Who Are People with Psychosis Delusional about? A Study of Social Agents in the Phenomenology of Delusions.

Delusions often involve strong beliefs about illusory social agents, yet the nature and identity of these agents in delusions have been understudied. An analysis of 205 electronic mental health records found that 83.4% of delusions explicitly referenced illusory social agents. Across 238 instances, 220 distinct agent identities were identified, averaging 1.17 agents per record. Most agents were humans (85.1%), commonly family members (31.0%), followed by acquaintances (17.2%), religious figures (13.2%), unnamed persons (12.8%), professionals (11.8%), and cultural figures (10.9%). Hierarchical clustering revealed two groups: socially proximate and socially distant agents. The findings indicate a social gradient in delusions, with socially closer individuals appearing more frequently. This gradient may be a general feature of delusions, not limited to specific syndromes.

Study at a glance

Abstract

Delusions frequently involve strong beliefs about, or interactions with, illusory social agents. Although such agents have been systematically described in hallucinations, few studies have investigated their nature and identity in delusions. We identified 205 electronic mental health records describing the content of delusions from the Clinical Record Interactive Search platform at South London and Maudsley NHS Foundation Trust. Delusional content was classified as non-social, implicitly social, or explicitly social. Descriptions of illusory social agents from explicitly social delusions were extracted and categorized using an upward coding analysis. A hierarchical cluster analysis identified distinct groupings of illusory social agents. Most delusions (83.4%) contained explicit references to illusory social agents. Across 238 instances, we identified 220 distinct agent identities, with a mean of 1.17 agents per record. The majority were humans (85.1%), most often identified as family members (31.0%), followed by acquaintances (17.2%), religious figures (13.2%), unnamed persons (12.8%), professionals (11.8%), and cultural figures (10.9%). Hierarchical clustering revealed two distinct groups: one including socially proximate agents and the other more socially distant agents. These findings indicate a social gradient in delusions. Social content predominates in delusions, with patients experiencing delusions about socially closer individuals more frequently. We discuss the extent to which social gradient may not be unique to delusional misidentification syndromes but may instead represent a general feature of delusions. Better characterization of social agents in delusions could inform mechanistic accounts of these symptoms and, in clinical practice, guide family support and risk assessment.