This chapter presents the book's conclusions, including the key findings from each previous chapter. It also presents the solutions for low performance in relation to social determinants, health funding and health expenditure which were not included in previous chapters, to consider what they can add to our findings. It includes a further discussion of the role of the Australian healthcare system, which has been found to be unusual in its patterns of causal factors and outcomes in previous chapters, as well as a consideration of what the US and UK can learn in terms of health system change. It concludes with a discussion of some general lessons that the book offers and, finally, the strengths and limitations of the book as a whole.

Chapter 2 was concerned with the social determinants of health. It found that countries with high health outcomes have ∼GINI as a necessary condition, and there are two sufficient solution pathways to that outcome, ∼GINI*EDUC, or BEHAV*∼GINI*HEALTHEXP, but with the former having higher coverage of countries, which were France, the Netherlands, Norway and Sweden. For high health equity, ∼GINI was again a necessary condition, and the solution with by far the largest coverage and unique coverage was ∼GINI*HEALTHEXP, which covered Germany, the Netherlands, Norway, Sweden and Switzerland.

For both high health outcomes and high health equity, ∼GINI*HEALTHEXP was a necessary condition, with ∼GINI*EDUC*HEALTHEXP forming the sufficient condition for the Netherlands, Norway and Sweden, but with Switzerland deviant for coverage, so also having high health outcomes and high health equity, but not forming a part of the solution set.

As such, for high health outcomes and health equity, low income inequality and high health expenditure were necessary conditions, and the root of the dominant sufficient solution. These two causal factors were therefore carried forward to Chapter 6, which explores the relationship between the most important factors, as well as widening the sample of countries to 31.

Chapter 3 was concerned with health funding and found GOV*∼VOL to be a necessary condition for high access, with that combination forming the core of the two sufficient solution pathways (GOV*HEALTHEXP*∼VOL and GOV*∼OOP*∼VOL), and the former covering Germany, the Netherlands, Norway and Sweden, and the latter Germany and the Netherlands (in common with the first pathway), along with New Zealand and the UK.

Introduction

When comparisons of different health outcomes are carried out, there is still often an assumption that the reasons for those differences must be due to the performance of the health services in those countries. Politicians and policymakers debate league tables of health outcomes as if the results are entirely dependent on what goes on in healthcare services, and plans are put in place to attempt to address what have been identified on problem areas (Greener, 2016). However, it may often be the case that the health outcomes differences between different countries may be due to factors outside of the direct control of healthcare services.

Healthcare services are undoubtedly important, and the book will explore how they are funded, and what the money is spent on, in Chapters 3 and 4. But however important healthcare is, our health depends on a range of other factors that fall outside the remit or control of healthcare organisations and institutions (Schrecker and Bambra, 2015).

In respect of our own lives, we are fully aware that health services are not the only, or perhaps even the most important, factors in determining our health. Whether we can access health services (or not) when we are ill or injured is clearly important. This will be especially the case where people have a serious injury or life-threatening illness, but is also the case for the millions of people with long-term health problems that may require medications or medical devices, as in the case of diabetes or asthma.

At the same time, there are a range of factors which are likely to affect our health, but which generally fall outside of the control or remit of most health services. These factors are often referred to as the ‘social determinants’ of health, and go from those that might come most quickly to mind, such as the levels of smoking and drinking, and other activities we have come to associate with poor health outcomes, as well as education levels and the quality of housing available. However, there may also be much larger social factors, such as levels of inequality, which those taking a more large-scale, social determinants approach suggest are extremely important.

In QCA set-theoretical relations are the central features of the method, so it is worth explaining these first.

A necessary relation exists where, starting with the outcome we are interested in, we find a range of causal factors that consistently also appear. A simple example will make this clear. To get a graduate job (the outcome), it is necessary to first be a graduate (causal factor). Being a graduate is a necessary condition of getting a graduate job.

However, necessary factors, by themselves, do not guarantee a specific outcome will occur (unless there are also sufficient – which we will explore in a moment). This is because whereas it is necessary to be a graduate to get a graduate job, it is not enough by itself. There are a host of other things that a graduate will probably also have to do – apply for the job, go through a selection and interview process, and so on. So necessary conditions tell us about factors that have to be present (they are ‘necessary’) to lead to an outcome, but they usually are not enough (sufficient) – by themselves, to achieve that outcome as other factors have to be present as well.

A sufficient relation exists where, whenever we have the cause (or combination of causes), then we also get an outcome we are interested in. It is ‘sufficient’ to know that, when these causal factors are present, we will also get the outcome. However, just because a cause is sufficient, it does not mean that it is the only way of achieving the outcome. We may have several sufficient solutions (which we will call ‘pathways’ to an outcome). This is because our causal factors (and others) can be combined in different ways to achieve a specific solution. This approach of having multiple pathways to a solution, rather than a single solution, is called equifinality.

The logic of a sufficient condition is therefore, at least in some respects, the opposite of that of a necessary condition. For a necessary relation we start with the outcome (a graduate job) and then look for consistently present causal factors (being a graduate).

This book compares the health systems of 11 countries in terms of their social determinants, health funding and health expenditure, and explores how the different configurations of these factors, in turn, relate to a range of different outcome measures. It also compares a wider range of countries in relation to the factors found most important for the 11 countries, as well as exploring the first-wave response to COVID-19 in 2020. By exploring health systems in terms of several of their most important aspects, we can assess what they have in common and in difference, and whether those commonalities and differences are linked to better or worse outcomes.

No empirical work takes place in a theoretical vacuum. Things that seem important are more likely to be measured, and those measures often already come in clusters, based on the relationships that we assume exist between them. It is therefore important to actively think about what it is we are trying to measure, what theories are explicit (or implicit) in those measures, and then whether the empirical findings that we find support or challenge those theories.

It is also really important in comparative research to have a method for linking together the existing data and theory, and for testing it in a robust and transparent manner. Not everyone will agree with the findings in this book, but they will be able to see exactly where they agree or disagree. I hope this can lead to debate, and in turn to greater understanding.

The starting point for the book is to identify what perspectives it will take – the different ways health systems will be explored – and, in outline, the debates that it will cover as a result. After that, this chapter moves on to consider the different outcomes each chapter will address, before explaining the method the book uses to achieve its comparative analysis.

Health systems

The dominant book on comparative health policy is that of Blank, Burau and Kuhlmann (2018), now in its fifth edition, which gives us an initial template for the topics or dimensions that this book should reasonably be expected to cover.

Introduction

This book has been about comparing health systems, and this chapter is structured around two final comparisons. The first takes factors which were important in terms of causality from the social determinants, funding and expenditure chapters, and sees how they combine for 10 countries (the inclusion of long-term care expenditure means that New Zealand must be omitted). The combinations of these factors are then explored in terms of their necessary and sufficient conditions in relation to health outcomes. The health outcomes measure has been chosen as the best overall benchmark against which health systems should be judged – if they are generating good health outcomes for their populations, then they are probably doing a good job. Other outcome measures can make a strong case for their importance as well, but as the UK demonstrates, strong health equity can also lead to poor health outcomes, and as the US demonstrates, strong care process measures do not necessarily lead to measures in other outcome measures.

Utilising the health outcomes measure also allows a final comparison, in which the sample of countries is expanded to 31 to see how the causal patterns in that wider sample compare to those in the original 10. Including an outcome measure for the 31 countries means finding a replacement for the Commonwealth Fund health outcomes measure, but a key composite of that measure is the OECD ‘preventative life years lost’ (PYLL) measure, which provides data for a wider range of countries and so allows a wider comparison to be made.

The chapter first outlines and justifies again which causal factors it will include in its analysis. It then performs QCA using those factors and the Commonwealth Fund health outcomes measure for 10 countries to generate necessary and sufficient solutions. Next, the dataset is expanded to 31 countries, with the same causal factors included, but the OECD's preventable years of life lost (PYLL) measure used as the outcome instead, and QCA is carried out again. Finally, the two sets of results are compared to see which causal outcomes seem to most reliably link to strong health outcomes.

Introduction

Typologies of health system expenditures tend to be based on their degree of publicness (Blank et al, 2018, p 73), or countries are compared on the basis of their total spend on healthcare (Kotlikoff and Hagist, 2005). However, there is still relatively little work which explores different categories of health expenditure and how these contribute to good or bad care, and whether that care, in turn, leads to better or worse health outcomes.

In terms of arguments around levels of expenditure, there is often a general assumption that greater healthcare expenditure allows the purchase of more health services, and that this should lead to better health outcomes. However, this clashes with critical work, perhaps best exemplified by Illich (1977b), suggesting that increased spending on healthcare may itself be detrimental (Blank et al, 2018, p 260), with medicine being portrayed as a ‘disabling profession’ (Illich, 1977a) that prevents us from trying to find our own sources of well-being. As well as the disabling profession critique, Illich argued that the toxic or dangerous effects of medicine (its ‘iatrogenetic’ dimension) were not being taken into account, and raised questions that more recent authors (O’Mahony, 2016) have used as a basis for questioning the legitimacy of many medical interventions, which they find fall short of the standards of evidence which medicine aspires to (Stegenga, 2018).

There have been significant debates on the implications of trying to shift expenditure between primary and secondary care, which has been explored both in terms of individual health systems, but also comparatively (Peckham and Exworthy, 2003). There is a general trend towards health systems becoming more ‘primary-care led’ and of care moving more away from high-cost hospitals into community settings where it can be delivered more responsively (and perhaps more cost-effectively), but conceptual and measurement problems abound of what qualifies as ‘primary care’ (OECD, 2019), and so, although this debate is an important one, it is not the main focus of this chapter.

When considering what health systems spend their funding on, there are significant challenges involved in trying to balance the acute health needs of people today with expenditure on those with long-term conditions, with changing demography leading to increasing numbers of people with conditions such as asthma or diabetes, which medicine currently often cannot cure, and so which potentially require life time support.

Introduction

The way in which health systems are funded is often based on a series of political decisions which were made in the early development of different nations’ health systems, and yet, through processes of institutional reproduction, have remained remarkably intact today (Immergut, 1992b; Wilsford, 1995). As health systems absorb such substantial levels of resources, and because access to healthcare is not only recognised as a human right, but is also an international business of enormous scale, methods of healthcare funding in a particular country will be the result of a series of compromises between competing interests. Key stakeholders are those working in health services (with doctors usually having the most influence), government, public, private or not-for-profit providers of care, as well as other organisations such as insurance companies, patient representative groups and regulatory bodies. At election time the public will also have a say, but generally only from the ‘menu’ of options presented to them. Should events occur with particular salience to the general public (such as a rogue doctor or nurse, or a vulnerable person not receiving the right care), this can also mobilise change, especially should those events occur near an election.

This chapter explores the different configurations of the funding of health systems among the 11 countries included in the book, and the relationship between these configurations and access to healthcare, as well as a measure of the efficiency of the health system.

It therefore aims to discover whether there are patterns of healthcare funding that have necessary or sufficient relationships with healthcare access and measured efficiency, as well as whether there are any health systems which achieve both of these outcomes.

Healthcare funding

Healthcare is funded through a range of sources, but at the highest level of abstraction this involves a mix of money from the government, from health insurance of various kinds, or from the public through private, out-of-pocket payments. Government-funded health systems may utilise compulsory social health insurance as well as general taxation as their source of money. Health insurance can be public (where it can turn into a government scheme), or be organised on a private or not-for-profit basis.

Introduction

This chapter utilises the same method as the rest of the book (QCA) but with a different dataset. During the book's writing, the COVID-19 pandemic began and spread across the world. This gave me two options – I could ignore it, as the pandemic was not in the original book proposal, or I could incorporate it, and see how different health systems had responded to the challenge that it offered. I have decided on the latter, but of course any analysis I can offer is limited in that, at the time of writing, the pandemic is far from over. This has resulted in some methodological choices about what I can and cannot write about, but I hope that the chapter offers an insight into the ‘first wave’ of the pandemic and so makes a contribution to the comparative analysis of health systems.

Understanding why some countries were more successful than others in responding to the pandemic in its first wave – with the analysis here running up to mid July 2020 – gives important insights into the relative importance of the structural influences which are now known to be important in containing the virus, as well as giving an opportunity to assess the success (or otherwise) of different countries’ COVID-19 testing regimes.

Comparative studies have the potential to bring insight into how COVID-19 risk factors and testing regimes interrelate, but there are significant data limitations in terms of what can and cannot be measured in a robust way at the time of writing. This necessarily means some compromises have to be made. It is clear that policy responses such as the extent and timing of lockdown restrictions, hygiene measures, border controls, availability of protective equipment, and COVID-19 testing regimes, all have important roles to play. But achieving robust comparative data capturing these factors remains extremely difficult.

It is clear, however, that several important COVID-19 risk factors can be measured. Research over the last six months has shown clearly that older people are more susceptible to the virus and that there are increased risks through obesity (Goldacre and OpenSAFELY Collaborative, 2020).

Introduction

This chapter brings together the literatures on policy-learning and lesson-drawing on the one hand, and intra-crisis learning on the other, in order to examine the UK's response to the COVID-19 pandemic. The policy-learning literature explores issues such as what lessons were learned by whom. The lesson-drawing literature examines the content and process of policy transfer, focusing on the fungibility or transferability of lessons. However, most existing work is based on ‘ordinary’ policymaking rather than ‘extraordinary’ or ‘crisis’ policymaking characterised by elements of threat, urgency and uncertainty such as that during the pandemic.

We critically examine three different ‘real time’ lenses, drawing on three main sources: political (government documents, and Hansard debates); scientific (minutes of advisory groups such as the Scientific Advisory Group for Emergencies, SAGE) and media (national news media). These three sources provide different perspectives on the rapidly evolving government agenda. Political sources provide a record of what was being discussed by policymakers, who often claimed that they were ‘following the science’, as well as debates between the government and opposition, providing an insight into the scientific sources they were making use of in their deliberations. The scientific sources explore the extent to which the advisory bodies were looking at emergent research from their own countries, from the past or from abroad. The news media provide a rapid (daily) commentary on issues, giving a contemporary record of what was happening in other countries and material from the past that could inform learning about the virus. As Wolfe et al (2013) suggest, which issues are on the agenda, which ones are not, when and why, are the central questions that drive agenda-setting in communications and policy studies, and become even more important during a period of extraordinary policymaking. Our study cannot trace the links between policymaking and media reporting, but it can throw light on how the media aims to influence the public and policymaking. In particular, we focus on the lessons that the media suggest the pandemic offers to policymakers.

The chapter focuses in particular on the early stages of the pandemic, especially during January to April 2020 when key decisions were being made about policy in respect of testing, lockdown and border controls – the three elements highlighted by prominent public health advisors such as Edinburgh University's Chair in Global Public Health, Devi Sridhar (see Appendix 2: articles on 15 March and 4 May) as being crucial to pandemic response.

Introduction

Policy-makers seem to find the urge to reorganise healthcare almost irresistible. Doing so, however, as governments across the world have repeatedly found out, is expensive and time-consuming. In the UK, the National Health Service (NHS), after experiencing relative stability between its founding in 1948 and its first substantial reorganisation in 1974, has been subject to substantial changes with increasing frequency. In the 1980s the NHS Management Inquiry (DHSS, 1983) led to changes attempting to make the service better run. In the 1990s, an ‘internal market’ was introduced (HM Government, 1989) that attempted to create a dynamic where a split between purchasers and providers would generate improvements to services. After their election to power in 1997, Labour engaged in an almost hyperactive series of changes to the organisation of the NHS in England (with devolution taking other UK countries down a different path). Such a period of intense policy-making offers us significant opportunities for learning, both in terms of the NHS, and for health policy more generally.

While it is relatively straightforward to try and draw lessons from individual, specific policy changes, trying to disentangle the effects of one change from another, especially because they came with such frequency under Labour, is a more significant challenge. Finding a method of achieving this is a difficult but important task. To try and address these challenges, this chapter adopts an approach based on Pawson's realism, especially in the context of realist review (Pawson et al, 2005; Pawson, 2006, 2013), in trying to extract contextually sensitive programme theories from Labour's reorganisation to learn lessons from the changes between 1997 and 2010 for policy today.

In contrast to more conventional approaches to evaluation and review, Pawson suggests that we need to consider not only evidence about what appears to have worked in specific instances, but also the context within which those changes occurred, the theory that they appear to draw from and the outcomes that resulted as a consequence. The patterns between context, mechanism and outcome can be used to compare evidence of what happened in each case with both policy-makers’ expectations and existing theories to generate learning that we might be able to use to inform future policy.