USMLE Biostats 5: Sensitivity and Specificity, NPV and PPV and More!
Vložit
- čas přidán 9. 09. 2017
- Want to support the channel? Be a patron at:
/ lymed 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...
Continuing our talk on biostats! We'll begin with some terminology. Let's imagine we have a test to look for HIV. If a patient has HIV and the test is positive, that's a true positive. If the patient doesn't have HIV and the test is negative, that's a true negative. If the test is positive, but the patient does not has the disease, that's a false positive. Lastly, if the test is negative but the patient does have the disease, that's a false negative.
Now let's discuss positive predictive value (PPV) and negative predictive value (NPV). Positive predictive value is the probability that subjects with a positive screening test truly have the disease. Negative predictive value is the probability that subjects with a negative screening test truly don't have the disease. Next up, let's talk sensitivity. Sensitivity refers to the test's ability to correctly detect patients who do have the condition. In a sensitive test, if it is negative, it rules out the disease. This is related to negative predictive value. When we discuss sensitivity, the worse case scenario includes a false negative. The more sensitive a test, the less false negatives. Now let's discuss specificity. This is the bbility for a test to detect a particular disease. If it is positive, it's likely the patient has that particular disease. This is related to positive predictive value. The worse case scenario occurs when there are false positives and that has to factored in. The last part of this video will discuss a common graph they use to synthesize all of these factors.
Our last topic will be on incidence and prevalence. Incidence looks at how many new cases there are in a population. Prevalence is the number of TOTAL cases in a population. Prevalence can change PPV and NPV. Know that incidence and prevalence are related to each other by time. In chronic cases, prevelance is often higher. In acute diseases, prevalence and incidence are often the same.
Thanks for watching! If you found these videos helpful, please consider supporting me at www.patreon.com/LYMED
Much love, -Mike
thank you! I was able to understand this concept. Life saver, GODBLESS! amazing
You are the most clear among all others. Great job!!!!!
Great video, great explanation. Thank you very much.
You are amazing
Your videos have helped me pass medical school ily
That was so great and easy thanx alot👍
You are great! man💕
Tysm 😁
thank you
Thank you.
I think the positive and negative predictive values here are wrong and divided down the wrong column of your chart. The positive predictive value is the percentage of those with disease who have a positive test so would be:
true positives / (true positives + false positives) which in your HIV case would be 95/96 = 98.9%
Check his chart again. You're correct in what you're saying, I just think you've confused the table. He did it correctly. In his example, true positive = 95, false positive = 5, true negative = 99, false negative = 1. So for PPV = TP/(TP+FP) = 95/(95+5) = 95%.
The numbers you're using would be 95/(95+1), or TP/(TP+FN), which is the sensitivity, not PPV.
Hey great video. I have 2 questions.
1) if the question says that the patients test results came back negative and the patient asks you what are the chances that I do have the disease. What do I have to calculate then?
2) In incidence and prevalence of acute cases, will the dead/curex be counted? Eg. Acute break out of ebola affected 50 people of which 10 died/cured within the year.
1. If the patient has the disease but tests negative, you have a false negative. A sensitive test will have no false negatives. So you calculate sensitivity, then subtract by 1 (1-sensitivity) which is the likelihood the test will incorrectly test a patient with the disease.
2. incidence in the year (assuming all total cases happened that year) is 50. It could be 30 for the year if 20 cases were seen 2 years ago.
Prevalence is a snapshot of who is still with disease and alive. Cured don't have disease so you subtract them from total cases. Dead are no longer alive so you subtract them from total. In this case, 40 would be the prevalence at the end of the year. Another example, if we look at the 6 month mark, maybe only 2 people are cured and 5 have died. If we have seen all 50 cases start by the 6 month mark, we would have 50-2-5=43. Prevalence would be 43 at the 6 month mark. Eventually more people get treated or die, and maybe in 4 years prevalence is 0.