Study setting and population
Institutional based cross-sectional study was conducted in Public Hospitals and health centers of West Arsi Zone. West Arsi is one of the zones in Oromia Regional State with total population of about 2,507,643. Shashemene town is the capital of the zone and located at 250 km to the South from Addis Ababa (West Arsi Zone health office statistics). The study was conducted in Public Hospitals and health centers of West Arsi Zone from March 1 to April 30, 2019. All women who have a child less than 12 months old in the Public Hospitals and health centers of West Arsi zone were our study population. Those caregivers other than the biological mother who brought the children for immunization were excluded and attempt was made to get biological mothers of the children.
Sample size and sampling procedure
Sample size was determined based on the single population proportion for cross sectional survey assuming 19.7 % [16] of the proportion of male partners had accompanied their wife’s to ANC, 49.2% of male partner involved in BPCR [17] and 26% of the institutional deliveries [6]. Finally, the minimum sample size of 804 was taken in order address all objectives
Simple random sampling was used to select five districts from the total districts of West Arsi Zone. Accordingly, shashemene, Arsi Negelle, Dodola, Adaba and Wondo Genet districts were randomly selected. The estimated number of mothers with a child of less than 12 months were obtained from zonal health office. Based on the obtained information, the sample size for each district was proportionally allocated. Then, the hospitals and health centers that are found in the randomly selected districts were listed along their average monthly client flow for immunization. Then, those public Hospitals and health centers with relatively large client load were purposively selected. Study subjects were again proportionally allocated for each selected institution.
The sampling interval was calculated by dividing the estimated mothers with a child of less than 12 months attending immunization in each institution during the study period for allocated sample size for the health institution. Finally, the study subjects were selected by systematic random sampling using every kth interval. First mother was selected by lottery method from institution and subsequent mothers were included randomly based on the interval.
Data collection instrument and procedures
The data collection was carried out using structured and pre-tested questionnaire. The questionnaire was adapted after reviewing related literatures [11, 13, 16]. The questionnaire contained closed ended questions and it was developed in English language and translated into Afaan Oromo and back translated to English to ensure consistency and accuracy. It contained socio-demographic characteristics of the women and her partner, male involvement in ANC and BPCR. Information about the place of delivery, type of attendant and decision making power were also included.
Study variables
The dependent variable is place of delivery (home versus institutional delivery). The main exposure variable was male involvement in ANC and BPCR. Other covariates were age, education, occupation of mother and husband, residence, religion, marital status, family size, number of under-five child, household assets, male accompany to ANC, HIV test, delivery, Decision making about place of delivery and Time of ANC Visit. Mothers with a child of less than 12 months attending immunization were used to minimize recall bias. The manuscript was prepared as per the STROBE guideline [25]
Data quality control
To maintain the quality of the data structured and pre-tested questionnaire was used to collect data. Pretest was done in a health center which was out of study area and appropriate corrections were made to the questionnaire. Training was given to data collectors for one day on how to conduct the interview and content of the questionnaire. Data collectors were recruited by the investigators. Questionnaires was checked for completeness during data collection and incomplete ones were sent back to the data collectors for checkup under supervision. The supervisors checked the filled questionnaire daily.
Operational definitions
Male involvement in ANC: Whether a partner had accompanied his wife for at least one ANC visit during the most recent pregnancy resulting in a live birth.
Male involvement BPCR: A male will be considered as involved in BPCR if she (his wife) reported that; her husband involved in three of the following; saving money, identified health institution with 24-hour delivery service, arranging delivery by SBA, identified place of delivery, arranged transportation, identifying early sign of emergency, and arranged blood donor for the last delivery [18].
Male involvement during delivery: Husbands physically present at the birthplace
during the time of delivery. Husbands could either be in the room with the woman while receiving the service or at another location within the premises where the service is received [8]
Decision-making: is whether the women can make decision, jointly with her partner or husband decides alone about the place of delivery and ANC attendance. For example; participants were asked “Who made the final decision of where to give birth?” The responses included “yourself”, “your spouse,” “jointly with your spouse,” “other family member,” “health professional,” “friend,” and “others”
Skilled Birth Attendance: A skilled birth attendant is a medically qualified provider with midwifery skills (Midwife, Nurse, Health Officer or doctor) who has been trained to proficiency in the skills necessary to manage normal deliveries and diagnose, manage, or refer obstetric complications.
Institutional delivery service utilization: a woman who gave birth in the health institution for the most recent delivery being assisted by skilled birth attendant
Data processing and analysis
Data was checked for completeness and entered in to Epi info version 7.0. It was exported to SPSS (statistical package for social sciences) version 21.0. Then cleaned, labeled and recoded in SPSS. Descriptive analysis was used to summarize the socio-demographic characteristics and the level of male involvement. Male involvement in ANC, BPCR and delivery care were assessed based on the women reports. Variables for which p-value is less than 0.25 by binary logistic regression analyses were entered into the model for subsequent multivariable logistic regression analyses. Involvement of the male in BPCR was considered as the primary exposure variable. Place of delivery was a dichotomous variable (home vs health institution) and considered as a dependent variable.
Principal component analysis (PCA) was used to compute wealth index. Seventeen household assets were used in computing PCA and all the assumptions (Correlation, KMO, Bartlett’es test of spherecity, MSA (measure of sampling adequacy), eigen value) were checked. Wealth index was divided into three categories as lowest, medium and highest. Multiple logistic regression model was done to identify the associations between husbands’ involvement in BPCR and women’s utilization of institutional delivery after controlling for other covariates. Stepwise forward logistic regression method was used for selection of variables in the model. Multi-collinearity was checked using standard error. Strength of the association was done using odds ratio (OR) with 95% confidence intervals (95% CI) at p value less than 0.05.