Welcome back. I'm Lea Drye, and now we're going to talk about how to evaluate the written literature. In this section, we have a list of questions from Curt Meinert, who again was the founder for the Center for Clinical Trials at Johns Hopkins. And he talks about how to evaluate what is written in a results paper from a clinical trial, so that you can decide whether or not you think the results are reliable. So when we read the introduction of the paper, we need to decide whether or not we think there was a legitimate state of equipoise before the trial began. Was it a fair comparison, or did we already have evidence that one of the treatments was better than the other? We need to read through the conflicts of interest so we can decide if we think the investigators were trustworthy, and we need to read carefully in the methods to determine whether there were adequate protections against bias. Did they use randomization? If it was possible, did they use masking? And even if they couldn't mask patients or, clinic staff, did they mask outcome assessment? Did they have a standardized follow-up design and was it executed properly? And in the analysis, did they use an intention-to-treat analysis? And even if they said they used an intention-to-treat analysis, it's important to read through the results and see what they actually did and who they actually included. We want to look and make sure that all events after randomization have been counted. And if there are variations between denominators in different tables, were those variations described, and do they make sense, are they consistent with good practice? When the authors describe their subgroup analyses, how did they interpret the analysis? Did they put a lot of weight on post-hoc analyses that came up to be statistically significant? Occasionally, you'll see a post-hoc analysis that the authors really focus on to the point where they even include it in the title. And while post-hoc subgroup analyses might give you interesting ideas of things that should be looked at in the future, you need to know how to appropriately weight them. And have the authors done an adequate analysis that really supports the results that they're stating? And do the authors recognize and discuss any potential weaknesses of their design and execution? There are weaknesses. No trial is perfect, and it's important to review what the authors think the weaknesses are. Okay, so we've finished with the questions from Curt, and we're moving on to what we think are marks of a good trial. So we need to have a relevant question, randomization, an adequate sample size that can detect a clinically meaningful difference in the treatment groups. A meaningful outcome measure that's important to several different stakeholders: important to the patient, important to the clinicians. We need to have an adequate period of follow-up so that we can actually observe the outcome. We need to use analysis by original treatment assignment, and have adequate bias control procedures, such as masking, if possible, but standardized follow-up procedures. We need to have adequate performance at our clinics. If there's a lot of loss to follow-up, then it can be difficult to interpret the results. If there are a lot of protocol deviations, it can be difficult to interpret the results. So, we need to actually perform the trial well. We need to have comprehensive reporting following guidelines such as the consort guidelines so that we include all relevant information that the reader needs. And we need to have timely reporting. Results need to be reported in a reasonable amount of time after the trial is finished. So now we've come to the end of this lecture on writing and reading clinical trials literature. And hopefully, this will be useful to you as a writer of the results of clinical trials, and also as a potential reviewer for papers and even applications for funding for clinical trials.