Computer generated transcript

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

OK. I think we'll go ahead and start. Um, the plan for the day is, uh, we'll start just by a little bit of an introduction about the research team and what kind of plans have you got? Um And then I'll introduce our speaker for today, Doctor Valerie Deep. Uh, he will be giving a presentation on systematic reviews. Um And then hopefully we can have a bit kind of a discussion. People if, if people have any questions, please feel free to, uh, post them in the chat. Uh, tell me if that's all right. Would you be able to keep an eye on the chart just to see if there are any questions for Valerie? Thank you. Um So my name is Noon. Uh I'm going to be, I'm the research lead for this year and I'm working with Temi. Uh, and we've got AAA lot in plan for what we want to do. Uh, in preparation for this application cycle. There are two things that we're focusing on. The first part is on research, which is part of what we're doing today. And the second part that we're trying to focus on is on quality improvements and audits, quality improvement projects and audits. Uh We're, we're, we're aiming to focus on that first because in the application and for, for portfolio discount for a lot of work so that we want to help um maximize uh or support people in maximizing their portfolio scores uh in for the upcoming application round. But we also hope that people will, what the knowledge that people will gain will be beyond just uh the portfolio scores uh in terms of the quality improvement projects and the audits, we have done one workshop uh in the past um in the past few weeks and that focused on um kind of what kind of the, we've talked a little bit about what they are and what um their definition, how to go about doing them. Um And we wanted to, we wanna follow up on that with workshops. Um And what we want to do today is again, talk about some re some research um methodology that people can um consider and we would want to support people if they're interested in starting a project or working on a systematic review, a scoping review, uh any uh research project with the intention of presenting it or publishing it. Um Today, like I said, would be the first session for the research uh club but Teme and I will be sending in the group in the next few days. Uh Some ideas about some research ideas that we're considering that people could um take on. And if also people have any ideas, please feel free to kind of share uh in the group. Also, if you already have some ideas, feel free to ask Valerie at any point about what you're thinking and to get some tips while we've got them. Um And we'll also be sending kind of further details about uh follow up workshops that we want to do. So, uh like I said, we're trying to help people um support people in maximizing their scores. Um And we will be sending further details about kind of support uh workshops that we're uh planning. Um So without kind of talking a bit more about this, I'll give this um the floor to our speaker today. Um We have here uh Doctor Valerie Deep. He is a clinician from Cameroon. He is also very passionate about research. Uh We were colleagues at some point and he's one of um the few people that I know who is extremely passionate about research very well um um kind of knowledgeable, extremely accomplished in terms of kind of publications. Also, he has in addition to his research, he also holds a number of kind of editorial roles. So not only has he had firsthand experience of research, he also has experience um kind of uh editing other people research advising. Uh in that sense, I think he would have um a wealth of experience that we can all learn from uh today. Um I'll hand over to you Valerie and thank you so much for um um accepting our invitation today. Thanks so much and thanks for inviting me and it's always a pleasure to talk about research and I will just pull up my slides for uh um just let me know if you can see. Yeah, which of my slides I can see your slides. The main slide. Yes. Ok. Ok. So thanks once again and today will be, you will be just having discussion, it will just be an introduction to systematic reviews. And so for you to appreciate why you need a systematic review. And if you, if you get to that stage where you need to do a systematic review for you to appreciate the, the different steps that are involved in, in conducting a systematic review, either for your own study or if you want to join a, a different research group so that you can have the basic ideas and the concepts. And I'm mostly hoping that you will have a discussion around it. So if there's anything that's unclear and you want me to explain it further, please stop me at any point. And then I'm going to explain further. And another thing I wanted to ask if people are familiar with um systematic review, if you have done it before or if you have seen it before. Yeah, I can't see the, the, the charts, but I'm just going to start with your presentation anyway. So the the main, so the main objective therefore will be for us to understand the need for a systematic review, what is meant to do and to learn how to formulate a question that can be answered using a systematic review. Um Then after we have decided on a question that we want to answer, we will need to know the different steps that are involved in conducting a systematic review and then understand how you you present the results or the methods of your systematic reviews. So just to to set it off, there are different reviews, there are different types of reviews. And so a review is just collecting a bunch of different studies that have been conducted on a specific area of research. And so there are, there are bunch of related studies that you are trying to combine all of the information into one to constitute the reviews and broadly, you can classify reviews into three. They are what they call narrative reviews that we are going to talk about. There are scoping reviews and then the there will be, there are systematic reviews w which, which is going to be the main focus for this presentation. So in native reviews, usually there is no documented methods in which you go about identifying or selecting the individual studies that we we have talked about. And this type of review is usually done by an expert on auto body in an area of research and it it's, it's usually quick to perform. Usually you have a essential idea or you have a, a hypothesis and then you go into the literature, you try to find individual studies that's, that support the essential idea to make a conclusion on a certain topic. But you can see that this, this can be problematic because it's usually subjective and the essential idea might, might not be correct. And usually if you have an idea, even if it's wrong, you usually find studies in the literature that support the idea. So this is not a gold standard because it's it's subjective and usually the the authors do not explain or give the step by step breakdown in which they went about identifying the different studies to come up with the conclusion that they are presenting with. So this is usually lower in terms of the strength or the level of evidence. So what we want to be doing is um systematic reviews that use more vigorous, systematic and transparent approach to gather and appraise or synthesize evidence to an a well defined question. The methods that you use to select or interpret the findings is usually transparent and are therefore less um prone to bias. Now, systematic reviews, the problem with them is that they are time consuming and they require a lot of resource, usually human resource. So for example, um if you are starting a systematic review for your first time, it might take an average of 18 months just to complete a systematic review. But that time, I might considerably shorten if you are doing a systematic review, people who are already experienced in, in conducting one and it can even be as, as short as three months or even six months. So hopefully, you appreciate at this point, the need to for a multidisciplinary team, usually you, you'll be walking with in an ideal scenario, you'll be walking with Liberians. You will see why Lib where Liberians come to systematic review and why they are instrumental, you'll be working with maybe epidemiologist and statisticians and then clinicians as well. So it's usually a multidisciplinary team and it's advisable to, to form this team when you want to conduct a systematic review rather than doing it alone. So why do you want therefore to, to perform systematic review? The the the main reason is to you want. So you have an an area of research, for example, there are many different findings on that particular area of research. There are studies that have shown, for example, that a particular drug will work for some disease and other studies are showing that it doesn't work. And so there is a controversy. So you want to, you want to summarize all of that evidence to come up with a conclusion with a systematic and logical conclusion, maybe there is already evidence that a particular drug doesn't work. And there are new studies coming up showing that the intervention works and you want to update the previous evidence and try to discuss why there is this discrepancy in the evidence. So those are some of the the scenarios where you need to do a systematic review. So at times you might be applying for, for a grant and you need to conduct a systematic review around the research question that you are applying for that grant to be able to support your grant application. So this is one area and finally, you might be doing a systematic review and we see this um for many different societies, like for example, the European Society of Cardiology, they conduct reviews in order to, to, to inform clinical guidelines. So we go, we, we switch gears a little bit and talk about. So we have talked about narrative reviews where it's usually subjective and there is usually no documentation in which you go about identifying the different studies that you include in the review. We have talked about systematic review, the endoscoping reviews. They are more, they are, they are a little bit similar to systematic reviews. Uh But there are some sorts of differences that we we are going to to talk about. And in scoping reviews, the aim is just to identify the nature and the extent of research evidence on a particular. Um um on a specific topic, we'll, we'll talk about that a little bit more simply put, we just want to know in this case, we just want to know what is out there. This is a little bit different from a systematic review where you have a very precise question that you are trying to answer. And so it's copy reviews, therefore involve, they also involve a systematic approach. So you have to document how you went about identifying the studies that you're included in your review, that there is less we about the quality of the studies that you are including in that review. And that is the key difference between scoping reviews and systematic reviews. So the first difference is in how you frame the question for the review between scoping and systematic review and the legal in the quality of the studies that you are including in your review. So it we go back again to the question of when do you therefore use a systematic review or a scoping review. So this is the first example, this is a systematic review. And the main aim here is I want you to appreciate how precise the question for the systematic review is compared to the the research question for the scoping review that we are going to talk about. So the for example, in this case, the authors want to investigate whether um clo clopidogrel and aspirin or double antiplatelet therapy has added benefits above and beyond aspirin alone in preventing recurrent thrombotic events in patient with a minor stroke or high risks of transient ischemic attack. So you, you appreciate you see how precise the question is, and this is the type of questions that you want to formulate to um for systematic reviews. And we are going to go into the systematic reviews questions a little bit more to try to break it down so that you see the different components of the question. So if if we look at this case, this one was a scoping review that, that we did a few years ago, you see that the the question for the scoping review is a little bit is is broader compared to the systematic review. So in this case, the authors just wanted to know summarize available data on the prevalence associated factors, etiologies, comorbidities, treatments and all of that. So you see that it's broader than the question for the systematic review. So it's less precise. So it's we just wanted to know what is out there concerning um chronic kidney disease and summarize that evidence. So it's hoping if you just know, want to know what is out there, the question is a little bit more open and good. And you want to summarize the evidence, then for systematic review, you have a precise question about a particular topic and you want to now take the gather the evidence and answer that that precise question. So in this case, the difference between the two reviews is how precise your question is. So let's move on to how at this stage, if if you have any question, please feel free to, to ask and we'll discuss a little bit further about it before I continue. So let's look at how you can conduct your own systematic review. And the first thing that most people overlook, even people, some people who are already familiar with conducted systematic reviews is to know if w what you are trying to do has already been done. And so you find people who have gone through all the different stages, they have written a protocol because ideally you want to write a protocol before doing the systematic review. They have already written a fullblown protocol for a systematic review. They invite me for example, for authorship and then you do it some searches and you find that what they are trying to do has already been done. And so all of that effort is wasted. So if you have a research question and you want to do a systematic review, we have talked about how intense and how long it takes to complete one systematic review. So you need to on, you need to know if what you are trying to do has already been done. So you want to search different databases. So if you are looking at for example, interventional studies or clinical trials or things that are trying to see if one intervention or a drug or a, a device works for a particular condition, you can look at um the Cochran database for systematic reviews is mostly for interventional studies. The Prospero um database is where you register a systematic review. So it's worth looking at the Prospero Database because mm the the authors might be working already on your ID even though it's not published. So you may want to see if they are already registered that um a review on what you are trying to work on. So it, it's definitely worth looking at the possible database and the other things you can look at, you can you can look at PMD, which is a database that contains, you contains um publication, health related publications. And our simple friend Gole might might also give you the answer. Now, another question is, so what if you find out that what you are trying to work on has already? So there are some authors walking on it already or they are published on it. So one thing you could do is you could say, OK, when was the that review published? Was it published? For example, 10 years ago, is it possible that the evidence has changed over time? And is there therefore a need to update the available evidence? It's one thing you could consider. So even if it has been done before, it's possible to update the available evidence. Another thing is to see, OK, uh can you ab adapt your question to, to answer some additional aspect of that, that topic that the calling you did not address is one thing one could consider or you Yeah. So, so the, these are some of the things if you find out that what you are trying to walk on has already been done, you can always find a niche if it's something that you are really interested in and, and, and add additional information on that topic. So after you, you have, you, you are searching that what you are trying to do has not been done before you, you need to go on to formulate the search. But before you go on to formulate the search, you need to to refine your question in such a way that it's, it's it can be answered using a systematic review, not the systematic review review question is very important because it is going to guide how you, you formulate your search strategies. So we'll talk about what the search strategy is. It is going to define um how the eligibility of studies that you are going to include in your study. So it's very important that you get the question right from the get go before you even study your search. So what is the the systematic search therefore aims to, to identify um it aims to identify all relevant studies that have been published in a particular area. So the aim, the key, the key goal of a systematic review is you have a question that you are trying to answer and you want to identify in a review in an ID scenario. Every possible study that is eligible and has been published on that topic. You want to identify all of those studies and then summarize the evidence and come up with a conclusion. I see. So and the way you do that, you need to formulate a clear question that we, we have talked about develop a protocol. Now, once you develop a protocol, you need to decide the, so the sources in which you want to search for these individual studies, we'll talk about different databases that you could use to search for studies that you include in your systematic review. So once you have decided on the the sources, you go on to develop the search strategy and then you see your search results. So let's talk up a little bit about how you formulate the question. So there is this framework that is called the PO. So it's the PO framework for systematic review and it guides you on the different components that are supposed to be included in your systematic review question. So the PCP is for, for the population or the participants that you you are concerned with. So for example, are they adults? Are they women of, of of the productive age? Are they Children? For example, or maybe a group of patients that have had a specific intervention or a group of patients with a specific condition, say diabetes, for example, I it could be I or E depending on the type of study that you're interested in So for example, if you are interested in systemat um randomized controlled trials, for example, you'll be looking at mostly interventions but if you are interested in observational studies, so an observational study will be those um is diabetes, for example, a risk factor for um myocardiac infarction. And you want to just look at risk factors, this will be observational studies. And instead of I you are going to have exposure and in this case, exposure is going to be diabetes. So you need a comparator group, not always, but you need a comparator group if you are looking at um intervention studies, um or a control group. So in, in the case of is diabetes a risk factor for your myocardiac infarction, you'll be comparing the exposure, which is those with diabetes to the controls, who, who, who are um those without diabetes. So you need a complete of group. The next thing that your question is supposed to contain is the outcome. What will be the outcome? We'll look at this in further detail. These two are, they're optional, it, it, they look at the time. Um So for example, are you interested in including studies after 2010 for some particular reason? Um So if you want to look at um studies on COVID-19, for example, you it's needless to start looking at studies that were published in say 2008, except you have some specific reason to look at COVID-19, which didn't even exist at the time really. And then the last is the study design. So what type of study designs will you be including in your studies? So I if you are concerned with the interventions and say, for example, there's a particular drug prevent some disease, you want to be including randomized controlled trials. Ideally, or if you are concerned with observational studies, you need to decide if you are including cross sectional studies, case control studies or prospective course studies, depending on what would be the best type of study to answer the question that you're posing. So if we look at an example of a systematic review question guided by that um the framework that we are talking about. So one question would be uh adults with rheumatic arthritis, all likely to have a heart attack if they are taking Celecoxib D diclofenac. So the participants or the population in this question will be adults with rheumatic uh uh rheumatoid arthritis, right? The intervention. So in this case, we are looking at the intervention and and anti intervention. So they say it makes sense to follow the the codes. The intervention here will be Celecoxib, which we are interested in. We are comparing Celecoxib to diclofenac, the outcome will be a little more likely to have a heart attack. So the heart attack is is will be a simple thing. But in this case, we we want to be more clinical and say myocardial infarction. The time will be a little bit relative. So for example, you can say if they take it for at least four weeks, so you want to include the participants who took it at for at least four weeks. Now, it depends on the question that you are trying to answer. So maybe for example, if they took it for less than a week, it has not been shown to have any detrimental effect compared to Diclofenac. And you want to look at a, a longer um period of exposure. And then of course, since it's an intervention study, the gold standards for to answer this question will be to include randomized controlled trials. Now, one thing about systematic reviews and the question is based on my experience is you want to, especially if you are starting off, you want to keep the question as simple as possible that it should be relevant. You don't want to overload your back at most of the times we are enthusi enthusiastic about what we want to do. I want to answer a lot of things. But if your question is just too broad, that review might end up in the be you might end up not finishing the review. So you want to keep your question as simple as possible. Although relevant then outside reliably like, like if you see the donkey in the background, don't, don't, don't bite more than what you can chew. So once you are clear with the question and you see that I it, it, it fits into the P code framework that we have looked at, looked at, you can go ahead and develop your protocol. So one good thing about systematic reviews is that you are encouraged, you are encouraged, although you usually most people don't, some people don't do it, but you are encouraged to develop a protocol on how you are going to conduct a review and present your findings and publish that protocol before. So publish it in a peer review journal or you can publish it on the poster website that we talked about before starting a review. So that in case you have this minimizes by us when you start looking at the data. So in case you have any deviation from your main publication, if your main publication deviate from the protocol itself, then you need to do some explaining. So it's important to publish a protocol beforehand. So the key elements of the protocol will be to provide some some summary of the relevant background of what the question you are. You are, you are trying to answer you state the review question clearly and then um you want to be clear about the eligibility criteria. So in the case of a review, it's different from individual studies where we recruit participants to to take part in our study. In the case of a review, the unit of analysis, uh those individual studies, not individual participant. So you need to be talking about the inclusion and the exclusion criteria of different studies, not the participants themselves. If that makes sense. Now there to, there's, there's a, a guide or a checklist called the Prisma Prima P for protocols. So this checklist can help you, it can guide you to guide you a protocol if you ever need to do to write one. And this, this is the link to, to the updated guidelines and you can always use the Prospero website to to publish your yeah protocol before starting the review. And it's always a good idea to to do the protocol because it forces you to think carefully about the question carefully about the studies that you want to include in that review and the data that you need to answer the question and how you are going to extract the data. So we are going to talk about data extraction shortly. So we have talked about developing the protocol developing the the review question. Um You need to after you develop the the review question before designing the search strategy. So the search strategy, before designing the search strategy, you need to decide which databases you are going to use to look for studies that you are going to include in the systematic review. So there are many different databases that you can look at. The the most common ones are are these ones for health-related studies. They are MEDLINE or met. You have M BS, you have the Cochrane library, usually if you are doing an intervention study or web of science, but most people, most people use um MEDLINE or P and M BS. But obviously the more the database you search, the more likely you are to capture all of the individual studies. But again, this if the more database that you include, the more studies you have to screen before, identify the studies are relevant for your review. So you need to strike the balance between the number of databases that you want to search. Because the more database you search, the more difficult it will be to handle the the defense studies for your review. Mm Again, other sources that you could look at if you are interested in intervention study is to ask pharmaceutical companies, you could look at um clinical trial.gov usually contains um studies on interven intervention studies. Like randomized controlled trials need to be registered on Clinical trial.gov. So you could go there and look at studies that in the pipeline, you it's important to, to know the different sources in which you can identify what we call great literature. These are mostly things that are not published in the conventional websites or they are not published at all. So it could be a thesis that is lying in, in the library, for example. And then one other thing is to do what is called hand searching. So for example, you can just if you identify a particular study that's eligible, you might just want to glance through or screen carefully the reference section of that, that paper because you might have cited some other studies that are relevant for your review. So once we have decided on, so with all of these things, usually, like I said, the main, the, the main, the main database is appropriate M BS, maybe web of Science and library. So once you have developed a search strategy, um once, once you, once you have identified the different databases to search, the next thing would be to develop. Now, the search strategy. So the search strategy, you can think of it as a computer code that you are passing to a database to give you these studies that you need for your particular review. Now, if the code is wrong, you are going to get the wrong studies or you are going to miss some studies. So it's it's very important that that search strategy is carefully designed because the studies that you get from that search is going to determine the quality of your review. So that's why Liberians are experts in designing such strategies. You must, you must not work with the Liberian. If you can get access to one, it would be great. You can also learn how to design different set strategies. So people who have been working on systematic review and have taken courses on how to design such strategy, we usually know how to do this but at times, it can also be a simple, it can be a simple research, but I will not go into the details of designing the search strategy, but just know that the search strategy is very important and how you define the different set them for your review is also important and is going to determine the quality of the review. So let's let's look at um how to define the the design, the sex strategy. So if we go back to, to the research question that we looked at earlier about um Celecoxib and diclofenac compounds, adults with rheumatoid arthritis. So the aim here is to say, OK, I want to design a search that's going to capture studies that has um keywords that include rheumatoid arthritis. This diclofenac a cardiac infarction and maybe studies that only randomized control trials. So I don't want to maybe include um studies that are called sectional studies or prospective code studies. They usually it's advisable to leave except you uh experience in the search strategy. Just keep the study design open. So we start with the population usually a key. The will include terms like it can just include rheumatoid arthritis as as 110. So in a block and use that for your age or you can use individual them. So R myoid simply you separated arthritis simply for a silica, you could use the aim to identify Celecoxib. For example, you might want to use different um synonyms or you want to use um different forms of the drug to be able to identify. So you can use the Coxib or CeleBREX or celebra, for example. So again, you go the same idea for Diclofenac, you want to look at, you want to look at, um you want to look at synonyms, you go to myocardial infarction. You want to, some studies might report heart attack, for example, instead of myocardial infarction, you might want to capture that. So in the end, it's just to look at the different synonyms and the different such things that other studies might use. Ideally, we want to, we want all studies to report myocardial infarction as myocardial infarction. But this is not usually the case and that variation in which different authors report the different terms, the different key terms in your systematic review question, you want to capture that variation in as a strategy. So I'm going to leave these other two because it's just advisable to keep them open. And when you get the defense studies, then you can exclude them based on whether the duration of the intervention was less than or more than four weeks or whether the studies included in there were observational or um intervention studies. So we have looked at how to define how to define equation, how to select the different possible databases that you can select and how to define your set strategy. One thing that I didn't mention is depending on the database that you. So just like how the the codes for um windows, for example, might differ from Linux, for example, the way you define the search strategy for one database is going to differ with another database. So that's why it's it's important for you to decide which database you are going to use before defining your search strategy. So once you have you, you have selected the database, you de design your asset strategy, then you need to go on to. So you design your assets strategy, you go ahead and then you search those databases once you search the database, if I can, I can just do a short demo for you guys. Um So if we can, if I'll just open P met, for example. So this, this is one of the databases that we looked at and this is spoke me and we can see um let's see once you answer the question is diabetes a risk factor for myocardiac infarction. And so in that such time, we we are going to include diabetes. Another component of the S would be um um myocardiac infarction. Oh So another form of diabetes will be it could be hyperglycemia. Maybe the others include hypoglycemia. For some reason, we are going to include them, combine with uh something that they call. Oh So this is a brilliant what they call bullion operators where you are, you are combined, you are combining different and 10. So yeah, the combining and synonym. So I want studies that include the term diabetes or hyperglycemia to represent diabetes or, I don't know, um, let's keep it simple. The next thing will be myocardiac infarction, which is the outcome that we are looking at. So I want something that doesn't just include diabetes or myocardiac infarction. I want a study that's talking about diabetes and the myocardiac infarction because it's, it's a little bit like sexual and you, you don't want to include if you include diabetes and it's just going to give you studies that are talking about diabetes and hyperglycemia. They will include and myocardial infarction. So maybe include diabetes or myocardiac infarction is going to give you any study that's talking about diabetes or myocardiac infarction including studies. I'm not talking about both of them, which is not what you need. So I want to search this as a single term. So you want to search this as, as a single term. You need to, to embrace it in, in, in code and you add them with an and to keep it simple, we can just do something like this and then we'll run, we run this search and then you will see that in this database from 1948 to 2004. It's giving us 29,000 now and get you to the result. Now, this studies it in you will need to. So after we run, we have search pop me, it has given us this study. The next stage now in, in this is to take all of these 29,000 records and then go ahead and start do what they call studies screening, which is what I'm going to talking about next. So first stage is to search the database. It's not going to give you exactly the studies that you are going to need. It's going to give you a bunch of study you need to go through. Now the that bunch of studies to find the exact studies that you need to include in your review. So as an example, for example, this first study is a review. Once you include in individual studies, we don't want to include review. So a study like this will not be eligible and we will need to be excluding them. So the next stage, therefore, after we have done the part, we have searched the database, it has given us those individual studies. The next thing is to do study screening and selection. So this process is as you have seen, for example, in that simple case, we had 29,000 studies, it is very easy if you are doing this cleaning alone to make a mistake or to miss an eligible studies. So, so to add more reliability to that screening process, so the screening process is where you go through all of those records that we we have identified from the database to try to find the studies are are eligible for a review. So that process should be done independently by two people to increase the reliability of that process. So the first teach is to, to do what is called type two and abs abstract screening. So it would be practically impossible to read all of the 29,000 studies. At the goal to try to identify the studies for you, it's not possible. So the first thing is to scan the type to scan the abstract. Does it talk about diabetes? And does it talk about diabetes and myocardiac infarction or whether diabetes is a risk factor? So you scan it if at the goal, it seems like it's talking about that. You keep it for the next step. That is the full text screening, which we'll talk about. But if it's obviously not talking about it, you exclude it. If it's a review, we are not, we are not including review, you excluded. If it's a case report, we are not including. So that's where you need to go back to what is eligible for my review. What type of study then you need to start excluding them at stage. Now, after you do the type and the abstract screening, now there are some studies that when you did the type and abstract screening, they seem like they might be eligible for a review. But you need to go in detail, you need to read the full text, you need to read the methods and see if it's actually eligible for a review. So that the next stage therefore will be what we call full text screening where you read. Now, the studies carefully to see whether they are ali they are eligible for for your review. Now, when two people are doing the screening, it is possible that one person might say, ok, let's include this study, this study is eligible, the other person says it's not eligible It of it happens very frequently. Now, the best thing would be for the two of them to look at the study again, discuss about it and see if there's any change in ideas and if both of them can agree to say, ok, I see a point. Let's include the study or, ok, this study is not, is not eligible. Let's exclude it. Maybe they cannot come to a consensus that you need to, to sort what, what we call an a adjudicator. So an adjudicator is just a third or that that third author is going to look at those studies that you have conflicts over and the person is going to give you his own verdict and say, ok, maybe the study, the study is eligible. If the person says the study is eligible, you include it. If the person says otherwise, then you exclude the study. So after you go through that screening process, you go through the type a screening, the full tech screening, you'll be left. Now with studies that you are setting that are eligible for your review, then then you are ready to move on to the next stage, which is data extraction. But before we go to data extraction, I want us to talk a little bit about the importance of identifying all possible studies in the uh the field of o of research. And why it's important not to look at only studies that have been published to try to identify on published studies as well. So there are two key concepts here to, to understand. The first one is so there are two key concepts. The first one is um publication bias. So it's a concept whereby studies that um that have negative results. So studies that for example, did not show any association are less likely to be published. So the authors will not even bother publishing it or some journals who reject publishing it. So if you, if you report only positive studies, it's likely that those positive studies are are bias. So it's, it's po it's very important to include all possible studies. Then another thing is selective outcome reports in whereby the authors look at only studies uh that are important or that show a an interesting result. And so and and publish that. So it's important to, to s to identify all published studies in uh in order to, to address these different forms of bias. And we have already talked about the different sources of unpublished data. So after you do this screening, you need to document the, the different stages of the screening. So for example, we looked at so after we run the database searches, we identify this number of studies, you need to report that we didn't talk about removal of duplicates. But if you search more than one database, it's likely that um more than one study will be found in multiple databases and you will need to exclude those to prevent duplication of effort. Then you need to report on the full text and for the full text each. Therefore, you need to, to explain or give you a reason why you exclude some particular studies and quantify those. And then you end with the number of studies that were included in in your review, then you go to the data extraction stage. So after you have done that in your protocol, it's important to outline the tables and figures that are going to appear in your results. And so this is going to guide the data that should be extracted for your view and then you make it some effort at least to identify the relevant data because the data that you are looking for that review might not have been reported in the study although the study is eligible. So it's important to contact the authors. We currently or several to try to get the data for the review. So there are different source is the most common source that we have seen is journal articles where you can get data for your review. Another source is um conf abstract, but the problem with abstract is the authors usually do not report details about the methods. So some some systematic review authors do not really feel comfortable um including conference abstracts in their review. And most of the times you might see that by the time the study goes through peer review stage and eventually gets published, the findings might be different from what, what they presented in the abstract. So what what we extract there? So usually want to extract mostly what we need because there's a lot of data out there. Do you want to extract what is relevant for your review? So for example, things like the study design eligibility criteria or the method of sampling the participants your age. So since we are not looking at individual participants, we are looking at studies, we want to ex uh extra on the mean age rather than the individual participants age or the percentage of females, for example, or the percentage of participants with some comorbidities. So the type of intervention, if you are looking at intervention study or the exposure, so it could be OK. Um How was diabetes, for example, assess or the duration of of exposure, the duration of diabetes if that's relevant for your question and the outcome, how the outcome was assessed the timing of measurement and and some miscellaneous information. So it's important to, to extract only what you need. Because the more information you need to extract if you have 1000 studies or maybe 200 studies to extract data from if you have more information, it the review is just going to take unnecessarily longer time to complete. So we have talked about approaching authors if you have if you find an eligible studies, but the authors did not report the data in the format that you want. So you can send some emails to them and then try to a explaining what you are trying to do via project and ask if they can provide you data, it will be helpful to ask them precisely the type of data that you need. You can also ask them the full data set so that you analyze it. But usually authors are are hesitant to give the data set. So you might just ask them exactly what you need. If they can analyze that I give you, that would be great. So the fine. So after we have, we have selected the study to include in the review, the the next thing would be to assess the quality of the studies that you are including in the review. And this is what colleagues or bias assessment. So you want to grade the defense studies based on their quality and then draw a conclusion, you can say for example, OK, does the conclusion change if I exclude low quality studies? So after you have, I included those studies, you need to do ELIXA assessment and this is one way in which a systematic review differ from, from um scoping reviews. And there are different types of B Elixia to this link down you if you're interested in this type of things. So the next stage therefore is you have done all of these things, you have extracted your data, you have assessed the quality of these studies. The next stage will be to report your findings. So you need to report details about the assets strategy for the different databases that you search, you need to report the the prisma flow chart. So the prisma flow chart is the flow chart that shows how you included the different stages that you took to include studies in your review. Then the characteristics of the individual studies included in the review and the quality assessment, which is all the risks of bias assessment and then you summarize the key results for that review. Now I just wanted to mention if there, there there's another instance where you might need to do um statistic analysis of the results in case you did not come up with a clear conclusion. And that's the scan analysis of results from a systematic review is what we call meta analysis, which I'm not going to go into much detail. But I want to end by saying that the quality of your systematic review determined depends on the quality of individual studies that you include in the review. So therefore, if you include garbage study or poor quality studies, you are going to get poor results. So it's the time that they call garbage in and garbage out. So that will be the end of the presentation. And I'm happy to take any questions if you have. Thank you so much Valerie for this very thorough presentation as always, uh I'm very happy you included garbage in garbage out. It's one of my favorite things about research and it's also my husband's like favorite thing to, to say about analysis. So I'm, I'm glad you included that. Um I've got a couple of questions um and tell me if you don't mind if you can keep an eye on the chat if anybody has any questions. Um When I introduced Valerie earlier, I mentioned that he's quite an accomplished researcher. He's been publishing for the better half of a decade. He's got over 70 different publications. Um And I wanted to ask you just to, to kind of open up a little bit about your research journey because um now you've um so Valerie has just submitted his phd thesis at the University of Oxford. Um And then just looking back at your journey so far, uh Can you tell us kind of the things? How did you get started with research? Uh Did you, did you have to work on your own? Um Can you tell us a little bit about maybe working within a team. Um And especially because, for example, A phd is quite a solo effort, but um you can also work with uh a team when you're working on uh things like systematic reviews. So I just wanted to, to kind of um get your personal take and research. How was it when you started out, how is it now? And what is it like working with a team when you're leading a project? But also when you're contributing to a project? Yeah. OK. Thank you so much. So I think it's, it's, it's, it's, it's really important to work in a team. So when I started the first exposure that I had with research is at the end of a medical training, we are supposed to do um a dissertation or a thesis. And so that was the first ever experience I had to research. But it was, it was not all that, that, that fun when, when I submitted it, but I met a friend when I started walking, he was publishing some research papers. And when I read his first publication, I was like, I'm interested in doing research. He, it wasn't fun because he was like, OK, the he gave me, I started with a case report. So he was the one who guided me on the entire publication process. How you even go about writing a case report, how you submit it for publication. So I learned everything through him. So that's 11 the need for mentorship in terms of, of research and that's how I got started. So when I published the first one, II just fell in love with it. It was, it was just fun. And then I continued, it wasn't the easiest thing to do because I didn't really know anything concerning research. But I wanted to learn more. I guess that was, that was the main thing I wanted to learn more. So I went on goa for example, if we, he had a ki a cross sectional study that he was doing, I'll go and read about cross sectional studies. What is it all about, what the odds ratios, what the hazard ratios and all of those things? And then that's how I started growing and I started identifying. So when we started publishing, I started identifying some other researchers, local researchers who were doing some things I was interested in and then I got in contact with them and they included me in some of the projects that were doing. One of those project was um a study, a systematic review to summarize um get available evidence on heart failure in Sub Saharan Africa. I know it's my face valley view and that review sets the, the path for my admission into Oxford and getting a place for the phd because that's what even took me to Oxford because I saw that they were doing something related to that. So in the end mentorship is key, but you, you need to really want to do this. You need to research is not the easiest thing if you are starting. So you need that resilience, but you also need the mentorship, mentorship makes it easier. Thank you for that. Um So, II guess what? I what I'm also like um curious about one of the examples that you shared with us was a scoping review um for someone who is um starting to do research or does not necessarily have a lot of experience doing research as, as some of the members in our group. Um What kind of tips would you give them in terms of starting early? I know you've talked about mentorship, which I can, I can uh second more. II completely think that's uh uh especially when you're starting with the research. You, you really need someone to kind of show you the ros uh what other things um looking back you think would be helpful for you to know, going into a systematic review or like a scoping review. All right. So I would just, I would just give the, the, the part that to the, the main thing is the main thing about learning all of these things is it's good to learn about it. When you are in the process of doing it, you learn more, you learn, you learn better that way because I learned about system I taking there a course on cause on introduction to systematic reviews which ii taken before that course was helpful, but I forgot practically everything within a few months. But once I got to start doing the reviewing itself, yeah, I needed to go back and say, OK, we are talking about the question, what was the element of the question? Like what we are doing here? Um What about the, the systematic age and all of that? So you need to do um Yeah, you need to be in the process of doing it. It, it's more helpful that way. Yeah. So if, if you are, if you are, if you are someone with that thing, I want to do a scoping review, the good thing about most of these things. If you do the first one, if everything is easier from there and some of these cases as are transferrable. So if you succeed to do a systematic review, either guided by someone else for the first time, the second one would be easier. So for this coping review, for example, we have been doing systematic reviews for, we had been doing systematic reviews and then we wanted to do something on a cardias in Africa, but we're not really sure about what was out there. And then II proposed to one of my collaborators that I had seen something called a scoping review. How they and they are published about the methods on how to conduct a scoping review. So when I got, when I read through the methods I saw there was practically similar to a systematic review just that you are not doing the leaks or bis assessment. So, yeah, most of the skills are transferable and there's a lot of help out there. If you do the first one, it will be easier to do subsequent ones as well. And can you tell us um a bit about your experience about working in research within a team? What kind of challenges um come up and how can people deal with it? Especially if they're starting to work on a project together? Um What kind of things should people keep in mind? OK, let me talk about systematic review things as who we are talking about. So walking, walk, you need to walk in it seem ii guess we have already established that for systematic review. But the m the major issue especially since you, if you are working with clinicians, which is likely what you are going to be doing is that people do not simply have the time to some don't have the patient to go through those 29,000 studies and screen them. So that's one thing you need to consider the time. And so if you have 29,000 studies, for example, one approach in which we have succeeded in doing is to split that into chunks of maybe 2000. And instead of getting two people to go through all of that, we can give it to 10 people and it pay those the screening and then we combine all of that. So you need to get more hands on neck to be able to complete it. So it's that time, it's that patient. You need people who are passionate about that. Maybe people will need also to publish because people need to publish for various reasons. And so you need people who are passionate and are able to get that done. So that's one challenge as as well. Um You need to walk with someone who is experience, experience is key because you it's important to see the end of the review when you are starting it. If you, if you see the end of the review, you can track back from there and anticipate some of the challenges that you are going to have along the way. But if you have never done it before or you have, you have not done enough reviews, it will be difficult to see those challenges and try to mitigate them from the get go. And so yeah, you need to walk with some people who are experienced as well. So to get into some journals, I have not really experienced this before. But if you are thinking about publishing in high impact journals, you might need to collaborate with people with experts in your field. Usually a clinician and I have have already published in that journal uh kinda have a voice and you may want to include them as well in your review. So those are some of the things to consider. Thank you so much for your answers. I have one more question but um I'll just see if uh if anybody has any question, please feel free to kind of jump in or um send a message in the, in the, in the chat. Uh Just to kind of make a comment on what Vale has been saying. um this year as part of the research team, we're trying to again help create that kind of community where people can come team up and work on projects and, and I will be sending further details in the next uh few days to two weeks about what we're what we're planning to do so that it's a bit more concrete, we have some ideas for research projects. But uh we would also welcome if you have an area of interest that they, they um would like to work on uh to and if they have any ideas that they can bring forward. Um And what Valerie has talked about today about forming a research question that could be an exercise we uh we could do so if you have any idea that's not necessarily formulated, uh please feel free to, to bring that and then we can work together to um to formulate that into more of a, a research question. Uh What my last question really? Uh And I'll just keep uh checking if we get any more question from uh our attendees. So uh part of what you do is you have multiple high salary. But one of the things you do is you do, you are an editor in some journals. So you do review a number of research and kind of, you know, have a say on what get gets published and what doesn't really get published um for people who are working, who are trying to do research with the intention of having that published. Uh What do you look for when you're reviewing a paper uh to, to make the decision of, is this, you know, useful is this novel? What kind of things do you do you look for? And one of the things that I'm hoping is um you could also answer is because people have this idea that for something to be published, it has to be like groundbreaking. Uh And if you could just shed a bit more light on, on that. Yeah. OK. So I guess it depends on the in terms of mobility, it depends on, on the journal and, and the priorities there, some journals that is just too difficult to get into and you need something groundbreaking to get into like like the lancet, for example, the there are other journals that do not really focus a lot on novelty and no, it doesn't really need to be something that has groundbreaking. So it could be something that has been done before in Western countries and you are doing it in some other country, um some nonwestern country as well. So that is novel because you are doing it on another population and that strengthens the evidence that's available out there. Another th one thing that I look at is, is the research question clear because if you, if you look at four quality people as it starts from the, the way the question has been asked. And if the research question is not clear, I already start having it some doubts about the people. And it's going to show down the line because it's going to determine on a lot of different things how the search was done, how the studies were included in the review. If it's an analytic study, how the data was analyzed and how they draw a conclusion in the end to support the question that was even not as well in the first place. So I look at the question and make sure that the question is clear, then go on to the methods to make sure that it it has been well done. If it's a systematic review, how many databases were were such, how is the search strategy, correct? Um So looking at the methods and the way the data was analyzed, then you go on, I'll go on to the way the or discuss the con the the results and then conclude based on that. So you most of the times it's easy for authors. It's common for authors to be too passionate about their findings and over extrapolated beyond what the data cannot support. So some of those things will lead to ejection because you are claiming what your data is not even supporting in, in the first place. So, so those are some of the things but one other issue as well is is the writing. If it's clear enough or not. At times, there are some writings that can be managed and that the editorial team can help to improve the grammar for clarity. But some writings are not just clear that it would take a lot of iterations to try to get it up to standard. It's just easier to ex to exclude some of those people as well. So, so that these are some of the things. Thank you so much, Valerie. These are all my questions. I'm just going to have a quick look. OK. Um I think that kind of covers all our questions. Really. Thank you so so much for coming today, talking to us and sharing your um experience. Uh If you have any kind of closing remarks, please go ahead. Um Or, and uh let me just have a quick look here. Yeah. Uh Thank you so much. We will send some feedback forms to the attendees. Uh And I'll, I'll share that with you in, in due time, but thank you so much and thank you for everybody who's attended today. Uh And please keep a look out in the group when we share further details about what we have uh planned next. Thank you so much. Thank you so much for helping me. Uh Just say if, if you have, if you have an idea for a systematic review, definitely go for it. It might help you just with the publication. That might also be the what is going to take you to that next level. You might, you don't know who is reading that somebody might just read your paper and that might be the breakthrough for you. So, and now you go for it. Thank you. Thank you so much. Enjoy your evening, everyone. Bye-bye.