How do clinicians choose antidepressants ?

Research suggests no difference in efficacy be tween tricyclic and related antidepressant drugs. The British National Formulary (1993) lists 21 different antidepressants. Reviews of antidepres sant treatment suggest that choice of antidepres sant should be made on the grounds of previous drug exposure and side effect profiles of different drugs. At present there are no clinical guidelines to indicate which antidepressant should be pre scribed in particular clinical situations, and no evidence to suggest differences in subgroups of depressives in their response to antidepressants. Study of the rate of fatality in overdose has led Beaumont (1989) to advocate limiting the use of the older tricyclic drugs. Clinical experience sug gests that psychiatrists adopt a varied approach to treatment, and it has been argued (Joyce & Paykel, 1989) that treatment with antidepres sants is often inappropriate or inadequate, or both. This study describes the antidepressant pre scribing practice of a group of Bristol psychia trists, and examines the reasons given by clinicians for their choice of antidepressant, and their attitudes to involving patients in choosing between alternative treatments.


The study
The instrument used was a 13-item postal ques tionnaire.It was sent twice to increase the re sponse rate, with a two month interval, to all Southmead Trust doctors who work permanently in psychiatry.This included consultants, senior registrars, clinical assistants and postmembership registrars.Trainees without mem bership were not included, as it was felt that they may be a different group, comprising many GP trainees.The questionnaire asked about current antidepressant prescribing practice, and for the clinicians to describe their first line and second line antidepressants, and usual doses pre scribed.It asked about the rationale for choosing particular antidepressants, whether seven specific factors had influenced choice, and for the clinicians to rank the three factors most influencing their decision.Attitudes to discuss ing treatment options with patients and to involving patients in treatment decisions were sought.

Findings
Of the 27 questionnaires sent out, 22 were re turned, a response rate of 81.5%.The respon dents included seven consultant psychiatrists, seven post-membership senior registrars/ registrars and eight clinical assistants.There were nine female and 13 male respondents.The mean frequency of initiating a course of antide pressants was 2.12 per week per professional.Two psychotherapists reported never prescribing antidepressants and have been excluded from parts of the analysis.Of the 20 prescribing respondents, 15 reported that lofepramine was their first-line antidepressant, others reporting prothiaden (3), doxepin (1) and amitriptyline (1).Second-line drugs included fluoxetine (6), trazodone (3), paroxetine (3), prothiaden (3), clomipramine (2), amitriptyline (2) and doxepin (1).The most commonly reported dose regime was lofepramine 70-210 mg, four respondents pre scribing daily doses up to 280 mg.The usual daily doses of other drugs varied widely, mean maximal daily doses equating to amitriptyline 135 mg (comparative dosage data from Bazire, 1993).
On the reasons for choosing a particular antidepressant, 11 of the 20 respondents reported that seeing the benefits of prescribing certain Psychiatric Bulletin(1994), 18, 597-599 drugs, while finding others ineffective, had been important.Most respondents were influenced by the seven specified factors, with the exception of cost and drug company advertising (see Table 1).Their ranking of the first, second and third most important factors was converted to scores of 3, 2 and 1 respectively.This led to a cumulative 'influence score' for each of the seven factors.The highest scoring factor was the differing side effect profiles of antidepressants, followed by the dan ger of antidepressants in overdose, and adverse personal experience of using certain drugs.Other factors were considered less important, drug company advertising considered one of the three most important factors by only one respondent.
Discussing treatment options with patients presenting with depression (i.e.drug treatment, ECT or psychotherapy) was regarded positively by 19 of 22 respondents.However, the idea of discussing the choice of antidepressant drugs, based on facts about them, was less popular, 12 of 22 respondents being in favour.

Comment
This study surveyed the practice of a group of 22 psychiatrists regarding their treatment of de pression.The response rate was reasonable for a postal questionnaire but refers purely to practice in a single, urban mental health trust.It is hard to know how representative this is.The reported rate of prescribing antidepressants indicates that each professional was responsible for initiating over 100 treatment courses per year.
Interestingly, these clinicians prescribed mainly the newer, more costly drugs.Cost and drug company advertising were considered relatively unimportant in determining choice of antidepressant.The cost of prescribing different antidepressants varies by up to one hundred times (Eccleston, 1993), and it seems unlikely that psychiatrists will remain unconcerned by such factors, given the current political climate and that 2.5% of the total NHS budget is spent on antidepressants (Slater, 1992).If doctors are influenced so little by advertising, it would be difficult to understand the amount drug companies spend on it.
These professionals varied in the extent they reported involving patients in decisions about treatment.Respondents were in favour of dis cussing the differing forms of treatment but less so involving patients in deciding the choice of antidepressant.Worrall (1989) advocates a formal model 'clinical decision analysis' for such communication with patients.Given widely dif fering practices, some considered inadequate by leading professional groups, it may be difficult to find a meaningful and consistent way of communicating such information to patients.
Reports of preferred antidepressant drug and dose varied.Fifteen professionals chose lofepramine as first-line treatment and of these seven reported the same dose range.Given the number of antidepressants available, the pre scribing pattern found was remarkably uniform.In over half, the usual upper limits of doses fell below amitriptyline 150 mg per day or equivalent -the dose recommended as the minimum effectivefor treatment.The prescribing practices of senior and junior psychiatrists were similar, and it is suggested that undertreatment of depressed patients may be occurring even by experienced psychiatrists.Alternatively, these psychiatrists may have found lower doses to be as effective in their clinical practice.
Over half of the respondents spontaneously volunteered that personal experience of efficacy had been a determining factor in the choice of antidepressant, often citing experience of having used ineffective drugs.By contrast, the psychi atric literature suggests no difference in efficacy between major antidepressants.'Modelling' of the prescribing behaviour of senior training psychiatrists also influences practice, as does observing trends in the psychiatric literature, in particular the overdose toxicity of drugs.
This study has demonstrated variety in psychiatrists' prescribing of antidepressants.Choice appears to be based on many factors, perhaps reflecting the experiences and person alities of the clinicians as much as scientific evaluation.Despite encouragement in the litera ture for a more uniform approach to treating depression, it seems that the process remains an individual one based on the relationship between doctor and patient in which antidepressants are only one of a number of tools available.This area of psychiatric treatment remains as much an art as a science.

Table 1 .
Factors affecting antidepressant prescribing practice