Dr Ian Anderson explains why the British Association for Psychopharmacology felt the need to develop new, evidence-based guidelines for treating depression


In 1993, the British Association for Psychopharmacology (BAP) published guidelines for treating depressive illness with antidepressants in the Journal of Psychopharmacology.1 These guidelines were based on a consensus meeting involving experts on depression and appeared at around the same time as statements from other authorities (e.g. the American Psychiatric Association2).

The early 1990s also saw the joint initiative from the RCP and the RCGP – the Defeat Depression Campaign – aimed principally at improving the recognition of depression as a treatable illness by both the public and doctors.

During the 1990s, treatment guidelines proliferated at both local and national level in the UK, as a result of political and professional priorities including rising healthcare costs, concerns about the quality of, and variation in, clinical practice and the advent of clinical governance and calls for evidence-based practice.


The Institute of Medicine has defined guidelines as 'systematically developed statements to assist practitioner and patient decisions about appropriate healthcare for specific clinical circumstances'.

There are obvious benefits in terms of setting standards, improving practice and informing patients and clinicians. Less obviously, guidelines may highlight deficiencies in the evidence and areas of uncertainty and therefore stimulate research or at least clarify which decisions are based on opinion rather than data.

Guidelines may, however, also be harmful for the service or individual patients if they are wrong or incomplete. It could be argued that this is inevitable because evidence is often inadequate or even misleading. Value judgments have to be employed in moving from evidence to recommendations and this is magnified if the evidence is poor.

Furthermore, even if recommendations are the best for patients in general, this does not mean that they are the best for an individual patient. This cautions against using guidelines as protocols.

There is also the spectre of an ugly medicolegal slippery slope if clinicians were to be penalised or sued for not following specific guidelines.

Guidelines therefore have to be applied intelligently; they do not replace clinical judgment but should be a means of strengthening it. For this to be possible there needs to be confidence in the guideline and its evidence base, and this can only occur if the user is aware of its pedigree and the status of its recommendations.


Traditionally, guidelines have been based on the views of a group of experts. These may fail to distinguish between quality of evidence and offer no guarantee that all, or even the most relevant, studies are reviewed.

In recognition of the limitations of this approach, there has been a move to promoting evidence-based guidelines. These are based on 'systematic' reviews in which defined literature searches are made to try to obtain all relevant studies and attempts are made to draw unbiased conclusions by emphasis on randomised controlled trials and meta-analyses.3

Given the inevitable limitations of the evidence available, it is also important that the process of drawing up evidence-based guidelines is explicit because the final recommendations will depend on judgments based on the evidence. Five steps in identifying evidence-based guidelines have been suggested (Table 1).4

Table 1: Steps in developing an evidence-based guideline
  1. Identifying and refining the subject area
  2. Convening and running guideline development groups
  3. Assessment of the evidence about the clinical question or condition by the group on the basis of systematic reviews
  4. Translation of the evidence into a recommendation within a clinical practice guideline
  5. External review of the guideline

From Shekelle et al4

In order to assess the evidence and be explicit about the strength of recommendations it is important to grade the quality of evidence and derived recommendations (see Table 2).

Table 2: Categories of evidence and strength of recommendation

categories of evidence for causal relationships and treatment

Ia Evidence from meta-analysis of randomised controlled trials
Ib Evidence from at least one randomised controlled trial
IIa Evidence from at least one controlled study without randomisation
IIb Evidence from at least one other type of quasi-experimental study
III Evidence from non-experimental descriptive studies, such as comparative studies, correlation studies and case-control studies
IV Evidence from expert committee reports or opinions and/or clinical experience of respected authorities
Strength of recommendation
A directly based on category I evidence
B directly based on category II evidence or extrapolated* recommendation from category I evidence
C directly based on category III evidence or extrapolated* recommendation from category I or II evidence
D directly based on category IV evidence or extrapolated* recommendation from category I, II or III evidence
* Extrapolation may be necessary because the evidence may be only indirectly related, cover only a part or the area of practice under consideration, or be contradictory
From Shekelle et al4

In addition, because evidence accumulates it needs to be recognised that guidelines are not set in stone and will need to be reviewed and revised in due course. One way to do this is to specify when the evidence or systematic review is to be updated.


Clearly the BAP 1993 guidelines needed to be revised. As was common practice at that time, they were developed at a consensus meeting on the basis of experts' reading of the literature, but were not explicitly evidence-based in current terms and clear distinction between evidence and recommendations was not made.

In fact the main recommendations have not altered, but in the intervening years there have been significant advances in the scope and nature of the evidence, particularly bearing on complexities in the range of presentation of depressive disorders and the threshold for treatment, new antidepressants, and the issue of problems with discontinuation of antidepressants. These factors led us to the conclusion that there was a need to review and revise the guidelines.

In 1998, members of the previous consensus meeting and a range of experts in depressive illness and its treatment were asked to identify areas of the old guidelines that were problematic and/or needed updating.

On the basis of this, a consensus meeting was organised in November 1998 at which presentations based on systematic reviews of key areas were given to a range of current experts in the field, including members of industry.

The views of users were represented but they were not part of the discussions at this stage. The evidence and strength of recommendations were evaluated as outlined in Table 2 (above).

The scope of the guidelines was defined and restricted to the most common depressive disorders, excluding psychotic depression and depression in bipolar affective disorder (manic depression).

I was then responsible for coordinating the evidence and expanding it where necessary to produce a draft document. This was circulated to members of the meeting, experts in evidence-based psychiatry and users.

The final guidelines, which incorporated feedback on the draft, were accepted by the BAP ruling body (Council) and published in March 2000.5


The task of producing a document that presented and evaluated the evidence, while being both comprehensive and short enough to be directly used in the clinic, defeated us. The resulting guidelines are therefore designed to be used as a basis for practical summaries.

The introduction to the guidelines reviews what we know of the epidemiology, nature and course of depressive disorders. The variety of presentations is covered, with the recommendation that a distinction is drawn between major depressive disorder (or illness), milder depressions and dysthymia (milder depression lasting 2 years or more) (Box 1).

Box 1: Recommendations for diagnosis

Identification of the following three presentations of patients with depressive symptoms guides treatment choice (A):
Major depression (major depressive episode according to DSM-IV criteria 3).
Milder depression (recent-onset depressive symptoms not meeting DSM-IV criteria for depressive episode).
Dysthymia, a chronic milder depressive disorder (at least 2 years of depressive symptoms not meeting DSM-IV criteria for depressive episode) which is not a consequence of a partly resolved major depression.
Subdivision of major depression into mild, moderate and severe, guides treatment choice (B) although precise definitions are difficult in practice.
Severe major depression is characterized by the presence of all, or nearly all, DSM-IV depressive symptoms to a clinically severe degree and marked functional impairment in all areas of life.
Mild through to moderate severity ranges from a threshold number of symptoms with minimal functional impairment through to marked symptoms and impairment of function.

This distinction is useful because there is little evidence that milder depressions (and the mildest major depression) respond specifically to antidepressants. It appears likely that the benefit from drugs increases with the severity of the depression.

The guidelines recommend psychiatric referral for certain conditions (see Box 2, below), but point out that there are no controlled data relating to the indications for psychiatric referral.

Box 2: Recommendations for psychiatric referral

Referral to psychiatric services is indicated if there is a:

Risk of suicide (D)
Psychotic symptoms (D)
A history of bipolar affective disorder (D).
Consultation with, or referral to, a psychiatrist (or a specialist in the treatment of affective disorders), is appropriate
When the practitioner feels insufficiently experienced to manage a patient's condition (D)
If two or more attempts to treat the patient's depressive disorder have failed or resulted in only partial response (D).

The average severity of depression seen in general practice is at about the threshold where benefits from antidepressants start to occur. A major reason is that many milder depressions are relatively short-lived and the natural recovery rate is high. This is important in avoiding unnecessary drug prescription. Box 3 (below) shows the recommendations regarding indications for antidepressants.

Box 3: Recommendations regarding indications for antidepressants
Major depression: antidepressants are a first-line treatment irrespective of environmental factors (A). At the mildest severity the benefit is uncertain
Dysthymia: antidepressants are a first-line treatment (A).
Acute milder depressions at initial presentation:
Antidepressants are not indicated (B)
Education, support and simple problem solving are recommended (D)
Monitor for persistence or for the development of major depression (D).
Persistent milder depression: a trial of antidepressants is recommended (D).
Milder depression with history of major depression: consider treatment with antidepressants (D).

Patients with dysthymia do, however, appear to benefit from antidepressants and, by extrapolation, so may those with milder depressions that persist. Assessment of milder depression is therefore a two-stage process – initial watchful waiting followed by a trial of treatment if it persists.

The guidelines consider the alternatives to antidepressants and review the evidence for specific psychotherapies such as cognitive behaviour therapy (good), non-directive counselling (poor), St John's wort (modest to good), exercise and bibliotherapy (modest).

In considering choice of antidepressant, the guidelines conclude that there is now sufficient evidence to prefer newer antidepressants as first-line treatment, with most evidence for selective serotonin re-uptake inhibitors (SSRIs) (see Box 4, below).

Box 4: Recommendations for choice of antidepressants

Match choice of antidepressant drug to individual patient requirements as far as possible, taking into account likely short-term and long-term effects (D).
In the absence of special factors, choose antidepressants that are better tolerated, safer in overdose and more likely to be prescribed at effective doses (C). There is most evidence for selective serotonin re-uptake inhibitors (SSRIs); lofepramine, mirtazapine, nefazodone, reboxetine and venlafaxine are also relatively safe and well-tolerated.
In severely ill hospitalized patients, and in other situations where maximizing efficacy is of overriding importance, consider an older tricyclic antidepressant or venlafaxine at a dose of 150mg or greater in preference to an SSRI (B) or monoamine oxidase inhibitor (B).
Factors to consider in choosing an antidepressant include:
Previous treatment response to a particular drug (D)
Tolerability and adverse effects of a previously given drug (D)
Likely side-effect profile (e.g. sedation, weight gain) (C)
Low lethality if history or likelihood of overdose (D)
Concurrent physical illness or condition that may make the antidepressant more noxious or less well-tolerated (C)
Concurrent medication that may interact with the antidepressant drug (C)
Associated psychiatric disorder that may specifically respond to a particular class of antidepressant (e.g. obsessive-compulsive disorder and SSRIs) (C)
Patient preference (D).
Dysthymia: treat using the same principles as for major depression (D).

This is based not only on controlled evidence of better tolerability (although the size of benefit is controversial) but also, and equally important, on naturalistic and survey data indicating that SSRIs are much more likely than tricyclic antidepressants to be prescribed at known therapeutic doses and for an adequate period.

The guidelines recommend the frequency of monitoring of patients started on antidepressants (every 1–2 weeks at first) and the use of educational methods to enhance treatment compliance.

How long to persist with initial treatment before altering course is also considered (4-6 weeks) and the options for which there is evidence of efficacy at that stage – essentially increasing dose (modest), changing drug (modest) or augmenting with another drug (good).

The guidelines reinforce the need to treat for an adequate length of time – at least 6 months after remission of an individual episode – in order to halve the significant relapse rate.

Less well recognised is the need for a substantial group of patients to receive maintenance (or prophylactic) treatment. This group includes those with three or more recent episodes of significant depression or more than five episodes altogether.

A feature of the guidelines is that consideration of the elderly is included and the general rule is that treatment should be vigorous but on a longer timescale in terms of judging response and continuing treatment.

Finally, the problem of discontinuation reactions on stopping antidepressants is considered. It is recommended that once antidepressants have been taken for more than a few weeks, at least a 4-week taper should be attempted if possible once the decision to stop has been made. Patients should routinely be warned about the possibility of discontinuation reactions, as many make their own decision to stop treatment suddenly.


Guidelines are only as good as their implementation. There is consistent evidence that the majority of depressed patients in both primary and secondary care receive less than adequate antidepressant treatment. Successful application of guideline-based treatment has been shown to improve outcome, at least in the short to medium term.5

However, it is often difficult to demonstrate that guidelines bring about any change in practice, and passive dissemination of information seems ineffective.6 The failure of guideline implementation may be linked to a questioning of the quality of the evidence base, and therefore the recommendations of depression guidelines.7 This can only be tackled by explicitly evidence-based guidelines such as ours.

Improvement has been demonstrated when guidelines are linked to alteration in service delivery or structural changes, including appropriate reminders. Guidelines, however good, are only the starting point.

  • Boxes 1-4 are reproduced from the British Association for Psychopharmacology evidence-based guidelines for treating depressive disorders with antidepressants,5 by kind permission of the BAP.


  1. Montgomery SA, Cowen PJ, Deakin W et al. Guidelines for treating depressive illness with antidepressants: a statement from the British Association for Psychopharmacology. J Psychopharmacol 1993; 7 :19-23.
  2. American Psychiatric Association. Practice guideline for major depressive disorder in adults. Am J Psychiatry 1993; 150 (Suppl): 1-24.
  3. Woolf SH, Grol R, Hutchinson A, Eccles M, Grimshaw J. Potential benefits, limitations, and harms of clinical guidelines. Br Med J 1999; 318: 527-30.
  4. Shekelle PG, Woolf SH, Eccles M, Grimshaw J. Developing guidelines. Br Med J 1999; 318: 593-6.
  5. Anderson IM, Nutt DJ, Deakin JFW. Evidence-based guidelines for treating depressive disorders with antidepressants: a revision of the 1993 British Association for Psychopharmacology guidelines. J Psychopharmacol 2000; 14: 3-20.
  6. Bero LA, Grilli R, Grimshaw JM, Harvey E, Oxman AD, Thompson MA. Closing the gap between research and practice: an overview of systematic reviews of interventions to promote the implementation of research findings. Br Med J 1998; 317: 465-8.
  7. Kendrick T. Why can't GPs follow guidelines on depression? Br Med J 2000; 320: 200-1.

British Association for Psychopharmacology
The British Association for Psychopharmacology (BAP) is one of the largest national psychopharmacology associations in the world and is independent of government and industry. Any doctor with an interest in psychopharmacology is eligible to join. For further information about the BAP, contact: Mrs Susan Chandler, Administrator, 6 Regent Terrace, Cambridge CB2 1AA, or look on the website at http://www.bap.org.uk

Guidelines in Practice, August/September 2000, Volume 3
© 2000 MGP Ltd
further information | subscribe