From The Scientific Review of Alternative Medicine, Vol. 4, No. 1, Spring/Summer 2000

Another controversial effort to establish the
medical efficacy of intercessory prayer

by Gary P. Posner, M.D.

A study led by Dr. William S. Harris and published in the October 25, 1999, issue of the AMA's Archives of Internal Medicine "attempt[ed] to replicate [Dr. Randolph] Byrd's findings by testing the hypothesis that patients who are unknowingly and remotely prayed for by blinded intercessors will experience fewer complications and have a shorter hospital stay than patients not receiving such prayer."1 Byrd's 1988 paper was the first to purport to establish the positive effects of intercessory (not personal) prayer, without the recipients' knowledge, on the hospital course of patients.2

Byrd had found that of twenty-six indicators in patients admitted to a coronary care unit (CCU), most differences between prayed-for and control patients were statistically insignificant. Nevertheless he reported a significant decrease in certain medical complications in the prayed-for group: less congestive heart failure, pneumonia, and cardiopulmonary arrest; less need for diuretic therapy, antibiotics, and respiratory intubation and/or ventilation. Byrd also devised a scoring system to rate the overall hospital course as "Good," "Intermediate," or "Bad," and reported a statistically significant decrease in the number of "Bad" outcomes among the prayed-for patients.

Despite its procedural and statistical shortcomings,3,4,5,6 the 1988 Byrd study has been frequently quoted as showing positive effects of prayer. Writers such as Dr. Larry Dossey credit it with spurring their own quests for further proof of the power of prayer.7 The authors of the recent Archives paper reported that their own efforts to "replicate Byrd's findings" were successful: "Our findings support Byrd's conclusions . . . " But, as the authors acknowledge, there is much more (or less) to this claim.

Over a 12-month period, about a thousand patients admitted to the CCU at Mid America Heart Institute (MAHI) in Kansas City, Missouri, were randomly placed either in the prayer or the control group under an assignment protocol based on chart numbering -- odd or even. As in the Byrd study, each patient in MAHI's prayed-for group was assigned to a team of "intercessors" who would pray for the patient on a daily basis. MAHI's teams consisted of five members each; Byrd's varied from three to seven. MAHI's intercessors prayed for "a speedy recovery with no complications" plus "anything else that seemed appropriate to them."

With regard to the matter of "speedy recovery," Byrd's study, in which the prayers had been targeted for "a rapid recovery and for prevention of complications and death," found no statistically significant differences in length of CCU stay, total days hospitalized, or number of deaths. MAHI also found no significant differences in these categories. Thus, with respect to Byrd's negative data, the MAHI study was able to replicate his results.

As for the development of medical "complications" during the course of hospitalization, the MAHI study failed to achieve any of the statistically significant beneficial effects reported by Byrd with regard to congestive heart failure, pneumonia, cardiopulmonary arrest, diuretic therapy, antibiotics, and intubation/ventilation. Further, using Byrd's "Good/Intermediate/Bad" scoring system to evaluate their own data, the MAHI researchers found "no significant difference between [the] groups." Thus, with respect to Byrd's positive data, the MAHI study was not able to replicate his results.

The MAHI authors do not claim to have replicated Byrd's findings -- which was their stated objective. Here is the entirety of their sentence only partially quoted from above: "Our findings support Byrd's conclusions [emphasis added] despite the fact that we could not document an effect of prayer using his scoring method." The MAHI researchers had their own "weighted and summed scoring system called the MAHI-CCU score [which is] a continuous variable that attempts to describe outcomes from excellent to catastrophic." This system was developed at their request by four other clinicians who practice at MAHI.

The MAHI-CCU system, like Byrd's, is "an unvalidated measure of . . . outcomes" (they could find no previously validated system in the medical literature). Yet the MAHI researchers were able to claim "findings . . . consistent with those of Byrd, who reported that intercessory prayer for hospitalized patients lowered the hospital course score." Overall, their prayed-for patients, as a group, scored 11% better than the others, with only a 1:25 probability that this difference in score is attributable to chance alone (P = .04). For comparison, when my skeptical colleagues or I test claimants of paranormal powers, we try to devise a test wherein the probability of "success" by chance alone is in the range of 1:10,000,000. The James Randi Educational Foundation would not confer its $1-million prize upon someone able, on a single occasion, to identify correctly a number between one and 25. As the axiom of science goes, extraordinary claims -- especially supernatural ones -- demand extraordinary proof.

Although they have "no explanation" as to why, even at the P = .04 level, the Byrd scoring system failed to find significance in their own data, the MAHI authors offer speculation. They note that their protocol was more thoroughly blinded than was Byrd's, in that neither the patients nor the medical staff were even aware that a study was being conducted. Such awareness among the Byrd patients assured that those with objections to such a study were able to opt out, thus indicating to the MAHI authors that "only 'prayer-receptive' patients were included in [Byrd's] final cohort." Additionally, the Byrd intercessors had been kept informed as to their patients' conditions and progress, whereas the only patient information given the MAHI intercessors was their first names. Translation: Byrd's scoring system may have been too tough for a more thoroughly blinded test involving patients not screened for prayer receptiveness. But would Byrd's data pass the MAHI-CCU scoring test? Presumably so, though the authors make no mention of this.

As with the Byrd study, one is left to ponder the cosmic significance of the MAHI conclusions. First, do they suggest the existence of a God who responds to prayer? The authors claim only that "when individuals outside of the hospital speak (or think) the first names of hospitalized patients with an attitude of prayer, the latter appeared to have a 'better' CCU experience." Moreover, "it is probable that many if not most patients in both groups were already receiving intercessory and/or direct prayer from friends, family, and clergy during their hospitalization." The authors acknowledge that what they studied was not the effects of intercessory prayer per se, but rather of "supplementary intercessory prayer."

Certain logical conclusions from such prayer studies are not mentioned. One conclusion from the MAHI study, as from Byrd's, is that responses to "supplementary" intercessory prayer appear nearly imperceptible. Also, the responses to prayer appear to be based on the number of prayers/thoughts being offered, independent of the character and religious beliefs of the patients. An "Adolph Hitler" in the prayed-for group would be expected to have a marginally "better" outcome than a "Mother Teresa" in the control group, or in another CCU not involved in a prayer study.

Nevertheless, the MAHI authors conclude that, given the "possible benefits of intercessory prayer" as suggested in their study and Byrd's, "further studies using validated and standardized outcome measures and variations in prayer strategy are warranted to explore the potential role of prayer as an adjunct to standard medical care." Although, as noted earlier, they could find no applicable "validated and standardized outcome measures" in the medical literature, I suspect that this will not deter continuing attempts to prove the medical efficacy of intercessory prayer.

Listen to my 35-minute presentation on prayer and healing at a national conference in 2001, in which this study was one of my principal topics


1. Harris WS et al. A randomized, controlled trial of the effects of remote, intercessory prayer on outcomes in patients admitted to the coronary care unit. Arch Int Med 1999;159(19):2273-2278.

2. Byrd RC. Positive therapeutic effects of intercessory prayer in a coronary care unit population. Southern Med J 1988;81(7):826-829.

3. Posner GP, God in the CCU? Free Inquiry. 1990;10(2):44-45.

4. Witmer J, Zimmerman M. Intercessory prayer as medical treatment? An inquiry. Skeptical Inquirer. 1991;15(2):177-180.

5. Sloan RP, Bagiella E, Powell T. Religion, spirituality, and medicine. Lancet. 1999;353:664-667.

6. Tessman I, Tessman J. Efficacy of prayer: A critical examination of claims. Skeptical Inquirer. 2000;24(2):31-33.

7. Dossey L. Healing Words. New York, NY: HarperCollins, 1993: xv. [Read my book review]

Read some critical Letters to the Editor following publication of this study (here | here | here | here | here | here | here | here)

Read Dr. Larry Dossey's "Commentary" defending the study in the face of these criticisms

Read my Letter to the Editor in response to Dossey's "Commentary"

Return to Posner's Publications Page

Return to Posner's Medically Related Articles Page

Return to Posner's Prayer-Related Articles Page