Prof. Timetest
cluded as an event in the score calculation but was omitted from Tables 1 and 3 in our article.Smith and Fisher are correct in noting that an even-odd medical record number randomization scheme is less than op- timal; in future trials, we would use, as they suggest, a sys- tem that is more impervious to detection. Nevertheless, there is little room for subjectivity in a chart review method that sim- ply records the presence or absence of a set of predetermined events. Thus, we do not believe that our findings were biased by this approach. These writers also raise the issue of ªfile- drawer bias,º ie, the reluctance of some investigators to pub- lish no-effect studies. We clearly have no control over what others may have done, and while this charge can be leveled at any field of inquiry, the fact that in this very young field sev- eral studies with negative findings have been published 3-5ar- gues against such bias. We hope that most investigators, in ad- dressing an important question and having designed their study to the best of their abilities, would make (as we did) an a priori commitment to publish their results regardless of outcome for the good of the overall scientific enterprise. Several letters raised questions regarding the theologi- cal implications of our study. As we noted in our article, we cannot draw any conclusions regarding the existence or na- ture of God from this trial. A critically important attribute of any scientist is open- mindedness, the willingness to objectively consider new or al- ternative concepts and hypotheses. There is a growing demand among patients that we acknowledge their need to be treated as whole persons who have not only physical but emotional and spiritual needs as well. Practicing as we do in a large metro- politan hospital among a wide variety of religious traditions, we are acutely sensitive to the need for a nonsectarian approach to addressing spiritual issues. This diversity is mirrored in the spectrum of religious practices among our authors, which ranged from a variety of Protestant and Roman Catholic traditions to Hinduism. Since spiritual factors may play some role in heal- ing, additional studies are needed to clarify the place of inter- cessory prayer in maintaining and restoring health.
William S. Harris, PhD Manohar Gowda, MD Jerry W. Kolb, MDiv Christopher P. Strychacz, PhD James L. Vacek, MD Philip G. Jones, MS Alan Forker, MD James H. O'Keefe, Jr, MD Ben D. McCallister, MD Kansas City, Mo 1. Byrd RC. Positive therapeutic effects of intercessory prayer in a coronary careunit population. South Med J. 1988;81:826-829.
2. Harris WS, Gowda M, Kolb JW, et al. A randomized, controlled trial of the effects of remote, intercessory prayer on outcomes in patients admitted to the coronary care unit. Arch Intern Med. 1999;159:2273-2278.
3. Joyce CRB, Welldon RMC. The objective efficacy of prayer: a double-blind clinical trial. J Chronic Dis. 1965;18:367-377.
4. O'Laoire S. An experimental study of the effects of distant, intercessory prayer on self-esteem, anxiety, and depression. Altern Ther Health Med. November 1997;3:38-53.
5. Walker SR, Tonigan JS, Miller WR, Corner S, Kahlich L. Intercessory prayer in the treatment of alcohol abuse and dependence: a pilot investigation.
Altern Ther Health Med . November 1997;3:79-86. See Correction below Correction Correction Errors in Results. In the Original Investigation titled ªA Randomized, Controlled Trial of the Effects of Remote, Interces- sory Prayer on Outcomes in Patients Admitted to the Coronary Care Unit,º published in the October 25, 1999, issue of the A RCHIVES (1999;159:2273-2278), the authors, Harris et al, were prompted by questions raised in postpublication correspon- dence to reevaluate their calculations and feel that 2 points need to be clarified. In Table 3 of their article, a percutaneous transluminal coronary angioplasty procedure (PTCA) with a stent and/or a rotablator appeared to count as one event. How- ever, when they calculated the unweighted score, they gave one point for PTCA and an additional point for stent and one for rotablator when these occurred in the same patient. Thus, a patient receiving all 3 procedures was given 3 points, not 1, as was implied in Table 3. Second, the need for a cardiovascular stress test (such as a thallium test or an echocardiogram) was included in the calculation of the Mid American Heart Institute±Cardiac Care Unit (MAHI-CCU) scores but was omit- ted from Tables 1 and 3 of their article. There were 44 of these events in the usual care group (8.4%) and 26 (5.6%) in the prayer group ( P= .11). The following tabulation clarifies how Harris et al arrived at the scores reported in Table 4: Usual Care Group Prayer Group Sum of points from Table 3 as published 1436 1173 Extra points for PTCA + stent 79 59 Extra points for PTCA + rotablator 5 0 Extra points for PTCA + stent + rotablator 4 *0 Cardiovascular stress test 44 26 Total events 1568 1258 No. of patients 524 466 Unweighted MAHI-CCU score as published 3.0 2.7² * Two patients 32 extra points each.
² P= .04. In the calculation of the weighted MAHI-CCU score, the need for cardiovascular stress tests was ranked as a category 4 event; if reclassified as a category 2 event, the mean SEM scores become 6.97 0.26 for the usual care group and 6.24 0.26 for the prayer group ( P= .05); the effect size remains 10% to 11%.
In Table 4, the number of patients in the Usual Care Group was incorrectly reported as ª(n = 52)º; it should have been ª(n = 524).º (REPRINTED) ARCH INTERN MED/ VOL 160, JUNE 26, 2000 WWW. ARCHINTERNMED.COM 1878 2000 American Medical Association. All rights reserved. at Colby College, on October 9, 2009 www.archinternmed.com Downloaded from
