Welcome to the all-new Vetlearn

  • Vetlearn is becoming part of NAVC VetFolio.
    Starting in January 2015, Compendium and
    Veterinary Technician articles will be available on
    NAVC VetFolio. VetFolio subscribers will have
    access to not only the journals, but also:
  • Over 500 hours of CE
  • Community forums to discuss tough cases
    and networking with your peers
  • Three years of select NAVC Conference
  • Free webinars for the entire healthcare team

To access Vetlearn, you must first sign in or register.


  Sign up now for:
Become a Member

Compendium April 2008 (Vol 30, No 4)

Guest Editorial: Love at First Cite — A Serious Reader's Guide to References

by Thomas K. Graves, DVM, PhD, DACVIM

    Several years ago, while reading a review article on treatment of feline hyperthyroidism, I came across a clear statement, duly referenced, that methimazole causes platelet dysfunction in cats. I was excited to see that the problem of nonthrombocytopenic bleeding in cats on methimazole had been solved, but I was a little disappointed because I had been working to find an answer to that problem in my own laboratory and had been unsuccessful. I turned to the last page of the article, anxious to check the reference, and found, surprisingly, that one of my own publications had been cited as the basis for the statement.1 In that reference (a book chapter), I had stated that some cats on methimazole exhibit abnormal bleeding despite normal platelet counts, a statement I had referenced carefully,2 and that studies were ongoing to investigate the effects of methimazole on platelet function. After I picked myself up off the floor, and feeling guilty, I went back to the original chapter to make sure I had not said anything misleading. In the chapter, I had suggested that methimazole might cause platelet dysfunction but had stated clearly that further investigation was needed.1

    Inaccuracies in journal article citations are common,3,4 not just in veterinary medicine, but probably in all fields. Careful scrutiny of references by editors and reviewers can reduce the rate of errors in citations5 but cannot eliminate them completely. Checking the actual content of each citation and determining its validity as a reference is an impossible job for an editorial staff. It is the responsibility of the authors. Authors, being human, often fail to meet this responsibility, leaving it to the reader to determine the validity of a referenced statement. A long list of citations might give the impression that a manuscript has been thoroughly vetted by its authors, but that is not always the case. Conversely, a short list of citations should not be assumed to be inadequate, because the references may be very good and accurately applied. So how is a reader to know if those little superscripted numbers should assure the validity of a statement or arouse suspicion?

    The best way to determine the quality of a reference is to read it. Obviously, that would be impossible for most of us. Even if I had that much free time, I doubt I would want to spend it searching for original publications. I always, however, make sure to at least check for some characteristics that might give me a clue about the overall quality of an article's citations. Here are the things I look for:

    Book chapters. Referencing book chapters is a common practice in the biomedical literature. I often do it in my own manuscripts, but it's not always justified. As an example, I recently read an article in a respected dermatology journal that contained a statement that hypothyroidism is the most common endocrine disease in dogs. This statement is found frequently in the veterinary endocrinology literature, and it's one of my pet peeves. My opinion is that hypothyroidism is much less common than most veterinarians think, although it may well be the most commonly diagnosed (not necessarily correctly) endocrine disease in dogs. The statement in the dermatology article was supported by a reference to a chapter on hypothyroidism in a textbook.6 The chapter says nothing of the kind. In fact, the chapter contains a table showing that diabetes mellitus is the most common hormone deficiency syndrome seen in the authors' practice. The citation in the dermatology journal was clearly misrepresented.

    In the same article, I found a statement that patients with hypothyroidism can present with bacterial dermatitis and otitis as the only signs. Finding that statement difficult to believe, I looked up the citation. The listed reference was another textbook chapter. I looked up the chapter in the sixth edition of the book and found that it referenced the fifth edition. The fifth edition cited the fourth edition. The fourth edition cited the third edition. In the third edition, the statement on which the suspicious claim was based was referenced by the author's unpublished data and by an obscure, 27-year-old research article on mice. I dug up the mouse article and found that it had nothing whatsoever to do with hypothyroidism and was an erroneous citation to begin with. This is an extreme example, but references to textbook chapters should always be viewed with a little suspicion, especially if a controversial claim is being made. Textbook chapters are not subjected to scientific scrutiny by peer reviewers. They often contain statements based largely on expert opinion. As an author of numerous textbook chapters, I am most certainly guilty of that.

    Nonrefereed peer-review publications. There are different levels of peer review. The highest level of scrutiny belongs to journals with official review by a review panel. These journals typically publish lists of reviewers. Other journals rely mostly on editors and editorial panels, sometimes using outside peer reviewers, but usually not publishing lists of reviewers. These are referred to as nonrefereed peer-review journals. While they can be of very high quality, the level of scrutiny is not always as great as with a refereed journal.

    Proceedings. Published proceedings from scientific meetings are often used as references for journal articles. This practice should be viewed as extremely suspect. Not only are proceedings usually not subjected to peer review, they are usually not even seen by an editor.

    Abstracts. Abstracts are usually subjected to some level of scientific review, but it is almost never as rigorous as that for refereed journal articles. Abstracts can sometimes be almost hidden in a reference list. Consider the following example:

    Gilor C, Graves TK. Serum fructosamine does not correlate with body weight, body condition score or age in cats. J Vet Intern Med 2007;21:595.

    To the unsuspecting eye, this looks like an article in a highly respected journal. It is not. It is an abstract from the annual ACVIM Forum. It should not be given the same weight as a regular manuscript. It is also an example of an abstract not being identified clearly as such, which it should be. Furthermore, an old abstract should probably be considered a fairly weak reference because abstracts that do not eventually become full-length manuscripts might not be completely reliable.

    Old references. If a statement is so basic that a decades-old reference will suffice to support its validity, it probably doesn't need a reference at all. Otherwise, old references risk being outdated.

    Theses. References to master's- or PhD-degree theses should sometimes be viewed with suspicion. A degree candidate's thesis is supposed to be reviewed by a graduate committee, but that is not the same as anonymous peer review. If an older thesis is referenced, one has to wonder why no refereed journal publications came from the thesis work being cited.

    Self-references. Sometimes an author is the only person published on a given topic and has no choice but to reference his or her own works. A reference or two by the main article's author is a good thing. It lets the reader know something about the author's background and expertise. Too much self-referencing, however, can be a sign of tunnel vision.

    Personal communications. References to personal communications provide just about the weakest possible support for a statement in an article.

    Unpublished observations. Similar to personal communications, references to unpublished observations should probably not be used unless they pertain to something that does not affect the meaning of the article or interpretation of data.

    Any given reference can be perfectly valid, but a little cynicism is a healthy thing when it comes to reading the scientific literature. Most veterinary journal editors do excellent work when it comes to policing the accuracy and appropriateness of references. Unfortunately, some references fall through the cracks. A high level of expertise is required not only of authors, editors, and reviewers but also of readers. Personally, I think that's a good thing.

    Thomas K. Graves, DVM, PhD, DACVIM

    University of Illinois at Urbana-Champaign

    Downloadable PDF

    Dr. Graves discloses that he has received financial support from Intervet, Inc., Pfizer Animal Health, Schering-Plough Animal Health, and Smart Cells, Inc.

    1. Graves TK. Complications of treatment and concurrent illness associated with hyperthyroidism in cats. In: Bonagura JD, ed. Kirk's Current Veterinary Therapy XII. Philadelphia: WB Saunders; 1995:369-372.

    2. Peterson ME, Kintzer PP, Hurvitz AI. Methimazole treatment of 262 cats with hyperthyroidism. J Vet Intern Med 1988;2:150-157.

    3. Hinchcliff KW, Bruce NJ, Powers JD, Kipp ML. Accuracy of references and quotations in veterinary journals. JAVMA 1993;202:397-400.

    4. Wager E, Middleton P. Technical editing of research reports in biomedical journals. Cochrane Database Syst Rev 18:MR000002, 2007.

    5. Wager E, Middleton P. Effects of technical editing in biomedical journals: a systematic review. JAMA 2002;287:2821-2824.

    6. Feldman EC, Nelson RW. Hypothyroidism. In: Canine and Feline Endocrinology and Reproduction. 3rd ed. St. Louis: Saunders; 2004:86-151.

    References »

    NEXT: Rectocutaneous Fistulas
    Stay on top of all our latest content — sign up for the Vetlearn newsletters.
    • More