nan_mahal Classification nan_rocplot

NaN Toolbox >> NaN Toolbox > Classification > nan_partest

nan_partest

This function calculate the performance, based on Bayes theorem, of a

Calling Sequence

nan_partest(X)

Description

Input: X is the following 2x2 matrix.

....................Affected(D+)..Healthy(D-) _______________________ Positive Test(T+) | True | False | | positives | positives | |___________|___________| | False | True | Negative Test(T-) | negatives | negatives | |___________|___________|

Outputs: - Prevalence of disease - Test Sensibility with 95% confidence interval - Test Specificity with 95% confidence interval - False positive and negative proportions - Youden's Index - Test Accuracy - Mis-classification Rate - Positive predictivity with 95% confidence interval - Positive Likelihood Ratio - Negative predictivity with 95% confidence interval - Negative Likelihood Ratio - Error odds ratio - Diagnostic odds ratio - Discriminant Power - Test bias - Number needed to Diagnose (NDD)

Examples

X=[80 3; 5 20];

nan_partest(X)

Answer is:

Prevalence: 78.7%

Sensitivity (probability that test is positive on unhealthy subject): 94.1%
95% confidence interval: 89.1% - 99.1%
False negative proportion: 5.9%

Specificity (probability that test is negative on healthy subject): 87.0%
95% confidence interval: 73.2% - 100.0%
False positive proportion: 13.0%

Youden's Index (a perfect test would have a Youden index of +1): 0.8107

Accuracy or Potency: 92.6%
Mis-classification Rate: 7.4%

Predictivity of positive test (probability that a subject is unhealthy when test is positive): 96.4%
95% confidence interval: 92.4% - 100.0%
Positive Likelihood Ratio: 7.2
Moderate increase in possibility of disease presence

Predictivity of negative test (probability that a subject is healthy when test is negative): 80.0%
95% confidence interval: 64.3% - 95.7%
Negative Likelihood Ratio: 0.1
Large (often conclusive) increase in possibility of disease absence

Error odds ratio: 2.4000
Diagnostic odds ratio: 106.6667
Discriminant Power: 2.6
A test with a discriminant value of 1 is not effective in discriminating between affected and unaffected individuals.
A test with a discriminant value of 3 is effective in discriminating between affected and unaffected individuals.
Test bias: 0.9765
Test underestimates the phenomenon
Number needed to Diagnose (NDD): 1.2

Created by Giuseppe Cardillo
giuseppe.cardillo-edta@poste.it

To cite this file, this would be an appropriate format:
Cardillo G. (2006). Clinical test performance: the performance of a
clinical test based on the Bayes theorem.
http://www.mathworks.com/matlabcentral/fileexchange/12705
nan_mahal Classification nan_rocplot