Bret S. Stetka, MD; Christoph U. Correll, MD


May 21, 2013

In This Article

Field Trial Results

As part of the DSM-5 field trials, 2246 patients with various diagnoses and degrees of comorbidity were interviewed (86% twice) on the basis of the DSM-5 criteria. Interviews were conducted by 279 clinicians in various disciplines who received training similar to what would be available to clinicians after publication of the DSM-5.[2,3]

The field trials tested the criteria for 23 disorders: 15 adult and 8 child/adolescent diagnoses. Reliability as a measure of agreement between 2 independent clinicians was measured with kappa statistics. For example, if 2 clinicians agreed on a diagnosis 85% of the time, the kappa value was 0.46 -- as it was for schizophrenia and conduct disorder. Overall, 5 diagnoses were in the "very good" range (kappa = 0.60-0.79), 9 were in the "good" range (kappa = 0.40-0.59), 6 were in the "questionable" range (kappa = 0.20 0.39), and 3 were unacceptable (kappa < 0.20) (Figures 1 and 2).

Figure 1.

DSM-5 field trials: diagnostic reliability in adults. Modified from Freedman R, et al.[2]

Figure 2.

DSM-5 field trials: diagnostic reliability in children and adolescents aged 6-17 years. Modified from Freedman R, et al.[2]

For 8 diagnoses, including several personality disorders, sample sizes were insufficient to generate precise kappa estimates. Among the 14 most reliable diagnoses were posttraumatic stress disorder, bipolar I disorder, borderline personality disorder, schizoaffective disorder, schizophrenia, bipolar II disorder, and alcohol use disorder in adults, as well as autism spectrum disorder (ASD), attention-deficit/hyperactivity disorder, bipolar I disorder, conduct disorder, and oppositional defiant disorder in youth.

Whereas new DSM-5 entries, such as major neurocognitive disorder, complex somatic symptom disorder, hoarding disorder, binge eating disorder, and ASD were among the most reliable diagnoses, disruptive mood dysregulation disorder was in the "questionable" reliability range, as were the unmodified major depressive disorder (MDD) and generalized anxiety disorder (GAD) diagnoses. The 2 diagnoses that fell into the "unacceptable" reliability category, mixed anxiety-depressive disorder and nonsuicidal self-injury, have been removed or included in Section 3, respectively.

Critics of the field trial results have focused on the fact that the unchanged disorders MDD and GAD were among the 6 disorders with questionable reliability, and that the kappa thresholds for "good" by the DSM-5 field trials were lower than traditional thresholds. Moreover, for conditions where several different sites contributed data, there were considerable variations in reliability.[3]