Faith Healing by Prayer: Review of a Questionable Study

Bruce L. Flamm M.D.
June 1, 2004

Can prayer influence the success of in vitro fertilization (IVF) embryo transfer? The lead article in the September 2001 issue of the Journal of Reproductive Medicine claimed to have demonstrated that it can [1]. The report’s authors, two of whom were faculty members at Columbia University’s College of Physicians and Surgeons, claimed to have demonstrated that distant intercessory prayer can double the success rate. In this study, 219 infertility patients in Seoul, South Korea, apparently required and underwent in vitro fertilization (IVF). Twenty patients were excluded prior to randomization due to incomplete data, leaving 199 study subjects. After randomization, 100 patients received prayer and 99 did not. Thirty additional patients were subsequently eliminated from the study due to failure of embryo transfer, leaving 88 patients who had prayer and 81 who did not. In vitro fertilization was performed in the usual fashion in both groups. The patients in the prayed-for group were not informed that Christian groups in the United States, Canada, and Australia were praying for them. The pregnancy rate in the prayed-for group (50%) was essentially twice as high as the pregnancy rate in the non-prayed-for group (26%). At first glance, this success rate may seem remarkable. But a close look suggests otherwise.

Any claim of successful treatment, including faith healing, must meet certain criteria. First, it must be established that some type of illness or medical condition actually existed. On any Sunday, revival-tent faith healers and televangelists claim to restore vision to the blind and limb function to the lame. To many people, the sight of wheelchair-bound invalids standing up and walking (or even running) away from their wheelchairs is convincing evidence. But no such miracle healing has been scientifically authenticated. When these cases are investigated, it is often found that the infirmity did not exist or was psychosomatic [2]. Furthermore, the human body has intrinsic healing mechanisms and will frequently recover from illness or injury even when no treatment is rendered. Any claim of successful treatment, medical or otherwise, must therefore account for these factors.

The study’s authors — Kwang Cha, M.D., Rogerio Lobo, M.D., and Daniel Wirth — stated that their results were preliminary, but that their rigorous study design and careful methodology made them valid. Nevertheless, several questions arise about the study design.

  1. Twenty of the initial 219 patients had to be eliminated from consideration due to “fragmentary e-mail transmission.” The nature of the data transmission problem is not explained. One wonders if other data transmission problems of a lesser degree might cast doubt on the remaining data. Were all problems with e-mail transmission of any degree limited to these 20 patients?
  2. The authors claim that “we set out with the expectation that we would show no benefit of intercessory prayer.” This is a questionable claim for at least 2 reasons. First, many studies have failed to demonstrate any medical benefit of intercessory prayer. It would seem doubtful that anyone would embark upon an international study of this magnitude to simply add another negative study [3,4]. Second, 1 of the study’s 3 authors has a long history of publications that appear to sup-port spiritual and prayer-based faith healing [5-11]. Several of these studies originated from an entity called Healing Sciences Research International located in Orinda, California.
  3. In an apparent effort to demonstrate a lack of bias, the authors point out that, “none of the authors are [sic] employed by religious organizations, and we were not asked by any religious groups to conduct this trial, nor did we seek religious advice at any time.” However, the authors did not disclose any other religious affiliations that could introduce bias. Most people engage in religious organizations as volunteers, not as paid employees. A later comment stated that, “most intercessors were known to one author.” The reader may wonder if any other authors actually had strong religious convictions and/or affiliations. The reader can surmise that at least one of the researchers had attachments to some of the participants, since most of these extremely religious people were “known to one author.”
  4. The study protocol seems unnecessarily convoluted and there is no explanation given for the bewildering study design. A straightforward test to evaluate if prayer is efficacious in increasing the success rate of IVF would be to simply have 1 or 2 people secretly pray for success in randomly selected study patients. Patients would consent to the study but would not be told if they were being prayed for. In contrast to this straightforward approach, the study involved at least 3 different levels of variously overlapping and intertwining prayer groups praying for entire groups of patients. Tiers 1 and 2 each consisted of 4 blocks of prayer participants. Prayer participants in tier 1, block A, received a single sheet of paper with 5 IVF patients’ pictures, called a “treatment unit.” The 5-patient “unit” was apparently chosen because 5 photos fit conveniently on a single page. Prayer participants in tier 1, block A, prayed in a directed manner with a specific intent to “increase the pregnancy rate” for patients in the assigned unit. Prayer participants had no information about the patients they were praying for except for a photograph (no name, no age, etc.) on a page with 4 other photographs. Prayer participants in tier 2, block A, were asked to perform 2 different types of prayer. First, they prayed for the prayer participants in tier 1, block A, with the intent to “increase the efficacy of prayer intervention.” In other words, they were apparently praying to increase the effectiveness of their colleagues’ prayers, whatever those prayers might be. The study design is no longer testing whether prayer works but rather testing whether prayer, when reinforced by overlapping layers of prayer participants praying for each other, works. Next, prayer participants in tier 2, block A, also prayed in a nondirected manner for the study patients with the “intent that God’s will or desire be fulfilled in the life of the patient.” Similar nondirected prayers apparently took place in all of the other prayer blocks. Finally, in addition to all of the above groups, tiers, blocks, and units, a separate group of 3 individuals prayed in a general, nonspecific manner with the intent that “God’s will or desire be fulfilled for the prayer participants in tiers 1 and 2.” In other words, these final 3 prayer participants were apparently praying to increase the efficacy of the second tier of prayer participants, who were in turn praying to increase the efficacy of the first tier of prayer participants, who were in turn praying for increased pregnancy rates in the study patients. The authors offer no explanation for this confusing study design.
  5. Including prayers asking that “God’s will or desire be fulfilled” introduced a vague and obfuscating concept that cannot be measured as an endpoint. This maneuver offers additional chances for successful outcome. In other words, if a given patient failed to get pregnant in spite of intercessory prayers, one might still conclude that the prayers were successful, since one could argue that the will of God was fulfilled and his will for that patient was for her to remain infertile.
  6. The authors made no attempt to discover how much prayer was being conducted, outside of the study protocol. on behalf of the study patients. Patients in the study had suffered from infertility for an average of 5 years. When infertility patients finally undergo IVF, their last hope for pregnancy, personal appeals to deities are not surprising. It is also extremely likely that friends and family members were praying for many of the study patients. The problem is that one has no way of knowing who was prayed for by whom, and how this could have affected the study results. Logistic regression was used to evaluate possible confounding variables such as patient age and duration of infertility but no attempt was made to determine if the study groups were biased by Other sources of prayer, the very thing the study set out to investigate. Koreans are approximately 1/3 Buddhist, 1/3 Christian, and 1/3 non-religious or of other religions [12]. Many or all of the patients in the group receiving Christian prayer may have also been receiving Buddhist or Shamanist prayers. Which prayers, if any, influenced the IVF success rates?
  7. The observed statistically significant difference may be misleading. As with homeopathic and psychic research, this kind of study is most likely a contest between 2 placebos. Depending on how many ways there are to win and how many times the study is done, a statistically significant result can be demonstrated. The chances of finding at least one “statistically significant” (p = .05 or less) difference between approaches 100% as more and more endpoints are measured and as the calculations are made in more imaginative ways. A humorous discussion of the statistical limitations of the Cha et al. study can be found in the Skeptic’s Dictionary [12].
  8. Although it is distasteful to ponder the possibility of intentional manipulation of study data, Occam’s razor (the principle that a simple explanation rather than a convoluted one is more often correct) demands that this possibility not be ignored when highly unlikely results are encountered. Simply stated, is it more likely that data were compromised, or that the study proved the existence a supernatural power and may be one of the most important scientific publications in history? From the onset of data collection to the final data analysis, everyone with access to study data could have been aware of the study’s importance. Although the authors explain that they used careful masking techniques to avoid this type of problem, it is always possible that, unknown to the authors, the data were compromised.

Recent Developments

Recent developments have raised questions about the study’s raw data. Dr. Lobo, identified by the New York Times and ABC News as the report’s lead author, now claims to have not been involved with the study until after its completion and to have provided only “editorial assistance.” Both he and Dr. Cha have refused to respond to phone calls or letters about the study. The remaining author, Daniel Wirth, has no medical degree but has published many studies claiming to support the existence of paranormal phenomena. Many of these studies originated from an entity called, “Healing Sciences Research International,” an organization that he supposedly headed. This entity’s only known address was apparently a Post Office box in Orinda California. Wirth holds an MS degree is in the dubious field of “parapsychology” and also has a law degree.

In April 2004, Wirth and an accomplice (Joseph Horvath) pleaded guilty to conspiracy to commit mail and bank fraud and agreed to forfeit assets of more than $1 million acquired through their schemes. Documents in the case indicate that the pair used assumed names, obtained bogus identifying documents, and obtained employment with a large financial institution from which Horvath improperly paid to Wirth for alleged services [13]. Wirth’s long pattern of dishonest behavior raises the question about whether the studies in which he was involved actually took place and, if so, whether the results were reported honestly.


  1. Cha KY, Wirth DP, Lobo RA. Does prayer influence the success of in vitro fertilization-embryo transfer? Journal of Reproductive Medicine 46:781-787, 2001.
  2. Randi J. The Faith Healers. Amherst, NY: Prometheus Books, 1989.
  3. Glickman R, Gracely EJ. Therapeutic touch: Investigation of a practitioner. Scientific Review of Alternative Medicine 2(1):43-47, 1998.
  4. Sloan RP, Bagiella E, Powell T. Religion, spirituality, and medicine. Lancet 353:664-667, 1999.
  5. Wirth DP, Cram JR. The psychophysiology of nontraditional prayer. International Journal of Psychosomatics 41:68-75, 1994.
  6. Wirth DP, Cram JR. Multisite electromyographic analysis of non-contact therapeutic touch. International Journal of Psychosomatics 40:47-55, 1993
  7. Wirth DP, Marrett MJ. Complementary healing therapies. International Journal of Psychosomatics. 41:61-67, 1994.
  8. Wirth DP. The significance of belief and expectancy within the spiritual healing encounter. Social Science & Medicine 41:249-260, 1995.
  9. Wirth DP, Richardson IT, Eidelman WS. Wound healing and complementary therapies: A review. Journal of Alternative and Complementary Medicine 2:493-502, 1996.
  10. Wirth DP, Cram JR, Chang RJ. Multisite electromyographic analysis of therapeutic touch and qigong therapy. Journal of Alternative and Complementary Medicine 3:109-118, 1997.
  11. Wirth DP, Cram JR. Multisite surface electromyography and complementary healing intervention: a comparative analysis. Journal of Alternative and Complementary Medicine 3:355-364, 1997.
  12. Johnstone, P. Operation World: The Day-By-Day Guide to Praying for the World. Grand Rapids, Mich: Zondervan Publishing; 1993.
  13. Carroll B. Mass media bunk. Skeptic’s Dictionary Web site, accesses May 28, 2004.
  14. Fourth superseding indictment. U.S.A. v. John Doe (FBI No. 034055NA9) and Joseph Wirth. U.S. District Court for the Middle District of Pennsylvania. Filed Feb 12, 2004.


Dr. Flamm is an obstetrician/gynecologist with Kaiser Permanente Medical Group, Riverside, California and Clinical Professor of Obstetrics and Gynecology, University of California, Irvine. Portions of this article were originally published in The Scientific Review of Alternative Medicine 6(1):47-51), 2001.