As I said in the previous video, this series of tutorials will mainly focus on the literature search process required by a systematic review. However, to put these videos in perspective, it is necessary to talk about the typical process of a systematic review, and where searching fits in the big picture. If you decide to write a systematic review, and you have the necessary resources for such a project, such as people, money, time, library support, et cetera - the first thing you should do is to define the question you are trying to answer with your review. What is it you are trying to find out about? What are the key concepts in your question. As I mentioned in my other tutorials, we typically use the PICO model to help us define such a question. Recall that "P" represents patient, population, or problem. "I" represents intervention, prognostic factor, or exposure. "C" represents comparison of intervention. The comparison part is optional - your question may not always have a specific comparison. "O" represents outcome you would like to measure or achieve. For example, this question here "in patients with morbid obesity, is laparoscopic bypass more effective than open gastric bypass in helping patients lose weight, improve quality of life, and reduce complications"? Here the "P" is "patients with morbid obesity". "I" is "laparoscopic bypass" and "C" is "open gastric bypass", and "O" is "lose weight, improve quality of life, and reduce complications". Now that you've PICO'd your question, you also need to consider what type of question it is. This is a table showing different types of questions, and the most useful study designs for each type. This will help you determine what study designs will best answer your question. The example above is a treatment question, and in most cases, a randomized control trial is best for answering such a question. So if this is your review topic, you should pay special attention to randomized controlled trial studies when you search the literature. Your medical librarians or other information professionals trained in performing systematic reviews can assist you with defining your question with PICO, and determining question types and the types of studies that best answer your question. They can also provide assistance in searching for existing systematic reviews to see if the question has been already addressed in the recent past, to make sure that yours is not a duplicate effort. Now that you have a clearly defined question, and you know what type of studies you're looking for, you need to assemble a team to work on the review. Systematic reviews do not need approval from the IRB, and they typically do not require expensive equipment, but they are usually quite time consuming to do. So they're most commonly conducted by a team of reviewers, rather than by one individual. As a matter of fact, some review standards actually require that you have at least two members on your team to avoid any bias brought in by a one-person team. When assembling your team, you should make sure that team members have the appropriate expertise--such as knowledge of the clinical content, systematic review methods, searching, quantitative methods, et cetera. The next step is to create a review protocol. This is arguably the most important step in the review process. The protocol should dictate how the review will be conducted. It should include a conceptual discussion of the review question itself, a description of the search process, criteria for the inclusion and exclusion of studies, process of assessing the quality of studies, data extraction procedure, data synthesis procedure, record keeping practice, and a project timetable. Your protocol should follow the Cochrane, IOM, CRD, PRISMA, or a similar template. The next step is to conduct a thorough search, which is what the following videos in this series of tutorials will focus on. In general, you should consider searching databases in disciplines relevant to your topic, for both published and gray literature. You should also consider hand searching major relevant journals, reviewing the reference lists of relevant articles, and contacting the experts in the field to verify your search. If your protocol requires specific types of study designs you should consider using validated study design filters in your search. During the search process, using a citation management program, such as EndNote, will be a great help. Once you have the search results, you and your team members will then select appropriate studies using the inclusion and exclusion criteria defined in your review protocol. You typically do a first pass by reading the titles and abstracts. The primary goal here is to check if each of the results actually matches your PICO model - is it about the same type of patients? Is it about treating the same disease with the same intervention? Does it measure the same outcome? It is often required that this is done by at least two reviewers to avoid any errors and biases. At this point you should also develop your strategy to resolve disagreements among yourselves, and keep a log of all the excluded studies, and the reasons for their exclusion. After the first pass, you will need to obtain and read the full text of the included studies, and further appraise the quality of those studies using the criteria to find in your protocol. As with the first pass, it is strongly recommended that each study be assessed by at least two reviewers - you can use a simple check list or a quality scale system to keep track of your assessment. To avoid bias, you should also consider blinding the author, institution, and publisher information of the studies to the reviewers. After the second pass, you end up with a group of qualified studies for your review. The next step is to extract the data from these studies, so that you can do a comparison. Before extracting data at full scale, you should design a data extraction form, and pilot it with some data. Like everything else, the data should be extracted by at least two reviewers and they should be blinded from the author, institution, and publisher information of the studies. Your systematic review provides a summary of the data from the results of the included studies. Tabulating the results from the studies is a common way to compare them. If the results of the studies are similar, you should consider a statistical method called a "meta-analysis" to combine the results and calculate an overall summary estimate. Results of a meta- analysis are traditionally displayed in a figure known as a "forest plot". The next step is to interpret the results. Do the results have limitations due to unavoidable publication bias or related biases? What is the strength of evidence presented in these results? How applicable are these results? What are the economic implications? What are implications for future research on this topic? Finally, it's time to write up the review, which should include all sections as detailed by the systematic review standard of your choice. In this entire process, the search is an important link, and is the basis of all the following steps, and directly impacts the validity of the review's conclusion. Because of that, it is highly recommended that you work with a medical librarian to perform this search. As a matter of fact, one of the standards, the IOM standard, requires that you work with a librarian or other information specialist trained in performing systematic reviews to plan your search strategy and conduct searches. That's it for today, from the next video on, we'll go through the entire search process step by step. Thanks for viewing, and I'll see you next time.