EBP Review Diagnostic Testing and Measurement Error

define sensitivity

proportion of patients with the target disorder who have a positive test result

-true positive

define specificity

proportion of patients without the target disorder who have a negative test result

-true negative

A positive test and a present condition is considered a ....

True positive

A negative test and a present condition is considered a ....

False negative

A negative test and a absent condition is considered a ....

True negative

A positive test and a absent condition is considered a ....

False positive

how can we read sn?

snOut

-when sensitivity is high, a negative test rules out the disorder

how can we read sp?

SpPin

-when specificity is high, a positive test rules in the disorder

Positive Likelihood Ratio

the likelihood that a given patient has the disorder

-sensitivity / 1 - specificity

what happens when you have a large LR+ IF their is a positive test?

the probability that the patient has the disorder increase markedly

what happens when you have a LR- IF their is a negative test?

the probability that the patient has the disorder becomes much smaller

Negative Likelihood Ratio

the likelihood that a given patient does not have the disorder

- 1 - sensitivity / specificity

whats the purpose of the likelihood ratios?

help indicate how much a given diagnostic test results will increase or decrease the probability of a target disorder

how can we use a nomogram?

to estimate the pre to post test shift in probability

pre-test probability

the likelihood that a patient exhibits a specific disorder before the clinical examination is performed

how can the prevalence rate of the condition be used?

used as an indication of pre-test probability

-needs to be estimated value

what is the meaning of the LR+ >10 and LR- <0.1?

Large and conclusive change from pre-test to post-test probability

what is the meaning of the LR+ 5-9.9 and LR- 0.1-0.29?

Moderate shift from pre-test to post-test probability

what is the meaning of the LR+ 2-4.9 and LR- 0.3-0.59?

Small but sometimes important shift from pre-test to post-test probability

what is the meaning of the LR+ 1-1.9 and LR- 0.6-1.0?

Small and rarely important shift from pre-test to post-test probability

how do we interpret reliability statistics?

With the ICC values

-continuous data (ordinal or interval)

What is the range set for the portney & watkins ICC values?

<.50 poor

.50-.75 moderate

>.75 good

what is the range set for Liporance for the ICC values?

0-0.02 poor

0.3-0.4 fair

0.5-0.6 moderate

0.7-0.8 strong

>0.8 Almost perfect

how do we interpret reliability statistics (kappa)? (the values)

<40% Poor to Fair

40-60% Moderate

>60% Substantial

>80% Excellent

why do we care about what type of data is being collected?

to ensure we can determine the reliability and validity of the data

define arbitrary

based on random choice or personal whim, rather than any reason or system

define standard error of the measure

error associated with a single measure

define minimal detectable change

error associated with the change in score

define minimally clinically important difference

amount of change which represents true change to patient

-value taken from the MDC

reliability equals

ICC

how are self-reported disability questionnaires used?

help eval a patient, objective information from the patient themselves

when do you use self-reported disability questionnaries?

initial evals

re-evals

before physician appointments

discharge