JMIR Publications

ABSTRACT

Background:

Neurodevelopmental disorders, especially ADHD and ASD, have seen a marked rise in public attention, yet research on public opinion remains limited. Social media analysis offers real-time, unfiltered insight into public perceptions, enabling empirical examination of public attitudes and opinions.

Objective:

This study aims to assess the evolution of public opinion on ADHD and ASD over the past decade by analyzing tweets from X (formerly Twitter; X Corp, San Francisco, CA), comparing perceptions across English and Spanish languages and against other mental health conditions.

Methods:

Tweets mentioning keywords related to ADHD and ASD, and control conditions (depression, anxiety, insomnia, bipolar disorder, schizophrenia, suicide, and substance use disorders) were collected from X between 2009 and 2023. The dataset included English and Spanish tweets. Machine learning algorithms were then applied to classify tweet content into predefined categories, including volume of tweets, engagement, personal experiences, trivialization, perceived causes, and perceived treatability. Parametric and nonparametric tests were used to assess for differences by language. Descriptive statistics were presented using tables and graphical representations.

Results:

A total of 852,990 tweets were analyzed, with 59.97% in English and 40.03% in Spanish. Overall tweet volume on mental health conditions increased significantly over time. In English, tweets about ADHD (18.98%) and ASD (14.59%) were among the most frequent, while in Spanish, ASD tweets accounted for 14.49%, significantly outnumbering ADHD tweets (5.34%). Engagement analysis indicated a notable increase in likes and retweets per tweet over time, particularly post-2019, with ADHD-related tweets in English experiencing peaks during the COVID-19 pandemic. ASD tweets, however, had comparatively lower engagement across languages. Tweets sharing personal experiences were more polarized in Spanish, with higher proportions of negative and positive experiences compared to mostly neutral English tweets. Trivialization of mental illnesses was less common in Spanish tweets than in English, particularly for ADHD (93.59%) and ASD (84.73%). User-perceived causes included multifactorial factors, biological/genetic factors, substance use, psychological susceptibility, acute psychosocial stressors, and COVID-19. Perceived treatability varied by language, but consistently included high perceived incurability, limited improvement despite professional help, and low perceived self-manageability except for anxiety.

Conclusions:

Analysis of social media discourse showed ADHD attracted significantly higher tweet volumes, particularly during COVID-19, often described with multifactorial causes including substance use and genetics. ASD consistently received lower attention. Both language groups showed low trivialization, awareness of the chronicity of the illness, and limited support for self-management of mental health conditions. These findings underscore social media’s value for capturing direct public perceptions to guide future educational and intervention efforts.