| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Statistical Concepts Series |
1 From the Department of Radiology, Riley Hospital for Children, 702 Barnhill Rd, Indianapolis, IN 46202 (K.E.A.); and Health Services Research and Development Service, Department of Veterans Affairs, Washington, DC (P.E.C.). Received October 9, 2003; revision requested October 14; revision received October 17; accepted October 29. Address correspondence to K.E.A. (e-mail: kiappleg@iupui.edu).
Index terms: Education • Statistical analysis
One should not go hunting for buried treasure, because buried treasure is found at random, and, by definition, one cannot go searching for something which is found at random.
Attributed to the Talmud; cited by Salsburg (1)
With this issue of Radiology, the Statistical Concepts Series of articles reaches its conclusion. We take this opportunity to thank each of the talented authors for sharing with us their expertise and patience. It takes considerable skill to translate complex concepts into a format understandable to a wide audience. Without their efforts and considerable time commitment, this series would not have been possible. Thank you.
In the current issue of Radiology, the 17th article in the series, by Dr Eng (2), provides an example of how we can use advanced statistical modeling to understand and predict what may work in radiology research (2,3). While this final article is complex, its sophistication provides a window into a world we radiologists rarely visit. The articles that preceded this final article were designed to provide readers of Radiology with an understanding of the basic concepts of statistics, probability, and scientific methods that are used in the medical literature (4). Because the Accreditation Council for Graduate Medical Education now requires of residents a basic grasp of statistical concepts as a component of the six core competencies (5), one must have a sound understanding of the methods and results presented in today’s medical literature, whether one is a resident or seasoned radiologist.
We are all consumers of information. Statistics allow us to organize and objectively evaluate empiric evidence that can ultimately lead to improved patient care.
Nearly all readers of the radiology literature know that understanding study results and determining their applicability to practice requires an understanding of statistical issues. The articles that compose the Statistical Concept Series in Radiology are meant to increase understanding of the statistics commonly used in radiology research.
Statistical methods revolutionized science in the 20th century (1). Statistical concepts have become an essential aspect of scientific inquiry and even of our common culture. The influence on society is evidenced by the use of statistics in the lay press and media; the use of terms such as probability and correlation in everyday language; and the willingness to collect data, accept scientific conclusions, and set public policy on the basis of averages and estimates (1).
In just over 1 century, statistics have altered our view of science. Together with the rapid evolution of computer capabilities, there are many new statistical methods on the horizon. Recent trends in medical statistics include the use of meta-analysis and clustered-data analysis (6). In addition, some statistical methods, formerly uncommon in medical research, are quickly becoming embedded in our literature. These include the bootstrap method, Gibbs sampler, generalized additive models, classification and regression trees, models for longitudinal data (general estimating equations), hierarchic models, and neural networks (6). Regardless of the sophistication of a technique, to take full advantage of their potential it is necessary to understand fundamental statistical methods. The challenge for physicians is to develop and maintain statistical "literacy," in addition to the scientific literacy of radiology and medicine.
We define a functional level of statistical literacy as that which includes an understanding of methods, the effect of statistics on research design and analysis, and a basic vocabulary of statistical terms (7).
As a profession, how can we encourage statistical literacy? First, we must educate ourselves by requiring the teaching of medical statistics in medical school and residency training; Second, we should encourage the development of consensus guidelines on the proper reporting of scientific research—for example, the CONSORT (Consolidated Standards of Reporting Trials) statement (8) for reporting results of randomized controlled trials, the QUOROM (Quality of Reporting of Meta-analyses) statement (9) for reporting results of meta-analyses and systematic reviews, and the STARD (Standards for Reporting of Diagnostic Accuracy) statement (10) for reporting results of diagnostic accuracy studies. Some journals have published statistical guidelines for contributors to medical journals (11), while others have statistical checklists for manuscript reviewers. In January 2001, Radiology became the first American radiology journal to provide statistical review of all published manuscripts that contain statistical content, to the benefit of both authors and readers (12). Third, we should continue to promote the learning of critical thinking skills and research methodology at our national meetings, such as the seminars held at the 2002 Radiological Society of North America and the 2003 Association of University Radiologists annual meetings. Fourth, we must continue to promote the value of scientifically rigorous reports, relative to that of less scientific ones, through our national organizations, the program content at our scientific meetings, and the support of these concepts through written and oral announcements by our leadership.
The goal of the Statistical Concept Series was to enhance the ability of radiologists to evaluate the literature competently and critically, not to make them statisticians. When contemplating the value of such a basic understanding of statistics, consider that Bland (13) argued that "bad statistics leads to bad research and bad research is unethical." We must beware of translating bad research into bad medicine and recognize that we have an essential role in increasing the evidence base of medical practice. Such an understanding is perhaps one of the most useful things that radiologists must learn.
 |
REFERENCES |
|---|
 TOP REFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
