In this section, I'll be going over some of the specific items that are included in a journal, the nuts and bolts of a journal article. What do you think is the most important part of the paper? What is the really the essential part that drives the success of a paper that it's going to be cited? Is it that wonderful graph? Is it all the tables with those fabulous results? No, it's not. It's the title, the abstract, and the key words. The reason that they are the most important part of the paper is because they're the lynchpin for editorial reviews, for search results. When you submit your paper, the editor is going to read the abstract and see if that looks like a paper that would be of interests for their journal. Electronic searching is based on the words that are in the title and the abstract, and what they're looking for is words that are in common with other articles on that subject and terms that are in the title or specified as key words carry even more weight in that matching. It's essential that the title and the abstract contain all the important terms that a potential reader may use in searching for the relevant literature. Another thing about the abstract is it may be the only freely accessible portion of your paper, and I'll talk about that more in a minute. As I mentioned before, you really need to get a very good draft of your paper before you write the abstract, and every time there's a revision of the manuscript, you need to go back and review the abstract and possibly revise it and certainly make sure that the numbers in the abstract match the numbers in the paper. There are CONSORT guidelines for preparing titles and abstracts. For titles they want them to be informative to include the target population, what interventions were used, also on the title page you will need to provide information on funding and other support for the trial and trial registration. The abstract, according to CONSORT, should be a structured abstract. On the right-hand side of this page, I've given you an example of a structured abstract where the paragraphs are titled to fit into those four sections of a paper. Basically the first two are the background and introduction, and then we have the next three paragraphs that are all the methods of the paper followed by the results and then the discussion. You have to make sure that you get the number enrolled in there that you have the number analyzed and that you have an estimate of the size of the treatment effect with confidence intervals or some estimate of precision. Abstracts are limited, you have 2-400 words, you should use short sentences, avoid acronyms because then you have to spell them out, and it should be a standalone document because often, this is the only part of your paper that's going to get read and even people that are going to use your data to include in their paper as a reference or in a systematic review maybe just pulling the numbers from your abstract. So it is very important that the abstract accurately reflects the content of the paper, and there shouldn't be anything in the abstract that is not in the paper. For little bit of a break, I thought I'd look at some of my own titles and see how well I had done with the CONSORT diagrams. I see in the clinical trial of low-dose theophylline and montelukast in patients with poorly controlled asthma, well, we manage to say it was a clinical trial, we did not say it was randomized, and in fact we actually don't have the control group which was fluticasone, so I wouldn't give myself an A for that, may be a C. Title 2 is from manuscript reporting the results of a trial of treatments for uveitic macular edema and it really has all of the information you might need about understanding that trial although many journals would not allow such a long title. The final one randomized comparison of strategies for reducing treatment in mild persistent asthma is actually my favorite of the group, but it is lacking the actual interventions and it would've been nice if we had included maybe a subtitle with the interventions. CONSORT has several guidelines for the methods section of the paper. Again, it's pretty straightforward here. There are critical items that you must include and you can use the consort checklist to ensure that you have included them. I think they are, for the most part, self-evident that you would have to talk about the participants, a description of the intervention including whether it was masked or blinded outcomes and measurements, details of randomization, and statistical methods. What may not be as clear is that if there's been any change in your protocol or how the study was done, you need to incorporate that in the methods. If the eligibility criteria was changed during the trial to expanded that help with recruitment, you need to state in the methods that an eligibility criterion had been changed and you may in the results want to talk about how that change affected your population. It is very important that all significant changes that occurred in the design and conduct of the study be reported in the methods, and again utilize the online supplements for details of how things were measured or particular laboratory procedures. CONSORT has quite a few guidelines for the results section as well. One of the important things is to set your study in a time-frame. You need to state the actual time period for recruitment and follow-up. As I mentioned, CONSORT diagrams are required. Most reports of clinical trials start with the table 1, which is the characteristics of the population so this helps the reader understand who these results are applicable to. Generally they shouldn't have P-values and I'll talk about that in a few minutes, reporting on how well the protocol was followed, and the adherence to the treatment. Did people stop taking treatments and why did they stop? Of course the treatment effects and with estimates of precision, if there were important subgroup analyses or ancillary analyses that were specified in the protocol, and sometimes during data analysis you may run across surprising finding that you didn't expect, maybe an unexpected subgroup effect that you did not anticipate. For example, you may have looked at the treatment effect by former versus current smokers and saw difference. That would be an exploratory analysis. You could mention the results of that analysis, but because it wasn't a part of the protocol, it cannot be considered as definitive test. Of course you need to specify the harms of the treatments, what were the adverse events that people experienced while they were on the assigned treatments. Here is the outline for the CONSORT flow diagram that you can find on their website and here's an example of one from one of the studies I was involved with. If you follow it, you can see that there are three treatment groups. Let's start from the top. We had 1,300 and some patients that were assessed for eligibility, that of those only 787 enrolled and 500 were randomly assigned. We have the reasons that people dropped off along the way. Then it shows how many people were assigned to each group, how many actually received their treatment assignment, who discontinued treatment or follow-up, and how many people were included in the analysis. You have a complete overview of the data that was analyzed for this trial, and this is an essential component of reporting your results. The discussion section can be the most challenging section to write. It usually starts with a short summary of what the key findings of your study are. You can't introduce any new results from your study at this time. The result that are included in the discussion have to have been presented in the results section. Then you can go on to compare the results of your study with prior studies and try to integrate the information from all of the studies conducted thus far and what new information was added by your study. Is it a different subgroup of participants that were included in your study? Did you use different doses of the drugs? Potentially you used a different outcome maybe prior studies had been based on surrogate outcomes and you actually used a clinical outcome. Those would be some of the reasons why your study is important. You should point out the strengths of your study, it was a randomized clinical trial, and the limitations. Because if you don't acknowledge the limitations, you can be sure someone will, whether it be a reviewer, an editor, or someone reading the study and writing to the editor. You need a concise bottom line of what was the impact and significance of the study, and nowadays you commonly need two other summary statements about your study; a precis and a summary for the public. The precis is about a three-sentence overview of why your study was important and what you found, and the summary for the public is again a short statement in non-technical terminology that people in the general public would be able to understand. Here's an example of a conclusion statement from a study I was involved with and we concluded that patients with asthma that is well controlled, so we said what the population was with the use of twice daily inhaled fluticasone can be switched to once-daily fluticasone plus salmeterol without increased rates of treatment failures. We have the control group and the experimental group. A switch to montelukast results in an increased rate of treatment failure and decreased asthma control. A switch to montelukast results in an increase rate of treatment failure and decreased asthma control. However, patients taking montelukast remained free of symptoms on 78 percent of the treatment days. Overall, I think that's a pretty good conclusion statement, but I think that we should have lost the qualification about montelukast. It makes it confusing to say that on the primary outcome montelukast was worse than the control, but then we say that people didn't have many symptoms. I think we probably should have skipped that qualification, although we thought it was an important qualification. Who gets to be an author? One person has probably spent a lot of time pulling the paper together, and when we see it published, it may have 20 authors on it. Well, there are criteria that journals have developed for including someone as an author that you have to be accountable for the accuracy and integrity of the work that you've actually reviewed and approved the final manuscript, and that you've participated and you have to be able to say that you were part of the concept and design of the study or you were involved in acquisition of data or analysis or interpretation of data and that you are involved in the drafting or critical review of the manuscripts and you contributed to the conduct of the trial. I don't really see how the third criterion is any different than the first criterion, but regardless you need to check off all three. Avoid gift authorships, that's how you get 20 people on a paper, people that really didn't contribute to the research but for some reason you feel it is necessary to include them because they're your boss or you hope to get favorable treatment in the future from this person or you think it will please them. Most of the people that you would give a gift authorship to don't need anymore authorships. There's also the phenomenon of ghost authorships, which means that there's really someone behind the scenes that has written the paper and other people are taking credit for it and you don't want to be on a paper where all of the authors are not named. Because with ghost authorships introduces the possibility of conflicts of interests that are undisclosed. There's lots of conventions about ordering authors who should be the first author since that's the most prestigious position, sometimes she can have two of them. The other prestigious position ironically is the last author because is considered the mentor or maybe the intellectual leader of the research. Everyone in between falls into the et al category because when the paper is referenced, they usually just have the first three authors and even the last author becomes an et al. I'd like to encourage you to think about corporate authorship for multi-center clinical trials that require involvement of several people and several institutions and it's a way that more people can take credit for the study, but to be consistent with the criteria on the left side of the slide you usually need to name a writing committee of the people who actually wrote the paper and will be held accountable for it. Regardless, recognize that authorship is a sensitive subject and it can be very important for people's careers and their professional growth. So you should approach it thoughtfully.
