PC: Evidence Based Practice Flashcards
(25 cards)
validity
the extent to which a test or measurement actually measures what it is intended to measure
*validity trumps reliability
reliability
the extent to which a test or measurement produces consistent and stable results over time
responsiveness
ability to detect change over time in the measured construct
PT examples of responsiveness
ROM/flexibility
muscle strength
Pain
outcome measures
true positive
a test that correctly identifies the presence of a condition
condition present
false positive
the test that incorrectly identifies the presence of a condition
condition absent
false negative
the test that misses the condition - it is present, but the test fails to detect it
condition present
true negative
test that correctly identifies the absence of a condition
condition absent
sensitivity
SnOUT: good for ruling out
“if someone has the condition, how likely is the test to catch it?”
given that the individual has the condition, probability that test will be positive
true positive / (true positive + false negative)
specificity
SpIN: rule in
“if someone does not have the condition, how likely is the test to say so?”
given that the individual does NOT have the condition, probability that the test will be negative
true negative / (true negative + false positive)
100% sensitivity
the test detects ALL true cases of the condition, but may same some healthy people have the condition (false positives)
100% specificity
if you test positive, you definitely have the condition, but it might miss some people who actually have the condition (false negatives)
what is the difference between a good diagnostic test and a good screening test?
high sensitivity: diagnostic
high specificity: screening
positive predictive value
given a positive test result, the probability that the individual has the condition
true positive/(true positive + false positive)
negative predictive value
given a negative test result, the probability that the individual DOES NOT have the condition
true negative/ (true negative + false negative)
limitations using predictive values
sample specific
depends highly on prevalence of condition in study population
why will the predictive values look like in a condition with low prevalence?
lower positive predictive values → many false positives
higher negative predictive values → few false negatives
likelihood ratios
combine sensitivity and specificity values to tell you how much a test result changes the probability of the disease
positive likelihood ratio
given a positive test result → increase in odds favoring the condition
the increased LR+ → the more certain the individual has the condition (rule in)
sensitivity/(1- specificity)
negative likelihood ratio
given a negative test result → decrease in odds favoring the condition
the decreased LR - (close to 0) → the odds that the individual has the condition is LESS (rule out)
(1-sensitivity)/specificity
interpreting likelihood ratios: Positive LR
LR+ > 10 Large evidence to rule in disease
LR+ 5–10 Moderate evidence to rule in disease
LR+ 2–5 Small but sometimes meaningful increase
LR+ 1-2 No diagnostic value
interpreting likelihood ratios: negative LR
LR- < 0.1 Large evidence to rule out disease
LR- < 0.1-0.2 Moderate evidence to rule in disease
LR- < 0.2-0.5 Small but sometimes meaningful increase
LR- < 0.5-1 No diagnostic value
minimal detectable change
“are you better than the error”
statistic used to represent amount of change needed to exceed measurement error of the test
- reliability measure of change
increase the reliability of the test → _____ MDC value in that population
decrease
*MDC values differ between different populations
