Partest

The test calculate the performance of a clinical test based on the Bayes theorem
15.9K Downloads
Updated 7 Oct 2024

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)
Answer is:

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!

Versions that use the GitHub default branch cannot be downloaded

Version Published Release Notes
2.0.0.0

inputparser and github link

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

more parameters added

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

Matthews Correlation Coefficient added

To view or report issues in this GitHub add-on, visit the GitHub Repository.
To view or report issues in this GitHub add-on, visit the GitHub Repository.