Computer generated transcript

Warning!
The following transcript was generated automatically from the content and has not been checked or corrected manually.

OK, welcome everybody. Thanks for joining us for the third session in our S Mra teaching series. Last week, we talked about writing a protocol and selecting inclusion and exclusion criteria. This week, we're gonna chat about creating a search strategy for your S MRA delivering the session today is Dr Bay Sheriff. Uh He's an incoming academic foundation doctor and he has a BSC in biomedical science, an MSC in global health and conflict and an M BBS from Saint George's. He's interested in global public and women's health and has completed projects in Kenya, Morocco and Sierra Leone amongst other places. He's founder and chair of Student Med A London and the current chair of the Students for Global Health Board of Trustees. Just to say, and many of you will know already, please fill in the feedback form to get your certificate of attendance. I'll put it in the chat part way through the session for you. So thanks for joining us and I'll, I'll hand over to you, of course. Um Thank you for joining again everyone this week, I'll just start off by sharing my slides. Um And I hope everyone can see them brilliant just a heads up at two points during the presentation. I'm gonna ask you guys to get out your phone and go on mentee, but the codes will be available on your screen. I'll let you know when that's going to happen. Um So let's get started. So today we are talking about the specific elements of how to create a cert strategy when you're thinking about your wider systematic review. What is the cert strategy? I'm sure you guys at this point of your medical academic careers have written like literal reviews or essays. And I'm sure all of those involve going to Google trying to get information or look for a paper or writing something along the lines of 15 year olds p and hoping the appropriate article just pops up magically. Now when you're doing a systematic review, everything needs a little bit more structure and needs a little bit more prep planning as we mentioned in our session last week on how to write research protocol. The search strategy itself is this organized structure you're giving to the search, which means you're using specific key terms, you're using specific databases and you wanna make sure that your search will account for all possible papers or gray literature. Within this specific review that you're doing, you don't want to miss out huge chunks of like research or papers or trials because you didn't try a specific phrase or because you didn't use a specific database. So that's essentially what your search strategy is when you're writing a systematic review. Now, what are the benefits similarly to last week? And the research protocol, it makes sure that every few thing you do is more systematic rather than, than you just doing whatever and hoping you get the appropriate papers. It also makes sure that your team works well together in general. If you're doing a large systematic review, you're probably going to have more than one reviewer. And to make sure that there's consistency between the team, the strategy will help direct you guys in terms of what you're searching for, where you're searching for it and how you're searching for it. And overall, at the end of the day, that means it prevents additional work for any specific individuals within your team. And it also reduces the risk of useless abstract inclusion in your overall systematic review when you guys are doing your screenings. Now you guys, this is the third week. Um So far in your first week, you did the most important part, which was learning how to actually formulate your research questions, how to make sure it's specific, how to make sure it's relevant and how to make sure it's something you're actually interested in. Um Then last week, we discussed how to write a research protocol specifically how to write your inclusion exclusion criteria, which become incredibly relevant when you write your search strategy and to do any of this at this point of the systematic review, you would have probably done a little literature review to just see if your idea is viable at all. So what are we doing today? We're looking at how to write a search strategy. We're looking at four main things. We're looking at some of the potential research databases out there and where you'd go to find information. Obviously, Google is amazing. But for systematic reviews, we're looking for maybe slightly more specific things, we uh will be looking at how to define your search terms and something called concept mapping, which I'll talk about a little bit later and then we'll talk about all the little tips and tricks you can use uh to essentially help with your search on the different platforms and then overall how to refine your search. Um the feedback to the last patient stated that some of you wanted me to give a bit of a break after each slide to give you an opportunity to ask questions. So that's just a reminder if you do have a question about anything I've said so far, just write it in the comments section and the moderator will interrupt me and let me know. So they're no stupid question. Feel free to ask any time. So yeah, the first element research databases, when we're thinking about research databases, when it comes to systematic reviews, there are generally two types. There are some really large broad research databases that you'd use for almost any medical or health care, systematic reviews. And you also have a few specific and niche ones. And essentially I want us to try a little exercise. This is a systematic review. I did a few years ago with my wife, um slides Kudo to her, she, she, she makes them essentially, I don't know how to make things look presentable or pretty. Um But yeah, the title for that systematic review was essentially eb positivity and program definitely. And one expression in diffused large B Cell improvement. So my question to you guys is if you are trying to do a systematic review on this topic, do you have any idea of what kind of research database you might use to find relevant literature or relevant papers that might be included on this topic? This is one of the points where I'll ask you guys to take out your phone. So if you could go to mental.com on your phone, on your laptop or tablet or whatever device you're using that you're hopefully not using to watch this and type in this code. And I will have a quick look to see what you guys have come up with. Great. Um I hope you guys can still leave my screen. So yeah, if you go to mental.com and use the code 3594 5921, great. I can see some common ones. I hope everyone is aware of PUBMED is my personal good to uh I feel like it's everyone's MEDLINE really good Google scholar, which I feel like people don't use. I think it's a great starting point. Places like scopers, we signs brilliant. So these are a range of databases which are generally specific to healthcare, medicine. And the science is when it comes to literal review. And these databases just essentially index publications from a range of different journals. Now back to this light. Now for this paper, some of you mentioned it, we used mbas, we used Web of science, we used Ovid, but we also used specific databases related to on. So it was things along the lines of the national cancer institutes, things along the lines of eu clinical trials register and also clinical drive.gov. So when you think about databases, obviously you're going to use uh the big ones, your P meds, your mbas, your O VS. But depending on what the topic of your systematic review is, there will be slightly more specific databases for that topic and also something you don't wanna exclude when it comes to research databases. It's considering the use of gray literature in your systematic review. Essentially when I'm talking about gray literature is essentially everything that isn't a peer reviewed publication in a like respected journal. So it could be something like conference papers, pre pens people, thesis, um official publication, policy papers, um government documentations and those should be included in your systematic review. If during your protocol, you realized there was information there that was useful when it comes to research databases. I would not discount your universities and essentially your local librarians, their wealth of knowledge on this tends to be incredible. You al also tend to have access to quite a large range of databases with your university logins and Athena. So always consider talking to your librarian if you're not sure where to start or if you've used your p and you don't know where to go next. Great. If anyone has any questions on research databases, please put them in the comments and we'll get an answer. So as part of your search strategy, you know your topic, you've chosen your questions and you have written out a list of potential databases, you're gonna search to hopefully find relevant literature and you've also looked at some gray literature which you also feel that it's liberate. Next thing you're going to do is you need to find out what search terms you need to use to find an appropriate literature and do something called concept mapping. So what that involves is writing out your research question. I have an example of one here which says, assess the impact of social media exposure on young people's overall mood and mental health. If you guys could go to mentee again, it's the same code. I will move on to the next question soon. And essentially what the way this works is you write your research question, you highlight either specific phrases, specific keywords or specific concepts in your question. And what you're trying to do is you're gonna try to find synonyms, alternate terminology, acronyms that can be used to replace the word. You just used an example for that would be if you use the word social media exposure as one of your concepts, you'd want publications that mention social media exposure, but also anything that's similar or relevant to that. So in this case, that could be just using social media platforms where like Instagram, Facebook, tiktok, you called merge social media exposure to be quite similar to exposure to phone use, same thing with young people. There are several different ways you can use say the word young people, adolescent, et cetera. So if you guys could go to men.com and essentially type and alternate words, you guys can think of that match the like essential concept in this research question, which would be social media exposure, young people, mood and mental health and I'll take us back to mental. Great. I think the next question should be available to you guys now. Great. So as you can see, we already have quite a few words that you could use to, for example, replace mood and mental health words being wellbeing, depression, anxiety, uh you have words replacing young people, which could be Children, young adults and you have words replacing social media such as platform internet in general, social networks. And it's really important in your third strategy that you do this concept, my pain because it allows you to make sure you're finding appropriate literature. Like if your research question is asking about young people with the mental health and social media, you don't wanna miss out on a paper because it's talking about um the relationship between depression and young people using social media because you didn't word depression or you didn't use the word anxiety or you didn't use the word social networks instead of social media or Instagram or Facebook. What this is just to give you guys an idea of how to do a simple concept map and how you can find either acronyms or alternate words for the word that you're using. And I'm taking us back to the slides. Now when you finish one, I'm using my review from 2020 again to give an example of what this would look like in the paper that was looking at the association between EBV program, Deft one and a large B cell lymphoma. The important concepts are obviously going to be EBV program deft one and the large B cell lymphoma. And there's several ways you can write these Epstein Barr virus can be written as EB VHE VE BR and MP. And they're like a range of different acronyms or ways it can be written and you wanna make sure that you have considered and covered all of those concepts when you're doing your search strategy. Now your next question is why it is relevant. Are you individually just gonna replace E BB uh defuse large B cell lymphoma or program def ligand for each of those words and search over and over again. No, you don't because once you've done this concept mapping and you've gotten all your alternate words and your acronyms and your terminology, what you do then is combine them because there's a range of like specialized search tools you can use on those databases in the platform. Um A really useful one are brilliant operators. So essentially in most cases, they and or are not, and they essentially allow you to use these additional terms you have found in your search structure when you're asking for questions to make it a piece of text. For example, in the topic we just covered where we had several versions of EBV. When you're searching for papers, you will then include E BV or Epstein Barr virus or HV. If that kind of makes sense and would be used if you want different articles that are related in the same topics. For example, in the Epsom V review, we're talking about, you're not just looking for ABV, you're looking for its relationship to and, and D legs. And then the last billion operator you could use is not, which essentially allows you to exclude a search general concept. If for example, you're looking at something very specific, but there's something related to it which you're not interested or you don't want to include it in your papers, you could then add the term, not in essence what this ends up looking like is this, you essentially combine your terms. So at this point of your third strategy, you've done a concept map, you found all these alternate words um for the question you've written and you've combined them. If you look at the top one in brackets, we have all the different ways we can see EBV, each of them alternated by all, which means when the database is searching, it's searching for EBV or Epstein Barr virus or HB four, any of those will be accepted, but it's doing it in combination with either DL BCL or lymphoma or B cell lymphoma or the large B cell lymphoma, which means all the results you should be getting are gonna contain an element of EBV, an element of the Frisbee large lymphoma and an element of program defla one. And you're essentially this way catching all of the relevant papers in your case. Another example of how this can be done, which is another review I was a part of as you can use the brilliant operators, not just between concept and keywords, you can use them within concept and keywords. So in this case, by using global investments or and global financing, it means we're happy to accept papers which are using the word global financing rather than the term global investment. But we're equally happy to accept both terms as part of the title. And again, that can be within specific key terms and concepts, but also separate from specific key terms of concepts. This is the part that can get a little confusing. Um But again, if anyone has any questions, please put them in the comments. So essentially, at this point, you have the bread and butter of your search strategy, you have taken your questions, you have highlighted all the important key concepts or key terms. You have broken them down and found alternate ways to write them, including acronyms. You have looked at all of the databases available. You obviously know all of the big ones like PUBMED MEDLINE Ovid. But you've also looked for specific databases around the theme of your systematic review and you've made sure you know them, you have now compared all of these terms that you're using using brilliant operators and you should be fine. Now, you know what you're looking for, you know, the databases you're looking at, what else can you do as part of research strategy? Now, the additional search operators you can use, you might have noticed in your previous light. I had little oh I forgot what is it called? Uh but around the words and essentially that is an operator on most databases, that means you're looking for that specific phrase or that specific word, which means it will only show you papers and articles that have that specific phrase. In this case, it would be apple. So if the journal doesn't have the word apple, it wouldn't be included. Um Another operator you can use is for truncated words. For example, instead of an insert strategy including the words vaccine, vaccination, vaccinated. Since the stem of the word is vax, you can add a truncation and in some papers that's essentially adding a little star after the word and it will automatically the search will then include the prolongation of those words. So if you use vac it would you include vaccine and vaccination and vaccinated. Another search operator you can use is for wild card searches. This is usually useful when you're not maybe exactly sure how to spell a word or if there's a difference between how a word is spelled in the UK or the US. Um because it's spoke English. And depending on if the researchers were American or British, they might have written it slightly differently. And again, in some search bases. Um in that case, you just use a question mark where you think there might be a difference within the word. And then it would make sure to search for both version of that word. Another search operator, you can use it for proximity searching. Let's say you're looking for papers that have the word physicians and relationship, but you're happy to have them in any combination of those words, but you essentially want them a certain amount of words together, you want them to have a certain proximity. In this case, the operator you can use in this case is a DJ three adjacent three, which means they need to be three keywords adjacent of each other. So it's not gonna count words like there or of or in so in this case, if you use physicians adjacent relationship, you will get words like physician, patient relationship, patient, physician, relationship or relationship of the physicians to the patients. Because again, it's not counting the words of the or two. And these are just like additional search operators you can use in your search strategy to make sure you're getting all the relevant like journals and papers that you're looking for. But just to be clear, the operators I used here are just examples of how they're used on some of the databases and their differences on, for example, how they might be used on MEDLINE compared on how they'd be used on word of sciences. Um This is like a really nice guide like produced by the Queen University Belfast that gives like examples of essentially which search operators are used on different platforms. As you can see in terms of truncation, the star is generally used everywhere. But when you're doing for specific free research showing um it's different based on what platform you're using and some platform might not have them at all. Fy Yes. Um Are you happy to take questions now or would you like to leave them to the end. I'm happy to take questions now. I feel like we should take them while we're on the line. Yeah. So we, we have a couple actually. So the first is how can you ensure that you haven't missed out any key words when doing searches? So to make sure you haven't missed out any keywords? Hm. This is the step that becomes important like this isn't something you produce in a minute or 15 seconds is you make sure to cause at this point you've done a basic literature review, you have decided your research question, you've written a basic research protocol, you're gonna write out your full research question and you're going to highlight all the relevant concept and key words, you then actually going to take your time for each of those concepts and key words to write every iterations variation acronym of that word that you can think of. And if you're working as part of your team, everyone in your team will be involved in that process. Obviously, that doesn't mean that there isn't any possibility that you might have forgotten something or there's something that you didn't include. But as part of this, once you've essentially created your third strategy, you'd also do a like basic trial run. You'd go on something like Google scholar or just PM try your search strategy and see what's in the abstract of the papers that coming up. For example, if your systematic review is on social media exposure of young people in their mood and mental health. And you have like the first draft of what your search terms are and you pop that into Google Scholar. And then on one of the abstracts you see in terms of mood and mental health, it's talking about, I don't know, personality disorders, mood disorders and you didn't include that. You'd be like, oh, ok. I need to update my search terms and keywords based on what I've seen is out there. If that kind of makes sense. Yeah, that makes sense. Um, a related question and I think you've probably already answered it is, can you tell us a bit more about this template that you've used in organizing the search terms and concept mapping? And can they make a personalized one of it? Yes. So this is just a template I think from potentially leads. You don't need to write it out this way. I have this template level available in a little folder essentially at the end of the session, I have an additional feedback session that I'd appreciate. If you guys filled and at the end of it, you'll have this week's slides last week's slides, but also essentially this basic template. You're welcome to make your own. It's just like a reminder to make sure you actually break down your question is a key concept and themes and make sure that you didn't break those down. Hm. And then the last is does the search tool look for search terms in the title of the paper or in the text of the paper as well? So you can play around with that in most the most advanced settings. But usually it will look for the title and the abstract, but it's also depending on what database you're using. But on one of the latest slides, I will also explain the way you can play around of where it looks for the relevant information. Ok, brilliant, thanks. Great. Where were we? Yes, this is where we were. But yeah. Um so yeah, different platforms essentially use slightly different operators just to be aware of that. Next, what else can we do? So at this point, you have your questions, you have split down the concept map, you have broke down the words and you know all of the ne tricks to make sure you're searching for all the relevant words and phrases and things you're looking for and additional things you can do is search for subject headings. So there are specific terms that are used in different databases that search for specific things. An example is if you're using MEDLINE, if you used mesh, we'll be looking for specifically medical subject headlines, an example of what that might look like. So this is you've gone on PUBMED and you have written your search strategy and you're looking for specific things. So there's several ways you can do this, you can apologies, you can search for your specific words and start adding them with an and with an or with a not. So it's already part of the permanent search system. But you can now also start searching for specific things. Are you looking for editors? Are you looking for specific mesh terms? Are you looking for subheadings? Which means now, not only is it looking at the papers abstract entitle, it will also look for additional things within the paper. So essentially when your full search structure is written, you will be including mesh terms on top of the key terms. For example, let's say our paper is relevant and we are looking for a subheading of um mental health. This would now include the specific term depression, an inclusion with mental health, but as a subheading if that kind of makes sense, but there is a risk, the more like specific like mesh subheadings you have the more like specific terms you're looking for in the paper overall, the more specific it becomes and then instead of potentially getting like a decent number of papers, you might get this. Yeah and back to the slides. So yeah, but again, these are again specific to the different databases, but you can specifically look for subject heading in the text and look for words in the text. It's not available on all of them, but you can do this in addition to the terms you've already used. And in addition to the tips and tricks you've learned here, things like uh wide search, proximity searching and searching for like exact and specific phrases and words. Um And the last tool which I think is the most useful, especially if you're doing a systematic review at your stages is citation searching. Um Essentially, it's amazing you can find a lot of really good papers. This way, essentially what that involves is if you have found what you think is a relevant paper to your topic, it's looking at the citation and references of the paper to find other articles that are also relevant. It will often help you find like more recent papers on the similar topic. Uh It will let you find maybe if you found a paper about a like meta analysis of like a range of like RCT S, you can then maybe find individual R CT S involved in the paper and what they have to say about it and places they reference and overall it will help you make your literature review more comprehensive. And you can do this either by just like manually looking through the references and citations that include in your paper or some databases have specific functions that allow you to do citation searching or reference searching as part of your whole search analogy. Now, we've gotten to the point we have our questions, we've done some concept mapping. We know exactly the additional words that we're searching for. We know which ones need to be truncated. We know which ones need to be specific and we've done a little trial on, on, let's say Google Scholar and we've either just gotten loads and loads of papers or only four papers have shown up and we're not sure how to move forward with our system much. You, if again, depending on how big your team is. If in your systematic reviews, you've put in your search terms and you have gotten 400,000 papers, that's obviously more that you can realistically look through. You need to narrow it down at all. So at this stage, what you do is you could add search limits and search filters, search limits will be along the lines of all the papers published between a certain time point and another or all the papers published before 2000. For example, if you wanted to look at the impact of pandemics on the global economy, but you didn't want to include the impact that COVID had had. Your search limit would include something along the lines of before the year 2019. You might also want to limit to specific languages. So this might be English or a specific publication type. So that might be RCT. But a lot of this would have been decided in the inclusion exclusion criteria. You would have written as part of your research protocol. Apart from setting specific search limits, you can also set search filters and you can essentially do that in two ways, you can, some databases will have a specific function that will allow you to set a filter to all the results that you're getting some of them don't. But you can essentially achieve the same function by adding to your key terms and whatever that filter is that you wanna add, if you have the opposite problem, for example, where you've typed on your question and it's telling you there are no papers available or there are only three available. First thing I will do is check your spelling, especially if you're looking for specific phrases because if you misspell a word, it just won't come up. Even if there are loads and loads of papers and articles that fit your criteria. If you spelled something wrong, it's not gonna come up. So make sure you actually check the spelling and you've used world searches if you haven't done that, if again, you are doing a systematic review. And so far you've only used MEDLINE, but it's specifically to and you haven't gotten that many results. Think about adding additional databases, preferably that specific to the topic that you're talking about. And if all of that doesn't work and there's still only three or four papers for your systematic review at this point, you need to revisit your research question and your concept map and really think about did I include every iterations or every versions of these key concept or key phrases that I'm using in my research question. Great. And lastly, so this isn't specific to how to write your um search. But it is relevant to think about when you are writing a search strategy. I think the session next week is about screening and it will probably cover it there. And this is essentially just a prism and it's essentially important as part of your strate strategy that you know, a who's going to be reviewing this. How many people are reviewing it? What criteria are they using mostly your inclusion and exclusion criteria? And how do they resolve this agreement? If reviewer A thinks, oh I think this is relevant and review B thinks, oh, that's not relevant. And you will usually include all of that in your prisma diagram. But that is more specific to actually have screen papers. But I think it's important for when you're writing or thinking about your search strategy to think. Oh, this is what I'm gonna think about including it in. So I wanna know the number of databases I'm looking at how many papers are you coming up for each? Um How many of them are repeating? So to just have that in your head essentially as you're doing your presentation, uh sorry as you're doing your search. Um And lastly, I will send a link of this in the questions, but you can take a picture of the QR code just to make clear this isn't part of the N MRA feedback that you guys have to fill out to get your certificate at the end. This is just for my own personal and professional development. I like to see um where I can improve in my teaching and how I deliver sessions like this. So I'd really appreciate it if you guys took the time to fill it out. Also, once you filled it out at the end, you guys will get a link to essentially this slide and the slide from last week and essentially the a little word document on how uh to essentially do the concept mapping. But obviously you can do that in whichever way you prefer to do. And lastly, the session you have next week is going to be on screenings that isn't going to be given by me, but I'm gonna guess it's gonna be a really great session and hope you guys are there as well. Um I'm gonna stop sharing my screen in a minute and essentially send a link to the feedback form in the group as well. And if you guys have any additional question, please feel free to put them in the comment section. Thank you. Thanks baffy. Um Just two questions. Do you have the link to your feedback form that we could actually put in the chat? Yes, I am doing that right now. Um And then some other questions just when you're ready. Um How many papers should they have minimum? I think it depends. Sorry. Wait, this is sorry. I was just filling out the Feever form. So it depends on your study. If you are doing a systematic review on a really small population on a really small condition that's on a neglected tropical disease. And you only end up with, I don't know, 45 searches, 45 results generally and your screening cuts them down. So four, that makes sense. That's acceptable. That's because of the topic you have chosen. But for example, right now, if you're doing a systematic review on COVID and its impact on health and you have less than 50 or 100 results come up, I will assume that maybe something didn't go quite right or maybe there's something wrong with the search term. There isn't like a hard or fast rules of, you need a minimum of like 50 papers or 100 papers or 1000 papers. It's gonna be very dependent on the topic that you're doing it on. I hope that makes sense. I'm sorry, I don't have a specific answer for that. Um Another question is how do I choose to go either specific or general when I search for studies? So this person is giving an example of looking at the effect of medication on obese people, how specific or general should they go? So I think this now goes back to the first session you guys had, which is on how to write your research question. So first of all is how important is this piece of research looking at the effect of medication of these people, then you do a quick literature review and see what's already out there because there's no point you doing a systematic review that already exists. And if you realize, oh, someone has already looked at the effect of specifically diabetic medication of these people, there's no point you repeating that which means it will make the term a medication more specific. So you will then think instead of just using medication, am I using, looking at a specific medication and doing a systematic review on that? And even in term obese people do, I mean obese people above a certain BM I, am I looking for a certain weight category? Am I looking at a specific region? So because your research question itself should hopefully be slightly more specific than this. It makes it easier when you come to writing a search strategy. Let's say you've done all of those. And your research question is literally the effect of medication on obese people. When you do your concept mapping, one of the important things would be medication and then you would have 500 different versions of medications. You'd have uh medication pills, you'd have hypertensive medication, you'd have to have diabetic medications. You'd have each of those specific key terms because they'd be relevant to what's coming up. So when you're searching overall in your and this and this and this section, you would have your broth things, but you would also have all of your specific things. Yeah, that makes sense. Um And then I think there's some confusion around the citation searching. So how is citation searching different than the other search? And why is it more helpful? So there are other search options. You're essentially just looking at the whole database and hoping that when you get to the screening, the things that come up are useful and relevant, but there aren't really connected. Citation searching is a bit like convenience sampling. If you're doing like our population research, because you already have a paper that's relevant that fits your inclusion and exclusion criteria. It's very likely the, the papers that journal or piece that article is referencing it citing will also be relevant. So the difference between citation searching is instead of just generally everything that fits those terms. I'm now in addition, looking at the paper that I know fits and see what it cited. I'm not sure if that makes sense. No, that makes sense. OK. OK. Um So, yeah, thank you very much for that session. That was very interesting. Um Thanks everybody for joining us if you wouldn't mind filling in the N Mra feedback form and also Fi's feedback form. That would be brilliant. Um And if there's no more questions and Fi's happy, um we can leave it there. Yeah. Yeah. Great, brilliant. Thank you. Thank you, everyone.