The usefulness of the likelihood ratio in the diagnosis of dyspepsia and gastroesophageal reflux disease.

Clinicians are familiar with the concepts of sensitivity and specificity to describe the accuracy of a diagnostic test. These measures do not always express the probability that a patient has a disease with a given test result as this will vary with the prevalence of the disorder in the population. The likelihood ratio is a more clinically relevant method of reporting accuracy, and the probability of having a disease after a positive or negative test can be calculated. The likelihood ratio can be applied to the clinical problem of dyspepsia management. This suggests that Helicobacter pylori (H. pylori) test and treat will detect and treat most peptic ulcers with only 0.5% of H. pylori-negative patients having peptic ulcer disease. Serology is possibly acceptable in populations with an H. pylori prevalence of approximately 50%. The urea breath tests are more appropriate in more extreme prevalence ranges. Once the prevalence of H. pylori falls below 10%, then the urea breath test becomes inaccurate, and screening and treatment may be less appropriate. The absolute probability of having peptic ulcer disease in a largely H. pylori-negative population will be very small, however, and the appropriateness of performing any investigation in these circumstances is debatable. Finally, likelihood ratios indicate that the clinical diagnosis of gastroesophageal reflux disease (GERD) is not straightforward. Traditionally, it is believed that patients with dominant heartburn are likely to have GERD. Likelihood ratios predict, however, that patients with these symptoms have a little over 50% chance of having GERD as defined by 24-h esophageal pH studies.