Читать книгу Critical Realism for Health and Illness Research - Alderson Priscilla - Страница 9
ОглавлениеRethinking theories: the basis of practical research and problems with paradigms
During a cholera epidemic in the Soho area of London in 1854, Dr John Snow hoped to find the cause of the illness.1 He mapped the households where the people with cholera lived and tracked their daily life and movements in the area. These centred on a water pump used by poorer families. The pump was next to a workhouse and a brewery that had their own water supplies and where people seemed to be safe from the disease. Snow questioned the dominant view that cholera and malaria (‘bad air’) are airborne. To test his theory that cholera is waterborne, he had the infectious handle removed from the pump. Numbers of cholera cases quickly fell.
Snow is also renowned for another innovation. He administered chloroform to Queen Victoria while she was having her eighth child, and her doctors at last agreed to grant her wish to be relieved of the pain of childbirth. He helped to transform public attitudes towards anaesthesia as well as to hygiene and public health. Public patronage from aristocratic and wealthy clients was as vital then, to develop and spread new ideas, as research grants and academic journals are today.
John Snow set examples of health and illness research that critical realism (CR) supports. He used three methods. Induction: when he observed many cases and began to form theories about the cause of cholera. Deduction: when he formed his hypothesis that cholera is waterborne and set out to test it through counting households and mapping people’s daily movements and habits. Thirdly, he used retroduction:2 this involves:
•searching beyond evidence for the unseen cause only seen in its effects;
•working out the simplest explanation;
•imagining new possibilities;
•taking seriously a new idea (that most doctors dismissed);
•asking ‘what must the world be like for this to occur?’ (that disease perhaps spreads through water); and
•rethinking and potentially rejecting older beliefs (about the actual nature of clean water and cross-infection).
Retroduction, a major CR concept, is ‘democratic’ in that scientists and everyone else in everyday life use it.3 Snow put his research idea into practice. By having the pump handle removed, he changed the world through helping to reduce and prevent disease and eventually to change medical knowledge and practice. He promoted equality by assisting disadvantaged groups, who had to rely on public water pumps. And his service to the Queen was part of his work to publicise and promote a medical innovation, anaesthesia, so that it might become widely accepted and used.
Many researchers consider they are too practical to need theories or to find time to think about them, apart from as working hypotheses. Yet all research is based on theories. Eugenics was a predominant unquestioned theory in the first half of the twentieth century.4 Until around 1970, most researchers seemed unaware of how strongly sexism and colonialism shaped their assumptions, language, questions, methods, analyses, findings and conclusions.5 Their research design almost inevitably ‘found’ certain groups to be inferior. More valid accurate findings depend on researchers critically rethinking their assumed theories.
Theories include paradigms or the world views that underlie each scientific method. Paradigms involve beliefs about what counts as knowledge and evidence, or cause and effect, why the findings of large-scale research are trusted more than small-scale findings and what human agency is. Paradigms include positivism, interpretivism and critical realism. The most practical way to begin research is to question its intrinsic theories.
This chapter will review: social and biomedical influences through the centuries; positivist health research and the problems with it; interpretive research and some associated problems; seven differences between positivism and interpretivism; realism in realist evaluation (RE); how realism and critical realism are similar and how they differ and combining realism with critical realism and discourse analysis. The chapter ends by questioning the present unhealthy state of health research. This will prepare for future chapters that will show how CR helps to draw on the strengths and overcome the limitations of these paradigms by combining them into larger analyses.
Social and biomedical influences through the centuries
For centuries, medicine was informed by the social study of widely varied cultures of health and illness, by folklore and religious beliefs. Before they could offer reliable cures, doctors’ sympathetic yet confident ‘bedside manner’, their personal relationships with patients, besides their enquiries into the cause of illness, and their cautious support for their patients’ and colleagues’ preferences were all informed by their skilful observations of social values, habits and relations. These helped doctors to maintain their reputation, their income and social position and their patients’ trust.6
Medicine began to be transformed from a socially based profession into a more scientifically based one, moving from the study of the individual to the study of the disease, during the nineteenth century. Led by Claude Bernard and others towards objective study and laboratory-based experiments, the movement was furthered when many poorer patients were collected into hospitals and clinics. Doctors could more easily study groups of patients and examine their bodies before and after death.7 Yet scientific enquiry brought new attention to social contexts in public health and epidemiology as Snow showed.
The transition to modern scientific medicine was slow and painful. Older doctors held on to their ancient paradigm of the four humours inherited from Hippocrates and they opposed newer ideas. Public approval and confidence in medicine grew when effective treatments were gradually developed: vaccines, anaesthesia and antisepsis in the nineteenth century and insulin to treat diabetes in 1921. Penicillin dates from 1928, though it was not widely used until the 1940s and 1950s, and chemotherapy for cancer began in 1946.
However, the main routes towards lengthening lives and improving health were less through treating illness than in preventing it. This was achieved from the late nineteenth century onwards in industrial countries by piped water and sanitation systems, cheaper food and better food hygiene, safer childbirth, less absolute poverty and fewer slums.
Yet these routes contained the seeds of the present rising threats to world health. Longer healthier lives have increased the world population, over seven billion people in 2020 and predicted to increase to nearly ten billion by 2050.8 Food production has become a global industry run for profit, which involves massive waste and promotes unhealthy diets that increase obesity, heart disease and Type II diabetes.9 By 2030, 250 million children may be obese, one-fifth of them morbidly obese.10 Global corporations thrive in the capitalist system of inexorably increasing inequalities, so that one in every eight people is very malnourished.11 Each year, fewer individuals own greater wealth. In 2020, the world’s richest 1 per cent owned more than twice as much wealth as 6.9 billion people.12 Each year, many more people live in stressful debt. Distribution of wealth correlates with health. Morbidity and mortality rates increase along with poverty, and in less equal societies even richer groups are less healthy.13
Rising populations, high consumption and wasteful lifestyles increase greenhouse gas emissions and global heating. These challenge the health of all interdependent species, through the growing incidence and severity of floods, cyclones, heatwaves, fires, droughts, soil erosion, desertification, ocean warming, habitat loss and species extinction. Waste and refuse in rubbish tips, rivers and oceans, with pollution of air, land and water, add to the problems for people and the planet, or rather for the fragile ecosphere of gases, fluids and living species that surrounds the planet. Glaciers are disappearing, and no longer absorb carbon or reflect the heat of sunlight away from the earth or feed the great rivers on which the lives of billions of people and animals depend.14 Scarcity of fertile land and clean water leads to more violent conflict over who will control them. The key cause of the problems, increase of greenhouse gases and global heating, is unseen, except in its effects. These include increased air pollution. The highest level deemed to be safe by the World Health Organization air quality index is 25. In Delhi, India, the level can rise at times to 500.15 London air quality regularly reaches toxic levels, when admissions to hospital for heart and lung problems also rise.16 Absence of water is another hidden contributing causal mechanism. Warfare occurs in highly water-stressed countries, such as in Iraq, Syria and Yemen.17 Health and illness have to be understood within all these global, physical, social and economic contexts,18 and CR offers concepts, frameworks and interdisciplinary research approaches to connect them.
Modern medical science is based on a broad consensus of theories about how the body works, the causes and nature of health and illness, and the means of promoting health and of treating and preventing illness. The consensus persists across the numerous subspecialties ranging from anatomy to genetics, epidemiology to oncology, neonatology to geriatrics. The hundreds of thousands of mainstream health research reports published annually work within this internationally endorsed medical paradigm, while developing it in new directions.
In contrast, as mentioned in the Introduction, social research about health and illness is broadly split between two main partly contradictory paradigms or views of the world and of knowledge – positivism and interpretivism. They will now be considered in more detail.
Positivist health research
What are the most useful methods for researching health and illness? Everyone concerned with planning, providing or receiving healthcare gains from factual, evidence-based and rigorously tested research findings. These are essential for promoting high standards and for preventing inefficient, harmful and wasteful services. Valid, reliable research is served by the main healthcare research tradition of positivism.
Positivists apply deductive methods to test hypotheses and research questions through laboratory experiments, trials, surveys and statistical analysis. To positivists, assuring unbiased observation is an important marker of quality in order to increase accuracy and levels of predictability. Positivists also analyse published reports for literature, policy, systematic and synthesised reviews. They collect, analyse and report facts and recognise independent factual realities, though accepting that we can only partly and fallibly (potentially inaccurately) know them. Knowledge is constantly revised and updated. Few researchers describe themselves as positivists, but they accept underlying positivist theories when they measure, test, compare, evaluate and predict.19
The rise of evidence-based medicine and the ‘gold standard’ of randomised controlled trials (RCTs) seemed to promise certainty, particularly on effective interventions to treat disease. Over past decades, they have greatly helped to transform medical care, increase healthy outcomes and longevity, and reduce and prevent illness and suffering. Randomisation is an attempt to construct a partly closed system. Large groups of people are sorted into different treatment groups (research arms) randomly, by chance like the toss of a coin. Randomly allocated groups are surprisingly alike statistically; that is, they divide the large original sample group evenly on all the present known measures (such as age, weight, ethnicity, smokers) and presumably on all the present unknown measures too. All things being equal, the only difference between the large groups is the new treatment given to one arm, tested against older treatments or placebos (dummy treatment) given to the other arms, or control groups. The closed system of RCTs can assess and compare the effects of each treatment.
RCTs work best when testing biochemistry, the effects of a drug on a disease. Social life is too complex to have closed systems. It exists in open systems where many causal influences converge or conflict. And in social interventions, such as helping people to lose weight, it is impossible to standardise the treatment given. Unlike a drug, a weight-loss programme depends on how expert or inefficient each course leader is and how they relate to each participant.
A review of qualitative research reports found that adults on weight-reducing courses believed that personal supportive relationships with course leaders and other group members could have more effect than the course content or goal-setting.20 All behaviour-modifying programmes from psychoanalysis to cognitive behaviour therapy may be most likely to succeed if the therapist is empathic.21 This cannot be standardised or easily noticed or measured in quantitative RCTs. Randomisation is meant to remove or even out all secondary influences, such as hope or confidence or course leaders’ being more or less skilful, in order to test purely the effect of the intervention. In double blind RCTs, neither doctors nor patients know who is in which research arm, so that no one has the placebo feeling of certainty that the medicine will help them. Placebo can be as therapeutic as the treatment itself.22 In blinded RCTs, hope, which could ‘contaminate’ the results, is therefore replaced by uncertainty. However, blinding is less possible with social interventions, such as comparisons of counselling courses, when participants know which version they are having and may discuss it with those having other versions.
Positivist researchers examine repeated correlations between variables or, in the philosopher David Hume’s term, ‘constant conjunctions’.23 They check which variables seem to be most closely associated and recurring together. For instance, in long-term health surveys, how do each respondent’s diet and weight correlate with their risk of developing Type II diabetes? Researchers run regressions to measure the relations-conjunctions between ‘independent variables’ (the input or cause) and ‘dependent variables’ (the output or effect). Regressions are used to make predictions and forecasts by showing how likely effects are to follow causes.
The guiding principle is the supposedly inevitable ‘covering law’: if A occurs then B will follow. Another version of the covering law is counterfactuals: If A does not occur, neither will B. A typical covering law is that when a dry match is struck and surrounded by oxygen a flame will light up. The counterfactual is that if the match is not struck, the flame will not light.
Some positivists believe that when there is a range of variables, such as lifestyles about diet and exercise, if they connect them all:
one might hope to achieve a completely closed deductive theoretical system, in which there would be a minimal set of proposition taken as axioms, from which all other propositions could be deduced by purely mathematical or logical reasoning [with] the completely closed deductive system as an ideal.24
This was written at a time of ‘physics envy’, when natural and social scientists hoped to find the presumed ideal of a closed, deductive, law-like system and possibly a single cause.25 Such views influenced the reduction of sociology into tracking law-like empirical propositions, which were equated with theories about causes. Yet clearly, smoking does not lead inevitably to lung cancer, and lung cancer does develop in non-smokers. In the social world, the covering law never has total results. It is therefore analysed by how likely things are to occur, for their statistical probabilities, when large samples are needed to support general conclusions.
While many trials show strong correlations, others support questionable claims. An example is a follow-up study of children aged 11 years to assess the long-term effects on their health of their under-5-years experiences with English preschool Sure Start centres. The researchers concluded: ‘There was a 30 per cent fall in hospital admissions for injuries among that group,’26 and press releases announced a clear cause-effect connection. The researchers claimed, ‘We use statistical techniques to robustly estimate the causal impact that Sure Start has on children’s outcomes.’ Yet the measure was only potential access to the centres, not actual use of them. ‘This strategy lets us isolate the causal impact of having access to Sure Start during the first five years of life on children’s later-life health’ (my emphasis). ‘Having access to’ means living in an area with enough Sure Start centres that the child was likely to have attended one. No records of actual attendance were checked. The covering law difference was that six or more years earlier the ‘healthier’ group had ‘one more [Sure Start] centre per thousand children’ in the local authority where they lived. This, it was claimed, ‘prevents around 5,500 hospitalisations per year’.
The connection seems tenuous. Many other events in their lives might influence hospital admissions of 11-year-olds. Many of them might have moved between different areas during their preschool years. Yet positivists would reply this is not relevant when a significant statistical association can prove a causal connection between variables. It is not clear if the hope of preventing future accidents was a salient reason for providing Sure Start centres.
Problems with positivist health research
Positivist social science provides much essential knowledge about the nature and incidence of health and illness, but it is subject to some criticisms.27 Large trials and long-term surveys are very expensive. By the time findings are reported, contexts and policies may have changed and the findings may have lost relevance. Predictions based on childhoods from decades ago may not fit today’s or tomorrow’s childhoods. Contact with respondents is hard to maintain during longer term follow-ups. Assessments may therefore be brought forward but then be too short term to be reliable. People may feel better six weeks after cognitive behaviour sessions but not always months or years later, so premature assessments can be misleading.28 Assessments may be conducted by potentially biased providers of the service being tested, instead of by independent researchers. The high ideals of positivist research are hard to achieve.
Social science is overly dominated by positivist theories of causality, when associations between variables are taken to be cause-effect explanations. Yet these are correlations between different effects, not causes. Powerful social structures such as class or ethnicity are presented as ‘variables’ clustered around each person, alongside personal agency, such as each person’s reported views and preferences. This can confuse structures with agency and strong influences with weaker ones. Chapter 3 will show why it is vital to avoid doing this.
Positivists’ attention to empirical evidence (things that can be experienced through the five senses) prevents them from examining unseen causes. For example, real unseen causes include biochemical reactions between tobacco and lung tissue, people’s motives for smoking and persuasive pressures from tobacco companies. To concentrate on the variables, people’s actual smoking behaviours and health levels, can only describe and measure effects. It cannot explain why these occur. The Sure Start study researchers discussed earlier are typical in seeming to assume that it is sufficient, when informing policymakers, to show a connection.29 They do not explain exactly how Sure Start helps children to be healthier. Positivists avoid questions of justice or rights, which to them can seem biased.
There has long been concern that epidemiologists are too concerned with methods and not theories of disease causation and that they over-rely on biomedical individualism. Some researchers stress the biological and others the social production of disease, instead of integrating different causes.30 Tacit beliefs pass unquestioned in much positivist research, such as that health is best promoted by altering individual behaviours and not economic determinants of illness. This may seem neutral and non-political but is inevitably political and has been criticised as a victim-blaming, toxic narrative.31 For example, analysis of a longitudinal birth cohort study found that adults aged 45 were more likely to be obese and to have diabetes and high blood pressure if they had become parents before they were aged 23. The researchers noted that poverty-related problems ‘may have shaped’ the respondents’ lives. Yet they concluded:
Our findings support the theory that the stresses of early parenthood on both men and women accumulate over time, and may be contributing to poorer health in middle age. Policies and public services for the sexual and reproductive health of young adults are critical.32
Better health services and economic policies to support this clearly disadvantaged group were not recommended. The advice on ‘reproductive health’ echoes eugenic policies that manage racial futures by means of ‘encouraging or preventing the heredity of desirable and undesirable traits’ by promoting or discouraging births.33 They connect to the long Malthusian tradition that blames, disciplines and criminalises the poor and makes them responsible for their own plight, which has fed into neoliberalism.34 Like demographics, eugenics developed in the twentieth century as a means towards regulating race across the British Empire. Eugenic values influence broader policies, which follow Malthus’s view that prosperous groups deserve their wealth and rightly deprive the feckless poor of the means to enjoy life. This includes: reducing taxes, welfare states and free health services; hoarding wealth in tax havens;35 the austerity politics that led to 120,000 premature deaths during 2010–20,36 and many more deaths later, after services to cope with COVID-19 had been reduced; and the ‘herd immunity’ plan in 2020.37
Systematic reviews, coordinated through the online Cochrane Collaboration,38 have greatly influenced policy and public opinion. Their key value, ‘evidence-based’, is now an everyday term and policy standard. RCTs, which worked well in biochemical clinical research, have spread into the entire range of education, criminal justice, and social and economic services and interventions, coordinated by the Campbell Collaboration.39 However, as noted earlier, social RCTs are far harder to conduct and analyse than biomedical ones, given the complex open social systems and conditions they attempt to control.
Positivists are among the first to criticise faulty positivist research rigorously. Many systematic reviewers of published research start with checking hundreds of research reports and reject almost all of them for not meeting the reviewers’ criteria.40 Each systematic review assesses all the selected papers by around six criteria and concludes that very few reports meet the required standards. The reports may not explain the methods clearly enough. The research design and methods may be inadequate, the samples too small, the statistics flawed and results uncertain or negligible. Leading doctors estimate that up to 85 per cent of the millions of published clinical research papers are misleading, worthless or even ‘bad’.41 Medical research worldwide costs over $200 billion annually, but an estimated half of all the research remains unpublished.42 Most papers lack ‘context placement, information gain, pragmatism, patient centeredness, value for money, feasibility, and [or] transparency’.43 Most projects have not been repeated, and when they are, the results are seldom replicated or the predictions supported.44 Researchers also report different findings and effects from similar (though not duplicated) interventions. They thereby at least challenge one another’s findings if they do not refute them. Many psychology findings are ‘buried in bullshit’, with great pressures on psychologists to overstate statistical significance, publish unfinished work and avoid attempting to test and replicate colleagues’ findings or to publish critical reports.45 Research into genetics and neuroscience is found to have similar problems.46 Retraction of papers for fraud or plagiarism is increasing.47
Systematic reviews of RCT reports are designed to separate rigorous transparent research reports from inferior ones. They serve a main aim of the Cochrane and Campbell Collaborations: to provide reliable, verified, cumulative evidence of any useless or harmful interventions so that these may be withdrawn.
Not all research is trustworthy or relevant and even highly cited studies may be challenged or refuted over time. It would be preferable to focus on systematic reviews that bring together research in an explicit and accountable way rather than relying on individual primary research studies … research may be used selectively … to support ideas, positions or actions already taken or decided upon, rather than being a rational basis for making a new policy or practice decision.48
Synthesised reviews of many RCTs are seen as superior to direct RCT reports because they can deal with ‘the fact that policy interventions often shift in their purpose and meaning’.49 The reviews concentrate on ‘the ability of [an] intervention to support the process of behaviour change … rather than the specifics of the intervention itself ’.50 Systematic reviewers explain their aims and methods clearly and can benefit readers with their expert review of research methods. However, the advice implies that there can be unanimous agreement among all researchers and readers on the research design, content and findings, ‘the recommendations and formal guidance’ of the chosen papers and the overall review.51 As pandemic politics showed, this involves value judgements and political positions, to be reviewed in Chapter 4.
Systematic reviews have been criticised. ‘Existing models of systematic reviews fall far short of this aspiration’ of wholly reliable authoritative overviews.52 There is no simple evidence-to-practice link. Some of the review standards may be arbitrary and ignore valuable reports. The claim that reviews are emancipatory because related groups such as patients are consulted does not ensure that all the groups involved have equal influence or that they all agree on how services are inadequate or should be reformed see (Chapter 4). Systematic reviews of published papers are locked into the past when the reviewed research has been conducted years and even decades ago about past practices. Insistence on evidence base inhibits forward thinking about new alternatives. Synthesised reviews may distort reports of papers, highlighting minor points and missing major ones to fit them into the reviewers’ agenda. Reviewers may not request the consent of the original researchers or their participants in the primary research. Their voices tend to be lost, and they might not all agree with how their data are being used. Researchers whose papers are criticised or rejected by reviewers do not have a right of reply or appeal.
Some positivists believe their work should be objective in terms of being neutral and apolitical. Others think they should critically examine the influences of politics and economics on health.53 This debate will be considered in Chapter 4.
Interpretive and constructionist research and some associated problems
Current dominant concerns to promote ‘what works’ and ‘evidence-base’ favour large-scale quantitative research, although there is growing interest in qualitative research.54 Yet there is general caution about case studies and other small samples. They tend not to be seen as valid sources of general insights or of causal analysis and could offer misleading findings.
Just as the term ‘positivism’ has been used to cover a range of approaches, ‘interpretivism’ here refers to its broad alternative. Interpretivists examine how individuals interpret or socially construct their experiences and attend to perceptions and hermeneutics rather than facts. An image of hermeneutics could be two people interacting in a to-and-fro relationship through which they construct and reconstruct one another. A nurse might treat a patient as helpless and needy or as a victim to be rescued from suffering or as a brave and resilient person. When the patient responds with the suggested behaviour, the pair may keep mutually reinforcing it. Interpretivist research includes constructionism, constructivism, postmodernism, poststructuralism, ethnomethodology and other approaches.
Many interpretivists value Verstehen, Weber’s respectful understanding of the views held by the individuals and groups being researched. Researchers may see their own values and judgements as relevant, or they may examine these to see how they might illuminate their work or else try to ensure that their values do not influence the findings.55 Constructionists examine how we construct and reconstruct our experiences of health and illness.56 A subgroup, of constructionist realists, aims to show how ‘the contrasting perspectives and evaluations of medical knowledge are under-determined by … an external reality, while over-determined by social processes’.57
Positivist questionnaires designed to measure yes/no responses are less useful than narrative interviews can be when researching complex ambivalent personal views. These may reveal depths that require intricate analysis of how each speaker interprets the world. The facts of the causes, symptoms and treatments of illness, the stark biological realities of morbidity and mortality, matter less in interpretive research than how these are perceived and experienced. The ill-health assemblage is the networks of biological, psychological and socio-cultural relations that surround bodies during ill-health.58 Fox argued for sociologists to reject the organic body-with-organs as the unit of analysis of health and illness research and replace it with the approach to embodiment influenced by Deleuze and Guattari.59 He applied three concepts: the body-without-organs, assemblages and territorialisation. He contrasted these with the (factual) biomedicalised body-with-organs in order to explore the shaping of the ill-health assemblage. Whereas positivists consider there is an independent factual reality that can be discovered, interpretivists concentrate on how we perceive, construct, interpret and invent our experiences.
Methods that contribute valuable interpretive studies of health and illness include semi-structured interviews, ethnographic observations, histories and case studies of participants’ complex interactions within contexts and processes of change over time. All these methods can complement and enrich findings from quantitative surveys, when ‘quantitative forms of knowing may reduce the complexity of human experience in a way that denies its very meaning’.60 Detailed study of a few people with cystic fibrosis can be ‘nested’ within, and illuminate, large surveys of hundreds of affected people. The surveys can indicate how typical or unusual the detailed cases may be. The interviews help to resolve contradictions found by surveys, such as people’s ambivalent views about when they comply with their doctors’ advice or partly reject it.
An example of differences between the paradigms is when positivists include ‘trick questions’ in their questionnaires, designed to check if all the replies are consistent and ‘honest’.61 If there are contradictions between the replies, these tend to be seen as unreliable and deviating from accurate true accounts. Interpretivists, however, may regard inconsistent replies as a sign of a successful interview. Interviewees often start with the ‘public account’,62 saying what they believe will be accepted, wanted or endorsed by their expert interviewer. Gradually, as trust grows, and as interviewees explore their views more deeply and perhaps develop new insights, they may give ‘private accounts’ to later questions. ‘I used to think that, but now I’m beginning to realise that maybe I think this.’
Some interpretivists see the later replies as nearer to the interviewee’s genuine or true views,63 beyond a superficial and potentially misleading public level. If uncritical public accounts are sustained throughout interviews and questionnaires, or are counted as the genuine replies by the researchers, they can be used to support rather than challenge current practices and policies. Interpretive researchers help interviewees to move towards giving private accounts in personal narratives that emerge during informal trusting encounters. Accounts may also seem inconsistent, though be genuine, when interviewees talk of their interactions with different people who have varying views of them and their illness.
Narrative researchers have shown the great value of listening to interviewees developing their deeper insights and of analysing underlying themes in their replies. Mildred Blaxter analysed Scottish women’s views on their ‘health capital’ traced through family histories.64 She found that illness is easier to talk about than health, like Michael Bury.65 He showed the importance of narratives for interviewees who try to establish their moral status when coming to terms with their chronic illness. Arthur Frank also found how their narratives help chronically ill people to make sense of their experiences, to repair and create their new self and to make new life maps.66 These three researchers listened not only to what people said but also to how they said it and how they reconstructed their world. This is valuable information for healthcare professionals who want to understand, assist and influence their patients.
Some interpretivists consider that there is no truth, or that truth cannot be known or assessed, and that each inconsistent reply is valid in its own context. They give equal weight to early replies, which fit discussions between strangers, and to contradictory later replies, which fit more intimate relations developed through the interview. They may support relativist views, which see any general standard of truth as irrelevant and believe there are many contingent truths.67 Researchers’ concern is then with how interviewees present and perform their accounts.68 The main problems with interpretive research include its sometimes tenuous contact with reality and, in some versions, its denial that reality exists beyond personal perceptions. This limits its relevance for healthcare practitioners, patients and policymakers.
Another problem is barriers to combining interpretive and positivist research. Researchers often combine the paradigms in mixed quantitative and qualitative studies. However, useful multimethod work is held back when researchers in each group criticise the other group. Qualitative researchers contend their work is valuable because they can explore complex questions and go beyond measuring, comparing and evaluating. Quantitative researchers are wary of ‘anecdotes’ and ‘bias’ and poor selection of cases to support misleading evidence. They consider their large databases produce more accurate, valid and reliable evidence. Some researchers claim that only their own methods are reliable and worth publishing. In 2016, 76 senior academics protested that the (positivist-based) British Medical Journal kept refusing to publish qualitative research because it was seen as ‘low priority’, ‘unlikely to be highly cited’, ‘lacking practical value’ or ‘not of interest to our readers’.69
An anti-positivist interpretivist text claimed:
Phenomena can be understood only within the context in which they are studied: findings from one context cannot be generalized from one setting to another … Evaluation data derived from constructivist inquiry have neither special status nor legitimation; they represent simply another construction to be taken into account in the move towards consensus …
We have argued that no accommodation is possible between positivist and constructivist belief systems as they are now formulated. We do not see any possibility of accommodation if [it] is to occur by having one paradigm overwhelm the other by the sheer power of its arguments or by having the paradigms play complementary roles, or by showing that one is a special case of the other.70
Although many health researchers observe a positivism/interpretivism opposition, many others reject what they see as extreme views at either end of the spectrum. Numerous researchers aim to work on a middle ground of moderate versions. Yet to work in either paradigm involves tacitly accepting that paradigm’s theories. And researchers who combine, for example, positivist ‘objective’ surveys with interpretive ‘subjective’ interviews have to cover over contradictions, which are set out in the next section.
Seven differences between positivism and interpretivism
People with Type I diabetes have episodes of high blood sugar levels when they feel hyper and low blood sugar levels or hypos when they feel weak. Similarly, positivists can be hyper-factual and may treat everything as a strong reliable fact. If ideas cannot be clearly defined, counted and measured, they may be rejected as not worth researching; they might not even exist. In comparison, interpretivists’ knowledge may be seen as weak if it is hypo-factual, with the doubt that few if any facts and truths exist at all. Policymakers, journalists and the general public seem to hold this view when they rely only on hard positivist scientists to advise on the pandemic policies.
Positivist approaches in social and natural science research assume seven tenets, which are summarised here. They may be applied to experiments, surveys, tests, evaluations, demographics, statistics, multivariate analysis, medical records and medical imaging such as neuro-scans.71 They may be represented in the image of the objective, detached scientist examining through a microscope, computer or other technology specific data isolated from the social context.
1.Detached researchers observe objective, self-evident, value-free facts and material that are
2.set apart from their social context, often as separate variables
3.independent, pristine and the same, whoever observes, reports or reads about them
4.having, therefore, essential inherent qualities and
5.stable lasting reality ‘out there’ in the world so that data – words, numbers, images –exist unchanged across time and space.
6.Social and natural science facts provide general laws, replicable findings and reliable predictions.
7.‘Evidence-based’ findings yield self-evident conclusions about causes-effects, to support effective policymaking and problem-solving.
The seven main tenets on which interpretivism is based are summarised next. Here, the image could be two participants, or the researcher and participant, in hermeneutic human interaction with one another and with their complex social backgrounds. Readers will note contradictions between each pair of numbered tenets in the two lists.
1.Researchers see people, objects and events as constructed through negotiated interactions, hermeneutics and perceptions
2.within specific social contexts, cultures and meanings.
3.Phenomena are therefore contingent
4.with few or no essential inherent qualities
5.and no general, lasting, universal reality or truth that transfer intact across time and space.
6.Without fixed realities, it is hard to compare or transfer meaning, to generalise or connect causes to outcomes.
7.Connections between data, conclusions, recommendations and policy seem tenuous.
Later chapters will show how positivism and interpretivism do not need to be opposed. They can complement one another’s strengths and limitations when combined within larger concepts of reality. The next sections compare CR with realist evaluation (RE). RE is considered here in some detail because it is widely used in health research. Readers who do not work with RE may wish to move on to the later section on combining paradigms.
Realism or realist evaluation
One paradigm intended to combine the strengths and overcome the weaknesses in the above two main paradigms is realism or realist evaluation. RE is explained on a website and in many publications,72 and it is used in some systematic and synthesised reviews, considered earlier. RE is often assumed to be like CR, whereas RE and CR differ.73 This section lists aspects of RE and then notes similarities and differences between RE and CR.
Realist evaluators consider that much positivist and evidence-based research and policy work:
•is too simplistic for complex and varied social interventions, contexts, systems and implementations;
•gives little idea as to why something worked or not in different contexts;
•needs to move beyond measuring and reporting effectiveness when evidence is mixed or conflicting;
•needs to involve the views of all stakeholders; and
•needs a research synthesis method to evaluate complex interventions.74
Eight features of RE will be summarised and then these will later be compared with CR.
1.1. RE accepts the ‘realistic’ factual positivist view of a real external natural and social world that we can discover and know through experiment and observation. Social structures and systems are seen as ‘real’ because they have real effects. RE also supports constructionist understanding of experience: everything we sense and experience is interpreted through our brain, so that we cannot be certain of the nature of reality. We amass and interpret and know and experience only partly and fallibly, filtered through our language, culture and memories. RE is ‘sitting between’ positivism and constructivism.75 Realists may involve a wide range of stakeholders in planning and reviewing their work to ensure their views are considered. Yet RE aims to avoid the relativist tendencies of interpretivism.76
1.2. The aim in RE is not simply to show what works (or does not work) but to show what works for whom, when, in which contexts and ways, and why it works. RE is ‘a methodological orientation, or a broad logic of inquiry that is grounded in the philosophy of science and social science’.77 RE reviews are confined into past evidence and the decisions and methods of the earlier primary and secondary researchers whose work is synthesised. This often non-realist work is evaluated and may be ‘recast’ and reworked into new CMO configurations. CMO stands for how Contexts and Mechanisms interact to affect or generate Outcomes. To show how and why things work in certain ways, there are complex RE methods to examine diversity and complexity in subgroups of interventions and participants.
Realism sees the human agent as suspended in a wider social reality, encountering experiences, opportunities and resources and interpreting and responding to the social world within particular personal, social, historical and cultural frames. For this reason, different people in different social, cultural and organisational settings respond differently to the same experiences, opportunities and resources. Hence … a complex intervention aimed at improving health outcomes is likely to have different levels of success with different participants in different contexts— and even in the same context at different times.78
RE researchers accept that the social world is in constant flux and change. People are reflexive, make choices and so are unpredictable. RE evaluates healthcare interventions or ‘programmes’. When local programmes and contexts are researched, they are ‘vulnerable to the intrusion of or invasion by more immediate external contextual conditions’ that can overwhelm the programme. Conditions include ‘political, population, transportation, administrative, economic or even climatic sources.’ There is also ‘an almost infinite range’ of cultural influences in open systems, which can never ‘be fully articulated’. All these limit RE’s predictive power. Programmes ‘by their nature are inherently fragile’.79 RE examines complexity by showing ‘associations and correlations in data from many types of evaluation’ to explore and explain why they occur using qualitative and quantitative data.80
In common with other theory-driven review methods, the realist approach offers the potential for insights that go beyond the narrowly experimental paradigm of the randomized controlled trial. It can do so in relation to complex, complicated or simpler interventions (for example, even a simple intervention, such as a drug, is prescribed, dispensed and taken – or not – in a particular social, cultural and economic context).81
So RE aims to evaluate a range of possible problems and varied personal responses through its more detailed and open methods than is usual in positivist research.
1.3. Ray Pawson, the leading realist, considers ‘the real starting point of science … lies in “theory”, our ideas on the nature of the problem and on the nature of its solution’.82 However, by ‘theory’, realists mainly mean their hypothesis: to show within the CMO framework what the tested programme is expected to do and, in some cases, how it might work. The RE method:
•makes theory explicit on the underlying assumptions about how something works and its expected effects (deduction);
•looks for empirical evidence to support, contradict or modify the theory;
•combines theory and evidence into results to explain CMO interventions;
•informs decision-makers about effectiveness and estimated predictions of risk and expectations if something will work or not; and
•does so by providing rich, detailed, practical understanding of the working of complex interventions.83
Yet realists accept that their programmes and theory-based understandings about ‘what works for whom, in what contexts, and how’ and their reported outcomes work differently in different contexts and through different change mechanisms. They cannot simply be replicated in each context.
1.4. RE aims to solve social problems by examining how interventions can benefit and alter the behaviours of different subgroups of research subjects.84 Rather than conducting RCTs, RE’s secondary analysis of RCTs examines the effects on subgroups within large trials and within synthesised reviews of many RCTs. In CMO theory:
Contexts include spaces and social, geographical and historical settings, with norms, values and interrelationships found in underlying social, cultural, economic or legal contexts. Contexts may also include the RE research programme design and staff. Contexts are the circumstances in which mechanisms can be fired and will operate or be prevented. Each research programme can have multiple mechanisms and contexts.
Mechanisms involve changing the reasoning of research participants, their values, beliefs and attitudes or the logic they apply to a situation, their ‘choices and capacities which lead to regular patterns of behaviour’.85 Mechanisms are the ‘entities, processes, or structures which operate in particular contexts to generate outcomes of interest and give rise to causal regularities … a guiding principle across many social and natural science disciplines’.86 RE programmes work when participants make and sustain new choices within their available resources (information, skills, material resources, support). Mechanisms include programmes and their resources, for example, a weight-reducing course, and how these change the participants’ reasoning through the ‘choices and the capacities they may derive from group membership’ or other resources.
The measured outcomes are the changes triggered through the interactions of contexts and mechanisms on participants’ reasoning and behaviours. Outcomes indicate whether to ‘mount, monitor, modify or mothball a programme’. Programmes produce multiple outcomes, which are examined by testing hypotheses on subgroups. ‘Outcomes are not inspected to see if [whole] programmes work, but are analysed to discover if the conjectured mechanism-context theories are confirmed.’87
We find the same combination of agency and structure employed generally across sociological explanation and we thus suppose that the evaluation of social programmes will deploy identical explanatory forms, reaching ‘down’ to the layers of individual reasoning (what is the desirability of the ideas promoted by a program?) and ‘up’ to the collective resources on offer (does the program provide the means for subjects to change their minds?).88
RE social programmes and complex interventions may change the micro resources, or opportunities available to participants, and the macro social context, such as legislation. To alter participants’ reasoning can mean helping them to want to lose weight and to sustain related changes to their daily life. When realists alter the context of opportunities to trigger the mechanisms of change, there will be both intended and unintended outcomes.
1.5. Realists use positivist covering law. Semi-closed systems, contexts and programmes are constructed to test if they work more or less well through different mechanisms or contexts or with different groups of people. By tracing more diverse and detailed conjunctions than RCTs and evaluations usually do, RE aims to show ‘the success, failure or mixed fortunes of complex interventions’.89
The aim is also to show why changes occur. RE ‘always has explanatory ambitions. It assumes that programme effectiveness will always be partial and conditional and seeks to improve understanding of the key contributions and caveats.’90 Realists assess patients’ or professionals’ behaviours before, during and after trial interventions and compare CMO configurations within programmes. Stakeholders’ views on how well programmes work are collected through interviews, focus groups, questionnaires and the DELPHI method.91 Their views help to refute or refine theories about how and for whom each programme ‘works’.92 Given the multiple CMOs and interactions between them, as noted earlier the findings are not replicable but they can be transferable. They might work ‘more or less well’ with certain groups and contexts, at certain times, and even differently within the same context.93
1.6. Auguste Comte, the founder of positivist sociology, promoted research to discover and explain covering laws in order to predict, and to predict in order to control society.94 Durkheim and Parsons developed this functionalist tradition, the view that whole societies function for the greater good. Like other ‘what works’ research programmes, RE aims to show ways to organise and improve the effective functioning of healthcare and social systems.
1.7. Realists aim to be objective and value-neutral, to concentrate on facts and exclude values.95 They mainly work on altering individuals’ observable behaviours, rather than addressing unseen structural influences on ill-health such as poverty or air pollution. This risks blaming unhealthy victims instead of pathologising policies. RE began with criminology and programmes to reform offenders.96 While overall outcomes of an evaluated programme did not show big gains, evaluations of certain interventions with certain subgroups of prisoners within the programme showed better outcomes. This led to RE’s more complex versions of analysing research trials. The model of reforming prisoners to reduce crime transferred to the model of altering patients’ behaviours to reduce illness. This fits the structural functionalist tradition,97 where illness is a form of deviance and patients should adopt the sick role, comply with treatment and strive to recover.
1.8. The huge amounts of extra data generated by RE analysis of subgroups can be confusing unless they are very well organised and explained. Realists admit that RE ‘can be difficult to codify and requires considerable researcher reflection and creativity. As such there is often confusion when operationalising the method in practice.’98 Research may ‘become bogged down in finely detailed lists [of Cs, Ms and Os and fail] to produce a coherent explanation of how these … were linked or related (or not) to each other’.99 RE researchers regret that their reports need to be longer than the 3,000–4,000 word total set by most academic journals.
How realism and critical realism are similar and how they differ
This section comments on the eight points and raises ideas to be developed in later chapters.
2.1. RE and CR share the theories noted in 1.1.
2.2. RE and CR share caution about researchers’ limited ability to make general predictions, given complex societies, choice-making individuals and open systems. RE examines empirical evidence of past events. CR is interested also in the present, the future and unseen influences.
Although realists claim to show why things work, they examine effects in empirical correlations but not unseen real causal mechanisms.100 RE merges causes and effects in timeless multivariate analysis. ‘Realists shun the successionist [sequenced through time] view of causation as a relationship between discrete events.’101 They see ‘causal powers’ embedded timelessly ‘in social relations and organisational structures which they form’.
Yet timing is vital in CR to understand causes that precede effects and also to track processes of transformative change. For example, unseen chemicals in traffic emissions precede the ‘new tobacco’ of toxic air,102 which increases children’s breathing problems and weight gain.103 Should RE studies of how walking to school might affect childhood obesity extend to also examining air quality and its effects?104 CR considers these kinds of political and potentially global influences, which are beyond RE’s controlled CMOs.
2.3. By ‘theory’, realists mainly mean their CMO hypothesis, and they are primarily concerned with methodology. Pawson criticised CR’s ‘totalizing ontology, its arrogant epistemology and its naive methodology’,105 but CR does not have a methodology. CR is primarily concerned with theories, philosophical questions, concepts and frameworks of analysis, though it can be combined with methods including RCTs, systematic reviews and evaluations.106
2.4. RE and CR both examine how agents interact with social contexts and structures in qualitative and quantitative research, but they do so in different ways. RE mechanisms merge the resources of structures with the reasoning of agents, and conflate agency and structure into people’s thoughts.107 CMO is criticised for being unclear about when context and mechanism refer to structure or to agency. Which has an empirical basis and which has a theoretical basis? How do socio-economic structures fit into CMO?108 How can advice that researchers should theorise links between variables work ‘if one variable is of a different epistemological nature to the other two’?109 Another confusion is advice that RE researchers should engage with context in their explanations but also strip out context when they identify a generalisable theory that can be tested in various contexts.110 Unlike RE, CR clearly distinguishes between inanimate structures and conscious agents. These exist separately but interactively.
RE and CR have different views of reality. Whereas RE works at the empirical and actual levels, Chapter 2 will explain CR’s three layers of reality. These are empirical experiences, actual existing things and the third level of unseen real causal mechanisms. In RE, the researchers’ programme, such as analysing RCTs about weight-reduction courses, is seen as a real cause. Yet in CR, a research programme is not seen as a real cause ‘but an artificial structure of inquiry that seeks to set up a relatively closed system. In itself it does not cause anything, it is a research method.’111 In CR, it is vital not to confuse the methods with the objects of research, such as the causal law method of analysis with the observed patterns of events. To critical realists, ‘the idea that research experiments can produce laws, not simply study them’, is as absurd as imagining that ‘human beings, in their experimental activity, cause and even change the laws of nature!’112
RE includes the programme, which is being managed and tested, among its contexts and mechanisms. This also raises confusion between the research processes-methods and the phenomena being researched. Another complication is when effects of the programme, such as responses to obesity and attempts to reduce it, are treated as if they are causes. RE’s attention to actual measurable events and behaviours further diverts attention away from unseen real causes. Critical realists consider that only if the real causal level is understood and addressed can we understand the causes of obesity, or of any problem, and therefore the means of preventing it (Chapter 2).
2.5. Whereas RE constructs semi-closed systems in order to test covering law, CR researches the real world of complex open systems and unpredictable choice-making agents. Pawson claims that ‘laws of nature are only produced in artificial closed systems’.113 CR denies there can be closed systems in the complex social and natural world. CR also denies (point 2.4) that closed systems and research programmes can ‘produce’ laws of nature, which can only be revealed and discovered, not invented.
CR celebrates open systems, whereas RE has a partly negative view of them when they ‘subvert or enhance’ RE programmes, ‘compromise’ empirical closure and ‘threaten even well-established CMO configurations’ so that they may ‘suddenly fail’.114 ‘Some have argued that syntheses based on critical realist meta-theory will ensure fuller explanations of social change than can be offered by current [RE] approaches to meta-analysis as these are based on the thin inferences of causality from standard experimental studies.’115
Despite the aim of objectivity, RE appears to allow for considerable subjectivity. RE is an ‘intellectual craft’ that leaves great discretion to the researchers. ‘The strength of evaluation depends on the perspicacity of its view of explanation.’116 If programmes become too complex, the breadth of areas and ‘depth’ (amount of detail) may need to be cut, in consultation with stakeholders. Influential stakeholders help to increase ‘maximal end-user relevance’.117 Increasing the relevance and usefulness of research is very worthwhile. Yet there are risks to impartial research of involving powerful advisers and working closely with funders, allowing them to alter the research processes, risks not addressed by all realist evaluators.118
2.6. RE’s functionalist cost-effectiveness, ‘what works?’ agenda, is strongly supported by funding and policy authorities.119 The findings help authorities to organise society and promote public health within current power structures. CR, however, follows critical research traditions. These aim to understand and change the world in order to promote justice and equality but are not generally welcomed or funded by authorities. Pawson regards the fundamental division between RE and CR to be ‘on the matter of whether social science should primarily be a critical exercise or an empirical science’.120 However, CR research can be critical, empirical and scientific.
2.7. RE researchers claim to be value-neutral.121 Pawson believes value-freedom is essential if evaluation science is not to ‘abandon analysis for ideology’, which he sees as the fundamental error of CR.122 Being value-neutral can involve attempts to avoid political and moral matters, but CR recognises that all social life is imbued with these matters and values (Chapter 4).123 As mentioned earlier, values have greater influence over researchers if they remain tacit and unacknowledged. One example is when RE ignores powerful socio-economic influences on health and concentrates on health promotion that attempts to get individuals to address their ‘social problems’ by changing their beliefs and behaviours.124 Critical realists are very interested in observing and analysing how individuals’ beliefs change (Chapter 3), but they do not try to direct individuals as RE does.125 Instead, CR concepts of health promotion include changing structures to advance justice and human flourishing generally.126 Although RE researchers would reject this as too political, their close ties to government raise political questions. They seem to endorse, or at least not challenge, government austerity and healthcare policies, despite the many ways in which these damage public health and wellbeing.
Caution is needed in RE health research to avoid echoing RE’s origins in criminology and deviance.127 These risk pathologising and blaming health service users as if they are deviant, and there is a risk of endorsing the merging of medical and police roles.128 Doctors increasingly have to certify access to paid sick leave and other benefits, disability payments and food banks and to check that non-citizens are denied free care. RE’s term ‘stakeholders’ misleadingly implies that everyone involved has equal power, an equal ‘stake’, and shared common interests. CR, however, critically researches inequalities and conflicts of interests.
2.8. RE and CR can both be complex and often need to be explained in longer papers than are accepted by most academic journals. On differences between RE and CR, Hinds and Dickson conclude:
the configuration C+M=O confuses researchers since it does not separate out the empirical, actual and real [the three CR levels of reality] clearly enough so that they can engage in the iterative and retroductive theorising that should take place between the different levels. We argue, therefore, that the shared methods of orthodox reviews and realist reviews confuse researchers into believing that a realist review is just another way of adding things up. We think that critical realism offers greater opportunities for genuinely trans-disciplinary explorations of social change that fully exploit its philosophy.129
Combining paradigms
Since CR is about theories, it works with a range of methods across positivism and interpretivism, to clarify them, fill in gaps and extend analysis. Instead of splitting different and opposing ideas into dichotomies, CR tends to draw them together into interacting dialectic. Dialectic is considered in Chapter 6, but meanwhile Example 1.1 combines RE, CR and dialectical theory of discourse to show the importance of recognising policies and values in community care research.
Example 1.1: Combining paradigms: an evaluation of a community-based integrated care service
Hannah Kendrick, University of Essex, UK
I found applying critical realist principles useful within my case study for connecting micro-level practices of managers and frontline staff with wider political and economic discourses. By integrating principles from realist evaluation with dialectical theory of discourse,130 I could show how assumptions and problems within the policy discourse of integrated care worked through discursive and non-discursive mechanisms. Beyond the levels of the actual and the empirical, I gained an understanding of how political factors at the level of the real were working dialectically with the contexts, mechanisms and outcomes normally explored through RE.
My PhD sponsor required me to evaluate the community-based integrated care service.131 Initially, rather than simply try to evaluate impact with an RCT, I planned to do a realist evaluation,132 due to its focus on how, for whom and under what circumstances interventions create change. Realist evaluation was also cited as being useful to policymakers and being equipped to deal with the complexity of healthcare interventions.133 However, during the first phase of research, I observed that the initial programme theories were normative assumptions about patient responsibility and the reduction of service use, without much reference to austerity and the political factors that were influencing this drive. Local policy implementers were also constructing certain actors within the service as ‘problem contexts’ to be overcome. They included ‘resistant’ patients who were too dependent on services, ‘self-serving’ staff who were resistant to generic working and nurses who were described as not being very ‘good’ at case management and self-management. I was left wondering what sort of work do the initial programme theories do if policymakers and implementers obscure political motivations within health service change or construct certain groups as ‘problem contexts’ to serve political aims? In what ways does this drive certain political agendas, with negative effects for those affected by the policy intervention, such as patients and health staff?
Chouliaraki and Fairclough’s dialectical theory of discourse understands social life to be produced through a range mechanisms working in dialectical relationship with one another.134 These mechanisms include discourse, material activity, social relations, power and institutions, mental phenomena (beliefs, values and psychological processes). This theory follows the critical realist assumption that each of the elements contributes its own distinctive generative powers to the production of social life. In that sense each mechanism ‘internalises’ the others without being reducible to any of them. Discourse is therefore a form of power, a mode of formation of beliefs/values/desires, an institution, a mode of social relating and a material practice. Social practice always has a reflexive/positioned nature to it.
RE would be interested in the mechanisms arising from material activity, social relations (interactions, interrelating and behaviours) and mental phenomena (beliefs, values, desires), but less interested in power and discourse. The benefits of looking at non-discursive processes working in dialectical relationship with discourse are that they highlight how those mechanisms analysed within RE have a reflexive and normative basis that cannot be analysed in a value-neutral or objective manner. Furthermore, the dialectical approach can show through empirical research the consequence of some of the more hidden political agendas driving service change for those involved in implementation, in their agential responses and lived experience.
My study shows that integrated care policy discourse works in ‘empty oppositional status’ to wide-ranging and diverse issues within the health service. This allows it to be presented as a policy solution to poor public finances, fragmented and disjointed care, lack of patient decision-making and high demand.135 The discourse ‘irons out’ the contradictions, dilemmas and antagonisms of integrated care as a practice in ways that accord with dominant interests.136 This can reinforce economic austerity with an appearance of government action on poor patient care, fragmentation between services and lack of respect for patient decision-making.
Three empirical examples in my case study demonstrate how contradictions between the drives either to create efficiency savings or to create less fragmented, and more respectful, patient-centred holistic care played out. Local policy implementers’ discursive assumptions allowed them to drive through the efficiency savings and worked dialectically with non-discursive mechanisms to mobilise economic austerity and undermine policy rhetoric about the benefits of integrated care.
Firstly, clinical staff decision-making was replaced with an auto-allocation scheduling system for community health visits. With interactional and behaviour mechanisms this increased stress and reduced the ability of clinical staff to treat patients holistically and to coordinate care and integrate with GP practices, but it increased the workload of individual staff. Secondly, nurses were made to be overly paternalistic and patients overly dependent on public services. This happened through managerial resources placing pressure on staff to discharge patients, resulting in conflicts between clinicians and patients and disempowerment for patients. Lastly, the move to a more generic or ‘integrated’ workforce model was assumed to be an easy common-sense process, which improved the holistic and coordinated nature of patient care. Staff resistant to this workforce model were problematised in terms of their age, their confidence and individual preferences, through top-down managerial practices and inadequate training. Some professionals resigned due to the changes, while those who remained felt very stressed and alienated from their managers. Patients also told me they thought the care remained very disjointed. Therapy assistants with nursing skills filled the resource gap within nursing, but then both nursing and therapy tasks proved largely impractical to perform within one visit. Some said they had not been officially signed off to give nursing care and they did not feel adequately trained.
In summary, the dialectical theory of discourse, which connects powerful, political, hegemonic discourses with realist evaluations, focused on contexts, mechanisms and outcomes at the level of the actual and empirical, allowed for greater explanatory power in determining real change within an integrated care service and connected micro-level practices with integral political factors.
The unhealthy state of social health research?
Malcolm Williams (not a critical realist) criticises social researchers for pursuing disconnected directions, like an orchestra of soloists. All the researchers write, he believes, as if their method is self-sufficient and stand-alone and compensates for the failings of all the others. Williams asks what each method is for ‘if not for itself?’ He regrets that he cannot see a cohesive ‘intellectual division of labour within the social research enterprise’.137
There is serious discord, contradiction and disconnection between social research paradigms,138 instead of a working together in critical coordination. Can and should the contradictions between paradigms be resolved? They split social research in major ways. Rather than mutual learning and exchange, there is too often mutual criticism between health and illness researchers. Yet research is not well served by this division and rivalry. If social researchers criticise one another, how can they expect everyone else to respect and apply their findings? Sociologists were notably missing from the official and independent Scientific Advisory Group for Emergencies, though fortunately there is Ann Phoenix, professor of psycho-social studies.
Social research is useful when it measures and describes health and illness, the effects of treatment and prevention, and the experiences of all concerned. But that is not sufficient. It is like a doctor dealing only with symptoms (effects), offering paracetamol for pain but not checking if there is a tumour. It can be like researchers watching water pouring through a kitchen ceiling. They map, measure and describe the flood and how people cope with it. Yet they say that they cannot trace the unseen source because that does not count as empirical evidence. Research theories are also needed to find the unseen leaking pipe and see how it can be mended.
The health of nations is enhanced by all medical, nursing and related specialties being united in their underlying paradigm theories of health, illness and effective treatment, as mentioned earlier (though of course they disagree on the details as knowledge develops). The wellbeing of nations could be served far more effectively if social researchers could likewise agree on basic common theories and meanings of truth, validity, knowledge and reality.
Bio-social-economic problems such as health are ‘wicked’ when they are complex and involve many different perspectives.139 The critical realist Leigh Price developed this concept to argue that wicked problems should be addressed with strong science and democratic retroduction (defined earlier). This involves everyone, experts and lay people, working out together what the world must be like if we are to promote general wellbeing.140
The realist evaluator Pawson criticised CR:
It is a strategy for lording over complexity rather than analysing it. It is a strategy with no use whatsoever in applied social inquiry. The science of evaluation [RE] starts by recognising that the fate of social policy lies in the real choices of choice makers and its task is to explain the distribution and consequences of those choices rather than to condemn them.141
Critical realists do not condemn the choices, but they aim to do much more than explain ‘the distribution and consequences’. They aim to:
•resolve related contradictions;
•combine and develop other paradigms (positivism, interpretivism and realism) into larger, more convincing and valid approaches;
•explain as well as describe and measure;
•show how small qualitative studies can give reliable explanations;
•provide frameworks to organise multimethod and interdisciplinary research across the social and life sciences; and
•show healthcare providers and users and policymakers how the research findings are useful and valid.
To convince others involves reaching agreement with them on the truth of evidence. Thomas Kuhn doubted this is possible when he traced the history of science as a series of incompatible paradigms.142 He argued that ‘the truth of a proposition depends on the framework [or paradigm] in which it is asserted’. He believed truth can only be understood and accepted within that paradigm. Yet this would mean everyone who is working in another paradigm could not believe that truth – including statements such as Kuhn’s that deny any general truth. When applied to itself Kuhn’s statement is self-defeating, like all relativist generalisations.
The following chapters show how to resolve this problem and establish shared common ground to support valid, convincing and useful health and illness research. This should challenge false research evidence, which supports useless treatment, delays better alternatives being introduced and therefore undermines people’s health, wellbeing and survival across the world.
I suggest that each social science paradigm is like a jigsaw puzzle. When we try to combine them, it is as if there are hundreds of jigsaw pieces scattered around but no guide to how they can fit together. The next chapters offer a guide, like the picture on the lid of the jigsaw box, although that picture will never be complete.
Questions to write or talk about with a colleague
1.Does your research draw on positivism, realism, interpretivism or another paradigm?
2.Does it combine one or more paradigms? If so, were there difficulties?
3.Can you list the main questions from your research so far that you hope this book will help you with?
After Chapter 7, you will be asked about the main ways in which you think this book has been helpful – or not helpful. So you might conduct an audit of each chapter.
