Reading scientific articles reporting the results of randomised controlled trials is a skill that requires practice. This post provides four quick tips to help you get started.
TIP 1: understand how trials are laid out
The typical article reporting trial results will include the following elements: Title, Abstract, Introduction, Methods, Results, and Discussion. However, it is not necessary to read all these elements or to read them in this order. The most important elements are the Title and Abstract (to decide whether or not to read the article) and the Methods and Results sections (to get a detailed understanding of what was done and the main findings). A video by Sketchy EBM explains this very nicely.
TIP 2: skim-read specific sections of the article to decide whether or not to read the article
You need to confirm that the trial matches your information needs and has been conducted robustly. The information you need for this should be in the Title and Abstract. Unfortunately, not all articles are well written, so you may also need to dip into parts of the Methods and Results sections. The things to look for are (1) what the trial was about (the patient group, intervention, and key outcome measures), (2) how well the trial was done (was there random and concealed allocation to groups, blinding of outcome assessors, adequate follow-up, and intention-to-treat analysis?), and (3) clear reporting of results (the difference between the groups). PEDro can help you to quickly identify trials that are more likely to be well-conducted and clearly report results. This is because the trials indexed in PEDro have been pre-appraised using the PEDro scale. Trials with higher PEDro scores are more likely to be valid and contain sufficient data to guide clinical practice than those with lower scores.
TIP 3: focus your attention on specific aspects while you read the Methods and Results sections in detail
If having skim-read the article you decide to read the paper more carefully, look for more detailed information about what the trial was about, how well the trial was done, and reporting of results than you gathered while skim reading. For the subjects, you need to understand the key characteristics required for inclusion in the trial and where they were recruited from – were the patients in the trial broadly similar to the patients you are treating? For the intervention, look for detailed descriptors of who, what, where, when and how much. Pay particular attention to the contrast (or distinctions) in treatment provided to the different groups in the trial. Trials are set up to test the effects of these contrasts in treatment – for example, in a trial comparing exercise and education to exercise only, the contrast is education and the trial is testing the effects of education (not the exercise provided to both groups). In addition to determining how subjects were allocated to groups, blinding of key people involved in the trial, completeness of follow-up and how the analyses were performed, also confirm that the trial was registered (eg, using ISRCTN Registry) and that the number of subjects recruited was sufficient to provide a robust answer for the question. For the reporting of results, it is common for multiple outcome measures to be used and for the trialists to identify which they consider to be the most important (or primary) outcome. Look for the difference between the groups at the end of the treatment and follow-up phases for the primary outcome measure, any harmful effects and the costs of providing the intervention. The estimate of the between-group difference and its precision (ie, 95% confidence interval) are more informative than statistical significance (p values). If the primary outcome is a continuous variable (eg, pain, walking speed) look for the mean between-group difference and its confidence interval. If the primary outcome is a dichotomous variable (eg, return to work) look for the relative risk reduction or odds ratio and its 95% confidence interval.
TIP 4: seek out more information
If you decide that the intervention has large enough effects to make it worthwhile to apply in clinical practice you will probably need to find out more about the intervention than was provided within the word restrictions of the article. Many trials now publish protocols and maintain web-sites that provide this information.
Your ability to read scientific articles reporting the results of randomised controlled trials will improve with practice. Make the commitment to read at least one article per month and share your reading with the global physiotherapy community in #MyPTArticleOfTheMonth.