Journal of Psychopharmacology 13(2) (1999) 136-143 C1999 British Association for Psychopharmacology (ISSN 0269-8811) SAGE Publications, London, Thousand Oaks, CA and New Delhi 0269-8811 [199905] 13:2; 136-143; 008435 Longitudinal patterns of antidepressant prescribing in primary care in the UK: comparison with treatment guidelines Rodney L. Dunn1, John M. Donoghue2, Ronald J. Ozminkowski3, Deborah Stephenson4 and Timothy R. Hylan5 'University of Michigan, Ann Arbor, MI, USA, PCS Health, Liverpool, UK, The MEDSTATGroup, Ann Arbor, MI, USA, 4Lilly Industries, Basingstoke, UK and Eli Lilly and Company, Indianapolis, IN, USA. The objective of this study was to determine whether patients beginning therapy on the most common tricyclic antidepressants (TCAs) and selective serotonin reuptake inhibitors (SSRIs) differed in their likelihood of having antidepressant treatment that was consistent with recommended treatment guidelines in the UK. An analytical file constructed from a large general practitioner medical records database (DIN-LINK) from the UK for the years 1992-97 was constructed. A total of 16 204 patients with a new episode of antidepressant therapy who initiated therapy on one of the most often prescribed TCAs (amitriptyline, dothiepin, imipramine and lofepramine) or SSRIs (fluoxetine, paroxetine and sertraline) were analysed. A dichotomous measure was defined to indicate whether subjects were prescribed at least 120 days of antidepressant therapy at an adequate average daily dose within the first 6 months after initiation of therapy. Only 6.0% of patients initiating therapy on aTCA and 32.9% of patients initiating therapy on a SSRI were prescribed antidepressant treatment that was consistent with treatment guidelines. After controlling for observable characteristics, patients who initiated therapy on a SSRI were much more likely (odds ratio=7.473, p<0.001) to have a prescribed average daily dose and duration consistent with recommended treatment guidelines within the first 6 months of initiating therapy than were patients who initiated therapy on a TCA. These findings suggest that initial antidepressant selection is an important determinant of whether the subsequent course of treatment is consistent with current national guidelines for the treatment of depression in the UK. Key words: antidepressants; depression; drug prescribing patterns; selective serotonin reuptake inhibitors; tricyclic antidepressants Introduction Antidepressant dose and duration are two important factors in achieving the goals of depression treatment which include not only symptom resolution, but also restoration of patients' functioning and prevention of relapse or recurrent episodes. Consensus guidelines have been issued by the British Association for Psychopharmacology (British Association for Psychopharmacology, 1993) and the Royal College of Psychiatrists (Royal College of Psychiatrists, 1992) for the management of depression with antidepressants. For treat- ment to be effective, they recommend that antidepressants should be prescribed at a dose shown by clinical trials to be effective in treating depression, and continued for at least 4 months beyond initial symptom resolution. These guidelines are broadly consistent with other consensus groups' recom- mendations (World Health Organization Mental Health Collaborating Centres, 1989; Agency for Health Care Policy Research, 1993). Most clinical trials of different antidepressants generally show equal efficacy (Effective Health Care Bulletin, 1993; Anderson and Thomenson, 1994; Song et al., 1994). However, studies of antidepressant use in clinical practice ('naturalistic' studies) have shown that most patients who initiate therapy on a tricyclic antidepressant (TCA) in primary care in the UK begin at suboptimal doses and remain there (Donoghue and Tylee, 1996; Donoghue et al., 1996; Martin et al., 1997). In addition, studies have also shown that most patients beginning treatment with a TCA are less likely than patients on selective serotonin reuptake inhibitors (SSRIs) to refill their prescrip- tions and often discontinue therapy by the third month (MacDonald, 1997). Naturalistic, observational studies of antidepressant use in clinical practice can complement the findings from controlled trials. A feature of observational studies is that the results are more generalizable to broader populations than those consid- ered in controlled clinical trials due to the lack of study conditions or protocol restrictions. However, unlike rando- mized controlled clinical trials, observational studies are subject to the influence of confounding factors. Previous observational studies of antidepressant prescribing patterns in primary care in the UK have been limited by their cross-sectional design 1 89-%58 0010022