Computer generated transcript

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

Hi, everybody. We're gonna start in a little bit. We're just gonna wait for a few more people to join. Um but welcome. I think we had quite a few sign ups. So should maybe wait a couple. Yeah, maybe wait a couple of minutes. I hope uh fourth year is going well for everyone and we're doing ok. We've just started June Sun's out. Do you have any questions as well? About fourth year? Um please do go go for it in the chart as well. Yeah, we might have some non UCL people as well to be fair. So maybe we don't just have four years. But yeah, if you wanna comment what year you're in or university feel free. Yeah. Any questions? Ok, good. Ok. So and can you just confirm? You can see that? Mm. It says exit full screen right now. Oh, ok. So I think you need to. Yeah. Uh can you make it bigger? Yeah, I think it's um I need to just share my entire screen rather than Yeah. Does that work now? Yeah, it's perfect. Should be in uh Yeah. All right. Amazing. Uh Welcome everyone uh to another teaching things. Session. So today we're going to be going through everything you need to know about statistics. Uh My name is Presha. Um and I've also got Milly on the call with me. We're both fifth year medical students at UCL. Uh So you're new here. Uh Welcome to teaching things. We do tutorials which are open to all and focused on sort of the core things you need to know about medical students. Um And we have lots of events coming up, so do sort of keep up to date on emails and group chats. Um So just a quick sort of thing about statistics. Uh it is quite a high yield topic um in a and your A KT, they do ask you to do calculations. And also it's important to practice um explaining statistics in lay person terminology as well because often patients, the scenario might be a patient comes in with a study and then you have to sort of explain the study design and explain why the results might be applicable to them or not applicable to them. So it is definitely a high yield topic for exams in terms of the things you need to know for statistics. So sort of basic definitions and then we'll go through some types of studies. Um And then we're going to go through all the calculations you also need to be able to do. So we have an exercise. Um Milly will be helping me out because I can't see the chat. So if you guys can match up, um, which sort of the definition to the terminology and let me know what you guys get. There's nothing in the chart yet, but I will let you know right. If you're not sure, just give it a guess. Um, no judgment here. Uh, oh, we've got one. Someone said B is one and A is two. Anyone else? Mhm. Someone else put a thumbs up. I had a few thumbs up. So a few agreements. Ok. Nice. All right. Yeah, let's not spend too long on this side. That is correct. Um, so if we were to define incidents versus prevalence, um, so the prevalence is sort of the snapshot, uh, or a head count in a particular moment. So it's a proportion of people who might have a specific disease or characteristic in a given, um, sort of time. Uh, if you were to maybe, like, I think analogies work well, quite work quite well with patients. So, for example, if you were to give an analogy for prevalence, you could say if we're in a hospital ward, how many patients currently have pneumonia, for example, and then incidents is the number of new cases of a characteristic develop that develop in a specified time period. So again, if we're using our ward analogy, um, it could be how many patients were newly diagnosed with pneumonia in the last week and that would be the incidence, um, of pneumonia. So, moving on, uh, talking about randomized controlled trials. So we've got four different studies on this slide. If you guys can tell me whether 123 or four, which one is, um, a randomized controlled trial? Ok. We've got someone in the chart. Um, they said two. Any other guesses? So, on that one, I got another one, a few more ones. Ok. Just chat to slow down. So. Ok. Amazing. So the answer is um study one. So a randomized controlled trial um essentially is um a trial where you have two cohorts of people. Uh One cohort is given the intervention and one might be given a placebo. Um You can either blind or double blind, the trial. So blinding means the patients don't know whether they receive the intervention or the placebo and double blind, meaning both the patients and the doctors involved or the researchers involved in the study also don't know which uh groups have received which intervention. Um So it's sort of like the gold standard. Uh the, the sort of best study to identify whether an intervention is helpful. Um Some of the limitations with a randomized controlled trial, it is quite time consuming and there's a lot of ethical issues involved, especially when you talk about giving a placebo to the other cohort group. And also it's important that the demographics of the sample that you're involving in the randomized controlled trial is going to also apply to sort of the wider population when you roll the intervention out. So another sort of common thing in acies um may be that a patient brings in a study. Um And for example, if it's a, it's a female patient bringing in a study about a new like hypertensive drug, but the study they've brought in was only testing within the male population. This is something you can bring up with the patient and say, you know, this intervention was only really research for this particular demographic. So it might not entirely be relevant for your particular case. So that's always important to also read um within the abstract as well. So other sort of terms related to a randomized controlled trial. So we've got intention to treat analysis. So this essentially means that even if patients drop out of the study, we would still analyze um sort of their data and their results to sort of mimic real world outcomes. This also means that we're not sort of biasing the study one way or another. Um So sort of to maintain that standard as well, you've got a term called composite primary outcome. So that essentially means when you combine lots of different end goals or end points into a single outcome. So for example, if you had a study, um I don't know if you, if you had a study exploring uh cardiovascular risk factors and the effects on or the chance of you getting a heart attack or, or dying, um, the chances of people.