|
nif:isString
|
-
Study design, setting and participants: A community based analytic cross–sectional study was carried out between February and March, 2015 in Esera woreda (district), Southwest Ethiopia. Esera woreda is located about 670 km from Addis Ababa, the capital city of Ethiopia. In 2015, the district had an estimated total population of 99,319 with 1:2 urban to rural proportion [24, 25]. It has four urban and 25 rural kebeles (the lowest administrative structure in Ethiopia) with an average of 4.8 persons per household (3.8 in urban and 5.0 in rural households) [24]. There were four health centers, 29 health posts and 117 health professionals in the district.
The required sample size was calculated using double population proportion calculation formula with the following assumptions: 52.3% prevalence of healthcare seeking for perceived illnesses of urban households and 29.6% prevalence of healthcare seeking for perceived illnesses of rural households [26], 95% confidence level, 5% margin of error, 80% power, 2:1 rural to urban ratio and 10% estimated non–response rate. Considering a design effect of 2, the calculated sample size of households was 388 (126 urban and 258 rural). Multistage sampling technique was used to recruit the respondents. The district was classified into two strata; urban and rural. Then two urban and eight rural kebeles were randomly selected. To identify households with perceived illness during the last two months, census of households was conducted in the selected kebeles and used as a sampling frame. Then using the sampling frame, households were selected via simple random sampling technique and head of households were interviewed at their home. The schematic presentation of the sampling technique is presented in Fig 1.
Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pone.0161014.g001 Diagrammatic presentation of sampling technique and procedure.Fig 1 shows the graphic presentation of recruitment of households included in the study. District stratified in to rural and urban; kebeles in both urban and rural selected via SRS; census was conducted among the selected kebeles; and households were allocated via PPS.
Variables in the study and its measurement: The dependent variable was healthcare seeking behavior coded as 1 if heads claimed sough healthcare in any modern health facility—hospital, health center and private clinic) or 0 otherwise when any member of the household was sick. The exposure variables included age, sex, education, occupation, marital status, family income, distance from health facility, cost of healthcare service, duration of illness, perceived severity of illness, use of traditional medicine and self-medication. Level of education was classified as illiterate (couldn’t read and write), literate (could read and write but received no formal education), primary (received education up to grade eight), secondary (received education 9–12 grade), and college or university. Duration of illness was measured as acute-if lasted for less than 14 days or chronic- if lasted for more than 14 days. Perceived severity of disease was measured using the question, “Did you think the illness was serious?” and self-medication using the question, “Would you prefer a self-treatment?” with “yes” or “no” responses in both. Data were collected using face-to-face interview using a structured questionnaire (S1 File) with the head of the households at home level. Data collectors got training on the aim, confidentiality of information, respondent’s right and procedures of interview prior to census and actual study data collection.
Data exploration, editing and cleaning were undertaken before analysis. The analysis of both descriptive and inferential statistics was conducted. Descriptive statistics included mean and standard deviation values for continuous data; percentage and frequency tables for categorical data. Logistic regression was used to identify factors associated with health seeking behavior. Bivariate logistic regression analysis was conducted to see the existence of crude association and select candidate variables (with P value below 0.25 were considered) to multivariable logistic regression. We checked multi-collinearity among selected independent variables via variance inflation factor (VIF) and none was found. P-value ≤0.05 was considered as a cut point for statistical significance in the final model. Fitness of goodness of the final model was checked by Hosmer and Lemeshow test and was found fit. Three models: urban, rural and both were developed to compare the health seeking behavior of the households. Data were summarized using odds ratio (OR) and 95% confidence interval. Data analysis was conducted using SPSS version 14 for windows.
The study was approved by institutional review board of college of health sciences at Jimma University (Approval number: RPGC/558/2015). Permission for the study to be conducted was also obtained from the kebeles. Participants were informed of the study and its purpose in their mother tongue. Study participants gave an informed consent before the commencement of each interview, and no personal identification was registered. We prepared an informed verbal consent that involved purpose of the research, expected duration of the interview, and a description that the participants could withdraw from the interview at any time, had no risk and no payment for their recruitment. This statement was read to each study participants before conducting the interview and requested their permission to be involved in the study. Verbal consent was proposed over written consent for the following reasons. Firstly, this was cross-sectional study that enquired descriptive data. Secondly, their responses had no personal, social or political consequences. Thirdly, there would not be significant risk/s to the participants. Lastly, a significant number of people living in the rural areas in Ethiopia have no educational status. The IRB approved the proposed verbal consent procedure. The Confidentiality of the data was ensured.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}