In the last decade, dozens of scientists have claimed they have found evidence of links between religion and health, and the idea has caught on. News articles with headlines like "Spirituality Is Good Medicine" and "The Prayer Cure" have reported the findings as if they added up to just another instance of science confirming the obvious: that spiritual faith helps people stay healthy and live longer.
More than 70 medical schools offer instruction in how to address patients' religious beliefs. An H.M.O. in Denver offers spiritual counseling to all members. And doctors are beginning to cross the traditional divide between religion and medicine.
But a psychologist at Columbia University, Dr. Richard P. Sloan, says this is a movement that should never have started. Dr. Sloan has been examining the research and has found it to be rife with methodological problems and statistical flaws.
"Nobody would dispute that for a great many people, religion provides comfort in times of distress, medical or otherwise," Dr. Sloan said. "But there is no really good compelling evidence that there is a relationship between religious involvement and health."
Dr. Sloan, who studies associations between psychology and heart disease, has made an avocation of debunking the scientific literature on religion and health. "My colleagues and I got tired of reading uncritical media reports about the putative benefits of religious activity," he said.
In a recent Annals of Behavioral Medicine, Dr. Sloan aimed at statements that religion had salutary effects on heart health and said they did not hold water.
One study from 1995 that he reviewed looked at whether religious activity lengthened the lives of 232 elderly patients who had undergone heart surgery.
The subjects were asked to describe their religious denominations, how frequently they attended services, how much strength and comfort they received from religion, the number of people they knew in their congregation and how religious they were.
Taken together, the five characteristics were not found to be associated with longer or shorter survival times. But one of the five, by itself, was. Respondents who said they derived strength or comfort from religion lived longer after surgery than those who said they did not.
Dr. Sloan called the single association insignificant, given that the other variables had no effects on survival. "You can take any data set, and if you cut it apart in a number of different ways, you'll find something eventually," he said. "But you can't just go in and pick out the one finding that's significant and say `Aha, here it is,' when the other findings are not significant."
He said the researchers set out to learn whether all five religious variables together had an effect on survival time. Only after the data were in did they decide to look at each variable independently.
This tactic, Dr. Sloan said, exemplifies what's known in scientific research as the sharpshooter's fallacy, analogous to the way a gunslinger might empty his six-shooter into the side of a barn and then draw the bull's-eye around the bullet holes.
Dr. Harold G. Koenig, a psychiatrist at the Duke University Medical Center in Durham, N.C., said Dr. Sloan misrepresented the study, making it seem weaker than it was. Dr. Koenig, who cited the study as positive evidence that spirituality enhances health in his "Handbook of Religion and Health," said comfort from religion was not the sole variable associated with longer life.
Two others, he said, attendance at services and overall religiousness, were also associated, only less strongly.
"The risk of mortality for those patients who were both low on participation and low on strength and comfort from religion was over 14 times greater than in those with high group participation and high comfort from religion," Dr. Koenig said. "But this is an example of what Sloan does with all of the studies he's reviewing. He selects out the negative and doesn't even mention the strengths of the study."
Another study that Dr. Sloan criticized looked at patients in a cardiac care unit to determine whether they fared better if a small group of born-again Christians prayed on their behalf.
Some patients in the study received the prayers and others did not, and neither the patients nor their doctors knew which ones were which, a double-blind situation.
The researchers looked at 26 possible outcomes, like new cases of congestive heart failure, pneumonia and cardiac arrest and new prescriptions for diuretics and antibiotics. They found that for six outcomes, the people who had prayers said on their behalf fared better.
Those findings were not valid, Dr. Sloan said, because the researchers had failed to analyze their data according to established principles of statistics. When so many variables are in question, he explained, it is quite likely that chance alone will cause a few of them to appear significant. Scientists usually employ statistical formulas to ensure that none of the associations they find can be attributed to chance, but the researchers here failed to do that, Dr. Sloan said.
Dr. Frederic M. Luskin, a clinical science research associate at the Stanford Center for Research in Disease Prevention, who cited this study in his review in 2000 of the research on religion and cardiac health, acknowledged the statistical problem. But he argued that despite that flaw, the study at least raised the suggestion of an association between intercessory prayer and health. "To dismiss it entirely," Dr. Luskin said, "is exaggerated."
Dr. Sloan faults some research for failing even to address the link between religion and health. He dismissed one study of a church-based weight control program that helped participants lose weight. "Just because a weight-loss program is conducted in a church doesn't make it a religious activity," Dr. Sloan said.
But Dr. Koenig argued that the weight-loss program in question had many religious aspects. "These are church members who are supporting each other," he said, "whose belief systems are involved in why they're doing this."
Dr. Koenig said this and other studies suggested links between religion and health that were at least strong enough to warrant further study.
Stronger evidence exists to suggest that religion is beneficial to mental health, he said. Because mental health is connected to physical health, he said, it stands to reason that religion would also be associated with physical health.
"It's just not as clean and easy to show an association with physical health as it is with mental health," Dr. Koenig said.
"That's a reasonable hypothesis," Dr. Sloan responded. "But then you've got to show it."
Dr. Luskin said he welcomed the criticisms because they would lead researchers to improve their methods. "There's not enough data out there to say, `Without a doubt we can prove this,' " he said. "It's just that there is an awful lot of smoke suggesting there's a fire. He's saying no one has proved there's a fire. I think that there's enough smoke that you can say to your patients that fire may be there."
Dr. Koenig said the research suggested doctors should pay attention to patients' spirituality. "Doctors shouldn't prescribe religion," he said. "But they should take a religious or spiritual history on all patients with significant medical illness."
If patients indicate that they are not religious, the doctor should let the subject drop, Dr. Koenig said. If a patient does find strength in religion, he said, "the doctor might consider supporting the religious beliefs that the patient finds helpful."
Dr. Sloan considers such an approach discriminating, because if a doctor thinks a treatment is useful, the doctor should use it no matter what a patient believes. "If a patient has an infection, we don't ask, `What are your feelings about tetracycline?' "
In another paper in the same annals, in an issue on religion and health, Dr. Walter L. Larimore, vice president for medical outreach at Focus on the Family, a religious organization in Colorado Springs, urged doctors to "encourage positive spirituality with their patients."
Dr. Raymond J. Lawrence, head chaplain at New York Presbyterian Hospital, argues in the same journal that if physicians become involved with religion, they will only cook up "a witches' brew that will embarrass medicine and trivialize religion."
Those who want doctors to pay more attention to religion, Dr. Lawrence said, underestimate the complexity of the subject.
Some patients, he added, believe that their illnesses are punishments from God, and well-meaning physicians may try to discourage such a belief.
"So which beliefs are you supporting and which are you not supporting?" Dr. Lawrence asked. "Religion can be a destructive force in the world, just as it can be a benign and health-giving force."