Speaker 1: You've developed a research question, conducted a preliminary search on your topic, identified some relevant search terms, and now you're ready to begin searching. Depending on your goals and your project, there are many different approaches that you can take to do just this. In this lecture, we'll look at a systematic approach that you can use in PubMed and other biomedical databases to find relevant articles on your topic. You'll learn how to use a combination of medical subject headings, MeSH, and keywords in your search, incorporate search techniques such as Boolean operators, truncation, and field tags, and run your search in PubMed one concept at a time. This approach is best suited for developing a comprehensive search such as the one you might use for a literature review. Let's get started. In a previous lecture, we conducted a preliminary search for the following research question. How are mobile apps currently being used to help individuals with high blood pressure eat a low-sodium diet? In this lecture, we're going to broaden the search to how are mobile apps currently being used to help individuals with high blood pressure eat a healthier diet? We can use this question to develop the type of search strategy that we would use if we were conducting a comprehensive literature search. From our question, mobile apps, diet, and blood pressure are our key concepts. To build the search, I'm going to work in a Word document, and I'll paste the final strategy into PubMed when it's complete. As informationists, we frequently use this technique because it allows us to think through all the terms and strategies that we want to use before we even get into a database. It also provides documentation of the search terms and strategy, which becomes vital if we want to rerun the search or let others know how we found our results. As a first step, we'll identify the relevant mesh for our search. As you might already know, mesh is a specialized subject headings or controlled vocabulary used in MEDLINE. All MEDLINE articles are indexed using this vocabulary. To better understand the value of mesh and why we include mesh in our searches, let's take a look at a common problem that occurs when searching. That is, that authors may refer to one concept using many different terms. For example, think about handwashing. If we want to find all of the articles available in PubMed on handwashing, just using the term could cause us to miss out on some key articles. This is because authors may refer to it as hand sanitation or surgical scrubbing. Sometimes they even write out handwashing using two words. If you just search for the term handwashing, your search won't retrieve citations where authors use these variations. To remedy this problem, databases such as MEDLINE, Embase, APA PsychInfo, and CINAHL use a controlled vocabulary. In MEDLINE, the controlled vocabulary is called MESH. Indexers who work for MEDLINE read each article and tag it with the appropriate MESH. They use the most specific headings that are available and apply somewhere between 5 to 15 MESH to each article. For articles that address handwashing, they are tagged with the MESH hand disinfection. What this means for you is when you search using the appropriate MESH, such as hand disinfection, you'll retrieve articles regardless of the terminology used by the authors. There are several ways to identify relevant MESH for your research question. First, if you found an on-topic article in PubMed and it's indexed in MEDLINE, such as this one, titled, Mobile Apps for the Dietary Approaches to Stop Hypertension, click on the citation's title. Next, scroll to the right-hand side menu bar and click on MESH. This will take you to a listing of the article's MESH. If you see a heading that looks appropriate, click on the term and then the Search in MESH link. You're now in the MESH database where you can look more closely at the heading. Show the term as a definition and the year it was introduced. Read through the definition and if everything looks good, click on the Add to Search Builder button on the right-hand side of the screen. Notice that PubMed puts the term in double quotes and adds the language bracket MESH bracket. This is the syntax that PubMed needs so that it knows to search the term as a MESH. Copy and paste exactly what you see in this box into your Word document. Alternatively, you can identify MESH by going directly to the MESH database from the PubMed homepage. Scroll down to the Explore heading and click on the MESH database link. While it may look like you're still in PubMed, you're actually in a different database. You can tell this because you'll see the word MESH next to the search box. Now let's look up the corresponding MESH for our concept Diet. I'll type Diet in the search box and click on the Search button. The results show several pages of MESH to choose from. It's worth scrolling through this list as you might find more appropriate MESH to use than the first few results. For example, if we were focusing on low-sodium diets, we could use Diet, Sodium Restricted. Instead, let's click on the heading Diet Therapy. This heading is defined as adjusting the quantity and quality of food intake to improve health status for an individual, which makes it a good choice for our topic. Again, I'll click on the Add to Search Builder. If you want to see the articles tagged with this MESH, click on the Search PubMed button and go into PubMed to see your results. From here, you can also copy the heading with the added syntax and paste it into your Word document. Let's move on to our last concept, Blood Pressure. There is the broad heading Blood Pressure that we could use, but I don't want our search to veer too far off the original question. So instead, I'm going to select Hypertension. Let's click on this heading and scroll down to the hierarchy. Here you'll see all the narrower MESH that fall below Hypertension. These headings are automatically included when you search the broader term Hypertension. The technical term for this is Exploding. But if you want to exclude these narrower terms, you can click on the button next to Do Not Include MESH Terms Found Below This Term in the MESH Hierarchy. Now when I add the term to the Search Builder, in brackets, you'll see MESH colon No EXP. Essentially, we're telling PubMed not to explode this heading. But in our case, I'm going to uncheck this box and just search with the MESH Hypertension. I'll copy and paste this into my Word document. Now we have three MESH for our research question. It's not uncommon to identify and use several relevant MESH for each concept, so we could keep looking through the MESH database. But for the sake of time, let's move on to the next step. In an ideal world, searching by MESH should be good enough to retrieve relevant citations on a topic. Unfortunately, this is not always the case. Sometimes there isn't a MESH for the concept you're searching. This is particularly true for terms related to technology and rare diseases. For example, there is no MESH for Sporadic Parencephaly. Also, new MESH gets added every year. If you're using one of these newer headings, like Augmented Reality, which was added to the MESH database in 2019, you'll miss older articles that were indexed under another heading. Furthermore, it can take several months for MEDLINE to index new articles. If you only search with MESH, you might miss the most recent literature on your topic. For these three reasons, if your goal is to be thorough, you'll want to search using a combination of MESH and keywords. Once you've identified keywords that you'll use in your search, you can begin applying search techniques such as Boolean operators, truncation, nesting, quotes, and field tags to these terms. For our example, in my Word document, I've already identified two to five keywords for each term. Depending on the topic and the scope of your project, you may keep going until you've identified every possible keyword for your topic. But what we have here is enough to get started. Looking at our first concept, Mobile Applications, I'll use an OR in between the synonyms. Remember, we want articles that address any of these terms, so OR is the correct Boolean operator to use. I'll type Mobile Applications, or Mobile Technologies, or Mobile Healthcare, or mHealth. Then I'll go through this list and look for terms that I can truncate. Mobile Applications might seem like an obvious phrase to truncate in order to find the terms Mobile Application, Mobile Applications, Mobile App, and Mobile Apps, but currently PubMed only allows truncation with stems that are four letters or longer. So while we can use an asterisk at the end of the N to get Application or Applications, we'll need to write out Mobile App or Mobile Apps. I'll also take the IES off of Technologies and use an asterisk to find Technology and Technologies. Next, I'll put double quotes around any multi-word phrases. So I'll put double quotes around Mobile Application, Mobile App, Mobile Apps, Mobile Technology, and Mobile Healthcare. I won't use double quotes around mHealth because it's a single word. Notice in PubMed I can now use truncation and double quotes at the same time. This is a recent change. Just make sure that the asterisk is inside the quotes. Last, I'll add field tags to my terms. However, for a comprehensive literature review, the TW or TIAB field tags are most frequently used. Test your search both ways to see which provide the best results. I'll start with the TW field tag for this search. So I'll put bracket TW bracket at the end of each term, making sure that I type it after the double quotes for my phrases. Before moving on to the next concept, double check to make sure the syntax is correct and that Boolean operators are in all caps. Even the smallest mistake, like missing an end quote around this phrase Mobile Technology, can throw off the results when you go into PubMed. I'll go through my other two concepts and apply the same treatment to them, adding ors in between the terms and then incorporating truncation, double quotes, and field tags. Finally, double checking for typos and errors. As a last step, I'll use an or and include my mesh into the searches. So the final search will look like this. Once you have the search the way you want it, you can go into PubMed and start to run your search. The method we're going to use is frequently referred to as the building block approach. That is, we'll build the search one block, or concept, at a time and then connect these pieces together at the end to see our final list of results. Start with your first concept. In our case, I'll copy the search for Mobile Applications and paste it into the PubMed search box. One of the benefits of searching one concept at a time is that you can catch any mistakes in these smaller blocks before you put everything together. You can see that I misspelled health in mhealth. PubMed notified me of this underneath the results. I'll go back to my Word document and fix this mistake. Then I'll run the search again. Don't be nervous about the large number of results. Remember, each time we AND in another concept, the results will get smaller. Now I'll paste the search strategy for diet. In addition to looking for spelling mistakes, it's also a good idea to look at the search details for each concept to make sure PubMed is searching exactly how you want it to. Click on the Advanced link and scroll down to the History and Search Details. Then click on the arrow next to Details. Everything looks good here, so I can go back to the PubMed search box and paste in my third concept, Hypertension. If you're on the Advanced Search page, you can also paste in your search strategy in the Query box and click on the Search button. Again, check the results for spelling and the search details to make sure each concept was searched correctly. Once you search your three concepts, it's time to combine them and see your final results. In the History and Search Details box, you'll see all of the searches that we've conducted so far. I'm going to delete this first search with our spelling mistake so it doesn't cause confusion. Click on the three dots under the Action heading and then click on the Delete link. Now I'm looking at the final searches for my three concepts. There are several ways to combine these searches. What I do most often is use the set numbers. You can see our first concept, Mobile Applications, is number 2. Our second concept, Diet, is number 3. And our third concept, Hypertension, is number 4. In the Query box, I can simply type number 2 and number 3 and number 4. Remember, I'm using an AND between the sets because I want articles that address all three concepts. Once the Query box is how you want it, click the Search button. As you look through the results, you should see your terms appear in the titles and abstracts. This is the first indication that your search is working correctly. Again, if you're in doubt, you can go back to the Search Details and take a look at how PubMed searched. If you don't like using the set numbers, you can click on the three dots next to each set and click on the Add Query link for your first concept. And for each set afterwards, you'll click on the Add with AND link. If you make any changes in PubMed, make sure to make all of these changes in your Word document. It's important to keep your Word document current so that you can report your search strategy and replicate it when needed. I often put my search strategy in a chart using the same number sets that I used in PubMed. This provides a clean and organized way to share exactly how the search was conducted. An important idea to remember when searching biomedical databases is that searching is an iterative process. Oftentimes you need to make major and minor adjustments along the way. For example, after you review your results, you may see more terms that you want to add or remove from your search. Or you may decide to change your field tags or use of truncation after you see what these are doing to the search. While this may feel like extra work, remember that each change you make is bringing your search one step closer to generating relevant results. In closing, with a little bit of legwork and patience, you will be well on your way to developing an effective search strategy following a systematic approach that uses a combination of mesh and keywords integrated with Boolean operators, truncation, quotes, and field tags is a key factor in building comprehensive searches that can be used for literature reviews and other research projects.