The researchers did a computer-based search of the medical literature and collected all of the articles that reported on testing the effects of long-acting beta-agonists in asthmatic patients. To be included in this study, the articles had to report that patients were assigned to a group that received a long-acting beta-agonist or to a group that received placebo (an ineffective substance used in place of the active medicine). Assignment to one group or the other had to be done completely by chance. In addition, the study had to have lasted at least 3 months. After retrieving the articles, the researchers judged the quality of each study. They then analyzed the pooled results from all the patients they had been able to gather from all the studies combined and calculated the proportion of patients who had had 1 of the complications listed earlier.