Sunday, April 02, 2006

Just barely not significant

I'm already doing way too many blog entries on this, but bear with me; I just need this dose of vindication. The L.A. Times reports that:
Overall, 59% of patients who knew they were being prayed for had complications, compared to 51% of the patients who did not receive prayers. The difference was not considered statistically significant.
Well yeah that's what my calculations told me. It isn't significant at the 99% level. I've just recomputed using a more rigorous method and the numbers still tell me it isn't statistically significant, although just barely. Because it's more or less a borderline case there may indeed be reason for the researchers to rack their brains and offer candidate explanations for the difference, such as the following:
Atrial fibrillation, a fluttering of the heart that can be related to stress, was the most common complication in all groups but was more likely to occur among patients who knew others were praying for them. ... "We conclude that telling people introduces the stress response," said Dr. Charles Bethea of Integris Baptist Medical Center in Oklahoma City and a study researcher. He surmised that patients thought, "Am I so sick that they had to call in the prayer team?"

On another front and in contrast to Koenig, Dr. Richard P. Sloan, a professor of behavioral medicine at Columbia University School of Medicine, offers a critique worthy of consideration.
One problem in the study, he said, was that in addition to the organized prayer, some patients prayed for themselves and received prayers from families, friends, people they work with or their congregations. "They have absolutely no idea how much prayer individuals in any of the groups received," Sloan said. "If we can't know that, we can't draw any conclusions whatsoever about the intervention."

No comments: