Related Experiment Videos
Application of Bayes' Theorem in Valuating Depression Tests Performance
Marco Tommasi1, Grazia Ferrara1, Aristide Saggino1
1Department of Psychological, Health and Territorial Sciences, Università degli Studi G. d'Annunzio Chieti e Pescara, Chieti, Italy.
Frontiers in Psychology
|August 8, 2018
Summary
The Bayesian approach offers a more accurate method for diagnosing depression using psychological tests than the frequentist approach. Most common depression scales lack diagnostic accuracy, with the Teate Depression Inventory showing promise.
Area of Science:
- Clinical Psychology
- Psychometrics
- Diagnostic Accuracy
Background:
- Clinical diagnosis validity is crucial for evidence-based practice in psychology.
- Incorrect diagnoses lead to negative consequences for individuals and treatment efficacy.
- Frequentist probability approaches using cutoffs for self-report tests may underestimate diagnostic risks.
Purpose of the Study:
- To evaluate the diagnostic accuracy of common depression self-report scales.
- To compare the frequentist approach with the Bayesian approach for clinical diagnosis.
- To assess the utility of the Teate Depression Inventory (TDI) developed with the IRT procedure.
Main Methods:
- Analyzed published data on sensitivity and specificity for Zung's SDS, HDS, CES-D, and BDI.
- Employed the Bayesian approach, utilizing Bayes' theorem to estimate posterior probabilities of pathology.
- Included the Teate Depression Inventory (TDI), developed using Item Response Theory (IRT).
Main Results:
- The Bayesian approach provides a more robust method for diagnosis than frequentist cutoffs.
- Most widely used depression scales (Zung's SDS, HDS, CES-D, BDI) demonstrated unsatisfactory diagnostic accuracy.
- The Teate Depression Inventory (TDI) showed potential for improved diagnostic accuracy.
Conclusions:
- The Bayesian method is a superior alternative for diagnostic decision-making in clinical psychology.
- Limitations in item selection and subject definition may contribute to poor accuracy in existing scales.
- The TDI's IRT-based development suggests a more effective approach to reducing false positives.