Wednesday, April 15, 2009
A short note on screening for cancer
My sister, a biologist and cancer survivor, has some choice words for cancer screening tests. They are not what you might expect after that set-up.
My own view -- and I also have some relevant, if dated, professional experience in the field -- is that we overscreen for certain cancers in certain populations and underscreen for others. It is not obvious, for example, that the massive increase in expense to the system that has come from theoretically more accurate cervical cancer screening in the last decade has actually had an impact on cervical cancer rates, morbidity, or mortality. But I'm not going to let that opinion get in the way of my regular colonoscopy or dermatological exam.
3 Comments:
By 
Escort81, at
Wed Apr 15, 10:12:00 PM:
PSA tests have always had a fair amount of Type I and Type II error (false positives and false negatives). I have to say, from kind of an "Oprah" standpoint, it always feels good to have a low number! Moreover, it was always a good thing when my nonagenarian father receives a low score, and it is only this year that we have discontinued it, after consulting with his urologist. An exam in still indicated, however (nobody's favorite moment).
There are significant differences between the two studies your sister cites -- one is American, the other European -- the American study uses quite a short time frame, and is not terribly useful to clinicians who are used to "watchful waiting" with issues such as BPH. The European study had somewhat different results, and most urologists would still like the PSA test to be done.
By Charlottesvillain, at
Thu Apr 16, 11:00:00 AM:
I'm willing to consider the somewhat counterintutitive results of this research if it will keep the doctor's big knuckle out of my ass.
By 
, at
Thu Apr 16, 05:02:00 PM:
The massive costs for cervical cancer don't justify themselves based on mortality for the same reason the massive cost of breast cancer screening doesn't. Time of diagnosis is essentially irrelevant to mortality, what matters is how fast the tumor grows. If you have a fast growing tumor you're going to die very soon, because even if you eliminate the cancer down to a handful of cells, what's left will rebound back into a full-sized tumor within 12-18 months. If you have a slow-growing tumor, one that was diagnosed in your 60s and has been slowly growing for the last 30 years, once reduced to a handful of cells won't come back for another 30. You'll be dead of something else long before the cancer becomes a problem.
We spend massive amounts of money on cervical and breast cancer screening because of sexism, not demonstrated effectiveness in improving mortality.
