Why do they bother? Yet another group has performed a study into intercessory prayer and yet again has concluded that prayer doesn’t do anything:
Praying for patients undergoing heart operations does not improve their outcomes, a US study suggests.
A study found those who were prayed for were as likely to have a setback in hospital, be re-admitted, or die within six months as those not prayed for.
I’m not saying that there aren’t studies that show prayer has an effect. There have been several such studies, but eventually they are shown to be flawed, and/or when the controls are tightened the effect disappears, and/or they have been shown to be fraudulent. I give three examples below.
One of the most frequently quoted “successful” studies was by Randolph Byrd of the San Francisco General Medical Center:
To evaluate the effects of IP (intercessory prayer) in a coronary care unit (CCU) population, a prospective randomized double-blind protocol was followed. Over ten months, 393 patients admitted to the CCU were randomized, after signing informed consent, to an intercessory prayer group (192 patients) or to a control group (201 patients).
The control patients required ventilatory assistance, antibiotics, and diuretics more frequently than patients in the IP group. These data suggest that intercessory prayer to the Judeo-Christian God has a beneficial therapeutic effect in patients admitted to a CCU.
Byrd claimed that the prayer group did better than the control in six categories.
But there were severe methodological problems with this study. First, no significant differences were found among the other twenty categories being evaluated, including mortality, despite explicit prayers for prevention of death. Byrd had cherry-picked the six categories that happened to support his hypothesis and ignored the others.
Second, in looking at outcomes in those six categories that did support his conclusion, Byrd ignored the interrelationships between the categories. For example, the development of congestive heart failure automatically leads to the need for diuretics; the development of pneumonia automatically requires the use of antibiotics. Since these were not independent events he hadn’t really even reported success in six categories; three at most.
And third, there was no significant difference in length of recovery period, despite explicit prayers for a rapid recovery. This is what counts as success, apparently.
Harris / MAHI
A later study by the Mid America Heart Institute attempted to replicate Byrd, and claimed
Remote, intercessory prayer was associated with lower CCU course scores. This result suggests that prayer may be an effective adjunct to standard medical care.
However, it is clear that this study failed to replicate Byrd’s so-called positive data in the six categories in which Byrd had claimed success. Additionally, there was no significant difference between the prayer and control groups in length of CCU stay, total days hospitalized or deaths, thus, replicating Byrd's negative data. So this study can hardly be considered a success either.
The granddaddy of fraudulent prayer studies has to be Elizabeth Targ’s study in the Western Journal of Medicine that was funded by the National Institutes of Health's Center for Complementary and Alternative Medicine to the tune of $1.5 million. This study originally showed no effect, so the experimenters un-blinded the study, data mined it for something, anything, that would show prayer was a benefit, found something, changed the study to make it appear they had been looking for those things all along, re-blinded it, then published. As Wired reported:
Targ asked him to crunch the numbers on the secondary scores - one a measure of HIV physical symptoms, the other a measure of quality of life. These came out inconclusive; the treatment group didn't score better than the control. Not what they wanted to find. In dismay, Targ called her father. He calmed her down, told her to keep looking. She had Moore run the mood state scores. These came out worse - the treatment group was in more psychological stress than the control group. Same for CD4+ counts. Targ flew down to Santa Fe to attend a conference at a Buddhist retreat run by her godmother. When she called back to Moore's office, Sicher answered. Moore was crunching the last data they had, hospital stays and doctor visits. "Looks like we have statistical significance!" Moore announced. Sicher told Targ, who turned and yelled out to her friends and the conference.
Later that week, Moore met with an AIDS physician at California Pacific Medical Center. This doctor thought distant healing was bogus but agreed to give advice. He remarked that the length of hospital stays wasn't very meaningful. Patients with health insurance tend to stay in hospitals longer than uninsured ones. He pointed Moore to an important AIDS paper that had been recently published. It defined the 23 illnesses associated with AIDS. He told Moore they ought to have been measuring the occurrence of these illnesses all along. Moore took this list to Targ and Sicher. There was only one problem. They hadn't collected this data.
They gathered the medical charts and gave them to their assistant to black out the names of the patients. This done, Targ and Sicher began poring over the charts again, noting the data they hadn't previously collected. Since Sicher had interviewed many of these patients (up to three times), Moore worried Sicher could recognize them just by the dates they came to the hospital and what they were treated for. Sicher admitted he could (there were only 40). He had also seen which group each patient was assigned to, treatment or control, but he swore he didn't remember and maintained he was therefore impartial. (Sicher remembers this differently. He insists he couldn't recognize the patients from their charts and never knew which group each was in.) Targ told her boyfriend she was worried about Sicher's impartiality, but she took him at his word, even though Sicher was an ardent believer in distant healing, by his own frequent admission. He had put up the money himself for the pilot study ($7,500), had paid for the blood tests. He had a vested interest in the outcome.
This isn't what science means by double-blind. The data may all be legitimate, but it's not good form. Statisticians call this the sharpshooter's fallacy - spraying bullets randomly, then drawing a target circle around a cluster. When Targ and Sicher wrote the paper that made her famous, they let the reader assume that all along their study had been designed to measure the 23 AIDS-related illnesses - even though they're careful never to say so. They never mentioned that this was the last in a long list of endpoints they looked at, or that it was data collected after an unblinding.
And that really is the best there is. Isn’t it about time we recognized that getting down on your knees and muttering platitudes to the invisible daddy in the sky doesn’t change anything, and stopped wasting time and money on these ridiculous studies? If there is a god, he doesn’t intervene in medical procedures based on the number of people who tell him how wonderful and powerful he is and who go on to ask for medical favors.