Sunday School Lessons – A Rumination — Critical Reading Helps Digest Neurosurgical Literature

John A. Jane Sr., MD, PhD, FRCS(C), has served as the editor of the AANS Journal of Neurosurgery since 1993. He is professor of neurological surgery at the University of Virginia in Charlottesville.

During my research year at the University of Virginia, with no danger of exceeding the soon-to-be-implemented 80-hour workweek, the tradition was for the laboratory residents to meet at Dr. Jane’s house every Sunday at noon, either before or possibly after lunch (I could never tell, since I rarely during my residency saw him eat) to review and discuss a neurosurgical journal in its entirety.

At first “Sunday School” was somewhat time-consuming, necessitating the rapid development of efficiency in quickly reviewing and evaluating the neurosurgical literature. While hardly groundbreaking, I humbly submit some lessons learned at the knee of a master, lessons that since have proven their worth.

What Is the Exact Question Being Asked?
Every research paper should properly start with some sort of question. Otherwise it starts with an answer, and the resulting paper is likely to be biased, since one is likely to attach undue weight to confirmatory evidence while ignoring contradictory evidence. Determining the question is usually self-explanatory, but particularly in complex randomized controlled trials, it may be much more specific than one would first think.

For instance, given the title “Anterior Versus Posterior Fusion for Thoracolumbar Burst Fractures” one might be tempted to assume that the question being asked is simply which approach is better for this type of fracture. However, after further reading, it may be that the actual question is, “In a neurologically intact patient with a thoracolumbar burst fracture of less than 50 percent canal stenosis, is there a benefit to either an anterior or posterior approach?” which can be further shortened to, “Does the method of fusion matter in a patient who could, because of the nature of the fracture, have either approach?”

How Important is the Question?
Each subspecialty in neurosurgery has its important and unsolved questions, be it sagittal balance, vasospasm, or gross total resection of gliomas. If the importance of the question is not obvious, the question either is not important or the reader lacks the information to put the paper into its proper context.

Also, is the intervention something that can be generally implemented in practice? If not—because it involves a procedure that requires special skills, technology, unavailable resources, or is just plain clinically impractical—the importance of the paper will be somewhat diminished.

How Well Has the Question Been Answered?
Entire books have been written on this topic. Suffice to say that a double-blind, randomized controlled trial trumps a randomized controlled trial, which in turn is more valid than a case-control trial, case series, and lastly, a case report and review of the literature. As previously noted, however, the scope of either a double-blind RCT or an RCT sometimes can be so narrow as to not be clinically useful.

Other important questions:

• Are the numbers large enough to minimize false positive or false negative errors?
• Is the treatment or intervention standardized in each patient?
• Is the measurement instrument well-known, valid, accurate, and reproducible, and is the effect measured, even if statistically significant, a clinically important effect? For example,given enough patients, a statistically significant difference of 1 point in the Oswestry Disability Index probably could be determined. However, this would not represent a significant difference in terms of how a patient actually feels after back surgery. Also, one should be wary if a paper uses a relatively unknown outcomes measure or statistical test to determine significance.

Are the Figures and References Appropriate?
I learned this lesson the hard way. Figures should be easily understood and readable, and should match what is written in the text. When a statement should be referenced, make sure it is referenced and check to see if the references are actually the seminal, important papers in the subject, rather than obscure papers obviously selected to fit the data or confirm the author’s viewpoint.

What Constitutes “Discussion”?
The discussion section rarely makes or breaks a paper but, particularly in more narrowly focused papers, “discussion” should place the current study in its proper context. Additionally, the results should be compared to the current literature, and potential biases and flaws in the paper, noted.

In conclusion, critical, efficient reading of the literature is important during training, when both time and the need to learn and assimilate new information are at a premium. The same skills also serve to improve the quality of one’s own research papers.

K. Michael Webb, MD, is in practice with Neurosurgical Associates PSC in Lexington, Ky.

]]>