What We Talk about When We Talk about Randomized Controlled Trials

criteria usually adopted by the main RCTs on COPD treatment. It really is questionable whether the results from RCTs based on such selection criteria can indeed be extended to a larger ‘real-life’ population of patients with COPD. Herland et al. [3] studied a large population of patients with obstructive lung disease who were classified as having asthma (38%), COPD (42%) or were included in a mixed group (20%). In this study, absence of comorbidity, a FEV 1 50–85% of predicted value, present or historical reversibility, being either a nonsmoker or an exsmoker with a smoke burden 15 pack-years) and absence of atopy were the selection criteria they considered. Only 17% of the COPD patients were eligible for the RCT. In line with these results, Travers et al. [4] found that in a group of 55 COPD patients undergoing treatment and identified by postal questionnaire and functional assessment, only a negligible percentage met the eligibility criteria of 18 RCTs cited in the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines. When we talk about randomized controlled trials (RCTs), we talk about rigorous interventional experiments that have been developed to control for possible biases that can be present in observational studies [1] . RCTs restrict the possible systematic errors by means of a well-defined and controlled setting, blinding, allocation concealment, randomization, strict inclusion and exclusion criteria and so on, thereby reaching a high internal validity, i.e. the ability to determine a cause-effect relationship. For this reason, RCTs fall into the trial category with an ‘explanatory’ approach according to the definition of Schwartz and Lellouch [2] . Well-designed RCTs carried out in large samples of a population provide international guidelines with the strongest category of evidence-based medicine. However, the same features that ensure the internal validity of RCTs can, on the other hand, severely limit their external validity, i.e. the ability to generalize the results in a clinical setting and in the general population. Notably, the strict inclusion and exclusion criteria followed by RCTs in the recruitment of the patients represent the chief limitation to generalizing their results. An advanced age, too-mild or too-severe airflow obstruction, no smoking habit, no long-term oxygen therapy, no other lung diseases and no comorbidities are the selection Published online: November 23, 2013