Developing and implementing National Health Insurance: learnings from the first try in Benin

Study design, data collection and data analysis

We used Walt and Gilson’s policy triangle framework to guide the design of the study and frame its analysis. The case study design was used because it allows in-depth investigations of many aspects of the RAMU policy—history, institutional, context, etc—to seek patterns and cause of behaviour of actors. Walt and Gilson’s framework of health policy analysis incorporates interactions between contextual factors—social, economic, cultural and political —content of the intervention—policy options, nature and scale of the policy—actors involved in decision making—individuals, organisations, formal and informal networks—and the process—agenda setting, policy development and implementation. Use of this framework helps understand the dynamics of these four elements that shaped the formulation and implementation of the RAMU.

We performed a mixed methods data collection using in depth interviews, quantitative data extraction from routine information systems and document review to collect both qualitative and quantitative data. In-depth interviews were done with the main stakeholders involved in the formulation and implementation process (decision-makers from ministries of Health, Finance, Social Affairs and Microfinance, NHI-Agency, health officials and healthcare providers working in the health districts involved in the implementation of the RAMU; development agencies; insurance companies; professional organisations and non-governmental organisations). Using the snowballing technique, we identified new respondents, based on their knowledge and involvement in the RAMU process. Twenty Individual in depth interviews were conducted between may and august 2018. Six participants were from the Ministries of Health, Finance, Social Affairs and Microfinance, five were from NHI-Agency, four were from the Technical Working Group of the NHI (TWG-NHI) and five were health officials and healthcare providers working at operational level. Among the participants, 30% were female. The full list of participants was anonymised and categorised by level of seniority and job roles (see online supplemental appendix 2). The interviews covered testimonies of the stakeholders’ direct participation and observations of the process; context of the emergence of the RAMU on the political agenda, appreciation of the governance of the process, the healthcare benefit package, reimbursement flows of providers, premiums, nature of the relations between the main actors in the process. The interviews lasted 35 min on average, and notes were taken.

Furthermore, we performed a document review related to the formulation and implementation of RAMU. We searched official policy documents, legal documents related to health insurance or particularly to RAMU scheme, implementation guides, workshops and study reports, publications in scientific journals from various sources including key informants and online digital literature repositories. Review focused on how the dynamics between stakeholders, contexts and processes structured the key functions of RAMU,22 23 namely mobilisation of resources (eligibility and target enrolment rules), pooling of collected resources (resource pooling architecture and management of revenues) and purchasing of healthcare providers according to contractual clauses (design of the healthcare benefit package and type of services covered, status of contracted providers, method of payment of providers).

Actors’ discourses were analysed through content analysis. The results are presented as key themes.

Patient and public involvement

Patients were not involved in this study. However, the development of the research question was informed by patients’ priorities and public interest as this research aims to contribute to a better accessibility of healthcare through an improved implementation of health insurance reforms. The findings of the study have been disseminated to the participants through local workshop gathering experts from local universities, policy makers from various ministries and civil society actors. The results will be also published in peer-reviewed journals.

Context, actors, content and process of the RAMU policy

The actors

The content

The key RAMU policy features that emerged from our findings are presented in table 1.

RAMU was financed by the annual premiums of affiliates, state budget including taxes allocated to health insurance (taxes on mobile telephony, taxes on financial transactions outside the West African Economic and Monetary Union and excise duties on heavy vehicles), the budget of local governments and subsidies from development partners.29 However, poor and vulnerable people, the majority of whom work in the informal sector, are exempted from paying the contribution. For the authorities, targeted free healthcare for the poor remains a strategy to ensure equity.

In its design, RAMU opted for a single pooling system for the resources mobilised, which theoretically maximises its distributive capacity. The compulsory nature of RAMU (law no 2015-42) implies that all socioeconomic categories of the population must pay contribution to NHI-Agency. This was considered as a mean to avoid problems of adverse selection and allow for a strong pooling of risks. Indeed, all socioeconomic groups, including employers and employees in the formal sector, were meant to contribute to its financing. Arguments against multiple pools revolved around efficiency and cost of running the system. However, in practice the other pooled funds (civil servants contributions, indigent funds, various fee-exemptions schemes) remained separated from the RAMU until it was stopped.6

The RAMU is a third-party payment system.19 So, there is a defined healthcare benefit package that is reimbursed if consumed by its members. The heathcare benefit package consisted of clinical consultations and hospitalisations, drugs, medical consumables, blood products, laboratory and radiology tests, surgical procedures, rehabilitation, compulsory vaccinations, locally manufactured orthopaedic equipment, transportation of the patient from one health facility to another in case of referral, and services related to pregnancy and childbirth (Article 29 of Law No. 2015-42). There is a consensus that purchasing healthcare will be through a common pool. A copayment that varies according to the level of the health system was introduced to regulate the demand for care and allow for cost sharing between NHI-Agency and consumers: 0% at the primary level, 10% at the intermediate level and 20% at the national level.30 The reimbursement of the providers is based on a fee- for-services system. However, the tariffs at which the NHI-Agency will buy healthcare services have been negotiated with the providers. During the period the RAMU has been implemented, only public health facilities have been contracted.

Although the RAMU law made it compulsory for all the population, the first phase focused on people from the informal sector and the extreme poor. The former should pay the NHI membership fees while the latter are exempted from it. The definition of poverty the government has adopted was food poverty. A household is considered as poor if its per capita food expenditure fell below the XOF400.3 per day cut-off. The targeting strategy used by the government consisted of a community-based method in which villagers’ committees selected potential poor households based on local perception of food poverty. Then eligibility was determined by conducting proxy mean test (PMT) on this limited list. This targeting strategy aimed to use the communities’ knowledge, while using the PMT as a check on potential elite capture. Any household with a total score below the eligibility threshold was declared extreme poor. These poor households were reported on a list. The list constituted the register of beneficiaries eligible for exemptions from NHI- membership fees and/or state subsidies. Ultimately, the households exempted from paying contributions are those whose names appeared on the final list of households recognised as poor at national level.

The RAMU reform process

Table 2 shows the historical development of the RAMU policy process with a focus on key events. The ideas of setting up a health insurance scheme started back in 2004 when some Benin government officials participated in regional health insurance workshops organised by the development agencies to build the capacity of health policy makers.30 With the support of the Programme of Social Protection sector of International Labour Organization, these government officials, particularly those from the Ministry of Labour and Civil Service, and the Ministry of Health started to propose, during policy discussions, health insurance as an alternative option to users fees (see context for more details on RAMU agenda setting). In 2008, the government set up the interministerial committee for proposing an effective design and implementation strategy of the policy. Many players said this committee lacked the technical capacity to achieve the missions the government assigned to it. Indeed, the policy document the committee was supposed to elaborate lagged. Due to this limitation, the government asked for the technical and financial support of several development agencies such as the French Agency for the Development, WHO, World Bank and Belgian Technical Cooperation (BTC) to carry out the first studies to support policy decision. Therefore, international consultants reviewed healthcare financing and feasibility studies and organised many workshops with the participation of several institutions, including the trade unions, representatives of the providers, representatives of the mutual health organisation and private insurance companies. A second joint expert mission from the World Bank, WHO and BTC was opened in 2012 with the aim to operationalise the RAMU implementation strategy. This mission assessed the progress made in terms of formulation and implementation of RAMU policy and derived some recommendations to the government such as need of a proper costing of the benefit package, a structured registration fee and dealing with the legal aspect of the universal health coverage. They also recommended to base RAMU on the existing mutual health organisations and finalised the formulation and the implementation guideline of the intervention.6 The interministerial committee was expected to draft RAMU policy formulation document using the results of the review and feasibility studies.

In 2011, the government appointed the director and deputy director of the NHI-Agency, with the missions of managing the RAMU policy reform, reporting to the government and establishing the NHI-Agency. The latter (created in 2012) launched the first operational phase and proceeded to symbolic delivery of the first insurance cards by the Minister of Health to mutual members of the Nikki Municipality in May 2012. The signature of agreements between the NHI-Agency and health facilities to provide healthcare services to the NHI members occurred later, on 18 March 2013. A ministerial decree was taken on 9 July 2013, to define the healthcare benefits package. Furthermore, through NHI-Agency, the government launched several awareness, information and social dialogue campaigns around RAMU concept and submitted the RAMU bill to the Parliament in December 2013. In 2014 (2 years after the launch), a communication strategy was elaborated with the objective of improving the stakeholders’ adhesion to the RAMU. This strategy was implemented (including communication for behaviour change and training sessions). However, there were shortcomings in terms of expression of the voice of stakeholders and knowledge transfer.6

The same year, the government adopted in the Council of Ministers the method for identifying the poorest to be subsidised. Then, a pilot implementation started in 14 health districts with support from the World Bank and BTC (six districts in 2013 and eight more in 2016). Moreover, several tools and technical documents (eg, for guiding the enrolment, patients management or reporting from the health centres) were designed and deployed by the NHI-Agency for supporting the RAMU’s implementation. However, these tools and documents were only available several years after the RAMU’s launch (many of them only being available in 2016-20176).

Table 3 shows some aggregated data on the financing structure collected from the NHI-Agency and related to concrete realisations of the pilot implementation. Over the period 2012–2017, the share of the national budget represented on average 72.93% of the funds mobilised by the NHI-Agency against 27.06% for the development partners. Membership contributions was 2.48% of the collected funds. The reimbursements to healthcare providers were marginal, falling from XOF23.44 million (US$39 848) in 2015 to XOF8.8 million (US$14 960) in 2017. According to our interviewees, the low level of healthcare consumption could be explained by the fact that the names of some beneficiaries who are up to date with their contributions did not appear in the members’ register located in the health facilities contracted by NHI-Agency. Similarly, many households that had paid their membership fees did not receive membership cards. In other cases, the entitled persons received the cards but were turned away when they visited the healthcare centres because some healthcare providers had not yet received the list of beneficiaries and the NHI tools.

With respect to the membership, the data we accessed reveals 48817 persons affiliated to NHI-Agency with 11293 extreme poor, 5170 dockers, 32354 informal sector workers.

Identify the factors affecting the realisation of the reform

The RAMU policy is well intentioned as it provided an opportunity for the Beninese to benefit from health insurance at low cost. However, the realisation of its objectives was limited by some weaknesses in policy implementation.

First, the NHI core managers and implementers did not have full operational and managerial skills in building health insurance. These gaps generated delays and misconception around the purpose and how the reform was implemented. For instance, despite the communication strategy, the public was still lacking clear and appropriate information regarding the benefits package, accredited healthcare providers, or mechanism for affiliation to the NHI-Agency. Also, there was not a formal complaint system, and the information system was not optimal.

Second, the country context may have not been the most favourable to the success of the RAMU policy. There was an ideological conflict between the political will from the President to implement health insurance and the overall pro-user fees culture in the country. The RAMU policy needed a paradigm shift with respect to this pro-users fees culture. The organisational changes (steering documents, changes of the attitudes and the way of working, etc) required for this paradigm shift did not occur. There was also a need to establish trust between providers, consumers and NHI-Agency as health insurance is a relative recent health financing mechanism in Benin. Another contextual issue was that the President lost the majority in the parliament between 2011 and 2016, which probably contributed to delay the adoption of the RAMU law.

Third, there was an uneven process in the reform’s implementation. The health insurance technical documents and tools were finalised very late, several months and even years after the launch. Moreover, although there was an active strategy to raise awareness by involving stakeholders during the development of the implementation guidelines, these discussions did not occur the before the launch of the pilot. The audiences targeted by these guidelines, in particular the healthcare providers, were somewhat sceptical about the capacity of NHI-Agency or the government to mobilise financial resources, including members’ contributions, and to organise the timely reimbursement of health facilities. This was because the district health managers and hospital directors interviewed have noted that in past experiences the government did not respect its commitment towards the health facilities. Moreover, the main actors of the RAMU policy at national level did not show a discourse of persuasion to convince healthcare providers on the interest of the RAMU’s success.

Fourth, no mechanism of incentives, persuasion, control and accountability was put in place to constrain both NHI-Agency and care providers to deliver the reform properly and improve their productivity. Appropriately designed, such mechanisms could generate good policy implementation outcomes. The transmission of decisions or directives from managing agency to the managers of health facilities and healthcare providers was problematic, reflecting its poor information capacity. Degree of compliance of health facilities with the RAMU objectives and the practice at district health level showed significant differences. The perceived mistrust affected active engagement of all key players and some actors interpreted the situation as a lack of government commitment to transparency and good governance. The lack of trust on government leadership including from private sector eroded the confidence that have been gained. Transparency and trust were factors that dampened the support and the willingness of the population to pay for their contribution.

Fifth, there was a glaring lack of collaboration between country stakeholders which led to missed opportunities for enhancing the RAMU approach and leveraging existing resources and expertise. For instance, policy makers from the Ministry of Health and the Ministry of Labour worked separately in proposing a Health insurance scheme for Benin. Also, the local authorities were meant to contribute financially in the RAMU process. However, they were not considered as major players and their role at the various stages of the process were not specified. No incentive mechanism (contract, strategic communication, advocacy efforts, full participation, etc) has been put in place to make them participate actively. The NHI-Agency could have contractually delegated some responsibilities to local authorities. That could concern not only the sharing of annual insurance premium subsidies for the poor with the government but also policy responsibilities in the expansion of coverage. The contract of objectives between the government and the local authorities would be accompanied by criteria for evaluating the performance of local authorities and officials. If these targets are not met, elected officials and local civil servants would receive low scores that will have negative impacts on their future promotion. This mechanism could have increased coverage and enrolment activities. Above all this, the lack of coordination of the donors’ activities did not help in getting the necessary effectiveness and efficiency for the schemes’ sustainability.

Finally, the financial sustainability of the RAMU was threatened by several factors. Although it opted for a unique pooling system, the RAMU has not been able to integrate the other pooled funds, probably because the policy makers lacked the time for the necessary reforms and because of the late adoption of the law. Furthermore, the benefits package (although generous) was not necessarily attractive to beneficiaries because of the poor quality of care in public health facilities, the only ones to contract with the NHI-Agency.31 In these conditions, and in a context where enforcing enrolment was not easy, it is not surprising that the members’ contributions fell from 23.4 million in 2015 to 8.8 million in 2017.6