(1994; 115 pages) [French] [Spanish]
Many prescribers have a problem in reading everything they would like. The reasons are lack of time and - in industrialized countries - the sheer volume of materials mailed to them. It's wise to adopt a strategy to use your time as efficiently as possible.
You can save time when reading clinical journals by identifying at an early stage articles which are worth reading, through the steps listed below.
1. Look at the title to determine if it appears interesting or useful to you. If not, move on to the next article.
2. Review the authors. The experienced reader will know of many authors whether they generally provide valuable information or not. If not, reject the article. If the authors are unknown, give them the benefit of the doubt.
3. Read the abstract. The main point here is to decide whether the conclusion is important to you. If not, reject the article.
4. Consider the site to see if it is sufficiently similar to your own situation, and decide whether the conclusion may be applicable to your work. For example, a conclusion from research in a hospital may not be relevant for primary care. If the site differs too much from your own situation, reject the article.
5. Check the ‘materials and methods’ section. Only by knowing and accepting the research method can you decide whether the conclusion is valid.
6. Check the references. If you know the subject you will probably be able to judge whether the authors have included the key references in that field. If these are missing, be careful.
It is beyond the scope of this book to go into the details of how reports on clinical trials should be assessed, but a few general principles are given here. Generally, only randomized, double-blind clinical trials give valid information about the effectiveness of a treatment. Conclusions drawn from studies of other design may be biased.
Second, a complete description of a clinical trial should include (1) the patients in the trial, with number, age, sex, criteria for inclusion and exclusion; (2) administration of the drug(s): dose, route, frequency, checks on non-adherence to treatment, duration; (3) methods of data collection and assessment of therapeutic effects; and (4) a description of statistical tests and measures to control for bias.
Finally you should look at the clinical relevance of the conclusion, not only its statistical significance. Many statistical differences are too small to be clinically relevant.
Sometimes conflicting evidence is presented by different sources. If in doubt, first check on the methodology, because different methods may give different results. Then look at the population studied to see which one is more relevant to your situation. If doubts remain, it is better to wait and to postpone a decision on your P-drug choice until more evidence has emerged.