Of course, as a CRT Consultant I would (and do) reverse the argument: what should we do? What works? Problematically (for me), a commissioner might pose the same question: why should we fund you, or fund you in your current form? Therein lies the key problem in crisis care: a lack of good clinical data.
The backdrop is their original design and intent (Carpenter et al 2013): ‘gatekeepers’ intended to reduce admission rates and durations, with an unspoken expectation that ‘good’ would be somehow contained within. Far more so than in 2000 when this was first established, we live in a time of bed crises: 2017 has already seen this as almost consistent headline news. So, it’s not an unreasonable starting point to have a team mandated with helping to resolve this. However, it sets up a clinical, management, and academic conversation that has spinning wheels as its focus: crisis teams, crisis teams, hey-hey-hey, how many admissions did you avoid today?
What has been lost? Care. I have no doubt crisis teams provide great care to many people, and are usually staffed by individuals who want to do their best, but we fumble along trying to figure out just what that is. A couple of years ago, my then-MSc student Becci Carpenter undertook a qualitative exploration (Carpenter & Tracy, 2015), of CRT patient hopes and expectations, and their language was of ‘compassion’, ‘support’, and ‘helping me look after myself’. Nothing about gatekeeping. It has preoccupied me since how we might better understand the ‘treatment’ somehow contained within home treatment teams.
Into this comes a timely paper by Werbellof et al (2017) in the Lancet Psychiatry, exploring the clinical factors that impacted on outcomes in crisis services across two large mental health Trusts in London.
Accessing the routinely collected electronic records of individuals who utilised Crisis Resolution Teams in these Trusts (covering a population of just under one and a half million) between 2008 and 2014, they identified over 17,500 patient contacts. Their sociodemographic and clinical details and subsequent engagement with crisis services (inpatient admission, CRTs, and crisis houses) were explored, with a view to understanding patient baseline characteristics, and factors that affected outcomes. Univariate and multivariable Cox regression models estimated hazard ratios for time to admission, controlling for year and team of index CRT episode.
- Patients had a mean age of about 40 during their index Crisis Resolution Team episode, and in the year after CRT care, the rate of return to crisis services was 50 per hundred person-years. However, the richness lies in the analysis beneath
- Those with non-affective psychotic disorders and individuals over 65 had greater such use, as had those with the highest-tertile HoNOS scores – namely those who were more unwell
- Anxiety disorders, first CRT contact patients, and longer duration of CRT care had reduced use
- Individuals of Black ethnic origin and those who were unmarried were more likely to have readmissions, but interestingly, against one’s clinical expectations, social deprivation per se was not associated with subsequent outcomes.
- Clinical and sociodemographic factors impact on Crisis Resolution Team outcomes. We always knew this, and now we can show it
- The type of problem one has, impacts on the success or otherwise of CRT care
- These data should facilitate CRT service reflection in terms of supporting or perhaps upending anecdotal team wisdom. Are there some individuals who might need more input or monitoring than we imagine, or some who we might be able to pass on to other services more comfortably?
Strengths and limitations
The main strength of the work is its size. Almost 18,000 patient contacts evaluated means that it dwarfs (Ed: did you have to mention them!?) other work on Crisis Resolution Teams. Further, we now have clinical data associated with outcomes. The work takes us past simplistic admission duration and Mental Health Act detention figures.
The data are limited to two NHS Trusts, both in London, so the model of care (a UK CRT) and the population (London typically being a very heterogeneous and multicultural mix) may not map onto other areas. That said, the findings are primarily related to clinical factors, which helps instill confidence about how generalisable they are.
The main limitations are, from my perspective, unfixable ones in its current form. It’s a retrospective analysis; outcomes are somewhat crude (tracking ‘readmission’, so back to the old problem of missing ‘care’); and we cannot link what clinical input(s) are helping or not (for example, is it time with a key-worker, medication monitoring, psychological support). Of course, the authors do not set out or claim to do this, but the paper self-invites these next steps.
Werbeloff et al’s work is doubly helpful. ‘Big data’ may be an annoying phrase by this point, but we are in the era of large-scale evaluations: the golden era of more classical randomised controlled trials may have past, replaced by such data-sets, complete with lots of rich noise to unpick. It is pleasing to see this style of work be applied to crisis care. Secondly, clinical markers are emerging: what works and for whom? We’re starting to evidence this. Clearly, much more, and more sophisticated, work in the field is necessary, but a flag has been planted in the sand.
2017 promises to be a good time for crisis care. University College London’s Crisis Team Optimisation and Relapse Prevention (CORE) programme, led by Professor Sonia Johnson, is producing more data (Lloyd-Evans et al, 2016) on factors that lead to higher quality care. The National Collaborating Centre for Mental Health is working with the Royal College of Psychiatrists and NICE to produce guidelines on achieving better access to crisis care (disclosure: I sit on the Expert Reference Group helping draw these up). We have heretofore lacked both standards, and the teeth to enforce them, but this should change: not only will there be national guidance on what one might expect in a crisis, but data will be published and benchmarked. Service users, carers and relatives, and CRT staff should be able to work together with this; where a service is not achieving expectations (often due to resources, not good will), everyone should be asking for, and expecting more from commissioners.
Some next steps that need to happen:
- The routine recording of patient outcome and experience measurements (PROMs and PREMs) needs much greater consistency within and across teams, with evaluation of who gets better, who does not, and consideration of why this might be happening
- More sophisticated work would digitize this, link with inputs (nursing, medical, psychology, social work etc), and try to refine resources
- This doesn’t need large scale academic centres: local teams can and should undertake this.
Can we really bottle sunshine, and ever find out how to quantitatively measure ‘care’? I’m not sure we ever have to; we just need to remember that it is usually the magic that matters, and should centre our clinical practice.
THANKS TO @MENTALELF FOR REPRODUCTION OF THIS BLOG WWW.NATIONALELFSERVICE.NET