Objectives: To characterize recently published literature on traumatic brain injury. Hypotheses: None. Participants: Scholarly papers indexed in 1 or more bibliographic databases. Methods: Publications (with or without abstract) were identified using key words and text words in PubMed, PsycINFO, and CINAHL, using an algorithm that optimized identification of papers on traumatic brain injury (TBI) in human adults. Abstracts were read by 1 of 2 experienced researchers, who coded as YES, NO, or UNCLEAR: nature of paper (quantitative empirical vs all other); subjects involved (humans vs all other); diagnosis (TBI vs other); age group involved (child vs adult); subject matter (intervention/secondary prevention vs all other); phase of healthcare (rehabilitation vs other). Results: Of 1679 papers, covering an estimated 70% of 2004 publications, 1000 (60%) were judged to concern (certainly or possibly) human adults with TBI. Of these, 696 (70%) were considered to possibly be quantitative empirical studies; the other 304 (30%) were reviews, theoretical papers, or (a very small group) qualitative research. Of the quantitative empirical studies, only 176 (25%) concerned intervention or secondary prevention, and of these, only 79 (45%) involved rehabilitation (defined broadly). Of the 304 nonempirical publications, 121 (40%) involved intervention or secondary prevention, and of these, only 64 (53%) involved rehabilitation. Conclusions: Quantitative empirical studies of rehabilitation interventions are a very small percentage (14.3%) of scholarly output. Most published papers, whether empirical or theoretical in nature, concern diagnosis, prognosis, ethics, costs, and other subjects. While clinicians require information on a number of subjects to competently discharge their responsibilities, it would seem that, for an interventive specialization as medical rehabilitation, there is insufficient emphasis on treatment research. Progress in TBI rehabilitation requires a shift in focus, however difficult implementing high-quality intervention research may be.