Accessibility and utilization of modern health care among rural community in Ethiopia: The case of Bako district, Western Oromia: Community based descriptive cross-sectional study

Objectives: Knowing determinants of health seeking behaviors of families is signicant in assuring access and utilization of health services. Less is known on the seeking behaviors and access of modern health services among populations in rural settings in Ethiopia. The general Objective of the study is to assess health care seeking behavior (Utilization) and associated factors among the community of in Bako Tibe District, Ethiopia. Community based descriptive cross-sectional study design was used on 750 head of the residents of ≥ 18 years old. Multistage sampling technique was used to get the households and data was collected by interview using structured questionnaire and analyzed using SPSS version 24. Binary and multiple logistic regression analysis were used to identify the factors associated with the health seeking behavior. Results: 495(66%) reported having morbidity and level of health care seeking behavior was 78.6% of which 66% sought care from government facilities. Sex of respondent, residence, family size, educational status, monthly income, disease condition, Perceived severity & access to needed health information were found to be signicant associated factors of healthcare seeking behavior of the respondents in multiple logistic regression analysis. seeking Access, Bako district, Ethiopia


Background
Health seeking behavior can be conceptualized health service utilization as outcome or the process of seeking care. Health care Utilization behaviors should be seen in wider contexts of social capital as communities' engagement in the health system and services. (1)(2)(3)(4). Knowing determinants of health seeking behaviors of families will be essential in assuring access and utilization of health services. Less is known on the seeking behaviors and access of modern health services among populations in rural settings in Ethiopia (5,6).
Studies indicate that, people who are in better socioeconomic status are in better utilization at household level. Study in rural Bangladesh showed that households in higher wealth quartile had better utilization of maternal and child health services and Relative poverty level played a key role in utilization of health services. Study from national survey data in Nigeria indicated that education level of individuals were main factor for individual level predictor of maternal service utilization and socioeconomic status and residence status were main predictors of utilization of services at household and community levels respectively (7)(8)(9). Usually, the poor seeks care from government facilities compared to the middle class and wealthy (10)(11)(12)(13)(14).
Contrary to this, on study India, the higher and the middle income individuals used public facilities more than the lower income and the lower income utilized more private facilities. Seeking care from private facilities in respective of income status may imply concerns of quality of care in public facilities. In South Africa, most rural and urban residents claimed adequate access to health care but they still face di culties in actual receiving of care (12,15,16).
According to study in Uganda, private health care providers' role is increasing in Uganda and other sub-Saharan African countries implying regulatory mechanisms, quality and affordability of services (18,19).
Studies from east African countries show that, preference for modern health care is almost universal but differed for different socioeconomic classes. Delay to seek care was higher for adult related cases and poorer groups. (20)(21)(22)(23)(24).
Studies also show that, during emergency and critical illness proximity of services is the main issue to choice facility and in case of chronic illness cost of service becomes a main concern to determine type of provider (23,25,26). This study intended to determine the level of health care seeking behavior of the community and to identify factors that affect healthcare seeking behavior of the community in Bako Tibe District. Figure 1 below shows the conceptual framework of the study. Study Design and period: Community based descriptive cross-sectional study design was used from August to September 2017.

Population:
All households residing in Bako Tibe district were considered as the source population and Household respondents who participated in the study were considered as study populations.

Inclusion and exclusion criteria:
Head of the households or family representatives of ≥18 years old who resided in the district for six or more months before data collection period.

Sample Size determination:
Sample size was calculated using single nite population proportion formula with the following assumptions: 58.4% general prevalence of healthcare seeking behavior for households, from a study (Begashaw B. et al 2016), 95% con dence level, 5% margin of error, and 5% estimated non-response rate, design effect of 2, the required sample size was 750 (177 urban and 573 rural). Considering 5% estimated non-response rate and a design effect of 2, n = 783 Sampling Procedure: Multistage sampling technique was used to get the households. The district was classi ed into two strata; urban and rural. Eight kebeles were included in the study (2 urban and 6 rural). The kebeles were selected using simple random sampling methods. Systematic sampling technique was employed for households' selection.

Data Collection Procedure:
Data was collected by interview using structured questionnaire. The data collection tool was prepared in English and translated to local language (Afaan Oromo). To check reliability it was translated back to English.

Variables:
The dependent variable was healthcare seeking behavior and the exposure variables were 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, residence and selfmedications.

Data Analysis and interpretation:
Data was entered in Epi data version 3.1, transported to SPSS version 20 and analyzed using it. Descriptive and analytical statistics were calculated. Binary and multiple logistic regression analysis were used to identify the factors associated with the health seeking behavior. Crude and adjusted Odds ratios were used to determine strength of association at 95 Con dence level and error ( = 5%). Summaries of the result were presented by textual, tables and graphs.
About 452 (60.3%) of households had family size greater than four. Only 252 (33.6%) of households have Community health insurance and majority of the respondents are Oromo 704(93.8%).
Health care seeking behavior: Among all respondents, 495(66%) reported having morbidity of which 335 (67.67%) of the illness was perceived to be acute illness. Among those who reported morbidity, perceived severity of illness was 351(70.9.8%). Distance from modern health facility (Public health center, private clinic &health post) is less than 10 km far from all households. The general level of health care seeking behavior was 78.6%.
About 66% of the respondents sought care from government facilities, the main type of facility being health center which was sought by about 52% of all those sought care. Only 70% of those who sought care reported immediate seeking of care after the perception of the illness.
Factors associated with health care seeking behaviors of households: Result of bivariate analysis to identify the candidate variables for multinomial logistic regression analysis is indicated below ( Table 1).
From adjusted model in multinomial logistic regression analysis below (Table 2), sex, residence, family size, educational status, income, disease condition, Perceived severity andaccess to needed health information were found to be signi cant factors for seeking healthcare of the respondents in multiple logistic regression analysis. Accordingly, the odds of health seeking behavior among male participants was about 5.7 times higher than female participants (AOR = 5.7, 95% CI: 3.0, 11.0) & the odds of healthcare seeking behavior among urban households was about 9.5 times greater than healthcare seeking behavior of rural households (AOR, 9.5; 95% CI, 3.6, 25.5). Healthcare seeking behavior among households with < = 4 family sizes were approximately 6 times greater than households with >4 family sizes (AOR = 5.8, 95% CI: 2.6, 12.8). Households with monthly income above 1,170 birr were about 9 times more likely to seek healthcare as compared to those who earn less than 1,170 (AOR = 8.97, 95% CI: 4.5, 17.7). Healthcare seeking behavior was approximately 2.8 times greater among household having secondary education & 8.5 times greater among households graduated from College and above when compared with illiterate ones (AOR = 2.8, 95% CI: 1.02, 7.5) & AOR = 8.5, 95% CI: 1.9, 38.9 respectively). The odds of health seeking behavior among those who perceived serious illness (Perceived severity) was about 3 times higher (AOR, 3.3; 95%CI, 1.7, 6.4) than those who didn't. Households who perceived illness (disease condition) as acute were about 2.8 (AOR = 2.8, 95% CI, 1.5, 5.2) times more likely to seek healthcare than those who perceived chronic illness. Households who have access to needed health information of healthcare for perceived illness were about 3.7 time more likely to seek healthcare than those who had no access to needed health information of healthcare (AOR = 3.7, 95% CI, 1.9, 7.4). (Table  2).

Discussion
Since past 3 decades, Ethiopian strived to increase the coverage of the modern health care for massive populations particularly the rural and remote communities through different strategies and potential health service coverage reached beyond 90% (27,28). As to the health seeking behavior of the community, about 78% of the respondents sought care for their illness. This nding is relatively higher than studies conducted in the country elsewhere. This can be the attributed to the decentralized and community based health care delivery strategy in the country. However still signi cant proportion (22%) of the population did not seek care for their perceived illness and about 30% who sought care were late in seeking care (4,5,8).
One of the impressive ndings of the study is that all of the households had access to physical facilities (Public health center, private clinic and health post) within 10 km of distance. Proximity of basic health care facilities is among key factors that utilization of modern health care. The structurally decentralized facilities should be functionally accessible in terms of quality and cost of health services. This nding is in agreement with study in South Africa which indicate adequate access of facilities for health care although there is di culty in actual access of care (11).
The health care facilities visited by the care seekers varied among the respondents. Majority of them (66%) visited government facilities. Health center, which is expected to provide basic health promotive, disease preventive and curative services, has been visited by majority of the participants (52%). This is in agreement with the policy direction of the country as the essential health services should be available to all population with in 10 km diameter. However, the proportion of the respondents who sought care from health post is very low (below 2%). The utilization of private health care institutions also contributed for the second major share, 23.9%. This implies the increasing share of private health care providers in preference of seeking modern health care by the community. This is contrary to the existing practice where there is limited support and regulation to the private health services in the country. The fact that this share is increasing may imply that the government facilities are not embarking on quality of health care that satisfy the health care needs of the country. This has been evidenced by other studies globally (9,11,15). Health seeking was affected by different level factors. Males seek care 6 times more than females. This is strong association which shows huge gender imbalance in health care utilization in the area. This can be attributed to the sociocultural, economic, burdens Ethiopian women face in the livelihood of the households (EDHS). Ample studies globally also show similar pattern of imbalance in health seeking behavior among male and female but in different level of strength (1, 2, and 14). This study also identi ed that urban residents had 9.5 times better health care seeking behaviors than the rural ones. This can be justi ed by better access of health information and access of health facilities of the urban population than the rural.

Conclusions
The overall healthcare seeking behavior of households for perceived illness was satisfactory. The private health sectors (clinics) are contributing a signi cant role in access of health care for the community in the study area and demands strong support from the public sector. Sex, residence, family size, monthly income, educational status of participants, perceived severity of illness, condition of illness and access to needed health information showed a signi cant association with health care seeking behavior. Family and community level interventions should be enhanced and CBHI should be scaled up especially the rural population.
Limitations seasonal variation, recall bias and social desirability bias might have affected the study.

Declarations
Abbreviations: EDHS: Ethiopian demographic and health survey; US: United States; CI: Con dence interval; AOR: Adjusted odds ratio; COR: Crude Odds Ratio; CBHI: Community based health insurance.
Ethics approval and consent to participate: Ethical clearance was obtained from Wollega University ethical clearance committee. Written consent was obtained from the respondents to participate in the study and con dentiality was kept along all process of this research work. The purpose of the study was also communicated and there is no serious harm the study poses to the participant expect a time they waste for interview. They were also told as they can withdraw from the study in middle of the responses if they think it is not fair to them.

Consent for publication:
Not applicable.
Availability of data and material: The data sets during and/or analyzed during the current study available from the corresponding author on reasonable request.

Competing interests:
The authors declare that they have no competing interests.

Funding:
Fund was not received from any rm or individuals.
Authors' contributions:All authors have participated in the research work and have read and approved the manuscript.
TA generated the research question, developed proposal, supervised data collection process, analyzed data and prepared research report.
ZD was main/senior advisor of the research proposal development and data analysis process.
MC was co-advisor of the research proposal development and data analysis process.   Figure 1 Conceptual Framework for Healthcare Seeking Behavior modi ed from Andersen's behavioral model.