Randy Neil Biostats 14: USMLE Step 1 - Viewer Requested Biostats Questions #1

USMLE purpose: Solve viewer-style relative risk and study-design questions by reading the final sentence first and identifying what comparison is being requested.

How to use this page

1. Read the final sentence first: what metric are they asking for?
2. For ratio questions, place the first mentioned group on top and the comparison group on bottom.
3. For study design questions, ask whether anyone had the outcome at the start.
4. Open the answer toggles only after attempting each question.

One-liner

For relative-risk comparison questions, the phrase “A compared to B” means calculate A ÷ B .

High-yield rules from this episode

Anki-style rapid recall

Practice Questions

Question 1

The risk is 0.5 in heavy vapers and 0.1 in non-vapers. What is the relative risk for heavy vapers compared with non-vapers?

• A) 0.2
• B) 0.5
• C) 5
• D) 10

Question 2

A study starts with vaccinated and unvaccinated patients, then follows both groups for migraine development. Which measure is most likely reported?

• A) Odds ratio
• B) Relative risk
• C) Prevalence
• D) Sensitivity

Quick Reference Summary

📌 Transcript note: The transcript is kept below for reference only. Use it if you want Randy’s original explanation after mastering the high-yield review above.

Full Transcript

Alright guys, so this is a quick video. These are two questions that we're sending by a viewer, DeVica. Thank you for sending these in. So anytime you guys get questions especially from the bio stats that are difficult or you just see some kind of fine tuning on how to work them. You know, go ahead and send them to the email and be in the description below and I'll see if I can kind of re-word them per se. And again, the more questions you go through, the more you can kind of see them being worked out, I think they're often be prepared for this step exam. So hope liked the video. Alright guys, so in a question that's really long, you just want to go to the last sentence and then read what the question is. It'll help kind of direct what you should be focusing in on. You know that it's a bio stats because basically there's some numbers involved. It says based on this information, which of the following is the relative risk of COPD before a heavy vapor here compared to a non-vapor. Okay, so they're talking about vaping and relative risk, okay. Now, so let's go to read the question. It says, a 45-year-old man presents to his primary care physician because of his family is concerned about his constant cough after vaping, you know, basically essentially smoking. The patient inquires about the risk of developing chronic COPD, chronic obstructive pulmonary disease, COPD, and adults who vape. A cohort study revealed that compared to smokers, the relative risk of COPD and non-vapers is 0.1 and the heavy vaping is 0.5, okay. So COPD correlation, non-vapers 0.1, heavy vaping 0.5. The patient considers himself to be a heavy vapor. Based on this information, which of the following is the relative risk of COPD for a heavy vapor compared to a non-vapor. Now, the key to this problem, the key to this problem is right here is that when again, and I said this before, but you get it,

when you read math problems, you read them top to bottom left to right. Whatever comes first goes on top, whatever said second goes on bottom because the key is right here. They want to know what's the relative risk for COPD for a heavy vapor compared to non-vapers. So which one was said first? Heavy vapor or non-vapers? Well, the first one was the one that said first was heavy vapor. So heavy vapor compared to or divided by non-vapers. And that's really all they're asking. So the relative risk and that's associated with the COPD and we'll talk about this in the next question a little bit more, but relative risk is going to be essentially, well, you're comparing the incidence of a disease of what occurs from the exposure versus the non-exposed or the heavy vaping versus the non-vaping. But again, whatever is said first goes on top. So in this sense, the heavy vaping, as we know, heavy vaping is, is the correlation is 0.5. The non-vapor is going to be 0.1. And of course, that's going to be, when you divide 0.5 to that, I'd point 0.1, it's going to be 1, which is going to be answer choice B. Now, they could easily flip this. I could say which of the following is the relative risk of COPD for a non-vapor compared to a heavy vapor. In that case, if they would have said non-vapor first, you'd put it on top and so on. That's the difference. Whatever is said first goes on top, correct answer B. And this question, obviously, it's another long one. So let's go ahead and read the question. It says, which of the following measures of association or the researchers most likely to report? You look at our answer choices. OK, odds ratio, relative risk, incidence prevalence, confounding. OK, well, when I see this, incidence number of new, over the course of period of time, and then confounding, that's kind of the obscure one. I can tell you this. Step one, step two, step three. They

really want you to know if you know the difference between odds ratio and relative risk. OK, so just to review back on these, if you see the word case control, you better be thinking the word odds ratio. And that's one number. I'm using letters, represent numbers. One number over one number. OK, if I said cohort, you better be associate that with the word relative risk. So any time in a question, if I see the word case control, you better be thinking odds ratio, and you better be using a formula like this, one number over one number. If I see the word cohort anywhere or a description of a cohort, I used the word relative risk. And that's one number over two. And if you have trouble understanding that, go back to the original bias stats videos that explains this a little more detail on a lot slower. All right, so let's think about this. Remember, case control has an A and an O. That means one group has the disease. The other group does not. Cohort has two O's. So neither of them, two O's are the same. Neither of them have the disease. That's your other key point. Now back to this question. A research study was performed to determine if there was an association between vaccination and migraine headache. In patients living in a northeast or a United States town, the researchers selected random patients who had received the new vaccine and a random set of patients in the same area who had not. So you got one group with the vaccine and one group that did not. The two samples of residents from the same town were assessed and the occurrence of migraine headaches after the administration of the vaccine which the following measures of the association are other researchers most likely to report. All right, so again, let's just, they're not really looking for the number of new cases per say. They're looking the difference between the people who got the vaccine and the people

who did not. But here's your big takeaway point, okay? At the beginning of this study, did anybody have the migraine headaches, right? They talk about a lot, but at the beginning of the day, did this study start with a group of people who had migraines and then no migraines? No, they did not, right? Nobody had the disease. When they started this, nobody had the disease. When I say disease, I mean migraines. The only thing that was different, one group had the vaccine and one group did not. Nothing about the migraines. So at the beginning of this one, they're looking at a cohort, all right? And if you say cohort, your next words out of your mouth better be relative risk, okay? And then you better know formulas, but for this question you don't need that. But that's the key. The key point is case control A and I know one group has the disease, one group does not at the beginning. A cohort study, neither group has disease. It's something like this where they're kind of anticipating, you know, what they, where they compare the incidents of a disease between the exposed, between the exposed and the not exposed. Okay? So on this one they're looking for be relative risk. So again, thank you for sending in these questions. It does kind of help to know where people are at and hope the video is helpful. Thank you.