From trials to treatments

Randomised controlled trials Alejandro R. Jadad BMJ Books, London; 1998 (Available through the CMA‚s Member Service Centre) 123 pp. $39.95 (CMA members: $32.95) ISBN 0-7279-1208-9

Alejandro Jadad‚s Randomised Controlled Trials is a little book that tries to do a lot. In 123 pages of text the author sets out to provide not only an introduction to the design and reporting of randomized controlled trials (RCTs) but also guidance on quality assessment, meta-analysis and the role of RCTs in policy formulation.

In his introduction Jadad notes that his motivation for writing the book came from a desire to have a satisfactory single-source treatment of the RCT and its role in health care decision-making. The book is aimed at busy readers who want to become conversant with the basic principles of the RCT, the cornerstone of evidence-based care. Only about half of the book is devoted to traditional issues relating to the strengths and weaknesses of trials; the remainder deals with broader issues of evidence-based health care. Jadad has set a difficult course for himself, not only in terms of the breadth of his subject, but also in terms of his target audience. Although he explains that, after considering the needs of both doers and users, he decided to produce ”an introductory guide for busy readers,” the tension of that choice is evident. The lack of certain caveats (e.g., concerning methods of randomization) and the inclusion of certain topics (e.g., whether or not to contact authors of trial reports in doing a meta-analysis) are surprising for an introductory guide. Although the inclusion of the latter can be considered a bonus, the exclusion of the former is somewhat troubling.

Section headings in the form of pertinent questions provide helpful signposts for the reader. In the first chapter, on RCT basics, we encounter fundamental questions such as: ”What is a clinical trial?” and ”How can randomisation be achieved?” Although this format is highly effective, there are problems with the corresponding content. Under the heading, ”Are the elements of RCTs very different from other studies?,” there is no mention of the ethical issues that arise when we intervene and randomly assign patients to different arms of a trial rather than simply observe the result of systematic treatment choices. The section on the means of achieving randomization offers suggestions about how to use coin-flipping and die-tossing as means of randomization without alerting readers to the fact that, although these methods serve as useful pedagogic models, in practice they are to be avoided.

Even though the book is aimed at busy readers it would have been in their best interest to leave a little more flesh on the bare bones. Without it, the risks of a nasty poke are not negligible. In the discussion of what is often called ”randomized consent” (in the section ”What is a trial with Zelen‚s design?”) readers are cautioned about the potential pitfalls of this study design. They should probably also be alerted to the fact that the National Institutes of Health has declared the original version of this study design unacceptable.¹

The book comprises eight chapters. Those dealing with RCTs per se concentrate on aspects of design and reporting. There is an entertaining and informative catalogue of potential biases, some unique to RCTs, some common to other research strategies, and several chapters oriented explicitly to important issues in evidence-based health care. These include a discussion about how to judge whether a trial‚s results are likely to be useful to the reader‚s particular circumstances, a very good treatment of reviews, meta-analyses and guidelines, and an examination of the links between information and decisions. Jadad has a great deal of experience in the study of evidence-based health care and has made major contributions to RCT quality assessment. This comes through clearly in the good practical advice he give us. But his discussions on the steps from individual trials to health care decision are the book‚s greatest strength. Unfortunately, the title of the book gives readers no indication of what they have to look forward to in these chapters.

Some additional fine-tuning is in order. I was a bit taken aback when I looked up ”ethics” in the index and was referred to pages with nary a mention of the topic. Other pages that should have been referenced were not. Similarly I was surprised to find that the occasional reference had gone missing.

Is Randomised Controlled Trials a useful addition to the literature? Unequivocally, yes. It provides a good survey of important issues in evidence-based health care, is easy to read, and is organized in a manner that helps the reader target specific areas of interest. Does it fill the gap the author alludes to in the introduction by providing ”a single source that could help [the reader] really understand what RCTs [are] about, their strengths and limitations, and how to use them while making health care decisions”? Not really. If one wanted to pursue that goal, a second edition with an expanded treatment of some of the basic topics would be a step in the right direction.

Figure.

Photo by: Fred Sebastian