Sensitivity: Compute sensitivity, specificity and predictive values

In ModelGood: Validation of risk prediction models

Description

Compute sensitivity, specificity and predictive values

Usage

Sensitivity(x,event,cutoff,comparison=">=",...)
       Specificity(x,event,cutoff,comparison=">=",...)
       NPV(x,event,cutoff,comparison=">=",...)
       PPV(x,event,cutoff,comparison=">=",...)

Arguments

x	Either a binary 0,1 variable, or a numeric marker which is cut into binary.
event	Binary response variable. Either a 0,1 variable where 1 means 'event', or a factor where the second level means 'event'.
cutoff	When x is a numeric marker, it is compared to this cutoff to obtain a binary test.
comparison	How x is to be compared to the cutoff value
...	passed on to binom.test

Details

Confidence intervals are obtained with binom.test

Value

list with Sensitivity, Specificity, NPV, PPV and confidence interval

Author(s)

Thomas A. Gerds <[e-mail]>

See Also

binom.test

Examples

set.seed(17)
x <- rnorm(10)
y <- rbinom(10,1,0.4)
Sensitivity(x,y,0.3)
Specificity(x,y,0.3)
PPV(x,y,0.3)
NPV(x,y,0.3)

Diagnose(x,y,0.3)

Example output

Sensitivity: 16.7 (CI_95:[0.4,64.1])
Specificity: 25 (CI_95:[0.6,80.6])
Positive predictive value: 25 (CI_95:[0.6,80.6])
Negative predictive value: 16.7 (CI_95:[0.4,64.1])
2x2 table:
    event
test 0 1
   0 1 5
   1 3 1

Diagnostic parameters:
     Parameter   Estimate CI.95     
[1,] Sensitivity 16.7     [0.4,64.1]
[2,] Specificity 25       [0.6,80.6]
[3,] PPV         25       [0.6,80.6]
[4,] NPV         16.7     [0.4,64.1]

ModelGood documentation built on May 2, 2019, 5 p.m.