A species not extinct: Publication of case reports and scientific knowledge

Authors

  • John C. Carey

    Corresponding author
    1. Professor of Pediatrics and Obstetrics and Gynecology, Division of Medical Genetics, Department of Pediatrics, University of Utah, Utah
    • American Journal of Medical Genetics, 127 South 500 East, Suite 620, Salt Lake City, UT 84102.
    Search for more papers by this author

In 2004 Dr. Leslie Biesecker, now Deputy Editor of the American Journal of Medical Genetics, wrote an important and provocative piece for the Journal entitled “Endangered Species” [Biesecker, 2004]. The title of the Editorial Comment referred to the “beleaguered” academic clinical geneticist currently beset with several challenges in the modern world of academics and mainstream medicine, not the least of which is “less time for academic pursuits.” Notwithstanding this eye-catching title, the major focus of the editorial had to do with Dr. Biesecker's thoughtful comments on the gathering, storing, retrieval, and publication of clinical data in the field of medical genetics. His experience in the discipline and as an associate editor had led him to rethink the publication of case reports in this context. Biesecker advocated two changes for the field: (1) “discontinuing the practice of publication of most (bold added) case reports,” and (2) “deposition of individual cases in a data repository (bold added) to facilitate analysis and pooling of cases.” His piece was timely as it arose from numerous discussions that have occurred at the editorial board meetings of the Journal since 2003. The discourse included the recognition of the view of some in the field who question the scientific value of the publication of case reports. This view is not so radical. In a study of journal publications, I examined recent tables of contents of 25 prominent journals in various areas of medicine, including medical genetics: eight of twenty-five (32%) of the journals do not publish reports of single clinical observations [Carey, 2006].

The purpose of this Editorial Comment is to summarize the issues surrounding the publication of case reports (a category of article that the Journal refers to as “Clinical Reports”) and to invite the genetics community to provide input and dialogue on these issues.

The theme of case reports is important for the Journal. In 2005 we received 346 submissions under the author's designation of “Clinic Report.” This figure does not include the many other clinical observations that come under the heading of “Research Letter” or “Correspondence.” Notably, “New Syndromes” (nine submitted in 2005) do not fall under the category of “Clinical Report,” even if the novel entity represents a single patient. In 2005 the Journal published 184 clinical reports, approximately six per issue. Clinical reports comprised one-third of our published journal articles for last year. (Of note, about 40% of the clinical reports comprise clinical observations of patients with cytogenetic abnormalities, demonstrating that the Journal has a significant role in the publication of individuals with chromosome syndromes.)

So what is at issue here? The reasons why a significant proportion of journals do not publish case reports may be numerous, but I will postulate two: (1) The view that case reports (and case series) represent a lower quality of evidence in the design hierarchy of studies, and (2) the increasing visibility and utilization of the impact factor (IF) in medical scientific publications as a metric. Let me address the first of these influences, that is, the place of case reports in design hierarchy of studies in evidence-based medicine. I would like to direct your attention to a recent publication in Obstetrics and Gynecology. In this paper, Dauphinee et al. 2005 documented a recent increase in the number of published articles in their journal of an analytical nature and a subsequent decrease in “anecdotal reports.” The authors clearly deem this direction as “improvement.” Their premise is based on the evidence rating for scientific studies that have been adopted by the American College of Obstetrics and Gynecology and other austere groups. This design hierarchy places “case series/case reports” as level III and having a lower place in the quality of evidence. This view has some foundation: Case Reports, even smaller case series, represent investigations that primarily present a numerator with no denominator. Determination of the successful treatment of a condition can hardly rely on the documentation of a single patient. Single cases cannot lead to an estimation of predictive value. However, it is important to consider that these evidence ratings are primarily used to assess the effectiveness of therapeutic agents and are not necessarily used for studies of etiology, natural history, or pathogenesis.

Let me touch briefly on the second influence mentioned above, that is, the IF. For those who are not familiar with this ratio, the Institute of Scientific Information (ISI) calculates and distributes an IF for over 3,700 scientific and medical publications. The IF is calculated by counting the number of cites to articles published in a journal in a 2-year period over the total number of articles published in that 2-year period. Citations used in the numerator can include all type of articles published in a journal while the citable types of articles used in the denominator are limited by ISI definition. The IF is now often used as a rating of a journal's overall quality and its wide ranging significance should not be underestimated. The IF influences academic endeavors such as submission placement and faculty evaluation of publications and peer reviewing. The economic influence of the IF may be perceived as only indirect but nonetheless can affect journal accessibility by impacting publication schedules and even institutional subscription choices. Within the context, there is an impression that case reports have the effect of lowering an IF because of their relatively small number of citations compared to the analytical studies or reviews. As mentioned above, about a third of papers in the Journal are clinical reports and are included in this denominator. The editorial board of the Journal and the publisher are performing a detailed analysis of the individual citations of clinical reports in our own data gathering. For a comprehensive discussion of the IF, including discussion of its limitations and how it is commonly misunderstood, see Dong et al. 2005.

Having mentioned potential factors for why journals do not publish or have abandoned publication of case reports, I will now address the benefits of the publication of case reports in the scientific literature. Firstly, the documentation of astute clinical observations has a long tradition; it relates the story of a patient, albeit in a rather standard biomedical style, but a patient's story just the same. The stories of our patients make up the essence of the tradition and culture in medicine. All of us are familiar with the statement we may have heard numerous times, “What an interesting patient. Let us write the case up.” Secondly, the exercise of documenting a clinical observation often provides a rich opportunity for students, residents, fellows, and young faculty who have an interest or are beginning their career in academic medicine. Thirdly, for those of us involved in the care of patients with rare diseases, the knowledge of natural history, or a therapeutic approach can be advanced with one or two observations. For example, take a recently published paper on the treatment of hypophosphatemic rickets by octreotide in the linear nevus sebaceous syndrome [Hoffman et al., 2005]. Given the rarity of this condition, a randomized controlled trial, level I evidence, could never be orchestrated to analyze treatment in this condition.

Fourthly, the etiology of disorders can often be determined by a small number of observations. In clinical teratology, the observation of only a few cases of patients with very rare outcomes associated with very rare exposures (termed the “astute clinical observation”) can lead to the inference of causation. For example, see the publication of only the fourth case of the fetal fluconazole embryopathy by Aleck and Bartley 1997. This paper helped to establish high-dose fluconazole as a human teratogen. Many clinical reports of single patients with chromosome disorders exemplify this notion of a single observation determining the causal basis of a phenotype. For example, see the recent paper in the Journal on distal 1q deletion [van Bever et al., 2005]. Both the New England Journal of Medicine and Lancet, regarded by most as two of the most prominent journals in medicine, recognize this potential of a single observation and publish case reports of this nature on a regular basis.

Lastly and importantly, the documentation of some clinical observations generates testable hypotheses regarding causation and mechanism. Following up on this point, I draw your attention to a quite interesting paper in the November 18th issue of Science by Tassabehji et al. 2005. The potential benefits of the observation of a single patient generating hypotheses and adding knowledge of etiology are illustrated nicely in this paper. The authors characterize the phenotype of an atypical Williams syndrome patient with modern craniofacial laser scanning, delimit the patient's small microdeletion and determine the missing genes in the region. The study of the patient led to experiments in a mouse establishing a model of the deleted genes causing a similar craniofacial alteration. The authors proceed to hypothesize about the genes involved in producing the phenotype. Here an astute clinical observation was combined with new approaches to phenotype analysis, state of the art molecular work, and an animal model to advance knowledge of the relationship of genotype and phenotype in Williams syndrome.

Let me conclude this part of the discussion by adding two other points: (1) In the study I mentioned above regarding journal publications, 32% of the journals reviewed (as well as the American Journal of Medical Genetics) still publish clinical reports on a regular basis and not in a modified format, such as, online only [Carey, 2006]. (2) The issue of the design hierarchy of studies mentioned above is an important and timely one. I certainly applaud the adaptation in recent years of this systematic approach to evidence. However, I would maintain that the discussion of clinical reports or even case series as a methodological approach to advancement of scientific knowledge has not been thoroughly evaluated. A recent paper [Dalziel et al., 2005] indicated that there were no previous studies empirically investigating the methodological characteristics of case series. They concluded by stating, “Further exploration of the differences between case series and randomized control trial results preferably using registry of comprehensive case series data would be valuable.” The point is that more study on these smaller cross-sectional investigations is needed.

So how do we as a journal, with our longstanding tradition in the field of medical genetics, address these emerging issues? First, let me turn to our recently articulated mission statement. It was drafted by the associate editors in recent months:

The American Journal of Medical Genetics is a scientific medical publication devoted to the study of the cause and pathogenesis, including molecular developmental analyses; delineation, including phenotype analyses, natural history and variability, nosology and new recognized syndromes; and, management, including genetic counseling, care guidelines, and treatment, of human congenital anomalies and genetic disorders. While recognizing the existence of thousands of human genetic conditions, the main focus is on phenotypically driven analyses of disorders of morphogenesis, metabolism, and neurogenetics, especially mental retardation. Other principal themes of the Journal include the epidemiology of congenital malformations and behavioral phenotypes of syndromes.

Well-written and well-focused clinical observations of single patients (or a small series of patients) have the potential to add to the literature in these noted areas of study. Thus, the Journal will continue to publish single cases as Clinical Reports. However, we anticipate more stringency in selection as to which ones will be published. Some of the cited examples above are examples of the best, and we are seeking the best of these observations that add to knowledge. Additionally, the editorial board has been discussing other strategies for publishing case reports including more brief communication of some types of clinical reports.

The editorial board of the Journal has not made a final decision on the publication strategy of clinical reports. We welcome input. I have appointed a working group of the board, now charged with considering these issues and making a recommendation to the entire board by July of this year. Opinions on this topic can be addressed to the editorial office (editor.ajmg@hsc.utah.edu) or as traditional Correspondence submitted electronically to the Journal. We will acknowledge all such correspondence and publish some in the Journal.

Ancillary