Study design and setting
Community based cross-sectional study was conducted in Mekelle. Mekelle is the capital city of Tigray region, found in northern Ethiopia. It is located around 780 kilometers north of the Ethiopian capital, Addis Ababa. Administratively, it is considered as special zone in the region, which is divided into seven sub-cities. These are Addi-Hak'i, Ayder, Haddinet, Hawelti, Qedamay - weyyane, Quiha, and Semien. The city is an economic, cultural, and political hub in the northern region of the country. According to projected central Statistical agency of Ethiopia, 2015 It has a population of more than 323,000 among these populations, 110,788 are females, 104,758 males and around 60,998 women are in reproductive age (15-49) years.
Study population and eligibility criteria
The study was conducted among husbands of pregnant wife or having less than one year child and who lived for 6 and above months before the study in the selected kebeles of Mekelle city. From each household unit one eligible husband was included and husbands who were unable to communicate due to disability or any other problems were excluded from the study.
Sample size and sampling procedure
The required sample size was determined by using a single population proportion formula; with 5% desired precision, 95% confidence level, 42.2% husband awareness of danger signs of obstetric complications (9), 10% of non-response rate, and 1.5 design effect was considered. The total sample size after computing for 10 % non-response rate and 1.5 design effects was 620. Multistage sampling technique was used to select study husbands. From the 7 sub cities in the Mekelle city, two sub cities were selected by lottery method which was Ayider and Hawelti. Each sub cities have five kebeles (the smallest administrative unit in Ethiopia) and all kebless with in the sub cities were included in the study. On selected sub cities before the actual data collection pre-survey was conducted to determine the number households. The households’ Ayder and Hawelti sub cities were 3898. A total of 620 households were proportionally allocated to Ayder (279 households) and Hawelti (323 households) sub cities. The first household was identified by health extension workers the remaining study households were included by systematic random sampling.
Data collection tools and procedures
A structured questionnaire was adapted and used from the survey tools developed by JHPIEGO Maternal Neonatal Health Program (2). Additionally, to address for all research questions further related literatures were used to develop the structured questionnaire. The questionnaire comprises 3 parts. The first part was about socio-demographic characteristics of the study participants. The second part was about level of awareness on obstetric danger signs on three phases with the source of information and the perception aspect included the importance of knowing danger signs and health seeking behavior of husbands. Ten BSc graduates’ midwifery students collected the data and two MSc midwifery students were recruited as supervisor. Moreover, four health extension workers were recruited as supporter for data collectors.
Data quality measurement
To maintain the quality of data; data collectors, supervisors and health extension workers were trained by authors for two days. The questionnaire was first prepared in English and then it was translated into Tigrigna (a local language and regional language for the study area) together with language experts for better understanding by respondents. Tigrigna version of the questionnaire was then translated back to English to check for its consistency. Additionally, definition of concepts and terms were harmonized with a local language of the district to avoid ambiguity. Pretest was done on 15 (5%) husbands at nearby town (wukro) in order to assess consistency and meaning of the instrument. Furthermore, collected data were checked by supervisor and principal investigator every day for its completeness.
Study variables
Awareness and perception on obstetric danger signs are dependent variable. Whereas, variables like Socio-demographic characteristics of the husband (age, educational level, household income, ethnicity, occupation and eligion), Socio-demographic characteristics of the wife’s (educational status, occupation of wife’s and age), obstetrics factors (place of last delivery, number of children and number of ANC visit) and source of information (health care providers, media , family and friend) were independent variables.
Operational Definition
Obstetric danger signs: these are signs and symptoms of obstetric complications which occur during pregnancy and childbirth and immediately after delivery.
Good awareness: refers to those participants who respond correctly to awareness questions and scored above the mean value.
Poor awareness: refers those participants who correctly respond to awareness questions and scored equal or below mean value.
High perception: refers to those participants who respond to perception questions and scored above the mean value.
Low perception: refers to those participants who respond to perception questions and scored less than mean value.
Data management and analysis
The data was entered and cleaned using Epi info (epidemiological information) version 7 and analyzed using SPSS 22.0. Data cleaning was done by running frequencies, cross tabulation and sorting among various variables. Results are presented in tables and figures by their frequencies and percentages. Both binary and multivariable logistic regression model was done to identify factors associated with awareness and perception towards obstetric danger signs. All factors with a p-value <0.25 in the binary logistic regression analysis was further entered into multivariable logistic regression to control confounding effects. Multiple logistic regressions were used to estimate the adjusted effect size of factors on awareness and perception towards obstetric danger signs. Magnitude of association was measured using odds ratio at 95% confidence interval and statistical significance was declared at p-value less than 0.05.