The editors of Clinical Orthopaedics and Related Research ® use the editorial pages of the Journal to raise the profile of many topics: The ethics of publication and of surgical practice, the education of surgical trainees, the conflicts of interest that arise in science and medicine, the language we use in scientific reporting, and numerous others. But perhaps the most-frequent subject we write about in editorials is scientific evidence. What constitutes high-quality evidence? How can readers evaluate it? How can surgeons put it to use in their practices? We return to these issues regularly because they are both important and deeply nuanced.

As an Editorial Board, we consider evidence quality as we evaluate each study for publication. We also think about evidence quality as a structural problem: What kinds of studies deserve readers’ attention more than others, and which sorts of studies need tougher scrutiny. Last year, we identified some characteristics of certain laboratory studies that set them apart as “product testing,” explained why the evidence mustered in studies of this design generally is not compelling, and indicated that our enthusiasm for publishing this kind of work is low [2]. This month in CORR ®, we say goodbye to a genre of clinical research—the case report—likewise because the Editorial Board concurs that this study design no longer meets our readers’ needs for high-quality evidence.

While case reports occasionally contain interesting safety messages, and while case-based learning undoubtedly is effective in medical schools, the fact remains that it is difficult or impossible to draw general inferences from one or a few cases. Even aggregating cases, as now is being done by a number of journals and databases, results in a heterogeneous hodgepodge of dissimilar entities of unproven value to clinicians. Notwithstanding suggestions from political scientists to the contrary [3], the plural of anecdote is not data.

Editors know this, and readers do, too. In fact, readers’ habits helped us to confirm that our decision to drop case reports is correct. Case reports rarely are cited in other research; on a per-article basis, case reports were cited about half as often as other content in CORR ®, and case-of-the month reports in our journal were cited about 10% as often as clinical research articles in the last 2-year period we surveyed. Perhaps citation is the wrong metric; would case reports not be worth retaining if people simply read them? Maybe—but it appears that they are not read very much, either. Readers download case reports far less frequently than other kinds of articles we publish. Readers seem (appropriately) to seek higher levels of evidence in the work they use to guide their practices.

I recently wrote on this page [1] that dramatic increases in numbers of articles submitted to journals require editors to make some difficult choices—some journals now charge manuscript-submission fees (CORR ® does not), and many use editors to screen manuscripts in order to determine suitability for peer review before committing reviewers to the task (CORR ® does). By contrast, the choice to eliminate case reports from our Journal does not seem difficult. They do not provide a strong basis for clinical decision-making, and eliminating them will allow our reviewers and editors to devote more attention and time to the kinds of research that readers prefer. Going forward, we will not publish case reports in Clinical Orthopaedics and Related Research ®.