AI Analysis of Digital Conversations Yields Information on Cultural Beliefs About Mental Illness
Abstract
Advanced digital data extraction techniques can play a key role in efforts to better understand the needs of Hispanic patients and design health care communications and public health programs.
Analysis of open-source digital conversations about mental illness using artificial intelligence can offer useful insights into cultural beliefs about stigma and seeking treatment, said Ruby Castilla-Puentes, M.D., Dr.P.H., M.B.A., at APA’s 2023 Annual Meeting.
“Our findings highlight the need to reinforce mental health disease awareness and education among the Hispanic community,” said Ruby Castilla-Puentes, M.D., D.Ph., M.B.A.
She is affiliated with the Community Leadership Institute at the University of Cincinnati and president and founder of WARMI-Women’s Mental Health. Other presenters were Tatiana Falcone, M.D., M.P.H., a staff physician with the Cleveland Clinic Neurological Institute, and Laura Daniela Jimenez, M.D., a medical graduate of the Universidad Nacional de Colombia.
“Advanced digital data extraction techniques and leading big-data tools can play a key role in the efforts to better understand the needs of Hispanic patients to design engaging health care communications and public health programs,” Castilla-Puentes said.
She described a study using AI-powered algorithms to analyze 543,000 online open-source discussions related to depression, of which Hispanics participated in about 43,000 over a 12-month period ending in January 2019. Most conversations (74%) occurred on message boards and sites related to mental health and depression, 16% occurred on social media, and 10% on blogs. All data about conversations were de-identified. The study was published in the Annals of General Psychiatry in November 2021.
CulturIntel, an AI company, extracted topical tag data and origin- and user-related information to create a large, unstructured dataset through advanced search techniques. “We analyzed 12 months of digital discussions about depression to identify patterns in the topics and barriers that correlate to dealing with depression and seeking help/treatment,” Castilla-Puentes said. The search was limited to conversations originating from U.S. internet protocol addresses and tagged based on self-identified profile descriptions.
The analysis found that people who identified as Hispanic expressed more concerns about social appearances and public opinion than those who did not identify as Hispanic, and stigma-related topics were discussed nearly twice as much. There were nearly twice as many conversations among Hispanic individuals about “living with” depression than among non-Hispanics. Non-Hispanics discussed causes and therapies more than twice as often as Hispanic people.
“Hispanic patients expressed more helplessness- or hopelessness-related online conversations and tend to approach a depression diagnosis with a struggling mindset,” Castilla-Puentes said. “During the pre-diagnosis stage, cultural beliefs—such as faith and fatalism—seem to foster a resigned, helpless mindset, which is reflected more among Hispanics than in the general population. That attitude seems to play a role in preventing patients from seeking help as they would rather live and cope with it than face public shame or labeling.
“Our findings highlight the need to reinforce mental health disease awareness and education among the Hispanic community,” she concluded. “There is a need to better communicate the high prevalence of depression as well as to continue to fight stigma. Culturally aligned public health campaigns and educational materials can contribute to empowering patients by addressing their misconceptions about the disease and promoting the positive impact that family, friends, neighbors, and other key members of the Hispanic community can play in supporting the patient and encourage opportune diagnosis and treatment.” ■
