BACKGROUND: In the last 10 years the use of antidepressants has increased drastically. Unfortunately, the epidemiology of these compounds has shown significant gaps between recommendations derived from randomised controlled trials and current clinical practice. METHODS: We argue for the need to develop and maintain clinical databases of patients receiving antidepressants as a way of bridging this situation. RESULTS: In addition to experimental data generated in selected patients and settings, observational databases of large cohorts of typical patients, followed in typical settings, should be developed and maintained. Clinical databases could collect information on patient social and demographic characteristics, clinical symptoms, diagnosis and pharmacological and non-pharmacological treatments. In addition, they can provide accurate estimates of probabilities of different outcomes and on factors that affect outcome. CONCLUSION: Clinical databases should not be seen as another expensive administrative task for busy doctors. Clinical databases should be developed, organised and utilised only by clinicians who are interested in monitoring their clinical practice and want to provide patients, relatives and the public with information on prognosis and outcome in their specific context of care. Maintaining clinical databases is a routine process, nested in everyday clinical activity, which aims at constituting a permanent link between research and practice.
BACKGROUND: In the last 10 years the use of antidepressants has increased drastically. Unfortunately, the epidemiology of these compounds has shown significant gaps between recommendations derived from randomised controlled trials and current clinical practice. METHODS: We argue for the need to develop and maintain clinical databases of patients receiving antidepressants as a way of bridging this situation. RESULTS: In addition to experimental data generated in selected patients and settings, observational databases of large cohorts of typical patients, followed in typical settings, should be developed and maintained. Clinical databases could collect information on patient social and demographic characteristics, clinical symptoms, diagnosis and pharmacological and non-pharmacological treatments. In addition, they can provide accurate estimates of probabilities of different outcomes and on factors that affect outcome. CONCLUSION: Clinical databases should not be seen as another expensive administrative task for busy doctors. Clinical databases should be developed, organised and utilised only by clinicians who are interested in monitoring their clinical practice and want to provide patients, relatives and the public with information on prognosis and outcome in their specific context of care. Maintaining clinical databases is a routine process, nested in everyday clinical activity, which aims at constituting a permanent link between research and practice.