Overall, the review included seven RCTs; three evaluations of existing services; two surveys of GPs views of counselling; two protocols for Cochrane reviews; two non-Cochrane published systematic reviews; and five reports.
One RCT favoured counselling in terms of benefit and decreased surgery attendance, psychotropic drug use and referrals.
There were no completed reviews conducted by Cochrane reviewers.
One published review (based on 19 RCTs identified from a MEDLINE search from 1974 to 1995) was not very clear and did not define some forms of treatments or some abbreviations. Patient groups were very heterogeneous, outcome measures were not clearly defined and intention-to-treat analysis was rarely used. The review concluded that the studies had significant methodological problems.
The other published review (based on 10 RCTs) concluded that the methodological rigour of the primary studies was questionable and evidence was limited.
1. An NHS Executive report in 1996 reviewed the variety of psychotherapy strategies used in treatment of adults and children in primary, secondary and tertiary care and reported gaps in research findings, weaknesses in research to uncover naturally occurring interaction in complex systems, and that the quality of outcome measure was not yet satisfactory.
2. A 1996 review of research into psychological services (based on 14 studies, including one meta-analysis of 12 previous studies) reported research to be heterogeneous and inadequate. The review was criticised on the grounds that few studies actually used counsellors.
3. A 1998 report, based on two reviews (one of which comprised 15 descriptive studies and 19 RCTs) found very little quality evidence of the relative effectiveness and efficiency of counselling. High levels of patient satisfaction with counselling were reported.
4. A feasibility study reported on a small scale pilot scheme of counselling in general practice and found average referral rates of 10.9 patients per 1000 practice population per year and very high rates (90%) of patient satisfaction.
A variety of counselling interventions were conducted by a variety of personnel.
Use of counsellors (3 unblinded RCTs, 521 patients): methodological flaws included large drop-outs rates, questions over altered GP care for those patients not receiving counselling, proportion of usual care patients referred for counselling, and failure to comment on apparent increase in the number of external referrals, lower drug costs and decreased need for GP time in the "usual care" patients. Outcomes were assessed between 4 and 9 months. None of the studies reported any difference between intervention groups and usual care, though sub-group analysis in one study suggested that depressed patients did benefit from counselling.
Trained nurses (2 RCTs): one RCT (70 patients) found that nurses trained in problem solving had no influence on symptom scores at 26 weeks but patients receiving the intervention reported higher satisfaction, decreased disability days and lower medication costs than those receiving usual care. The other RCT (577 patients) reported that the use of nurses trained in supportive care (no counselling provided) did not alter antidepressant prescribing, GP visits or symptom scores compared to usual care. Cognitive behaviour therapy and/or fluvoxamine for panic disorder and agraphobia (one RCT, 193 patients): CBT was effective and the most effective combination was CBT and fluvoxamine.
Evaluation of existing services (3 papers): all were methodologically flawed but seemed to suggest that counsellors have no effect on overall psychotropic prescribing.
GPs' views of counselling (2 surveys, 739 GPs): GPs considered that counselling generally reduced their workload, reduced prescribing of psychotropic drugs and reduced referral to psychiatric and psychological services. The other survey found that 60% of GPs would select counselling as their first choice for psychological treatment of elderly patients with depression. However, clinical psychological services had long waiting lists, were not available locally, and there was a lack of transport.