Friday, 26 February 2016

Risky business

Another nerdilicious blogpost. This one is about a central part of the NHS England-led plan ‘Building the right support’ , which is the latest Transforming Care reboot. ‘Building the right support’ makes the welcome point that reducing the number of people going into inpatient services will require much more extensive, earlier, and more useful community support.

As part of this, the plan says that there are an estimated 24,000 people ‘at risk of being admitted to hospital without the right support’ (page 36). My understanding is that there is quite a lot of head-scratching around how to identify in practice who these 24,000 people are, and then to work out what services are needed to support them so that they don’t end up in inpatient services.

Personally, I’m not convinced about this ‘risky cohort’ idea, and I think it might have some unhelpful unintended consequences.

First, the 24,000 figure is extrapolated from a study examining the prevalence of challenging behaviours amongst a population of people with learning disabilities in a large area of Wales (total population 1.2 million) in 2007 (Lowe et al., 2007). It is a relatively large, well-conducted study, with similar findings to previous research. However, using this as a benchmark to construct a ‘risky cohort’ raises a number of problems, including:

1) There are no nationally comprehensive data on who is currently showing ‘challenging behaviour’, and trying to construct such a dataset is likely to be highly expensive, difficult and time-consuming.

2) We also don’t know if the cohort of people with the types of ‘challenging behaviour’ used to derive the 24,000 number make up all/most of the cohort of people who end up in inpatient services. For example it’s likely to miss out relatively able people who end up in inpatient services via ‘offending’ routes, where even standard forensic risk measures don't seem to be brilliant at predicting what people do once they arrive in an inpatient service (O'Shea et al., 2015). From the sparse research that has been done it’s also likely to miss out some people with learning disabilities who end up in inpatient units with a diagnosis of a mental health problem - it's unclear how many of these people would be picked up as showing 'challenging behaviour' (Chaplin et al., 2015; Devapriam et al., 2014) .

3) In its original form, the definition of ‘challenging behaviour’ is a social construction, dependent on what the person is doing only within the context within which they are doing it. Hand someone a pair of glasses and you can be put in a temporary cell - flatten someone running past you and you can be a hero (both of these things happened to Burnley FC mascot Bertie Bee - see this article for a brief explanation of context ). Staff who don’t or choose not to really understand a person may very quickly reach for a ‘challenging behaviour’ label and punitive response that is completely unnecessary, establishing a punitive spiral with no ready exit. Identifying a risky cohort of relatively fixed quantity locates the ‘risk’ squarely within individuals, rather than paying careful attention to people’s histories, what matters to them, and the influence of environments (and other people).

4) We don’t know what impact changing patterns of/reductions in the group of people supported by/known to social services will have on who is visibly ‘at risk’. It’s entirely possible that the more able ‘offending’ cohort of people will stop being identified as people with mild learning disabilities at all (either in education or adult services). Although a lack of support might make this group more likely to get into trouble with the law, and might mean they are more likely to end up in prisons or even mental health inpatient services, they will become invisible to learning disability services and drop out of the ‘Transforming Care’ remit altogether. Which will buff up the Transforming Care target indicators no end, while doing a grave disservice to that group of people.

5) We don’t know what impact changes to inpatient services themselves will have (assuming that there ever are any!) on the quantity of people that could be admitted and the ‘problems’ that would make someone ‘eligible’ for an inpatient service, and therefore what a ‘risk cohort’ would look like. Over a ten year period, some of the 'reasons' for admission to a single inpatient unit, and what happened to people once admitted, changed quite substantially (Oxley et al., 2013).  And of course, if community support services do increase and improve, then this would also change who is ‘at risk’.

6) The boundaries between ‘inpatient’ services and other residential services (the infamous Daisy residential service) are becoming increasingly fuzzy, as local areas look to claim that their inpatient numbers are going down. The ‘risk cohort’ might be less likely to go somewhere called an ‘inpatient service’, but they might be more likely to end up somewhere with a different label but serving a similar function (residential special schools are part of this too, I think). As Mansell and colleagues said in 2010: "The large number of patients in low secure services raises the question whether this type of service is recreating the intellectual disability institutions which government policy intends to close" (Mansell et al., 2010).

So I think if you want to know anything meaningful about the group(s) of people who might be ‘at risk’, there are a number of questions you need to ask:

1) Does having an ‘at risk’ cohort, who are maybe getting extra support as a result of their risky status, create more de facto eligibility hurdles? Do people and families have to show a ‘problem’ to get this extra support, and what if the extra support reduces the ‘problem’ – does the support stop? How are the boundaries to be policed? Personally I think support needs to be offered universally as the only way to get round this gatekeeping, avoid unnecessary labelling, and develop coherent support that is rooted within meaningful communities.

2) ‘At risk’ of what? At risk of admission to an inpatient service is too restrictive (for the reasons I’ve outlined above) and will lead to ‘gaming’ the system from unscrupulous types. I think instead there needs to be a recognition that trying to get a definitive ‘at risk’ list of people will be likely to be highly expensive and slightly self-defeating as circumstances change. Instead, I would propose some more pragmatic steps to be done on an ongoing basis to monitor what’s happening, for example:
a.      We don’t know much at all about the histories of people who are being admitted to inpatient services. It should be relatively straightforward to find out some details of people’s histories and collate them either at the point of admission and also for people ‘averted’ from inpatient admission. This would need to be ongoing to look for potential changes in ‘risky’ histories as the pattern of inpatient services changes.
b.      We need to do something similar for people in residential special schools (and to examine their onwards trajectories too).
c.      We need to look at people who get/acquire some label of learning disabilities when they come into contact with the criminal justice system (police contact onwards) in terms of histories and what happens to them (diversion to forensic services, for example?).
d.      Within annual health checks we need better coverage of mental health problems and ‘risky’ behaviours.
e.      We need to find a way of understanding what is happening to the group of people formerly known as people with mild learning disabilities.

Within local areas, we need to build a comprehensive picture of what is happening to people that is monitored in real time. This shouldn’t need a complicated extra special database of ‘risky’ people. It needs to get a whole picture of where people are (including people who are sent out of area). Are apparent reductions in inpatient services accompanied by increases in emergency hospital admissions, or people getting into the criminal justice system, or into forensic services? As the number of people identified as eligible for learning disability services decreases, are there local increases in emergency hospital admissions or generic mental health inpatient admissions? I think in areas that have their act together, most of this information  will already be available – it is just very rarely put together in ways that make any kind of sense. I think this is much more achievable than identifying an ‘at risk’ cohort of people with learning disabilities.


Chaplin, R., Roach, S., Johnson, H. & Thompson, P. (2015). Inpatient Children and Adolescent Mental Health Services (CAMHS): outcomes of young people with and without intellectual disability. Journal of Intellectual Disability Research, 59, 995-998.

Devapriam, J., Alexander, R., Gumber, R., Pither, J. & Gangadharan, S. (2014). Impact of care pathway-based approach on outcomes in a specialist intellectual disability inpatient unit. Journal of Intellectual Disabilities, 18, 211-220.

Lowe, K., Allen, D., Jones, E., Brophy, S., Moore, K. & James, W. (2007). Challenging behaviours: prevalence and topographies. Journal of Intellectual Disability Research, 51, 625-636.

Mansell, J., Ritchie, F. & Dyer, R. (2010). Health service inpatient units for people with intellectual disabilities and challenging behaviour or mental health problems. Journal of Applied Research in Intellectual Disabilities, 232, 552-559.

O'Shea, L.E., Piccioni, M.M., McCarthy, J., Mason, F.L. & Dickens, G.L. (2015). Predictive validity of the HCR-20 for inpatient aggression: the effect of intellectual disability on accuracy. Journal of Intellectual Disability, 59, 1042-1054.

Oxley, C., Sathanandan, S., Gazizova, D., Fitzgerald, D. & Puri, B.K. (2013). A comparative review of admissions to an intellectual disability service over a 10-year period. British Journal of Medical Practitioners, 6, a611.

No comments:

Post a Comment