The study

A national survey found that 31% of general practices employed a counsellor (Reference Sibbald, Addington-Hall and BrennemanSibbald et al, 1993), but there have been few evaluations of their impact on patient management and outcome. Fletcher et al (Reference Fletcher, Fahey and McWilliam1995) found less prescribing of psychotropic medication among practices that employed counsellors. This finding was supported by another study, which also concluded that patients receiving counselling were less likely to be referred on to secondary mental health services (Reference Boot, Gillies and FenelonBoot et al, 1994). A comparison of referral rates has shown an increase in referrals to clinical psychology, but not to psychiatric out-patient clinics (Reference Cape and ParhamCape & Parham, 1998) from practices that employed a counsellor compared with those that did not. There have been no studies of the impact of general practice counsellors on referrals to community mental health teams (CMHTs).

Our study aimed to demonstrate whether the presence of a practice counsellor was associated with a difference in referral rate and whether the appropriateness of referrals was affected.

All general practices within Merton, Sutton and Wandsworth Health Authority were sent a structured questionnaire seeking information about the practice (list size, number of partners, presence of partner with special interest in psychiatry, presence of a practice counsellor) and the partner with most interest in mental health was asked to complete the questionnaire. We used the definition of a practice counsellor adopted by Sibbald et al (Reference Sibbald, Addington-Hall and Brenneman1993):

The definition excluded support provided by staff as part of their routine work.

Referrals received by the 14 CMHTs operating within the health authority were collected over a 1-year period and a random sample of case notes from each CMHT was reviewed by an experienced clinical researcher (N.B.) to obtain demographic details and diagnoses. An assessment of caseness was made as a measure of the appropriateness of the referral; this was categorised as case (experiencing a mental or psychological disorder), non-case (no evidence of a mental or psychological disorder) or borderline case (where the evidence for the presence of a mental disorder was uncertain).

Findings

Assessment of caseness

The distribution of caseness of referrals is shown in Table 1. There was no significant difference between the two groups in the rates of caseness of the referrals, with fewer than 10% in both groups rated as non-case.

Table 1. Assessment of caseness of referrals

	Referrals from practices employing a counsellor (n=76) (%)	Referrals from practices not employing a counsellor (n=104) (%)
Case	60 (78.9)	78 (75.0)
Borderline case	10 (13.1)	18 (17.3)
Non-case	6 (7.9)	8 (7.7)
Total	76 (100)	104 (100)

Referral rates

The mean rate of referral from practices that employed a counsellor was over double that of the ones that did not. The distribution of referrals between practices was skewed, but analysis using non-parametric tests showed a significant difference (P=0.003).

Table 2. Referral rates from practices employing and not employing a counsellor (per 1000 patients on practice list)

	Practices employing a counsellor	Practices not employing a counsellor
Mean referral rate/1000	0.516	0.204
Median referral rate/1000	0.556	0.000
25th centile	0.000	0.000
75th centile	0.867	0.287

Comments

The only differences revealed by our study between the two groups of practices were higher rates of fund-holding and presence of partners with a special interest in psychiatry among practices that employed a counsellor. These factors are unlikely to be independent and probably reflect the ability of fund-holding practices to direct resources towards the priorities of partners; partners with a special interest in mental health would be more likely to see the employment of a practice counsellor as a priority area.

Our data showed a lower rate of employment of practice counsellors than data reported by Sibbald et al (Reference Sibbald, Addington-Hall and Brenneman1993) (20.3% v. 31%). We looked at practices within an urban/suburban setting only, whereas the previous survey was nationwide. In urban areas, which tend to have higher levels of deprivation and higher rates of severe mental illness, the provision of counselling for more minor psychological problems may be given a lower priority by general practitioners (GPs).

Our data showed that the presence of a counsellor was associated with an increased rate of referral to mental health services, which is in contrast to the commonly held assumption that it should lead to a reduction. This will have significant clinical implications for the workload of CMHTs, with more time being spent on assessment of new referrals than working with the long term case-load of people with severe mental illness.

Levels of caseness of referrals were not affected by the presence of a counsellor. It would appear, therefore, that the sensitivity of detection of mental distress is increased in practices that employ a counsellor, without causing a reduction in the specificity, indicating that substantial psychological morbidity remains undetected in practices that do not employ a counsellor. This is consistent with the work of Johnstone & Goldberg (Reference Johnstone and Goldberg1976) on the detection of morbidity in primary care. Training in interview techniques has been shown to improve GPs' detection of depression (Reference Goldberg and HuxleyGoldberg & Huxley, 1992), and many cases of depression and short term mental health problems should be managed in primary care.

Our study covers an area with well-developed CMHTs (where psychologists work as integral members of the multi-disciplinary team) and shows that, in this setting, the presence of practice counsellors is associated with an increase in the rates of referral to the secondary mental health service as a whole. This is in contrast to the findings of Cape & Parham (Reference Cape and Parham1998), who demonstrated an increase in direct psychology referral rates only from practices that employed counsellors, but in a less integrated mental health system.

There are limitations to this study, particularly in the assessment of referrals' caseness. This was achieved entirely by a review of case notes and not by interview with the patient or consultation with the CMHTs. The three categories used in the assessment of caseness do not allow for grading of the severity of mental health problems, only the likely presence of a disorder. The sample is restricted to routine written referrals and does not include urgent telephone referrals, but many of the CMHTs follow a model of home-based assessment (Reference Burns, Beadsmoore and BhatBurns et al, 1993) and have a rapid response to written requests.

This study has shown a higher rate of routine referral to CMHTs from practices that employ a counsellor. We cannot, however, draw any conclusions as to whether the referrals were inappropriate or not. Our data support evidence from previous research that a special interest or previous training in mental health is associated with improved detection of psychological morbidity (and hence higher rates of referral). An appropriate response would be closer links between the CMHTs and GPs to provide advice and training. This would enable patients with short term or minor mental disorders to be detected and treated within the primary care setting. Practices with lower than average referral rates to CMHTs may also benefit from training in clinical skills to enhance their ability to detect psychological disorders.