“Stepping out of time?” Maybe not

10 Jun

Last week the Ellisnelson blog ran a piece titled “Stepping Out of Time.” The post included this bit about a study on the effectiveness of prayer:

“Leonard Leibovici, MD conducted a double-blind, randomized trial of 3393 hospitalized patients all suffering with a sepsis infection. Leibovici was interested in whether prayer could affect patient outcome. The patients were divided with half being prayed for and half not being prayed for. Dr. Leibovici collected data on the length of fever, length of stay in the hospital, and death as a result of infection. Turns out the prayed for patients had an earlier reduction in fever and shorter hospital stays. The death rates for both groups were not statistically different. The results may shock some, but science has been doing prayer studies for quite a while ( Healing Words-The Power of Prayer & The Practice of Medicine by Larry Dossey, MD). The truly mind- numbing thing about the study is that those praying in 2000 were praying for patients who were hospitalized in the period 1990 to 1996. The conclusion drawn here was that patients who were prayed for in 2000, actually got better in the 1990s.”

This is an extraordinary claim, so I looked up the abstract on line. Just from reading the abstract I could see that there might be problems with the study so I posted a comment to the blog. However, Nelson did not post my comment. Of course, one of the great things about the internet is that I can have my own blog.

I looked up and read the entire paper and it is much worse than I originally thought. The paper is badly written and unclear on several critical points. The author did not specify in advance which outcome variable was of interest. This is a problem because of the possibility of HARKing (Hypothesizing After the Results are Known).

Note that difference reported between the two groups was for the length of hospital stay. There was no difference between the groups in terms of mortality, that is, you were equally likely to die whether you had been prayed for or not. In other words there was no effect for the variable that we would care most about.

In addition, the variable hospital stay is operationalized as days until release from the hospital or death. In other words, some of the shorter hospital stays could have been the result of the patient dying.



Enhanced by Zemanta

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: