December 2004 Archives

The Bell Curve: Does It Affect Us All?

| No Comments

The implications of the article on cystic fibrosis (CF) treatment in November 23rd's New Yorker go well beyond medicine.

The progress in treating CF in the last forty years is surprising in itself. In the early 1960s, the average patient died by the age of three, but in 2003, life expectancy with CF had risen to thirty-three years nationally. Even more surprisingly, at the best center it was more than forty-seven..

In some sense it shouldn't be surprising that the results of treatment follow the famous bell curve. But 14 years between the best and the average!?! That's much greater than I expected.

What makes the situation especially puzzling is that our system for CF care is far more sophisticated than that for most diseases. The hundred and seventeen CF centers across the country are all ultra-specialized, undergo a rigorous certification process, and have lots of experience in caring for people with CF. They all follow the same detailed guidelines for CF treatment. They all participate in research trials to figure out new and better treatments.

The difference between the average and the best turns out to be relatively easy to understand. At the top institution (in Minneapolis), they go to extraordinary lengths to make the important measures. In CF, it turns out that lung capacity is one of those measures. At this institution, they aren't content if their patients' lung function is eighty per cent of normal, or even ninety per cent. They aim for a hundred per cent — or better

Why? Because, over the years, a small difference in lung capacity affects a patient's chances of surviving noticeably. Lung capacity is just one measure, but it's important.

In addition to a focus on the measures, there's a drive to the doctors that leads them to try out new approaches and think up new ideas. What the best may have, above all, is a capacity to learn and adapt — and to do so faster than everyone else.

What gives me pause for thought that if these things are true for a field as complicated as medicine, they're probably also true for software development.

One could expect to see large differences between software teams and that those differences can be seen in the results. (Anecdotally, we all know this is true.)

Part of excellence comes through specialization. There are more than a hundred centres for cystic fibrosis in the United States. They are all much, much better than any non-specialists could hope to be. (However, they are still within their own bell curve.)

One would expect the best teams to be know the important measures and be driven to improve them. (In the CMM model, this is something that is postponed to maturity level five. )

The best would frequently experiment and try new approaches, even crazy ideas, all in the interests of doing better.

Now, IT isn't medicine, so few people are going to live longer lives as a result of any software. However, most IT shops I know of emphatically are not like the top CF clinics. In a highly competitive world, I wonder if one would ever want to work in a centre that does not strive to be the best.