Determinants of Enrolment for Community Based Health Insurance Scheme among Agro-pastoralist communities of Aw-barre District in Somali region of Ethiopia: Unmatched case control study.

Ethiopia launched community-based health insurance scheme in 2011 as part of the revised health care nancing strategy to ensure universal health coverage and implementation has started in most part of the country since the launching of the scheme. However, the roll out of the scheme started in Somali Region in 2020—much later the rest of the country. The aim of this study was to assess determinants of enrollment of community-based health insurance among households in Awbarre Woreda, Somali Region, Ethiopia Methods Community based unmatched case control study using a mixed approach of quantitative and qualitative methods was conducted between March and April 2021 and the study participants were selected using multi-stage sampling technique. The quantitative method used interviewer administered structured questionnaire among 216 participants (54 enrolled and 162 non-enrolled), while the qualitative method used key informant interview and focus group discussions in two rural and two urban kebeles of the woreda. The quantitative data was analyzed using SPSS version 20 and thematic analysis was used for the qualitative data. Multivariable logistic regression was used to determine the determinants of enrollment for the community-based health insurance and statistical signicance was set at p value of <5%.


Introduction
Community Based Health Insurance (CBHI) is a type of health insurance program that provides nancial protection against the cost of illness and improving access to health care services for communities engaged in the informal sector. Universal health coverage means that all people have access to health services they need when and where they need them without nancial hardship [1]. It includes the full range of essential health services, from health promotion to prevention, treatment, rehabilitation, and palliative care [1,2]. The progress towards this aspiration seems poor [3] particularly for countries whose scal capacity is low and whose social health insurance for the employed sector is absent or very small, thus limiting the mobilization of additional resources from payroll contributions. Financing health care in most developing countries greatly relies on out-of-pocket payments which contributes to unacceptably high burdens of preventable diseases and deaths [4,5] with most donors and global health initiatives such as the Global Fund focusing on speci c diseases or interventions rather than the broader health system.
Countries with a high share of out-of-pocket payments are more likely to have a high proportion of households facing catastrophic health expenditure-de ned as spending more than 40% of household consumption expenditure, excluding food, on health, more than 25% of non-food consumption expenditure of households on health, or more than 10% of total household consumption expenditure on health [6]. The available evidence clearly demonstrates that health insurance can be an alternative to user fees as a health nancing mechanism. The strong evidence that Community Based Health Insurance( CBHI) and Social Health Insurance (SHI) can improve nancial protection and enhance service utilization patterns [7], but the weaker evidence that CBHI and SHI can foster social inclusion of speci c vulnerable groups in Low middle income countries (LMIC) [8,9].
Ethiopia endorsed a health care nancing strategy in 1998 that envisioned a wide range of reform initiatives launch of health to remove nancial barriers, enhance equity, increase health service utilization rate and improve quality of care by increasing resources available for health facilities [10]. In its revised strategy (2017-2025), Ethiopia envisioned to ensure universal health coverage through primary health care by 2035 [11]. The Per capita spending is $33.2 which is far below the globally recommended $86 per capita estimated to make essential health care services available in low-income countries. In terms of spending, the government spends 8% of government budget on health sector which is below the Abuja target of 15% and the out-of-pocket expenditure is also still high standing at 31% of total health expenditure [12].
In Ethiopia, Community-based health insurance (CBHI) scheme was launched and piloted in 2011 and is among the avenues designed to contribute to reduction of out-pocket payments and realization of universal health coverage. The contribution (premium) is collected from the households at the pre-set at rate meaning equal amount of payment is levied to every household regardless of their household characteristics such income and family size [33]. With the assumption that 85% of the total population is involved in informal sector and the country has 100+ million populations, CBHI enrollment coverage in 2012 scal year (2019/20) accounts about 37% of population in informal sector. Geographically, the number of districts implementing CBHI in Ethiopia was 1100 districts in 2020 which translated to a coverage of 75% compared to the 2020 target achieving 80% of all districts to have 80% CBHI coverage. However, majority of these CBHI woreda are concentrated in four big regions and the capital city of the country [13].
Somali region is one of Ethiopia's four Developing Regional States (DRS) and predominantly inhabited by pastoralists (85 per cent), faces developmental inequities and has shown less improvement or lagging in key health outcomes compared to the national average [14]. CBHI was launched in four districts (4%) of this region in 2020-almost ten years behind the other regions of the country. However, implementation has just started only in Aw-barre district with an estimated household of 56,896 and a very low enrolment rate of roughly 16% unlike the national target coverage of 80% set for districts [15]. Thus, it is against this background that this study was aimed at identifying determinant factors for CBHI scheme enrolment in the district and to inform strategies to be implemented in improving the enrolment rate of the scheme in the region.

Study area
Aw-barre district is 72 km North of Jigjiga-the capital of Somali region and is rst district to start CBHI implementation in Somali region of Ethiopia. Administratively, the district is composed of 44 Villages (kebeles) arranged in six CBHI clusters and its population is projected to be 382,569 based on the 2007 national census conducted by the Central Statistical Agency of Ethiopia [15]. The number of households enrolled in CBHI scheme are 10,100 which roughly translates in to 16% registration rate (District estimation). The district has four health centers and 46 health posts providing primary health care services to district population.

Study duration
The study was conducted in Aw-barre district between March 24 to April 24, 2021.

Study design
The study design was community based unmatched case control study (enrolled and non-enrolled members) using a mixed approach of quantitative and qualitative methods.

Source and study populations
The source population were all households residing in Aw-barre district at least in the last six months before the study.

Sample size estimation
The sample size was computed using Epi Info TM 7 software [16]with the following assumptions: 80% statistical power with a level of signi cance at 5%, Odds ratio of 4.16 [17] and a case to control ratio of 1:3. The estimated nal sample size was 216 (54 enrolled and 162 non-enrolled) using the doubleproportions formula.

Sampling methods and procedure
The quantitative part of the study was conducted using multi-stage cluster sampling technique. In the rst stage, ten kebeles(villages) were selected by simple random sampling out of 44 kebeles which are under six CBHI clusters. In the second stage, 216 HHs were selected using probability proportional to size of households in villages. A "Kebele" is the smallest governmental administrative unit, and on average has a population of 5000 people (Fig1).
The qualitative data was collected using focus group discussions (FGDs) and Key Informant Interviews (KIIs) at district and regional level to triangulate with the ndings from quantitative part. Four FGDs were done in two purposively selected kebeles (one rural and one urban) for both enrolled and non-enrolled households. Key informants were woreda/district health o ce representative, woreda CBHI coordinator and Regional Health Bureau focal point for CBHI.

Data collection tools and techniques
A structured questionnaire was administered by trained enumerators to head of selected households and a Semi-structured interview guide used for the FGD and KII. The questionnaire and interview guide were prepared rst in English and translated to Somali and then translated back to English to keep its consistency.

Data analysis
Data were entered and analysis was performed with SPSS V.20. Descriptive statistics was calculated to describe the study objectives in terms of appropriate variables. Binary and multi-variate logistic regression analyses were performed to identify determinant variables for enrollment of CBHI scheme.
Variables with a p-value of ≤0.25 on bivariate logistic regression analysis were used for multivariable logistic regression analysis to nd the adjusted odds ratio at 95% CI and p-value <0.05.
For the qualitative data, notes were taken, and the entire discussion was audio recorded and later transcribed for analysis. Analysis was done using a thematic analytic approach. Emerging themes were summarized under different thematic areas and quotes from participants were presented textually.

Results
Page 6/19 Socio-demographic pro le of Respondents    Four FGDs were conducted -two among CBHI enrolled households and two among non-enrolled households in rural and urban kebeles. The common reason for joining CBHI was the need to get accessible health care services free of charge for family members.
"I registered to get medicines and free health care services for my family" [CBHI member in Lafa-issa kebele of Awbarre woreda].
For non-CBHI members, the commonest reasons for not being enrolled were not heard of CBHI followed by unaffordable payment.
"It is through this FGD session that I rst hear community-based health Insurance scheme" [Non-CBHI member in Lafa-issa kebele of Awbarre woreda].
Regarding indigents targeting, FGD members from different kebeles had varied views. Whilst some FGD members claimed that the poor were identi ed fairly in their kebele and their payments were covered, other FGD members from the other kebele claimed that indigent identi cation and targeting did not happen at all.
"Membership registration in this kebele was based only on the ability to contribute 250 Birr, no single indigent was registered".
"The selection process for the poor households were transparent and fair in my view. And I think the process was based on the economic severity of the households. I can say needy households were not left out" [CBHI member in Lafaissa, Aw-barre district].
Key informants at regional and district level have described commendable progress of the scheme within a short period and the challenges that came along with it.
"Some of the major achievements include helping the community understand the bene ts of insurance, managing many poor households and people with chronic illnesses to get enrolled into the scheme and receive free service including the laboratory services. But some challenges such as community misunderstandings require further work" -[Awbarre woreda, CBHI coordinator]

Discussion
This study identi ed determinants of household enrolment for community-based health insurance (CBHI) in Aw-barre district of Somali region, Ethiopia. Determinants of enrolment include awareness about CBHI, Household income, perceived affordability of the premium and being a member of a solidarity group.
The study found that having information about CBHI was a key determinant identi ed to affect enrolment of the CBHI scheme in the study area. Households with information about CBHI were more likely to get enrolled into the scheme and the poor enrollment is related to inadequate sensitization at community level even though the scheme is at the early stage of implementation. This is similar to ndings in other studies done in Ethiopia [16,17,19,25,30], Nigeria [18], Uganda [22] and Cameroon [27] which identi ed inadequate knowledge and understanding of insurance and CBHI principles are some of the obstacles to enrolment. Another study also found inadequate client education, and limited community engagement as factors that threaten the scheme's sustainability [39]. Similarly, a previous study in Ethiopia found that knowledge of CBHI not only determines the enrolment but also associated with a drop out from the scheme [38].
This study also found that being enrolled in a solidarity group was an important determining factor for household enrolment of CBHI with those who are members of any form of a solidarity group more likely to get enrolled into the scheme. This goes in line with a study that reported solidarity and trust encourages groups susceptible to risk to put together their resources for common use [20] and is also consistent with other studies done in Ethiopia [17] and Cameroon [27]. Moreover, community solidarity is a form of a social capital at community level that positively impact households' decision for health insurance, which in turn increases the demand and sustainability of CBHI and could serve as a platform for sensitization on CBHI scheme [20,27,28,42].
Household income was found to be another determinant for CBHI enrolment. Households with higher annual income tended to be enrolled more into the scheme compared to those with less income. This is comparable with other studies with similar ndings [17,22,24,26,29,30]. A systematic review of factors affecting uptake of CBHI in low-and middle-income countries demonstrated low levels of income and lack of nancial resources as major factors affecting enrolment [23]. Household income directly contributes to the affordability and is positively correlated with willingness to pay the premium as found by different studies [23-26, 29, 34, 37]. This nding is akin to the nding in the evaluation of Ethiopia's CBHI which documented that 39 percent households mentioned registration fees and premium are not affordable [19]. Other studies conducted in Ghana, Nigeria, Mali, and Senegal also reported that at-rate premium restricted enrolment decision of the poor [23,31,[35][36]. This is the very reason why there is a targeted subsidy for the poor in Ethiopia's CBHI design (an estimated 10% in each district) [33] although there is always a huge challenge in identifying who the poor is or the indigent population expected to bene t from the subsidy [40].
On the contrary, a previous study in Ethiopia reported that household's socioeconomic status does not inhibit uptake of CBHI and found that most food-insecure households were substantially more likely to enroll [32]. This was done in a setting where poor households were participating in a productive safety net program which targets chronically food insecure households. Another study reported that participation of households in both social protection programs i.e. community based health Insurance and productive safety net program were associated with a 5 percent increase in livestock and a 27 percent decline in debt which in turn affords greater protection to vulnerable households against multiple risks [21].
The ndings in the study depended on the feedback provided by the respondents which may be subject to some form of respondent bias. The strength of this study is the use of combined quantitative and qualitative approach which helped in triangulating the information provided by the respondents and gave better clarity and understanding of the quantitative information. In addition, this is the rst study done in this region of Ethiopia on this topic to our knowledge.

Conclusions
The enrollment for community-based insurance was determined by being well informed about CBHI, Household income, and being a member of solidarity groups at community level. Given the early stage of implementation, enhancing sensitization of the community about the scheme using various community platforms, promotion of the existing community based solidarity groups/associations, diligent targeting the poor households and whilst ensuring linkage with any existing social protection program would help to increase enrolment for the scheme. Ethical approval and consent to participate Ethical clearance was granted by Ethics Review Committee of Somali Regional Health bureau, Ethiopia. Support letter was given by Bureau addressed to study districts and verbal consent was taken from study participants after being given adequate information about the study. More-over, anonymity and con dentiality were kept by omitting personal identi ers.

Figure 1
Schematic presentation of the sampling procedure.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. CBHIDATA.xlsx