Review articles are an important staple of medical and other scientific literature that are liked and sought after by both readers and journals. The common perception is that they provide credible and reliable information about a specific area to readers and bring readership to journals. During the last decade or two we have seen a departure from the traditional narrative and integrated review articles to more specialized review articles such as systematic reviews with or without meta-analyses, umbrella reviews, viewpoints, and scoping reviews. It is not always clear what a review article and its different variations and permutations are, and also their goals, attributes, and utility.
Are All Review Articles Useful and Have Clear Goals?
Three recent articles [1-3] on the seemingly very similar topic of coexistence of anxiety and depression are a case in point of the lack of clarity of what a review article is and what its attributes and goals are The first one by Saha et al.  is a systematic review and meta-analysis which “found robust and consistent evidence of comorbidity between broadly defined mood and anxiety disorders. Clinicians should be vigilant for the prompt identification and treatment of this common type of comorbidity.” The second article by Cosci and Fava  (Editor and Associate Editor of this journal, where I am member of the Editorial Board) also states that depressive and anxiety disorders are frequently associated, and that depression may be a complication of anxiety and vice versa. They argue that the selection of treatment of coexisting anxiety and depression “should take into account the modalities of presentation and be filtered by clinical judgment.” The third review by Goodwin  states that: “Many antidepressant agents are also effective for symptoms of GAD [generalized anxiety disorder], including selective serotonin reuptake inhibitors (SSRIs)” and “Likewise, drug classes used to treat GAD are also effective in the treatment of depression with anxious symptoms (e.g., SSRIs, serotonin-noradrenaline reuptake inhibitors, tricyclic antidepressants, and agomelatine).”
I do not want to discount what is presented in these articles, but the question that comes to my mind (and I hope to the mind of other readers) is: what did we learn from these three articles and who benefits from them? Most clinicians know that anxiety and depression frequently coexist in spite of the fact that our nomenclature system has ignored this until the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders . Thus, stating the obvious as a conclusion raises the question about the goals or aims of the article by Saha et al. . Does any clinician need “an up-to-date list of studies that have examined comorbidity between broadly defined mood- and anxiety-related disorders and to meta-analyze the risk estimates according to key design features related to types of prevalence estimates for the two disorders” [sic]? The article by Goodwin  does not really provide anything new, or even a synthesis of previous data. It does not mention monoamine oxidase inhibitors as a treatment for coexisting anxiety and depression. This article also lists only one antidepressant – agomelatine - by its name among key words and in Key Summary Points, which considered together with the funding sources for the article may raise question about the article purpose. Cosci and Fava , in their “Innovation,” remind us of the complexity of clinical psychiatry and point out that: “Reviews generally base their conclusions on randomized controlled trials that refer to the average [italics mine] patient and may thus clash with the variety of clinical presentation.” They emphasize that diagnostic criteria “do not include patterns of symptoms, severity of illness, effects of comorbid conditions, timing of phenomena, rate of progression of illness, response to previous treatments that demarcate major prognostic and therapeutic differences among patients who otherwise seem to be deceptively similar since they share the same psychiatric diagnosis” (p. 309). They apply this discussion to coexisting anxiety and depression symptomatology. They call for a renaissance of psychopathology as the basic neglected method of clinical psychiatry and dimensional psychopathology that needs to be integrated into the diagnostic process. They also warn about the limits of existing guidelines that refer to a hypothetic or average patient, a strategy that is not always helpful in more or even less complicated clinical cases. This article is based on a new model of reviewing literature which integrates details of both biology and biography with the ultimate goal of “precision medicine” . Yet their article  feels more as an “opinion piece” rather than solid review article.
Thus, in my opinion, out of three articles dealing with a similar topic, only one seems to be useful to clinicians . Two of these articles [2, 3] would probably not fulfill usual journals’ criteria for a review article (see below), and one is not really clinically useful . The utility of another one  is also doubtful. I am not saying that all review articles are not useful or do not fit what some envision as a review article. The goals of two of these articles [1, 3], though stated, are not very clear. These three examples [1-3] demonstrate the wide variety of what could be considered a review article, and that the utility of these types of articles can questionable. Thus, it seems that the review article and its attributes and goals require a closer look.
How Do We Define a Review Article and What Are Some of Its Key Attributes?
There is no one generally accepted definition of a review article. We all have certain ideas of how a review should look like and what it should accomplish. Some authors and organizations have formulated definitions of a review article. For instance, Ketcham and Crawford  note that in pragmatic terms the Institute of Scientific Information (ISI) Web of Knowledge Science Citation Index categorizes a paper as a review if it has more than 100 references, or if it appears in a review journal or the review section of a research journal, or if the paper states in the abstract that it is a review paper. That does not seem like a definition of what most readers and editors expect from a review paper, as, based on this definition, a review article could be just a simple listing of the main findings of a number of papers. Maggio et al.  provide a definition that is probably closest to an acceptable definition of a review: “A synthetic review and summary of what is known and unknown regarding the topic of a scholarly body of work, including the current work’s place within the existing knowledge.” They also point out that a literature review should provide context, inform methodology, help maximize innovation, and help avoid duplications. Some outline attributes of a good review article, proposing that such a review should provide synthesis and critical evaluation of available knowledge in a clear and concise way, while using rigorous and consistent methodology. A good review should also be even-handed in presenting various views and interpretations with a focus on answering a clinical or research question(s) in a clear and useful fashion. A solid review article should be sharply focused on a well-defined issue(s).
Though there is not one widely accepted precise definition of a review article, especially in view of newer variations and types of review articles, there seems to be a general notion of what a review article is and what the attributes of a good review article should be. Nevertheless, that is not the only factor that should be considered and addressed in writing a good and useful review. Other factors to be considered in writing are the type of review article, its purpose, intended audience, journal (general, clinical, educational, research), and the expertise of the author(s), their experience, and conflict(s) of interest (COI). In considering the author(s’) COI, one needs to pay attention to all possibilities of COI, and not just the relationship with pharmaceutical companies, but also for instance the author(s’) intellectual “investment,” such as patents, grants, and theoretical orientation (e.g., cognitive behavioral vs. psychodynamic orientation). Some journals also provide their own definition of what a review article should be. Some journals are focused only on reviewing and synthetizing the existing literature (e.g., UpToDate) or on reviews providing continuing medical education.
Types of Review Article
As noted, there is a growing number of permutations of what is called a review article. These include historical review, narrative review, systematic review with or without meta-analysis, integrated review, qualitative research synthesis, and viewpoints, to more specialized ones such as scoping review, rapid review, umbrella reviews, mapping review, meta-synthesis, mixed-method review, overview of reviews, diagnostic accuracy review, review of complex interventions, and network meta-analysis. The discussion of all these types is beyond the scope of this editorial; however, this list illustrates the complexity and confusion regarding review articles.
In their Editorial, Conn and Coon Sells  provide a brief outline of some of the more commonly known and used reviews. Narrative reviews “describe primary studies without using an integrated, meta-analysis, or meta-synthesis approach to examine results of various studies; historical review s trace development of an area of science by assessing the chronological order of studies and focus on possible temporal patterns; meta-analysis is the synthesis of previous quantitative studies and involves statistical approaches to aggregate effect size across primary studies to determine patterns of differences across studies related to key concepts; and umbrella reviews summarize findings of previous reviews. Scoping reviews are characterized as identifying the extent and range of previous studies by creating a profile of existing studies but placing little emphasis on study findings or methodological features . However, in my view it seems that scoping reviews are rather preliminary and short reviews, clarifying concepts and identifying gaps, and are limited in their time and scope. Systematic reviews may or may not use meta-analysis and are focused on systematically analyzing studies related to a certain topic/question with particular attention to the methodology of studies, and synthetizing results using strategies to reduce bias and random errors.
It is important to realize that all types of reviews have their strengths and weaknesses and do have a different place in the evidence hierarchy. For instance, some may feel that meta-analyses provide a great systematic and statistically sound summary of available information, yet others point out their weaknesses, and the fact that meta-analyses of the same subject could provide different results  and could be of little clinical use [10, 11]. Meta-analyses rely on published data, do not necessarily include non-statistically significant results, select non-uniform statistical approaches, may have weak inclusion standards, and frequently have an agenda-driven bias. Contrary to the frequent perception that meta-analyses and systematic reviews are relatively free of COI after the financial COI related to the pharmaceutical industry is addressed, these articles may also be biased by non-financial COI . It is also important to realize that the magic of the term systematic review fades away when one considers the clinical usefulness of this type of review article. As noted , systematic reviews require increasingly complicated and cumbersome procedures, and expert clinicians and expert critical clinical advice are not always part of the systematic review enterprise.
Purpose, Goals, and Attributes of a Review Article
The overall purpose of a review article should be to provide a valuable, solid, informative, critical summary of a well-defined topic/area to the reader. I am emphasizing the reader, as some (e.g., Ketcham and Crawford ) discuss the value of the review article for the author and for journals, and the reader is left out. The issue of value, though not always easy, is also important. It is related to the question that we may frequently have: Does the world need yet another review article on this topic?
Related to the reader is the audience. The audience is related to the reader, but not the same. Is the targeted audience a clinician, researcher, educator, student, resident, patient, families, or others?
Similarly important is the topic, or the question to be “answered,” or explored. The topic should always be clearly stated in the abstract and at the beginning of the text, to help the reader (and before him/her the editor and reviewers) to decide whether this article is of his/her interest. The topic should be as simple and narrow as possible, remembering that “less is more.” Nevertheless, while the topic should be narrow, the final product should be complex (not complicated), nuanced, and consider various opinions, views, and interpretations. The complexity and nuances are especially important in clinically oriented articles, as current clinical thinking is frequently superficial and reductionistic (as demonstrated for instance by the article by Goodwin ).
The next attribute to consider is what type of review article should one select. That is somewhat intertwined with the type of audience and journal. A general audience would probably still most appreciate either the classic narrative (a better term would be critical narrative review) or historical review, followed by a solid systematic review. Most of the other types of review articles (e.g., mapping review, scoping review) are for more specialized audiences and journals. However, narrative reviews should not be the present-day summary of conclusions of research studies and case-reports. They should provide a critical, practical view of highly intellectual content/writing. These articles, especially in clinical medicine, should be written by real experts in the field who could appropriately and skillfully mix so-called evidence-based literature data, their clinical significance and application, and real-life experience and examples.
Selection of the journal is another issue, important for both the reader and the author. Some journals publish only or mostly review articles, some journals have a dedicated section for reviews, and some journals publish review articles occasionally. The reputation of the journal (including the magical and not always useful Impact Factor) should be considered, though it is not the most important factor. Journals with a great reputation can publish mediocre and not very useful reviews, while some journals with not such a great reputation may publish a very useful review(s).
Finally, the issue of usefulness is significant, namely for the reader. It is difficult to gauge easily. The reader (and author and editor) should consider this, especially in clinically oriented review articles. It is connected to the type of reader and the topic and scope. The reader should ask: what did I learn, and how can I use it.
A few more words about review articles. The articles should always have some educational value, no matter what the topic and audience is. Second, the language used should be clear and concise. Many of us have been dismayed by the lack of clarity, coherence, lucidity, and simplicity of psychoanalytic articles in the past. However, concerning language attributes, the language of some neuroscience articles seems even worse than the language of those old articles to me.
Review articles are a significant, vital part of medical literature. They have various formats and different goals. Review articles should be meaningful, and not just for improvement of the impact factor or H-index. That would be “l’art pour l’art.” Review articles should provide state-of-the-art critical reviews of particular topics, and should target specific readers. They should provide high educational value to readers. Journal editors should pay more attention to what review articles are and whether they are of value to their readership, and the readership should carefully consider what is worth reading and useful for one’s education.
Conflict of Interest Statement
The author declares no conflicts of interest.
There were no funding sources for this work.
Richard Balon, MD, is the sole author of this editorial.