I am a korean dentist.your lecture is good for teaching methods of avoiding overdiagnosis ln dental cavity.thank you for excellent lecture which gave me well information.
Sensitivity identifies patients WITH the disease so it picks up people with disease (true positive) better if it has a strong sensitivity. However it can have the faults of picking up false positives because it is too sensitive. SNNOUT rule: If The test is negative then it helps RULE OUT disease. So it helps identify those with disease if positive and rule out disease if negative???
Sensitivity identifies patients WITH the disease so it picks up people with disease (true positive) better if it has a strong sensitivity. However it "MAY" have the faults of picking up false NEGATIVEs (meaning, it failed to pick up on people with disease. So that's bad, because if a pt has an underlying disease like Cancer, and the test showed a false negative, that pt doesn't get proper treatment and in worst case scenario die.) "SNNOUT rule: If The test is negative then it helps RULE OUT disease. So it helps identify those with disease if positive and rule out disease if negative???" --Only if the Sensitivity is high. Example: Test A has 100% sensitivity. You test 100 pts with Disease A with Test A. Everyone in that group will test positive because the sensitivity of Test A is so high, that there's no chance for it fail. Now you use Test A to test 50 pts with no diseases (healthy). All 50 should be negative, meaning they do not have the disease If the sensitivity of the test is lower, lets say 50%... then a pt with a disease may be falsely diagnosed as having no diseases. Then that whole thing of not getting proper treatment occurs. Does that help?
Thank you very much for the explanation. I think the slide on comparison between PPV and NPV has a little correction, the abbreviation for Negatives predictive value was written PPV instead of NPV.
Thank you for your observation. It's amazing how often one can look at an item like this an not see the error, even though it stands out clearly. I do plan to redo this video at sompe point and I'll make sure to fix that. Thanks again. Jim
So Sensitivity and NPV tell you the same thing? You said that "If i have a test that is 95% sensitive, if that test is negative i know with a 95% assurance that i.ve ruled out that disease" And at the NPV you said " NPV is the likelihood of not having the disease when the test is negative "
Thanks for the question Pop Tudor. I think about the difference between predictive values and sensitivity/specificity in terms of who is the most interested in the result. Predictive value is related to the population on which it was developed. It tends to be of most interest to patients if they belong to the population. It gives them the likelihood of having or not having the disease. On the other hand, sensitivity and specificity are of more interest to clinicians because they are important to the diagnostic process. Sensitivity and specificity are independent of the population on which they were developed. That said, while sensitivity and specificity are important for making the diagnosis, predictive values are useful to the clinician for educating the patient about their condition. I hope this helps. Take care, Jim
Dear Jim, I did two different test for detection of a bacteria in the same population. Is it possible to calculate PPV, NPV, Sensitivity and Specificity of each test Separately with out gold standard ? Could you guide me please. Thank you for your consideration and confirmation would be appreciated. Sincerely, Narjes
Narjes, In order to do these calculations, you will need to know the number of false negatives and false positives in your population/sample. Let's say you have 100 subjects, all of which you know have the bacteria. You test all 100 and find that 80 are positive. Because you know that all 100 have the bacteria, you now know that the 20 negatives are false negatives. Likewise, if you have 100 subjects, and you know that none of them have the bacteria and your test is negative for 80, you know that you have 20 false positives. Combining these examples, you have enough information to do the calculations sensitivity, specificity, PPV, & NPV. Now let's say you have a sample and you don't know the how many have the bacteria. You use two separate tests, one of which is highly sensitive (rules out when negative) and one that is highly specific (rules in when positive), and both of those are negative for the sample. You now have a higher assurance of accuracy of the negative results because the highly sensitive test was negative and the highly specific test was NOT positive. I hope you find this answer helpful.
Dear Dr.Cropper. First, Thank you very much for response. It is very nice of you. As i read from other articles for my work, I calculate sensitivity, specificity, PPV & NPV each of two tests form the test three; I mean from the test three I can understand true positive, true negative, false positive and false negative. The test 3 is so expensive, So I want to know is it any way like new formula, that use it instead of the test three ( gold standard)? Thank you in advance, Narjes
Narjes, I am not aware of any formula or process that precludes the use of a gold standard for confirmation when establishing sensitivity, specificity, PPV & NPV. However, not being in your field, I am not well-equipped to advise you on specific issues. I would recommend that you consult somebody in your field who would likely be able to give you better and more specific advice than I can. Maybe somebody will know of a less expensive substance or source to use for testing. Sorry I can't be of more help. Best regards, Jim
Thank you for your question, and sorry for my tardiness. The rotator cuff is a set of muscles at the shoulder that work sometimes together (synergistically) to achieve a certain movement, and sometimes in opposition (antagonistically) in order to stabilize the shoulder. I hope this help. Best regards, Jim
Sorry for my tardiness. I would say yes, total population is fine as long as it is clear in the context that it is for those who were tested for the condition.
It's interesting how biostats also has its own twist on this topic; epi presents it one way, and biostats gives it in another. But both end up boiling down to mean the same thing. Each of them helped me learn it better in the other's "language" though.
