Assessing the use of retrospective databases in conducting economic evaluations of drugs: the case of asthma.

When evaluating drug substances, the traditional clinical study setting does not allow scope for observing real-life behaviour since all alternative actions are determined beforehand. However, a study based on prospective or retrospective databases containing real-life data can examine how patients and physicians behave in a real-world setting and can investigate the relationship between the introduction of a drug and the amount of healthcare used in actual practice. We reviewed the quality and potential policy application of published retrospective database studies in which an economic evaluation of the use of drugs in asthma was conducted. A search in literature databases found 16 such studies, which were reviewed and evaluated according to a published checklist. No article fulfilled all the criteria for a 'good' economic evaluation. The results of many of the evaluations may be informative, but not transparent enough to deliver policy conclusions. This may limit the use of the currently published retrospective database studies as a base for policy decision, compared with randomised controlled trials, despite the additional value of these database analyses when well conducted. A greater transparency when presenting material and results is therefore called for, to increase the usefulness of database studies.