Introduction

Globally, 150 million people per year suffer from financial catastrophic shock, and 100 million are pushed into poverty because of direct payments for health services. In the majority of African countries, more than 40% of their total health expenditure was constituted by out-of-pocket payment (OPP) and this resulted in scarcity of funds for health [1]. As a result of this, developing health financing system was a common agenda for all countries [2,3,4].

In Ethiopia OPP accounts 37% of the total spending in health care. This leads to financial catastrophic shock [5]. To overcome these challenges an efficient and equitable health care financing strategy was formulated [6]. The immediate goals of universal health coverage (UHC) are access, utilization and financial protection, while health status is a longer-term indirect result (although probably the ultimate goal), and one of the key strategies is reducing direct payments [2].

The government of Ethiopia has introduced Community Based Health Insurance (CBHI) as strategy for reducing financial catastrophic shock in our country, and it has been piloted in selected districts of the country. The utilization of scheme by the community members were affected by different factors such as socio-demographic, economic and health related factors [7].

A pilot study in Ethiopia has indicated that the scheme provides a benefit package of all available services in the health centres and hospitals excluding tooth implantation and eyeglasses with the provider payment mechanism of fee-for-service. Accordingly, the membership for the scheme could be all core family members [8, 9].

According to the Health Sector Financing Reform (HSFR) project report, the overall enrollment in the pilot schemes was 48%. It varies from district to district and ranges from 25% in Deder to 100% in Yirgalem, and households registered as indigents is 7%, and ranges from 1% in Deder to 15% in Tehuledere and Yirgalem [10].

Understanding the community’s willingness to pay (WTP) for the CBHI and determining the amount of money is useful for policy makers to implement the program and to expand the program nationally. Therefore, the objective of this study was to assess the WTP for CBHI and associated factors among rural households in Bugna District, Ethiopia.

Main text

Methods

Community based cross-sectional study was conducted from February to March, 2016 among households in Bugna district. Bugna district is found in Amhara National Regional State which is 354 km far from Bahir Dar (capital city of the region). The district has thirteen rural and one urban kebele. According to the district health office report, it has a population of 91,750 in 2016 [11]. The district has 4 health centers and 13 health posts that provide health service for the dwellers.

All households in the rural kebele of district were the source population whereas all households in the selected rural kebele of the district were the study population. Those permanent residents of the community with household head aged 18 years and above were included in the study. Households with heads or spouses that have been employed in the formal sectors were excluded from the study.

The sample size of the study was calculated using single population proportion formula with the assumption of; proportion of WTP for CBHI in Fogera district (p = 80%) [12]. The total sample size was 541 after considering 95% confidence level, 5% margin of error, design effect of 2 and 10% non-response rate.

Multi-stage systematic sampling technique was employed to select study participants. First, five out of 13 rural kebeles were selected using lottery method. Then, the sample was proportionally allocated among the selected kebeles based on the number of the households. Finally, systematic random sampling technique was used to select participants.

The dependent variable was WTP for CBHI, whereas socio- demographic characteristics (age, sex, marital status, educational status, family size, religion and ethnicity), economic factors(wealth index and occupation); environmental factors (distance from health institution in walking time);health and health-related factors (health condition of illness, chronic illness and disability, medical treatment for the recent episode, health care cost of the recent treatment, perceived quality of the health care service in the area); knowledge related factors(awareness about CBHI and social trust were independent variables of the study. The CBHI scheme (insurer) covers all costs of the members related to transportation, drugs, laboratory tests, hospitalization, and other necessary imaging examinations, but expenses of some selected chronic non-infectious diseases such as diabetes mellitus, heart failure, cancer, and treatment from abroad weren’t covered by the insurance. WTP for CBHI was measured using the bid contingent value method through asking a specific amount of money (180 ETB), and probing the question using a higher or lower bid value depending on the respondent’s response to the first question until the maximum amount of money participants were WTP. A pre-tested and standardized questionnaire was used to collect the data. Most of the tools were adapted from the previous studies of Benchi Maji [13] and Fogera [12]. The wealth item questions were adapted from the Ethiopian Health Insurance Agency CBHI evaluation report [8]. Pre- test was done on 10% of the subjects at Lasta District.

The collected data were cleaned, coded, entered into EPI-INFO version 7 software and exported and analysed using STATA software package version 14. Variables with p-value of less than 0.2 during bivariate analysis were considered for the multiple regression analysis [14,15,16]. Regression coefficient with 95% CI, t-value and p-value were used to measure the strength association.

Tobit model was used to identify factors associated with WTP and the maximum amount of money that individuals were WTP. This model has an advantage over other discrete choice models (linear, logistic, and probit) in that, it reveals both the probability of WTP and the maximum amount of money the respondents are WTP.

$${\text{y}} = \left\{ {\begin{array}{*{20}l} {1\quad if\;MWTP =\upbeta_{0} + {\beta^{\prime}}Xi + {\text{e}} > 0} \\ {0\quad if\;MWTP \le 0} \\ \end{array} } \right.$$

where y = outcome; MWTP = Maximum willingness to pay; Xi = explanatory variables; β0 = Slope, β′ = Coefficient; e = error term; 0 = No and 1 = yes.

The model also estimates marginal effect of an explanatory variable on the expected value of the dependent variable. A p-value = 0.05 was used to determine statistical significance.

Results

Socio-demographic characteristics of households

A total of 532 participants participated in the study with response rate of 97.4%. The median age (IQR) of the respondents was 35 (14) years, ranging from 18 to 75 years. The majority of respondents were males (66%), farmer (70.3%), illiterate (55.3%). Moreover, 71.6% of the participants were married and the mean family size of the participants were 4.8 (± 1.8) (Table 1).

Table 1 Demographic and socio-economic characteristics of study participants in Bugna District, North East Ethiopia, 2016 (n = 532)
Full size table

Health related characteristics of study participants

One hundred ninety-six (36.5%) respondents evaluated their family’s health status as medium. Eighty-seven (16.4%) of the member of the households had at least one member with chronic illness while 41.0% had at least one member who had encountered acute illnesses 3 months prior to data collection. The median expenditure of the 145 households which got treatments was 60 ETB with range of 10 to 1500 ETB. Sixty-eight (46.9%) of these households reported that it was difficult to cover payments for treatments. As a result, 43.4% of them have borrowed money from someone to cover their medical costs (Table 2).

Table 2 Health related characteristics of study participants in Bugna district, 2016
Full size table

The mean amount of money household heads willing to pay was 233.3 ETB (95% CI = 225.6, 241.4) per house hold annually or 11.1 USD (Additional files 1 and 2).

Factors associated with willingness to pay for CBHI

For each an additional household member, the WTP value of households was increased by 0.4 USD other conditions being held constant (ß = 0.408, 95% CI (0.092, 0.724).

The study also revealed that respondents who attended formal education were WTP 3.2 USD more than those who didn’t attend formal education, holding other variables constant (ß = 3.20, 95% CI (1.87, 4.53).

Households who had awareness about the scheme were WTP 2.9 USD more than those who had no awareness about the scheme, holding other variables constant (ß = 2.96, 95% CI (1.61, 4.30).

Households who had history of illness in household member were WTP 2.5 USD more than the counter factual, holding other variables constant (ß = 2.52, 95% CI (1.61, 4.30). Similarly, households with higher economic status were WTP 5.5 USD more than those who are in lower socioeconomic status, holding other variables constant (ß = 5.55, 95% CI (4.19, 6.90) (Table 3; Additional file 3).

Table 3 Maximum likely hood of Tobit econometric analysis of factors associated with WTP for CBHI in Bugna district, 2016
Full size table

Discussion

This study aimed to assess WTP for CBHI and associated factors among households in Bugna district, Northeast Ethiopia. About 77.8% of participants were WTP for CBHI. The average amount of money WTP per household per annum was 233 ETB ($11.12 USD) for proposed CBHI. This implies the communities were WTP higher money than the proposed payment during the study period which was 140 up to 195 ETB. But the amount of money respondents WTP was consistent with the newly proposed premium which is 240. The mean amount of money WTP in the study is greater than study in Nigeria [17] and Ethiopia [9]. However, this was lower than study done in Cameroon [18], Nigeria [19], Namibia [20], Bangladesh [21] and Burkina Faso [22]. The discrepancy might be due to the difference in study period, area, design and participants.

The study revealed that households with larger family sizes were WTP a higher amount compared to households with smaller family size. This finding is supported by other studies conducted in Nigeria [17], Ethiopia [8], Bangladesh [21], china [23], and India [24]. The possible explanation might be as a result of the huge financial burden faced by households with a large size when they seek health care services.

Educational status of the respondents was found to be another factor to increase the amount of premiums WTP for CBHI scheme. This is consistent with a studies done in Nigeria [25], Cameron [18] and Burkina Faso [22]. Educated household heads have better knowledge about the advantage of making regular insurance payments to avoid the risk of catastrophic medical expenditures at the time of illness.

Respondent’s awareness about the CBHI had positive and significant association with WTP for the CBHI. This finding was consistent with studies conducted in Nigeria [26], Cameron [27], and Myanmar [28]. This might be due to knowing the catastrophic effects of health problem and the benefits of joining the insurance scheme earlier.

The wealth status of the families had positive and significant association with WTP. Which is similar with other studies conducted inEthiopia [12], Nigeria [19], Bangladesh [21], China [29], St. Vincent [30], and India [24]. The possible explanation might be having more wealth is associated with high asset losses if an unexpected event occurs that leads to be households more WTP for the insurance than the poorer.

The history of illness in the last 3 months in household members had significant effect to WTP for the CBHI. This finding was consistent with other studies in India [24], and Bangladesh [21]. This might be due to the risk-averse individuals are more likely to enrol in the insurance.

Political will and governmental commitment towards the scheme were the opportunities for CBHI implementation in the country. Accordingly, the presence of clear action plans, national scope of implementation and existence of regulatory frameworks were the strengths before introducing and start to implement the scheme. On the contrary, the decision for CBHI enrollment was made at the kebele level as opposed to the household level and a general subsidy from the federal government is also provided for all scheme members without giving especial emphasis for those members who cannot afford to pay. Furthermore, premium load was decided only by their family size without considering their level of income were the limitations of the scheme [8, 9].

Overall, this finding indicated that the proportion of households willing to pay for the scheme was low as compared to the goal of UHC. WTP for CBHI were influenced by educational status, history of illness, family size, awareness and wealth status of the households. Therefore, the government had better to consider the household family size and wealth status for setting the premium load. Moreover, strengthening awareness creation at the community about the scheme to enhance utilization of the scheme.

Limitation

Even though this study lies on the large sample size with high response rate, the study has its own limitation on response biases which may overestimate or underestimate the results of WTP due to the use of self-reporting.