Friday, June 29, 2007

Statistics is scary!

So as I promised, I will be discussing the controversies around lung cancer screening trial lead at NYP and now many other places around the world by Dr. Henschke. So Traditionally to evaluate the merits of a treatment/screening trial (I will be focusing on screening here), clinicians set up what is called a Randomized Control Trial (RCT). In an RCT people are divided randomly into two groups: one which will be screened and one which will receieve either no screening or another form of screening. Clinicians design a regimen of screening for both groups, then screen them for a specified period of time and later on perform a follow up. The efficiency of the screening will be evaluated by what is called the rate of mortality reduction due to screening, which is the number of people who will die of a particular disease of interest over the total number of deaths (in the period of screening).
RCTs have a lot of merit in clinical treatment trials, however, they do not seem to be very informative and efficient as far as screening is concerned. I am simplifying so much here but what I just said seems to be the matter of controversy in the medical community. In the 1970's three RCTs, Mayo Lung Project (MLP), Memorial Sloan-Kettering Lung Project (MSKLP), and Johns Hopkins Lung Project (JHLP) looked at the effect of screening people with a high risk of developing lung cancer with chest radiograph (CXR) and sputum cytology. The MLP (the largest of all three and funded by NCI) after 6 years of screening and about 14 years of follow up, concluded that screening with CXR does improve the survival but does not reduce mortality rate compared to the controled groups. In other words, CXR may be detecting a lot of biologically cancerous, but clinically benign cases that will improve the survival but it is not actually saving lives and that's why the difference in mortality between the two groups did not turn out significant. This result was followed by same results from MSKLP and JHLP. As a result mass screening for lung cancer, at least for individuals with high risk of developing it did not turn into public policy (as opposed to other cancers screenings such as breast cancer or cervical cancer).
So here is the question: it makes intuitive sense that the sooner you catch cancer, the better your chances of treating it. So why the the above RCTs prove the opposite? It's also worth mentioning that a few years ago the same controversy was involved in questioning the merits of annual mamography for women to screen for breast cancer. In this case again a few RCTs disprove the value of mamography screening while a lot of clinicians disagreed with the results. So where's the mystery? What seems to be inconsistent here? stay tuned.......:)

No comments: