Where are the hypotheses when you need them?

Summary

Poor staff morale is a pressing problem in UK mental health services, especially for acute in-patient wards, community mental health teams and social workers. Instead of interpreting low morale using a demand–control–support model, it is suggested here that simply being honest about what should be expected of staff and stopping constant criticism and reorganisation might be more fruitful.

Johnson and colleagues completed a Herculean undertaking to report on the morale of the English mental health workforce. ¹ They surveyed 100 admission wards and 36 community mental health teams (CMHTs). From 3545 questionnaires distributed they received 2258 valid returns, a remarkable success introducing the study and encouraging engagement with it. The magnitude of this achievement is brought home by the list of measures in the questionnaire (of which more later). There is no estimate of how long it took each participant to complete, but it cannot have been brief. We are also told that this paper only lists the measures used in the analyses presented; there will be several more in subsequent papers. So, an exemplary exercise in engaging clinicians.

The authors set out the aim of their work with clarity: ‘[…] to address this need for large-scale evidence on morale in the mental health workforce’. ¹ They outline three specific objectives: seeking a shorter list of higher-order indicators for variations in morale, testing for variations between specific staff groups and subspecialties, and testing for variations between service settings.

Damned if you do and damned if you don't

The paper could, however, be accused of telling us what we already know. The survey shows that overall we do well in avoiding cynicism and have acceptable levels of personal accomplishment. Emotional exhaustion is highest in acute general wards, CMHTs and among social workers. Many of us might have guessed this. These three groups have least control over the demands made on them and have to continuously confront patients’ distress head on. ‘Emotional strain’ and ‘positive engagement’ ¹ each have a handful of associations with either service or personal characteristics. However, we should not be dismissive of these findings. Mental healthcare is littered with our ‘knowing’ things that turn out not to be so; ² getting the data matters. Having some of our preconceptions confirmed may also help us get our bearings in this volume of information.

There is so much in these results that it is a challenge to work out what is important and what is not. We have 8 different work settings, 7 professional groups and by my reckoning at least 78 individual variables in the questionnaires. Not surprisingly these generate several intriguing associations. In deciding how to present them, the authors are damned if they do and damned if they don't. If they do not present them all, they may be suspected of drawing conclusions from partial data derived from leading questions. If they do present them all and compare them with an accepted generalist model of professional morale, the demand–control–support model, ³ their conclusions might be perceived as vague or unconvincing (what exactly is ‘management support’?). The authors have already provided some support for Karasek's demand–support–control model's fit with the reports of this staff group in a previous paper. ⁴ However, it is not the only model for such an understanding (e.g. the prosocial motivation literature ⁵ ), so it would help those not so familiar with this area to have had the choice explained some more.

In praise of hypotheses

Without clear and relevant hypotheses the risk is that we simply use the findings in a pick-and-mix manner to confirm our prejudices. Where they comply with our assumptions (e.g. the emotional exhaustion in acute general wards, CMHTs and social workers) there seems no problem interpreting them. Or is there? This is where it might be so useful to have some targeted hypotheses and to know which were considered and which rejected (and why they were rejected).

Is it just that being a social worker or working in acute admission wards or CMHTs is so relentless and uncontrollable (‘high demands’, ‘low autonomy’)? What other interpretations have been, or could have been considered? One that surely deserves attention is that these three services have been subject to the most consistent, harsh and, I would argue, unjustified public and professional criticism for the past decade and more. Community mental health teams are the failed ‘old model’ of care that should be swept aside ⁶ to make space for shiny new alternatives or, to add insult to injury, to act as the control condition in trials of their replacements. The ‘pioneer’ effect ⁷ which exaggerates outcome differences in community psychiatry research may act, in part, through the demoralisation of the controls from being called just that, ‘controls’. That these CMHTs have inconveniently shown themselves consistently to be about as good as their proposed alternatives ⁸ does not seem to filter back either to them or to their commissioners. Acute in-patient wards find themselves similarly stigmatised as ‘old fashioned’ or ‘untherapeutic’ for which new alternatives must urgently be found. ^9,10 Even worse, they were declared ‘unsafe, overcrowded and uninhabitable’ by the former President of the Royal College of Psychiatrists, who added: ‘I would not use them, and neither would I let any of my relatives do so’. ¹¹ Yet the acute in-patient ward is the one component that all mental healthcare services across the world seem to rely on despite other differences. Could it be that it is getting something right? Criticising social workers has become such a national pastime that it is testimony to the human spirit and basic humanity that young people still train for this vital and unappreciated role.

Could these three groups be so vulnerable to this endless criticism because of their lack of autonomy and control? The findings reported in Johnson et al's study would support this interpretation but is this explanation adequate? If it were the full explanation, then why do we not observe the same low morale in, for example, Salvation Army (or other religious charity) staff in homeless shelters? Could the more toxic aspect of these open-ended jobs be the lack of a clear practice model with realistic aims and standards, rather than just the volume of work? Community mental health teams and in-patient wards evolved without a strong, explicitly defined care model, nobody ‘owns’ them or advocates strongly for them. Increasingly, expectations and goals are imposed on them by others.

Careful anticipatory monitoring and simple emotional support are two of the most frequent activities of these ‘low-morale’ staff. ¹² This is no easy job with resentful community patients or acutely disturbed in-patients; it requires tact and tolerance of the very highest order. These are surely core skills of mental health workers. Yet where do they appear in job descriptions and what status do they get? Such staff are often left with the feeling that they have underperformed because they are not spending their time doing cognitive–behavioural therapy or motivational interviewing or some other easily labelled, highly technical intervention. The Salvation Army care worker has a much more realistic understanding of what he or she has to do, and what he or she can do.

What to do about low morale

Johnston and colleagues have provided us with much needed data. They rightly identify the need for even further work (quantitative and qualitative) to understand the antecedents of the associations they found. This is particularly so if interventions to address low morale are to be developed and implemented. It would be a shame if that debate were restricted prematurely to the framework of the demand–support–control model. This carries a real risk of potential oxymorons such as increased management support for clinical autonomy or even (heaven protect us from) well-meaning team-building exercises.

Nearly every motorist thinks he or she is a ‘better than average’ driver. In my experience, nearly every healthcare system thinks its mental health provision is ‘much worse than average’. The result is endless change and meddling and, not surprisingly, confused and demoralised clinicians. Before trying to reduce demand, increase autonomy or provide more support, two things could be tried. First, agree and publicise a realistic definition of what CMHTs, acute wards and social workers should do. Second, hold off on endless criticism and reorganisation (which always carries an implicit criticism) for a decent period. If they do not make a difference, then sophisticated interventions for improving morale may well be needed.