Recommendations for Reporting of Clinical Research Studies

Abstract

As clinical research methods are evolving and becoming more sophisticated, there is a growing need to standardize the reporting of clinical research studies. Transparent and concise reporting helps the reader appraise the quality and applicability of research findings and compare findings across studies. Recently, several checklists and guidelines have been developed that list key elements for reporting in studies of different designs (Table 1). These checklists guide authors in manuscript preparation and editors, reviewers, and readers in critical assessment of the findings. Furthermore, general recommendations for statistical reporting enhance rigor in the design, execution, and interpretation of a study. This editorial clarifies expectations for manuscripts of original investigation submitted to the American Journal of Kidney Diseases (AJKD) and emphasizes issues related to statistical reporting. Full instructions to authors and links to the recommended formats for reporting are provided on the AJKD Web site (http://www ajkd.org). Readers will find the checklists helpful when reviewing an article.Table 1Recommended Formats for Reporting of Clinical Research StudiesStudy Design/TopicRecommended Format(s) for ReportingRCT/efficacyParallel group design with individual randomization: CONSORT (checklist and patient flow diagram)6Moher D. Schulz K.F. Altman D.G. CONSORT GroupThe CONSORT statement: Revised recommendations for improving the quality of reports of parallel-group randomized trials.Ann Intern Med. 2001; 134: 657-662Crossref PubMed Scopus (968) Google Scholar, 7Altman D.G. Schulz K.F. Moher D. et al.The revised CONSORT statement for reporting randomized trials: Explanation and elaboration.Ann Intern Med. 2001; 134: 663-694Crossref PubMed Scopus (3063) Google Scholar, 8Egger M. Juni P. Bartlett C. CONSORT Group (Consolidated Standards of Reporting of Trials)Value of flow diagrams in reports of randomized controlled trials.JAMA. 2001; 285: 1996-1999Crossref PubMed Scopus (274) Google Scholar, 14CONSORT. Strength in Science, Sound Ethics. Available at: http://www.consort-statement.org/. Accessed November 2, 2006Google ScholarCluster randomized trials: extended CONSORT (checklist and cluster/patient flow diagram)10Campbell M.K. Elbourne D.R. Altman D.G. CONSORT GroupCONSORT statement: Extension to cluster randomised trials.BMJ. 2004; 328: 702-708Crossref PubMed Scopus (1138) Google ScholarNoninferiority and equivalence randomized trials: extended CONSORT (checklist and patient flow diagram)9Piaggio G. Elbourne D.R. Altman D.G. Pocock S.J. Evans S.J. CONSORT GroupReporting of noninferiority and equivalence randomized trials An extension of the CONSORT statement.JAMA. 2006; 295: 1152-1160Crossref PubMed Scopus (1040) Google ScholarRCTs of herbal interventions: extended CONSORT (checklist and patient flow diagram)11Gagnier J.J. Boon H. Rochon P. et al.Reporting randomized, controlled trials of herbal interventions: An elaborated CONSORT statement.Ann Intern Med. 2006; 144: 364-367Crossref PubMed Scopus (451) Google ScholarRCT/harmsExtended CONSORT12Ioannidis J.P. Evans S.J. Gotzsche P.C. et al.Better reporting of harms in randomized trials: An extension of the CONSORT statement.Ann Intern Med. 2004; 141: 781-788Crossref PubMed Scopus (1110) Google ScholarNonrandomized clinical trial/effectiveness of behavioral and public health interventionsTREND (checklist and cluster/patient flow diagram)15Des Jarlais D.C. Lyles C. Crepaz N. Improving the reporting quality of nonrandomized evaluations of behavioral and public health interventions: The TREND statement.Am J Public Health. 2004; 94: 361-366Crossref PubMed Scopus (1410) Google ScholarObservational study: cohort, case-control, cross-sectional/natural history, prognosis, association, prevalenceSTROBE17STROBE Statement. Strengthening the Reporting of Observational Studies in Epidemiology. Available at: http://www.strobe-statement.org. Accessed November 2, 2006Google ScholarObservational study/gene-disease associationSTREGA (forthcoming)19Strengthening the Reporting of Genetic Associations Workshop. http://www.hugenet.ca. Accessed November 2, 2006.Google ScholarLittle et al20Little J. Bradley L. Bray M.S. et al.Reporting, appraising, and integrating data on genotype prevalence and gene-disease associations.Am J Epidemiol. 2002; 156: 300-310Crossref PubMed Scopus (338) Google ScholarStudy of diagnostic accuracy/diagnostic accuracySTARD (checklist and flow diagram of test results)22Bossuyt P.M. Reitsma J.B. Bruns D.E. et al.Towards complete and accurate reporting of studies of diagnostic accuracy: The STARD initiative.BMJ. 2003; 326: 41-44Crossref PubMed Scopus (1161) Google ScholarMeta-analysis of RCTsQUOROM (checklist and flow diagram of study selection)24Moher D. Cook D.J. Eastwood S. Olkin I. Rennie D. Stroup D.F. Improving the quality of reports of meta-analyses of randomised controlled trials: The QUOROM statement Quality of Reporting of Meta-analyses.Lancet. 1999; 354: 1896-1900Abstract Full Text Full Text PDF PubMed Scopus (3945) Google ScholarMeta-analysis of observational studiesMOOSE (checklist and flow diagram of study selection)25Stroup D.F. Berlin J.A. Morton S.C. et al.Meta-analysis Of Observational Studies in Epidemiology (MOOSE) groupMeta-analysis of observational studies in epidemiology: A proposal for reporting.JAMA. 2000; 283: 2008-2012Crossref PubMed Scopus (15504) Google ScholarSystematic review of observational studies on gene-disease associationHuGENet™ Working Group handbook26Little J, Higgins JP (eds): The HuGENet™ HuGE Review Handbook, version 1.0. Available at: http://www.hugenet.ca. Accessed November 2, 2006Google ScholarDecision analysis/cost-effectiveness analysisUS Public Health Service panel on cost-effectiveness in health and medicine28Siegel J.E. Weinstein M.C. Russell L.B. Gold M.R. Panel on Cost-Effectiveness in Health and MedicineRecommendations for reporting cost-effectiveness analyses.JAMA. 1996; 276: 1339-1341Crossref PubMed Scopus (0) Google ScholarBritish Medical Journal Working Party on economic submissions27Drummond M.F. Jefferson T.O. The BMJ Economic Evaluation Working PartyGuidelines for authors and peer reviewers of economic submissions to the BMJ.BMJ. 1996; 313: 275-283Crossref PubMed Scopus (1506) Google ScholarAbbreviations: CONSORT, CONsolidated Standards Of Reporting Trials; MOOSE, Meta-analysis Of Observational Studies in Epidemiology; QUOROM, QUality Of Reporting Of Meta-analyses; HuGENet™, Human Genome Epidemiology Network; STARD, STAndards for Reporting Diagnostic accuracy; STREGA, STrengthening the REporting of Genetic Associations; STROBE, STrengthening the Reporting of OBservational studies of Epidemiology; TREND, Transparent Reporting of Evaluations with Nonrandomized Design. Open table in a new tab Abbreviations: CONSORT, CONsolidated Standards Of Reporting Trials; MOOSE, Meta-analysis Of Observational Studies in Epidemiology; QUOROM, QUality Of Reporting Of Meta-analyses; HuGENet™, Human Genome Epidemiology Network; STARD, STAndards for Reporting Diagnostic accuracy; STREGA, STrengthening the REporting of Genetic Associations; STROBE, STrengthening the Reporting of OBservational studies of Epidemiology; TREND, Transparent Reporting of Evaluations with Nonrandomized Design. A clinical trial is an experimental study to assess the effect of an intervention or compare effects of 2 or more interventions. The term clinical trial includes randomized controlled trials (RCTs) with random allocation of subjects to active and comparison group(s), nonrandomized controlled trials with nonrandom allocation, and uncontrolled trials without a control group.1The Cochrane Collaboration: Glossary of Terms in the Cochrane Collaboration, version 4.2.5. Available at: http://www.cochrane.org/resources/glossary.htm. Accessed November 2, 2006Google Scholar An RCT represents the gold standard for a study that tests the efficacy of an intervention. In 1996, the CONsolidated Standards Of Reporting Trials (CONSORT) statement proposed a standard for the reporting of parallel-group RCTs.2Begg C. Cho M. Eastwood S. et al.Improving the quality of reporting of randomized controlled trials The CONSORT statement.JAMA. 1996; 276: 637-639Crossref PubMed Google Scholar It comprises a checklist of items to be reported and a diagram to detail the flow of study subjects through a trial. CONSORT has been widely accepted in the medical community. Its success rests on supporting evidence showing the influence of various methodological criteria on RCT results.3Schulz K.F. Chalmers I. Hayes R.J. Altman D.G. Empirical evidence of bias Dimensions of methodological quality associated with estimates of treatment effects in controlled trials.JAMA. 1995; 273: 408-412Crossref PubMed Scopus (3989) Google Scholar Endorsement of CONSORT by medical journals is associated with improved quality of reporting.4Moher D. Jones A. Lepage L. CONSORT Group (Consolidated Standards for Reporting of Trials)Use of the CONSORT statement and quality of reports of randomized trials: A comparative before-and-after evaluation.JAMA. 2001; 285: 1992-1995Crossref PubMed Scopus (851) Google Scholar, 5Devereaux P.J. Manns B.J. Ghali W.A. Quan H. Guyatt G.H. The reporting of methodological factors in randomized controlled trials and the association with a journal policy to promote adherence to the Consolidated Standards of Reporting Trials (CONSORT) checklist.Control Clin Trials. 2002; 23: 380-388Abstract Full Text Full Text PDF PubMed Scopus (138) Google Scholar The CONSORT guidelines have been updated for RCTs of parallel designs6Moher D. Schulz K.F. Altman D.G. CONSORT GroupThe CONSORT statement: Revised recommendations for improving the quality of reports of parallel-group randomized trials.Ann Intern Med. 2001; 134: 657-662Crossref PubMed Scopus (968) Google Scholar, 7Altman D.G. Schulz K.F. Moher D. et al.The revised CONSORT statement for reporting randomized trials: Explanation and elaboration.Ann Intern Med. 2001; 134: 663-694Crossref PubMed Scopus (3063) Google Scholar, 8Egger M. Juni P. Bartlett C. CONSORT Group (Consolidated Standards of Reporting of Trials)Value of flow diagrams in reports of randomized controlled trials.JAMA. 2001; 285: 1996-1999Crossref PubMed Scopus (274) Google Scholar and adapted for RCTs with noninferiority or cluster design, herbal interventions, and harms reporting.9Piaggio G. Elbourne D.R. Altman D.G. Pocock S.J. Evans S.J. CONSORT GroupReporting of noninferiority and equivalence randomized trials An extension of the CONSORT statement.JAMA. 2006; 295: 1152-1160Crossref PubMed Scopus (1040) Google Scholar, 10Campbell M.K. Elbourne D.R. Altman D.G. CONSORT GroupCONSORT statement: Extension to cluster randomised trials.BMJ. 2004; 328: 702-708Crossref PubMed Scopus (1138) Google Scholar, 11Gagnier J.J. Boon H. Rochon P. et al.Reporting randomized, controlled trials of herbal interventions: An elaborated CONSORT statement.Ann Intern Med. 2006; 144: 364-367Crossref PubMed Scopus (451) Google Scholar, 12Ioannidis J.P. Evans S.J. Gotzsche P.C. et al.Better reporting of harms in randomized trials: An extension of the CONSORT statement.Ann Intern Med. 2004; 141: 781-788Crossref PubMed Scopus (1110) Google Scholar Although poor quality of reporting of an RCT cannot be equated automatically with poor methodological quality,13Huwiler-Muntener K. Juni P. Junker C. Egger M. Quality of reporting of randomized trials as a measure of methodologic quality.JAMA. 2002; 287: 2801-2804Crossref PubMed Scopus (334) Google Scholar assessment of methodological quality and applicability is contingent on complete reporting. Thus, RCTs submitted to AJKD must address all points in CONSORT. The CONSORT Web site provides the most updated versions and hyperlinks with examples illustrating each item.7Altman D.G. Schulz K.F. Moher D. et al.The revised CONSORT statement for reporting randomized trials: Explanation and elaboration.Ann Intern Med. 2001; 134: 663-694Crossref PubMed Scopus (3063) Google Scholar, 14CONSORT. Strength in Science, Sound Ethics. Available at: http://www.consort-statement.org/. Accessed November 2, 2006Google Scholar Studies that examine the effect of a behavioral or public health intervention may use nonrandom allocation. Guidelines for reporting the latter are provided in the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) Statement.15Des Jarlais D.C. Lyles C. Crepaz N. Improving the reporting quality of nonrandomized evaluations of behavioral and public health interventions: The TREND statement.Am J Public Health. 2004; 94: 361-366Crossref PubMed Scopus (1410) Google Scholar The CONSORT flow diagram also should be completed. An observational study is a study in which the investigators do not deliver an intervention or manipulate its use, ie, they do not assign patients to treatment and control groups. Instead, investigators observe and describe patients based on their exposure to an intervention or risk factor.16National Information Center on Health Services Research and Health Care Technology (NICHSR): HTA 101: Glossary. Available at: http://www.nlm.nih.gov/nichsr/hta101/ta101014 html. Accessed November 2, 2006Google Scholar Observational studies address a wide spectrum of clinical questions relating to prevalence, natural history, risk relationship, or pathogenesis. With the greater range of hypotheses and research designs, as well as the more exploratory nature of observational research, it may not always be possible or desirable for an observational study to fit a particular mold for reporting. The challenge is to strike a balance between setting minimal standards, yet allowing freedom to explore new ideas and evolving technologies. A recently formed group for STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) provides a generic outline for the reporting of longitudinal cohort studies, case-control studies, or cross-sectional studies.17STROBE Statement. Strengthening the Reporting of Observational Studies in Epidemiology. Available at: http://www.strobe-statement.org. Accessed November 2, 2006Google Scholar Studies that evaluate gene-disease associations constitute an increasingly common subtype of observational study. These studies deserve special mention because the fast evolution in molecular and genetic technology is accompanied by a risk for spurious findings.18Ioannidis J.P. Bernstein J. Boffetta P. et al.A network of investigator networks in human genome epidemiology.Am J Epidemiol. 2005; 162: 302-304Crossref PubMed Scopus (100) Google Scholar We look forward to the recommendations by the Working Group for STrengthening the REporting of Genetic Associations (STREGA), expected to be released in late 2006.19Strengthening the Reporting of Genetic Associations Workshop. http://www.hugenet.ca. Accessed November 2, 2006.Google Scholar They will provide guidance for reporting of gene-disease association studies. Meanwhile, investigators can follow the checklist proposed by Little et al,20Little J. Bradley L. Bray M.S. et al.Reporting, appraising, and integrating data on genotype prevalence and gene-disease associations.Am J Epidemiol. 2002; 156: 300-310Crossref PubMed Scopus (338) Google Scholar although it does not explicitly distinguish between criteria for reporting and quality appraisal. In addition to following the general principles for statistical reporting detailed below, gene-disease association studies should report results of testing for Hardy-Weinberg equilibrium, address the issue of population stratification, and report the method used for dealing with multiple testing. For each gene polymorphism, the rs number, Celera hCV number, or flanking sequences should be provided.21Single Nucleotide Polymorphism (SNP) Database. Available at: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD=search&DB=snp. Accessed November 2, 2006Google Scholar A study of diagnostic accuracy compares the performance of 2 or more diagnostic tests. The STAndards for Reporting of Diagnostic accuracy (STARD) statement provides a checklist to follow and a flow diagram to record test results for reporting diagnostic accuracy studies.22Bossuyt P.M. Reitsma J.B. Bruns D.E. et al.Towards complete and accurate reporting of studies of diagnostic accuracy: The STARD initiative.BMJ. 2003; 326: 41-44Crossref PubMed Scopus (1161) Google Scholar Other studies of diagnostic tests may not fit the suggested format, including studies that assess the impact of following different diagnostic strategies on patient management and future health outcomes.23Fineberg H.V. Bauman R. Sosman M. Computerized cranial tomography Effect on diagnostic and therapeutic plans.JAMA. 1977; 238: 224-227Crossref PubMed Scopus (154) Google Scholar Assessment of the clinical usefulness of a test depends not only on its diagnostic accuracy, but also on the availability, accuracy, and costs of alternative testing strategies; the risks associated with diagnostic misclassification; and the effectiveness of available treatments. Therefore, a study of diagnostic accuracy should include a discussion of the potential relevance of the test for patient care. A single study in isolation often is not sufficient to answer a clinical question with certainty. Systematic reviews follow an explicit protocol to systematically identify and synthesize the findings of studies that address a similar question. A meta-analysis is a systematic review that quantitatively combines results of several studies to draw a more precise conclusion. Decreasing the variability of reporting in individual studies facilitates the conduct of systematic reviews. Systematic reviews themselves should follow an explicit format for reporting, with clear explanations of how studies are selected and results are combined. Guidelines for meta-analyses of RCTs are presented in the QUality Of Reporting Of Meta-analyses (QUOROM) statement,24Moher D. Cook D.J. Eastwood S. Olkin I. Rennie D. Stroup D.F. Improving the quality of reports of meta-analyses of randomised controlled trials: The QUOROM statement Quality of Reporting of Meta-analyses.Lancet. 1999; 354: 1896-1900Abstract Full Text Full Text PDF PubMed Scopus (3945) Google Scholar and for meta-analyses of observational studies, in the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) statement.25Stroup D.F. Berlin J.A. Morton S.C. et al.Meta-analysis Of Observational Studies in Epidemiology (MOOSE) groupMeta-analysis of observational studies in epidemiology: A proposal for reporting.JAMA. 2000; 283: 2008-2012Crossref PubMed Scopus (15504) Google Scholar For reporting of a systematic review of gene-disease association studies, the reader is referred to the handbook published by the Human Genome Epidemiology Network (HuGENet™) Working Group.26Little J, Higgins JP (eds): The HuGENet™ HuGE Review Handbook, version 1.0. Available at: http://www.hugenet.ca. Accessed November 2, 2006Google Scholar A decision analysis models different possible strategies and their projected consequences to determine the optimal choice or to inform decision-making. A cost-effectiveness analysis is a type of decision analysis that compares the costs of different strategies with the projected outcomes expressed in non-monetary terms.16National Information Center on Health Services Research and Health Care Technology (NICHSR): HTA 101: Glossary. Available at: http://www.nlm.nih.gov/nichsr/hta101/ta101014 html. Accessed November 2, 2006Google Scholar For guidance on the reporting of cost-effectiveness analyses, we reference 2 consensus statements.27Drummond M.F. Jefferson T.O. The BMJ Economic Evaluation Working PartyGuidelines for authors and peer reviewers of economic submissions to the BMJ.BMJ. 1996; 313: 275-283Crossref PubMed Scopus (1506) Google Scholar, 28Siegel J.E. Weinstein M.C. Russell L.B. Gold M.R. Panel on Cost-Effectiveness in Health and MedicineRecommendations for reporting cost-effectiveness analyses.JAMA. 1996; 276: 1339-1341Crossref PubMed Scopus (0) Google Scholar Regardless of study design, investigators should follow basic principles for reporting of statistics. The 1988 edition of the Uniform Requirements for Manuscripts Submitted to Biomedical Journals includes recommendations for reporting of statistics29International Committee of Medical Journal EditorsUniform requirements for manuscripts submitted to biomedical journals.Ann Intern Med. 1988; 108: 258-265Crossref PubMed Scopus (146) Google Scholar; these were further expanded upon in an accompanying editorial.30Bailar 3rd, J.C. Mosteller F. Guidelines for statistical reporting in articles for medical journals Amplifications and explanations.Ann Intern Med. 1988; 108: 266-273Crossref PubMed Scopus (316) Google Scholar All reports with complex statistical methods considered for publication in AJKD will undergo statistical review, and comments from statistical reviewers will be provided to the investigators to assist them in the revision process. The following paragraphs highlight several points regarding appropriate reporting of statistical methods and results. In general, methods should be described sufficiently so that work is reproducible. The study population should be described clearly and inferences should be made only to the population from which the sample is drawn. There must be a clear definition of exposures and outcomes. The total number of events, as well as numbers per group, must be presented. Sufficient information should be provided that the data potentially can be used in a systematic review or meta-analysis. Tables and figures must be used appropriately, and text should not repeat messages already in tables and figures; in particular, numbers stated in tables need not be repeated in text. The threshold for assuming statistical significance should reflect the number of comparisons made, such that P values are adjusted for multiple testing. Appropriate methods of longitudinal data analysis should be used to deal with multiple measurements. In reporting results, descriptors for a variable must be presented as mean and SD for approximately normally distributed variables or median and quartiles for skewed variables. Measures of precision (confidence intervals) should accompany estimates of magnitude of the association/effect (eg, coefficients, odds, or hazard ratios). P-values alone is not a sufficient descriptor of effect. The relationship between confidence limits and clinically meaningful differences should be discussed. Results from models using variable transformations should be translated back to the original scale to allow easier clinical interpretation. Reports that use multivariable models must clearly state the model-building approach, including all variables tested for inclusion, criteria used for covariate selection, and final variables in the model. Diagnostics used in checking model assumptions and a summary of their results must be presented. Appropriate sensitivity analyses should be performed to test the effect of model assumptions; in particular, the effect of influential points and outliers should be evaluated and appropriate interpretations made. These sensitivity analyses should be summarized briefly, especially if they relate to points made in the introduction or discussion. Investigators should be careful not to overinterpret a statistically significant result when conducting multiple tests because such statistically significant findings can occur by chance. Conversely, lack of a statistically significant result does not imply equivalence unless a clinically meaningful difference can be excluded with high probability. Care must be taken to avoid the possibility of reporting bias. This occurs when investigators neglect to publish nonsignificant results. Accordingly, AJKD will endeavor to give equal consideration to both negative and positive studies. It should be noted that an analysis of power is crucial for a negative study. However, investigators should avoid reporting post hoc power calculations intended to support the possibility that a null finding may have resulted from an insufficient sample size. The correct conclusion in most of these cases is that the data are insufficient to support conclusive inferences. Improved reporting will enhance accountability and transparency in clinical research and facilitate its translation into evidence-based clinical practice. As editors of AJKD, we endorse these recommendations for reporting of original research presented in AJKD. Recommended formats will help investigators in manuscript preparation and enhance the efficiency of the review process. Given the rapid progress in this field, we expect that recommended formats will undergo revisions based on empiric evaluation and feedback from users. As new conventions for reporting are established, we will include them in the AJKD Information for Authors and update our Web site. One exciting initiative to follow is Enhancing the QUality And Transparency Of Research (EQUATOR), which aims to lead to greater consistency in the development and implementation of reporting guidelines by providing an umbrella organization for about 20 different reporting initiatives (J.P. Ioannidis, personal communication, June 19, 2006). The authors thank V. Balakrishnan, Inga Peter, and Madhu Rao for critical review of the section on gene-disease association studies and Amy Earley for editorial assistance.