Dr Edward Mascha and Dr Thomas Vetter outline A&A’s innovative series designed to strengthen the research skills of anaesthesia professionals worldwide.
During the past 10 years, Anesthesia & Analgesia has published three distinct but complementary series of statistical articles to benefit readers and prospective authors. From May 2017 to March 2019, Thomas R. Vetter, MD, MPH, the Journal Deputy Editor-in-Chief, Edward J. Mascha, PhD, the Journal Statistical Editor and Patrick Schober, MD, PhD, the Journal Associate Statistical Editor, wrote a series of 21 “Statistical Tutorials” to educate the anesthesiology community on the fundamentals of biostatistics, epidemiology, and study design. Drs. Vetter and Schober followed the tutorials with 24 “Statistical Minutes” that described, in a very concise manner, the specific design or analytic technique applied in a highlighted published paper. At the same time, Dr. Mascha and other statisticians have continued the “Statistical Grand Rounds” series, which started in 2011, and now contains 20 in-depth articles on more advanced and cutting-edge study design and analysis topics.
Why did we do it?
Consistent analytic rigour is a requirement yet a challenge for all medical journals. Researchers and authors often need guidance but do not have dedicated biostatistical support. Therefore, Anesthesia & Analgesia identified its major responsibility to educate all its stakeholders about basic study design, statistics, and data reporting. The Journal collection of 21 focused statistical tutorials covers these fundamental concepts in a practical way for individuals not formally trained in statistics or epidemiology. The tutorials also parallel the content recommendations of the 2014 Statistical Analyses and Methods in the Published Literature (SAMPL) Guidelines. Each tutorial cites specific published papers that represent prototypical examples of study design, statistical analysis, and/or data reporting. Our tutorials have been noted as being especially helpful for researchers in a low- and middle-income country (LMIC) where a resource-challenged medical environment often makes conducting clinical research difficult. These statistical tutorials, along with the other 2 series, thus fully support the desired global reach of the Journal.
Our Statistical Minutes series is intended to catch the attention of the busy clinician-reader using an attractive figure or diagram and a to-the-point short summary of a key statistical method—perhaps prompting the clinician-reader to read more on the topic in a Statistical Tutorial or Statistical Grand Rounds. Each of the Statistical Minutes accompanies a paper in that print issue, and focuses on a specific topic in a succinct manner, with the goal of assisting the reader to understand and extrapolate the statistics applied in the highlighted study.
Our Statistical Grand Rounds comprise in-depth, advanced expositions on appropriate design and statistical methodology for topics of major and practical importance for researchers. The goal is to advocate for, enable, and guide best statistical practices for researchers and data analysts. Each presented topic is motivated to address recurring issues in the literature or to demonstrate improvements to currently utilized methods. Applications to actual study data are typically included, with programming code to facilitate implementation.
Who are these resources for?
All three series are geared towards investigators with various levels of statistical or research experience. They include applications across anesthesiology, perioperative medicine, pain medicine, and critical care, which can easily be generalized to any clinical speciality. While they focus on clinical research, most of the articles would also be helpful in the design and analysis of basic research studies.
Our Statistical Tutorials are particularly helpful to researchers with limited experience in study design, including formulating hypotheses, choosing the statistical methods and sample size, analysis, and scientific writing. Because scientific writing is quite distinct from other writing forms, several Statistical Tutorials focus on topics like hypothesis formulation and writing a manuscript. Together they have proven to be a strong resource for attending physicians, trainees, and student learners. Our Statistical Minutes are geared toward a similar audience and for those looking for a quick and easily understood summary of a topic. Both series have been helpful for Journal editors as well as educators of medical researchers.
Our Statistical Grand Rounds articles are aimed at the senior author of a clinical research paper and the involved statistician/analyst. For that reason, they are not overly theoretical and typically include a non-technical introduction as well as practical applications to real study data. That said, the “meat” of the paper may be best digested by someone with statistical or epidemiological expertise. This series is also utilized by researchers in an LMIC: for example, when designing or analyzing a study to assess the potential impact of an intervention using a before-and-after design when a randomized trial is not feasible.
What areas have been covered?
Each series is expansive in itself and provides a comprehensive resource. Key topics covered by the Statistical Tutorials include: study design options, hypothesis formulation, defining outcomes, sample size/power, basic statistical methods (e.g., t-test, correlation, regression, agreement, repeated measures, survival, systematic review, control charts, diagnostic testing), reporting (e.g., descriptive statistics, effect summary measures), and writing a manuscript. The Statistical Minutes also cover a wide range, including linear and logistic regression, repeated measures, confounding, and multiple testing.
Our Statistical Grand Rounds series of more advanced topics include, for example: multilevel modelling, appropriate modelling of non-normal data, equivalence/noninferiority testing in regression, composite outcomes, joint hypothesis testing, mediation, Wilcoxon test sample size, learning curve analyses, error grid analyses, Type I error control, propensity score methods, quantile regression and segmented regression.
What to do with it?
Simply read and apply! Our goals are practical. These 3 series are intended to facilitate and guide authors with planning, conducting, and reporting on their research in order to help improve the rigour of manuscript submissions to the Journal. A close reading of the relevant articles will help authors design their research studies, calculate an appropriate sample size, develop a statistical methods section that appropriately addresses the study objectives, and accurately report their results. In short, the potential impact of each series lies with prospective authors, readers, and educators to study and discuss the relevant articles and use their content in their research, teaching, peer-reviewing, or editing. The positive feedback received thus far from many stakeholders has been validating and gratifying.