Study area and period
The study was carried out in Abuna Gindeberet district from February to March 2012. Abuna Gindeberet is one of the districts in west shoa zone of Oromia regional state. The administrative town of this district is located 176 KM from Addis Ababa. Based on the national census of 2007, the district had an estimated total population of 126,049 of whom 64,285 and 61,764 were males and females, respectively. The district has 43 ‘kebeles’ of which two urban and 41 rural ‘kebeles’. All the ‘kebeles’ have health extension workers. According to information from district health office, there were about 91 health extension workers and 51 other health workers rendering the health service to the community at large. The ratio of health extension workers to households was 1:289.
A community based cross–sectional study was conducted on randomly selected heads of households who fulfilled the following inclusion criteria: (1) households living for more than one year in the study area and (2) households’ heads age of 18 years and older.
Sample size determination
Sample size was determined by using single population proportion formula n = (Zα/2)2 × P (1 - p)/d
2] with the assumption of 50% proportion since no previous study done, at 95% confidence interval, margin of error 5%, 10% none response rate and design effect 2. Finally, the required sample size was 806 head of Households.
The study had employed multistage sampling technique. By using simple random sampling technique, 10 from 41 rural ‘kebeles’ were selected and the sample size was distributed proportionally to the size of their households. By considering households registration numbers from registration book as a sampling frame, Systematic random sampling technique was employed to select households from selected kebeles. The study participants were selected every eight (8th) household intervals, by dividing the total number of households to the allocated sample size.
Revisit of two times was made in case where eligible respondents where not available by the time of the survey. In case where illegible subjects in the selected households where not available the next household was interviewed.
Health extension service utilization.
For this study, the independent or predictor variables used were age, sex, marital status, religion, educational status, occupation and family size, knowledge of the community on health extension package, community involvement/participation in planning of health extension activities and graduation of model family. The detail descriptions of some variables were as follow;
Age refers to the respondents age at the time of the survey, has six categories ranging from ≤19, 20–24, 25–29, 30–34, 35–39 and Above 39 years. Educational status refers to the highest educational level the respondent attained and it was categorized as Illiterate, Read & write, 1–8, 9–12 and above grade 12. Religion was classified according to the previous literature as orthodox, protestant, Muslim and traditional religion. Occupation of the respondent has four categories like farmer, merchant, government employed and daily labors. Marital status were also classified as single, married, widowed, divorced and separated based on the answer from the respondents. Family Size refers to the number of family members the head of the household have during the survey time and has three category ≤5, 6–10 and above 10. The rest variables were described in the operational definition.
Knowledge of HEP
Was measured based on respondent's ability to respond the questions related to health extension package. Respondent's score of below the mean (<50%) were classified as having unsatisfactory knowledge and those who score above or equal to the mean (≥50) were considered as having satisfactory knowledge on health extension package/services.
Health extension service utilization
Was measured using respondent's utilization of selected health extension services (services given by health extension workers at health post and outreach). Respondent's score above or equal to the mean were considered as utilized and respondent's score below the mean were classified as not utilized.
Are those households that have received theoretical and practical training on 16 health extension packages for at least three months or 94 hours to adopt healthy practices and serve as ‘models’ in their neighborhood.
Data collection procedures
Data was collected through face-to-face interview by using pretested structured Afan Oromo version questionnaires adopted from review of different related literatures. The questionnaire was first prepared in English and translated to Afan Oromo and back to English to check its consistency by respective language experts. In addition, it was pretested and revised accordingly before undertaking the main study. Five supervisors who were diploma holder nurses and twelve completed data collectors with previous experience of data collection and fluent speaker of the local languages were recruited Additional file 1.
Data quality control
To ensure data quality, adequate training was given to data collectors and supervisors to increase the reliability of the data collectors. During data collection, data collectors were supervised on how they are administering questions and randomly visiting household. The supervisors also checked all the filled questionnaires for completion, clarity and proper identification of the respondents every day. The principal investigator double-checks randomly for the completion each day. Incomplete and unclear questionnaires were returned to the subject (one who fill it) to get it completed for the next day using the codes given to the questionnaires and households during the data collection. After data collection was completed re-check-up was made in each kebeles before leaving the area. Finally, the data was cleaned thoroughly and double entered before analyses.
Data processing and analysis
The data was first checked manually for completeness and entered to EpI Info version 3.5.1 computer software. The entered data was transferred to SPSS version 16 computer software program for further processing. Descriptive statistics such as proportions, percentages, means were used and data was presented with tables & texts.
Bivariate analysis was conducted primarily to check which variables have association with the dependent variables individually. Variables found to have association with the dependent variables at 0.2 probability were then entered in to multivariate logistic regression for controlling the possible effect of confounders and finally the variables which had significant association were identified on the basis of OR, with 95% CI and 0.05 p-values and fitted into the final model.
The survey was conducted after the ethical clearance was obtained from research ethical committee of Haramaya University College of Medicine and Health Sciences. Informed written consent was obtained from respondents after explaining the objective of the study. They were informed that their participation was voluntary. Data was collected after assuring the confidential nature of responses and consent was obtained from the study participants.