Randomised controlled trial of acute mental health care by a crisis resolution team: the north Islington crisis study

BMJ 2005; 331 doi: http://dx.doi.org/10.1136/bmj.38519.678148.8F (Published 15 September 2005)
Cite this as: BMJ 2005;331:599

Recent rapid responses

Rapid responses are electronic letters to the editor. They enable our users to debate issues raised in articles published on bmj.com. Although a selection of rapid responses will be included as edited readers' letters in the weekly print issue of the BMJ, their first appearance online means that they are published articles. If you need the url (web address) of an individual response, perhaps for citation purposes, simply click on the response headline and copy the url from the browser window.

Displaying 1-10 out of 22 published

Dear Sirs,

we read with interest the randomised controlled trial by Johnson et al (1) . We are pleased to finally have a robust short to medium term trial which goes a long way to finally settling the efficacy in terms of prevention of admission to hospital, however, doubts may still linger regarding a possible Hawthorne type effect and the longer term sustainability of the model .

Johnson et al also demonstrated (2) that the 6 week post crisis admission rate fell by 22% and that patients prefered the experience of crisis intervention . Crisis Teams have been in existence since 2000 with the introduction of the NHS Plan by the Department of Health promising 335 such teams by 2004. This target is now exceeded .There has also been interest displayed in home treatment teams working collaboratively with early intervention services in treatment of first episode psychosis (3) .

I am interested in the huge emphasis placed on these specialist services with very little information available about their sustainability . A controlled study by Audini 1994 showed that home treatment teams suffer low morale after 20 months and issues of “burn out” and job satisfaction are linked to stress in community based staff which is said to be higher than in-patient units(4) (5).

An initial satisfaction survey (unpublished results available from author) of a mature crisis intervention team based in a large inner London psychiatric hospital showed that it was more critical of its own performance whereas the usual community mental health teams were extremely satisfied . Reasons for dissatisfaction are numerous. Because of the implications of these findings and the narrow evidence base for the efficacy of crisis teams, the sustainability of these teams demands deeper investigation.


1.Johnson S, Nolan F, Piling S, Sandor A, Hoult J, McKenzie N, White IR, Thompson M, Bebbington P. Randomised controlled trial of acute mental health care by a crisis resolution team : the north Islington crisis study. BMJ2005;331:599. (17 September )

2.Johnson .S, Nolan.F, Hoult.J, White.I.R, Bebbington.P, McKenzie.N, Patel,S.N, Pilling.S. Outcomes of crises before and after introduction of a crisis resolution team . The British Journal of Psychiatry 2005 187: 68 -75

3.Gould.M, Theodore.K, Bebbington .P, Hinton.M, Johnson. Initial treatment phase in early psychosis : can intensive home treatment prevent admission . Psychiatric Bulletin 2006 30:243- 246.

4.Audini .B, Marks. IM, Lawrence . RE, Connolly J, Watts.V Home –based versus out-patient /in-patient care for people with serious mental illness. Phase II of a controlled study . British Journal of Psychiatry 1994 165;204-210

5.Prosser. D, Johnson S, Kuipers .E, Szmukler .G, Bebbington .P and Thornicroft. G. Mental Health “burn out” and job satisfaction among hospital and community –based mental health staff . British Journal of Psychiatry 1996 164; 334-337

Reuven M Magnes Specialist Registrar reuven.magnes@nhs.net St Clements Hospital , 2a Bow road , London E3 4LL.

John Lowe Consultant Psychiatrist St Charles Hospital , Exmoor Street , London W10 6DZ.


Competing interests: None declared

Competing interests: None declared

Reuven M Magnes, specialist registrar

John Lowe

St Clements Hospital , London E3 4LL

Click to like:

Dear Sir, I read the report by Johnson et al(1). with interest. However, unusually for a randomised controlled trial published in the BMJ, the dose of the active intervention (the size of the crisis resolution team) was not specified, nor the weight of the patient (the size of the population weighted by a proxy measure of psychiatric morbidity). Crisis resolution teams are already in place in many areas of the country. They were set up before this evidence that they can be effective, at least over a short period, and though adverse events may be more frequent. However, anecdotal evidence (rumour) suggests that the size and skill-mix of such teams in relation to the population which they serve differ widely.

I would be most interested to know the number of Whole Time Equivalent professionals making up this crisis resolution team, their discipline, and grade. I would also be most interested to receive more information about the population from which the study sample was drawn i.e. the number of adults between the ages of 18 and 65 in the study population served by the crisis resolution team. Further, a number of proxy measures of psychiatric morbidity are available. None are entirely satisfactory, but the Mental Illness Needs Index (MINI)(2)is perhaps the most widely used. The MINI weighted population size would be valuable. This can then be used in conjunction with the information about the team to inform service planning.

Yours faithfully,

Tom Hughes Consultant Psychiatrist

References 1. Johnson et al. (2005) Randomised controlled trial of acute mental health care by a crisis resolution trial: the North Islington crisis study. BMJ 331, 599. 2. Glover GR, Robin E, Emami J, Arabscheibani GR. (1998) A needs index for mental health care. Soc Psychiatry Psychiatr Epidemiol. 33, 89-96.

Competing interests: None declared

Competing interests: None declared

Tom Hughes, Consultant Psychiatrist

Holly House, St Mary's hospital, greenhill Road, Armley, Leeds LS12 3QE

Click to like:

EDITOR-It was interesting to read the paper of Johnson et al on the north Islington Crisis study {BMJ 2005;331,599}. We have similar findings with the Warwick home treatment/crisis team {which is based at Yew Tree house in Leamington Spa). Congratulation for both!

Competing interests: None declared

Competing interests: None declared

AK Al-Sheikhli, Loc Consultant General Adult Psychiatry

Yew tree house,87 Radford Road,Leamington Spa,UK

Click to like:

Dear Sir,

I read with interest, the north Islington crisis study. I congratulate the authors on their attempt to do a randomised control trial on the subject. I did have some doubts regarding the data presented in the study.

1. Was a baseline BPRS done to ascertain the severity of psychiatric illness, because a crisis may in no way be due to the psychiatric illness, but may be situational. It would also help us to see if the patients succesfully treated by the Crisis teams were significantly different from those who needed admission, despite crisis intervention.

2. BPRS is generally applicable to psychotic disorders, but since the sample consisted of Major Depressive Disorder and substance use disorder etc, it would have been wise to use other measurements as well.

3.The duration of crisis intervention given should have been mentioned as Dr Debashish Basu notes, if it was given for more than 8 weeks, then it would not be an ideal outcome measure.

4. Technically, since this is a randomised control study, people who were admitted in the experimental group should be considered as "DROP OUTs" due to inefficacy of the intervention. One interesting thing would be to see how many people dropped out from hospital admission,(if there is an equivalent to that) and compare the two, since we want to see the effectiveness of the intervention.

Rajeev Krishnadas

Competing interests: None declared

Competing interests: None declared

Rajeev Krishnadas, SHO in Psychiatry

South Tyneside disrtrict Hospital, South Shields, Tyne and Wear NE34 0PL

Click to like:

29 September 2005

My response is not only to the particular study, but to the concept of crisis teams in general.

While it is important that NHS has to act in such a way so as to best utilise its resources, patient choice has also to be given its due importance. I wonder when we talk of the role of crisis teams to "prevent" admissions, we appreciate the need to give patients the choice as to either of the two alternatives. For this particular study, it would be interesting to know if the team evaluating the patient in crisis setting involved the patient in deciding about crisis team support or admission. This is especially important as there are no definite studies to prove that either of the two approaches have a better long term outcome.

Doctor – patient relationship: While crisis teams seem to do a wonderful job at keeping people out of hospitals, it is also worthwhile to note that most of these teams are not composed of doctors. Hence, it would seem as if during the period of acute illness, with the crisis team support, the patient was managed without being as frequently assessed and treated by doctors as had the patient been admitted. While this would help decrease the workload of the doctors, it does raise the ethical question as when the patient most needed help (during period of acute illness) they were seen more by trained nurses than doctors. While trained nurses are fully competent in assessing the patients, the patients might appreciate it more if they are seen by their doctor, during the worst part of their illness.

In the end, the patient should have the right to decide.

Competing interests: None declared

Competing interests: None declared

Nikhil Bhandari, Senior House Officer

Gloucestershire Partnership NHS Trust, Gloucester, GL1 3WL

Click to like:

28 September 2005

Its nice to learn that the crisis management teams lower the levels of burden on hospitals and their respective expences, but are all the beds in hospitals totally occupied by needy patients?Many a patients,after being treated, are still in the hospital due to some other reasons like housing problems,financial crisis etc.Some just land up again in wards because they have repeated long term compliance issues.Aren't there any ways to solve these non-medical problems,at a much faster rate and a place other than hospitals, so that the people who prefer hospitals over home management can get their due spaces?

Competing interests: None declared

Competing interests: None declared

shreyasi sharma, clinical attachae,psychiatry

sharma vikas

norwich ,NR6 5BE

Click to like:

EDITOR-The study by Johnson et al.1 is an important contribution to the area of home treatment for acute mental health crises vis-à-vis usual care standards. This will no doubt have an important bearing on policy and implementation decisions. Before these results are accepted for their wider implications, however, certain queries and comments need to be responded to.

First, I am unclear regarding the validity of one of the two primary outcome measures, i.e., hospital admission rates. Preventing hospital admission is an essential component of the intervention offered by home treatment itself, by its very definition. Thus, so long as the intervention is in operation (i.e., the crisis resolution and home treatment team is operating actively), prevention of hospitalisation cannot be an outcome measure because this is part of the intervention itself, just as the number of patients taking active medication during a placebo-controlled trial cannot be considered an outcome measure, since the dispensing of the active medication is the intervention itself. In essence, an intervention cannot be an outcome as well. Hospital admission rate does become an outcome criterion once the crisis team has done its job and decided to formally disengage, i.e., once the intervention is over. Unfortunately, we are not informed of this most crucial variable, i.e., the duration of the crisis intervention during the study period, and whether the hospital admission statistic was considered only after the crisis team had disengaged from the patient after its intervention. It is further interesting to observe in this regard that there was hardly any difference in additional admissions between the experimental and control groups after the first eight weeks till six months (for psychiatric ward admissions: 7% vs. 8%; for crisis house admissions: 5% in both groups; and for any type of admissions: 11% vs. 8% respectively). This supports the notion that the majority of the differential rates in admission was achieved early on during the trial, when the operation of the crisis team itself, by its very nature of intervention, kept the admissions at bay. This intervention, understandably and as admitted by the authors, did not help the more severely ill patients (who were underrepresented in the trial; further, the experimental group had significantly lower baseline symptom severity and total baseline HoNOS scores than the control group, a fact mentioned in the full online version of the paper but not in its printed version); neither did it help to significantly reduce involuntary detentions.

Second, it is noteworthy that more than 40% of the ‘control group’ patients did not actually require admission to a psychiatric ward (and nearly one-third were not hospitalised either in a psychiatric ward or a crisis house) during the first 8-week phase, although, again by definition, all of them were admissible. This again is an important sample characteristic regarding the severity of the patient group under consideration.

Third, the figure showing the flowchart of participants (shown in the online version) indicates that 125 out of the 135 patients allocated to the experimental group actually had contact with the crisis resolution team. It is not clear what happened to the remaining ten patients in the experimental group. The results are reported for all 135 for the eight- week period.

Fourth, contrary to the authors’ claim, a randomised controlled trial in this area was indeed published from the Institute of Psychiatry 13 years ago in this very Journal.2 Although formal Community Mental Health Teams were not in place at that time, the study did have a similar aim, design, methodology and setting, and reached similar, though perhaps somewhat less enthusiastic conclusions (“home based care may offer some slight advantages over hospital based care for patients with serious mental illness and their relatives”). Citation of this important study could have helped the present one to be placed in a correct historical perspective.

In the passing, please note that the ‘Table 1’ mentioned at the bottom of the Table 2 of the printed version actually refers to the Table 1 of the online full version (baseline characteristics) and not to the Table 1 of the printed version (main outcome measures): this was a little confusing for me.

In summary, I believe we need more clarity regarding the nature and duration of the crisis intervention period itself, we need to be assured that the admission rates cited as ‘outcome measure’ was not an artefact of the intervention process itself, and the relatively milder nature of the patient groups and their crises for which the intervention was apparently successful need to be clearly emphasized.


1. Johnson S, Nolan F, Pilling S, Sandor A, Hoult J, McKenzie N, et al. Randomised controlled trial of acute mental health care by a crisis resolution team: the north Islington crisis study. BMJ 2005; 331: 599-602.

2. Muijen M, Marks IM, Connolly J, Audini B. Home-based care versus standard hospital-based care with severe mental illness: a randomised controlled trial. BMJ 1992; 304: 749-54.

Competing interests: None declared

Competing interests: None declared

Debasish Basu, Consultant Psychiatrist

Mersey Care NHS Trust, Hesketh Centre, 51-55 Albert Road, Southport, Merseyside PR9 0LT

Click to like:

25 September 2005

The problem we seem to have locally with rapid response or crisis intervention isn't the help they give to the patients they do see, more with those they exclude.

It's all well and good quoting reduced admission rates for those they see but what about the patients that they don't see?

Competing interests: None declared

Competing interests: None declared

Andrew P Moltu, GP

LE19 2DU

Click to like:

The previous study by this team (British Journal of Psychiatry 187: 68-75) showed a twofold difference in violence in the crisis group, not a statistical difference but an interesting one. It would be helpful if the raw data on violence was available for this study beyond the comment 'no significant difference'. Perhaps this could be included on the website version.

Competing interests: None declared

Competing interests: None declared

Neil J Hunt, Consultant Psychiatrist


Click to like:

25 September 2005

We are pleased to see that our paper regarding Crisis Resolution Teams (CRTs) has provoked substantial interest and debate. The following are replies to some of the main points raised in the rapid responses:

a. Reasons for limited differences in satisfaction: We agree that this should be explored further, and have carried out an accompanying qualitative investigation exploring patients’ views about CRTs in more detail. This will be published separately, as will a health economic analysis led by Paul McCrone and a study carried out by SJ and Tanya Nelson regarding burnout, satisfaction and views about work of CRT staff throughout London.

b. Inclusion and exclusion criteria: Neither personality disorder nor substance misuse were exclusion criteria: as table 1 in the full version shows, substantial numbers of participants had these problems.

c. Admission as an outcome measure: Admission is a controversial outcome measure, as there are good arguments for not always regarding it as a poor outcome – sometimes hospitalisation may be the best way of meeting a patient’s needs. However, recent literature on acute admission in the UK suggests that it does not enjoy much popularity among service users (1), and it thus seems very desirable that alternatives should be made available provided that they are at least as safe and clinically effective as hospital. The findings that symptoms improved at least as much in the experimental as the control group and that the difference in admission rates was as great at 6 months as at 8 weeks suggest that delayed admission is unlikely to explain our findings, but this is worth examining over a longer timescale. As the last paragraph of the discussion reports, our findings are also supported by routine data, which indicated a sustained reduction in bed use in the year after the trial.

d. Representativeness of the study setting: Islington is not represntative of the UK as a whole. CRT implementation in this area may in fact be an especially stringent test of the model, as home treatment may be more difficult in a socially isolated population in which substance misuse and poor living circumstances are prevalent. However, evidence regarding the model’s impact in other settings is certainly needed.

e. Definition of a crisis: the target group for the study was patients experiencing a crisis severe enough for hospital admission to be considered. An operational definition of a crisis severe enough for hospital admission to be an appropriate response had been developed in our previous naturalistic study(2), and was used by the CRT in assessing whether patients referred by other mental health services reached this threshold.

f. Crisis houses: The crisis houses in the study are residential services based in the community. They are staffed 24 hours and offer a variety of forms of practical and psychological support, though not formal psychotherapy. Their staff teams do not include psychiatrists. The mixed sex crisis house is staffed mainly by workers without mental health professional qualifications, while some nurses and other professionally qualified staff are employed in the women’s crisis house. Both crisis houses work closely with other local mental health services, especially the CRTs. Published descriptions are available of Drayton Park, the women’s crisis house, and of the population using it and service user views about it (3,4).

g. Blinding: As clinical staff were involved in the recruitment and randomisation process, it was not feasible to prevent them knowing when the trial was in progress. We agree that this may have introduced a bias, although we feel it is unlikely that this can account entirely for the large effect we have found, especially as our impression was that the local CMHTs were also motivated to prevent admission whenever they could, and indeed succeeded in managing more than 40% of crises without initial recourse to admission. Given the various methodological limitations of our trial, other forms of evidence should also be considered. In particular, we suggest this paper be read in conjunction with the paper reporting the quasi-experimental study we conducted in a neighbouring area of Islington (2), which uses a more naturalistic design and also suggests a reduction in voluntary admissions.

h. The role of CMHTs in home treatment: The CMHTs in our study were well established, integrated health and social care staff and often carried out home visits. At the time of the study, they had no very severe staffing problems and caseloads were below 25 per worker. Thus they were probably representative of reasonably well established and resourced UK CMHTs. However, like most CMHTs, they were available only Monday to Friday 9am to 5pm and usually did not visit more than once weekly, though slightly more intensive contact could sometimes be maintained for short periods in crises. Thus they did not have the capacity to offer out of hours response to crises or acute home treatment of any intensity. We do not know whether an augmented CMHT, with a capacity for at least daily visiting and extended hours availability, could successfully deliver intensive home treatment and achieve outcomes similar to the CRT in our study. Such augmented CMHTs are not currently the norm in the UK, but have been described in other countries – an example is Alan Rosen’s North Shore service in Sydney, which has now been operating for more than two decades (5). Given our study findings and the strong advocacy of organisations that represent service users for 24 hour availability of a response to crises in the community, it seems to us difficult to justify a return to community services which are mainly available 9 to 5, close their doors at weekends and can visit only once or twice a week in a crisis. If this is accepted, a central question for future research and service development is whether CRTs are normally the best way of delivering intensive home treatment and a 24 hour response to crises or whether augmented CMHTs may be able to do so as or even more effectively. Augmented CMHTs offer potential advantages in terms of continuity of care, but may find it more difficult to sustain a focus on intensive community management of emergencies and to organise and staff extended hours rotas (some collaboration between CMHTs is likely to be needed to achieve 24 hour cover). It may also turn out that one size does not fit all: for example, rural catchment areas may be best served by augmented CMHTs with a capacity for intensive home treatment while CRTs may be more appropriate for metropolitan areas.

(1) Rose, D. Users' Voices: the perspectives of mental health service users on community and hospital care. London: The Sainsbury Centre.

(2) Johnson, S., Nolan, F., Hoult, J. et al. (2005) Outcomes of crises before and after introduction of a crisis resolution team. Br J Psychiatry 2005;187: 68-75.

(3) Johnson, S., Bingham, C., Billings, J. et al. (2004) Women’s experiences of admission to a crisis house and to acute hospital wards: a qualitative study. Journal of Mental Health 13: 247-262

(4) Killaspy, H., Dalton, J., McNicholas, S. et al. (2000) Drayton Park: an alternative to admission for women in acute mental health crisis. Psychiatric Bulletin 24: 101-104

(5) Rosen, A., Diamond, R. J., Miller, V., and Stein, L. I. (1997) Becoming real: from model programs to implemented services. New Directions for Mental Health Services, 74: 27-41

Competing interests: None declared

Competing interests: None declared

Sonia Johnson, Senior Lecturer

Fiona Nolan

University College London W1W 7EY

Click to like: