of this study. The prospective study participants were recruited from the framed source population. Those who met the eligibility criteria were included in the study population. We obtained the data of details of the existing government structure, the number of Households, and the total number of people living in each administrative unit of the Town from Hosanna Town Municipality, Kebele Administrations, and Urban Health Extension workers. The research team included field supervisors and data collectors.
The first eligibility criterion was living in Hosanna Town. Since this study was carried out in Hosanna, people who have been living in Hosanna Town for at least six months were included in the current study. We assessed this criterion by asking participants. The ever-married individuals with a marriage history: currently married (in marital union during the data collection period), divorced, or widowed were included in the study without age and sex restriction. In contrast, a few individuals who refused to participate in the study and were unable to provide adequate information due to health problems at the time of data collection were excluded from the study.
Variables and definitions
The prevalence of marital dissolution is the outcome of this study. The prevalence of marital dissolution was determined from the study sample. In other words, we estimated the point prevalence of marital breakdown (divorce or sedation) from the study sample that was identified during the data collection period. The marital specific characteristics prime reasons to get married, age at marriage, type of marriage, forms of marriage, parental marital dissolution history, General health status (presence of known chronic illness), behavioral correlates(substance use, Communication problem, conflicting behavior, marital commitment), and socio-demographic factors(sex, religion, level of education, house ownership) were the independent variables.
In this study, marriage was defined as the legal or formal union of two people (of the opposite sex), a man and a woman, as partners in a personal relationship. Moreover, divorce was defined as the legal termination of a marriage, the separation of husband and wife which confers on the parties the right to remarriage according to the laws of each country [33, 34]. Whereas, separation refers to the termination of a marriage on the basis of civil, religious, and/or other traditional provisions without conferring on the parties the right to remarry [33, 34]. Correspondingly, marital dissolution is defined as the termination of a marital relationship as a result of divorce or separation [33, 34].
Data sources and measurement
In this study, we collected raw data directly from the study participants. Thus, the study participants were the data sources. Data were collected using a series of forms completed using face-to-face interview techniques. The form includes demographics, general health, and marriage-specific characteristics.
Data were collected using a structured questionnaire. A questionnaire was developed by reviewing relevant literature [22, 23, 33,34,35,36,37,38,39]. It was prepared in English and translated into Amharic. The questionnaire had three parts, the first part contained four questions and was about the socio-demographic characteristics of the individual participants, the second part contained three questions designed to collect data on general health and health risk behavior, and the third part contained eleven items designed to assess the marital specific characteristics and marital dissolution.
The questionnaire was pre-tested on 10% of the sample in an adjacent Town (Durame, capital town of Kembata Zone) where the study didn’t take place. Therefore, a modification was made based on the results of the pre-test, and the modified version was used for actual data collection. The internal consistency of the items in the questionnaire was acceptable with the value of Cronbach’s alpha (0.71), exceeding the index of 0.7 [40, 41].
Prior to data collection, the data collectors and supervisory teams contacted officials of each administrative unit with official letters. The households and individual participants were selected with the help of a guide from the respective administrative units. After contacting these individuals, the details of the concern of the team were explained to each participant by the assigned team leader, and the process of informed consent was secured.
Prior to the participant recruitment and data collection process, the research team received two days of training (basic principles of research ethics, data collection tools, and the roles and responsibilities of each team member). Six first-year health science students collected data. Two public health professionals were assigned to supervise the data collection process. The investigators of the research project coordinated the entire fieldwork.
Bias
Depending on the study design different techniques were undertaken to ascertain both the selection and information bias in this study. In the selection stage, the study participants were randomly selected based on the pre-determined criteria and included in the study. During the data collection process, revisits were scheduled to complete the missed data and reduced the information bias [42]. Furthermore, several individuals participated in the data collection to enhance the depth of the findings. The training was given to data collectors to familiarize them with the local culture, research instruments, and principles of research ethics. Statistical procedures were also performed to treat information bias due to missing data [42].
Study size and sampling techniques
We used a single population proportion formula to determine the number of individuals to be included in the study [43] that is appropriate for the estimation of a single proportion [44]. The proportion of marital dissolution was obtained from a previous study (45%) [20], estimated with 95% confidence and 5% precision, and took into account a 20% non-response rate to determine the sample size. Consequently, the sample size was 459.
This sample size was proportionally allocated to each Kebele based on the number of households in each Kebele (Fig. 1). The town of Hosanna is divided into six kebels. We used this pre-arranged structure of the town to frame the current study. We used the STAT CALC program of the EPI INFO statistical package to calculate the sample size.
A simple random sampling technique was employed to select the study participants. The list of Households in each sub-cities that were documented in the respective administrative units was used to select the study participants. We first randomly assign a numeric code from one to six to each of the six sub-cities. Arada 1 is the first English alphabet letter, Bobicho 2, Heto 3, Sech Duna 6, and so on. Depending on the number of Households in each Sub-cities, we assigned four-digit alpha-numeric codes for each household. For example in Sech Duna Kebele there are three thousand six hundred and twenty-nine registered Households, so, the code for the first Household was 6 and the code for the last Household was 30,629. Correspondingly, there were forty-thousand two hundred and fifty Households in Bobicho Kebele, the code for the first Household was 2 and the code for the last Household was 40,250. A similar procedure was followed to code and select study participants in all Kebele.
Finally, among all households, randomly selected households using computer-generated numbers were included in the study. In situations where there was more than one person who satisfy the inclusion criteria in a given household, an individual was selected using a simple random technique (lottery method) and included in this study These codes, which we posted on the existing family archives, were removed as soon as the data collection process was completed.
Statistical methods
The collected data were coded, cleaned, and entered into IBM SPSS 25 (International Business Machines Corporation (IBM) Statistical Package for the Social Sciences (SPSS) for Windows version 25 for analysis. We described the sample using frequencies, percentages, and diagrams. The distribution of the data set was tested using statistical tests (Shapiro–Wilk test, and Kolmogorov–Smirnov test) and graphical (histogram) methods [45].
We used logistic regression (bivariate and multivariate) analysis to assess whether there was a significant association between the associated factors and dependent variables. Before fitting the final model and reporting the results, we performed the necessary evaluations, including Multicollinearity and goodness-of-fit tests. Therefore, the Variance Inflation Factors (VIFs) test was used to assess the Multicollinearity among the independent variables, and those that showed no Multicollinearity were fitted to the multivariable logistic regression model through a backward stepwise method to reduce the effects of cofounders. The variables with a p-value of < 0.2 in the Bivariable analysis were considered for multivariable logistic regression analysis.
The Hosmer–Lemeshow goodness-of-fit statistic was used to check the model fit. We used the adjusted odds ratio with a 95% confidence interval to examine the strength and direction of the association between the independent variable and the outcome variable. A P-value of less than 0.05 [46] was used to define statistical significance. Finally, the findings were presented in the form of tables, graphs, and text. STATA 14 software package (Stata Corporation, College Station, Texas, 77,845, USA) and IBM SPSS 25.0 was used for data analysis.