It is generally accepted that psychiatric education should be broad-ranging, with exposure to clinical, biological, psychological, social, and other approaches. A major resource in psychiatric education is the scientific literature, as reflected in psychiatric journals. This literature is important not only in the early phases of training, but also in later continuing medical education. There has been much debate as to whether there is an overly biological focus in contemporary psychiatry (1, 2). Major psychiatric journals, as suggested by their impact factors (3), influence practice and education. If published articles are primarily biological in orientation, this could influence psychiatric practice. This question can be tested empirically by conducting a bibliometric review of current psychiatric and internal-medicine journal trends as compared with previous such studies.

A survey of publication trends in American psychiatry from 1844 to 1960 found that biological (“genetic-organic”) studies predominated, as compared with “psychosocial” ones, peaking at the turn of the 20th century (4). In the 1920s–1940s, there was a dramatic decrease in biological studies, and by 1945, as psychoanalytic influence grew, there was an equal distribution with the psychosocial studies. From 1955 to 1960, coinciding with the rise of psychopharmacology, there was a return to the biological focus.

One study found a biological predominance in research published in the American Journal of Psychiatry in the years 1975, 1980, and 1984, and a reduction in the number of publications focused on epidemiology and therapies (e.g., behavioral and psychosocial interventions) (5). A similar study, examining publications from 1969 to 1990 in the American Journal of Psychiatry and the Archives of General Psychiatry, reported an increasing emphasis on biological studies, as well as a more disorder-specific orientation (6). More recently, in 2008, one study described a decrease in published articles on the “medical humanities” (arts, literature, sociology, philosophy, religions, cultural and gender studies, ethics, education, and history) in psychiatric journals, while noting the opposite phenomenon in internal medicine (7).

We examine previous studies of psychiatric journal content, utilizing a new analysis. These are compared with internal-medicine journals, which serve as a control group, given that they are widely viewed as mainly biological in focus. If psychiatry has become increasingly biological in focus, as previous bibliometric studies suggest, a comparison with internal-medicine journals may enlighten us as to the degree to which psychiatry has become biological, in contrast to internal medicine. We hope to ascertain whether the type of biological research conducted in psychiatry is similar to the research conducted in other branches of medicine.

We reviewed all published articles in the 2008 calendar year in two of the most prominent and highly-cited non-subspecialty English-speaking psychiatric journals, The Archives of General Psychiatry (AGP) and the American Journal of Psychiatry (AJP). These were compared with their internal-medicine counterparts, The Archives of Internal Medicine (AchIM) and Annals of Internal Medicine (AIM). Journal selection was based on their impact factor and shared publisher (2007 Impact factors: AGP: 15.9; AIM: 15.5; AJP: 9.1 and AchIM: 8.3) (8).

Only original articles were included, rather than opinion articles or editorials. For the AGP, there were a total of 121 articles; for the AJP, 97 articles; AIM, 120 articles, and AchIM, 137 (in a 6-month period, from January 1st to June 30th, because of a higher number of publications per month in AchIM). The articles were categorized by the lead author, KS, and then reviewed and discussed with SNG to reach a consensus.

Classification of articles was twofold: biological versus non-biological focus, and based on methodological definitions. Articles were defined as biological if they had a primary focus on genetics, physiology, or imaging. Physiological studies included topics such as electrophysiology, respiratory volumes, and cellular or molecular markers. Neurocognitive and neurobehavioral studies were also classified as physiological. Clinical, epidemiological, and review treatment studies, including drugs evaluating clinical outcomes (e.g., symptoms), rather than biological variables (e.g., neurotransmitter levels), were considered non-biological.

From a methodological perspective, studies were classified as epidemiological, clinical, reviews, or other. Clinical studies were defined as observational studies with a small sample (<100) or interventional studies (clinical trials). Clinical trials were divided further into randomized or not, and treatment or not. Biological non-treatment clinical trials, classified as focusing on measuring biological markers (e.g., molecular markers, eye-blink reflex, or changes in MRI) in patients receiving a non-treatment intervention (e.g., non-therapeutic drug, sleep deprivation, cognitive tasks, facial expression processing). Non-biological, non-treatment clinical trials referred to those in which there was an intervention such as change in delivery of medical service. Epidemiological studies were defined as observational, large-population studies primarily focused on characterizing medical conditions by describing their risk factors, prevalence, course and outcome, as well as, studies with a focus on quality assessment, diagnostic tool assessments, prediction rules, economy, and pharmacology. Having excluded opinion articles, almost all review articles were systematic. Any article that did not fit into a previously described classification was categorized as “other.”

Comparison Between Psychiatric Journals

The main results are provided in Figure 1. The two major psychiatric journals analyzed were approximately half biological and half non-biological in focus, with AGP being somewhat more biological then AJP (52.9% versus 42.3%, respectively: relative risk [RR]: 1.25; confidence interval [CI]: 0.93–1.66).

Regarding study methodology, psychiatric journals were about equally clinical and epidemiological, with AGP being more epidemiological (62% versus 35.1%; RR: 1.76 [CI: 1.30–2.39]) and AJP more clinical (56.7% versus 34.7%; RR: 1.63 [CI: 1.21–2.20]). There were no major differences between these two psychiatric journals regarding types of clinical studies, which were similarly distributed in terms of treatment versus non-treatment, presence or absence of randomization, and large versus small sample sizes.

Comparison With Medical Journals

Psychiatric journals published twice as many biological studies as medical journals (48.2% versus 22.2%; RR: 2.17 [CI: 1.66–2.83]). Medical journals published mostly epidemiological studies, followed by reviews. In contrast, psychiatric journals published three times more clinical studies than internal-medicine (44.5% versus 13.2%; RR: 3.46; [CI: 2.43–4.92]).

When compared with medicine, psychiatry had three times the number of genetic and imaging studies (genetics, RR: 2.71 [CI: 1.28–5.71]; imaging, RR: 3.3 [CI: 1.6–6.9]), but fewer than half the number of physiology studies (RR: 0.34 [CI: 0.23–0.48]). Clinical studies in psychiatry tended to have a more biological focus than internal-medicine (58.8% versus 14.7%; RR: 3.94 [CI: 1.73–8.99]), and they were more observational (RR: 2.10 [CI: 0.78–5.63]).

Although randomized drug trials predominated in both fields, psychiatric studies were more likely to be non-randomized (42.8% versus 6.6%; RR: 6.42 [CI: 1.64–25.19]). Psychiatric journals published almost four times the number of non-treatment trials, as discussed in the Methods, studies measuring biological markers to a non-pharmacological or non-therapeutic intervention, for example, eye-blink reflex when exposed to a painful stimuli (38.6% versus 10%; RR: 3.86 [CI: 1.26–11.74]). Finally, trials with small sample sizes greatly predominated in psychiatric versus medical literature (61.4% versus 3.3%; RR: 18.4 [CI: 2.6–127.72]). Regarding epidemiological studies with a non-biological focus, there were more quality-assessment (QA) and pharmaco-epidemiological studies in medicine than in psychiatry (QA: 32.7% versus 5.79%; RR: 4.7 [CI: 1.73–12.7]); (pharmacology: 17.6% versus 8.7%; RR: 2.09 [CI: 0.86–4.70]). Epidemiological studies with a biological focus were primarily genetic in psychiatry (75%) and mostly physiological (82.9%) in medicine.

It is a commonplace view that psychiatry has become more biological (9–11), and this bibliometric study is consistent with that impression. About one-half of psychiatric articles are biological in methodology, twice as frequently as in internal-medicine journals. Psychiatric studies—by focusing on genetics, neuroimaging, and physiology—tended to be more etiological and pathophysiological, in contrast to internal-medicine articles, which were more therapeutically-oriented and epidemiological.

These results suggest that published articles are primarily biological in content. Additional publications might expose readers to the full panoply of issues surrounding mental illness, such as environmental risk factors, public health initiatives and perspectives, cultural implications, and conceptual and philosophical analyses.

Our definition of “biological”—focusing on genetics, imaging, and physiology—may be seen as overly narrow. Future studies could include clinical studies on humans that assess drug response. Although clinical research can involve biological illnesses or treatments (e.g., pharmacological therapies), clinical studies, by definition, involve human beings and clinical variables as outcomes. Generally, clinical studies do not emphasize or assess biological variables as their primary outcome; however, we have classified those that do as biological.

Another critique might be that all things are, in some way, biological—even clinical studies of, for instance, social isolation in relation to a clinical symptom. Obviously, biological variables can mediate the latter, but our analysis seeks to ask a narrower question about directly biological studies, such as genetics, versus clinical studies without direct biological focus either as main predictors or outcomes.

An emphasis on biological research may challenge the importance of a broader understanding of psychiatry. Decades of attention and advances in neuroscience have led to little progress in pinning down etiologies of psychiatric clinical syndromes. One could argue that treatment advances do not necessarily depend on etiological understanding; for instance, hand-washing, in the 19th century, was a major contributor to reduction in puerperal fever long before microbial etiologies were even conceptualized (12). Thus, despite the continued importance of research on etiologies, psychiatry is at risk of neglecting epidemiological research, therapeutic studies (especially non-pharmacological interventions), and quality measures: three areas of special focus in contemporary internal medicine.

Psychiatric education may be influenced by these results. If the major psychiatric journals tend to publish mostly biological studies, then readers of those journals would likely be influenced, at least to some degree, in a biological direction. This is not inherently harmful, but it may be associated with some harm if other important aspects of the field are neglected.

One obvious interpretation would be to see these results as supporting a need for more direct psychosocial teaching in psychiatric education. In addition to this possibility, we suggest another interpretation. The dichotomy between the “biological” and the “psychosocial” may not be the key; rather the divergence may be between “clinical” and “non-clinical,” whether the content of the research is biological or psychosocial. In other words, much of the published literature is non-clinical—involving animals, genes, or neuroimaging—without any direct relevance to clinical factors. This research is hard to relate to learning the clinical practice of psychiatry; hence, the common disjunction between the “science” and an increasingly opinion-based “practice” of psychiatry (13).

One conclusion might be that journals themselves may need to take more responsibility to publish a variety of types of articles, with special attention to increasing clinical articles, so as to enhance the ongoing psychiatric education of all practitioners.

Limitations of the Current Study

This study represents a cross-sectional look at publication patterns in only two of the top journals in the fields of psychiatry and internal medicine; thus inferences about other journals may not follow. Also, our findings may not be generalizable to all publications used in psychiatric education. Furthermore, the journals studied were limited to the United States, and may not reflect European approaches to psychiatry. A larger sample of journals, as well as a look at trends over time, would be follow-up steps for further research to either extend these findings or refute them.