I have a general question about sensitivity and specificity (and by extension PPV and NPV).
So lets say a test has a sensitivity of 85%, and I get that test and it's negative...If I repeated that test a second time, and it was also negative, does that improve my negative predictive value?
For example, lets say I'm concerned about ovarian torsion. I order a pelvic US and it's negative, but I'm still concerned that the patient has an ovarian torsion. From previous EmRap episodes, I know the sensitivity of this study is not the greatest. Lets also assume that the patient does not want to stay in the hospital overnight. If I repeat the US, does that help me in any significant way?
I guess the question can be reduced to the idea are there intrinsic features of the patient (or pathology) that limit the effectiveness of the test, or, is it pure chance and thus two negative tests are additive and improve the NPV (as compared to a single negative test).
Jeff - very complicated and I'm not a stats expert but I'll try to explain as best I can. We really need both sensitivity and specificity to know how the result affects our thinking. Additionally, we need to have made a pretrest probability (your gestalt on how likely the disease is). We can derive likelihood ratios if we have sensitivity and specificity which are more useful in talking about how a test affects the patients risk of disease. If the test has a strong negative likelihood ratio and it's negative, it significantly changes your post-test probability. Depending on how high your pretest probability was, this may "rule-out" the disease (actually just makes it very very very unlikely). As far as doing the same test twice to further decrease your probability of disease, I'm not sure. If the test (US in this case) is done twice by the same person, not sure the second one adds anything unless the clinical picture has changed. If the patient's symptoms worsen, another diagnostic test may be very helpful. One thing that isn't well worked into test characteristics is stage of disease. The sensitivity and specificity of CT for appy is probably near 100% when the patient has significant symptoms (later in disease). Early on, the CT may be negative because the findings are subtle. Hopefully, this helps a bit. I would also recommend going back to 2013 and listening to the mini-Journal Club segments on sensitivity, specificity and likelihood ratios if you haven't listened to them before.
Can anyone at EMRAP point me to a good video or images showing ear bolster dressing technique? I've had a few of these and I'd like to get better at suturing in the dressing but can't find a whole lot of useful videos showing proper technique. Thanks!
Hi Jason, The first stop I would make is a procedure textbook.
Next stop is EMRAP HD on how to do an auricular block https://www.youtube.com/watch?v=6ZiB_9eNpcA
As far as suturing the ear bolster, here is one from Brown Emergency Medicine. It's a little different than I usually do (which is wrap dental rolls in petroleum gauze and put on each side of the ear, this one uses exclusively petrol gauze). https://www.youtube.com/watch?v=sIlm7vPs3q8
We make a lot of fuss about the exact way to do this, but to prove the point that compression is compression, here is an example of using coat buttons. I had never heard of this, but I think demonstrates the principles of compression in a simple fashion https://www.youtube.com/watch?v=hENwObuw_iM
This is a video of an MMA guy getting his auricular compression dressing removed. It's the end product but you can see another variation of how it's done. You'll notice that the dressing is much smaller than in the video from the Brown website. There is no best answer to this. One question that comes up is to suture through the dressing or loop around it. Who knows. My preference is to suture around it but that's just how I was taught. https://www.youtube.com/watch?v=0DLzSSPrdGg
Jeff S. - August 3, 2016 10:17 PM
I have a general question about sensitivity and specificity (and by extension PPV and NPV).
So lets say a test has a sensitivity of 85%, and I get that test and it's negative...If I repeated that test a second time, and it was also negative, does that improve my negative predictive value?
For example, lets say I'm concerned about ovarian torsion. I order a pelvic US and it's negative, but I'm still concerned that the patient has an ovarian torsion. From previous EmRap episodes, I know the sensitivity of this study is not the greatest. Lets also assume that the patient does not want to stay in the hospital overnight. If I repeat the US, does that help me in any significant way?
I guess the question can be reduced to the idea are there intrinsic features of the patient (or pathology) that limit the effectiveness of the test, or, is it pure chance and thus two negative tests are additive and improve the NPV (as compared to a single negative test).
Thanks.
Anand S. - August 4, 2016 7:11 AM
Jeff - very complicated and I'm not a stats expert but I'll try to explain as best I can.
We really need both sensitivity and specificity to know how the result affects our thinking. Additionally, we need to have made a pretrest probability (your gestalt on how likely the disease is). We can derive likelihood ratios if we have sensitivity and specificity which are more useful in talking about how a test affects the patients risk of disease. If the test has a strong negative likelihood ratio and it's negative, it significantly changes your post-test probability. Depending on how high your pretest probability was, this may "rule-out" the disease (actually just makes it very very very unlikely).
As far as doing the same test twice to further decrease your probability of disease, I'm not sure. If the test (US in this case) is done twice by the same person, not sure the second one adds anything unless the clinical picture has changed. If the patient's symptoms worsen, another diagnostic test may be very helpful. One thing that isn't well worked into test characteristics is stage of disease. The sensitivity and specificity of CT for appy is probably near 100% when the patient has significant symptoms (later in disease). Early on, the CT may be negative because the findings are subtle.
Hopefully, this helps a bit. I would also recommend going back to 2013 and listening to the mini-Journal Club segments on sensitivity, specificity and likelihood ratios if you haven't listened to them before.
Jason M. - October 12, 2016 8:57 AM
Can anyone at EMRAP point me to a good video or images showing ear bolster dressing technique? I've had a few of these and I'd like to get better at suturing in the dressing but can't find a whole lot of useful videos showing proper technique. Thanks!
Rob O - October 12, 2016 9:20 AM
Hi Jason,
The first stop I would make is a procedure textbook.
Next stop is EMRAP HD on how to do an auricular block
https://www.youtube.com/watch?v=6ZiB_9eNpcA
As far as suturing the ear bolster, here is one from Brown Emergency Medicine. It's a little different than I usually do (which is wrap dental rolls in petroleum gauze and put on each side of the ear, this one uses exclusively petrol gauze). https://www.youtube.com/watch?v=sIlm7vPs3q8
We make a lot of fuss about the exact way to do this, but to prove the point that compression is compression, here is an example of using coat buttons. I had never heard of this, but I think demonstrates the principles of compression in a simple fashion
https://www.youtube.com/watch?v=hENwObuw_iM
This is a video of an MMA guy getting his auricular compression dressing removed. It's the end product but you can see another variation of how it's done. You'll notice that the dressing is much smaller than in the video from the Brown website. There is no best answer to this. One question that comes up is to suture through the dressing or loop around it. Who knows. My preference is to suture around it but that's just how I was taught.
https://www.youtube.com/watch?v=0DLzSSPrdGg
Jason M. - October 13, 2016 6:30 PM
Thanks Rob. Very Helpful!