Twenty-eight years of liaison psychiatry at a general hospital

Aims and method A questionnaire was used to measure physician satisfaction with the unit and comparison of referral numbers and diagnoses with 1967, when the unit was new, and in 1985 when it was re-audited. Results Despite physicians9 satisfaction with the service, the referral rate remains unchanged. Clinical implications Liaison psychiatry should not concern itself with trying to increase referral rates. Research into patient outcome should be a priority and guide its future direction.

The liaison psychiatry unit to be described in this study was one of the first in Scotland. It was audited in its early days by Mcleod & Walton (1969), and again 18 years later by Brown & Waterhouse (1987). Despite the enthusiasm of the initial review, Brown & Waterhouse showed that in some respects the unit had failed to live up to its earlier promise. In particular they were concerned that despite the fact that a large proportion of patients in a general hospital setting have been shown to have psychological morbidity (Shepherd et al, 1960;Maguire et al, 1974), only a small proportion were referred for psychiatric assessment. They, like many others, considered that the fault lay with psychiatrists' style of relating to their colleagues. They con clude quoting Mason (1975) "that psychiatrists are seen as 'not readily available, remote in thought and tending to express opinions in a style which alienates physicians'". Ten years on we describe a further audit of liaison psychiatry at the Western General Hospital, Edinburgh.
The aims of the present study are to assess by questionnaire the views of the hospital's con sultant staff on the quality of service provided by the department; to review referral numbers and diagnoses; and to compare these to the rates described in 1967 and 1985.

The study Liaison psychiatry in the Western General Hospital
The study took place in the Department of Psychological Medicine of the Western General Hospital in Edinburgh. The Western General is a 600-bed teaching hospital in which most medical specialities are represented. It has been subject to much internal change over the past eight years as part of ongoing rationalisation of acute services within Lothian.
The Accident and Emergency Department, which took a third of emergencies from the city, became unsustainable after the closure of orthopaedic, paediatric, gynaecological and obstetric services and it closed in 1991. A Geriatric Assessment Unit was opened in 1990.
The Breast Unit from the, now closed, Longmore Hospital moved to the Western General Hospital in 1991. Haematology services have also been centralised at the Western General. These moves, along with existing oncology beds, have led to the forming of a regional centre. Respira tory medicine and rheumatology moved from the Northern General Hospital to the Western Gen eral Hospital in 1991 following the Northern's closure. The Department of Clinical Neuro science has expanded to become a 72-bedded regional centre. Renal medicine moved to the Royal Infirmary in 1995 (Eastwood &Jenkinson, 1995).
The liaison psychiatry team, like the hospital, has been subject to much change. It currently consists of two consultants (working three and two sessions per week), one part-time senior registrar, one registrar, one senior house officer (usually a general practice trainee) and secre tarial staff. One of the consultants is specifi cally dedicated to the department of clinical neuroscience. The psychiatric trainees both spend two sessions per week seeing cases of deliberate self-harm at the Royal Infirmary of Edinburgh.
The work of the unit consists of seeing in-and out-patients referred from the rest of the hospi tal, and participating in liaison meetings with specific units. The unit does not deal with cases of deliberate self-harm as they are all managed at a regional self-harm unit based at the Royal Infirmary of Edinburgh. In addition, a general psychiatrist from the Royal Edinburgh Hospital runs an out-patient clinic once a week in the department, which trainees participate in. The unit has no access to beds in the general hospital, but can negotiate admissions to the Royal Edinburgh Hospital when necessary.
A clinical psychologist works closely with the department and there is an established link with the hospital social work department.

Data collection
A brief questionnaire was designed and distrib uted to all consultant clinicians in the hospital for anonymous completion. The questionnaire concentrated on accessibility of the service, helpfulness and clarity of communication of diagnosis and management. It used a combin ation of Likert and semantic differential scales.
Records of all in-patients referred to the department over a six-month period (April-September 1995) were examined. Where the psychiatric notes were unavailable general med ical records were consulted. Data were gathered on the number of referrals, the reason for referral, psychiatric diagnosis, medical diagnosis and disposal. The results were compared with the two previous studies of the unit's activity.
The previous studies differ slightly from each other. The 1974 data includes a small number of out-patients whereas the 1985 data excludes out-patients. They both use a similar April-September time-period.
Psychiatric diagnoses were categorised accord ing to ICD-9 (World Health Organization, 1978). Although this has now been superseded by ICD-10, ICD-9 was in use during the study period and was also used for classification in the 1974 and 1985 studies.

Findings
Of the 70 consultant clinicians in the hospital, six were out of the country at the time of the survey. Of the 64 who were sent questionnaires, 54 replied. Seven responders had not used the service so were unable to comment on it. Reasons given for this included being new in post and not feeling the need for referral. The following results report the results of the 47 consultants who completed questionnaires.
In assessment of ease of access of the service (scale 1 (low)-4 (high))the mean rating was 3.38. On impression of psychiatrist's helpfulness (scale 1 (low)-4 (high)) the mean rating was 3.44. Promptness of response had a mean rating of 3.45 (scale 1 (low)-4 (high)).
There was overall satisfaction with clarity of communication when rated on a five-point semantic differential scale (5 (positive)-1 (nega tive)).The mean ratings are described in Table 1.
During the six-month study period, 92 pa tients were referred to the department. This is a referral rate of 1.1% of patients admitted to the hospital compared with 1.0% in 1985. Table 2 shows the referral rates for 1967, 1985 and 1995. Data were available on all patients who had been referred. The stated reason for referral is shown in Table 3. The majority of patients continue to be referred for assessment. Table 4 shows the psychiatric and medical diagnoses of patients referred. There is an increase in the number of patients diagnosed   1967 1985 1995 1967 1985 1995 1967 1985 1995 1967 1985 1995 1967 1985 1995 39 49   as neurotic and endogenously depressed. There continues to be a decline In the number of patients diagnosed as personality disordered. The number of patients diagnosed with organic states has also fallen. Schizophrenia continues to play a negligible part in the work of the department. A large increase in neurological referrals can be noted.

Comment
The questionnaire survey of consultant clini cians showed general satisfaction with the quality of the liaison psychiatry service and particular satisfaction with the psychiatrists' abilities to communicate their opinions to their colleagues. Despite this, there had been no increase in the overall referral rate for psychiatric opinion of general hospital patients.
Some changes in the patterns of referral from different units within the hospital were noted. This may be due to changes within the liaison department.
In particular one of the current consultants is designated to work solely with the Department of Clinical Neuroscience. Spe cific liaison meetings with units tend to increase referral rate and hopefully the quality of psycho logical care by medical staff.
Given the increasing throughput of the hospi tal and the increasing age of the population, the decrease in referrals of patients with organic states is unlikely to be due to a decrease within the hospital. We think that the setting up of weekly liaison meetings with the Geriatric Assessment Unit may have contributed to an increased competence in dealing with acute confusional states (delirium) without the need for further referral.
Overall, hospital referral rates are not increas ing substantially, despite evidence of the integra tion of psychiatry. While it is important that psychiatric morbidity is picked up and treated, it would be wrong to always see this as a need for referral to liaison psychiatry. Indeed if every patient with a mental disorder in the hospital, assuming a prevalence of 25-30% (Mayou & Hawton, 1986), was referred for review this could lead to up to 2000 new referrals per year! A more realistic model for looking at and assessing the role of liaison psychiatry may be to modify Goldberg & Huxley's views on mental disorders in the community (1980,1992) and the niters to specialist psychiatric care. Patients suffering from schizophrenia, affective psycho sis, organic states and substance misuse pro blems are all more likely to be referred for specialist attention. By contrast the work of the liaison psychiatrist lies more with what they describe as the 'common mental disorders', ones which in a primary care setting are much less likely to be referred for specialist help. This model predicts a 1% referral rate, with hospital doctors and primary care dealing with the less severe cases. We would suggest that this is an appropriate referral rate and that there is little to be gained from trying to increase it. This, read in conjunction with physicians' satisfaction, would suggest all was well.
However, the evidence in the literature and clinical impression suggests to us that when a patient is not referred, this usually equates with no treatment. Although there are a few laudable exceptions, there is little to suggest that physi cians are making attempts to treat patients who are psychologically disturbed. Indeed we wonder if physicians, and more importantly the public, are really ready for this from a sociological point of view. We are unaware of any anthropological studies looking at this issue and while we are aware of the growth in complementary therapies, we are far from convinced that people wish anything other than traditional treatment styles when visiting a physician. In an ongoing study by A.J.C, only four of 100 consecutive out-patient attenders in neurology have expressed any desire for a psychological treatment (further details available from the author upon request).
A more critical interpretation of our findings is that liaison psychiatry is failing to demonstrate its potential usefulness to physicians. Therefore, they are content with its role in dealing with conspicuous morbidity and emergency situ ations. This is disappointing when recent stu dies have highlighted the fact that psychiatric morbidity can have more effect on physical outcomes than more commonly examined phy sical parameters (Sullivan et al 1997). Liaison psychiatry must show that it can improve clinical outcome.
We see there being two strategies for achieving this. Most importantly is through shared re search studies. Liaison research is often carried out by psychiatrists who then tell physicians how they should do their job better. This is not always popular. If ownership of the idea, the study and the outcome is communal then physicians may be more likely to follow the recommendations. A good example of this joint practice would be the management of chronic fatigue syndrome by cognitive-behaviour therapy (Sharpe et al, 1996); following the trial of therapy the physi cians' were impressed with the results and found the money to employ a psychiatrist to provide a service.
The second approach is that liaison psychiatry moves into the physicians' out-patient clinic; possibly via the supervision of nurses trained in cognitive-behaviour therapies. These techniques can be simply applied and have been shown to be effective in many chronic conditions of both known and unknown aetiologies. This would prevent layers of abnormal illness beliefs being superimposed on the original disorder during the passage of time from extra investigations 'just to exclude. . . .'to referral to a specialist psychiatry clinic.
Locally, there has been the appointment of a senior lecturer in psychological medicine, which will serve to raise the profile of liaison research. Studies have commenced in collaboration with the departments of neurology and oncology and are planned with molecular medicine.
We have made links with the Cystic Fibrosis Unit, via a trainee psychologist, and palliative medicine via a senior registrar developing a liaison service as a special interest (Mitchell, 1998). There has also been the circulation of a handbook entitled The Psychological Care of Medical Patients.
We should continue to assist our colleagues in their detection and management of mental disorder. Education, especially of junior doctors, on interview techniques, simple basic treatment strategies, guidance on when to refer and the development of treatment protocols for use with in the hospital can all help. The department now participates in MRCP training and in the Scottish Higher Training for Neurologists programmes. A further suggestion is to increase the emphasis on neurotic and depressive disorders at under graduate level of medication education. We suggest that with regards to the detection and treatment of neurotic and depressive conditions, liaison psychiatrists are well placed to take a lead in medical education, at both under and post graduate levels. In 1938, a Lancet editorial deplored "the present system of occasional demonstrations at a distant hospital" and instead advocated the setting up of liaison psychiatry units to assist with medical students' education (Anonymous, 1938). After all these years we think psychiatry should follow this advice.

Discussion
This service was one of the pioneering liaison psychiatry units in this country. In its infancy emphasis was placed on integration with the rest of the hospital and increasing referral rates. Now that the unit has reached maturity it appears to have accomplished the first task with little change on the second. We think that the referral rate has been appropriate all along and the drive to increase numbers has little impact on patients outcome and is not the correct direction for liaison psychiatry to travel. Instead we would advocate an evidence-based approach to improv ing outcome in defined conditions, particularly in out-patients, through shared links with specific clinics.