Baseball & Breast Cancer

A certain confluence of reading came across my desk over the past two days, leading me to make a connection that I otherwise likely would have never considered. For some reason, I got very interested in looking up a book that Buzz Bissinger wrote about Tony LaRussa and the St. Louis Cardinals a few years ago, Three Nights in August, and while it contained some incomprehensible analogies, it was a decent enough book, which I finished yesterday evening. Written a few years after Michael Lewis’s seminal Moneyball, much of the editorializing Bissinger indulges himself in is refuting the ‘bunk’ that sabermetricians like Bill James have more or less proven statistically, a position that websites like the recently retired Fire Joe Morgan have ridiculed for its ignorance.

In this morning’s New York Times Sunday Magazine, mathematician John Allen Paulos illustrates why the government’s task force on breast cancer screening’s recommendation that asymptomatic women under the age of fifty need not undergo mammograms is mathematically a sound one. As anyone who even marginally keeps up with the news already knows, this caused a furor and was quickly sucked into the ever declining discourse on national healthcare. Despite what seems counterintuitive, tests for a relatively rare condition can have a false positive rate of about 1%, which can be quite misleading when the rate of an occurrence is less than 1%. I won’t go into his math, but it’s a short article and easy to understand.

And Paulos is right to remind us that most people don’t think probabilistically, nor do they respond correctly to very large or very small numbers. Think of the uproar ten years ago when mice being force fed aspartame in relatively gargantuan amounts got cancer. It doesn’t mean that other people might not come up with different conclusions using similar data, but to argue against the recommendations from the government panel, one must present facts in his or her argument, not merely use invective.

And this brings me back to Bissinger and statistics. I understand that many things the so-called Moneyball people say seems counterintuitive, and so the average person might be quick to reject them because they just don’t seem to be possible. Take for instance the argument that the player with the best on base percentage on a team be batted in the leadoff position because over the course of the game, that player will have the most at bats. The rest of the lineup would be structured accordingly, the player with the second highest on base percentage being batted next, etc. This makes sense on the surface because it is impossible to score runs without getting people on base, and it is impossible to win games without scoring at least one run. Having the most people on over the course of night should lead you to have the greatest ability to score runs and give you the greatest chance to win the game.

So why doesn’t anybody do this? Well, baseball managers and GMs don’t think probabilistically much more than the typical person. But let’s say they did. For as long as anyone can pretty much remember, the idea has been for a team to try and get a couple of quick men on base before the third-fourth-fifth part of the lineup, the three players with the most power, come up to try and knock them in. It doesn’t seem to matter that over the three games covered in Bissinger’s book, Kerry Robinson (OBP .281 in 2003) led off all three games while Albert Pujols (.439), one of the greatest players to ever put on a uniform, hit third, and thus got fewer chances at making an impact with his bat.

Let me run with this for a minute. Let’s say that LaRussa decided to buy into this strategy and placed Pujols as the leadoff hitter. Fans and the media, again like most of us not used to thinking probabilistically, would ridicule the move because even though it makes a certain amount of sense when you actually do the math, most people who listen to or broadcast on sportstalk radio aren’t doing the math. So unless it works, and works quickly, LaRussa (and the GM who let him do it) may be ordered to switch back or risk losing their jobs. But what would that result be based upon? Invective, and nothing more than the argument that ‘it’s never been this way before, so if this new plan was going to work someone smarter than LaRussa would have figured it out already.’

What is so frustrating to people like me, people who don’t think probabilistically but still understand that statistics are a science, is that the opposing argument not only adds nothing to the discussion, it is dismissive. Just as popular opinion has been against the recommendations of the breast cancer panel, it is against the sabermetricians who are looking at baseball in new ways. And while it may make a certain amount of populist sense to side with popular opinion, siding against science is always going to leave one standing in opposition to facts, a position from which it is very hard to win an argument.

Unless, of course, you can yell very, very loudly.

Better by Atul Gawande

If we wanted to save more patients’ lives in the medical system, is it more important to fund research that could perhaps find cures, or would it be more appropriate to invest time and money in improving the standards already in place? The tendency for us to say ‘more research’ is almost a given, but Atul Gawande, surgeon and staff writer for The New Yorker, argues that the later can have far more drastic effects.

In Better, Gawande explains that in medicine, as in nearly all human endeavors, variations in performance create a bell curve where most participants are merely at or below average. In this collection of essays, he studies this idea in the medical community and looks to find what separates the positive deviants from the rest.

Gawande writes about such the importance of hand washing, something one would think is a given in hospitals and doctor’s offices, yet shockingly staph infections in hospitals are transmitted to 30% of patients, a number that could be reduced dramatically using tools already in place. The doctor also covers ethics in medicine, from the use of chaperones when examining patients of the opposite gender and the role of doctors in capital punishment.

In covering medical interventions in slightly abnormal pregnancies, he makes a strong case that many caesarian sections are given when the old method of using the clamps on an infant would work just as well with an equal or better chance of complications. When studying the differences between a first class treatment center for cystic fibrosis and an average one, Gawande argues that the main difference is the ability of the medical staff to treat the person more than the disease and to be willing to think outside the box when it comes to diseases with which we have made relatively little progress on a cure.

In a stirring conclusion in which he offers five pieces of advice to medical students on making a difference in patients’ lives, Gawande says that it ‘often seems safest to do whatever everyone else is doing, but a doctor must not let that happen—nor should anyone who takes on risks and responsibilities in society.’ Technology provides many solutions and enables advances in areas previously thought impossible. But it is human ingenuity that underpins technological advance, and sometimes it is simple human practices that have the biggest impact.

Better is an entertaining and informative collection of essays with lessons that go beyond the specifics of practicing medicine. I look forward to reading more from Dr. Gawande in the future.