I recently read a great article demonstrating biases in the evaluation of screening tests. This is timely as last week I had a 75 year old female visibly upset that she no longer needed regular pap smears. She sighed and said “I guess no one cares if I die of cervical cancer.”
Cancers have different growth rates—which determines their potential to be detected by screening. Also there are different characteristics of the disease. How fast does the cancer grow? Might the early-stage abnormalities regress on its own (without treatment)? Is there effective and acceptable treatment available? Are patients asymptomatic for a time during which detection and treatment will significantly reduce morbidity and mortality? Do we have an effective screening test during the time that we could “catch” the problem in time to save their life?
What are the characteristics of the screening test? Is the test sensitive enough to detect the disease during the asymptomatic period? Is the test specific enough to minimize false positives (a false positive test shows you have the disease, but indeed, you do not)?
What are the characteristics of the screened population? At what age is which screening test appropriate? (Should we test 10 year-olds for colon cancer?) Are patients willing to comply with subsequent tests and therapy if needed?
What does all this mean? Researchers are pooling studies to give us more insight into which tests should be performed in which patients. This is so that there is less chance of false positives (which leads to patient worry and further work ups) and more chance to find those most at risk. I think medical school ingrained in me a perpetual weighing of risks and benefits.