Unit 1.3: Clinical evidence (5% of final) Flashcards
What is the purported truth of a test? Is that all to the story?
1) Positive means they have it
Negative means the don’t
2) No; actually. . . There’s much more to the story
What clinical evidence should you consider when deciding if to test?
-Honing your differential with tests
-Considering the “treatment/ test threshold
What is the zone of clinical uncertainty? What do the margins of this zone depend on?
-The area in between where one might consider testing.
-The margins of this zone are dependent upon a few factors:
-How much uncertainty do you need to rectify before you initiate treatment?
-How invasive is the test?
-How serious is the treatment?
-How serious is the illness?
What questions should you consider when deciding the testing/ treatment threshold?
-What is the level of-uncertainty based on history and physical alone?
-Are you almost certain that your patient has the disease in question?
-If so, then why bother testing?
-Move on to treatment
What questions should you consider when deciding how invasive a test is?
-Is this a matter of a simple blood draw?
-Just hooking up an EKG?
-Or does this require sedation and placement of a scope internally?
-Does it involve contrast dye, or other measures with possible side effects?
What questions should you consider when figuring out how serious a treatment is?
-If we begin treatment without a definitive test, what are the possible negative outcomes?
-Are we giving a little ibuprofen OTC?
-Are we initiating lifestyle modifications?
-Or are we initiating life-long immunosuppressants?
What is the final thing you should consider when deciding what is beneath the treatment threshold? Give an example
The severity of the disease
Ex: How low would you make the testing threshold for ruling out MI vs bacterial sinus infection.
In plain terms, you will likely hesitate less when testing for something severe
What are the factors of the treatment/ testing threshold?
1) What is the level of uncertainty based on history and physical alone?
2) Hone your differential with tests
3) How invasive is this test?
4) How serious is the treatment?
5) Severity of the disease
If you’ve determined you’re above the testing threshold, what’s the next step?
We must next consider whether the test we are performing has the ABILITY to rule disease in or out!
How do you determine the efficacy of a test? (5 things)
1) Does the test accurately identify whether a patient has a disease?
2) The gold standard test
3) Sensitivity and Specificity
4) Positive and negative predictive values
5) Prevalence of disease
1) What question should you ask when evaluating test validity? What does this involve?
2) After that question is answered, what’s the next question?
1) Does the test accurately identify whether a patient has a disease?
2) This involves comparing the test against a gold standard.
-The question becomes, how does the test in perform against our best means of establishing true diagnosis
Define gold standard test and give an example
The gold standard test is the modality that best demonstrates accuracy (ex: the full psychiatric eval in pt screened for depression)
What is the significance of the gold standard? What should you compare it to?
-Let the gold standard represent whether the patient has or does not have the disease.
-Consider how the test we are scrutinizing does against this
What does sensitivity of a test mean? What is the formula?
1) The probability that a person with a disease tests positively; aka true positives
2) A/(A+C)
True positives divided by all the folks with the disease
What does specificity of a test mean? What is the formula?
1) Does a person without the dz test negative?
2) D/(D+B)
True negative divided by all the folks without the disease
1) Do very sensitive tests rule diseases in, or out?
2) Why?
1) Out
2) A “sensitive” test has given you a negative result, which means your patient does not have the disease
Why does a sensitive test mean a patient doesn’t have the disease it’s testing for?
-The sensitive test is very SENSITIVE; it’s good at going off anytime someone has the disease
-Therefore, if your patient DID have the disease, the test would have been positive!
-It DIDN’T show up positive. So, your patient is cleared!
Why not just use highly sensitive tests?
1) Because if they are only sensitive, they’re not good at ruling IN.
2) The test is good at being positive. . . But if it’s not also specific then other things besides the disease you’re looking for can make your test positive
1) Do very specific tests rule diseases in, or out?
2) Why?
1) Very specific tests rule IN
2) Your “specific” test has given you a positive result.
This means your patient has this disease.
Why does a specific test mean a patient has the disease it’s testing for?
-These tests are good at being negative.
-So, if your patient was disease free, the test would have been negative.
-It wasn’t negative. So, your patient probably has the disease.
1) Define predictive value
2) Define positive predictive value
3) Define negative
1) Predictive value, in a phrase, determines the percent of the time that a test is “right”
2) Positive predicative value looks at probability that person with + test has the disease
3) Negative predictive value does the opposite
Define positive predictive value in mathematical terms
-PPV = A/(A+B)
-The true positives divided by all the positive tests
-In other words, “How many of these positive tests were right”
Define negative predictive value in mathematical terms
-NPV = D/(D+C)
-True negatives divided by all the negative tests.
-“How many of these negative tests were right?”
What is the catch when it comes to prevalence?
-Seems like a good way of evaluating tests, but there is a skew in the data.
