Using an Artificial Intelligence–based Analysis of Online Discussions From People With Schizophrenia to Inform an Understanding of Disease Burden
Introduction: Although the clinician perspective of schizophrenia is commonly reported and patients complete questionnaires about their disease, the spontaneous, subjective experience of people living with schizophrenia is not well understood. An artificial intelligence–based semantic analysis of social media posts was used to gain insights into the subjective experiences of people living with schizophrenia who participate in online communities and to understand how this population perceives disease burden.
Methods: Posts from 23 online health communities across 4 languages (English, Japanese, Chinese, German) were retrieved between August 2020 and July 2024. Experiences of people with schizophrenia were detected using a generative artificial intelligence–based tool and analyzed using Natural Language Processing. Data were collected and processed anonymously and securely.
Results: Overall, 263,046 posts from 8219 unique people were collected. Of 1217/8219 (15%) people whose gender could be identified, there was an even distribution of males (51%) and females (49%). Of 1860/8219 (23%) people whose age could be identified, the most common age group was 20–30 years. People reported information on treatment experiences (n=3447/8219; 42%), symptom burden (n=5251/8219; 64%), and quality of life (QoL) aspects (n=6650/8219; 81%). Recreation and leisure (42%), cognitive capabilities (37%), and sleep and rest (30%) were among the most mentioned QoL aspects.
Conclusions: Semantic analysis of social media posts from online health communities is an innovative method to inform on the subjective experiences of people living with schizophrenia in a real-world setting. Topics frequently discussed by people included treatment experiences, symptom burden, and QoL.
Funding: Boehringer Ingelheim.
