Monday, November 03, 2014

Getting more detailed about predicting outcomes

According to Laura Landro, WSJ, Sept 15, 2014, physicians rely on studies and their own experience as well as gut feelings to tell patients how they will do on a given therapy. In the case of cancer, the disease is often "staged"--one through four--and therapy outcomes predicted from there.

But this does not take into account the type of cancer, the patient's age, gender, and other characteristics.

So docs are turning to more sophisticated methods. One of the physicians pioneering this, Michael Kattan, Cleveland Clinic, was diagnosed with cancer of the lymph nodes in his early 20s. He was lumped in with older patients in wheelchairs and on oxygen.

Now, he uses "nomograms," statistical models, using large databases from medical studies.

These nomograms are usually for doctors' use, but one on prostate cancer risk is on the Cleveland Clinic website. Still, a physician is needed to interpret it.

A friend recently found out she had the breast cancer gene--she is considering having her ovaries removed. I wonder if there is a nomograph to help make that decision.

I know when I had my detached retina bleeding into my eyeball--the blood oozing up and down like a lava lamp--the doc said I had a 90% chance of using the eye after the surgery. After the thing detached again, eventually requiring four surgeries, my eye is blind.

I often wonder if the doctor pulled that 90% out of his ...well, you know.

No comments: