738 British Journal of General Practice, September 1999
on behalf of The Nottingham Counselling and
Antidepressants in Primary Care (CAPC) Study Group
Major depression can be treated effectively with antidepres-
sants. However, in the United Kingdom, patients with
depression are often referred to counsellors, and surveys
indicate that public opinion favours this approach. We car-
ried out a literature review to determine the evidence for the
effectiveness of counselling for depression in primary care.
Because no studies were identified in which counselling
had been evaluated specifically in relation to treating
depression, we examined indirect evidence from studies
evaluating the overall effectiveness of generic counselling in
primary care, and studies evaluating the effectiveness of
psychological treatments, other than counselling, for
depression. Methodological problems influencing the inter-
pretation of such studies are discussed. We conclude that,
while specific psychological treatments have been shown to
have equivalent effectiveness as antidepressants, there is
currently insufficient evidence to recommend that generic
counselling should be used alone in the treatment of
patients with major depression.
Keywords: depression; antidepressants; counselling; gener-
detection and appropriate management may reduce the severi-
ty, alleviate distress, and possibly reduce the risk of recurrence or
chronicity.2,3General practitioners (GPs) play a key role in this
process since 90% of depression is managed in primary care.4
EPRESSIVE disorders are common and disabling.1Early
There is clear evidence that antidepressants are both effective
and relatively safe in the treatment of major depressive
disorders.5However, public opinion is largely negative concern-
ing their use: a recent survey of public attitudes following the
national Defeat Depression Campaign (1992–96) found that, of
approximately 2000 people interviewed, 74% believed antide-
pressants to be addictive, and only 60% believed them to be an
effective treatment for depression.6Only 24% believed that peo-
ple with depression should be offered antidepressant treatment.
Professional attitudes to antidepressants also appear to differ:
there is wide variation in prescribing rates between GPs,7and
patients are frequently prescribed courses at subtherapeutic
dosages or for insufficient duration.8,9
By contrast, public opinion strongly favours the use of psycho-
logical treatments for depression. The Defeat Depression survey
found that 86% of responders believed ‘counselling’ to be an
effective intervention for depression, and 90% agreed that people
with depression should be offered it.6‘Counselling’, although
poorly defined, is one of the most widely available forms of psy-
chological therapy within general practice, having developed
over the past 20 years. Over one-third of GPs have direct access
to counselling services, most of which are provided by specific
‘counsellors’ as distinct from other health professionals employ-
ing such techniques.10-12Although counsellors deal with a range
of conditions, surveys have demonstrated that between 20% and
74% of their workload consists of patients with depression.11,13-16
It therefore appears that GPs consider counselling to be an appro-
priate form of management for depressive illness, and that this
view concurs with public opinion.
We carried out a literature search to identify published evi-
dence for the effectiveness of counselling in treating major
depression. In particular we aimed to seek evidence for the rela-
tive effectiveness of counselling in relation to the current recom-
mended ‘gold standard’ of antidepressant therapy.
We searched MEDLINE and EMBASE electronic databases cov-
ering the period 1980–98. We also searched the Cochrane
Library and hand-searched index lists of identified sources. Our
strategy was intended to identify references to trials of psycho-
logical treatments for major depression for adults in primary
care, and incorporated the following thesaurus terms: ‘primary
care’, ‘family practice’, ‘general practice’, ‘depressive disor-
ders’, ‘unipolar depression’, and ‘depression’. Terms used to
specify broad categories of psychological treatments were as fol-
lows: ‘counselling’, ‘counseling’, ‘psychotherapy’, and ‘psychol-
ogy — clinical’.
We selected randomized controlled trials performed in United
Kingdom (UK) general practice in which a psychological treat-
ment was compared with either another active intervention or
with ‘usual treatment’. We excluded studies of postnatal depres-
sion, depression in the elderly, co-morbid depressive disorders,
psychotic depression, or bipolar affective disorders. Where
appropriate, reference has been made to meta-analyses or studies
performed in other settings.
Should general practitioners refer patients with
major depression to counsellors? A review of
current published evidence
R Churchill, MB BS, MSc, MRCGP, lecturer in general practice, Division of
General Practice, University of Nottingham. V Gretton, PhD, research
assistant, Trent Institute for Health Services Research, Nottingham.
C Chilvers, DSc Hon, MFPHM, professor of epidemiology and director,
Trent Institute for Health Services Research, Nottingham. M Dewey,
PhD, senior lecturer in health services research, Trent Institute for Health
Services Research, Nottingham. C Duggan, MD, MRCPsych, PhD, professor
of forensic psychiatry, University of Leicester. A Lee, MA, MBBS, FRCPsych,
consultant psychiatrist and special senior lecturer, Queen’s Medical
© British Journal of General Practice, 1999, 49, 737-743.
British Journal of General Practice, September 1999739
R Churchill, M Dewey, V Gretton, et al
Direct evidence for the effectiveness of counselling for
We identified no published studies that specifically address the
issue of the effectiveness of counselling for major depression in
general practice. Indirect evidence was, however, available from
two collateral sources: first, studies evaluating the overall effec-
tiveness of counselling in general practice; and secondly, studies
examining the effectiveness of psychological treatments, other
than counselling, for depression. This review therefore examines
Studies evaluating counselling in general practice
We identified five randomized controlled trials of generic coun-
selling in UK primary care (Table 1). All compare the outcomes
of counselling with routine management by GPs for a range of
disorders. We examined the results of each to determine if there
were any indicators for the effectiveness of counselling specifi-
cally for depressive disorders.
Findings from the Leverhulme Counselling study13have never
been published, possibly because of the methodological prob-
lems encountered, which included concerns about incomplete
data collection. Analysis was based on only 34% of those origi-
nally randomized, and was not carried out on an intention to treat
basis, although cases and controls were matched by age and sex.
There were no differences in any of the main outcome measures
between baseline and 12 months, or in General Health
Questionnaire (GHQ) scores at 12 months. However, counselled
patients who were taking tranquillizers at baseline were signifi-
cantly more likely to have stopped at 12 months than controls,
while there were no differences in the proportion of patients who
stopped taking antidepressants between the two groups. This
finding would be consistent with the possibility that counselling
was less effective as an adjuvant therapy for depressive illnesses
than for other types of morbidity.
Boot and colleagues employed the GHQ as an objective out-
come measure at both baseline and six weeks.14At follow-up,
counselled patients had significantly lower mean GHQ scores
than those who had received ‘treatment as usual’, with both
groups having significantly lower scores than at baseline. Nine
per cent of the counselled group were prescribed antidepressants
compared with 23% of controls, and the former were also signifi-
cantly less likely to have been referred to outside agencies.
Patients who had received counselling were more likely to report
that they were coping better, feeling happier, and satisfied with
the treatment received. This study was limited by a short dura-
tion and an imbalance in numbers randomized to each group,16
although baseline demographic and morbidity characteristics
were similar. Although many of the participating patients were
subjectively depressed at baseline, there was no formal assess-
ment of the presence of major depression and no indication of
which groups of patients showed greatest improvements with
More recently, Friedli and colleagues performed a comprehen-
sive evaluation of non-directive Rogerian counselling.15
Audiotapes of the counselling sessions were independently eval-
uated to ensure adherence to the therapeutic model. Overall there
were no significant differences in outcome — which included the
Beck Depression Inventory (BDI) — between the two groups,
except that those receiving counselling reported significantly
more satisfaction. All measures of severity decreased significant-
ly over time. There was a significantly greater fall in BDI scores
in counselled patients than those receiving treatment as usual
who were among the sub-group whose initial BDI score was
greater than 14. This suggests that patients with depression may
respond better to counselling than to ‘treatment as usual’.
This study has advantages over previous research in that the
attrition rate was relatively low, there was an attempt to stan-
dardize the intervention, a range of validated outcome measures
was employed, and patients were followed up after completing
treatment. However, while a high proportion of patients had
depressive features, it is uncertain how many would have ful-
filled criteria for major depression. Only one-fifth of patients
receiving routine general practice care were prescribed anti-
depressants, which is lower than might be expected in the con-
text of patients with such relatively high Beck Depression scores.
Thus, as with all studies using ‘treatment as usual’ for compari-
son, it is unclear whether participants in the control group
received ‘best possible’ treatment. A further criticism is that the
counselling intervention was tightly controlled, and the results
therefore only reflect the outcomes for non-directive Rogerian
counselling, when, in practice, such approaches are often inte-
grated with other techniques.17
Hemming’s study18did not report counselling outcomes in
relation to specific reasons for referral, and so no inference can
be made about its effectiveness for depression. Overall outcomes
were similar between intervention and control patients; however,
half of the control group were referred to outside sources of psy-
chological treatment as part of the routine treatment by GPs.
Harvey and colleagues evaluated the effectiveness of generic
counselling in general practice using a range of outcome mea-
sures including the Hospital Anxiety and Depression scale.16
There were no significant differences in outcome between
patients receiving counselling and those receiving ‘treatment as
usual’ on the depression sub-scale of this measure. However, as
with other studies, ‘treatment as usual’ was not defined and may
have been suboptimal in terms of the proportion of patients
receiving adequate dosages of antidepressants. The study includ-
ed an economic evaluation that demonstrated no clear cost
advantage for either intervention.
Studies of psychosocial treatments for depression other
Counselling is a relatively poorly defined psychological inter-
vention, with primary care counsellors using a range of therapeu-
tic styles from Rogerian non-directive counselling to behavioural
therapy and psychodynamic psychotherapy.11Many use an
‘eclectic’ or ‘integrative’ approach, employing features from dif-
ferent models. We therefore examined the evidence for the effec-
tiveness of specific types of psychological treatment for depres-
sion in primary care. Such treatments vary in their theoretical
basis, content, and duration (Box 1).
Cognitive therapy (CT) has been most extensively evaluated.
Studies from a variety of settings have been subject to meta-
analysis that have concluded that, at the end of treatment for mild
to moderate depression, CT is at least equivalent in effectiveness
to treatment with antidepressants and may also reduce the risk of
relapse.5,19,20Two meta-analyses concluded that CT was superior
to pharmacotherapy and other forms of psychotherapy, but both
used the BDI as an outcome measure, which, since it focuses on
cognitive symptoms, would tend to be biased towards a positive
outcome for cognitive therapy.19,20
Studies of CT that have been conducted specifically in UK pri-
mary care are harder to interpret since they have used different
comparison groups, outcome measures, and durations of follow-
up (Table 2).
Overall, CT appears to be more effective than routine treat-
ment by a GP at the end of a course of treatment, with any signif-
icant differences disappearing on follow-up. The exception is
British Journal of General Practice, September 1999
R Churchill, M Dewey, V Gretton, et al
Table 1. Randomized controlled trials comparing brief generic counselling with treatment as usual by a GP in UK general practice.
Main outcome measures
and timing (in italics)
depression (%)ReferenceSetting Intervention Sample size/follow-up ratePrincipal results
GP consultation rate
GHQ (end point only)
Baseline & 12 months
726 randomized: 453
controls and 273 counselling
Data available and
analysed for 248 (34%)
43% (GP recorded)
(73% had depressive
No overall differences
between groups except
reduction of tranquiliser
Boot et al,
GP consultation rate
Baseline and six weeks
192 randomized: 68 controls
and 124 counselling
108 followed up (56%)
improved GHQ scores,
fewer psychotropic drug
prescriptions, and greater
satisfaction at follow-up
Friedli et al,
Four counsellors —
Between one and 12
Beck depression inventory
Brief symptom inventory
Revised clinical interview
Modified social adjustment scale
Baseline, three months,
and nine months
136 randomized: 66 controls
and 70 counselling
110 (81%) followed up
at three months
117 (86%) followed up
at nine months
52% (patient self-report)
51% (GP reported)
74% (based on BDI
threshold of 14)
No differences in clinical
Satisfaction was significantly
higher in intervention group
Subgroup of patients with
higher BDI did better
Up to 14 sessions
Inventory of interpersonal
GP consultation rate
Baseline, four months, and
188 randomized in ratio of
2:1 cases to controls
154 (82%) followed up
at four months
100 (53%) followed
up at eight months
No details providedNo differences in main
Control group was more likely
to have been referred for
Harvey et al,
Hospital Anxiety and
Dartmouth COOP Charts
Delighted-terrible faces scale
Baseline and four months
162 randomized (2:1 ratio):
51 controls and 111
122 (75%) followed up
24% (GP recorded)No differences between
groups in outcomes or costs
Subjects in all trials were aged 16 years and above with social, psychological, or emotional problems and excluding serious psychiatric morbidity. GHQ: General Health Questionnaire.
British Journal of General Practice, September 1999741
R Churchill, M Dewey, V Gretton, et al
Scott and Freeman’s study suggesting that antidepressants, pre-
scribed by a psychiatrist, may actually result in more rapid
response.21Differences between studies may therefore reflect the
use made of antidepressants in the ‘treatment as usual’ compari-
Problem solving therapy has been shown to be effective in
treating emotional disorders in primary care22and has recently
been evaluated in major depression.23Mynors-Wallis and col-
leagues undertook a randomized controlled trial of problem-solv-
ing treatment versus antidepressants or drug placebo in general
practices in Oxfordshire. Treatment in all arms was provided by
a psychiatrist and two specially trained GPs who provided a sim-
ilar amount of contact time for all three groups. At 12 weeks,
60% of patients receiving problem-solving treatment had ‘recov-
ered’, according to the Hamilton rating scale threshold, com-
pared with 52% taking amitriptyline and 27% on placebo (based
on intention-to-treat analysis). The authors concluded that prob-
lem-solving treatment was at least as effective as antidepressants
and superior to placebo in the management of major depression
at six and 12 weeks.
Social problems are prevalent among patients consulting with
depression,24and social work interventions may involve practical
help as well as supportive listening and advice. The Edinburgh
primary care depression study compared efficacy and cost of a
social work intervention with antidepressants, cognitive therapy,
or routine GP care for patients with depressive illness.21After 16
weeks, the severity of depressive symptoms had declined in all
groups, and differences in efficacy were not commensurate with
differences in length or cost of treatment. Only the social work
intervention was superior to GP care, although the baseline
severity of depression was lower in this group than the others.
Social work counselling was rated most helpful by patients.
Interpersonal therapy has been widely evaluated in the United
States of America but not, to our knowledge, in the UK. It
appears to be as effective as antidepressant treatment for mild to
moderate depression but less so for more severe forms.25,26There
is evidence that continued treatment may reduce the risk of
Cognitive therapyAimed at modifying the negative thinking
that occurs in depressive disorders and that
is believed to contribute to their onset and
Problem-solving therapy Aimed at helping patients to use their own
skills and resources to deal with present
and future problems by means of identifi-
cation and clarification of the problem,
goal setting, brainstorming and selecting
solutions, clarifying implementation, and
Social work intervention A potentially broader approach that may
encompass non-specific counselling and
directive advice and provision of practical
help to deal with adverse circumstances
Interpersonal therapy Aimed at reducing some of the social
problems that may provoke or prolong
depression by helping the patient to identi-
fy and understand his or her interpersonal
problems and conflicts, and develop more
adaptive ways of relating to others.
Box 1. Psychological interventions for depressive disorders.
Table 2. Randomized controlled trials evaluating cognitive therapy (CT) for major depression in UK primary care patients.
Main outcome measures and duration of follow-up
Sample size/follow-up rate
39 randomized: 24 (62%)
CT superior to AD at 20
et al, 198134
Baseline, then two to three times weekly to 20 weeks
‘Treatment as usual’
BDI, MADS, HRS
CT superior to TAU at end
et al, 198433
Baseline, post-therapy, and three months later
77% followed up
of treatment but no significant
difference three months later
Waiting list ‘treatment
51 randomized: 19/30 (63%)
CT superior to TAU at
as usual’ (TAU)
Baseline, post-therapy, then three months,
of immediate therapy group
end of treatment
six months, and 12 months later
1. Antidepressants (AD)
91 randomized in these three
AD superior to other groups at
2. ‘Treatment as usual’
Baseline, four weeks, and 16 weeks
arms: 92% follow up to 16 weeks
four weeks but no significant
difference at 16 weeks
Scott et al,
‘Treatment as usual’
DSM-III-R, HRS, BDI
48 randomized: 34 (71% )
CT superior to TAU at end of
Baseline, seven weeks, three months, six months,
followed up at seven weeks;
treatment with no difference
and 12 months
and 24 (50%) at 12 months
three months later
Subjects in all trials were aged between 18 and 65, and fulfilled either Research Diagnostic Criteria or DSM-III-R Criteria for major depressive disorder. HRS: Hamilton Rating Scale for Depression;
BDI: Beck Depression Inventory; MADS: Montgomery-Asberg Depression Scale.
742 British Journal of General Practice, September 1999
R Churchill, M Dewey, V Gretton, et al
Much of the available evidence of the effectiveness of psycho-
logical treatments for depression comes from studies of limited
duration, with no details of relapse or recurrence rates. Common
difficulties of study design include limited sample size; relatively
high attrition rates; the use of differing outcome measures; the
use of atypical, highly skilled, and motivated therapists for the
psychological intervention; and the failure to use antidepressant
treatment as the ‘gold standard’ for comparison.
The evidence presented suggests that a range of specific psycho-
logical treatments may be as effective as antidepressants in treating
depression in primary care in the short term. This raises the issue
as to whether the effectiveness is a result of the specific type of
treatment employed or of a more generic effect that could equally
apply to other psychological interventions including counselling.
In support of this, several studies have suggested that outcomes of
psychological treatment are unrelated to the theoretical basis of
therapy and are more closely correlated with non-specific aspects
of the interaction such as exploration and warmth.29-32
If therapeutic contact per se is the active ingredient, then it
might be expected that this would show a ‘dose-response’ rela-
tionship, with duration of treatment correlating with outcome.
Dobson failed to demonstrate this in one meta-analysis of studies
of cognitive therapy for depression.19Equally, Teasdale and col-
leagues,33reviewing past literature, concluded that ‘therapeutic
attention’, or time spent with the therapist, does not necessarily
produce a reduction in symptoms in patients with major depres-
sion. The evidence for a non-specific effect of psychological
treatments in depression is therefore equivocal and, at present,
insufficient to suggest that any psychological intervention, includ-
ing generic counselling, will be an effective form of treatment.
Studies of primary care counselling provide only weak evidence
of a specific benefit in depressive disorders, and results of stud-
ies of other specific psychological treatments for depression can-
not readily be extrapolated to generic counselling. However, the
increasing availability of counselling services, positive lay atti-
tudes about the benefits of counselling for depression, and
reported higher levels of satisfaction from counselling in com-
parison with usual GP care, all increase the pressure for GPs to
refer patients with depression to a counsellor. In the absence of
clear evidence, unless a counsellor is known to employ specific
techniques of proven benefit, GPs might consider restricting
referrals for counselling to patients with major depression in
whom antidepressants are unacceptable, poorly tolerated, or inef-
fective. The results of trials comparing the effectiveness and
cost-effectiveness of counselling with antidepressants in patients
with major depression are awaited.
1. Wright AF. Depression: recognition and management in general
practice. London: Royal College of General Practitioners, 1993.
2. Shea MT, Elkin I, Imber SD, et al. Course of depressive symptoms
over follow-up: findings from the National Institute Mental Health
Treatment of Depression Collaborative Research Program. Arch Gen
Psychiatry 1992; 49: 782-787.
3. Scott J, Eccleston D, Boys R. Can we predict the persistence of
depression? Br J Psychiatry 1992; 161: 633-637.
4. Paykel ES, Priest RG. Recognition and management of depression in
general practice: consensus statement. BMJ 1992; 305: 1198-1202.
5. Freemantle N, Song F, Sheldon TA, et al. The treatment of depres-
sion in primary care. York: NHS Centre for Reviews and
6. Paykel ES, Hart D, Priest RG. Changes in public attitudes to depres-
sion during the Defeat Depression Campaign. Br J Psychiatry 1998;
7. Pharoah PD, Melzer D. Variation in prescribing of hypnotics, anxi-
olytics and antidepressants between 61 general practices. Br J Gen
Pract 1995; 45: 595-599.
8. Kerr MP. Antidepressant prescribing: a comparison between general
practitioners pereceptions’ and the relevance of clinical psychology.
Br J Gen Pract 1994; 44: 275-276.
9. Matthews K, Eagles JM, Matthews CA. The use of antidepressant
drugs in general practice. A questionnaire survey. Eur J Clin
Pharmacol 1993; 45: 205-210.
10. Sibbald B, Addington-Hall J, Brenneman D, Freeling P. Counsellors
in English and Welsh general practices: their nature and distribution.
BMJ 1993; 306: 29-33.
11. Sibbald B, Addington-Hall J, Brenneman D, Obe PF. The role of
counsellors in general practice. A qualitative study. [Occasional
paper no 74.] London: Royal College of General Practitioners, 1996.
12. Clark A, Hook J, Stein K. Counsellors in primary care in
Southampton: a questionnaire survey of their qualifications, working
arrangements, and case-mix. Br J Gen Pract 1997; 47: 613-617.
13. Ashurst PM, Ward DF. An evaluation of counselling in general prac-
tice. [Final report of the Leverhulm Counselling Project.] London:
Mental Health Foundation, 1983.
14. Boot D, Gillies P, Fenelon J, et al. Evaluation of the short-term
impact of counselling in general practice. Patient Education and
Counseling 1994; 24: 79-89.
15. Friedli K, King MB, Lloyd M, Horder J. Randomised controlled
assessment of non-directive psychotherapy versus routine general-
practitioner care. Lancet 1997; 350: 1662-1665.
16. Harvey I, Nelson SJ, Lyons RA, et al. A randomized controlled trial
and economic evaluation of counselling in primary care. Br J Gen
Pract 1998; 48: 1043-1048.
17. Tylee A. Counselling in primary care. Lancet 1997; 350: 1643.
18. Hemmings A. Counselling in primary care: a randomised controlled
trial. Patient Education and Counseling 1997; 32: 219-230.
19. Dobson KS. A meta-analysis of the efficacy of cognitive therapy for
depression. J Consult Clin Psychol 1989; 57: 414-419.
20. Gloaguen V, Cottraux J, Cucherat M, Blackburn I. A meta-analysis
of the effects of cognitive therapy in depressed patients. J Affect
Disord 1998; 49: 59-72.
21. Scott AIF, Freeman CPL. Edinburgh primary care depression study:
Treatment outcome, patient satisfaction, and cost after 16 weeks.
BMJ 1992; 304: 883-887.
22. Catalan J, Gath DH, Anastasiades P, et al. Evaluation of a brief psy-
chological treatment for emotional disorders in primary care. Psychol
Med 1991; 21: 1013-1018.
23. Mynors-Wallis LM, Gath DH, Lloyd-Thomas AR, Tomlinson D.
Randomised controlled trial comparing problem solving treatment
with amitriptyline and placebo for major depression in primary care.
BMJ 1995; 310: 441-445.
24. Sireling LI, Paykel ES, Freeling P. Clinical features and comparisons
with outpatients. Br J Psychiatry 1985; 147: 119-126.
25. Weissman MM, Prusoff BA, Dimascio A, et al. The efficacy of
drugs and psychotherapy in the treatment of acute depressive
episodes. Am J Psychiatry 1979; 136: 555-558.
26. Elkin I, Shea MT, Watkins JT, et al. National Institute of Mental Health
Treatment of Depression Collaborative Research Program. General
effectiveness of treatments. Arch Gen Psychiatry 1989; 46: 971-982.
27. Frank E, Kupfer DJ, Wagner EF, et al. Efficacy of interpersonal psy-
chotherapy as a maintenance treatment of recurrent depression. Arch
Gen Psychiatry 1991; 48: 1053-1059.
28. Kupfer DJ, Frank E, Perel JM, et al. Five-year outcome for mainte-
nance therapies in recurrent depression. Arch Gen Psychiatry 1992;
29. Zeiss A, Lewinsohn P, Munoz R, et al. Non-specific improvement
effects using interpersonal skills training, pleasant activities sched-
ules, or cognitive training. J Consult Clin Psychol 1979; 47: 427-439.
·Psychological treatments for depression are favoured by
patients and employed by GPs.
There is good evidence that certain specific forms of psy-
chological therapy are as effective as antidepressants in
treating depression in primary care.
No studies have been published evaluating the effective-
ness of generic counselling for depression, despite its
widespread availability and use.
General practitioners should be cautious in referring
patients with major depressive illness to counsellors as
their sole mode of treatment, unless specific therapeutic
models are used.