There are numerous books on Survival Statistics. Here's a good one. And this one that covers trials that end before all patients have the disease. The hazard rate is the failure rate, and it's the key opposite of the success rate. It has simple statistics to determine it's departure from randomness and estimated confidence intervals. I don't like to use cox regression as it's a mixed bag of log, and multiple comparisons that isn't meant for the distributions we deal with. Nor does it do anything but add another 3 names to the 30 average names in all medical papers these days. It obscures the expected life expectancy from the treatment which should always be the end point rather than the phony stuff about succumbing to disease or not. Similarly the expected duration from failure to success which is in the program that so many on this list have relieved from me, is not a very good measure either as who cares how long if the expectations don't jive.





Speak your mind


Resources & Links