9. Screening & Diagnostics

A ____ ________ test is an accepted test that is assumed to be able to determine the true disease state of a patient (often expensive or highly invasive)

gold standard

_____ _______ occurs when a test is positive for a disease, but no disease is present

false positive

_____ _________ occurs when a test is negative for a disease, but the disease is present

false negative

____ _________ can lead to unnecessary anxiety or unnecessary treatment

false positives

______ ________ can result in failure to detect early, increasing the health risks to the patient, and potentially others

false negatives

To determine which screening test should be used, we must first consider two properties of the test: ________ and ________

sensitivity

specificity

The ________ of a screening test is the conditional probability that the test is positive given the person actually has the disease

sensitivity

Sensitivity =

P(Test(+) | Disease(+))

High sensitivity for a test is (good or bad)?

good

If sensitivity is high, the false negative rate is ___

low

The probability of a false negative is:

FNR =

1 - sensitivity

1 - P(Test(+) | Disease(+))

P(Test(-) | Disease(+))

The ________ of a screening test is the conditional probability that the test is negative given the person does not have the disease

specificity

Specificity=

P(Test(-) | Disease(-))

High specificity for a test is _____

good

If specificity is high, the false positive rate is ___

low

The probability of a false positive is:

FNR =

1 - specificity

1 - P(Test(-) | Disease(-))

P(Test(+) | Disease(-))

If we wish to minimize the false negative rate, we should select __________

→ minimizes the chance that thyroid cancer is not detected, when in fact it is present

→ Cost: Higher FPR. . . more people who do not have cancer may test positive, causing them to be unnecessarily alarmed and may cause them to have unnecessary surgery

ultrasound

If we wish to minimize the false positive rate, we should select __________________

→ Minimizes the chance that thyroid cancer is diagnosed, when in fact it is not present

→ Cost: Higher FNR. . . more people who do have cancer may test negative, leading to lower chances of successful treatment

fine needle aspiration

While very useful, the sensitivity and specificity of a test still do not tell us:

→ If someone gets a positive result, what is the probability that they actually have the disease?

→ If someone gets a negative test result, what is the probability that actually are disease free?

→ Predictive values converts the test result into the ____ probability of the event

real

_______ __________ ______ is the conditional probability that disease is present given the test is positive

positive predictive value (PPV or PV⁺)

PPV=

P(Disease(+) | Test(+))

_________ __________ ______ is the conditional probability that disease is absent given the test is negative

negative predictive value (NPV or PV^-)

NPV=

NPV = P(Disease(-) | Test(-))

We desire tests with ____ predictive values

high

Predictive values are a ______ criterion for test selection than sensitivity and specificity since:

→ we don’t know a patients true disease status when they are screened, but,

→ we do know the result of the test

better

In practice,

Sensitivity and specificity ___ be estimated directly from data

can

Predictive values often _______ be estimated directly from data

cannot

The predictive values depend on what proportion of all positive tests can be expected to be true positives

→ this depends on three pieces of information:

1) _________ of the test

2) __________ of the test

3) the _____________ of the disease is in the population

sensitivity

specificity

prevelance

The ___________ of a disease is the proportion of a population who have the disease in a given time period

prevalence

The _________ of a disease is the probability that a randomly selected member of a population has the disease

prevalence

___________ is estimated from large-scale prospective epidemiological studies

prevelance

the sensitivity and specificity of the test are measures of overall _____ _______

But, because predictive values are dependent on the prevalence of the disease in the population, they can be thought of as measures of how well the test works in a particular population

test quality

Sometimes, it is possible to estimate the predictive values directly form a table

→ this can only be done when the data were generated from a ____ prospective study

large

Data generated from case-control studies yields _______ estimates of the prevalence of disease because the investigator chooses the number of cases and controls to include in the study

biased

Many screening tests are based on a clinical measurement that can assume a range of values…

→ That is, many medical tests produce _________ numerical data, not just two categories (+ or -)

continuous

The __________ __________ _________ (ROC) curve is a plot of the sensitivity on the y-axis against (1 - specificity) on the x-axis

→ The distinct points on the curve represent different screening cutoff values used to designate a positive test

→ ROC curves provide a way to assess a test that produces continuous numerical data

Receiver Operating Characteristic

The quality of a screening test using a particular marker can be assessed by calculating the _____ _____ ___ ______ (AUC) of the ROC curve

area under the curve

The area under the ROC curve will generally lie between ___ and __

½ and 1

The closer the area is to 1, the ______ the marker for screening

better

Diagnostic tests with

AUC ≥ 0.9

is considered “__________”

excellent

Diagnostic tests with

0.8 ≤ AUC ≤ 0.9

is considered “_____”

good

Diagnostic tests with

0.7 ≤ AUC ≤ 0.8

is considered “____”

fair

Diagnostic tests with 

AUC < 0.7

is considered “____”

poor

Where should we set the screening value for ADA?

→ If the cost of a false positive exceeds the cost of a false negative, then we want to have a more ______ test

specific

Where should we set the screening value for ADA?

→If the cost of a false negative exceeds the cost of a false positive, then we want to have a more ________ test

sensitive

If the costs of false positives and false negatives are equal, we choose the screening value so as to _________ sensitivity and specificity

equalize