by Mitch Gordon*
Clinical studies test how drugs or medical devices work in humans before the drugs or devices are ever commercially marketed. Such studies must be designed so that the results will be robust enough to demonstrate the investigational product’s safety, and often its efficacy.
Researchers nearly always use statistical analysis to express and understand clinical study results. Because of this emphasis on statistics in clinical research, regulatory medical writers need a general background in statistics. They also need an understanding of how biostatistics is applied to clinical study design and analysis.
Typically, the statistician ensures that the study design is adequate to provide valid results, and interprets the clinical data in statistical terms after study completion. The medical writer generally writes the clinical study protocol and, later, the study report, in collaboration with other project team members. Both documents tend to require extensive statistical analysis and discussion.
If you’re a medical writer who uses statistics, but statistics wasn’t exactly your strongest subject in college, you may need a refresher. Application of statistics to clinical study data is a subject of its own, too–and you probably didn’t learn it in school.
To address my need for understanding biostatistics in a clinical study environment, I found two terrific, reader-friendly books. These tomes have taught me what I needed to know, without overwhelming me with technical detail that I don’t need.
The first of these is Intuitive Biostatistics, by Harvey Motulsky, MD (the book is about $45 on Amazon). Dr Motulsky is a pharmacology instructor at UC San Diego, and he’s an excellent writer (probably an enjoyable professor, too). According to the preface, his book is “…a non-mathematical introduction to statistics for medical students, physicians, graduate students, and researchers in the health sciences.” He gives extensive explanations of all the important concepts in basic statistics: confidence intervals, P values, odds ratios, hypothesis testing, correlation, regression, and various standard tests. But he only presents equations when necessary, and almost never goes through the derivation steps. Instead, he explains the logic behind the concepts and how they interrelate.
Although Intuitive Biostatistics
is intended to be non-mathematical, I found it to be very thorough. Every statistical concept that you may need is explained completely and in a step-by-step fashion. The author has an engaging, unassuming writing style that makes the reader comfortable–a plus, if the subject matter makes you nervous. You’d never know that Dr Motulsky is a physician, except from the inside cover, and I suspect he had a great bedside manner in his clinical practice days.
The strength of Dr Motulsky’s book is in how well he presents a good, general understanding of statistics as a toolset. He uses examples from the medical literature, and offers one chapter specifically on clinical research.
But, after you’ve finished his book, I think it’s important to move on to a book that explains the application of statistics specifically to clinical research. For this topic, I chose Statistical Thinking for Non-Statisticians in Drug Regulation by Richard Kay (it’s about $70 on Amazon).
Richard Kay is a biostatistics consultant in the UK. The audience he identifies in his preface is the same as Dr Motulsky’s: everyone in clinical research who works with a statistician, but isn’t one. His stated goals are to improve our communications with statisticians, help us critically review reports and publications, and make more effective use of statistical arguments in the regulatory process. At all of these objectives, he succeeds admirably.
Like Dr Motulsky, with his thorough book on basic statistics, Mr Kay systematically works his way through the statistical concepts used in clinical studies, but with great sympathy for the non-mathematical reader. While there’s a little overlap with the Motulsky book, much of the information is fairly specific to clinical study design and data analysis. He delves deeply into concepts that include randomization, multicenter studies, analysis of covariance, intent-to-treat, statistical significance versus clinical importance, interim analysis, meta-analysis, and the all-important power of a study.
After you read through these two books, you’ll understand a lot more of what is going on around you when you write with a clinical team. And that’s a great way to build your confidence and
the confidence the team has in you.
*Republished with permission from the author. Originally published in Vol 2, No 34, of the The Biotech Ink Insider, a free medical writing career newsletter. Subscribe to the Insider!
About the Author
Mitch Gordon has been a professional writer for 16 years. He is completing his Masters degree in Regulatory Affairs, to be finished in early 2010. His specialties are regulatory and clinical documentation, as well as a wide range of other writing and editing roles that support life-science companies.
Medical and Regulatory Writing