Chest pain is common: one in four of the population have an episode annually. Of those who present to hospital, nearly two-thirds have noncardiac chest pain. More than half of these cases might have gastroesophageal reflux disease. Opinion differs over what is the most appropriate application of current investigatory methods. Evidence suggests that, once cardiac disease is ruled unlikely, empiric use of a proton pump inhibitor is an option; if acid suppression fails, detailed investigations as clinically indicated can be considered. A range of esophageal investigations is available, including 24-hour or 48-hour esophageal pH testing and esophageal manometry, as well as provocative tests, but there is no consensus as to which methods are the most useful. Psychiatric evaluation is not routine, but psychiatric or psychological disorders are common. Musculoskeletal disorders are also common, but are frequently overlooked. It is possible to subject patients to a comprehensive set of investigations before empiric therapy, but recent studies have failed to demonstrate an improved outcome using this exhaustive approach. A new tactic is required, with less attention spent on absolute diagnostic accuracy and more emphasis on optimizing the long-term clinical outcome in patients with noncardiac chest pain. It is possible that the targeted use of multiple drug trials in a policy of 'therapy as investigation' might be a superior methodology.