The “universal” in UHC and Ghana’s National Health Insurance Scheme: Policy and implementation challenges and dilemmas of a lower middle income country

Background: Despite universal population coverage and equity being a stated policy goal of its NHIS, over a decade since passage of the first law in 2003, Ghana continues to struggle with how to attain it. The predominantly (about 70 %) tax funded NHIS currently has active enrolment hovering around 40 % of the population. This study explored in-depth enablers and barriers to enrolment in the NHIS to provide lessons and insights for Ghana and other low and middle income countries (LMIC) into attaining the goal of universality in Universal Health Coverage (UHC). Methods: We conducted a cross sectional mixed methods study of an urban and a rural district in one region of Southern Ghana. Data came from document review, analysis of routine data on enrolment, key informant in-depth interviews with local government, regional and district insurance scheme and provider staff and community member in-depth interviews and focus group discussions. Results: Population coverage in the NHIS in the study districts was not growing towards near universal because of failure of many of those who had ever enrolled to regularly renew annually as required by the NHIS policy. Factors facilitating and enabling enrolment were driven by the design details of the scheme that emanate from national level policy and program formulation, frontline purchaser and provider staff implementation arrangements and contextual factors. The factors inter-related and worked together to affect client experience of the scheme, which were not always the same as the declared policy intent. This then also affected the decision to enrol and stay enrolled. Conclusions: UHC policy and program design needs to be such that enrolment is effectively compulsory in practice. It also requires careful attention and responsiveness to actual and potential subscriber, purchaser and provider (stakeholder) incentives and related behaviour generated at implementation levels. The Volta region, one of the ten regions of Ghana, was purposively selected for the study because the co-sponsor for the study, the Korea Foundation for International Healthcare (KOFIH) was involved in supporting Maternal and Child Health programs in that region. Additionally, it was one of three regions where phase two of the scale up of the NHIS per capita provider payment system for primary care was planned, and findings from the study were considered of relevance to the scale up. In the 2010 national population census the Volta region had a population of a little over two million, about 8.6 % of Ghana’s population [18]. Its inter censal growth rate of 2.5 % was the same as the national average. Ghana has a relatively young population with 43 % of the population aged less than 18 years and the Volta region is no different. A municipality and a rural district adjoining each other, were purposively selected to study contrasting contexts of an urbanized district with relatively good health service access and a rural district with poor health service access. For reasons of research ethics and confidentiality, we do not use the names of the districts or the communities in which the study was conducted in this publication. The Municipality had a population of almost three hundred thousand in the 2010 population and housing census. It had relatively good social, economic and healthcare infrastructure, water and sanitary conditions, and fairly high literacy rates compared to the rest of the region. It had 45 health facilities including a regional and municipal hospital, with 43 run by the Ghana Health Service, one mission and one privately owned. Formal sector economic activity comprised mainly of employment in the public service, private services sector and private construction companies. Non formal sector economic activities were petty trading, subsistence farming, animal rearing, artisans and vocations such as hairdressing and dressmaking. It had an NHIS district scheme office. The population of the rural district was about 64,404 in the 2010 population and housing census. The district lacked basic social and economic infrastructure such as road networks, telecommunication, industry, banking etc. and had serious problems with access to healthcare, water and sanitation facilities. The main occupations were farming, Kente weaving and petty trading [19]. The Administrative Capital had one health centre with a trained midwife, nurses, dispenser and other supporting staffs. There was no hospital. Other sources of modern biomedical health care in the district were two health centres and several small clinics and Community Health Planning and Services compounds. We employed an exploratory, mixed methods cross sectional case study design. Data collection involved extraction of routine management information system data from the district insurance scheme records, community focus group discussions (FGD) and key informant (KI) interviews. We also reviewed NHIS policy and program document such as acts of parliament, legislative instruments and annual reports. Within each district, two communities each were selected for the community member FGD and key informant interviews. Criteria for selection were that in each district, one community should be within half an hour’s walking distance and the other an hour or more’s walking distance of the hospital (municipality) or health centre (rural district). Within each of the study districts, a mix of purposive and snowball sampling was used to select respondents for key informant interviews to get the views of key stakeholders namely purchaser (NHIA district scheme staff), providers, district assembly (local government) and clients. NHIA district scheme staff interviewed, were the district scheme manager (head of the team), Public Relations Officer, Management Information Systems officer. Within each district assembly (local government), those interviewed were the district Chief Executive, Finance Officer, Coordinating Directors, Assembly men (elected community representatives of the district assembly), heads of unit and area committees and opinion leaders. Providers interviewed were the district director of health services, and health insurance claims officers at the health facilities. NHIA regional management staff were also interviewed. In total 35 in-depth interviews were held. Interviews lasted between 30–70 min and were held in the offices and communities of the respondents. There were some rare situations where interviews were arranged and held at the homes of respondents for their convenience. Community focus group discussions were held with adult groups of currently enrolled, previously enrolled but currently uninsured because of failure to renew, and never insured in each of the four communities. Participants were purposively selected to ensure a mix of male and female and range of ages between 18 and 82. Community FGD were also held with insured pregnant women. They were recruited from pregnant women attending antenatal clinics and from the study communities. Pregnant women who participated in the FGD were between the ages 17–45. All focus group discussions were held in the community in the dominant local language, Ewe, and lasted between 30 – 60 min. In all 12 focus group discussions of between 8 – 11 persons per group were held. All Interviews were moderated by members of the research team assisted by research assistants who were recruited and trained for the purpose. Interviews were tape recorded in addition to the interviewers notes and transcribed verbatim into English so all team members could read the notes. All interviews and recordings were done with informed consent (Additional file 1). Qualitative data was analysed manually using thematic content analysis of the text around the study questions guided by the analytical framework for the study. Routine management information systems data on enrolment was analysed in Excel for percentages, patterns and trends.