How to get started in Clinical Research in Medical School

Um, tell me what kind of questions you are thinking about. Curiosity. I wonder—

I don't Like I don't have like something specific now, but I will think about it.

Yes, think about it. ممكن لو you think about it, you can drop it in the WhatsApp group.

Okay. I will.

راح تكون دايما عندي لما تيجي عندي تسأليني I want to do research. أول سؤال راح أسألك. What is your I wonder question? فهمت علي؟ I wonder if the new medication is as good as the old medication. I wonder if when they have a cold, how long do they wait before they come to the doctor? I wonder women that have breast cancer. When was the last— the first time that they saw the nodule, and how long did it take them to seek medical attention? I wonder what made it— took them so long. Is it because they don't want to know? Is it because they were scared? Is it because they were displaced? Is it because women are usually caretakers, so they take care of others before themselves? And so on and so forth. I wonder how many people Are late in taking care of the breast cancer because they have children compared without children. Children, are they a factor to delay breast cancer? I wonder. Yani, you get you get the information. خلي imagination the imagination go everywhere. Don't worry about if it's a scientific question, not a scientific question. It'll be fine. Okay. طيب. شغلتانية إنه once you have said to yourself, once you have to said to yourself, I wonder. Now you are you need to. Now you have to ask yourself, how can I make a good research article? عندي طريقة ببتدي بيها and I'm going to teach you how to do it. If you are going to submit a project of research للأستاذة ولا للكلية ولا medical journal. Somebody in the receiving end is going to judge it. Mohic, فنييتي إنه أحطكن as the judge. How do you judge that a research article is bad? Okay, مستعدين.

Yes.

So this is— can you see my screen? So this is 10 tips on how to spot a bad research article. أولاً the question. What is the question? وبنستعمل one of many frameworks. Is it feasible? ممكن نعملها. Like I want to do research on arsenic, but it's not feasible to give arsenic to people. So is it feasible? No, it's not. I want to study the infection that is carried by the mosquito that zetsay. Is it feasible? No, it's not. Is it interesting? Yeah, you're gonna find that this medication drops blood pressure compared to this medication, but it drops it with 1 point. Is it interesting? Not really. Instead of 12 over 5, it's gonna be 11 over 5. Well, whoop-de-doo! That does not make any difference. Is it novel, or is it just recycled from another research? Does it have good logic? And we did a research على الحبة السوداء وبنقول هل الحبة السوداء بتشفي من كورونا فيروس؟ Does it make logic? You actually have to tie up the حبة السوداء to its effect on the immune system and see if those effects are affected by COVID in order to have a logic. Is it ethical? I'm not going to give a medication to a pregnant woman to see if it's gonna cause child deformities. It's not ethical. And is it relevant? ممكن إن تكون the science is good, but no one uses that tool anymore. So is there a science to the diabetes in the urine and whether it tastes this way versus that way? There is a science, but we don't diagnose diabetes by tasting the urine anymore. We have other ways, so it's not relevant. So we use the acronym FINGER: feasible, interesting, novel, good logic, ethical, and relevant. ثانياً بنشوف ما اللي بنسأله إنه طب table one table one is demographics. Are the subjects on this study equal to my subjects? And لما بتشوف for example to the right side that was hydroxychloroquine for COVID-19. بس إما بنشوف table one. You find, for example, current smokers three percentage three percent. If you are in Palestine, you probably most of your patients are seventy percent smokers. So this study cannot be extrapolated to them. For example, you see 50% female, 50% male. So that's representative. You see that their weight is between 64 to 86 kilograms. So if your population is thin, thinner than this, it may not be representative. إلى آخره. So does the population represent My population. Then we ask, what is the sample size? لو مم لو في إنه أشوف is intervention that I did helpful to the patients. I have a formula. I have to ask what is the effect size I'm looking for. The example I gave you. Was a medication that dropped blood pressure by one point. That's an effect size. معناته إن ويدات إفكت سايز، you have this is the formula at the bottom. You use the power, you use the standard deviation, and you create this formula to calculate how much sample, how many patients you have to get into the trial powered by 90% here in order to show this difference. But then that changes. For example, if I was to use a test that measures your intelligence, IQ, for you and for me. I need to know what is the standard of deviation of that test. If the standard of deviation is 15, which is one standard deviation, we need 23 per group in order to determine the effect size I'm looking for. If alpha is 0.05, power is 90%. This is the formula. But if The difference I'm looking for is 10 points, which is 2/3 of the standard deviation. It's a smaller difference. I need bigger group to show the smaller difference. If I'm trying to show a difference of only 5 points, that means now I have to have 191 per group. And if I'm just trying to show 1 point difference, that's 4,730 per group. So when you have decided your question, I wonder if— and you have a question and you've determined, well, that question is about mortality, that question is about change in blood pressure, that question is about what are the barriers for women to access care for breast cancer, whatever that question is. You need to find out what is the definition of success. They do have access. They don't have access. They do have access means this. You define it. They don't have access means that. This blood pressure medication is effective. What does that mean, effective? Difference of 1 point from the standard of care, 10 points from the standard of care. That is what we call the effect size. You plug it into the formula and it calculates the sample size that you need to enroll in order to create research. What happens to those who drop out? Sometimes you give habas soda and nobody likes it. Some people drop out just because they don't like the way it tastes. It causes them some reactions. They don't want to come back because your schedule of visits are too frequent and they have to pick up their children from school, whatever that might be. You need to find out why there are dropouts. Were they too sick to stay in the study? Are there side effects? How do they compare with the placebo, etc.? Is the study design biased? Meaning, who's the comparator? And if you chose the comparator, somebody from a region that is very much aware of whatever you're studying versus one that is not aware, then there's a bias already. Maybe you are testing something about how people understand a native language of a region, and you chose People that grew up in that region versus people that did not grow up in that region. Well, there's an inherent bias because those that grew up in the region are familiar with the language you're studying. So the comparator group is very, very important. Blinding is very important. Every day when I assess clinical trials for sponsors, I have to ask them, how is the blinding? And they do studies about how does it taste? Is the color different? Can they generate something exactly similar? Okay. How good are the results? Just because it's statistically significant doesn't mean that it's clinically significant. It might be statistically significant to drop the blood pressure by 1 point, but clinically, that's not really significant. So you need to understand how good are the results. Do the results matter? Yes, we found that anyone who jumps without a parachute from a plane crashes and dies. The results don't matter because we don't let people jump off the airplane without Parachutes. So it's a moot point. Or, for example, they tell you that this study is really positive because it made this inhaler— the patient breathes better. By what? Are they breathing better because you measured it by a questionnaire? Are they breathing better because you measured their oxygen when they walk? Are they breathing— meaning, What is the endpoint you measured to translate it into the conclusion you're giving me? When we started with coronavirus and we wanted to see if Lhabba Sauda makes them better, what does make them better mean? So we looked at questionnaires assessing improvement of viral illness. And we only found the flu questionnaire. Well, COVID is new, so we had no questionnaire. So we had to take the flu questionnaire, do a small study, and validate it in the flu and the COVID Once we did that and we validated it, then we could use it for our study. Are you still with me? Thank you. Is this complicated? Okay, I'm gonna take this as a go. So when we look at any trial, What is the question? Are the subjects like my patients? What is the sample size and how do I calculate it through the effect size? What happened to the dropouts and why? Is the study design biased, blinding and all? How good are the results? Are they just symbolic or are they really clinically meaningful? How are missing data addressed? If I went and saw the patient but their blood pressure machine was broken that day and that machine is not giving me the blood pressure, what did they put instead? The computer does not do well with missing data. So, there are statistical methods to deal with missing data. Does the measurement matter? If I measure the blood pressure using my palpation, does it matter? Or do I need both numbers using the cuff? Check the figures. Are they figures that are inflated just so that it can look like the difference is big when in fact, If you just narrow them smart, small, they look small. Check the figures. And then how did the authors manage the confounding biases? Yes, you showed me that this improves blood pressure, but you did not account for smokers versus nonsmokers. That is a confounding bias. You did not include it in your study. Now that you know how we judge these studies, you now know that when you are studying and designing your own studies, you can make it better. So how does it make it better? Well, Let me share my screen. Downloads. Come on now. Download as PDF. I thought I did. Yes, it already exists. So let's go to desktop. And here it is. So let me share. And we are almost done, by the way. I hope that this is actually easy to grasp because we'll be doing some exercises later with it. Can you see the screen?

Yes. Okay.

I see it's good. So we're gonna continue. So what is the roadmap? The roadmap is, once you start with the I wonder, you need to have a foundation, which means, I wonder, how do I translate that into a research question? I wonder if there is any problem for those people with sleep apnea who don't take their CPAP trial to their hospital and they stay in the hospital. Are those 2 to 3 days in the hospital dangerous for them? Do they have any cardiovascular problems? That's the I wonder question. How do I turn it into the perfect question? Well, turning it into the first purpose question is first, what is the indication? Sleep apnea. What is the setting? Hospital. What is the population? Those that have sleep apnea and brought their CPAP machine, comparing to those that have sleep apnea and did not bring this. What is the outcome that I am Looking for cardiovascular. How do I define cardiovascular? Well, that's what you need to define as the author. Do I want to define a composite that asks heart attacks and high blood pressure and cardiac arrest and arrhythmias? Or do I choose one of them? That is the endpoint that makes you measure the outcome. Then how do you intend to do this? Do I want to— and it depends on just you. I don't know. Some people just want to do their research projects and get out of there. Some people want to become researchers. Some people just want to check a box and move on. So you could do it as a, you know, cross-sectional study. I show up in the morning and I go around the hospital bed by bed and I ask, Do you have sleep apnea? Do you have sleep apnea? Do you have sleep apnea? And then if you do have sleep apnea, did you bring your— how long are you here for? What is the reason you came? Why did you not bring it? How long have you been on it? So this is a cross-sectional study versus pull up the charts and do a retrospective study between now and last year versus I'm just going to start today and anybody who has sleep apnea, I'll follow them for the next 6 months. That's the methodology. And the execution is you write it and you submit it for peer review, and then it gets you back. You change whatever you do, and then you present it. لساتكم معي. So what we have discussed today is hopefully in the last 20 minutes is you didn't know where to start. Now, you know, the best way to start is start from the end. What do I want to prove? What do I like? Which area of medicine I like? And when I am in that area of medicine, then I can start to be open and say, I'll find something as I'm going around. And then one day you're like, I wonder. The minute you say that to yourself, you're like, well, that's a research question. So let's treat that a I wonder, and let's move it into a foundational research question. And that is the methodology, the outcome, the population, the endpoint, and so on. But that's all in your head. How do you make it better? By doing a literature search. Here's the mistake that most of us do. They tell you to do the more the researcher search. That's when you start. You don't start with the literature search. You start with the end. You craft your medical record or your medical question. You go back and then you do research literature about that. But then also you do research about the methodology you used. You do research if somebody else has done cross-sectional study. Maybe you can learn from them. You do research to see if that has been studied already. You do research to see, has anyone used this methodology instead of that methodology? You do research to find, I'm going to use this questionnaire. Has it been validated before? It's not all sleep apnea research trial. It's also all the stuff that you do in order to use it. So what we do is this is actually a game. And we'll play this game. in a little bit through the WhatsApp. But what I wanted you to finish is knowing something. You have forming the research question, and that research question is, what is the topic? What is the problem? What is the question? And then you apply the FINER criteria, which we did. Is it feasible? Is it interesting? Is it novel? Is it ethical, and is it relevant? Then after that, you decide about the PICO structure, which is which population are we doing it on? What is the intervention? Is it a surgery? Is it counseling on smoking cessation? Is it about vaccine? What is the comparison? You're comparing them to what? You can compare them to another city. Look, same, similar weather, similar number of inhabitants, similar. So you can do that. What is the outcome? What are you measuring? There's the outcome and there's the endpoint. The outcome is I want to see if it makes you breathe better. The endpoint is how do I measure if it makes you breathe better? 6-minute walk test, oxygen saturation, arterial blood gas. walking up the stairs? What is the endpoint that you use to measure makes you feel better? And then the time. Is it a year study, 30-day study, 1-day study? Okay. Another framework is population exposure and outcome. Then you find the gap. Usually a lot of people have done that research. So what we do is try to find the gap, what has not been studied. And how do you do that? You ask yourself, What do we know? And what is the gap? What's the missing knowledge? Maybe you do have the CPAP, sleep apnea, but you do not know the barriers by which people did not take it. Maybe they don't know. The gap is the education versus the gap is not accessing power outlets in the hospital. The question becomes, what is the specific inquiry that directly addresses the gap? What do you need to avoid? Don't try to do everything. You can't say factors affecting health in Morocco, too broad. Or the opinion, do you like telemedicine? Or yes and no dead question, is diabetes a problem? The answer is obvious. Do people die when they jump off airplanes? Yes. So don't waste your time on yes or no dead ends. And then the impossible standard. Will this cure cancer? You should say, will this improve the quality of life? Will this, you know, change the prognosis? And then we can go into some checklists. So I wanted you to at least have how do you get started? How do you make the question? How do you do a literature search? Not just the question, but also The tools, the endpoints, the methodology, the population. All of these need to be searched. And then you build that search strategy. I'm looking for all of these articles that are within these parameters for Middle Eastern, for Palestinians, for women, for all people that are within 15 to 25. So you decide that. And then once you are— once you have done that research, you identify those seminal papers, and you look at their references. You start to review them and summarize them. And then eventually what you end up doing is you have all of this stuff, and now you can meet with your mentor and say, I have an idea. Here's my question. I researched it. Here's what has been done before. But there's a gap. The gap is X. If I were to do it, this is how I would do it. But I found somebody else did it this opposite way. I think my way might give us some things, and you take all of that and you give it to a proper mentor, a friend, and you ask them, "Can you be honest with me?" And تنقذ يعني أعطيني تقييم النقد to my to my Project. The project is rejected, not you. So don't take it personal. You can always go back and change them. And what I, as a project mentor, would say to you, are the results valid? What are those results, and are they going to help patients? And then, did you look at all of the bias? How did you select these patients? How did you randomize them? Did you use a randomization on the internet, or did you use a randomization that someone can have tampered with? Did all the patients do all of the questionnaires you gave them? And attrition bias over time. Did you start to lose patients? And so on. So basically what I wanted you to go through is when you start a study, what you do is it's always starting with, I wonder. From there you turn it into a research question using those 10 things that we discussed. Once you turn it into a research question, you go into a literature review, and the literature review is not just in the topic. It's also the endpoints. It's also other people that may have done the same thing as you. It's also about how did they treat missing data in other places? Did they Use the same comparator as yours. Once you have that done, then you try to make a summary of that, and then you take it to your mentor and discuss with them. That's how you start research. There's plenty more, meaning then after that, what? How do you start to put together tools? How do you start to create a database? But this is the first steps of research that I wanted to add. Do you have any question for me, Ohtaya? Good. Okay. So what I'd like to do is in the next few days I will be adding some things on the chat to engage our thought process. Don't worry about getting it right, but be open to hearing about what does it mean to take this into a research question. When you give it to me, let me show you the things that you can make it better. So that's going to be our game in the next week. Give me an idea. How do you turn it into a research question? What is missing in that according to the FINGER framework? And then take it and put it in a project according to the PICO framework. And then from there, Can it be giving us something useful, or is it irrelevant? And then we can take the next step as to now what we got all of this. So what's the next step? Are we good? What year are you, Ayan? لو سمحت لي بالسؤال.

The third year.

The third year?

Yes.

ما شاء الله. ووانتم مختلفين يعني كان عندكم delay فامتحت حتكمل الthird year إن شاء الله.

أنا حالياً بداية الفصل الثاني.

وباقي زملائي المتأخرين في الفصل الأول فصل أول فصل تاني وهالأيام شو عم تدرسه بالسنة الثالثة؟ حالياً مديول الهيماتوليمفايد هيماتوليمفايد وهذول الأصدقاء اللي متأخرين عليك ولو إنه في السنة الثالثة شو عم بيدرسوهن؟ I think the endocrine. Endocrine. طيب. Excellent. فإن شاء الله الأسبوع الجاي حكون عامل the topic على ال endocrine and على ال ال hematolymphoid. We'll do half hour here and half hour here. And لو سمحتي ولو تشاطري مع أصدقائك my way of doing it. Is to start putting in the WhatsApp chat questions, games. Let it fallu maun. Let it rassakh ilkon concepts. Okay? If something is not there, don't ignore it. Correct me and say this is not what we're studying. Here's what we're studying. وانا ما هيرو لأن هدول كلن الالكن. I'm not borrowing anything from anybody. I make this every day fresh. For you.

تمام؟ إن شاء الله.

إن شاء الله. الله معكن. شكراً لحضورك. I will put this in the YouTube video, and then I will put the link of the YouTube video and the slides on the WhatsApp group and on the LinkedIn group. عندكم LinkedIn؟ Link. Yes, no. Am tismaini? Okay. You do not know what the LinkedIn is. فا أنتوا عارفين إيه ال LinkedIn مو هيك؟ It's a social media but professional. أوكي فأنا عملكن LinkedIn اسمه غرصة أمل لكم لوحديكن and this is what it looks like خليني أوريك so شطريه مع أخواتك لو سمحتي because I don't just pause in post in the WhatsApp. I also post in the LinkedIn for people to know what is happening. So this is what it looks like. And let me share my screen. Why can't I find the share screen? Here it is. Do you see my screen? Yes. So in my screen, هذه هذا ال page LinkedIn لإلكم مو لأي واحد تاني. ف I post in it the announcements. مبارح I posted about our session today. And what my— I post my session here, another event one month ago. But what I do, I post the YouTube video, and then I post the slides that I used. Hadikani ta'ali al-GI. Abliha, can another in, in event. And in this event we talked about OB-GYN. Okay. These are for this was GI. فهذول slides اللي عملتهم and this is the YouTube video for that. قبليها كنا عملنا كتير. ف it's been a year. So a long one on heart murmurs and how to amplify the sounds of hearts. How to do a cardiovascular examination, the anatomy of the heart supplies, how to take vital signs under pressure, and things like that. How do you summarize in rounds? The art of a one-liner. وبيجي معنا ال YouTube كمان. فهمت علي؟ Yes. So please share this with everyone. اللي عملناه اليوم كمان ساعة حيكون هناك recording of the video and the slides, and then I will continue to use the WhatsApp just to create engagement, little games, questions. اللي يترسخ هالأشياء هناك. Sounds good? طيب الله معك سلم لي على إخواتنا المسلمين كلهم هناك وعلى عيلتك والله يكون معكم ويجيب لكم الأمان والهدايا ويعني بندعي معكم كل قلبنا يلا السلام عليكم