Screening for Disease: Making Evidence-Based Choices

Abstract

Screening for illness should be an evidence-based activity. Screening tests are useful only if they reduce mortality or morbidity. Therefore, healthcare professionals must know how to evaluate research about screening tests to be sure that, in fact, the tests actually accomplish their goals. Tests that generate many false-positive results may cause harm from anxiety and unnecessary procedures. Tests that generate many false-negative results may worsen outcomes by leading to delayed diagnosis and treatment. Characteristics that make a disease amenable to screening include a significant negative impact on health, an identifiable asymptomatic period, and improved outcomes with early intervention. A useful screening test must have sensitivity and specificity for the disease being screened. It also must be cost effective and acceptable to patients. Sensitivity, specificity, and disease prevalence all interact to determine a test's positive predictive value--the likelihood that a positive test result indicates that the disease is present. Several types of test bias can undermine the validity of a screening trial. Screening bias occurs when the sample of patients used in a trial to evaluate a screening test is not representative of the patient population to be screened. Another bias results from the fact that indolent disease is more likely to be detected in a screening program than aggressive disease. The apparent improved outcome that results is called length bias. Finally, lead-time bias occurs when survival of a screened population is measured from the date of screening, whereas survival of an unscreened population is measured from detection of symptomatic disease. In screening for illnesses, the goal must not be merely to do something. It must be to do something useful.