Latent class analysis in a transdiagnostic psychiatric population to identify classes of psychosocial functioning:...

Current diagnostic classifications in psychiatry show symptomatic overlap, and considerable symptom diversity in diagnosed individuals. Data-driven clustering techniques may help to clarify this complexity. In a psychiatric transdiagnostic population we aimed to 1) identify transdiagnostic latent classes of psychosocial functioning, and 2) examine whether the classes are differently associated with mental healthcare utilization and functioning at one year follow-up.

Routine healthcare records of 9062 patients (55.8 % female, mean age 38.4 years (sd = 13.4)) from four Dutch mental healthcare institutes were used. Latent Class Analysis (LCA) was applied to identify distinct classes of psychosocial functioning, with the optimal number of classes determined by statistical fit indices, adequacy criteria, and overall interpretability. LCA indicators included the clinician-rated Global Assessment of Functioning (GAF) and the subscales of the patient-reported Outcome Questionnaire-45 (OQ-45).

The best model identified seven classes of psychosocial functioning: Severely impaired (N = 143, 1.6 %), Highly impaired (N = 1816, 20.0 %), Highly impaired with normal social role performance (N = 408, 4.5 %), Moderately impaired (N = 3313, 36.6 %), Slightly impaired (N = 1485, 27.4 %), Non-impaired with anxiety (N = 720, 7.9 %), and Non-impaired (N = 177, 2.0 %). The Highly impaired with normal social role performance class showed elevated impairment-levels on all subscales of functioning, except the social role performance scale. All other classes showed similar elevations on all subscales. Classes differed in distribution of sex-assigned-at-birth, primary diagnoses, and number of psychiatric comorbidities. After one year of care, the classes differed in psychosocial functioning, number of treatment sessions, and days receiving treatment.

Our classification of psychosocial functioning is associated with differences in mental healthcare utilization and prevalence of diagnostic classifications, but no direct association was found between psychosocial functioning and specific diagnostic classifications. The highest functioning classes had the most severe mental disorders, supporting the point of different severity levels in the most severe disorders. Incorporating psychosocial functioning classifications alongside diagnostic classifications could enhance healthcare management.