Since the advent of tricyclic antidepressants, the treatment of depression has been dominated by pharmacotherapy. This seems set to continue following the introduction of new drugs, particularly the selective serotonin reuptake inhibitors. These pharmaceutical products are widely accepted to have at least equivalent efficacy as tricyclic antidepressants, yet have a lower risk of toxicity and adverse side-effects, (!J.,12.) leading to potentially greater compliance and consequent reductions in the use of health and other services. Against this, the newer drugs carry considerably higher acquisition costs. There is consequently an important cost-effectiveness question to be considered: are the higher acquisition costs worth paying in terms of the reduced toxicity, adverse side-effects, and need for service inputs? A variety of study designs and methods have been used to address this cost-effectiveness question, specifically (a) prospective studies, (b) retrospective studies, and (c) simulation studies.
The sole example of a prospective randomized controlled trial with an economics component (in fact, a cost-consequences design) compared the costs and outcomes of fluoxetine (a serotonin reuptake inhibitor) with imipramine or desipramine (tricyclic antidepressants) as first-line drug treatment for a depressed sample of American primary care attendees (n = 536, two-thirds of whom met diagnostic criteria for major depression'13'). Six-month assessments of symptoms and quality of life revealed similar improvement in all three groups, whilst overall direct treatment and care costs were also not significantly different. Informal caregiver support and the indirect costs of depression (lost employment, reductions in well being) were not costed. This inconclusive result led the authors to suggest that the choice of antidepressant should be left to preference, with the implication that serotonin reuptake inhibitors would be most preferred (the discontinuation rate over 6 months for fluoxetine was 10 to 20 per cent lower than for the tricyclic antidepressants).
A number of quasi-experimental studies have been conducted^. 15 and 19 which suggest the superior cost-effectiveness of serotonin reuptake inhibitors over tricyclic antidepressants. For instance, Forder et al.(l4> found that treatment with sertraline produces better outcomes in terms of reduced symptoms of depression yet the costs are marginally lower. However, the usual limitations of retrospective analyses (selection bias and the presence of other potentially confounding factors) are apparent in these studies, leading to a reduction in the confidence with which results can be held. Also, lost productivity costs were not included, although Forder et al. did account for informal care costs.
Decision-analytic techniques have been employed to 'simulate' the costs and outcomes of depression, as opposed to measurement on the basis of actual clinical practice. Data on clinical efficacy and tolerability have been largely based on meta-analyses of randomized control trials, and I9) whilst data on service use and their attendant costs have typically been based on single studies or panels of experts. Although the majority of studies were favourable to the serotonin reuptake inhibitors, the methodological limitations of this simulation approach again impose considerable uncertainty on the findings. The simulation study that has attracted most interest (and controversy) is that by Jonsson and Bebbington,(6) who compared the cost-effectiveness of treatment with imipramine (tricyclic antidepressant) versus paroxetine (serotonin-reuptake inhibitor). They concluded that the cost per successfully treated patient was 20 per cent lower for paroxetine than for imipramine. Yet this finding has been contested by Woods and Rizzo, (29 among others, who revised a number of key assumptions used in the model and arrived at a reversed conclusion, namely that the tricyclic antidepressant was more cost-effective than the serotonin reuptake inhibitor.
Many of the studies indicating the superior cost-effectiveness of serotonin reuptake inhibitors over tricyclic antidepressants have adopted decision-analytic or retrospective designs, which carry important methodological limitations. Together with the studies that suggest that tricyclic antidepressants are equally or more cost-effective in comparison with serotonin reuptake inhibitors, including the only prospective controlled cost-effectiveness trial, it is only possible to conclude that there is no definitive evidence of superior cost-effectiveness for either treatment strategy. A (guarded) conclusion reached by Woods and Baker (2J) in their review of evidence is that first-line use of serotonin reuptake inhibitors may not be cost-effective from the perspective of the health-care system, and that there is insufficient evidence to address the question from a societal perspective. Hotopf et al.(22) also conclude that there is no evidence to suggest that serotonin reuptake inhibitors are more cost-effective than tricyclic antidepressants, and argue that the debate will only be concluded when a prospective cost-effectiveness study is undertaken in the setting of a large primary-care-based randomized control trial.
Was this article helpful?