Overdiagnosis in Medicine
I recently saw a book titled Overdiagnosis. It discusses in some detail the effects of lowering the cut-off for a number of tests including PSA, glucose, TSH, cholesterol and blood pressure.
As you know, the cut-off is that value for a diagnostic test that determines whether the test is positive or negative. As you also know, it is possible that a test can be falsely positive or falsely negative. The lower the cut-off, the more false positives there will be. There is no way to get around this. There is no question but that more people will be diagnosed by lowering the cut-off. The question raised in the book is whether the diagnosis of, say, diabetes when the ‘fasting' glucose is 112 is a good thing for the a possible diabetic and what are the effects, if any, on the many persons who are not diabetic but have a glucose of 112 or 113 at that time.
In 2008 the NIH stopped a study comparing outcomes of patients whose glucose was kept "normal" while another group was kept below 140 mg/dL. The study was stopped because the group kept within the ‘normal range' had a 25% greater risk of dying than the less tightly controlled group.
There is a lot to be said about lowering limits as these examples suggest. What do you think?