Sunday, January 25, 2009


this post is in response to another post by a real medical writer friend regarding screening tests for certain diseases. see the related post here.

basically, the question is, when should you start looking for diseases in people even if they have no symptoms (hence, screening)? a reader on the other post commented about ovarian cancer. she is a little on the younger side for women who get the disease; furthermore, it seems she actually had symptoms which is a separate issue. anyway, the reader suggested getting ultrasounds on all women but the question about screening turns out to be a lot more complex than it seems.

the first problem starts with the disease itself. the earlier you catch something, the harder it is to tell if it's a real problem. is that "spot" on your lung something that will one day turn into cancer or will it just stay a "spot"? is that cyst on your kidney or ovary just a cyst or are there features that make it worrisome for something more sinister?

then it gets into the realm of the actual test and interpretation. people tend to think a ct scan or an ultrasound is foolproof. but the fact is, there is no test in medicine, blood work, scan, algorithm, prediction score, you name it, that is 100% accurate. NOTHING. not even a pathologist examining something under a microscope is 100%. sometimes things that look like cancer turn out to be nothing and things that appear harmless turn out to be malignant. and not to malign pathologists, but radiologists are probably even more notorious for disparate readings. "impression: normal ct of the aorta. addendum [days later]: aortic dissection noted." don't get me wrong, these guys are smart, usually among the smartest in their class but we all miss subtle and not-so-subtle findings.

okay, so far you have a disease you're not sure is there, a scan that you're not sure how accurate it is in picking up the disease, and a radiologist who might be reading the scan incorrectly but it gets worse!

unfortunately, then the economic side of things rears its ugly head. let's use the ovarian cancer example. this is unfortunate. we've all seen it. i've seen metastatic colon cancer in patients in their 20's although it's a disease that you typically screen for when you're 50. the question then is, when do you start screening for ovarian cancer? no matter what age you start (heck, unless you start in prepubescent kids), you will probably miss some cases. there are always outliers. so, in order to catch that one in a million patient, can you justify starting screening at, say, age 30? 25? 20? who is going to pay for all this?! that, in today's economy and presidential administration, is the million dollar question. these tests can get expensive and you simply cannot screen all people for everything without completely obliterating the finances of this country.

so you start ovarian cancer screening at age 20. then how frequently do you repeat the testing? when is it safe to stop? how do you know when to act on some small finding? this now gets into the criteria needed for a screening test to be a good one. among them, the test has to be pretty accurate (namely, good "sensitivity," meaning you minimize the number of patients who have the disease but have normal tests, i.e. "false negatives"), the test isn't too costly, there are minimal side effects from the test itself and many other criteria.

then there's the whole question of whether the disease would have killed you anyway. now, in the patient with metastatic ovarian cancer, this is not as relevant since, again, it seems the patient had symptoms already. another dirty little secret: cancers detected incidentally or by screening that are not causing any problems do not always need to be treated. now that statement at face value sounds very controversial but the fact is, it's true. check out this controversial study. furthermore, take for example prostate cancer. unfortunately, many men will die of prostate cancer but many more men will die of heart disease and other cancers and stroke. frequently, some screening blood test or exam will reveal some abnormality which then requires biopsy or some other type of treatment that frequently will lead to incontinence or impotence or a myriad of other problems when there potentially may have been little or no benefit to the screening from the beginning. this is why there is such controversy in the realm of prostate cancer screening.

i think that's enough about screening but the thing is, so much more could be written. you see why it's such a complex topic when it seems so simple. the fact is, at least with cancers, some patients will get a fairly aggressive form and at a much earlier age than expected. most, however, will not and right or wrong, the policy is to try to implement the appropriate screening programs that are most cost-effective to benefit the most number of people while leading to the least amount of harm. primum non nocere--first, do no harm.