At the risk of sounding self-righteous, there has been a discipline that
has offered solutions to decision making in the presence of uncertainty
for more than a century now. That discipline is Statistics.
While some of the abuses associated with indication creep and prevention
creep have been caused by naive use of statistics, a careful
consideration of the costs and probabilities associated with false
negative and false positive findings produces a simple solution to the
problem. If there is a problem with wasteful medical practices, it has
been because doctors have been using too little statistics.
Statistics also offers solutions to the problems of targeting
individualized care to a patient with specific risk factors. That work
is being put into practice by some of my colleagues at Saint Luke's Mid
America Heart Institute:
It's also possible to adjust for individual patient's perceptions of the
costs and benefits of particular treatment options, but this done less
frequently. With Statistics, you can, for example, develop methods for
eliciting individual perceptions of the tradeoffs between length of life
and quality of life. There's some empirical data (I can't find the
reference right now) that suggests that doctors place too much emphasis
on quality of life relative to what their patients want.
Many sources of uncertainty are difficult, if not impossible to
quantify, but even here Statistics can play a big role. There's a lot of
work out there that I like to think of as meta-research--research about
the research process. So we learn a lot from the studies by Hrobjartsson
debunking the myths about placebos, by Concato showing that
observational studies perform much better than expected when matched up
with randomized trials, by LeLorier contrasting meta-analysis and large
scale randomized trials, and a whole bunch of work by Ioannidis. All
this work helps sharpen our understanding about how to apply research
findings in a real world context.
We don't need a bunch of math geeks for doctors, but if our medical
schools insisted on a decent level of quantitative literacy integrated
throughout the curriculum (rather than segregated into a single
throw-away class), we'd avoid many of the problems noted by Drs.
Djulbegovic and Paul.
Steve Simon, [log in to unmask], Standard Disclaimer.
Sign up for the Monthly Mean, the newsletter that
dares to call itself average at www.pmean.com/news