Bias, research, and complementary & alternative medicine.
Article Type: Report
Subject: Alternative medicine (Research)
Research bias (Control)
Medical research (Media coverage)
Medical research (Social aspects)
Medicine, Experimental (Media coverage)
Medicine, Experimental (Social aspects)
Author: Klotter, Jule
Pub Date: 02/01/2009
Publication: Name: Townsend Letter Publisher: The Townsend Letter Group Audience: General; Professional Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2009 The Townsend Letter Group ISSN: 1940-5464
Issue: Date: Feb-March, 2009 Source Issue: 307-308
Topic: Event Code: 310 Science & research; 290 Public affairs
Product: Product Code: 8000200 Medical Research; 9105220 Health Research Programs; 8000240 Epilepsy & Muscle Disease R&D NAICS Code: 54171 Research and Development in the Physical, Engineering, and Life Sciences; 92312 Administration of Public Health Programs SIC Code: 8730 Research and Testing Services
Organization: Organization: Association of Health Care Journalists
Geographic: Geographic Scope: United States Geographic Code: 1USA United States
Accession Number: 194825349
Full Text: My experience at the Association of Health Care Journalists' 2005 conference (Chapel Hill, North Carolina) incited much puzzling and questioning about bias. At first, I felt like an interloper. After all, years of reading little-acknowledged information on complementary and alternative medicine (CAM) topics had definitely skewed my view of health care. I no longer mindlessly nodded in agreement, like a bobble-head doll, as the Centers for Disease Control (CDC), Food and Drug Administration (FDA), and Journal of the American Medical Association (JAMA) issued their latest edicts. Too, I had internalized stories of persecution and prosecution of nonconventional practitioners, reported in the Townsend Letter over the years.

When I talked with other writers and editors, I found an unexpected enthusiasm and curiosity about CAM. Even among the cautious and skeptical, no one cast stones. I began to wonder if anti-CAM bias is as prevalent as CAM advocates believe. If bias exists, where is it found and how does it express itself? Has past experience caused CAM advocates to project more prejudice into situations than is actually there? Every human being is biased. Each of us has a particular lens, colored by temperament, experience, and beliefs, through which we view the world and make decisions. Scientific journals set up systems--most notably peer-review--in the hope that objective fact will prevail despite personal bias. How well do these publishing systems prevent the encouragement or selection of one outcome over another, leading to erroneous conclusions?

Positive Bias

"Published record is biased toward positive results," Catherine DeAngelis, editor-in-chief of JAMA, told journalists at the Chapel Hill conference. She attributed this bias to researchers who end studies that are moving toward a negative outcome before their completion. Some researchers change study objectives midstream to match the data and gain a positive outcome. Also, study sponsors have denied their permission to publish negative studies that refute the sponsors' point of view or interests. Consequently, most submissions to journals tend to have positive outcomes.

To curtail this type of selective reporting, the eleven members of the International Committee of Medical Journal Editors, which includes JAMA, New England Journal of Medicine, Annals of Internal Medicine, and The Lancet, adopted a requirement that any research article submitted to their journals must be registered in a public registry. The committee's September 2004 statement reads: "Honest reporting begins with revealing the existence of all clinical studies, even those that reflect unfavorably on a research sponsor's product." (1) Such a system helps editors see if the initial study objective is fulfilled. It also provides other researchers and the public with a record of topics being researched.

Five registries meet the committee's requirements: clinicaltrials. gov (US); ISRCTN.org (UK); actr. org.au (Australia); umin.ac.jp/ctr/index.htm (Japan); and trialregister. nl (the Netherlands). (2) Each registry is managed by a not-for-profit organization and accessible to the public for no charge. These registries assign unique identification numbers to each accepted study. Each study's file is supposed to identify the intervention(s) being studied, the study hypothesis, definitions of the primary and secondary outcome measures, target number of participants, funding source, key trial dates, and contact information for the lead investigator.

Although registration is free, it may not be as open as one would hope. Britain's ISRCTN, for example, provides identification numbers only to randomized, controlled clinical studies that compare two (or more) health-care interventions to assess efficacy. At this time, the International Committee of Medical Journal Editors does not require the registration of studies designed to assess unknown toxicity or to determine pharmacokinetics (phase 1 trials). Registration is also not required for studies that focus on the biology of a disease or that are meant to provide preliminary data for a larger study. Study authors, however, are charged with convincing editors that they have a good reason for not registering.

While the randomized, controlled study design readily compares the efficacy of two drugs, multifactorial protocols found in Ayurveda, Chinese medicine, and acupuncture, and other holistic CAM studies do not fit as easily. Also, large randomized controlled trials are typically the most costly. Since many CAM therapies are not patentable, funding for these studies is challenging. Nonetheless, I found studies involving chiropractic, acupuncture, homeopathy, and supplements registered at all three English-language registries. Registration, however, does not mean that the study's design is sound. Too low a dosage, too few participants, too short a trial, an active placebo, or compromised treatment can minimize a treatment's effect. Most of the CAM studies registered at www.clinicaltrials.gov are sponsored by the National Institutes of Health (NIH) National Center for Complementary and Alternative Medicine (NCCAM).

While the requirement for registry makes the process more open, it does not force sponsors to submit negative studies nor does it address the bias of editors and peer-reviewers. (3) Just as authors are more likely to submit positive reports, editors are more likely to accept manuscripts that show a large effect from a new treatment or strong equivalence of a new treatment to a known treatment.

Peer-Review

At peer-reviewed journals such as JAMA, reviewing editors either reject a manuscript outright or send it on for external peer review. Peer-reviewed medical journals are held in high esteem. The articles that appear in their pages are given more weight than journals whose manuscripts do not undergo peer review by fellow researchers. Reviewers do not accept or reject individual manuscripts. Rather, they give editors and study authors "an informed opinion about a manuscript." (4) Reviewers are supposed to look for mistakes in procedure or logic. A paper's conclusion should reflect the study's results. Citations from previous work should be accurate. Most reviewers receive recognition, not money, for their service. They also have the satisfaction of contributing "to the maintenance of high standards and veracity in their specific areas of research." (5) They receive no training in how to critique a manuscript, although some journals provide guidelines. Reviewers do not duplicate studies.

While the peer-review process helps editors pare down the numbers of submitted manuscripts, it does not ensure accuracy. A study published in Annals of Emergency Medicine used a fictitious manuscript with ten major and 13 minor purposeful errors to evaluate the journal's 262 reviewers. (6) Of the 199 who returned their critique with a recommendation for dealing with the manuscript, 117 (59%) recommended rejecting it. This group also identified the most errors: 39.1% of the major and 25.2% of the minor errors. The 15 reviewers who recommended accepting the manuscript found 17.3% of the major errors and 11.8% of the minor ones. (The remaining 67 reviewers recommended revision.) Only 32% of the reviewers noted that the conclusion was not supported by the results. The performance of these peer reviewers is not unusual. "Peer review influences what projects get funded and where research is published," say the authors of a literature review on peer review for the American Journal of Physical Medicine & Rehabilitation. "However, peer review has been criticized as lacking standardization and objectivity. Variability in the performance of reviewers is widely acknowledged." (7)

Regardless of the presence or absence of errors, lack of objectivity will skew which manuscripts will be accepted or rejected unless editors and reviewers remain conscious of their personal biases. Reviewers have demonstrated a bias against researchers from other countries and particular institutions, according to Wagner et al. (8) Conflict-of-interest bias may affect reviewers as well as researchers. Reviewers may examine a manuscript that conflicts with their own research, clinical activities, or beliefs more carefully than a manuscript that supports their views. Also, studies with strong results and immediate clinical application appeal to reviewers as well as editors.

Hypotheses or studies that challenge strongly held beliefs or powerful interests (such as a nonviral etiology of AIDS, homeopathy, and orthomolecular medicine) may be dismissed regardless of the evidence. In a 2000 German study, 398 reviewers, blinded to the study, received one of two versions of a "short report" submitted for publication. The first report focused on a conventional treatment for obesity. The second report was identical to the first, but the tested intervention was unconventional. Only 41.7% of the reviewers responded, resulting in 141 complete assessment forms. The study's abstract reports "a wide range of responses to both versions of the paper, with a significant bias in favour [sic] of the orthodox version." The authors believe that "technically good unconventional papers may ... be at a disadvantage in the peer review process. Yet the effect is probably too small to preclude publication of their work in peer-reviewed orthodox journals." (9)

CAM Studies and Bias

As previously stated, editors and peer reviewers tend to be more interested in studies with positive results. But does this form of positive bias hold true for CAM studies in orthodox journals? Are CAM studies more or less likely to have positive results than other studies in the same journal? I performed a survey of human (in vivo) research studies in three peer-reviewed journals: JAMA, Annals of Internal Medicine, and the Journal of Alternative and Complementary Medicine. The latter is the official journal of the International Society for Complementary Medicine Research, and is indexed and abstracted in MEDLINE. Originally, I had planned to survey just conventional journals, but an article by three British researchers made me curious about the positive bias displayed in CAM journals. (10)

For my survey, I gathered abstracts, published in January 2002 through June 2006, that tested conventional and CAM therapeutic modalities. I subdivided the CAM group into lifestyle (diet, specific foods, exercise, reading), supplements (vitamins, minerals, amino acids), and other (herbal extracts, acupuncture, homeopathy, etc.). I then categorized the results of each study as positive, negative, or mixed/equivocal. Lifestyle is a subset that both orthodox and CAM practitioners claim. Because diet and exercise have long been cornerstones of naturopathic medicine, I chose to include lifestyle measures in the CAM set. I also charted study size. Larger studies are believed to show more accurate results. (11)

This informal survey gave me some insight into the status of CAM in the early 2000s research environment. (12) Of the 140 JAMA articles included in the survey, 95 (68%) focused on conventional therapies (primarily drugs and some surgical procedures). Of these conventional studies, 56 (59%) had positive outcomes, 30 (32%) had negative outcomes, and nine (nine percent) had mixed or unclear conclusions.

Twenty-three of the total 140 JAMA articles (16%) dealt with lifestyle measures. Lifestyle accounted for 51% of the CAM category. All 23 of these lifestyle studies had positive results. With JAMA's interest in lifestyle measures, I wonder if this high number of positive studies indicates a movement from fewer drugs to morelifestyle recommendations. (If only patients will take more responsibility and practitioners have the time to educate!)

Supplements accounted for eight percent (n = 11) of the total 140 JAMA studies and 24.5% of the CAM category. In JAMA's supplement category (vitamins, fish oil, arginine), two of 11 (18%) had positive results, two had mixed results (18%), and seven (64%) had negative results. One of the positive studies dealt with folic acid-B12-B6 supplementation and major adverse events after percutaneous coronary intervention. The other tested folate and B12 in reducing hip fracture risk in elderly patients who had a stroke. Most of the negative studies involved vitamin E.

Eleven JAMA studies (eight percent) looked at CAM therapies such as acupuncture, herbal extracts, chelation for heart disease, magnetic insoles, and isoflavones. Only one of the 11 (the effect of DHEA on abdominal fat and insulin action in 50 elderly people) was positive. One on the use of acupuncture in treating migraines in 302 patients had a mixed result. In that study, acupuncture and "sham" acupuncture had no statistical difference but both were more effective than a waiting list control. Nine of the 11 CAM studies (82%) were negative, compared to 32% of the conventional studies. I do not know if researchers are more likely to submit negative CAM studies to JAMA or if JAMA's peer-review editorial process simply prefers negative CAM studies. However, people who rely on JAMA for CAM information may be getting a very skewed view. Eighty-two percent negative is high, especially compared to the second conventional journal in my survey.

Annals of Internal Medicine had a total of 67 human (in vivo) studies published during the same period. Fifty-four of the studies (81%) involved conventional therapies--higher than JAMA's 68%. Fortyfour of the conventional studies had positive results (81%); six (12%) were negative; and four (seven percent) had mixed or unclear. Although this journal published fewer CAM and lifestyle studies than JAMA, the results of the ones that they did publish tended to be positive. Of the six studies on lifestyle factors, four (67%) were positive, and two (33%) were negative. The sole article on a multivitamin-mineral supplement was positive. Also, four (67%) of six CAM studies (herbs, acupuncture, yoga, manipulative therapy) were positive. Overall, Annals of Internal Medicine published more studies with positive results than JAMA--81% vs. 59% in conventional category and 69% vs. 58% positive in the CAM category (which includes the high number of positive lifestyle studies in JAMA).

Although it published fewer CAM studies than JAMA, Annals of Internal Medicine had a series of informative articles about chiropractic, homeopathy, acupuncture, herbs, CAM credentialing, and research challenges. These articles tend to be positive (but sometimes conservative), often written by people affiliated with the National Center of Complementary and Alternative Medicine. JAMA's discussion of CAM was limited to editorials calling for supplement regulation and letters to the editor that critiqued (pro and con) published studies.

The Journal of Alternative and Complementary Medicine has the highest percentage of positive studies. Fifty-six (84%) of the 67 surveyed studies were positive; four (six percent) were negative; and seven (ten percent) were unclear or mixed. Again, this high proportion of positive results may stem from researchers' tendency to submit positive studies as well as editor and peer-review preference. Editors can only make choices from submitted material. Botanical and acupuncture studies appeared most often. Energy modalities (qigong, tai chi, healing touch), yoga, massage, homeopathy, and meditation/stress reduction were also the subjects of more than one study. Only one dealt with supplementation (MSM) and two with lifestyle--albeit with choices that mainstream America tends to ignore: vegan diets and yoga philosophy. Unlike the two older journals, the Journal of Alternative and Complementary Medicine had several animal and in vitro studies--reflecting the early stage of research.

Study size varied in these three journals. As mentioned previously, larger studies are believed to show results that are more easily generalized to the population as a whole. Studies in the Journal of Alternative and Complementary Medicine are considerably smaller than studies published in the other two. Fifty-one percent of the studies in the CAM journal had 50 or fewer participants, the size of a pilot study. Only one study had over 300 participants. In contrast, Annals of Internal Medicine published a wider range of studies. Pilot-sized studies, with 50 or less participants, accounted for only 16% (n = 11) of the Annals of Internal Medicine studies with a nearly equal percentage having over 1000 participants (n = 10). Nearly half (47%) fell between 51 and 300 participants. The studies that appear in JAMA tend to be much larger than either of the other two journals' and, therefore, more costly to perform. Twenty-nine percent of the JAMA studies (n = 40) had more than 1000 participants. The next largest group fell between 101-300 participants (25%, n = 35). Only four percent (n = 6) had 50 or less.

Media Coverage of CAM Studies

How mainstream journalists portray a published research study--if they can find it--is yet another story. One of the comments that I heard repeatedly from other journalists concerned the lack of CAM research. What few realize is that MEDLINE, a major information source for researchers, practitioners, and journalists, does not index all journals. MEDLINE (www.ncbi.nlm.nih.gov/entrez/) is a database of biomedical subjects, produced by the US National Library of Medicine and the NIH. MEDLINE indexes 5,164 journals (as of January 2007). This total reflects about 35% of all biomedical journals published worldwide and as little as ten percent of all CAM journals. (13) While testifying before the House Committee on Government Reform (March 10, 1999), Donald A.B. Lindberg, MD, director of the National Library of Medicine, explained that journals are chosen by a committee of outside experts that meets three times a year, reviewing about 120 titles per session. (14) At the time of his statement, MEDLINE indexed 74 complementary and alternative medicine journals, including journals devoted to acupuncture, chiropractic, homeopathy, mind-body therapies, and other subjects that fall under the CAM umbrella.

Recently, the Journal of Orthomolecular Medicine (JOM) was rejected by MEDLINE for the fifth time since 1989. (All JOM articles are now archived at http://orthomolecular.org/library/jom/index.shtml.) JOM is the official journal of the International Society for Orthomolecular Medicine. For nearly 40 years, this journal has published studies and papers about the use of high-dose vitamins and minerals to treat physical and mental ailments, including eight papers by Linus Pauling. JOM has an editorial review board and also uses outside reviewers. Its editor-in-chief, Abram Hoffer, MD, PhD, conducted randomized controlled clinical studies of high-dose niacin and schizophrenia in the 1950s. Andrew W. Saul, JOM assistant editor, says that the National Library of Medicine has given no specific reason for the low scoring the journal received but invites the organization to try again.

JOM is indexed in the British Library's Allied and Complementary Medicine Database (AMED), but Internet access requires a subscription fee. JOM is also listed among "All Journals" in the NIH Office of Dietary Supplements' database IBIDS (International Bibliographic Information on Dietary Supplements). When I tried to find a 2001 JOM article on vitamin B3 and hypochlorhydria, the IBIDS search engine reported no results. In addition, IBIDS produced no results for "schizophrenia + niacin," "vitamins + depression," and "vitamins + schizophrenia." I also searched for authors who had published in JOM. I found no articles from JOM. Selective indexing helps explain why doctors, researchers, journalists, and consumers are unaware of megavitamin therapy and its supporting research.

Most of the research that appears in mainstream media comes from one of the large conventional journals such as JAMA, the New England Journal of Medicine, or The Lancet. These journals e-mail previews of upcoming articles to journalists to encourage timely reporting of studies. Are research studies, with their various nuances, reported accurately in mainstream press? The question is important because consumers and practitioners get much of their information from lay publications.

Reporting on GAIT

On February 23, 2006, findings from the Glucosamine/chondroitin Arthritis Intervention Trial (GAIT) were published in the New England Journal of Medicine. This multi-center, double-blind, placebo-and celecoxib-controlled study was funded by two NIH agencies: the National Center of Complementary and Alternative Medicine, and the National Institute of Arthritis and Musculoskeletal and Skin Disease. Researchers randomly assigned 1583 patients with knee pain and radiographic evidence of osteoarthritis to one of five orally administered treatments: 500 mg of glucosamine hydrochloride three times daily, 400 mg of sodium chondroitin sulfate three times daily, 500 mg of glucosamine plus 400 mg of chondroitin sulfate three times daily, 200 mg of celecoxib (Celebrex, Pfizer) daily, or placebo. In addition, participants could take up to 4000 mg of acetaminophen (Tylenol, McNeil) every day, except for the 24 hours before their clinical evaluation for joint pain. Evaluations occurred at baseline, four weeks, eight weeks, 16 weeks, and finally at 24 weeks. At the outset, the majority of participants (n = 1229) reported mild pain according to the WOMAC scale. The remainder (n = 354) had mild to severe pain. Patients who experienced a 20% decrease in the WOMAC pain subscale by week 24 met the primary outcome measure, showing response to treatment. The abstract's conclusion (often the only information available to nonsubscribers) reads:

"Glucosamine and chondroitin sulfate alone or in combination did not reduce pain effectively in the overall group of patients with osteoarthritis of the knee. Exploratory analyses suggest that the combination of glucosamine and chondroitin sulfate may be effective in the subgroup of patients with moderate-to-severe knee pain." (15)

In the article's discussion about the study's limitations,, the authors posit that the mild degree of pain experienced by most participants may have affected the results. Classic osteoarthritis studies use patients with "a disease flare after the discontinuation of NSAIDS"--that is, with moderate-to-severe pain. (16) Measures used in these studies (such as WOMAC) may not be sensitive enough to identify decreases in mild pain. The high rate of placebo response (60.1%) and low response to celecoxib (compared to previous studies) also indicates that something may be amiss.

In GAIT's subgroup of patients with moderate-to-severe pain (WOMAC pain score, 301-400), the glucosamine-chondroitin sulfate group outperformed the celecoxib group in meeting the primary outcome. Fifty-seven of 72 (79.2%, P = 0.002) in the glucosamine-chondroitin group vs. 50 (69.4%, P = 0.06) in the celecoxib group had a 20% decrease in WOMAC pain score. P values measure the probability that the results might have been the product of pure chance. P = .06 indicates that six percent of the results may be falsely attributed to the celecoxib. P = 0.002 indicates that less than one percent (0.2 %) of the positive results in the glucosamine-chondroitin group are likely to be false positives. The supplement combination group also showed consistency in secondary outcomes. It had better WOMAC function score than celecoxib and was the only group of the five that did not show an increase in acetaminophen use.

Media articles and headlines about GAIT vary so greatly in tone that I had to double-check that they were talking about the same study. The headline for the New York Times article by Gina Kolata reads, "Supplements Fail to Stop Arthritis Pain, Study Says." (17) Kolata's article emphasizes the importance of this study and downplays the significance of the moderate-to-severe-pain findings. In fact, that subgroup merits one sentence, six paragraphs into article, after she states twice that the supplements had no effect and included, "But ... patients who took celecoxib had a statistically significant improvement in their symptoms." (18) Quotes from rheumatologists emphasize the study's importance but do not mention the limitations voiced by the study's authors. Kolata includes pro and con opinions from biostaticians and rheumatologists about the authors' call for further research with people with the moderate-to-severe pain. But she gives the last word to Donald Berry, chairman of the department of biostatistics at Houston's M. D. Anderson Cancer Center: "'This is a spurious subset result if I've ever seen one,' Dr. Berry said. 'I wouldn't spend a nickel trying to confirm it.'" (19)

Not all newspapers have staff journalists who write articles on medical studies. Many newspapers, television stations, and websites use articles prepared by services such as Reuters Health and HealthDay News. The GAIT article, found on www.forbes.com and produced by HealthDay News (a division of ScoutNews LLC), has a more positive tone than that in the New York Times. Instead of the strongly worded "Supplements Fail" headline, HealthDay News's headline reads, "Glucosamine, Chondroitin Not Much Help for Arthritic Knees." The article begins, "The dietary supplements glucosamine and chondroitin ... don't do much to help those with arthritis, a carefully controlled study finds." (20) The next sentence offers "there's some evidence," not Kolata's weaker "hope," that the supplements may help people with moderate-to-severe pain. After describing the study, HealthDay News focuses on the subset by quoting Daniel O. Clegg, GAIT's lead author, and Marc C. Hochberg, the author of an accompanying editorial published in the same issue of NEJM. Both doctors contend that the combination may help those with severe arthritic pain. Clegg is more cautious and intends to research this subset further. Hochberg states that GAIT is flawed, because it did not use glucosamine sulfate, the form of glucosamine that previous studies had shown to be effective. (Clegg explains that the researchers were unable to find a sulfate form that met federal pharmaceutical requirements.) Hochberg recommends that arthritis sufferers try glucosamine sulfate, 1500 mg/day for up to 12 weeks, by which time it should be obvious that it's working or not. The end of this article sounds conciliatory. Hochberg says, "In general ... the use of such supplements for pain relief 'is a difficult subject for consumers and difficult for health-care providers.'" (21) To my surprise, MedPage Today, a medical news service for physicians, ran a positive headline: "Glucosamine-Chondroitin Sulfate Can Ease Severe Pain." (22) MedPage Today provides professional medical analysis of consumer medical news. It also collaborates with the University of Pennsylvania School of Medicine Office of Continuing Medical Education to provide Teaching Brief articles for clinicians and consumers. The GAIT article, written by senior editor Peggy Peck and reviewed by Zalman S. Agus, MD, begins with a statement that the supplements "do not significantly reduce mild knee pain from osteoarthritis, compared with placebo" and then adds, "But in a subset of patients with moderate-to-severe knee pain, the combination of glucosamine and chondroitin sulfate significantly reduced pain and improved function. ... " Unlike the other two articles, this one focuses on specifics from the study and Hochberg's editorial.

Not long ago, virtually every mainstream article about a CAM therapy ended much like Kolata's piece. The skeptic had the last word. The ease with which I found the HealthDay News and MedPage Today tells me that times are changing. A pair of articles in USA Today by Rita Rubin further confirm my optimism. The first, "Study Casts Doubts on Cohosh's Effect," reported that a December 2006 National Institutes of Health-funded study found no difference between black cohosh and placebo in treatment for hot flashes. (23) Less than one month later, Rubin followed up with "An Herbal Menopause Therapy Gets a Closer Look." This article presents concerns about the formulation and dosage used in the black cohosh study. Physician Tieraona Low Dog, director of education for the Integrative Medicine Program at the University of Arizona's Health Sciences Center, gets the last word: "I can't help but observe that for those who think herbal medicine is passe or ineffective, this study provides further confirmation ... For those who believe herbal medicines are effective--this study is just one more example of how researchers get it wrong. And consumers--well, they just end up confused." (24)

Conclusion

My study of medical reporting made me realize how messy science can be. Personal point of view affects what questions get asked and how research proceeds. Mistakes are unavoidable--so is bias. The more I learned about the reporting of medical research, the more I understood that personal bias affects every step of the process. At the same time, I found an unexpected openness toward CAM. I also found monoliths of the status quo, fighting any possibility that unconventional therapies may be beneficial. As in politics or any other controversial field, some publications take a more balanced view than others. Only time and experience will show where the truth lies. In the meantime, I intend to remain aware that every report--including the ones that I write--is shaded by very human bias.

Notes

(1.) DeAngelis C, Drazen JM, Frizelle FA, et al. Clinical trial registration: a statement from the International Committee of Medical Journal Editors. Annals Int Med 2004;141(6):477.

(2.) The WHO International Clinical Trials Portal at www.who.int/ictrp/about/details/en/index.html joined this list in June 2007.

(3.) Since this article was firstoriginally published, the International Committee of Medical Journal Editors has released "Uniform requirements for manuscripts submitted to biomedical journals: writing and editing for biomedical publication," which addresses conflict of interest: www.icmje.org/#conflicts.

(4.) Benos DJ, Kirk KL, Hall JE. How to review a paper. Advan Physiol Educ. 2003;27:48. Available at: www.ajpadvan.physiology.org. Accessed November 20, 2006.

(5.) Ibid. p. 47.

(6.) Baxt WG, Waeckerle JF, Berlin JA, Callaham ML. Who reviews the reviewers? Feasibility of using a fictitious manuscript to evaluate peer reviewer performance [abstract]. Ann Emerg Med. 1998:32(3 Pt1):310-7. www.ncbi.nlm.nih.gov. Accessed November 20, 2006.

(7.) Wagner AK, Boninger ML, Levy C, Chan L, Gater D, Kirby RL. Peer review--Issues in physical medicine and rehabilitation. Am J Phys Med Rehabil. 2003:82(10):792. www.physiatry. org/publications/Peer%20Review%20-%20 Wagner.pdf. Accessed November 20, 2006.

(8.) Ibid. p.794.

(9.) Resch KI, Ernst E, Garrow J. A randomized controlled study of reviewer bias against an unconventional therapy [abstract]. J R Soc Med. 2000;93(4):164. www.ncbi.nlm.nih.gov/entrez/query. Accessed December, 7, 2006.

(10.) Schmidt K, Pittler MH, Ernst E. A profile of journals of complementary and alternative medicine [Abstract]. Swiss Me Wkly. 2001;131(39-40):588-91. www.ncbi.nlm.nih.gov. Accessed March 8, 2007.

(11.) Cohn V. News & Numbers. Ames, IA: Iowa State University Press; 1989:20.

(12.) A database of the abstracts that I used for this survey is available upon request.

(13.) Shekelle PG, Morton SC, Suttorp MJ, Buscemi N, Friesen C. Challenges in systematic reviews of complementary and alternative medicine topics. Annals Int. Med. 2005;142(12-Part 2):1042.

(14.) Assistant Secretary for Legislation. Statement by Donald A. B. Lindberg, MD. March 10, 1999. Available at: www.hhs.gov/asl/testify/t990310b.html. Accessed February 16, 2007.

(15.) Clegg DO, Reda DJ, Harris CL, Klein MA, O'Dell JR, Hooper MM, Bradley JD, et al. Glucosamine, chondroitin sulfate, and the two in combination for painful knee osteoarthritis. New Eng J Med. 2006;354(8):795.

(16.) Ibid. p. 806.

(17.) Kolata G. Supplements fail to stop arthritis pain, study shows. The New York Times. February 23, 2006. www.nytimes.com/2006/02/23/health/23arthritis.html. Accessed December 27, 2006.

(18.) Ibid.

(19.) Ibid.

(20.) HealthDay News. Glucosamine, chondroitin not much help for arthritic knees. February 22, 2006. www.forbes.com/lifestyle/health/feeds/hscout/2006/02/22/hscout531135.html. Accessed December 27, 2006.

(21.) Ibid.

(22.) Peck P. Glucosamine-chondroitin sulfate can ease severe pain. MedPage Today. February 22, 2006. www.medpagetoday.com. Accessed December 27, 2006.

(23.) Rubin R. Study casts doubts on cohosh's effect. USAToday.com. December 19, 2006.

(24.) Rubin R. An herbal menopause therapy gets a closer look. USAToday.com. January 11, 2007.

Bias ... an inclination of temperament or outlook ... prejudice ... systematic error introduced into sampling or testing by selecting or encouraging one outcome or answer over others.

--Merriam Webster's Collegiate Dictionary, 10th edition

Jule Klotter, who has a master's degree in Professional Writing from the University of Southern California, began writing reviews, "Shorts," and articles for Townsend Letter in 1991. When she isn't working, Jule enjoys writing and attending plays and taking her beagle, Sophie, for long walks in the woods. She now lives in South Carolina.

[ILLUSTRATION OMITTED]

This article originally appeared in the July 2007 issue of the Townsend Letter.
Gale Copyright: Copyright 2009 Gale, Cengage Learning. All rights reserved.