Monday, November 03, 2003

Power of Prayer: My wife (that's Mrs. Enobarbus) does a lot of work with research methods, statistical analysis, and statistical significance. Thus, we had a good topic to discuss when, two years ago, the news broke that research had shown a statistically significant benefit from prayer:
Patients admitted to hospital with heart problems suffer fewer complications if someone prays for them, according to scientists in the US.

The study, carried out at Duke University Medical Center in North Carolina, found that patients who received alternative therapy following angioplasty were 25% to 30% less likely to suffer complications . . . [a]nd those who received "intercessory prayer" had the greatest success rate.

I never doubted that this was anything more than an anomaly, a bit of statistical noise. It was a small study, first of all; secondly, and more importantly, I didn't see how researchers could effectively work past all the variables that might confound attempts to isolate what amounts to a "paranormal" phenomenon. In other words, we can know (to some extent) why a single heart patient might get well. He may change his diet, start an exercise regimen, quit smoking, etc. But there's no telling why a group of patients might get better (at least not without extensive data collection), and attempting to paste a question of prayer onto such a problem is an open door to the misinterpretation of the data.

So I was unsurprised to read this follow-up:

The world's largest study into the effects of prayer on patients undergoing heart surgery has found it appears to make no difference.

The results of the controversial study contradict earlier findings from the same team which suggested a drop of a quarter or more in "adverse outcomes" - including death, heart failure or heart attack.

However, that trial involved only 150 patients, and the more extensive research, completed this year, found no evidence of any benefits.

But the point is that this research is really no better. It came up with an accurate answer (that there's no reason to believe prayer helps) but the methodology is still so flawed that the whole line of research should be thrown out. I can't believe adults are studying this with straight faces.

No comments: