Computer generated transcript

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

Hello, everyone. My name is Raymond Long. I'm the national research director for Scotland this year. Welcome to the first part of our critical appraisal rub in our series. Today's topic will be study design, and we are very happy to have a GI Joe's the as our speaker today. She's a final year medical student in Glasgow, and she is also our age national research director for Scotland. She asked me to be careful various journal clubs and critical appraisal talks so indeed a very bright senior look up to please feel free to type in any questions that you have at any time during the presentation. You could tap your question here if you have verified your medal account, or you could also type it on slider. Using the length in the chat. A GI will be addressing your questions in the queue and a session at the end after the Q. And it'd be great if you could fill out the feedback form, which I'll be. I'll be sending out the link later in the chat. A certificate of attendance could be automatically generated for you after completing the form. So without further ado, let's invite a GI to share her presentation today on study design. Thank you very much. Hi, everyone. Thank you for joining in today. Could I just check if everyone could see the slides? I think so. Yeah, I think it could. Yeah. All right. Great. So, um, so we'll just crack on then. So before I start, I'm just going to talk about the topic itself. So today's talk is just about a brief, brief description about study designs. And, uh, what are they? Why is it important? And the types of study designs. So this is the layout, So I'll start off with the description of study designs. What are they, um, And then why the important In terms of research and then find it's the type of study designs and also a brief description of each different types of study designs. So first things first. What our study designs so study design are basically a framework created by a researcher, which involves a set of methods of procedures. So this is usually the methodology section, and these methods are important for the collection of data and also the analysis of data for the results section in the end, and the reason why study designs are important because study designs, uh, are not always appropriate for different types of outcomes. So, for example, if you're looking to, uh, determine the efficacy of the treatment, then usually you do a randomized controlled trial. Because the methodology and randomized controlled trial is more appropriate in assessing the treatment outcome's. And in this way, the different types of study designs then determine the quality of research. It also sets boundaries in the methodology area as well, and this and that this and then proves the validity of results as well. So before we go on to the types of study designs, this is a pyramid that I wanted to show you all. So this is an evidence based medicine pyramid or a pyramid of level of evidence. So if you look right at the bottom, it has the editorials and an expert opinion. Whereas if you look right at the top, it's got a systematic review there. So at the bottom, that's the lowest level of evidence, because it's just one person's opinion on a certain topic. Whereas right at the top, the systematic review. It's a collection of different types of studies and the researchers analyze them and then finally come out with a conclusion. So that is the highest level of evidence. And in today's talk, we'll go through all of these studies. So this slide over here, it shows the different types of study designs which category they fall into. And so, for descriptive studies, they are basically a study which describes the features of the population or features of a participant. So these usually involves a case report or a case series, so they're both identical. But they're also different, different in a different way. So I will talk about this later on and the other slides and then there are also analytical studies, and these can be divided into observational studies or experimental studies. So for observational studies, the researchers do not intervene in any way. So what they do is they just observe the participants and these involved studies such as cross sectional study, case controlled study and cohort study versus for experimental studies. In these studies, the researchers would intervene in terms of, uh, treatment arm or the exposure arm, and it primarily involves randomized controlled trials, clinical trials, crossover trials or end of one trials. So for today's talk, we'll be talking. I'll be talking to about the randomized control trials, crossover trials and also the end of one trials. And so these are also like ky terminologies to look at. So when you're reading a paper and an abstract, you would come across terminologies like this. So I'll just briefly describe each one of them. So long longitudinal studies. They're usually dealing with one patient or or a few patients, and then they sort of follow up for a few days or a few months or a few years. So this is usually prospective, so dealing with the present and the future. Whereas like cross sectional studies, it's sort of like a snapshot of a single patient or a group of patients that just one point of time, so meaning, for example, like in this week, what has happened to the patients and various retrospective studies. It's basically looking at today and then going back a few months or a few years. So it's basically dealing with data that has already been, uh, present, and then we've got pragmatic studies as well. So these are studies which takes place in like a hospital setting and These studies are ideal for looking at the efficacy of an intervention versus the current standard of care, for example, whereas explanatory studies, they take place in a proper ideal setting so usually like a research lab. So this is to look at, for example, like a treatment arm versus placebo and to see its efficacy. All right, now we'll move on to the first type of study design. So this is the case report. So this is a type of observational study, and it always involves a single patient. So what researchers are looking for in case report is basically something interesting or something rare that they've seen. So it's basically about, uh, for example, like, let's say, a patient comes up with a really rare disease, which the doctors are not really sure about. They will write up about this case so that eventually it can be sort of further research upon as well. But the problem with studies like this is that there's a lot of bias because it doesn't really have a proper methodology, and it could also be just chance that the patient is presenting with symptoms such as this. And then we move on to case series, so the case series is basically a case report, but it involves more than one patient. So it's a group of patients, Um, and it's very good for studying rare disease because you want to follow up with these patients and see what's been going on with them. Or you want to basically let everyone know that there's this sort of rare disease, and it's presenting in this sort of population and then moving on to a cohort study. So for a cohort study, I would put this below a randomized controlled trial. But it also holds a good level of evidence because there's a proper methodology in in order to do a cohort study, it still falls under the observational study category. But what happens here is patients are selected based on exposure, for example, so you expose these patients to a certain type of risk, and then you follow up with them. So it's a prospective study, so you follow up for months or years and then finally see whether, um, they have an outcome or not. So, for example, um, this is just an example. So let's say you've got to court so one cohort. You expose them to smoking, so you make them smoke, and then the other cohort, you make sure they don't smoke and then you follow up with them and see which cohort develops lung cancer. So that's a cohort study. Um, so it's very good because because you're following up with the group so you can actually collect good quality of records. So you have good data collection, but the issue is that it takes a really long time. It's expensive because you have to follow up with all these patients. And there's a very high risk of attrition bias. So meaning patients dropping out and with attrition bias you. The data, like the outcome of the results, can sort of skewed to a certain range. Um, so that's the issues with court studies. And where is the case controlled study is sort of the opposite of a cohort study. So in the case control study, you already have the outcome. So what you're doing is you're looking back and you're trying to see what exposures or what risks these patients have been exposed to. So it's it's always a retrospective study, and it's very good to investigate any rare disease or anything new. So you have. Like, for example, you have these patients with, um, let's say, breast cancer and another group without breast cancer. And then you look back and you questioned them about all the exposures that they've been exposed to, any risk factors that they've been exposed to. So the good thing about this study is it's quick, It's cheap, Um, but the bad thing is that it's prone to recall bias. So recall bias is basically when you're asking someone, they recall things. So when you look at a group who has been diagnosed with a certain disease, for example, cancer, they tend to over think about everything that they say that they've been exposed to. And they tend to list down all these things where as a group which has not been exposed, which has not been diagnosed with cancer, they don't tend to over think So. This is where the recall bias comes here because you're not really sure whether it's that particular risk which is leading to that sort of cancer. And then this is cross sectional study. So this is basically a snapshot, so meaning, for example, you just want to study in this period of time. So for this one month, what's been going on with these participants? So it's very good because you can determine the prevalence. So prevalence, as in, like how, like the number of diseases or the people diagnosed with this disease is in this specific time frame. But the issue with this is that it lacks, um, so there is because there's no time component, So you're not following up with them or you're not looking back at how long this has been going for so lax temporality. So there's no relationship with time, basically, and, um, but the good thing is, you don't need to follow up. It's quick and cheap, and it's very good for, like, immediate, um, data collection and results and moving on to randomized controlled trial. So this would sit high up in the level of evidence pyramid, because this requires a proper framework, proper methodology and procedures. So this is a type of experimental study. So this is where the researchers would intervene. And, uh, so what happens is if you look at the graph or the figure on the right. Sorry. Um, so you've got two groups and they're randomized, so they're randomly assigned to either the treatment group or the control group. And usually it's done by, like, a random number generator or a statistician would do this for you, and, uh, and then they go on to get intervention, or they're going to get the control, and then you finally see what the outcomes are. So the issue with the randomized controlled trial is it requires a lot of ethical approval because you're giving patients 11 group of patients, the treatment arm and then the other group of patients the control, which might not be the standard of care as well. Uh, and another thing is, it's It's really time consuming. It takes years. It's difficult as well, and it's very expensive as well, because you have to follow up for a long period of time. But the really good thing is that in terms of assessing efficacy of the treatment, So, for example, if a farmer company comes out with a new drug and they want to assess whether this drug is good enough or not, should it be, you know, used in the public? This is the gold standard study that they would use to assess the efficacy of this sort of treatment and then now moving on to cross over trial. So this is also an experimental study, so the researchers would be, uh, would intervene as well. But the thing with the crossover trial is, so the participants at the start they would be assigned to two different groups, and then, after a certain period of time, they will swap and you'll get two different treatment. So if you look at the figure on the right, so you can see that the participants eventually get both the treatment as well. But there is also like a washout period in the middle because you don't want the participants to bring forward the side effects of the first treatment that they got. And this is This is useful because you want to assess which treatment works better in a certain group. And it's also very good when when there's a lack of patients in the study. So, for example, like for a certain sort of like, um, disease, you're looking for a drug treatment and not many patients are diagnosed with the disease, so what happens with this is you get a single patient would provide data for two different treatments. Um, so and this is the end of one trial. So this is similar to the crossover trial, but this is just looking at one person, and this is this also involves two different treatments. So you give the patient treatment A for a short period for a period of time. And then after that, there's a washout period, so they do not take treatment A during this time. So that's to minimize the side effects that would move on neutron. And also it's determined based on the half life of the drug as well given, and then once the washout period ends, you want to give them treatment be, and this is very good because it evaluates which treatment works the best for this patient. Um, the downside of something like this is it also requires a long time as well to see which one to see which treatment the patient responds best to, and which treatment doesn't provide too many side effects to the patient as well. And then so this is an audit. So an audit is basically it's basically to look at the current standard of care and to also improve the current standard of care. Um, so what happens is you identify an issue and then you collect the data and then you compare and see what has been going on and then see whether you can make any changes or implement any policies, and then you reorder it again. So this is, for example, in the hospital or in a certain ward. You're looking at antibiotic therapy for, let's say, lung infections and you want to see you got these two different antibiotics and you want to see which one is better than the other one, or whether the current antibiotic care that you're giving to the patient is good enough or not. So an audit is very good for local guidelines. You've also got national audits good for national guidelines as well. But the most important thing to take away from this is that it improves the current standard of care. It doesn't give any new knowledge. The new knowledge is basically given by, like randomized control trials, but this one gives it helps with quality improvement, basically, and in order to have a proper audit, that has to be a minimum of two cycles in the same sort of audit that you're doing. Um, well, the cons about this is that because you need a minimum of two cycles and also a period in between where you are, where the policy implementation is going on. So it does take quite some time to actually see whether it's application or not. And also it requires lots of data collection as well. So, for example, if you're looking at the region so let's say let's take Scotland, for example, loads of patients with lung infections and requiring requiring antibiotics therapy. So it's loads of data collection and lots of time consuming as well. All right, Okay. And then this is just a Caribbean metabolizes. So this stands right at the top of the pyramid. So this is the highest level of evidence. So the systematic review on the market analysis, what they basically do is they look at, um, for example, they look at a certain disease and then they look at the types of studies done on the disease, for example, so a bunch of randomized control trials that were done on, for example, the treatment for that disease, and then they see whether they compare all the studies and they sort of see where the outcome lies it and they draw a conclusion. So, for example, if you look at the figure on the right, so on the left, you can see studies. So there's, um, Oakland blog door in so that those are all the studies, and then it shows them comparing steroids with placebo and the deaths of the number of deaths as well. So you can see that they all have different datas for the same type of, um, outcome. So what? This this is a Forrest Forrest plot. Sorry. And this forest plot is basically Sinemet analysis, and it compares, um, but the you know where the data lies that so you can see right at the bottom. There's a diamond. So the diamond is basically, um it consists of all the data that you see all the lines at the top, and it shows whether, um, the data is significant or not significant or not. And basically, this is a comparison of systematic review and met analysis. Um, so, in terms of guidelines, the way they include their patients, excluded patients will get the papers. It's similar, but only difference between the meta-analysis and a systematic review is that a meta analysis is always quantitative, whereas the systematic review can be quantitative and qualitative. So meaning meta analysis always requires statistics. So a lot of calculations will be done to draw out the grafts or the figures or the tables. Whereas the systematic review it doesn't always necessarily need, um, statistics of calculations, it can always be qualitative. So, for example, you can look at papers, for example, barriers and facilitators, and then draw a conclusion on what are the barriers for implementing this? And what are the facilitators for implementing this? So the key take away from this slide is Montana's. This requires calculations and statistics, and systematic review doesn't always require it, right? So in conclusion, we talked about talked about the study designs. What are they? It's a framework. It has a set of methods to collect data and told us to analyze data. And this is important because it is required for high quality research and for the appropriate, um, disease or outcomes that you're intending to to investigate. And we also looked at the different types of observational studies, experimental studies and also the different types of study designs individually. And why the important as well? All right, so that concludes my, um, talk. Thank you all very much. And, um, I'm open to any questions. Yep. So please feel free to type in any questions that you have in the chat If you could, or you could type it on over slide as well. Which the link could be fine in the chat. All right. So I don't see any questions. I mean, if you guys don't have any questions, just feel free to fill up the feedback form. And then, uh, we can call it a day then. And thank you for joining in today as well. Yep. So So just a bit for the next talk. Um, Yep. So our part two of the webinar series will be held on next Saturday on the 20th, which a GI will speak again. And the topic will be on quantitative paper. So please look out for a promotion on our instagram, Um, at times a dash Scotland in the coming week. We hope to see you again very soon. And thank you again for being the speaker for today. And thank you very much for joining us today and have a good night. Thank you very much. Thank you, Raymond. Thank you. Yeah, yeah. If you could just feel in the feedback from and there'll be a certificate generated for you if you have the event on account and it will store in in your account, Right. Thank you very much. You have a good night.