Catherine D. DeAngelis

Clinical trial registration: looking back and moving ahead

In 2005, the International Committee of Medical Journal Editors (ICMJE) initiated a policy requiring investigators to deposit information about trial design into an accepted clinical trials registry before the onset of patient enrollment (1). This policy aimed to ensure that information about the existence and design of clinically directive trials was publicly available, an ideal that leaders in evidence-based medicine have advocated for decades (2). The policy precipitated much angst among research investigators and sponsors, who feared that registration would be burdensome and would stifle competition. Yet, the response to this policy has been overwhelming. The ICMJE promised to reevaluate the policy in 2 years after implementation. Here, we summarize that reevaluation, specifically commenting on registries that meet the policy requirements, the types of studies that require registration, and the registration of trial results. As is always the case, the ICMJE establishes policy only for the 12 member journals (a detailed description of the ICMJE and its purpose is available at www.icmje.org), but many other journals have adopted our initial trial registration recommendations, and we hope that they will also adopt the modifications discussed in this update. The research community has embraced trial registration. Before the ICMJE policy, ClinicalTrials.gov, the largest trial registry at the time, contained 13 153 trials; this number climbed to 22 714 one month after the policy went into effect (3). In April 2007, the registry contained over 40 000 trials, with more than 200 new trial registrations occurring weekly (Zarin D., personal communication). The 4 other registries that meet the ICMJE criteria have also grown as scores of journals have adopted the ICMJE clinical trials registration policy. In response to burgeoning registration, many investigators, sponsors, and government agencies have asked the ICMJE to recognize their local registries as databases that meet the policy. Fortunately, the World Health Organization’s (WHO) International Clinical Trial Registry Platform (ICTRP), which was nascent when the ICMJE began to require trial registration, has matured rapidly and provides options for those that desire a wider array of registries. The ICTRP has taken the first steps toward developing a network of primary and partner registers that meet WHO-specified criteria (4). Primary registers are WHO-selected registers managed by not-for-profit entities that will accept registrations for any interventional trials, delete duplicate entries from their own register, and provide data directly to the WHO. Partner registers, which will be more numerous, will include registers that submit data to primary registers but limit their own register to trials in a restricted area (such as a specific disease, company, academic institution, or geographic region). The ICMJE strongly supports the WHO’s efforts, through the ICTRP, to develop a coordinated process for identifying, gathering, deduplicating, and searching trials from registries around the world, thus eventually providing a 1-stop search portal for those seeking information about clinical trials. In addition to the 5 existing registries, the ICMJE will now also accept registration in any of the primary registers that participate in the WHO ICTRP. Because it is critical that trial registries are independent of for-profit interests, the ICMJE policy requires registration in a WHO primary register rather than solely in a partner register, since for-profit entities manage some partner registers. As previously, trial registration with missing or uninformative fields for the minimum data elements is inadequate (1). Initially, the ICMJE required registration of all clinically directive trials, which it defined as “any research project that prospectively assigns human subjects to intervention or comparison groups to study the cause-and-effect relationship between a medical intervention and a health outcome” (1). In May 2005, the ICMJE clarified this definition to exclude preliminary trials designed to study pharmacokinetics or major unknown toxicity (phase I trials) (5). However, the ICMJE recognizes the potential benefit of having information about preliminary trials in the public domain, because these studies can guide future research or signal safety concerns. Consequently, the ICMJE is expanding the definition of the types of trials that must be registered to include these preliminary trials and adopts the WHO’s definition of clinical trial: “any research study that prospectively assigns human participants or groups of humans to one or more health-related interventions to evaluate the effects on health outcomes” (4). Health-related interventions include any intervention used to modify a biomedical or health-related outcome (for example, drugs, surgical procedures, devices, behavioral treatments, dietary interventions, and process-of-care changes). Health outcomes include any biomedical or health-related measures obtained in patients or participants, including pharmacokinetic measures and adverse events. As previously, purely observational studies (those in which the assignment of the medical intervention is not at the discretion of the investigator) will not require registration. The ICMJE member journals will start to implement the expanded definition of clinically directive trials for all trials that begin enrollment on or after 1 July 2008. Those who are uncertain whether their trial meets the expanded ICMJE definition should err on the side of registration if they wish to seek publication in an ICMJE journal. Over the time during which registration of trial methods has become common practice, several forces have begun advocating for registration of trial results. We recognize that the climate for results registration will probably change dramatically and unpredictably over coming years. For the present, the ICMJE will not consider results posted in the same primary clinical trials register in which the initial registration resides as previous publications if the results are presented in the form of a brief, structured (<500 words) abstract or table. The ICMJE favors a standard abstract format for results reporting, and the CONSORT (Consolidated Standards for the Reporting of Trials) group’s forthcoming guidelines for abstracts related to trials may be one such option. The ICMJE believes that parties interested in results registration should consider requiring the deposition of such an abstract in the registry 24 months after closure of data collection if results are not published in a peer-reviewed venue by that time. The registered abstract should either cite any related full, peer-reviewed publications or include a statement that indicates that the report has not yet been published in a peer-reviewed journal. Researchers should be aware that editors may consider more detailed deposition of trial results in publicly available registries to be prior publication. When submitting a paper, authors should fully disclose to editors all posting in registries of results of the same or closely related work. Three years ago, trials registration was the exception; now it is the rule. Registration facilitates the dissemination of information among clinicians, researchers, and patients, and it helps to assure trial participants that the information that accrues as a result of their altruism will become part of the public record. The WHO’s global efforts towards comprehensive trials registration and the ICMJE’s requirements for registration aim to increase public trust in medical science. Disclaimer: Dr. Sahni’s affiliation as a representative and past president of the World Association of Medical Editors (WAME) does not imply endorsement by WAME member journals that are not part of the ICMJE. Potential Financial Conflicts of Interest: Employment: Dr. Godlee was previously editorial director of Clinical Controlled Trials, which owns the ISRCTN (International Standard Randomised Controlled Trial Number) trials register. Mr. Kotzin is employed by the National Library of Medicine, which produces ClinicalTrials.gov; Mr. Kotzin is not responsible for activities or policies regarding ClinicalTrials.gov. Expert testimony: F. Godlee. Other: R. Horton (Co-Chair, WHO ICTRP Scientific Advisory Group); J.M. Drazen (member, WHO ICTRP Scientific Advisory Group); H.C. Sox (member, WHO ICTRP Scientific Advisory Group); M.B. Van Der Weyden (member, government advisory committee for the Australian and New Zealand Clinical Trials Registry).

Resolving Unreported Conflicts of Interest

CONFLICTS OF INTEREST HAVE TAKEN AN INCREASingly prominent role in politics, business, and medicine. High-profile examples of undisclosed or incompletely reported financial conflicts of interest have been well publicized. There have been recent investigations by academic centers and lawmakers into unreported conflicts of interest by physicians and recommendations for more transparent reporting of potential conflicts of interest by faculty and researchers. Despite this increased attention, episodes of unreported financial conflicts of interest continue to occur. JAMA editors take issues of undisclosed conflicts of interest very seriously and investigate such allegations rigorously. In investigating such allegations, we follow our standard policies and procedures, which involve conducting our own independent investigation and contacting the authors whose disclosures are challenged. We require the authors to verify or refute the allegations and to provide a detailed explanation to support their position. In addition, we also ask all other authors listed on the article to verify that their published disclosure information was complete and accurate. Once the investigation is complete, when appropriate (not all allegations prove to be accurate) we publish a letter in JAMA that reflects the results of the investigation, including the authors’ explanation for why the conflicts were not initially reported as well as an apology from the authors, along with a formal correction, and thereby correct the record and the medical literature. We have never hesitated to do this and have published several such letters following these same procedures. However, these investigations require due diligence to ensure that the allegations are evaluated fairly and the record is corrected as necessary. This approach is important to ensure due process and also to ensure that no one’s reputation is unfairly damaged due to inaccurate charges and at the same time to make certain that all relevant disclosures—including those that might not have been included in the original allegation—have been identified and addressed. Moreover, because of the sensitive nature of these investigations, we have conducted them confidentially and ordinarily have not revealed the identity of the individual bringing the allegations to the person being accused of failing to report conflicts of interest. The results of the investigation ordinarily are not revealed to any third parties until the letter of explanation and the formal correction are published. Clearly, with this level of diligence, these investigations into undisclosed conflicts of interest are time-intensive and require careful attention. To avoid misunderstanding about our policies and procedures, we are clarifying certain aspects regarding our approach for investigations into allegations of unreported potential conflicts of interest. JAMA requires that the individual bringing the allegations provide a written detailed explanation of the unreported conflicts of interest and provide documentation to support the allegation. We will explain to the person bringing the allegation that gaining full cooperation of all parties with knowledge of the facts is likely to be enhanced by maintaining confidentiality while the investigation is under way. This approach has proven to be vital in delicate situations, such as those requiring the assistance of deans. Furthermore, the person making the allegation will be informed about progress of the investigation and will be notified when the investigation is completed. In addition, once the investigation into unreported conflicts of interest is completed and the letter of explanation and the correction (if necessary) are finalized, those documents will be immediately posted online and linked from the article, and then subsequently published in the print journal. The duty to disclose and report potential conflicts of interest hinges on trust and patience, and a common bond among authors, deans, editors, and readers recognizing that reporting the best available biomedical science matters most. Pressures to publish rapidly, reports in the news media, and comments on blogs and advocacy sites must not overwhelm the process of thorough and fair investigation when reputations are at stake. A rush to judgment may spark heat and controversy, but rarely sheds light or advances medical discourse. JAMA has been a leader in developing and enforcing policies on conflicts of interest. We have dedicated our professional efforts to doing all we can to ensure that articles

Financiación, autoría y responsabilidad

1247 Como editores de publicaciones médicas somos conscientes de que la publicación de los resultados de investigaciones clínicas en revistas científicas con impacto constituye un factor determinante en la elección del tratamiento médico a utilizar. En el debate público sobre los resultados de seguridad y eficacia de un estudio se asume tácitamente que la recogida y tratamiento de los datos se ha realizado de forma objetiva e imparcial. El respeto de este principio es vital para la praxis científica de la medicina: la publicación de resultados influye tanto en la elección del tratamiento por parte del médico como en la política sanitaria, bien sea pública o privada. Nos preocupa especialmente que la objetividad, indispensable y valiosa, se vea amenazada por el marco intelectual en el que se conciben parte de las investigaciones clínicas actuales, por los criterios de inclusión de los participantes en los estudios y por la forma en que los datos son analizados y comunicados (o no comunicados). Todos reconocemos que los estudios clínicos son herramientas poderosas y que, por ello, merecen ser tratados con cautela. Los estudios clínicos permiten a los investigadores comprobar en pacientes la validez de tesis biológicas, con el potencial de llevar a modificaciones del tratamiento convencional. El impacto económico de dichas modificaciones puede ser importante. Los estudios bien realizados, publicados en revistas científicas de gran alcance, son frecuentemente utilizados para la comercialización de medicamentos e instrumental médico que resulta en beneficios sustanciales para el promotor del estudio. Por otro lado, los pacientes, en general, participan en estos estudios por razones altruistas, es decir, colaboran sin ánimo de lucro en la mejora de los tratamientos existentes. Según este hecho, la utilización de estudios clínicos con fines fundamentalmente económicos defrauda el fin propio de la investigación clínica y supone una desviación en el uso originalmente esperado de estas herramientas. Hasta hace poco tiempo, los investigadores académicos independientes tenían un papel fundamental en el diseño y desarrollo de los estudios clínicos y en la interpretación de los datos. El ámbito de trabajo de estos investigadores, el centro médico académico, era el eje de esta actividad, de forma que muchas instituciones han desarrollado infraestructuras complejas dedicadas al diseño y desarrollo de ensayos clínicos. La iniciativa académica ha sido protagonista indiscutible en la introducción de nuevos tratamientos y, sin lugar a dudas, ha contribuido a la calidad, rigor intelectual e impacto de los ensayos clínicos. No obstante, debido a la creciente presión económica, esta realidad puede formar parte ya del pasado. El objetivo último de una parte importante de los estudios clínicos actuales es facilitar la aprobación del uso de uno u otro medicamento o de este o aquel material médico. No se trata, pues, de probar la validez de una hipótesis científica innovadora. Como consecuencia del aumento de la sofisticación en el diseño y desarrollo de los ensayos clínicos y del estrechamiento del margen para introducir un nuevo producto en el mercado se ha producido un incremento en el tamaño de los estudios y, en consecuencia, un incremento en los costes del desarrollo de nuevos medicamentos. Se estima que el coste medio de la introducción de un medicamento en el mercado es de alrededor de 500 millones de dólares americanos. La industria farmacéutica se ha visto en la necesidad de reducir gastos y ha descubierto que los grupos u organizaciones privadas de investigación (CRO, acrónimo del inglés Contract Research Organisations, organizaciones no académicas) pueden desarrollar el trabajo por menos dinero y que, además, ponen menos trabas que los investigadores académicos. Durante los últimos años, las CRO han recibido buena ED I TO R I A L E S

The importance of independent academic statistical analysis.

The profession of medicine in every aspect—clinical, education, and research—has been inundated with profound influence from the pharmaceutical and medical device industries (DeAngelis and Fontanarosa, 2008). With industry-sponsored biomedical research, there are numerous examples of high-profile and widely publicized occurrences of data misrepresentation, inappropriate statistical analyses, and selective reporting of major trial results. In 2001, JAMA instituted policies in an effort to counteract such problems, and since then we have continued to strengthen these procedures (Fontanarosa and others, 2005). Unfortunately, critics continue to misunderstand, misconstrue, and misrepresent the details, intention, and application of these policies (Keiding, 2010; Rothman and Evans, 2005). To help ensure the integrity and veracity of data reported in research papers published in JAMA, authors must fulfill 2 requirements. First, for all research reports containing original data (regardless of funding source), at least one author must state that she or he “had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis” (http:// JAMA.ama-assn.org/misc/ifora.dtl). For industry-sponsored studies, this statement must be provided by an investigator who is not employed by any commercial funding source. This requirement is intended to ensure that a responsible academic author is aware of and has full access to “all” data collected as part of the investigation. This investigator vouches for the completeness of the data set, so that if issues or questions should arise about missing data or incomplete reporting of data, this academic author will be held accountable and responsible. Second, industry-sponsored studies in which the data analysis has been conducted only by statisticians employed by the company sponsoring the research or by other for-profit entities will not be accepted for publication in JAMA. For these studies, an independent analysis must be conducted by a biostatistician who has a faculty appointment at, and is operating under the auspices of, an academic institution or who is employed by a government research institute. For these analyses, the entire raw data set must be provided to the independent biostatistician, along with the study protocol and the prespecified statistical analysis plan. The independent biostatistician verifies the appropriateness of the analytic plan and conducts an independent analysis of the raw data. The results of these analyses are reported in the manuscript. Details of this independent statistical analysis, as well as the name and academic institution of the independent statistician and whether compensation or funding was received for conducting the analyses, are also included in the published article. This requirement is not intended to cast aspersions toward biostatisticians employed by industry or to imply any doubt about their analytic capabilities. However, there may be unique considerations for employees of for-profit study sponsors because of the vested financial interest those companies have in

Sponsorship, authorship, and accountability.

As editors of general medical journals, we recognize that the publication of clinical-research findings in respected peer-reviewed journals is the ultimate basis for most treatment decisions. Public discourse about this published evidence of efficacy and safety rests on the assumption that clinical-

Child and Adolescent Health—A Call for Papers

HEALTH CARE IS A VITALLY IMPORTANT ISSUE FOR ALL nations. However, despite health care spending in theUnitedStates rising6.7%in2006, exceeding

Toward more uniform conflict disclosures--the updated ICMJE conflict of interest reporting form.

2 trillionor16%of thegrossdomesticproduct, there arestill47millionAmericanswithouthealthinsuranceandglobally theUnitedStates ranks42ndin lifeexpectancy,downfrom 11th 2 decades ago. So what can be done? One possible way to mitigate the problem is to alter priorities from treatment of disease to prevention and start in early childhood. Inarecentcommentary,Woolfseemedalmostprescientwhen he said, “turmoil in health care and the economy may be shiftingthedynamicsforhealthpromotionanddiseaseprevention.” According to Woolf, the potential benefits of disease prevention are profound in modifying prevalence and severity of disease and costs, yet only 1% to 3% of US health care expenditures are used for prevention. One of the most cost-effective interventions is immunization. A recent ranking of clinical preventive services, based on clinically preventable burden and cost-effectiveness, ranked immunizing children highest. However, despite the huge success of vaccination, prevention of illnesses in children (as in other areas of medicine) lags behind that of adults. In the most recent Guide to Clinical Preventive Services containing recommendations by the US Preventive Services Task Force between 2001 and March 2008, 21 services are recommended for adults but only 6 for children. 5 These recommended preventive services for children and adolescents include prescriptionoforal fluoride topreschoolerswhoseprimarywater source is fluoride-deficient to prevent dental caries; prophylactic ocular topical medication for all newborns to prevent gonococcal ophthalmia neonatorum; HIV screening of adolescents at increased risk of infection; routine iron supplementation for children aged 6 through 12 months who are at increased risk for iron-deficiency anemia; sickle cell screening in newborns; and visual impairment screening in children younger than 5 years. The reason for so few preventive service recommendations for children and adolescents is because studies to prove effectiveness or cost-effectiveness are not available or are of low quality. For example, the task force found insufficient evidence to recommend for or against routine screening for overweight in children and adolescents. In explaining the paucity of pediatric recommendations, Maciosek et al stated that “many interventions have yet to be studied in children and adolescents. Effective clinical interventions for preventing or changing negative youth behaviors related to use of tobacco, alcohol, and drugs; physical activity; and nutrition are especiallyneeded.These interventionshave thepotential tobeboth consequential and cost saving.” Despite being understudied, health promotion and disease prevention have huge potential in children and adolescents because the ramifications could extend over the entire life span. How much more effective would be preventing a physical disease or mental illness starting in childhood as opposed to adulthood? How much more would be gained from stopping a child from smoking rather than getting an adult to stop, or from teaching good nutrition or making exercise a habit from an early age? What would be the effects on asthma, obesity, or diabetes if breastfeeding rates were increased? What would be the impact on lifelong neurological or orthopedic morbidity if childhood injuries were avoided or mental illness were prevented or treated early in life? Because of the great potential of health promotion and diseasepreventioninchildrenandadolescentstoimpactthehealth of children and adults and the costs of health care, JAMA will devoteaspecial themeissue to this topic in June2009.Authors from all nations are invited to submit papers on all related topics,especiallyhigh-qualityoriginal researchthatpresentsnovel findings, as well as systematic reviews and commentaries. Papers submitted by February 2, 2009, will have the best chance of being considered for the theme issue. Authors should consult the JAMA Instructions forAuthors forguidelinesonmanuscriptpreparationandsubmission. All submittedmanuscripts will undergo our usual rigorous editorial review. We look forward to receiving your manuscripts.

Formal Courses and Conferences

of Interest Reporting Form The great variability in the processes that different journals use to ask about and report authors’ potential conflicts of interest creates confusion for authors, readers, and the public. To help lessen this confusion, the International Committee of Medical Journal Editors (ICMJE) developed an electronic uniform disclosure form and placed it in the public domain in October 2009. The ICMJE member journals piloted the form, encouraged other journals to use it, and invited feedback. We recognized that the reporting of competing interests is complex and nuanced and sometimes contentious, and thus anticipated modifying the form on the basis of feedback received. We are grateful to the many authors, editors, and other interested parties who took the time to comment on the form and its implementation. The issues raised ranged from technical problems about the correct deployment of the form (to function, it requires the user to download the free Adobe Reader, version 8.0 or higher [Adobe Systems, San Jose, California]), to concerns about the ethics of inquiring about nonfinancial associations. The Committee considered these valuable comments and revised the form at our most recent meeting. We made several modifications. The major change in the reporting instrument is the removal of the queries about potential competing interests of authors’ spouses and minor children and about nonfinancial competing interests. We made this change on the basis of the largely negative feedback that we received about these sections. People who commented about this issue made it clear that there is immense difficulty in defining competing interests beyond those that involve the direct exchange of money from an interested party to an individual author or the author’s institution. Because the Committee continues to believe that there are situations in which indirect or nonfinancial factors could influence (or appear to influence) the conduct or interpretation of work, we replaced the specific questions with a single open-ended query (new Section 4) that asks, “Are there any other relationships or activities that readers could perceive to influence, or that give the appearance of potentially influencing, what you wrote in the submitted work?” This change places the onus on the person completing the form to identify and report appropriate nonfinancial competing interests. It has the advantage of being less intrusive than the previous queries, while providing a locus where authors can report nonfinancial relationships that may be perceived as potential conflicts of interest. In response to comments about the clarity of the form, each field in the form now has a numeric designation. We have modified the language in the instructions and in the individual queries. To make the form more useful to nonnative English speakers, we are creating a glossary of terms used in the form and will be posting guidelines for translation of the form’s instructions into multiple languages. The translation of this form is particularly challenging because translations must capture the essence of the queries rather than their literal meaning. The glossary and instructions will be available at www.icmje.org in the next few months; translations will be posted on the ICMJE Web site as they become available. The new form, in English, is currently available on the ICMJE Web site and the Web sites of our member journals. Authors who have completed the older version of the form in conjunction with a journal submission need not complete the new form, but the new form will be the standard for new submissions. We welcome continued input from the user community. Comments can be sent via the “contact us” link at www.icmje.org. The Committee will consider comments received before 1 May 2011, when we prepare the next iteration of the uniform conflict of interest disclosure form. The complexity, subjectivity, and emotionality of conflict disclosure assure that some will consider this vehicle for reporting to be excessively burdensome, while others will think it falls short in one area or another. We cannot, however, let the perfect be the enemy of the good. We hope that the revised ICMJE form will be another step toward simplifying and standardizing reporting of conflicts of interest. A more uniform reporting process will alleviate the confusion that prevails when multiple journals use different reporting formats and ease the reporting burden on the members of the biomedical research community so they can pursue the research that will improve the care that we can deliver to our patients. With these thoughts in mind, we encourage all journals to adopt the new version of the uniform disclosure form.

Sponsorship, Authorship, and Accountability

Formal courses and conferences are the second most popular method of continuing medical education, after reading. 2-4 Numerous studies have shown that physicians learn facts from such formal instruction. 5,6 Courses that focus on physician performance in specific clinical situations, however, can not only transfer facts but can also alter that performance, even though such changes are difficult to measure.