# Partest

Version 2.0.0.0 (23.1 KB) by
The test calculate the performance of a clinical test based on the Bayes theorem
Updated 19 May 2022

Syntax: PARTEST(X,ALPHA)
Input:
X is the following 2x2 matrix. [true positive false positive; false negative true negative];
ALPHA - significance level for confidence intervals (default = 0.05).

Outputs:
- Prevalence
- Sensibility
- Specificity
- False positive and negative proportions
- False discovery and discovery rates
- Youden's Index and Number Needed to Diagnose (NDD)
- Positive predictivity
- Positive Likelihood Ratio
- Negative predictivity
- Negative Likelihood Ratio
- Predictive Summary Index (PSI) and Number Needed to Screen (NNS)
- Test Accuracy
- Mis-classification Rate
- F-Measure
- Test bias
- Error odds ratio
- Diagnostic odds ratio
- Discriminant Power

Example:

x=[731 270;78 1500]
Calling on Matlab the function: partest(x)

DIAGNOSTIC TEST PERFORMANCE PARAMETERS

Prevalence: 31.4% (29.6% - 33.2%)

Sensitivity (probability that test is positive on unhealthy subject): 90.4% (89.1% - 91.5%) False negative proportion: 9.6% (8.5% - 10.9%) False discovery rate: 27.0% (25.3% - 28.7%)

Specificity (probability that test is negative on healthy subject): 84.7% (83.3% - 86.1%) False positive proportion: 15.3% (13.9% - 16.7%) False omission rate: 4.9% (4.2% - 5.9%)

Youden's Index (a perfect test would have a Youden's index of +1): 0.7510 Number Needed to Diagnose (NND): 1.33 Around 14 persons need to be tested to return 10 positive tests for the presence of disease

Precision or Predictivity of positive test (probability that a subject is unhealthy when test is positive): 73.0% (71.3% - 74.7%) Positive Likelihood Ratio: 5.9 (5.7 - 6.2) Moderate increase in possibility of disease presence

Predictivity of negative test (probability that a subject is healthy when test is negative): 95.1% (94.1% - 95.8%) Negative Likelihood Ratio: 0.1138 (0.1094 - 0.1183) Moderate increase in possibility of disease absence

Predictive Summary Index: 0.6808 Number Needed to Screen (NNS): 1.47 Around 15 persons need to be screened to avoid 10 events (i.e. death) for the presence of disease

Accuracy or Potency: 86.5% (85.1% - 87.8%) Mis-classification Rate: 13.5% (12.2% - 14.9%) F-measure: 80.8% (79.2% - 82.3%)

Test bias: 1.2373 (0.9474 - 1.6160) Test overestimates the phenomenon Error odds ratio: 1.6869 (1.2916 - 2.2032) Diagnostic odds ratio: 52.0655 (39.8649 - 68.0002)1.0968 Discriminant Power: 2.2 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.

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

### Cite As

Giuseppe Cardillo (2024). Partest (https://github.com/dnafinder/partest), GitHub. Retrieved .

##### MATLAB Release Compatibility
Created with R2014b
Compatible with any release
##### Platform Compatibility
Windows macOS Linux

### Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Version Published Release Notes
2.0.0.0

1.15.0.0

in the roseplot subfunction I wrote axes instead of axis

1.14.0.0

minor improvement in roseplot

1.13.0.0

change in description

1.12.0.0

1.11.0.0

I synchronized both plots

1.10.0.0

in plots and results false discovery rates were inverted. The bug is fixed

1.9.0.0

There was a problem with the previous upload: the m-file was missing

1.8.0.0

I added the F-measure and the roseplot

1.7.0.0

Changes in description

1.6.0.0

Help section was updated; computations are more efficient and output is more rational. Statistics are fully detailed.

1.5.0.0

Added new 95% confidence intervals for positive and negative predictive values

1.4.0.0

confidence interval for sensibility and specificity added

1.3.0.0

Mistake correction in discrimination power formula (Thank you Prof. Hans Winkler!)

1.2.0.0

Added a plot of the main results

1.1.0.0

Changes in help section

1.0.0.0