1/17
Looks like no tags are added yet.
What is a “pre-pre analytical error”?
These are errors before samples are even collected, including ordering the wrong test, ordering an unnecessary test, or not ordering a necessary test
What is a “post-post analytical error”?
These are errors that clinicians do with the result, like misreading something, not noticing something, etc.
True positive
Has the disease, the test catches it
False negative
Has the disease, the test misses it
True negative
Doesn’t have the disease, the test says no
False positive
Doesn’t have the disease, test wrongly says yes
How to not get confused by these 4?
Remember the last word is what the test says.
Eg. positive means the test said yes
Then the first word is what makes it true or false
Eg. false positive means test said yes when the patient wasn’t sick
What is sensitivity?
From all the people who have the disease, what fraction does the test correctly flag as positive
A highly sensitive test rarely produces ___
False negatives
What is specificity?
From all the people who don’t have the disease, what fraction does the test correctly call negative
A highly specific test rarely produces ___
False positives
SnNout and SpPin
SnNout: A highly sensitive test, if Negative, rules out the disease
SpPin: A highly specific test, if Positive, rules IN the disease
What is a test’s predictive values?
It can be positive or negative, and it’s used to tell what the patient’s true status is:
PPV: From everyone who’s positive, how many actually have the disease?
NPV: From everyone who’s negative, how many people are actually healthy
Prevalence rules
If prevalence goes up: PPV goes up, NPV goes down
If prevalence goes down: PPV goes down, NPV goes up
To rule a disease out, use a
Sensitive test
To rule in a disease, use a
Specific test
What is pre-test probability?
It’s how likely the patient has a disease before we see a test result, based off of their demographic, risk factors, history and examination
What type of pre-test probability best requires a test?
Intermediate pre-test probability