
Copyright © 2021
Inclusive Society Institute
50 Long Street
Cape Town
South Africa
8000
235-515 NPO
All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means without permission in writing from the Inclusive Society Institute
DISCLAIMER
Views expressed in this report do not necessarily represent the views of the Inclusive Society Institute or those of their respective Board or Council members or Members.
Authors: Percept Actuaries and Consultants
Edited: Daryl Swanepoel
17 June 2021
Foreword
South Africans deserve access to quality and affordable healthcare. This notion is indeed entrenched in the Constitution’s Bill of Rights that grants everyone the right thereto. It furthermore obliges the state to take reasonable measures to progressively achieve the right.
To this end, the National Department of Health has embarked on a process to establish a National Health Insurance (NHI) Scheme for South Africa. It is a health financing system that is designed to pool funds to provide access to quality, affordable personal health services for all South Africans based on their health needs, irrespective of their socio-economic status. These ideals cannot be faulted. Of concern to the Inclusive Society Institute (ISI), given South Africa’s current economic realities, is the affordability of NHI and whether it will be implemented incrementally, as funds become available. On the one hand, one would not want to disrupt the status quo without a clear understanding of the costs attached to the ultimate goals of the NHI, and on the other, the existence of the current world class health infrastructure needs to be assured for future generations. As the former German Federal Minister for Health, who oversaw the implementation of the national health system in Germany, Hon. Ulla Schmidt, puts it: Do not break down what you have until you have an alternative in place that works.
There is a dearth of studies into the costing of the proposed NHI. To this end, the ISI has ventured such a costing. The process has involved extensive consultation with a wide spectrum of health sector stakeholders, the discussions of which have culminated into a presentation of four alternative models to secure universal and affordable quality healthcare for all in South Africa. And, in an attempt to give direction, each alternative, together with the Department’s NHI proposal, has been costed. It is this costing that is contained in this report.
The report is not a criticism of the Department’s NHI proposals, neither does it profess to be accurate in all respects. What it is, is an attempt to ensure feasibility and sustainability, a fresh look at the approach to implementing universal healthcare, and an assessment as to how best to ensure the progressive realisation of the basic right to quality and affordable healthcare for all. As regards the costing: undoubtedly there is room for improvement. The public debate in this regard is bound to identify gaps that could improve the modelling. That said, the Institute is confident that the report is a sound starting point on which dialogue can build.
The Institute chooses not to be descriptive as to the final path that legislators need to take. It however hopes that the policymakers will accept this contribution as a further evaluation tool to ensure that the finally chosen funding model delivers for future generations, a legacy that is indeed capable of delivering on its goals.
Executive summary
This document sets out the costs of the four alternative universal health coverage (UHC) scenarios that were presented in the previous paper released by the ISI, entitled ‘Reimagined Pathways for UHC in South Africa: A critical policy assessment of NHI choices’ (2020). The scenarios we’ve analysed are described in Table 1. The four scenarios are alternatives for the fifth scenario, NHI as Described in the Bill.
The proposed reforms as they stand in the Draft NHI Bill represent a combination of policy building blocks sequenced in a particular way. This paper aims to unpack how and where these policy blocks could be reshuffled or re-envisioned to better achieve UHC, given the current South African health and economic outlook.
There are several reasons why it’s important to consider the costs of alternative scenarios relative to the likely costs of NHI as it’s currently envisioned. Firstly, it will allow policymakers, citizens and civil society to be cognisant of the trade-offs between cost, quality and access. Without transparency in trade-offs, it’s impossible to fully grasp the long-term consequences of a particular policy choice relative to others. Secondly, once the trade-offs and considerations related to a particular UHC path are transparent, policy choices can be reconsidered, if needed. Understanding what drives the costs and outcomes of each scenario allows for the adjustment of policy options, which will help South Africa reach UHC as efficiently and effectively as possible.
The alternative UHC scenarios presented in the previous report are shared below.
Table 1: Alternative UHC scenarios explored in ISI (2020)

How the model was constructed
The cost model presented here uses the current health system structure and the combined total health expenditure in the public and private sectors (including out-of-pocket spending) as a starting point. We believe it’s a pragmatic approach to work with what we have, rather than with what we wish we did.
As a reminder, we include a summary table to tell the story of what is likely to happen to the South African health financing sector as we currently know it under each scenario.
Table 2: Summary of how different components of SA health financing and delivery system are affected in different scenarios

We make assumptions about the main demographic and structural cost drivers to arrive at different costings for the five healthcare scenarios, while also considering macro-level indicators such as the consumer price index (CPI), healthcare cost inflation, population growth and GDP growth over the 2021-2040 period. Given that there are many uncertainties about how these variables will play out over time, we’ve provided our best estimates given the available information.
Changes in total health expenditure (THE) can be understood in terms of three main categories of expenditure:
Publicly funded: centralised public expenditure (a form of risk pooling, even if not explicitly managed as a risk pool);
Private risk pooling: the vast majority of this is located in medical schemes; and
OOP (out-of-pocket expenditure) and (non-medical scheme) private insurance by households for private and public services not covered by either medical scheme funding or public-sector expenditure. This includes, for example, private pharmacy expenditure by those who consult the public sector, or private GP visits for those who don’t have medical scheme cover.
Collectively, these three categories form the total funds currently available for health expenditure, unless other funding sources – such as a mandatory payroll tax, income tax, health taxes or value-added tax – are enacted to increase the available amount for spending on health services.
The cost model creates a framework that makes the trade-offs between different policy choices evident, and this is the real value of the model. It shows how total health expenditure, and its various components are likely to play out over time if we do or don’t take certain actions now. It makes the long-term costs of inaction evident, and in doing so, provides valuable information that can be used for NHI policy development.
Detailed assumptions about the overall financing structure and cost drivers for each scenario are shared later in this report.
What the model says about the costs of the various scenarios
By 2030, Status Quo Gold Standard is the most expensive model in terms of total healthcare expenditure. In 2040, NHI as Described in the Bill is the most expensive model. In contrast, Power to the People is the least expensive model in 2030 by a slight margin, while Reorienting Towards Value is the least expensive scenario by 2040. We also see that Reorientating Towards Value starts off relatively expensive due to the investments needed to reorientate a health system. However, by 2035, this scenario is the most affordable and sustains this position in 2040.
Why does NHI as Described in the Bill emerge as the most expensive scenario? UHC implemented through a single-risk fund is not necessarily expensive. However, the current description in the Bill, with its lack of detailed and clear information on accountability structures, creates uncertainty about whether it will be implemented with the necessary attention and rigour. The NHI Rejigged scenario shows the potential for NHI to achieve its goals, if it’s implemented with the current reality and context in mind, which would make the milestones more achievable and realistic. NHI Rejigged is less expensive than NHI as Described in the Bill over the longer term as the greater accountability and information and data systems with which it is implemented will greatly reduce administration costs and wastage (through corruption). Because NHI Rejigged has a longer preparation phase during which the quality of public health facilities will be improved, there will be more buy-in from system users and lower out-of-pocket expenditure later on.
In calculating the relative costs of estimated total healthcare expenditure, three different growth scenarios are used: mid-growth (1% real growth), low-growth (0% real growth) and high-growth (3% real growth). We only share the results of the mid-growth scenario here. In the mid-growth scenario, the most expensive health financing scenario, NHI as described in the Bill constitute 13.2% of the GDP by 2040. Such a large percentage of GDP being allocated to total health expenditure relative to current expenditure levels (around 8.7% of GDP) is unlikely to be affordable relative to other social priorities, but this must be viewed considering low to modest economic growth and a system that would have pushed many people to increase their OOP expenditure (and therefore total healthcare expenditure) by that time. The least expensive scenario costs R638 billion (real 2020 Rands), with the costs of the other four scenarios ranging from R720 to R885 billion (also real 2020 Rands).
What the costs and scenarios mean for policy choices?
The model’s results help to illustrate what is gained and lost through different policy choices, relative to the objectives being pursued by moving to UHC: equity, access to quality care and quality outcomes, efficiency, stewardship and governance.
These gains and losses are not just about the financial costs of system change, but ultimately also about healthcare outcomes and overall system resilience.
Equity

The goal of equity should take into account both where funds come from (with contributions based on the ability to pay), and where funds are going (with more funds directed to those with the greatest need). The option that fares worst in terms of equity is Status Quo Gold Standard, where it’s assumed that the public and private sectors mostly continue operating as they are, albeit with improved quality and efficiency. NHI as Described in the Bill doesn’t fare well either due to the high OOP expenditure dependency. Those who can pay out-of-pocket will likely be able to access better care, resulting in a highly inequitable outcome.
Access to quality care and quality outcomes

Comparing policy options based on cost can mask what really matters – namely whether we can expect the future system to help the population thrive. But which clear signs indicate that the system is performing? System performance is visible in overall outcome metrics (such as life expectancy and maternal mortality rates), disease-specific outcomes metrics (which our system currently lacks) and through secondary measures (such as the extent of medical malpractice claims).
Apart from considering how many units of care a system can afford, there is a range of dynamics that impact a system’s ability to deliver high-quality care. They include:
Investment in the service delivery platform – including infrastructure and IT systems. Financing reforms have limited scope to strengthen the quality of care if the service delivery platform is compromised. This is illustrated in the cost comparison of NHI as Described in the Bill and Status Quo Gold Standard.
The balance between primary and hospital care. A weak system will have too many resources directed towards relatively expensive hospital care due to a bypassing of referral pathways, weak preventative care, and late intervention. This is illustrated in the cost comparison of NHI as Described in the Bill, Power to the People and Reorienting Towards Value.
Systems that are patient-oriented, that have greater bottom-up accountability and that measure and incentivise quality care tend to produce better health outcomes – and ultimately deliver greater value for the money invested in the system. This is illustrated in the cost comparison of NHI as Described in the Bill, Power to the People and Reorienting Towards Value.
Strong systems that support health outcomes. From a patient perspective, this means strong linkages between care, better continuity of care and strong care coordination. System performance is supported by system stability (i.e. the absence of system shocks), a balance between centralised support and ground-level responsiveness, and ongoing investment in the service-delivery platform. Big-bang reform (NHI as Described in the Bill) is likely to weaken system resilience in an already fragile system. Policy options that support ongoing quality improvement and system strengthening are likely to outperform over the long term. This was clearly illustrated by the global impact of Covid-19, where countries with resilient and unified health systems have outperformed countries with weak ones.
Healthcare worker capacity and satisfaction. Healthcare outcomes rely on having healthcare workers who deliver quality care. A large monopsony creates the risk of a system that isn’t sufficiently oriented towards supporting healthcare workers and enabling supply-side innovation.
Efficiency: Accountability to users (bottom-up) and societal buy-in as levers

Much of the rationale for NHI as Described in the Bill is the creation of a single purchaser to achieve efficiency through strategic purchasing. There’s no doubt that the current system is weak from a strategic purchasing perspective. This is likely to continue in Status Quo Gold Standard, and care needs be taken in NHI Rejigged to improve the structural impediments to strategic purchasing in both the public and private sectors. The Health Market Inquiry (HMI) recommendations are very relevant in realising this goal.
Strategic purchasing is necessary, but not sufficient for improving system efficiency. Efficiency can also be supported in the following ways:
Offering those who use the system a choice of funder and the ability to move. Having multiple funds can create competitive pressure based on strategic purchasing (if, for example, funders have to publish value metrics). The risk of a single fund is complacency and a lack of customer-centricity. This is illustrated in the comparison of NHI as Described in the Bill and Power to the People, where individuals can choose between joining different UHC funds.
Value-based approaches shift the responsibility for both quality and cost to healthcare providers – who happen to be best placed in the system to innovate the ways in which care is delivered. This removes layers of administration centred on managing providers of care – with the emphasis shifting to empowering providers. Value-based systems are most likely to have supply-side innovation, which allows for a move towards more efficient models of care over time.
Minimising the cost of corruption, waste, and abuse. Central control of a system seems attractive from an efficiency perspective. However, large institutions increase the risk of large-scale corruption. A system with strong bottom-up accountability and empowered healthcare providers may be more efficient in the long term than a single fund system that’s susceptible to bureaucracy and governance failure.
Stewardship and governance

Table 3 and Table 4 show the additional cost burden for public healthcare and for total healthcare, respectively. These are shown per scenario and are considered relative to various sized potential payer groups, ranging from the total population (referred to as per capita) to only those who are active taxpayers. We present the additional cost burden in 2040, relative to the baseline cost(a) burden in 2040. We do not, however, aim to determine the optimal financing source for additional health expenditure.
The additional cost burden can be spread across groups of many sizes. Considering just the public health expenditure then the largest group (the total population) would have a financing implication of R461 (2020 real Rands) per month per capita by 2040 for the most expensive scenario. For the smallest group (active income taxpayers), there would be an additional financing implication of R4,267 per month per taxpayer. If we consider the total health expenditure (as opposed to just public) then we get a different picture. For the most expensive scenario, there would then be a financing implication of R229 per month per capita (the total population), or R2,121 per month per active taxpayer (smallest group). The most expensive option for public health expenditure is NHI Rejigged, while for total health expenditure it is NHI in the Bill. The differences are driven by OOP and private risk pools where appropriate.
Table 3: Additional PUBLIC health expenditure burden of different scenarios relative to potential payer groups

*(a) Baseline cost in 2040 is calculated by growing the 2020 expenditure by the population, i.e. it is how much we would spend in 2040 if each we continue to spend the same amount per capita.
Table 4: Additional TOTAL health expenditure burden of different scenarios relative to potential payer groups

The imperative for economic and employment growth is evident when we take a long-term perspective and consider how best to achieve health-system objectives within fiscal constraints. An expansion of the economy will permit much-needed investment in the health system, whilst employment growth will allow for a more robust tax base to enable income cross-subsidies and sustainable social solidarity. Any changes to the health financing system cannot be consider without understanding the strong need for economic and employment growth. Without this growth, far more gradual choices will have to be made.
Content
Foreword
Executive summary
List of figures
List of tables
Acronyms and abbreviations
1. Introduction
2. What are we trying to achieve? First principles
3. What do we have to spend on health services? Recent trends in healthcare expenditure
4. Previous NHI costings
5. NHI and the four UHC alternatives
6. Assumptions informing cost-modelling for each scenario
- Status Quo Gold Standard: No purchaser-provider split and strengthening the
public sector
- NHI Rejigged: NHI, but sequenced differently
- Power to the People: Purchaser provider split but with multiple purchasers
- Reorienting Towards Value: a value-based approach to UHC
- NHI as Described in the Bill: the policy proposal currently on the table
7. Approach to cost modelling and key cross-scenario assumptions
8. Results of cost modelling for the scenarios
9. What do the model results mean for policy? The trade-offs between NHI and other UHC
options
- Why does NHI as Described in the Bill cost so much more than the other models?
- Considerations in weighing the different policy options
10. Conclusion
References
Appendix A: Out-of-pocket health expenditure in South Africa
Catastrophic health expenditure
Expenditure in absolute terms
Appendix B: Additional detailed model results
Appendix C: Model assumptions by scenario
List of figures
Figure 1: Annual public health expenditure per capita uninsured (Real)
Figure 2: Public health expenditure per capita uninsured, by province (Real)
Figure 3: PHC utilisation rate by province
Figure 4: Professional nurses per 10k population uninsured
Figure 5: Total healthcare expenditure by scenario (Rands, standardised so NHI Bill scenario is R100 each year)
Figure 6: Real total healthcare expenditure (2020 Rand billions) by scenario for 2030 vs. 2040
Figure 7: Breakdown of public sector costs by scenario for 2030 and 2040
Figure 8: Total real expenditure (2020 Rand values) per capita per year over time
Figure 9: Real total health expenditure growth over time vs. population and medical inflation growth combined
Figure 10: Total health expenditure as % of GDP (1% growth, mid-growth scenario) by scenario over time
Figure 11: Total health expenditure as % of GDP (3% growth, high growth) by scenario over time
Figure 12: Total health expenditure as % of GDP (0% growth, low growth) by scenario over time
Figure 13: Real total costs (2020 Rand billions) by scenario over time
Figure 14: Real expenditure over time (Rand billions)
List of tables
Table 1: Alternative UHC scenarios explored in ISI (2020)
Table 2: Summary of how different components of SA health financing and delivery system are affected in different scenarios
Table 3: Additional PUBLIC health expenditure burden of different scenarios relative to potential payer groups
Table 4: Additional TOTAL health expenditure burden of different scenarios relative to potential payer groups
Table 5: Costs and implications of various NHI costing models
Table 6: Alternative UHC scenarios explored in this document
Table 7: Assumptions driving costs for Status Quo Gold Standard in model framework
Table 8: Assumptions driving costs for NHI Rejigged in model framework
Table 9: Assumptions driving costs for Power to the People in model framework
Table 10: Assumptions driving costs for Reorienting Towards Value in model framework
Table 11: Assumptions driving costs for NHI as Described in the Bill in model framework
Table 12: Model variables
Table 13: Equity
Table 14: Access to quality care and quality outcomes
Table 15: Efficiency: Accountability to users (bottom-up) and societal buy-in as levers
Table 16: Stewardship and governance
Table 17: Potential payer groups
Table 18: Additional PUBLIC health expenditure burden of different scenarios relative to potential payer groups
Table 19: Additional TOTAL health expenditure burden of different scenarios relative to potential payer groups
Table 20: Concepts and indicators used in OOP health expenditure studies
Table 21: OOP health expenditure in South Africa
Table 22: 2030 detailed model results
Table 23: 2040 detailed model results
Table 24: Model assumptions by scenario
Acronyms and abbreviations

1. Introduction
South Africa currently has a draft NHI Bill before Parliament1. The period for submissions on the Bill has ended, with a record number being received2. The significant number of submissions indicate the concern among civil society, funding bodies, health sector stakeholders and citizens about the massive step-changes that have been proposed for the health sector3.
The Covid-19 pandemic continues to highlight the extent of inequities between the public and private sectors – as well as between provinces and districts4–6. Furthermore, the pandemic-related impact on health system resilience provides an impetus to deepen universal health coverage (UHC).
The set of reforms in the draft NHI Bill is only one way to achieve the policy goals that would put South Africa’s health system in a stronger position to handle the impact of Covid-19, as well as other challenges and stressors. However, there are also other ways of conceptualising and implementing a move to UHC.
In a previous report7, four alternative scenarios for achieving UHC for South Africa were shared. They are:
· Status Quo Gold Standard;
· NHI Rejigged;
· Power to the People; and
· Reorienting Towards Value.
At their core, each scenario addresses the five UHC policy objectives embedded within them (see Section 2) and present a reimagining of the ‘how’ for UHC.
This report aims to stimulate debate and discussion about alternative policy pathways. Investigation the potential long-term costs of policy change is a useful way to bring the implications of both action and inaction to light.
We’ve modelled the potential costs of the four scenarios and NHI as Described in the Bill and provided high-level costings for the five scenarios from 2021-2040, with 2030 as a key midway point. Most scenarios improve equity and resource-sharing in healthcare between the private and public sectors. It is also demonstrated that while some scenarios may perform better over the medium term, their real success in terms of cost, quality, efficiency, stewardship and governance can only be assessed over the longer term (2040). Long-term costings are by nature highly uncertain. The costings presented here are illustrative and intended to highlight the underlying factors driving costs, and the trade-offs between policy decisions.
What the model offers:
The cost model presented here uses the current health system structure and combined health expenditure in the public and private sectors (including OOP) as a starting point. We believe it’s a pragmatic approach to work with what we have, rather than with what we wish we did.
We make assumptions about the main demographic and structural cost drivers to arrive at estimated costs for the five healthcare scenarios, while also considering macro-level indicators such as the consumer price index (CPI), healthcare cost inflation, population growth and GDP growth from 2021-2040. More details on these cross-cutting variables (not specific to any one model) are provided later in the report. The assumptions informing the variables are based on available evidence and aim to illustrate the differences between the policy choices. They should be seen as point estimates in a wide range of possible outcomes. By nature, they’re subjective, and useful to stimulate discussion and debate.
Considering the likely costs of each scenario, we looked at the total cost of the health system, which is consistent with the idea of a unified system. It’s also helpful in terms of considering the total envelope of funding directed towards healthcare, regardless of the source of the funding.
Total healthcare expenditure encompasses three main categories:
1) Publicly funded: centralised public expenditure (a form of risk pooling, even if not explicitly managed as a risk pool given that resource allocation is decentralised, as well as the absence of a fully articulated basic benefit package);
2) Private risk pooling: the vast majority of this is located in medical schemes; and
3) OOP expenditure and (non-medical scheme) private insurance by households for private and public services not covered by either medical scheme funding or public-sector expenditure. For example, the private pharmacy expenses of those consulting the public sector, or the GP visits of those who don’t have medical scheme cover.
Collectively, the three categories represent a ceiling on funds currently available for health expenditure, unless other ways are found to increase the available amount – such as an earmarked health tax or increases to payroll tax, income tax or VAT. A large proportion of this current funding envelope is discretionary, with households and employers making voluntary contributions to medical schemes or paying for care on an out-of-pocket basis.
The cost model allows for compelling storytelling that clearly highlights the trade-offs between the different policy choices. This is the real value of the model: to show how the total healthcare costs and their various components are likely to play out over time if we do, or don’t, take certain actions now; or, if certain much-needed steps aren’t taken in future. It highlights the long-term costs of inaction, and therefore, helps to inform both policy and implementation. Previous costings of the NHI proposals have compared costs to the current system (doing nothing to change the system) – we hope to make it clear that there are other possible counterfactuals.
There are several reasons why it is important to consider the costs of alternative scenario relative to the likely costs of NHI. Firstly, it will allow policymakers, citizens and civil society to be aware of the trade-offs between equity, cost, quality and access. Without transparency in trade-offs, it’s impossible to know the long-term consequences of a particular policy choice relative to others. Secondly, once the trade-offs and various considerations of the various UHC options are clear, it will be possible to reconsider policy choices, if necessary. Once it’s clear what is driving the costs and outcomes of a policy direction, the policy can be adjusted to improve costs and outcomes. This could entail tweaking the overall financial structure and approach of a UHC policy option, or opting for a completely different approach altogether.
What the model is not:
While offering projections for the five scenarios, this is not a detailed bottom-up costing model. We have not evaluated the finer points of a basic benefit package that are needed for a bottom-up costing, since no such package for NHI was set out in the draft NHI Bill – or in any other official government document. It is useful to keep the intention of the model in mind while reading the results of the five scenarios, while also paying attention to its limitations.
The hypothetical nature of the UHC scenarios, as well as the NHI Bill, also constrained our ability to accurately estimate the associated costs. For example, the NHI Bill provides scant information on the single-purchaser model, which is needed for a detailed bottom-up costing. We therefore often had to make bold assumptions about both NHI as Described in the Bill as well as the other scenarios to determine ultimate costs and illustrate the differences between policy choices. The results are normative as they’ve been informed by a large set of hypothetical assumptions. They can’t be viewed as objective cost assessments of the policy options given the multi-layered and interconnected nature of the cost assumptions. While the model is potentially subjective, it does provide a much-needed framework informed by logic.
In the absence of real policy alternatives that have been clearly articulated in terms of service and funding design, the estimated future costs for the different scenarios provide a starting point for us to collectively imagine alternative ways to achieve our UHC objectives.
Report structure:
Section 2 of this report provides a summary of the five objectives of UHC (as identified in the previous ISI report, ‘Reimagined Pathways for UHC in South Africa: A critical policy assessment of NHI choices’ (2020). Section 3 sets out the current total healthcare expenditure in South Africa. Section 4 offers a review of previous NHI cost models in the public domain. Section 5 provides high-level descriptions of each of the five scenarios. Section 6 details the assumptions that underpin the cost modelling for each scenario. Section 7 sets out the structure of the model and the nature of the cross-cutting variables. Section 8 presents the results of the model and in Section 9, the implications of the results for policy are articulated. Section 10 provides a conclusion.
2. What are we trying to achieve? First principles
South Africa has been struggling to find its path to UHC for more than two decades. Despite differences in policy, rhetoric and plans, the intention has always been the same: to improve equity between the provinces and health sectors, to ensure quality care is provided at the lowest possible cost, and to ensure that the health system is governed optimally and with sound leadership. To support this report, we conducted a comprehensive literature and stakeholder review of NHI in South Africa8. Based on the review, five policy objectives were identified that have been consistently present across policy documentation as South Africa has been trying to move closer and closer to UHC. These policy objectives are:
To improve equity in the health system, including the sharing of resources (human and other) across the public and private health systems.
To address escalating costs in the private health sector and contain future escalations in costs across the health sectors.
To provide universal access to quality healthcare.
To ensure efficiency in service provision and administration.
To ensure good governance and stewardship.
These key objectives of the health reform process will be used as benchmarks to assess the feasibility of various policy choices and pathways that were put forward during the UHC debate. This report may also assist in guiding the sequencing of current reform proposals. Throughout this document, we propose that the combination of cost, quality and efficiency could collectively be reconceptualised as orienting the South African healthcare system towards value.
3. What do we have to spend on health services? Recent trends in healthcare expenditure
Since 2012, the public health budget in South Africa has become severely constrained as a result of the deepening austerity after the 2008 global recession9. Prior to 2012, the National Treasury was able to shield the social sectors from the declining fiscal environment, but from 2012 onwards, health, education and other social service sectors also bore the brunt of South Africa’s ongoing fiscal austerity measures. Figure 1 shows the real expenditure per capita uninsured from 2010/11 to 2019/20 in South Africa10. The dip in health spending in 2012/13 is evident, as is the erratic nature of the public health allocation over the 10-year period.
Figure 1: Annual public health expenditure per capita uninsured (Real)

Our analysis shows little predictability in the health budget year-on-year, which impacts the sector’s ability to plan appropriately. The analysis also shows that in the decade between 2010/11 and 2019/20, the overall per capita spend for the uninsured population has only grown by 15%. This growth is worrisome given that medical inflation constantly supersedes average general inflation (CPI) and we would anticipate at least a CPI-linked increase for each year, which would result in a growth rate well above the 15% shown. Once the above-CPI price increases have been taken into account, it has not left much additional growth to account for a growing disease burden and greater health needs. This has no doubt contributed to the dire conditions in the public health sector in terms of its dilapidated infrastructure, limited human resources for health and rising quality concerns – the Eastern Cape being an example of this decline11.
Figure 2 shows the average expenditure per capita uninsured by province. The erratic nature of budget availability is clear over the period, highlighting the difficulties that an uncertain budget creates for health sector planning. The impact of this on a long-term plan to rollout NHI is immediately evident; there is no way to be certain that next year’s budget will be able to include new, major health reform costs.
Figure 2: Public health expenditure per capita uninsured, by province (Real)

Figure 3 provides a clear picture of the impact of this declining budget on service delivery. All the provinces show a substantial decline in primary healthcare (PHC) utilisation rates between 2010/11 and 2019/20. Accessible and quality PHC is a critical part of keeping a population healthy – with the reduction in utilisation of these services, one can expect the population to become sicker as the system is unable to perform key prevention and health promotion activities.
Figure 3: PHC utilisation rate by province

Another casualty of the declining budget has been the sector’s inability to hire more human resources. Professional nurses are the backbone of any health system and they’re particularly integral to a well-functioning PHC system. Figure 4 shows the professional nurses per 10k population uninsured per South African province. Almost every province shows a declining trend.
Figure 4: Professional nurses per 10k population uninsured

The public health budget has been declining since 2012/13. This has placed severe pressure on the sector as it attempts to maintain service delivery and respond to new crises of non-communicable diseases. Expenditure has increased by only 15% over the past decade, pointing to a severely constrained fiscal climate. The effect of this is evident in both utilisations, the overall health status of the population and rising medical negligence claims as facilities are unable to render effective and quality care with the resources at hand.
4. Previous NHI costings
There have been several NHI costing models in South Africa, only some of which are in the public domain. For those in the public domain, this report provides an overview of the costing approach used, as well as a high-level summary of both the absolute and relative costs.
Econex(12) used a demand-based approach (in 2010) to estimate the utilisation (demand) for different sub-groups of the population, and for different cost components, and then aggregated it. Based on a comprehensive package that incorporated all the essential features of NHI at the timeb, utilisation patterns of the uninsured population, unchanged unit costs and a conservative addition of R8 billion for NHI public health administration costs, Econex estimated the cost of a full NHI to be R287 billion (in 2008 Rand values). This upper threshold of their costing estimate (base case) was more than four times the total amount spent by the state on healthcare at the time (R62.8 billion in 2008 Rands), and almost double what was spent on total healthcare at the time ((R144.5 billion in 2008 Rands). In addition to the base case, Econex estimated five alternative costing scenarios that made varying assumptions about the degree of savings and rationing. At the lowest costing thresholdc, the total NHI cost was estimated to be R174 billion (in 2008 Rand values). If the costing scenarios were extended to account for the impact of HIV/AIDS, the authors added R6 billion (in 2008 Rand values) to their cost estimates.
McLeod, Grobler and Van der Berg(13) used an actuarial costing method in 2009 to estimate the cost of NHI. The model accounted for both the prevalence of various conditions and the cost of treating each, taking into account the age and sex distribution for a reference population. The reference population was drawn from the ASSA2003 provincial model14, as recommended by McLeod d 15.
McLeod, Grobler and Van der Berg’s model estimated NHI costs for five different packages across four levels of delivery efficiency. Their findings showed that a fully comprehensive package (with all healthcare benefits) delivered at the medical scheme efficiency level (100% of cost) would amount to R334 billion (in 2009 Rand values). This was the upper-most limit of their NHI costing range. At the lowest threshold of the range, a package of Prescribed Minimum Benefitse delivered at the staff model efficiency level (50% of cost) was estimated at R78 billion (in 2009 Rand values) for the same year. These costings excluded NHI administrative and managed-care costs, and didn’t fully account for the impact of HIV/AIDS. For these reasons, the authors regard their NHI costing as preliminary.
The McIntyre NHI costing approach of 201016 could be described as demand-based, as cost estimates were derived from the product of population, utilisation rates and unit costs. For the universal coverage scenario, the model estimated the cost of public healthcare if the population were to be covered by the public sector. The benefit package used by McIntyre leaned more towards the public sector framework of service delivery than the private sector framework, so the services covered were limited and explicit. However, the unit costs in this model were higher than public sector unit costs at the time – to account for the required improvements in public sector resourcing or the option to purchase services from the private sector. McIntyre also assumed a substantial increase in health services utilisation, particularly at the primary healthcare level. NHI administrative costs as well as the cost of HIV/AIDS were accounted for f in the modelling. Costing estimates for the universal coverage scenario g ranged from R193 billion to R200 billion (in 2010 Rand values).
It’s clear from all three models that the estimated cost of NHI (in GDP terms) was for the most part incredibly close to estimated total health expenditure (THE), except for a small margin that was allowed for OOP and other private risk pooling schemes. Given uncertainties about the elements that would constitute the basic benefit package, some of the relative upper-cost thresholds were more than total healthcare expenditure, going as high as 14.1% of the GDP. Such a high allocation of total GDP towards healthcare expenditure means that resources would have to be directed from other types of expenditure – possibly education, social development or even housing.
*(b) Universal coverage of the population (irrespective of contribution), comprehensive cover, service to be sought from the provider of choice, no co-payment.
*(c) After accounting for severe rationing and savings.
*(d) This policy brief strongly recommended that all NHI costing work use the ASSA2003 provincial model (or updated versions thereof).
*(e) These are the benefits that medical schemes are required to delivery by law.
*(f) Used the ASSA2003 AIDS and Demographic model 43.
*(g) Inclusive of those who chose to remain on scheme cover over in addition to their universal cover.
In addition, there are several NHI costing models not in the public domain that have been commissioned and produced by different organisations:
The PWC model produced for the Ministerial Advisory Committee;
The 2016 model by the Clinton Health Access Initiative (CHAI) and Insight Actuaries & Consultants;
The Government Technical Advisory Centre (GTAC) 2018 model;
The GTAC 2019 model; and
The 2018 Actuarial Society of South Africa model.
Table 5: Costs and implications of various NHI costing models

*(1) Own calculations: numerator sourced from respective NHI costing model; denominator sourced from relevant National Treasury Budget Reviews
*(2) Own calculations: numerator sourced from respective NHI costing model; denominator sourced from relevant National Treasury Budget Review
5. NHI and the four UHC alternatives
In the earlier report 7, the four scenarios that were presented (Status Quo Gold Standard, NHI Rejigged, Power to the People, and Reorienting Towards Value) allowed the reader to reimagine UHC implementation in South Africa. We conducted a comprehensive literature review before the report and also used interviews and documentation to differentiate the scenarios.
Each scenario has the five policy objectives embedded into its core, and therefore represent a reimagining of the ‘how’ for NHI (see Table 6). The scenarios may help the South African government and various UHC stakeholders to continue furthering the important UHC agenda, without risking the public purse or service continuity.
Table 6: Alternative UHC scenarios explored in this document

6. Assumptions informing cost-modelling for each scenario
The assumptions driving the cost model structure for each scenario are set out below. Each discussion starts with a narrative summary of the overall structure and anticipated changes.
Status quo gold standard: no purchaser-provider split and strengthening the public sector
Summary
The public sector is improved while the private sector is left to continue as is. Medical schemes remain operational in their current form and OOP expenditure from private sector users follows current trends, while reducing for public sector users. The state continues its role as funder for the dependent population, with provinces as purchasers. Public facilities are improved and regularly assessed against a set of standards to monitor quality. If facilities fail to meet the standards, there is an intervention to swiftly address issues. The public sector budgeting process is done in a transparent manner, contributing to a reduction in corruption.
What drives costs in this scenario?
Under this scenario (Table 7), costs relative to the current health expenditure will initially increase due to expenditure on health information systems (DHIS2 and PERSAL, as examples), while focused quality improvement approaches in the public sector will also lead to initial cost increases. Over time, however, these investments pay off and quality improves dramatically. A lack of competition in the purchaser space may result in sluggishness and sub-optimal performance and total health expenditure costs may not be as low as possible.
Table 7: Assumptions driving costs for Status Quo Gold Standard in model framework

NHI rejigged: NHI, but sequenced differently
Summary
NHI is eventually implemented as envisaged in the NHI Bill, but it occurs more gradually. Medical schemes are phased out and the public sector serves the entire population through the NHI Fund. While still relevant, the private sector is regulated, largely in line with the main recommendations of the Competition Commission’s Health Market Inquiry 17. The public sector is improved as outlined in Status Quo Gold Standard and develops the capacity to contract from private providers. The process to achieve NHI is incremental and by the time it is implemented, there will be a greater capacity to do so successfully and efficiently. OOP expenditure decreases over time as the state covers a greater proportion of needs. The incremental implementation of the fund provides time to establish governance structures that lead to a reduction in corruption.
What drives costs in this scenario?
Similar to the previous scenario, many of the initial costs for NHI Rejigged (Table 8) will be driven by data system investments as well as concerted efforts to improve quality in the public sector through certification and a broader quality improvement process. The implementation of the Health Market Inquiry’s recommendations will over time lead to relative cost reductions in the private sector. Relative to the two multi-payer scenarios (Power to the People and Reorienting Towards Value) the eventual monopsony payer (NHI Fund) may lead to a loss of efficiency due to the absence of competition.
Table 8: Assumptions driving costs for NHI Rejigged in model framework

Power to the people: purchaser provider split but with multiple purchasers
Summary
NHI end results are largely achieved as set out in NHI Rejigged, with the primary difference being the multi-purchaser environment. A purchaser-provider split is still maintained. Provincial Departments of Health (PDoHs), local municipalities and private providers are providers of care. There is a process to apply to be a purchaser, but the envisaged purchasers are the government body (through a fund such as the NHI) and medical schemes (that would be appointed by the relevant body to act as purchasers for UHC) as they have experience in performing the purchaser role. Users can select their preferred purchaser, who is then allocated funds accordingly. There is a clearly defined benefit package in place and these are used in the contracting between purchasers and providers. As with NHI Rejigged, the public sector is improved, medical schemes in their current form are closed and OOP expenditure and corruption decrease over time.
What drives costs in this scenario?
This scenario (Table 9) has many of the same cost drivers as the NHI Rejigged scenario (e.g. information systems and quality improvement costs), with some additional cost drivers. The big potential cost saving factor associated with this scenario the choice given to citizens between multiple purchasers. This potentially allows for better bottom-up accountability, compelling purchasers to move to greater efficiencies, which could lead to lower administrative costs and lower health expenditure over time.
Table 9: Assumptions driving costs for Power to the People in model framework

Reorienting towards value: a value-based approach to UHC