Abstract
Christopher P. Alderman, Repatriation General Hospital, Adelaide, Australia:
In a recent review of issues of legal liability relating to prescribing [1], Beerworth and Tiller remind us that the principles of evidence-based medicine should guide the choice of drug treatment offered to patients. Although the authors describe this review as a study, the quality of some information presented falls well short of that required for application in the style of systematic review used in evidence-based, decision-making processes. The result is a review that may create doubt and confusion for many readers, without providing robust evidence to clarify the issues examined.
The quality of the information presented by Beerworth and Tiller merits close scrutiny. Citing evidence from a meta-analysis as a means to demonstrate the clinical superiority of one form of treatment over another is fraught with difficulties. The results of a meta-analysis are profoundly influenced by the processes used to select trials for inclusion. Negative studies, such as those sponsored by interested parties but which do not demonstrate the superiority of new antidepressants relative to older agents, are unlikely (or at least less likely) to proceed to publication, and are thus unlikely to be included in a meta-analysis. A recent meta-analysis of the safety, efficacy and toler-ability of antidepressants in the elderly [2] revealed no advantage for newer antidepressants over tricyclic antidepressants (TCAs), even though this patient subgroup is one where tolerability advantages could be expected to significantly influence therapeutic outcomes. This finding underscores the importance of study selection in shaping the results of a meta-analysis.
Throughout the review, particular references are repeatedly cited to support the views of the authors. Examination reveals that several references quoted in multiple sections of the review are actually letters to the editor of medical journals, or unreferenced editorials/reviews. Indeed, at least one of the letters cited as a reference was written by the regional medical director of a pharmaceutical manufacturer marketing a popular new generation antidepressant. This type of information may represent opinion, but would not necessarily be regarded as evidence.
The authors of the review have used imprecise terminology (e.g. ‘cardiac patients’), and in some cases have not accurately reflected the findings of studies cited as references. For example, a study by Veith et al. [3] has been cited to illustrate ‘concern expressed in medical journals about the use of TCAs with cardiac patients.’ In fact, the investigators in this study actually concluded that in patients with heart disease, imipramine had no effect upon left ventricular ejection fraction, and that premature ventricular contractions were actually reduced during imipramine therapy. When dealing with a patient with heart disease, a discerning prescriber would be well advised to specifically consider the precise type of cardiac condition present, as in some cases certain types of antidepressant may be preferable.
A range of issues have also received unbalanced coverage in the review. The potential for a withdrawal syndrome with TCAs is highlighted, but the selective serotonin re-uptake inhibitor (SSRI) discontinuation syndrome (recently discussed in a review by Tiller [4]) is ignored. No mention is made of the dramatic increase in sedative/hypnotic prescribing which has accompanied the introduction of the SSRIs and a reversible inhibitor of monoamine oxidase A into the Australian market-place [5]. Perhaps most misleading of all is the presentation of data relating to the community use of antidepressants in Australia. Usage data for the newer anti-depressants in 1990 and 1991 have been included in the review, but bear little relevance to the discussion (as most of the newer antidepressants currently available were yet to be marketed in Australia at that stage). Moreover, although the review was revised and accepted for publication in March 1998, the most recent utilisation data presented in the manuscript were from 1994, when the use of TCAs considerably exceeded that of SSRIs. At the time that the manuscript was revised, data from 1996 were available [6], and clearly demonstrate that the usage of SSRIs (measured in terms of the number of defined daily doses per 1000 population/day) exceeded that of TCAs at that time.
The review contains considerable discussion of the lethality of the TCAs and monoamine oxidase inhibitors in overdose and the relative advantage of the newer drugs in this regard, but even this issue is not clear cut. When a prescriber assesses suicide risk a range of factors must be considered. For example, in a large study of suicide among patients prescribed antidepressants, self-inflicted violence (44.7%) was the most commonly used means of suicide, and carbon monoxide poisoning accounted for 18.1% of suicides in the cohort [7]. When compared to dothiepin as an index agent, fluoxetine was associated with a higher relative risk for suicide in this study [7]. Even more illustrative is the observation that although the new generation antidepressants were introduced in Australia in the early 1990s, we have not observed a precipitous decline in the suicide rate since that time.
Evidence-based medicine has been described as the explicit and judicious use of current evidence in making decisions about individual patients, integrating individual clinical experience with the best available external clinical evidence from systematic research [8]. This description rightly emphasises that the clinical circumstances surrounding each individual case should play a critical role in the decision-making process in medical care. In asserting the potential for liability on the part of a prescriber for not selecting one particular type of antidepressants preferentially, Beerworth and Tiller create the impression that liability in prescribing choice can be reduced simply by favouring a particular type of medication while avoiding others. In fact, a medical practitioner has a duty to exercise reasonable care and skill in the provision of professional advice and treatment [8]. In the case of antidepressant prescribing (or the prescription of any medication), reasonable care is exercised when the clinician objectively analyses all available treatment options, integrating knowledge of the characteristics of the drugs involved with consideration of the individual circumstances of the patient.
In selecting antidepressant therapy for a patient a clinician should consider a range of factors, including the history of previous antidepressant drug response, previous allergies or hypersensitivity reactions to antidepressants, the adverse effect profile of the drug, the toxicity of the agent in overdose, the likelihood that the patient may indeed take an overdose of the drug, concurrent medical and psychiatric conditions, potential for drug interactions, and practical issues such as simplicity of compliance and cost. Superimposed upon these considerations is the need to engage the patient in the decision-making process, taking into account the acceptability of the treatment to the person who will actually take the medicine. The clinician who can demonstrate that these factors have been considered carefully, and with due deference to unequivocal scientific evidence, is most unlikely to attract liability in prescribing choice.