USMLE Biostats 4: 2x2 Table, Odds Ratio, Relative risk, NNT, NNH and more!
Vložit
- čas přidán 8. 07. 2024
- Want to support the channel? Be a patron at:
/ lymed *Mistake @ 13:05: I should say that if you exercise, your chances of getting a MI is 25% of that who doesn't.
Welcome to LY Med, where I go over everything you need to know for the USMLE STEP 1, with new videos every day.
Follow along with First Aid, or with my notes which can be found here:
www.dropbox.com/sh/an1j9swvjx...
This video will introduce 2x2 tables. This table simplifies data and can be used in many ways to help you with your research studies. Recall some research studies include experimental and observational studies (which includes case control, cohort, and case sectional). Let's discuss the case control, in particular the odds ratio. Recall in this research type, we take a group of people WITH the disease and look retrospectively for risk factors. The formula to calculate this is (ad) divided by (bc). This is the formula for the odds ratio. Now how can we interpret the odds ratio? If OR is greater than 1, there is an increase chance of disease due to the risk factors. If OR = 1, then there is no link between the disease and the risk factor. If OR is less than 1, then there is an inverse and decreased risk.
Let's discuss cohort studies. This is when you take a group of people with RISK factors and follow them prospectively and see if they develop the disease. We like to look for relative risk in these studies.
You may say that there are multiple risk factors that lead to a disease, how can we only look at the risk factor in question? You will have to exclude any outliers, and that is called attributable risk (AR). We just subtract our risk in exposed and risk in unexposed. Now to calculate this even further is the number needed to harm (NNH), which is 1 divided by the attributable risk.
Our last topic will be on a 2x2 table that looks at beneficial treatment of exposure. In the video, we will calculate the relative risk (RR), as well as the relative risk reduction (1-RR). We will also calculate the absolute risk reduction and finally we will look at the number needed to treat (NNT) and that is 1 divided by the ARR.
Thanks for watching! If you found these videos helpful, please consider supporting me at www.patreon.com/LYMED
Much love, -Mike
sjhs
Your vids helped me ace my step 1! Got a 253! Lots of love and prayers towards you brother!
Great job!
I was so frustrated because I couldn't a word in biostatistics. And then i stumbled across your videos, thank you so much!
You basically saved me!
Ive watch your videos 4 yrs ago. I came back to watch again❤😁✌️
What a lad. Not doing the course you are explaining for but a similar Health based one in Pharmacy. Was able to completely link all the factors i struggled with. Thanks mate
First, thank you so much. You motivated me with these videos. Making this pretty uncomplicated! With what I learned I detected couple minor mistakes that I’d like to share for better understanding of all. Hope it helps and thanks again!
13:08-13:13 I think you worded it incorrectly. You mention here that not excercising increases the risk of MI 25% BUT you used in the numerator (of this equation: ad/bc) the upper row which is “yes” to excercising. So I believe what you meant here was actually: “if you DO excercise you’ll have a 25% relative risk for MI as compared to not excercising” (this would mean that out of 100 persons who do not excercise and develop MI, only a fraction: 25%- 25 people that do excercise will get an MI).
- 13:21-13:41: Here you are explaining Relative Risk Reduction. This equation isn’t a porportion, so just interchange word “over” for “minus”. Guess you meant-> 1 mimus 0,25: 0,75 (Since 0,25 was the relative risk). So for Relative Risk Reduction you have to REDUCe (subtract) the RELATIVE RISK-> 1- RR.
- 14:21: Absolute Risk Reduction-> think of Absolute Vodka (distilled/ pure-> take out confounders and purify data). Absolute: purify. Refuction: subtract. So subtract the RRisks. ARR: (RR no excercise)- (RR excercise)-> 0,15/ 15%.
- 15:30: NNTreat (beneficial effect) /NNHarm (most: harmfull). In this case excercise is beneficial so it’s NNT. NNT: 1/ ARR : 1/0,15 : 6,6-> rounds to 7. The NNT is 7. Read as: For every 7 people who excercise there is 1 MI prevented.
you're a gem ive been crying over biostats. you've made it so easy for me :') lyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy
I was so lost about this topic, until i found this channel.
This really helped me understand the formulas more and taught me that they actually are not as hard as they seem to be so thanks for that !
Made this easy to understand!
AWESOME EXPLANATIONS! Thank you!!!
Love your enthusiasm!
This is absolutely amazing!! Thank you!
Thank you!
AWESOME explanation
Great video - really helped.
amazing! thanks!
Good stuff bro !
Amazing!!
this was very helpful thank you
Tq for the vid...it very helpful....
thank you so much!!!
Thank youuuu!
Why don’t just every lecturer in the planet explain statistics like you do!
Ly Med. if you are given the odds ratio for patients that were treated with Treatment A vs non-treated patients of OR= 0.40. From the group of people who do not have treatment A, 12 are not sick, 15 are sick. How we can calculate the risk of having the event for a patient who received Treatment A.? The risk ratio for the treated vs non-treated patients? and
How many out of 20 patients are expected to have the event if they are treated with Treatment A?
You made me realize I learned all this in an ass-backwards way (i blame boomer teachers). Thanks for simplifying it!!
The way he talks reminds me of Chandler from Friends. Loll great video.
perfect
best tab.
Very handsome teacher , thank you so much
rad nice one. 3x3=1.
3x3 table.... fire
You do realize that you are God-send, right?? Thank you so much
did you come up with this way of thinking or did you learn in from someone else?
If I study from ur notes and videos and u world can I pass ?
Good boy
So if one performs calculations NNH and NNT on the same data set, would you get the same value?
might be but mostly not, two different concepts/"relative risks". NNH is a measure of harm or adverse effects, NNT is a measure of how many patients needed to be treated in order for one to benefit. (www.statisticshowto.datasciencecentral.com/number-needed-to-harm/)
من طرف الملباري لايك
ODDS RATIO = TP/FP / FN/TN or (Exposed with Disease/Exposed without Disease) / (Not Exposed with Disease / Not Exposed without Disease). Your ODDS RATIO Equation is Wrong. The ratio is A/B / C/D in your 2x2 table. You got lucky since you did the cross-product, but you conceptionally are wrong in your set up, understanding, and explanation. You can verify this in any biostats textbook.
There are two equations used for odds ratio. The exposure odds ratio (which is the one he used) and the disease odds ratio (which is what you are describing). They are equivalent. It does change how you describe the outcome. He is still correct as he was speaking of how the exposure changes your odds of getting the disease. In your example, you would describe people with the disease and their odds of having been exposed.
In disease odds ratio (what @cmetube describes) is equal to the (odds of disease given exposure) divided by (odds of disease given no exposure). Disease odds ratio = [(a / a+b )/(b / a+b )] / [(c / c+d )/(d / c+d)] = [a/b] / [c/d]. This is not conceptually the same as odds of exposure given disease.
try using a clipboard
Thank you !