Psychiatric diagnosis: Are clinicians still necessary?☆

In this narrative review, we survey recent empirical evaluations of AI-based language assessments and present a case for the technology of large language models to be poised for changing standardized psychological assessment. Artificial intelligence has been undergoing a purported “paradigm shift” initiated by new machine learning models, large language models (e.g., BERT, LAMMA, and that behind ChatGPT). These models have led to unprecedented accuracy over most computerized language processing tasks, from web searches to automatic machine translation and question answering, while their dialogue-based forms, like ChatGPT have captured the interest of over a million users. The success of the large language model is mostly attributed to its capability to numerically represent words in their context, long a weakness of previous attempts to automate psychological assessment from language. While potential applications for automated therapy are beginning to be studied on the heels of chatGPT's success, here we present evidence that suggests, with thorough validation of targeted deployment scenarios, that AI's newest technology can move mental health assessment away from rating scales and to instead use how people naturally communicate, in language.

As the Diagnostic and Statistical Manual of Mental Disorders fifth version (DSM-5) was published, the Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime version (K-SADS-PL) was modified to adapt the new version (K-SADS-PL DSM-5). We translated it to Chinese (K-SADS-PL-C DSM-5) and described its reliability and validity.

A total of 154 groups of 6 to 18-year-old children and their guardians were included. Trained interviewers interviewed subjects using the K-SADS-PL-C DSM-5. Interrater reliability was assessed by audio recording. Parent-reported scales, like child behavior checklist (CBCL), the Chinese version of Swan-son Nolan and Pelham, version IV scale-parent form (SNAP-IV), social responsiveness scale (SRS-1), and children-reported scales like depression self-rating scale for children (DSRSC) and the screen for child anxiety related emotional disorders (SCARED) were used to examine the validity of depressive disorder, ADHD, ASD, and ODD.

The K-SADS-PL-C DSM-5 had fair to excellent interrater (0.537–1.000) and test-retest (0.468–0.885) reliability of affective disorder and neurodevelopment disorder. The convergent validity of affective disorder and neurodevelopment disorder was good, and their divergent validity was acceptable.

i) Clinical questionnaires were insensitive in classifying disorders and had limitations in derived diagnoses. ii) Samples only came from clinical environment, iii) covered limited disease species, and iv) were small.

The K-SADS-PL-C DSM-5 can support reliable and valid diagnoses for children with affect, neurodevelopmental, and behavioral disorders in China.