Research studies on screening tests

Abstract

Screening tests are done to diagnose asymptomatic disease in apparently healthy people with the aim to reduce mortality and morbidity from the disease. Certain criteria need to be fulfilled before we adopt population-level screening for any disease. Several biases exist in evaluating screening studies, and the ideal study design would be a randomized trial with hard endpoints such as mortality and morbidity.

Keywords: Diagnosis; mass screening; research methodology.

Figure 1

Lead time bias. X-axis represents time in years. Two scenarios, namely where the disease is diagnosed in the usual clinical situation (after the occurrence of symptoms; top) and when the disease is diagnosed by screening (bottom), have been compared. Even though there is no real change in survival following early detection of disease by screening, the perceived duration of survival is increased, due to a lead time effect

Figure 2

Length bias. X-axis represents time. Y-axis represents tumor size or stage. The disease onset is at zero and each oblique line represents tumor growth in an individual, with lines on the left for persons with rapid tumor growth and those on the right and bottom for those with slower tumor growth. The persons with rapid growth are diagnosed based on symptoms (D) and die (M) rapidly. By comparison, the individuals with slower growth are more likely to be diagnosed where they get screened (S), and they would live longer