The Hidden Psychometric Model Underlying the DSM: What Happens When It Fails to Fit Psychiatric Data

What is it about?

Diagnostic criteria for a DSM psychiatric diagnosis are typically met by a (within-criterion) symptom count, which presumes each symptom has equal weight in reflecting underlying disorder severity. Conversely, advanced forms of factor analysis and item response theory models have been useful tools in reflecting and demonstrating how underlying psychiatric severity can be estimated in a manner that reflects differences in the relative weights of symptoms. While "symptom-weighted" scores are often highly correlated with symptom counts, graphically we show that patients with the same number of PTSD symptoms can have wide variation in their underlying severities, and their estimated PTSD severity depends on the combination of symptoms. Further, when diagnostic classification takes into account the relative weight of symptoms, many patients without a DSM diagnosis of PTSD show greater psychiatric severity than patients with a diagnosis.

Featured Image