Who should refer to psychotherapy

The referral rate to the Psychotherapy Department at the Warneford Hospital, Oxford, has increased by 84% in the last three years, creating a bottleneck at the assessment stage. As part of the response to this wecarried out the audit reported here. Assessment for psychotherapy is costly in professional time. If any group of referrers is significantly better at referring patients who need, and make use of psychotherapy, we would have to consider selectively accepting referrals from these sources. We thus asked whether different referrer groups, and different "levels" of

The referral rate to the Psychotherapy Department at the Warneford Hospital, Oxford, has increased by 84% in the last three years, creating a bottleneck at the assessment stage. As part of the response to this wecarried out the audit reported here. Assessment for psychotherapy is costly in professional time. If any group of referrers is significantly better at referring patients who need, and make use of psychotherapy, we would have to consider selectively accepting referrals from these sources. We thus asked whether different referrer groups, and different "levels" of referral (secondary or tertiary), were more or less likely to provide patients who ultimately entered treatment.
Psychotherapy's organisational standing in the wider health service has, by tradition, been flexible. In our direct work with patients, we are sometimes viewed as super-(or sub-)specialists, taking tertiary referrals from other mental health professionals. At the same time, we deal directly with GPs, and psycho therapists have often fostered productive direct contacts with primary care. This raises an important question. Is psycho therapy a sub-specialty of psychiatry, or do we pro vide an alternative perspective on psychological health that should be available directly to the pro vider of primary care? In concrete terms, this is the question "who should refer to psychotherapy?". Any answer provided by the structure of a mental health service will have profound implications for the role of the psychotherapist. The referrer can be seen as a "gate-keeper" for our service. If our gate-keepers are limited to one homogenous group, in the current intellectual climate within psychiatry they may have a very different model of mental dis-ease. Thus, the scope and application of the specialty would be determined by a "gate-keeper" philosophically at odds with the service itself.
Often such complex matters are represented as apparently simple questions of procedure and econ omics. In the "purchaser/provider" environment, the question is one of how we liaise with "pur chasers" of our service: indeed, who will these pur chasers be ? Will we look towards general practice, defining our services as related to the needs of popu lations and their primary carers? Alternatively, on organisational grounds, and in the hope of a more efficient use of scarce resources, should we only take referrals once people have been seen, and selected, by general psychiatrists?

The study
A retrospective case-note study was made of a cohort of patients referred for psychotherapy in 1989. This cohort was selected since all planned treatments were well established or completed by the time of the study.
For each we mapped their "pathway" to the psychotherapy department, using a modification of the 'encounter form' used in the WHO 'Pathways to Care' Study (GÃ¢ter& Goldberg, 1991). We then recorded whether or not an assessment was offered, and the outcome, and whether or not the patient subsequently entered treatment. No distinction was made between types of treatment. Referrals came from all doctors, mental health professionals, and a small number of other health workers and agencies. There was no routine "feed back" or information supplied to any group of referrers, although we are now developing this aspect of our work.
Acceptance of referrals was decided on available clinical information. If a patient clearly had complex psychopathology difficult to treat within the avail able resources, a referral would be declined after dis cussion with the referrer. Referrals were also passed elsewhere within the mental health unit if obviously more appropriate.
Assessment was by a consultant psychotherapist, or senior registrar under consultant supervision: assessors were familiar with a broad range of psychopathology, and psychological and psychiatric treat ment, so were able to recognise patients best treated elsewhere. The department itself offers individual and group psychodynamic psychotherapy, brief focused therapies (including CAT), and systemic family and couple work. There is also a special programme for victims of childhood sexual abuse.
A "successful" referral was judged to be someone who received, and completed, treatment. This is a limited criterion, ignoring important "consultation" and therapeutic aspects of psychotherapy assess ment. However, the provision of treatment was the explicit concern in over 80% of the referral letters. For analysis, outcome of referral was recorded on the two dimensions "assessed or not", and "treated or not". Numbers were large enough to allow valid conclusions using the x2 test.
Case-note information was subsequently linked to the records of a clinical and demographic question naire completed routinely on all patients assessed in the department since 1988, as part of the Oxford Psychotherapy Audit Project. This allowed some comparisons of the clinical features of patients referred from different sources.

Findings
Between 6 January and 22 December 1989 147 indi viduals were referred to the department. Of these, 114 were assessed, and subsequently 66 patients started treatment.
The three major groups of referrers were GPs, gen eral psychiatrists, and NHS psychologists, working in their usual organisational settings. Referrals from these professionals working in other settings, other types of professional, and self-referrals, were con sidered as a single group, to provide a large enough sample for statistical comparisons, with the reser vations that generalisations about this group as a whole should be cautious.
After referral, "filters" (Goldberg & Huxley, 1980) occur before assessment and before starting treat ment. Patients not passing through these filters include both those where there was an active decision not to provide care in the department, and those who in some way withdrew themselves. Thus the two dimensions, "assessed or not" and "treated or not" reflect the proportion of patients in each group that pass through each filter.

Referrer groups and outcome of referral
There were no statistically significant (/><0.05) differences in outcome of referrals from the four groups. The number of GP referrals where no assess ment was offered was 9 out of 56 patients, compared with 2 out of 53 referrals by general psychiatrists. Six of these were patients with phobic symptoms referred on for behavioural treatment: a simple intervention, which was economical of resources, and provided an opportunity to educate the referrer. Of the patients actually assessed, 23 of 40 (57.5%) GP referrals started treatment, compared with 26 of 45 (57.7%) general psychiatry referrals, 4 of 6 (67%) psychology referrals, and 11 of 20 (55%) of referrals from the 353 "other" group: the similarity of the percentages was striking.
"Pathway" to care and outcome of referral The number of filters passed through before referral might also affect subsequent uptake of treatment, by providing a more highly selected population. The data were re-examined in terms of the number of stages (i.e. referrals from one professional to the next) in the pathway to the department. Of 147 referrals, 66 had seen one professional only; there were 61 two-stage referrals, 16 three stage referrals, and 4 had seen four professionals. The bulk of one-stage referrals are from general practitioners: this group (49 patients) was distinguished from one-stage referrals from other sources (17 patients).
The two-stage referrals included 39 where the second professional, who made the actual referral to psychotherapy, was a general adult psychiatrist: pro portions of referrals from this subgroup passing each filter were not significantly different from the two stage referrals as a whole, indicating that psy chiatrists were not significantly better (on the limited criteria used) at selecting patients than any other group.
The only significant (P<0.05) finding was the failure to take up treatment of the 17 patients who only saw one person, not a GP, before referral. After assessment only 23.1% of these patients took up treatment, compared with a range of 50-73% for the other groups.
These non-GP single stage referrals were a hetero geneous group. Four were self-referrals by patients, accepted because of previous contact with the depart ment. Two of these completed a therapy, one defaulted after assessment, and for one it was thought that further treatment would be unhelpful. Three were primary referrals from outside agencies. Of these, one did not come for assessment, and the other two were assessed, but not offered therapy. The remainder (10 patients) came from within the mental health unit, with the shortened path due to departures from usual procedure. It thus appears probable that the rate of failure to take up treatment reflects dynamic issues related to the particular group of patients, and their unconventional use of services, rather than direct characteristics of the referrer.

Clinical features
The clinical and demographic questionnaire contains items relating to past and current difficulties. Comparing the incidence of these between the refer rer groups, no significant differences were found. Overall, the patients who actually received treat ments had more difficulties than the group of referrals as a whole.

Comment and conclusions
There was no evidence that restricting the right of referral to any professional group would signifi cantly increase the efficient use of resources in our assessment procedure. Once the decision was made to offer an assessment, referrers appeared equally able to select patients in need of, and likely to benefit from, the psychotherapies we offer. The pathway data did, however, confirm the clinical impression that accepting patients through "unconventional" channels was likely to be wasteful of resources.
Given this finding, in terms of overall service planning, there are important considerations. Long "pathways" use expensive professional resources, so without clear evidence of other benefits, extra "steps" must not be introduced. The cost to the patient of repeated, often unavoidably distressing, assessment by unfamiliar professionals, who will not be seen again, must also be remembered.
The study emphasised the role of the GP as primary physician, making appropriate referrals to our ser vice despite lack of detailed technical psychiatric or psychotherapeutic knowledge. The skill of this group of doctors at recognising and dealing with psycho logical problems, often based on intimate knowl edge of the patients and their families, must not be underestimated.
Throughout the NHS the present reforms are emphasising the inextricable link between ideological issues and the debate about efficient use of resources. As noted above, limiting the sources from which referrals are accepted, would have profound effects on the service, not least in terms of the debate within the profession between the concretely "scientific" and psychodynamic viewpoints. Patient choice could be reduced, except for those who could demand it from the system (which many cannot), or buy private care. Alternatively, psychotherapy could burgeon within other professional groups, confining the psychiatrically trained psychotherapist to a circumscribed role.

Maloney
If psychotherapists do not actively provide new models for their services, they will be provided from outside: the "no change" option is not easily avail able. Yet, in this instance, a detailed audit project using considerable resources has shown no case for change. We must proceed carefully, and only alter our service if a need can be clearly established. Change must neither be implemented nor resisted on the basis of assertion alone.
Psychotherapists have to take the lead in planning the changes to their services, since they understand them best. There can no longer be the concentration on the needs of the individual alone, which is still possible in private practice. NHS psychotherapists are responsible for whole populations, and for pro viding what the people who make up these popu lations need and want. In the new service this issue will have to be addressed at the purchaser/provider interface. To negotiate in these new settings, we need models for rigorously organised, efficient, auditable psychotherapy services within a state health service. We need these models now.