The study was conducted in Harar Urban Health and Demographic Surveillance System (Harar Urban HDSS) which is located in Harar town, Harari region, Eastern Ethiopia. Harar town is located at a distance of 510 kms from Addis Ababa, the national capital and it is the capital city of the Harari region. The region is boarded with different districts of the Eastern Hararghe zone of Oromiya regional state and divided in to 36 kebeles (19 urban and 17 rural kebeles) . According to the 2013 population projection, the total population of the region for the year 2017 was estimated at 244,711 of which 49.5% were females and 55.7% were urban dwellers .
Originally Harar Urban HDSS was established in Harar town to represent the Eastern part of Ethiopia and Harar town was selected among other towns in the Eastern part of Ethiopia due to the diversity of its population in ethnicity (there are around 50 ethnic groups in the town) and diversity in terms of religious affiliation of the population . Moreover, 12 out of 19 kebeles (sub- districts, smallest administrative unit in Ethiopia) in the town are included in Harar Urban HDSS.
In 2013 the total population of Harar Urban HDSS was 30,055 (52.2% females and 47.8% males) and the sex ratio was 91.4%. Crude birth rate was 20.3 births per 1000 midyear population, general fertility rate was 64 births per 1000 women of reproductive age and total fertility rate was 1.9 births per woman in 2013 .
A community based matched case-control study was conducted between September 2018 to March 2019 in Harar Urban Health and Demographic Surveillance System.
Cases: A married couple of whose wives were in the reproductive age (15-49 years) and non-pregnant during the data collection period and who were found effective communicator.
Controls: A married couple of whose wives were in the reproductive age (15-49 years) and non-pregnant during the data collection period and who were found non-effective communicator.
Matching: Cases and controls were matched by date of birth within ten years interval.
Sample Size and Sampling
The sample size was computed using PS (power and sample size calculation) software. The sample size was calculated by assuming: the prevalence of contraceptive use among controls to be 50% ; 1.3 times higher odds (OR) of contraceptive use among couples who had effective communication about family planning ; power of 80%, a 5% significance level, a 1:1 control to case ratio; a correlation coefficient (Փ) for exposure between matched case and control being unknown and hence considering a 0.2 (Փ) coefficient, as suggested by Dupont, 1988, a minimum of 1146 cases and 1146 controls were required for the study. This sample was distributed among the 12 kebeles according to the proportion of currently married couples whose wives were in the reproductive age and non-pregnant at the survey. All the 12 kebeles of Harar Urban HDSS were included in the current study. The complete list of currently married couples whose wives were 15–49 years (reproductive age) and non-pregnant was taken from Harar Urban HDSS database and used as a sampling frame. Married couples were then randomly sampled from each kebele, based on a computer generated random number list using the allocated sample size.
Data were collected using structured questionnaires. For male and female respondents, separate questionnaires were administered. But the contents of the questionnaires were similar which include socio-economic and demographic variables such as age, duration of current marriage, ethnicity, religion, education, occupation, number of living children, desired number of children, household ownership of assets and household income; knowledge/attitude to family planning variables such as knowledge of family planning methods, attitude towards family planning, attitude towards large family size and approval of family planning; couples’ participation in family planning use decision and their exposure to family planning message through mass media such as radio, television, newspapers, posters, pamphlets, etc and through interpersonal communication with health personnel, friends, relatives, neighbors. And couples’ family planning communication and their ever and current modern contraceptive use were also measured.
The survey instruments were adapted from a validated questionnaire and were considered reliable . Questionnaires were pre-test among one percent of the total sample in an area other than the study site, but with a similar set-up. Twelve male and twelve female data collectors participated in the study and were supervised by two field coordinators. Data collectors were recruited from the local community. The interview was conducted in a private location, each couple at a time (first woman and next man) but separately keeping interviewee privacy. The interview was conducted if both spouses agreed to participate in the study.
Spousal family planning communication was measured based on the question "Have you ever discussed about family planning with your husband/wife in the last 12 months?" Hence, the responses were coded as 1 if both spouses agree they discussed or the husband alone reported ever discussed or the wife alone reported ever discussed and 0 if both spouses agree they never discussed .
Effective spousal family planning communication was measured by taking three variables, i.e., spousal family planning discussion in the last 12 months prior to the survey date; couples' approval or disapproval of family planning and spouses' perceptions about their partner's approval of family planning and couple's fertility desires which were defined through their responses regarding the number of children that each spouse would choose to have for his or her entire reproductive life (ideal family size). Hence, effective communication about family planning among couples exists when couples discuss about family planning in the last 12 months prior to the survey date; when husband's perception to his wife's family planning approval matches with wife's response to family planning approval question and when wife's perception to her husband's family planning approval matches with husband's response to family planning approval question and when responses of couples' to their desired number of children concords . Couples' who communicated about family planning were coded as 1 otherwise 0; When husband's family planning approval response matches to his wife's perception about his family planning approval, it is coded as 1 otherwise 0; when wife's family planning approval response matches to her husband's perception about her family planning approval, it is coded as 1 otherwise 0; when the ideal family size of husband and wife matches, it is coded as 1 otherwise 0. Using these three variables composite variable is created, effective communication and non-effective communication. Those couples scored 1 for all variables were effective communicators otherwise non-effective communicators.
Wealth index was calculated using the principal component analysis (PCA) method. Items were assessed owing to household facilities including ownership of house, materials of the floor of the house, materials of the wall of the house, materials of the roof of the house, electricity, TV, Radio, Refrigerator, Chair, Table, Sofa, Electric Mitad, Bicycle, Motorcycle, Car, Sewing machine, Source of drinking water, Kinds of toilet, Fuel for cooking food and the number of animals owned (Cows, Oxen, Mules, Goats, Sheep, Camel, Donkey, Chickens). Nine components were extracted based on examination of Eigen-values, scree plots and the cumulative proportion of variance explained by each component. The first component consisting of several heavily loaded variables and accounted for the largest variation in the data was categorized into quintiles. Each household falls into a category with lowest score representing the poorest and the medium score and the highest representing the richest households .
The data were double entered, validated and cleaned using Epi-Data Software version 3.1 and analyzed using stata version 12. Simple descriptive analysis was done to explore levels of effective spousal family planning communication by various socio-economic, demographic, knowledge and attitude towards family planning. Bivariate analysis used to assess the association between socio-economic, demographic, knowledge and attitude towards family planning and effective spousal family planning communication. We applied conditional logistic regression model for estimating odds ratio and to identify predictors of effective spousal family planning communication. Variables significant at P-value of <0.05 in the bivariate analysis were considered in the multivariable analysis and odds ratio (OR) along with 95%CIs were estimated and a P-value <0.05 was used to declare the statistical significance. Multi-collinearity between each explanatory variable included was checked using variance inflation factor (VIF). Accordingly couples' frequency of family planning discussion and men's family planning counseling by health workers were not included in the model due to multi-collinearity with couples' family planning discussion with their friends/relatives and women's family planning counseling by health workers, respectively.
This study was conducted according to the principles of the declaration of Helsinki. Ethical clearance was obtained from Institutional Health Research Ethics Review Committee of College of Health and Medical Sciences, Haramaya University (IHRERC/158/2018, dated 30 May 2018). To safeguard the autonomy of the study participants, objectives of the research was clearly communicated and informed, voluntary, written and signed consents were obtained from the study participants prior to the data collection. To maintain anonymity and confidentiality, names of the study participants were not mentioned in the questionnaires. No person had access to the information collected from the study participants except research team. Privacy of the study participants was maintained during the interview.