Biostats-4 Flashcards
What is the Positive Likelihood ratio?
- Formula?
- what indicates high dx value?
Likelihood of Seeing this symptom in an affected patient divided by likelihood of seeing this symptom in any other patient
Formula:
Specificity/(1-sensitivity)
More than 10 is high Dx value (aka it changes the pretest probability)
What is the Negative Likelihood Ratio?
- Formula?
Likelihood of Not Seeing this symptom in an affected patient divided by likelihood of not seeing this symptom in any other test
Formula:
1-Sensitivity/Specificity
Less than 0.1 is high Dx value (aka it changes the pretest probability)
Why are LRs useful?
- Summarize Sensitivity and Specificity in a single #
- Not affected by the prevalence of disease in a pop.
- Can be used at multiple levels of Test Results
- Useful for describing the overall probability of disease in serial testing
You can combine the pretest probability determined by the patients characteristics with the likelihood ratio to come up with a post-test probability.
You can combine the pretest probability determined by the patients characteristics with the likelihood ratio to come up with a post-test probability.
What is lead time bias?
- Occurs because screening tests detect illness at ealier points in time
For example, it may appear that people with lung cancer live longer now then they did 20 years ago, but maybe we are just diagnosing lung cancer sooner so it appears that they’re living longer when really that were just in an earlier state of disease
What is length biased sampling?
- Diseases vary in how they carry out their natural course.
- If we detect a disease that remains asymptomatic for a long time then rapidly progresses it may appear that screening for this disease is effective allowing earlier intervention. However, other people who aren’t screen may have the same outcome regardless of treatment because we don’t know how long they’ve had the disease before they become symptomatic.
