The political and professional context
Now here was an unlikely, though none the less welcome, ally in the cause to which this book is ultimately dedicated:
The commitment to evidence-based medicine increasingly pervades modern medical practice. This kind of commitment should be extended to the social services world. (Stephen Dorrell, Secretary of State for Health and Social Services, Guardian, 20 February 1997)
Those of us who had been shouting this into what seemed like deaf ears for some considerable time were well advised to pause here, not to write the pained letter, but to reach for our long spoons instead. Probably the worst time to try to write a book on research-based approaches in social work was five or six years ago, when the previous administration (lovely phrase) was at its most bullish; the social work profession, on the other hand, was in a state close to learned helplessness as a result of the aggressive application of commercial principles to community care and the peak of the great epidemic of Mad Reorganisation Disease (MRD) which accompanied this. In 1996, a reviewer, describing my book Cognitive-Behavioural Therapy (Sheldon 1995) as ‘sectarian’, advanced the view that then was certainly not the time to discuss which methods of helping might prove more effective (perish the thought) since all but the most basic problem-containment procedures were off the agenda, and probably for good. It could, of course, be argued that had we taken opportunities to debate and act upon the positive and the negative findings of effectiveness research from the last forty-odd years, then we might not have been in this particular fix in the first place.
The situation, now and for the immediate future, strikes me as somewhat less threatening, for the following reasons.
Fashions come but fashions go, and the 'sieve of history' has large holes. In those local authorities which sought head prefect status from the previous government by eagerly implementing 'internal market' procedures ahead of the 1991 deadline (I remember predicting that, as in the National Health Service, it would all end in tiers), we now see clear signs of reappraisal, particularly in the child care field – which, remember, can easily get one's name in the papers. Co-operation and prevention via 'lighter touch' approaches rather than competition and 'targeting' are now the orders of the day. Indeed, what else could rationally be done in the face of official figures showing an 80 per cent 'false positive' rate from 20,000 child abuse investigations, with only 5 per cent of families receiving any subsequent help? Imagine a medical screening programme producing such results and what the public reaction would be. Perhaps I am being over-optimistic here, but the smell of burnt fingers is quite distinctive.
Requests from senior managers for academics in the know to talk about the case for evidence-based social services are now commonplace. Of all the things that should not be a craze, this idea is fast becoming one. For the sake of our clients and for the profession we should probably go with this flow, providing we keep a sense of balance and pragmatism and remind ourselves that Mary Richmond (1917) was addressing these matters on the eve of the Russian Revolution, and that the first controlled experiments into social work effectiveness were conducted in the 1940s (Lehrman 1949).
It is only an impression, but hard-pressed social workers now seem more willing to take students again. Once an eccentric-seeming decision ('What's in it for us?' was a common response to such requests; answer: 'the survival of the profession to which you belong because someone did it for you'), practice teachers are now seeking specialist qualifications, and appear willing to ally themselves with those courses which offer good support and involvement in all parts of the curriculum. Debates about what the content of practice should be increasingly appear on the agenda of practice teachers' meetings and curriculum development committees.
A consortium made up of Exeter University, social services departments in the South and the West, and the Department of Health have between them invested £2.5 million in the cause of developing evidence-based social services. My erstwhile colleague, Professor Geraldine Macdonald, has a grant from the Northern Ireland Office to incorporate into the prestigious Cochrane database a systematic review of controlled trials of the effectiveness of social work. Straws in the wind perhaps, but plenty of them, and in a breeze growing stronger by the day. I am afraid that we shall never see 'the end of ideology' (present-day politicians trumpet it, but human beings as a species probably are not up to it), but such initiatives may help to contain some of its more distracting influences on our field.
This little political foray has been necessary for two reasons. First, because law, policy and regulation impinge upon the day-to-day behaviour of social workers more than is the case for any of the other helping professions, leading sometimes to a state close to 'frozen watchfulness' because of the range and variety of risks they are expected to manage. Second, because it is vital for ethical, technical, and again, for political reasons, that we are seen to be developing a cumulative knowledge base (getting better and better at what we do, not just getting different) (Sheldon 1978; Macdonald 1996). If you accept this, it means that we should explicitly select approaches to the problems which come our way on the basis of their research track record, rather than simply affiliating to 'schools', following 'par for the course' local conventions, or just 'doing what comes naturally'. In this regard, I teach and make some use of cognitive-behavioural approaches because of the trends in empirical research, not because of any prior ideological commitment.
Adopting such a position means that occasionally one has to change one's mind and professional behaviour in the light of new evidence. I know how uncomfortable this can be, because I have had to do it on several occasions: upon encountering the first reviews of effectiveness research showing that the methods in which I had been trained (psychoanalytical casework) had proven useless under test (Mullen and Dumpsen 1972; Fischer 1973 and 1976); being beguiled by the mysteries of R.D. Laing's theories about the origins of severe mental illness (Laing 1960), but then being unable to square the circle with the steady trends in genetic and other biological research (see Gottesman 1991; Sheldon 1994); most recently, feeling well disposed to the 'cognitive revolution' in psychology and psychotherapy and signing up to 'New Behaviourism', but then having to pick my way through empirical comparisons with 'Old Behaviourism', and feeling as a result only cautiously optimistic (cf. Feske and Chambers 1995).
John Maynard Keynes, when berated by a reporter for altering his position on economic policy, once said: 'When the facts change then I change my mind, pray Sir, what do you do?' 'Facts' are hard to come by in our epistemologically challenged discipline, but this general stance on evidence and practice seems to me exactly right. But why do we find this process of considered adaptation to new findings so difficult? Some possible factors are described below:
Social work requires an emotional commitment to sometimes unpopular causes. It is easy for such 'these truths we hold to be self-evident' values and sentiments to influence the rather more intellectual process of selecting knowledge for use. When this happens, the congeniality of results often takes priority over due consideration of the means by which the evidence for them was gathered. A kind of methodological relativism ensues, within which 'all propositions and findings are created equal', which in turn implies that 'all investigators have won, and all must have prizes'.
Social workers lead anxious lives. It is therefore understandable that they sometimes reach for the security blanket of grand, all-explanatory, Protean, irrefutable theories, rather than devoting scarce time to the demanding business of critically evaluating research papers in pursuit of useful trends. Reviews help here, but only if the authors 'show their working out' alongside their answers and if readers are duly critical of what is being pressed upon them. This is something we are getting better at, but with still some way to go (see Macdonald and Sheldon 1992; Macdonald and Roberts 1995).
The most pessimistic results which have a bearing on these issues come from enquiries (conducted, one imagines, with fingers crossed) into the reading habits of social workers. Over the years these have revealed that, but for a minority of eccentrics, social workers do not have reading habits (Brown and McCulloch 1975, Fisher 1996). (By the time this book is published, we should have further empirical information on this issue as a result of a large survey conducted by the Centre for Evidence-Based Social Services at the University of Exeter.) This has not been seen as much of a problem because of the distracting idea, previously popular among senior managers, that services can be provided by 'remote control', via (allegedly) improved departmental structures alone. In reality, day-to-day practice requires informed judgements to be made about highly individual combinations of problems, and catalogue-shop fantasies about pre-packaged, standardised services are likely to lead to ineffectiveness in the face of all but the most basic and predictable patterns of need. If evidence-based social services are to become a reality, then improved information facilities, with informed and active librarians on tap, and with ready access to electronic databases, will be increasingly necessary. More importantly, ways will have to be found to reinforce the a priori notion of considered action, rather than the short-term, budget or policy-driven, ad hoc containment of problems (see Jordan 1996; Hutton 1995).
The problem with the idea of evidence-based practice is that sometimes research findings come to the 'wrong' conclusions. Nowhere is this more the case than with cognitive-behavioural approaches. Reviews of outcomes in clinical psychology, psychiatry, psychotherapy and social work (Bergin and Garfield 1994; BABCP, 1993; Blackburn et al. 1981, Stuart and Bowers 1995; Macdonald and Sheldon 1992) point to the fact that if therapeutic work is needed, then across the range of problems we encounter, combinations of cognitive and behavioural methods are likely to be more effective than anything else. Moreover, these findings contradict the pattern usually found, that the stricter the methodology employed (for example, randomised controlled trials), the less favourable the 'signal to noise' ratio. An influential review of child protection research described the methodologies typically used to evaluate behavioural approaches as 'streets ahead' of the typical means of assessing the impact of other methods (Gough 1993).
What we are up against here is an 'antibody reaction' produced by ignorance, stereotype and the influence of our complicated history. Books such as this may help remedy matters, particularly given its concentration on live case material. However, the urgent need is for those of us who know the relevant research and its implications to combine to shift the postqualification training agenda away from 'What Works?' soirées to 'This Works' courses - with subsequent arrangements for consultation and supervision built in.
The best predictor of a believably positive outcome in a study of the effectiveness of social services is that the approaches under test have a close 'logical fit' with what is known from good-quality research about the aetiology of the problems in view. We know quite a lot about relapse-prevention in schizophrenia (Falloon 1984; Birchwood and Tarrier 1992; Falloon and Coverdale 1994; Kavanagh 1992), about depression (Blackburn et al. 1981; Hollon et al. 1991; Dobson 1989; Scott 1989), about panic states and phobias (Beamish et al. 1996; Mathews et al. 1981); about the origins of child management difficulties (Kazdin 1994; Sheldon 1995), and about the effects of partnership and family support work in child protection (Fraser et al. 1991; Thoburn et al. 1995). Some of these results are of such rigour and clarity that it is almost obvious what should be done at a therapeutic level. In my view, if anything does, the huge body of empirical research on learning falls into this category.