Psychodynamic programmes for personality disorders: residential versus community treatment


Down in the woodland, all the elves are very passionate about improving the experiences of those labelled with ‘personality disorder’.  We were very sad when we wrote our last blog about the lack of specialist services back in March, but at least we took some comfort that services had expanded and improved over the years.

We get furious when people are taken from the woodland, sent miles from their family and friends and generally lose their accommodation while being compelled to receive treatment they don’t want. With this in mind we were very excited to read this new paper, which compares community treatment, step-down treatment and residential treatment over a 4 year period (Chiesa et al, 2017).

'Personality disorder': a real diagnosis or an unhelpful label?

‘Personality disorder’: a real diagnosis or an unhelpful label?


The Cassel Hospital is synonymous with the therapeutic community ethos and has been banging out research about its efficacy and cost effectiveness for many years (Chiesa et al, 2002a, 2002b, 2002c, 2003, 2006). In this prospective (non-randomised) study they compared patients diagnosed with 1 or more personality disorders who were referred to one of three possible care packages:

  1. CBP (n=30): Community Based Psychosocial programme. Contracted for 2 years, but often increased by up to a year, with 2 group sessions and 2 nursing contacts per week;
  2. RT (n=45): Residential Treatment for 1 year;
  3. RT-CBP (n=87): Residential Treatment lasting 6 months, which then stepped down to the CBP community programme

I’m told that the study had an observed power of 81% for a one way omnibus fixed effect analysis of variance. Someone else is going to have to translate that, but I think it means we can trust what the authors are telling us while also being a bit wary of proclaiming it as a solid truth.

Clients were rated at 0, 6, 12, 24, 36 and 48 months. At baseline they were given a diagnostic assessment and rated as clinically severe or not, based on the number of personality disorder diagnoses they met, their criteria for and their general intake psychiatric distress scores. They also took part in the Brief Symptom Inventory General Severity Index. What’s interesting here is that the CBP and RT groups had the same percentage of people rated as severe, while the CBP group had a higher average general severity index score. This will become important later so just to be clear, the people in the CBP community group were rated as more severe than those who were inpatients in the RT group.


Compared to residential treatment (RT), patients found the Community Based Psychosocial (CBP) programme and the Residential Treatment step-down to community (RT-CBP) programme easier to complete. The RT group had 40% drop-out within 5 months. While no-one in the CBP group died by suicide, 1 person did in the RT-CBP group and 5 people did in the RT group.

In terms of the General Severity index (GSI), the CBP group magnitude of improvement was large. For RT-CBP it was moderate and for RT it was small. The researchers found that severity of presentation was not a significant predictor of outcome, both in general and when looking at individual programmes. Where people were identified as most severe, CBP had the most dramatic effect on lowering the GSI rating. You might want to read that bit again.

For self harm, the CBP group saw a marked reduction after 12 months compared to RT. CBP was also found to reduce the likelihood of suicide compared to RT. Read that again too.

This study suggests that residential treatment not followed by outpatient treatment was associated with high dropout and inferior improvement in psychiatric distress and self‐harming behaviour.

This study suggests that residential treatment not followed by outpatient treatment was associated with high dropout and inferior improvement in psychiatric distress and self‐harming behaviour.


The authors themselves identify a number of limitations in the study that could inform how the results are interpreted:

  • The sample was naturalistic (those who turned up) as opposed to being randomised.
  • Also there was a high level of drop out, particularly from the RT stream, meaning that those results are probably underpowered.
  • The biggest limitation for me when drawing conclusions about the efficacy (or not) of residential treatment is that the Cassel is somewhere that people come voluntarily. Here I think you have to have different expectations about engagement and therapeutic alliance than you would with my usual experience of people going to placements, where people are carried through the door by a phalanx of burly staff.

Those caveats aside, let’s have a look at the conclusions the authors came up with.


The first was that community treatment was significantly more effective than residential treatment. It’s also more likely to keep people engaged. Paradoxically, the lower levels of support led to reduced self harm. The authors note:

non-significant changes in deliberate self injury and suicide attempts in RT raise questions about the effectiveness of long term RT to reduce the risk of self harm.

Obviously in a more restrictive setting than the Cassel, reductions in self harm can potentially be achieved by removing every object possible to self harm with, high levels of observation and regular restraint. What this does in terms of building trust and independence is another matter.

The authors suggest that part of the lack of success of RT over 48 months was due to the very contained nature of RT treatment: in for a year and out again, potentially back to nothing. The authors speculate that organisations that opt not to provide specialist community treatment, send people to RT then return clients to the lack of specialist treatment that required their RT in the first place are effectively pouring their money away. The ethical implications of compelling people to go to such units for what is likely to be ineffective care is chilling.

The authors note that distress and self harm increased in the first 6 months of RT and suggest this may indicate the possibility of iatrogenic effects (help making things worse) for a large number of patients. I hope that this can make a dent in the conventional wisdom that hospital is a safe place and help us realise that by compelling people to go to RT we may be increasing their risk of self injury and death (Paris, 2004).

This study found that distress and self harm increased in the first 6 months of residential treatment. An example of iatrogenic harm?

This study found that distress and self harm increased in the first 6 months of residential treatment. An example of iatrogenic harm?

Implications for practice

The authors feel their work adds to the body of studies indicating intensive community treatment is at least as effective as residential treatment (Arnevik et al, 2010; Gullestad et al, 2012). They call for treatment models to shift from low volume/high intensity (locking a few people up) to specialist high volume, moderate intensity (treating lots of people in the community). In these times of austerity this is vital given that community treatment can be 6 times cheaper than residential care. If the efficacy is similar we need to confront some challenging issues in organisations that prefer to spend millions treating a few, when treating more people at home would be as effective, cheaper and potentially safer. In The Ailment (1957) Tom Maine suggested that giving patients sedatives was a marvellous intervention for staff. If efficacy is similar or better in the community, if the costs are cheaper in the community and if residential treatment can result in more self harm and suicide than in the community, surely we need to be very, very wary when we are compelling people to receive residential treatment? It needs to be because it is the best clinical decision rather than an intervention to pass on a problem, avoid blame for an organisation or make staff feel better.

In the absence of clear evidence of superior efficacy, we suggest that long term residential treatment should either be discontinued or confined to cases where the cost-benefit of treatment outweighs the cost-benefit of maintaining the client in the community or when an effective treatment is not available.

So a residential placement generally won’t give you any change out of £200,000. That is some community package that is going to be more expensive than a placement. The authors kindly suggest that RT can be considered when an effective treatment isn’t available. I’d suggest that for £200,000 a year, there are very few treatments that couldn’t be made available.

The authors note that other studies have found benefits in admitting clients with a BPD diagnosis for ‘containment’ and ‘stabilisation’, but this is a very different option from packing someone up and sending them away for ‘a year of treatment’ that miraculously turns into 3 years; where they put on 6 stone in weight and find themselves on a whopping dose of Clozaril. The authors also note that the subjects of the above studies tend to be less severe than those in this one, including those in the CBP program. They conclude that paradoxically it is the severe PD groups who are more likely to experience benefit from specialist community treatment. This is something that every Trust in the country needs to write on the staff room walls, because usually the higher the severity, the greater the urge to physically contain, restrain and control kicks in. This fits with another study by the same authors (Chiesa et al, 2011), which found that people who chronically self harmed were the most likely to experience iatrogenic effects when treated as inpatients.

In conclusion, the authors say that community treatment is effective and that long term residential treatment offers no advantages over community treatment. They feel that the long term role of residential treatment needs to be reconsidered and that rather than condemning everyone with severe difficulties to compulsory treatment, we should carefully consider which people are most likely to benefit and which are likely to be harmed.

The authors of this new study say that community treatment is effective and that long term residential treatment offers no advantages over community treatment.

The authors of this new study say that community treatment is effective and that long term residential treatment offers no advantages over community treatment.

Personal view

So what are we to make of these conclusions? Firstly, let’s take my perspective on this with a huge bucket of salt. I am not a neutral party in this, having railed against what I see as the injustice of forced inpatient treatment in my own blog and having set up a company to try and provide an alternative to what I view as ineffective, coercive and unethical practice. But…

I think healthcare staff are trained to make people better. When people don’t get better our temptation is to do more. Tom Maine (1957) talks of how our desperation to be effective results in an ‘heroic surgical attack’ where we do whatever we can, however dangerous to MAKE our patients better. For me, this explains why you see someone with a personality disorder diagnosis on 12 different medications heading to the ECT suite. When everything else is exhausted it seems that teams get stuck in; the idea being that the only answer is to send the patient to an idealised elsewhere. Other factors come into play here; very often the “but what if they kill themselves?” question. This stops people thinking straight and can lead to clinically dubious decisions, that can at least be defended with the retort “But we sent them to a specialist unit, we did everything we could, so it’s not our fault.”

This new study drives a coach and horses through that defence. It forces us to be very careful about the decisions we make with (or inflict upon) individuals with complex problems. We need to bear in mind that regardless of severity (according to this study and previous research), community treatment is no worse and likely more effective than residential treatment. We need to bear in mind that residential treatment may increase the risk of self harm and death. When services compel people to go to residential treatment against their will, they need to be comfortable that they can be making a decision with the worst clinical outcomes and placing the client at greater risk. This will be bad news for the 8% of English Trusts who opt not to follow NICE guidelines for BPD and an awful shock to the majority of Welsh Trusts who do the same.

£1 million to treat 5 people for a year is a statistic that few countries can afford, especially when that treatment is less effective than something six times cheaper. Trusts and CCGs need to reign in the urge to send somewhere better (or at least somewhere else) and spend this money in a way that can offer better outcomes for hundreds, rather than poor outcomes for a few.

This paper adds to a growing body of work suggesting that for some, hospital can be potentially lethal (Large et al, 2017). For those who can just about stay alive in the community, but are utterly reliant on staff to save them in hospital, residential treatment might kill them. For those who cut in the community but ligature in hospital, residential treatment might kill them. For those who are passive in the community but confrontational/aggressive on the ward, residential treatment might mean they are locked up forever. The complexities of working with people who have lived through multiple traumas are extreme, but this paper reminds us of the need to act with thought, rather than react with emotion.

We need to bear in mind that residential treatment may increase the risk of self harm and death.

We need to bear in mind that residential treatment may increase the risk of self harm and death.


While all the above is an accurate (or at least as accurate as I can make it) summary of the study by Chiesa et al (2017), the paper itself is perhaps not the most accurate reflection of the residential treatment the Cassel offers. The residential service adapts according to the latest research and is more like a 9 month treatment. Contrary to the original paper (Chiesa et al, 2017), a referral isn’t accepted unless a step down treatment/service is available. People go home at the weekend and the whole ethos is around negotiation and consensus. This is in stark contrast to the services who describe themselves as specialist, but have nothing to back it up, other than a CQC rating that doesn’t look at how special its specialism is at all. In general these ‘specialist’ services or locked rehab services (another term that luckily doesn’t have any standards to denote its quality) will assess people once on a ward and the service user will go with the lowest bidder. As opposed to the strict time limit at the Cassel, these services will hang on to clients for as long as the money is coming in. There is no incentive for them to return clients to where they came from and a substantial disincentive to do so.

My point in this article was to say “this is a paper about the Cassel, imagine what happens in services without this level of collaboration, quality and oversight.” I do imagine this and it terrifies me.

Conflict of interest

Keir Harding is an Occupational Therapist and DBT therapist who specialises in working with people with difficulties often labelled as personality disorder. He is also the Founder of Beam Consultancy; a company that provides training, consultancy and intervention for those living and working with complex mental health issues.


Primary paper

Chiesa M, Cirasola A, Fonagy P. (2017) Four years comparative follow-up evaluation of community-based, step-down, and residential specialist psychodynamic programmes for personality disorders. Clin Psychol Psychother. 2017 Jul 26. doi: 10.1002/cpp.2109. [Epub ahead of print] [PubMed abstract]

Other references

Paris J. (2004). Is Hospitalization Useful for Suicidal Patients with Borderline Personality Disorder? Journal of Personality Disorders: Vol. 18, Suicide and Borderline Personality Disorder, pp. 240-247.

Arnevik E, Wilberg T, Urnes Ø, Johansen M, Monsen JT, Karterud S. Psychotherapy for personality disorders: 18 months’ follow-up of the ullevål personality project. J Pers Disord. 2010 Apr;24(2):188-203. doi: 10.1521/pedi.2010.24.2.188.

Gullestad FS, Wilberg T, Klungsøyr O, Johansen MS, Urnes O, Karterud S.Is treatment in a day hospital step-down program superior to outpatient individual psychotherapy for patients with personality disorders? 36 months follow-up of a randomized clinical trial comparing different treatment modalities. Psychother Res. 2012;22(4):426-41. doi: 10.1080/10503307.2012.662608. Epub 2012 Mar 15.

Chiesa M, Sharp R, Fonagy P. (2011) Clinical associations of deliberate self-injury and its impact on the outcome of community-based and long-term inpatient treatment for personality disorder. Psychother Psychosom. 2011;80(2):100-9. doi: 10.1159/000320975. Epub 2010 Dec 23.

CQC (2017) The state of care in mental health services 2014 to 2017 (PDF) Care Quality Commission, 2017.

Chiesa, M., Bateman, A., Friis, S. & Wiberg, T. (2002a) Patients characteristics, outcome and cost-benefit of hospital-based treatment for patients with personality disorder: A comparison of three treatment programmes, Psychology and Psychotherapy 75: 381-393

Chiesa, M. & Fonagy, P. (2002b) From the therapeutic community to the community: A preliminary evaluation of a psychosocial outpatient service for personality disorder, Therapeutic Communities 23: 247-259

Chiesa, M. & Fonagy, P. (2003) Psychosocial treatment of personality disorder: 36-months follow-up, British Journal of Psychiatry183: 356-362

Chiesa, M., Fonagy, P., Holmes, J., Drahorad, C. & Harrison-Hall, A. (2002c) Health service use costs by personality disorder following specialist and non-specialist treatment: A comparative study, Journal of Personality Disorders 16: 160-173

Chiesa, M., Fonagy, P. & Holmes, J. (2006) 6-year follow-up of three treatment programmes for personality disorder, Journal of Personality Disorders 20: 493-509

Large MM, Chung DT, Davidson M, Weiser M, Ryan CJ. (2017) In-patient suicide: selection of people at risk, failure of protection and the possibility of causation. British Journal of Psychiatry Open May 2017, 3 (3) 102-105; DOI: 10.1192/bjpo.bp.116.004309

Could psychiatric inpatient admission cause suicide?

Photo credits

Share on Facebook Tweet this on Twitter Share on LinkedIn Share on Google+