Background
According to the World Health Organization (WHO), health spending remains unequal across countries. The United States alone accounts for 42% of global health spending. More than 60% of global health spending is accounted for France, Germany, Japan, the United Kingdom and the USA [
1]. In 2015, the WHO estimated that, globally, around 400 million people lack access to at least one essential health service and that approximately 100 million are impoverished every year because of healthcare costs [
2]. Universal Health Coverage (UHC) means that everyone in the population has access to preventive, curative and rehabilitative healthcare when they need it and at an affordable price [
3,
4]. UHC aims to ensure that healthcare benefits are distributed based on the need for care and not on the ability to pay [
5,
6].
The achievement of UHC requires a commitment to three fundamental principles: (i) mobilizing adequate resources to ensure coverage, (ii) providing quality care through strengthening the health service delivery system and (iii) ensuring that health services are accessible to all impoverished and vulnerable individuals [
7]. The protection of people from catastrophic health expenditures is widely accepted as a desirable objective of health policy. Therefore, catastrophic payment from individuals’ available income can drop many households into poverty [
8].
Previous findings support that the poor have a higher incidence and higher out of pocket payments so that they are more likely to incur catastrophic health expenditures than the well-off [
6,
9‐
12]. To reduce the incidence of catastrophic health payments, the World Health Organisation recommends that total out-of-pocket expenses should not exceed 15–20% of national health expenditures [
8]. Lower-income countries lag behind on the road to UHC, especially many African countries. Although they have experienced good economic growth in the past two decades, improvements in health outcomes have been slow and uneven in many African countries [
13]. In 2018, the average government spending on healthcare in lower-income countries was only US$ 9 per capita, representing 1.2% of Gross Domestic Product with a marginal contribution of social health insurance which was only greater than US$ 5 per capita in some countries [
1].
In Africa, health-financing strategies vary broadly by geographic region and social context. According to [
14] the notable differences between countries regarding their health financing strategies show how health systems are influenced by social, cultural, economic and political factors resulting from the country’s context-specific. Although Universal Health Coverage (UHC) has become a political priority for many African countries, it has been challenging to achieve.
In the Democratic Republic of Congo, per capita, health expenditure remains low and largely below what other low-income countries have invested [
15]. According to [
16], at USD 13 per capita, DRC spends less than one-tenth the average of the rest of sub-Saharan Africa on health. In 2019, the country has allocated only 3.5% of the GDP and 8.5% of its budget to health financing. At a rate of 846 deaths per 100.000 births over 2007–2014, DRC’s rate of maternal mortality was higher than the average of the sub Saharan Africa of 510 [
16]. In their research, [
17] also emphasised the need for well-trained human resources for health who do not meet international standards in DRC’s rural areas. In the same rural areas, health facilities encounter large shock-outs of essential drugs and the population travel long distance to attend health facilities. Therefore, the financing gap for health is high as the country aims to achieve the UHC strategic objectives [
17].
Out-of-pocket payments account for more than 90% of household health expenditure [
16] as most Congolese are not part of any risk-sharing systems. If at all, the population of this country relies on voluntary community-based health insurance schemes [
17]. Therefore, on the path towards UHC, the DRC opted for a social system based on health insurance, in which community health insurances have a predominant role [
18]. This policy is an element of the healthcare reforms with the global agenda for universal access to health. In March 2016, along with other countries, the DRC agreed on a revised roadmap towards UHC in Brazzaville with a key UHC policy drawing attention to insufficient coverage in the country and to the fact that the contribution of informal sector households cannot meet the financing requirements [
19].
In 2009, the Democratic Republic of Congo joined the International Health Partnership. In 2016, IHP membership was transferred to UHC 2030. On the one hand, this commitment towards UHC holds the DRC accountable to provide financial risk protection by reducing household out-of-pocket payments, which contributes to the impoverishment of many Congolese. On the other hand, the DRC government has used this commitment as a critical instrument to hold donors more accountable for their obligations.
Our study examines changes and trends over time since committing to achieve the UHC on financial risk protection and health outcomes indicators. Specifically, we examine indicators related to the second and third principles of the UHC, which aim to ensure that health services are accessible to all impoverished and vulnerable individuals; and that the country is mobilizing adequate resources for health financing. It is crucial to examine the time-related effects of different health public policies adopted by the country to understand progress towards UHC in the context of DRC’s poverty, conflicts and fragility.
Although we acknowledge the existence of other essential plans and policies like Plan Directeur de Développement Sanitaire (PNDS) 2000–2009, PNDS 2011–2015 and PNDS 2016–2020, in this research we devote much attention to the time-related effects of the DRC’s commitment towards the UHC on financial risk protection and health outcomes indicators. The three strategic plans recognize the importance for DRC to align with the recommendations of the UHC. However, DRC’s strategic plans do not clearly define how a household can easily access health services. All the strategic plans do not specify a clear policy that allows the population to access healthcare through health insurance. As a result, roughly 90% of health financing in the country originated from households out of pocket.
Hence, our study answers the following question: Is the course of financial risk protection and health outcomes indicators associated with DRC’s commitment to UHC? This study highlights the changes and trends over time since joining the UHC on financial risk protection and health outcomes indicators.
Methods
Study design
Universal Health Coverage is an initiative aiming to provide all people with access to needed health services. DRC joined this initiative in 2009. We collected two types of indicators to measure the progress towards the UHC. The first category of indicators measures financial risk protection, and the second measures health outcomes. The main objective of our research is to analyse whether changes have occurred in the financial protection and health outcomes indicators between 2000 and 2018. If applicable, we also analyse when these changes happened.
Data sources
Our data originated from the World Development Indicators (WDI) of the World Bank and was complemented by the World Health Organization (WHO) database. The WDI is a compilation of comparable statistics about global development from officially international sources such as the WHO. The WHO database is a wide range of global health and well-being data sourced from their members’ states. The data collection is monitored by the World Bank (WB); the WHO data and the United Nations Inter-agency Group data are collected on an annual basis. Database closing is conducted at the end of each year with regard to the antedating year. Our study is based on data collected by the end of 2018.
Similarly, financial risk protection data are collected annually. These are aggregate data by groups computed by the WB based on the groupings for the respective fiscal year in which the data was released by WHO and WB. However, the main source of health outcomes data are vital registration systems and direct or indirect estimates based on sample surveys or censuses. This data is collected annually and compiled by the United Nations Inter-agency Group. Our research uses annual data from 2000 to 2018 to ensure balanced observation periods before and after the declaration of the DRC to adhere to UHC.
Variables
Table
1 exhibits the variable used for our models. We retained 5 indicators to capture financial risk protection while the health outcomes component is measured by 3 indicators. Overall, we have 8 indicators that define the basis for our parametric breakpoint regression approach which is applied to estimate changes and trends over time in relation to the year when DRC joined the UHC initiative.
Table 1
Definition of variables
OOP | Out of pocket as % of current health expenditures. | This indicator estimates how much are households in DRC spending on health directly out of pocket. It calculates the share of out-of-pocket payments over the total current health expenditures. | WHO | Financial Risk Protection |
GHE | Government health expenditure as % of general government expenditure | According to the WHO, this indicator reflects the extent to which healthcare is a priority for a country. | WB |
HI | Health Insurance (CHI) as % of Current Health Expenditure | Health insurance is a combination of voluntary HI and social HI (compulsory and from the government). | WHO |
TRADOM | Transfers from domestic government revenue (allocated to health purposes), as % of current health expenditure | This indicator refers to the funds allocated from domestic government revenues for health purposes. It shows the role of central and local governments in providing revenues of health financing schemes. | WHO |
TRAFOR | Transfers distributed by the government from foreign origin, as % of current health expenditure | This indicator shows transfers originating abroad (bilateral, multilateral or other types of foreign funding) distributed through the general government. For the financing scheme receiving these funds, the provider of the fund is the government, but the fund itself is from a foreign origin. | WHO |
LRMD | Lifetime Risk of Maternal Death | The lifetime risk of maternal death is the probability that a 15-year-old girl will die from complications of pregnancy or childbirth over her lifetime; it takes into account both. | WB | Health Outcomes |
PDC | Probability of dying among children ages 5–9 years (per 1000) | This indicator shows the likelihood of dying among children with ages comprised between 5 and 9 years old. | |
IMMUN | Immunization, DTP3 (% of children ages 12–23 months) | Immunization, Diphtheria, Pertussis, and Tetanus (DTP3). Child immunization measures the percentage of children ages 12–23 months who received vaccinations before 12 months or at any time before. | WB |
Control Variables |
GDP | GDP growth (annual %) | The Gross Domestic Product growth shows how fast the economy is growing. | WB | |
GGFC | General government final consumption expenditure (% of GDP) | This indicator consists of expenses incurred by the government in its production of non-market final goods and services (except Gross Fixed Capital Formation) and market goods and services provided as social transfers in kind; in % of GDP. | WB | |
EDS | External debt stocks (% of GNI) | This indicator shows in % of GNI the debt owed to non-residents repayable in currency, goods, or services. | WB | |
Statistical methods
Data on financial risk protection and health outcomes are described using descriptive statistics of central location (mean, median), the variability (standard deviation, interquartile range), and minimum/maximum. To examine our research question, we first compared the two observation periods (2000–2009 vs. 2010–2018) for all of the outcomes using the non-parametric two-sample Wilcoxon rank-sum test [
20]. This test requires at least ordinal measurements and computes the i
th rank of the pooled samples:
$${W}_{n_1,{n}_2}=\sum_{i=1}^{n_1+{n}_2}R\left({X}_i\right)$$
The test makes use of the U-statistic for each sample [
21]:
$${U}_1={n}_1\ast {n}_2+\frac{n_1\ast \left({n}_1+1\right)}{2}-{R}_1$$
$${U}_2={n}_1\ast {n}_2+\frac{n_2\ast \left({n}_2+1\right)}{2}-{R}_2$$
Here, R1, R2 represent the sum of ranks in each sample and n1, n2 the respective number of observations. The minimum of U1, U2 is compared with critical values to accept/reject H0 : P(X < Y) = P(X > Y) vs. H1 : P(X < Y) ≠ P(X > Y). Due to ties in the data approximate p-values were calculated. The Wilcoxon test allows conclusions on differences in the overall distribution prior to and after the declaration of the DRC to adhere to UHC.
In addition, the parametric approach of breakpoint regression has been applied [
22]. Synonyms of breakpoint regression are: change point, join point, or piecewise regression [
23]. This modelling approach enables for the identification of changes in the outcome determined as trends or slopes and the location of change points [
24]. Breakpoint regression has found several scientific applications: it has been applied, e.g., for the detection of changes in climate over time [
25] and in time-series data [
26]. One study examined annually aggregated data of suicides and found that changes of suicide methods over time coincided with governmental interventions, i.e. the results suggested strong associations of an intervention imposed on the population and the observed data on suicides [
27]. To identify breaks or location(s) of changes over time, the linear predictor of a linear regression model is specified, e.g. for one breakpoint, by:
$$y={\beta}_0+{\beta}_1\ast time+{\beta}_2\ast {\left( time-\psi \right)}_{+}+\epsilon$$
Here, the breakpoint is represented by the parameter
ψ,
β0 is the model intercept,
β1 the slope of time, and
β2 the difference in slopes introduced by a change of an association over time. The latter is computed by (
time −
ψ)
+ = (
time −
ψ) ×
I(
time >
ψ) with
I(·), the indicator function, being one if time is greater than
ψ the breakpoint. We used the R package
segmented [
28] in our study which estimates the optimal breakpoint in an iterative approach requiring only a vector of start parameters for
ψstart, the length of this vector corresponds to the number of assumed breakpoints. Several linear models are then fitted until convergence to find the optimal estimate of a breakpoint
\(\hat{\psi}\) [
28]. To define the optimal number of breakpoints the use of the bayesian information criteria
\(BIC=-2\mathit{\ln}\left(\mathcal{L}\right)+k\ast \mathit{\log}(n)\) [
29] is recommended for breakpoint regression [
28,
30]. Only estimates and respective confidence intervals will be provided for the parametric approach as the computation of
p-values has been shown to be slightly biased [
28]. For all outcomes, we compared models using none, one, two, or three breakpoints and presented adjusted R squared as well as BIC statistics. All statistical analyses were conducted using R version 4.1.1 [
31].
Conclusion
DRC’s UHC journey started in 2009. This paper has examined the changes in financial risk protection and health outcomes indicators during a 9-years interval (2010–2018) after DRC has committed to UHC. We implemented the parametric approach of breakpoint regression to detect whether the UHC journey has brought changes and when exactly the changes have occurred. Although OOP has improved, the results from the breakpoint regression support an adverse effect of DRC’s commitment toward UHC on household direct spending on health. Health insurance coverage is still deficient in DRC; we observe positive effects on health insurance coverage with progress over time. DRC’s government should invest a lot in improving the conditions of the health insurance market by defining new health insurance programs while setting norms and regulations for the overall system. The effect on health insurance and OOP reliance is minor due to the adoption of initiatives that focus on and favour wealthier people than the poorest, which roughly represent 73% of the population.
Our results suggest an improvement in the government’s transfers from domestic origins while the government’s transfers from foreign sources decrease. In 2008, the funds allocated to health from domestic origins increased up to 2012 when a negative breakpoint is identified. This trend shows a lack of consistency in setting health as a priority for the country.
The result of this study supports the target defined by the Abuja declaration as we found that DRC’s government does not allocate sufficient funds towards achieving UHC. While the increase in the government’s transfers to health from domestic origins is not sustainable, the effect on the transfers from foreign sources was not identified, suggesting that either the government did not properly allocate donors’ funds or donors did not meet their commitment.
Overall, the risk of maternal death has decreased significantly. Similar results appear for the probability of dying among children. However, the changes observed post-2009 are not statistically significant. These changes might be attributable to important investments from the private sector and NGOs in building hospitals and health centres in remote areas.
There is no conclusive effect on the indicator related to children who received vaccinations. In DRC, the vaccination rate of children mostly originates from donor initiatives such as GAVI, which has supported access to the vaccine among children in the country. These donor initiatives are important but mostly unsustainable and unpredictable. Moreover, DRC’s government should undertake additional efforts and provide infrastructure like electricity, roads, and refrigerators in short supply to deliver vaccines to remote health centres.
Therefore, our work suggests that although DRC’s UHC journey has slightly contributed to improving the financial risk protection and health outcomes indicators, much effort should be undertaken. In general, we make an essential empirical contribution that is relevant to the development and public policy towards the achievement of UHC. Using the case of DRC, we illustrate that many developing countries adhere to international initiatives but do not align their national policies to meet and monitor progress toward the set global agenda.
Our results reveal that the prevention of catastrophic health expenditure is still not a priority for the country and mostly for the majority of the poorest even after the DRC’s UHC journey has started. Hence, as we are approaching 2030, DRC’s government should pour into a well-structured and realistic operational plan the target goals to reduce the household’s financial burden of health expenses on households with a clear focus on the informal sector and poor households. The operational plan should focus on access to health insurance policies, which the country lacks. Achieving Universal Health Coverage in the Democratic Republic of Congo will require a set of differentiated policies for social groups. It is obvious that the poorest will not be able to pay out-of-pocket fees or an insurance premium. Thus, it might require mobilizing additional revenues to extend health coverage to the poorest. There is growing awareness that the “poorest of the poor” definitely require national (and international) solidarity. However, in DRC the majority of people are near-poor, i.e., they do not belong to the “poorest of the poor” but are constantly at risk of falling back into extreme poverty once an adverse event strikes them. For instance, the majority of the rural population of DRC works in the informal sector or in subsistence farming. Under “normal conditions” they can survive, but even a mild disease of the bread-earner of a major disease of a family member might bring them back beyond the poverty line. These people might be able to pay a token for social protection, but they will definitely also require a subsidy of their social protection contribution. To generate additional revenue to accelerate the agenda of UHC, the country should invest more in an effective tax collection system. With the additional revenue, the government could define an incentive programme to subsidize health insurance and healthcare for the majority of the poorest and the near-poor in the informal sector and rural areas.
For coverage expansion of the wealthy quintile of the population and the minority of the middle class, mandatory health insurance might be an option in the long term and should help to generate additional funds to support the government incentive programme.
The use of technology to support health policies should also be prioritized towards achieving UHC as the country faces logistical and infrastructure issues to monitor UHC’s progress in a vast country of 0.23 times as big as Europe. Setting technology at the centre of every health policy will enable the country to collect important data, which mostly are not available at the national level. Therefore, the country can rely on available data to promote information sharing and build accurate policies. Finally, DRC’s government should implement appropriate health insurance policies through well-organised social health insurance (voluntary community-based health) as many Congolese are involved in the informal sector to reduce OOP spending. The government can also subsidise hospital services for the poor. Implementing a robust monitoring and evaluation strategy will also ensure that the household’s health outcomes are improved as the country strives to implement appropriate and context-based health insurance regulations.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.