Testing for Diagnosis and Screening Flashcards
Sensitivity and specificity
Sens - percentage of people who do have a disease that test positive
Specificity - percentage of people that DO NOT have the disease that test positive
Screening and confirmatory tests
Screen - maximize sensitivity
COngirm - maximize specificity
Advantages of sequential testing
Lowers overall sensitivity and raises specificity
How does prevalence change
Higher prevalence - means great PPV
Lower prevlaence - means lower PPV and higher NPV
Likelihood ratio
Used to asses value of a given diagnostic test
Positive and negativel ikleihood ratio
Sensitivity/1-specifcity for positive
1-sensitivity/specificity for negative likelihood
Can use Fagan nomogram
Strong moderate and weak
Also unhelpful
Strong - 10-infinity or .1 to 0
Mod - 5-10 or .1 to .2
Weak - 2 to 5 or .2 to .5
.5-2 is generally unhelpful
Diseases that can be screened
Important health problems (high incidence or mortalitiy)
Tx more bebenficial in early stages compared to later
Early dx impoves QOL or survival
Appropriate screening tests
Simple, safe, and precise
Can detect the latent or early sx stage
Beneift outweight physical or psych harm
Screening benefits
Disease incidence or mortality
NNS
Increase in life expectancy (standard is over 1 month gain)…avoid screening if life expectancy less than 10 years
Screening harms
Adverse effects of screening
Overdx of disease
False positive screening tests
Screening program rationale
Cost is balanced in relation to expenditures
Systematic plan for monitoring program
Participant edu about benefits and disadvantages
Simultaenous testing
Screening test emphasis on sensitivity and sacrifice PPV
Cotesting will raise sensitivity and lower specificty
Overdiagnosis bias
Makes it look like mortality lower because we are including people who were asymptomatic in the orginal group…you find the true positives who are asymptomatic
LEad time bias
Outcome is the same but it looks like the survived with the disease longer because the diagnosis was made earlier in their life
Could artificially inflate the benefit of the results
Length bias
Due to different types of diseases
Will capture more slowly progressive cases than rapidly progresing
All ture positives and negatives
Maximizes sensitivity while minimizing specificty
Clinically meaningful ositives and negatives
Minimizes sensitivity while maximizing specificty
RCTs vs cohorts
Cohorts - healthy used
RCTs - outcome should be cancer mortality in the overall population and NOT survival or cancer case fatality rate
