The Puzzling Unidimensionality of DSM-5 Substance Use Disorder Diagnoses

Area of Science:

• Psychiatry
• Psychometrics
• Computational Statistics

Background:

• Ongoing debate exists regarding criteria for diagnosing substance use dependence in the DSM.
• Analysts frequently report evidence supporting the unidimensionality of the diagnostic checklist.

Purpose of the Study:

• To challenge the assumption of unidimensionality in DSM substance use disorder diagnoses.
• To propose alternative measurement models for a more scientifically grounded diagnostic system.

Main Methods:

• Critique of traditional reflexive measurement models for DSM diagnoses.
• Advocacy for formative measurement models to represent the checklist creation process.
• Recommendation of network or causal models for diagnostic system development.

Main Results:

• The reported unidimensionality of the DSM substance use disorder checklist is likely spurious.
• Traditional reflexive measurement models are ill-suited for characterizing DSM diagnostic criteria.

Conclusions:

• A formative measurement model better reflects the institutional process of DSM checklist creation.
• Network or causal models offer a more scientifically robust foundation for future diagnostic systems in substance use disorders.