Deliberate self-harm, of which self-poisoning is the major component, is estimated to cost general hospital budgets between £45 and £50 million per year (NHS Centre for Reviews and Dissemination, 1998). Deliberate self-poisoning itself accounts for approximately 150 000 emergency admissions annually into general hospitals in the UK (NHS Centre for Review and Dissemination, 1998).

It is recommended that individuals who have deliberately self-poisoned should receive a comprehensive specialist psychosocial assessment (Central and Scottish Health Service Councils, 1968; Department of Health and Social Security, 1984; Royal College of Psychiatrists, 1994). The Department of Health and Social Security (1984) guidelines recommended that along with doctors, other professionals could become involved in this work, a recommendation that was also supported by Johnson and Thornicroft (Reference Johnson, Thornicroft, Jenkins, Griffiths and White1994). Despite such recommendations and evidence that nurses can perform as well as doctors in the assessment of self-poisoned patients (Reference Catalan, Marsack and HawtonCatalan et al, 1980), the introduction of nurse-led assessment services has been slow to take off.

The six-bedded Cardiff poisons unit serves a population of approximately 410 000. Individuals aged 14 years or more who have self-poisoned and require in-patient treatment are admitted to the unit. Details of those admitted appear to be representative of other self-poisoning populations within the UK and have been reported elsewhere (Reference Bialas, Reid and BeckBialas et al, 1996; Reference Scorer, Beck and BissonScorer et al, 1999). Traditionally the psychosocial assessment has been conducted on the day after the self-poisoning episode by the duty liaison psychiatrist (a senior house officer in psychiatry) with supervision provided as required from the on-call senior registrar or consultant psychiatrist.

The duty liaison psychiatrist works within an on-call rotation system and may visit the poisons unit once every 10-12 days. This has the potential for a lack of continuity for both patients and staff at the unit. There may also be a variation in the style and standard of assessment when a different doctor attends every day. The system also falls short of that recommended by the Royal College of Psychiatrists (1994) and has been criticised as being an inappropriate training experience by a Royal College of Psychiatrists approval visit.

In order to address some of these problems we were keen to assess the feasibility of replacing the doctor on-call system with a nurse-led assessment service. The aim of this study was to compare psychosocial assessments of individuals admitted as a result of deliberate self-poisoning carried out by a psychiatric nurse with those carried out by a duty liaison psychiatrist.

The study

During a short-term placement within the liaison psychiatry service, an experienced qualified psychiatric nurse (G.G.) was attached to the poisons unit to carry out psychosocial assessments alongside the duty liaison psychiatrist. Before conducting patient assessments independently, the nurse received training and supervision as recommended by the Royal College of Psychiatrists (1994). This involved an introduction to the principles of assessment and the standardised measures used in Cardiff. The nurse then observed an experienced psychiatrist assess five individuals following an overdose before assessing five individuals herself in the company of an experienced psychiatrist. Issues about the assessments were discussed. Thereafter the nurse and the duty liaison psychiatrist attended the poisons unit daily and divided the assessment work between them. There were no criteria applied for allocating each patient to either the nurse or the doctor, which was done on a first come first served basis.

The patient interviews were conducted with the aid of a form specially designed for assessment of deliberately self-poisoned patients in Cardiff (copies of which are available from the authors upon request). The duty consultant psychiatrist for the day was available for supervision and the nurse received ongoing weekly supervision from the consultant liaison psychiatrist. After each assessment the nurse and doctor would implement a management plan that was recorded along with basic demographic information, details of the overdose and the assessment.

Findings

One hundred and forty-five patients were recruited to the study. Their background variables are shown in Table 1. There were significantly more female and unemployed patients assessed by a doctor than by the nurse. Of those with a previous history of overdose, nine had two previous admissions, two had three and one had four. Variables concerned with the index overdose are shown in Table 2. The only significant difference between those assessed by a doctor and those assessed by the nurse was that more of the nurse-assessed patients had taken alcohol at the time of their overdose. The management of patients appeared to be consistent.

Table 1. Comparison of background variables between doctor and nurse assessed patients

Variable	Doctor mean (n=68)	Nurse mean (n=77)	Mean diff/odds ratio (95% Cl)	P value
Age	31.8 (12.97 s.d.)	31.7 (11.9 s.d.)	0.096 (-4.01, 4.20)	0.96
Gender: female	46 (68%)	33 (43%)	2.75 (5.84, 1.34)	0.0051
Previous overdose	40 (59%)	47 (61%)	1.10 (2.25, 0.53)	0.92
Other deliberate self-harm	18 (26%)	29 (38%)	1.66 (3.63, 0.78)	0.21
Current psych	18 (26%)	31 (40%)	1.85 (4.03, 0.87)	0.12
Unemployed	61 (90%)	53 (69%)	3.76 (11.02, 1.46)	0.0041
Married	13 (19%)	17 (22%)	1.19 (2.91, 0.50)	0.82
Separated/divorced	22 (32%)	23 (30%)	1.12 (2.41, 0.52)	0.89
Single	32 (47%)	36 (47%)	1.01 (2.05, 0.50)	1.00
Past psychiatric history	22 (32%)	26 (34%)	1.06 (0.47, 1.88)	1.00

Table 2. Comparison of variables concerned with the index self-poisoning between doctor and nurse assessed patients

Variable	Doctors mean (n=68)	Nurses mean (n=77)	Mean diff/odds ratio (95% Cl)	P value
Beck inv total	8.9 (6.3 s.d.)	9.1 (6.7 s.d.)	0.182 (-2.26, 2.63)	0.88
Alcohol involved	22 (32%)	43 (56%)	2.61 (5.53, 1.27)	0.0081
Psychiatric diagnosis	56 (82%)	66 (86%)	1.28 (3.43, 0.48)	0.75
Psychosis	3 (4%)	5 (6%)	1.42 (8.32, 0.30)	0.85
Consultant advice	27 (40%)	24 (31%)	1.45 (0.33, 1.44)	0.37
In-patient disposal	9 (13%)	8 (10%)	1.31 (0.25, 2.32)	0.78
CMHT referral	38 (56%)	51 (66%)	1.54 (3.21, 0.75)	0.27
CAU referral	8 (12%)	14 (18%)	1.63 (4.72, 0.60)	0.40
General practitioner follow-up	21 (31%)	18 (23%)	1.46 (0.31, 1.52)	0.41

Discussion

This study provides further evidence that psychosocial assessments following self-poisoning can be provided by appropriately trained and supervised mental health nurses. There appeared to be no significant differences in the outcome of assessments performed by a duty liaison psychiatrist and those by a mental health nurse. There were limitations to this study, the most important of which were the absence of random allocation of patients to doctor or nurse assessment and no independent measure of the quality of the assessments. In addition there were minor differences between the patients assessed by a doctor and those by the nurse, but these seem unlikely to have made a significant impact on the outcome of the assessments. Possible explanations for the difference in alcohol involvement are chance, or that G.G. was better at eliciting an alcohol history than the duty liaison psychiatrists. Strengths of the study include an adequate sample size and a sample that reflects actual clinical practice in Cardiff, which is likely to be representative of other areas.

We believe that there are significant advantages to a well-trained, well-supervised mental health nurse-led assessment service. The service was very popular among the poisons unit staff. The presence of a named professional with specific responsibility allowed her to build up a specific expertise in this area and to develop a positive working relationship with the local poisons unit. It appeared to be well received by patients and by junior psychiatrists, who found the duty liaison psychiatrist role more manageable as a result. It is hoped that this initiative will lead to improved working protocols and result in ongoing training for non-psychiatric nursing staff involved in the care of patients following deliberate self-harm. This is particularly important given the large proportion of patients who do not receive a formal psychosocial assessment following deliberate self-harm. This group has been identified in previous research as being at particular risk of repetition (Reference Crawford and WesselyCrawford & Wessely, 1998).

One potential criticism of a nurse-led assessment service is that it could prevent junior psychiatrists receiving adequate training and experience in the assessment of individuals who have self-poisoned. We believe the contrary to be the case. Most services, including Cardiff, have traditionally relied on junior doctors to perform self-poisoning assessments with limited training and supervision. A key function of the nurse-led assessment service attached to the liaison psychiatry service should be the development of a comprehensive training and supervision programme for trainee psychiatrists. This would ensure that they gain experience in a planned way in the assessment of individuals who have self-poisoned, initially accompanied by a mental health professional who specialises in this work with input from the consultant liaison psychiatrist. This is in line with the recommendations of the Royal College of Psychiatrists (1994), allowing a move away from a duty that often becomes a service duty rather than a useful training experience.