Self-referral admissions

The liaison meetings appear to have been quite successful and meetings have continued with most of the practices since the pilot project ended. The time and usefulness ratings suggest that the balance of topics was generally about right, though two of the practices seem to have been less satisfied (a possible explanation is that these two groups were concerned about the recent loss of an out-patient psychiatric clinic which used to be held in their town). The most popular aspect of the meetings seemed to be the chance to forge closer relationships between GPs and the psychiatric team. Of the more specific topics, discussion of assessment and psychological management seemed to be the most valued. Nearly three-quarters of the GPs said that they would continue attending meetings in their present form. Long-term continuation of this scheme is likely to necessitate some changes, as the importance of GPs and the psychiatric team getting to know each other will diminish as relationships are strengthened. Options identified either by ourselves or by GPs include, first, that (as several GPs felt) the meetings could be more structured; and second, that GPs and psychiatric team members might work together on particular cases, either by seeing patients for joint assessments, as in the 'tripartite' model of Mitchell

"the legally responsible person" and that it would be easier to discuss service deficiencies if he were there.
Sixteen GPs said they would continue to attend meetings in the present format, while six said they would continue only if the meetings changed in vari ous ways. The most common requested were more frequent attendance of the consultant, and more structured meetings, with more preparation of cases beforehand.

Comments
The liaison meetings appear to have been quite successful and meetings have continued with most of the practices since the pilot project ended. The time and usefulness ratings suggest that the balance of topics was generally about right, though two of the practices seem to have been less satisfied (a possible explanation is that these two groups were concerned about the recent loss of an out-patient psychiatric clinic which used to be held in their town).
The most popular aspect of the meetings seemed to be the chance to forge closer relationships between GPs and the psychiatric team. Of the more specific topics, discussion of assessment and psychological management seemed to be the most valued. Nearly three-quarters of the GPs said that they would continue attending meetings in their present form.
Long-term continuation of this scheme is likely to necessitate some changes, as the importance of GPs and the psychiatric team getting to know each other will diminish as relationships are strengthened. Options identified either by ourselves or by GPs include, first, that (as several GPs felt) the meetings could be more structured; and second, that GPs and psychiatric team members might work together on particular cases, either by seeing patients for joint assessments, as in the 'tripartite' model of Mitchell (1979), or by interested GPs having more intensive training or supervision in psychological management of their cases.

A cknowledgemen ts
We would like to thank the GPs in West Oxfordshire who took part in this project, all the other members of the visiting psychiatric team (Anita Damle, Catherine Gardiner, Roy Overall, Rob Poole, Bill Trotter and Jane Webb), and Joan Kirk for her comments on an earlier version of the manuscript.  Little is known about those patients who bypass their GP and self-refer directly to hospital. There have been studies of referrals to emergency clinics (Lim, 1983;Haw et al, 1987;Kehoe & Newton, 1990) and to community centres (Hutton, 1983;Boardman & Bouras, 1989). All these studies found that more men self-referred than women. Hutton (1985), Boardman & Bouras (1989) and Kehoe & Newton (1990) found that self-referrals often had a non-psychiatric precipitant and rarely required acute psychiatric intervention.
The process of self-referral to hospital differs from self-referral to community units or emergency clinics in that patients are usually discouraged from the former but encouraged in the latter.
We conducted the current study to establish the features of patients who self-refer to psychiatric hos pital and to identify any differences between such patients and those following conventional referral routes.

The study
The study was carried out at Highcroft Hospital, Birmingham, which serves 466,000 people and is situ ated in its catchment area. It was conducted over the six months commencing 1 August 1989. After each duty period, one of us interviewed the duty doctor and any self-referred admissions in order to attain demographic and clinical data. We then compared the demographic and clinical characteristics of self-referral admissions with a control group of admissions referred by the GP.

Findings
The total number of acute admissions over the six months was 429; 14.7% were formal. Nineteen selfreferred hospital admissions were identified, includ ing one patient who self-referred twice; 34 patients self-referred but were not admitted. Only limited data could be collected on these: 26 male, 8 female; mean age 39 years, range 17-89 years. Eighteen other patients (9 male, 9 female) telephoned the duty doctor directly requesting admission. They were advised to contact their GP.

Comparison of self-referral with GP referral admissions
There were more males in the self-as compared with the GP-referral group (58% v. 42%). The mean ages were similar (38 v. 36 years) as were the percentages that lived alone (47%) and were single (26%). The mean distance of home from the hospital was similar for the two groups (2.4 v. 2.6 miles).
The average length of stay was approximately 20 days for both groups. Very few of either group (5% of the self-v. 16% of the GP-referrals) were formally detained. Similar percentages were discharged with medical approval (68% v. 58%) and approximately 70% of both groups were offered follow-up. Similar percentages (68% v. 53%) had been receiving psychiatric care before admission.
Of the self-referrals, 79% had previous psychiatric admissions as compared with only 47% of the GP-referrals. This finding approaches statistical significance (0.05 < P < 0.1).
Of the self-referral group, 31% had a case notes diagnosis of a primary mood disturbance and 42% of a non-organic psychosis; 27% had other diagnoses. The situation was reversed in the GP-referral group with 42% having a case notes diagnosis of a primary mood disturbance; 26% of a non-organic psychosis, and 32% having other diagnoses.
A rather disturbing finding was that four of the GP group have since died; two by natural causes and two by suicide.

Comment
The increased number with previous admissions in the self-referral group was the only difference that approached statistical significance. There were trends towards an increased diagnosis of non-affective psychosis in the self-referring group and an increased diagnosis of mood disturbance in the GP-referral group. An explanation could be the chronic course that non-affective psychoses often run, making repeated contact with the hospital more likely. The increased death rate in the GP-referral group may reflect increased severity of illness in that group.
Some of the similarities between the two groups were also interesting. It was surprising that distance from the hospital was so similar and that there were no major differences in length of stay or the percent ages offered psychiatric follow-up and discharged with approval.

Conclusion
Our study (albeit limited by small sample size) suggests that patients admitted by way of the selfreferral route are broadly similar to those admitted by the conventional GP-referral route. However, we found trends towards a higher proportion of males, an increased diagnosis of non-affective psychosis, and a greater number of previous admissions in the self-referral group. A larger study is needed to investigate further the significance of these trends.