Research paperSymptom severity and disability in psychiatric disorders: The U.S. Collaborative Psychiatric Epidemiology Survey

- Psychiatric symptoms are not interchangeable indicators of severity.
- The severity of psychiatric criteria was estimated and its validity examined.
- The four most disabling DSM disorders. Representative data of US population.
- Symptom severity predicted disability over symptom count in three of four disorders.
- The use of this additional information does not imply extra costs.

BackgroundWhile most psychiatric diagnoses are based on simple counts of symptoms, some symptoms may be sign of a more severe mental syndrome than others. This calls for validated estimates of the relative severity specific symptoms imply within a disorder. We focused on four diagnostic disorders: Manic Episode (ME), Major Depressive Episode (MDE), Post-traumatic Stress Disorder (PTSD) and Generalized Anxiety Disorder (GAD). Symptom-specific severity parameters were estimated, and validated by examining their association with levels of self-reported disability in daily activities over and above the number of symptoms.MethodsData from the cohort study of the U.S. Collaborative Psychiatric Epidemiology Surveys (CPES) was used, which comprises the National Comorbidity Survey Replication, National Survey of American Life, and the National Latino and Asian American Study. The four analytic datasets included respondents who endorsed disorder-specific pre-screening symptoms according to the World Mental Health Survey Initiative's version of the Composite International Diagnostic Interview. Disability was measured using the WHO Disability Assessment Schedule. Item Response Theory and Tobit models were implemented.ResultsFor ME, PTSD, and GAD (not MDE) symptom severity based on psychometric Item Response Theory predicted disability outcomes after adjusting for symptom count. For PTSD, symptom count was not associated with disability.LimitationsThe analytic sample for each psychiatric disorder was based on a pre-selection stemming from index criteria (e.g. sadness or pleasure loss for MDE), which implies that our results are only generalizable to those individuals at risk rather than for the entire population. Additionally, we acknowledge that the use of unidimensional models is only one of the several options to model psychopathological constructs.ConclusionsThe same number of symptoms may be related to different levels of disability, depending on the specific symptoms from which the person suffers. Diagnostic procedures and treatment decisions may benefit from such additional information without extra costs.