Original research articles in JAMA are selected for publication because the results are valid and findings provide important new clinical, research, or policy-related insights. To be current, clinicians must read and understand the primary research literature. By implication, this means also understanding increasingly complex methodologies and statistical analyses now used in clinical research. Clinicians may not be familiar with research methods introduced after they completed training. Because relatively little emphasis is placed in medical school on research methods and statistics, clinicians may never have learned enough about these topics to properly understand current research articles.
As an aid for readers, in this issue of JAMA, we introduce the JAMA Guide to Statistics and Methods. This new series of articles will provide explanations about statistical analytic approaches and methods used in research reported in JAMA articles, and they will be written in language practicing clinicians can understand. These explanations will be published concurrently with research articles that use the statistical test or methodological approach, thereby providing an example of the topic being discussed.
The challenge in balancing statistical rigor with reader comprehension dates back to one of the first uses of a χ2 test in a JAMA article. A randomized clinical trial evaluating azacyclonol for schizophrenia treatment used χ2 analysis to demonstrate a statistically significant treatment effect.1 The author concluded that since the P value of .0003 was less than .05, sufficient evidence existed to establish a hypothesis. A letter published in response to this paper pointed out that P values do not establish hypotheses: “ No p, however small, can ever establish that a hypothesis is correct…p merely is the probability that if a given hypothesis is correct, then chi-square will be found at least as large as it was in fact found. The distinction may be made clear to nonmathematical readers by the following example from the game of bridge. The chance that if a deal is honest a particular player should be dealt 13 hearts is only 1 in 635,013,559,600; but if he is indeed dealt such a hand it would be quite erroneous—and perhaps even fatal—for him to conclude that the probability is only 1 in 635,013,559,600 that the deal was honest, that is, that it is virtually certain that the deal was crooked.”2
Recognizing the need to help clinician readers better understand how to interpret scientific articles, JAMA launched the Users’ Guides to the Medical Literature series in 1993.3 The Users’ Guides help clinicians better understand the medical literature. There are articles about how to search the literature, how to interpret studies, and how to understand the nuances characteristic of review articles. The JAMA Guide to Statistics and Methods complements the Users’ Guides by providing a more granular and specific discussion about statistics and research methodology used in an individual article.
The first JAMA Guide to Statistics and Methods article discusses intention-to-treat (ITT) analysis4 as it relates to a research article in this issue.5 In addition to providing a general description of the ITT principle, the Guide to Statistics and Methods in this issue explains how this principle was applied in the research study it accompanies. By pairing a JAMA Guide to Statistics and Methods with a specific article, we hope the learning experience will be enhanced.
JAMA Guide to Statistics and Methods articles will be written in plain English, avoid complex mathematics, and present material graphically whenever possible. We distinguish between statistics and methods: statistics are mathematical approaches to describing collections of data whereas methods refer to how a study was designed or some other general approach to how a study was organized and conducted.
JAMA Guide to Statistics and Methods articles will explain why a particular test or method was used, what its limitations are, discuss risks of bias, and examine why the study authors used the particular test. These articles will explain how the findings from statistical tests should be interpreted in the accompanying JAMA research article. Also discussed will be the limitations of interpreting the data given the methodology used to examine it.
Because medical information is vast and rapidly expanding, physicians must pursue life-long learning. This requires reading and understanding research articles published in medical journals. Research articles cannot be assessed if the statistical analysis and research methodology used are not understood by the readers. Along with the Users’ Guides to the Medical Literature, the new JAMA Guide to Statistics and Methods will help readers better understand clinical research reports that in turn will help them provide better patient care.
Corresponding Author: Edward H. Livingston, MD, JAMA, 330 N Wabash Ave, Chicago, IL 60611 (email@example.com).
Edward H. Livingston. Introducing the JAMA Guide to Statistics and Methods. JAMA. 2014;312(1):35. doi:10.1001/jama.2014.7991