4.4. Data collection procedures (Instrument, personnel, data quality control)
Instrument: Data were collected using a study format and structured questionnaires.
The principal investigator, with the help research advisor, developed a study format for health facility record review and a structured questionnaire for household survey in English. Two public health students with good language ability of English and Amharic translated the English versions to Amharic. The translated Amharic versions were then translated back to English by two other public health students looking for a possible gap in the contents of the original and the second translated versions. Based on this, wording and way of questioning of the developed questionnaire were corrected.
Pre-testing of the developed structured questionnaire was carried out on December 15,2004, in 'Hatu' pa of Meskan District, which located 5km to west of Butajira town. This rural pa shared somehow similar economic, geographic, cultural and socio-demographic characteristics with study PAs. During the pre-testing the questionnaire was assessed for its clarity, understandability, length, completeness, validity and reliability. At the same time, the performance of the data collectors was assessed. A total of 24 households were selected for pre-testing, it took 30 minutes on average to administer the questionnaire. By this, some previously unidentified but necessary skip patterns were found. Questions with higher proportion of non-responses and missing values were removed, and some modified. The flow of questioning was then rearranged. The identified skip patterns were also corrected.
Data were collected by two groups of data collectors. Ten health workers and one supervisor who were recruited locally, and oriented for one day collected the health facility data. The household data were collected by 10 high school complete data collectors and two supervisors who were selected from Butajira, and trained for two days during December 13 and 14/2004 and data collection was carried out for 10 days during January1-10, 2005.
Data quality control: For health facility data one supervisor was supervising the data collection procedure daily. She checked each and every completed form for completeness and consistency and 2% of randomly selected information in the register. For the household survey, two supervisors were supervising the data collection procedure daily. They checked each and every completed form as well as visited 5% of randomly selected households weekly. Additionally, the researcher was providing onsite technical assistance and guidance.
4.7. Data Analysis procedures
Data were coded, entered into computer using SPSS version 11.1 package and cleaned. Analysis was done using SPSS/PC package. Associations between variables were assessed using Crude and Adjusted OR and significance of association was tested using Chi square test and multiple logistic regressions. Finally, the results of the analysis were presented in tables for further description and interpretation.
4.8. Hypothesis to be tested
It is assumed that health service utilization is affected by various socio-demographic, economic, geographic factors among which:
Sex of household head: Sex of the household head is believed to influence the decision as to use or not to use Health services. It is hypothesized that households with a female head are more likely to use the service than those headed by a male.
AGE: Similarly the increase in age of the respondent is expected to reduce utilization of health services. Hence it is hypothesized that the younger age groups used the service than those older groups.
OCCUPATION: It is also hypothesized that a family with non-farmer household head has positive association with utilization of health services.
Family Size: A household with large number of family size has to allocate higher share of its income on basic requirements other than health service. Hence a household with large family size is associated with less utilization of health service.
EDUCATION: With an increase in the level of education, utilization of health services is expected to increase. it is hypothesized that Mothers having at least some schooling are more likely to have their children vaccinated and to receive maternal preventive services than those with lower levels of education.
MONTHLY INCOME: Economic theory suggests that income and quantity demanded are positively related. Therefore, a significant positive association is expected between income level and utilization of health services.
Wealth rank: With an increase in the Wealth rank, utilization of health services is expected to increase.
HOME and LAND OWNERSHIP: It was assumed that a household that possesses its own house and own private farm land is relatively better off and are more likely to use the service than those who did not possesses of these assets.
Place Of Residence: It was assumed that urban residents are more likely to use the service than rural residents.
Distance From health facility: With an increase in distance from health facility, utilization of health services is expected to decrease.
Decision maker to use health services: It is hypothesized that households in which decision to use health services is made by the mothers are more likely to use the service than those in which decision to use health services is made by husbands alone, or others.
4.9. Ethical Considerations
Verbal consent to participate in the study was secured before conducting the interview. Appropriate advice was given to households with un-immunized child encountered during the survey.
The study proposal had secured the necessary ethical clearance from the ethical committees of the Faculty of Medicine, Addis Ababa University before the commencement of data collection.
A formal letter was written from Department of Community Health, Addis Ababa University to the SNNPR Health Bureau, Gurage Zone Health Desk and the Meskan ans Mareko Woreda Health Office for securing the commitment of the sector and local officials in facilitating the study process.