The Spanish language Twitter dataset of 28,268 tweets included 40% (n = 11,336) positive, 22% (n = 6,064) negative, and 38% (n = 10,868) neutral tweets. In contrast, among the 205,774 English tweets analyzed 55% were positive (n = 113,173), 18% (n = 37,059) were negative, and 27% (n = 55,542) were neutral sentiment tweets.

Four overarching themes across the 15 topics emerged among the Spanish-language tweets: awareness, self-care, lived experience, and service providers. See Table 2 for high-probability words and example tweets for each of the 15 topics. The awareness theme was the most prominent and encompassed six topics, including intimate partnerships and mental health (Topic 2), youth mental health (Topic 4), mental health advocacy (Topic 5), World Mental Health Day (Topic 6), welfare advocacy for mental illness (Topic 9), and mental health services advocacy (Topic 10). Within the self-care theme, we observed topics such as self-care and well-being (Topic 1), astrology and well-being (Topic 3), mindfulness self-care (Topic 13), and fostering resilience (Topic 15). The lived experience theme included topics related to referrals to blogs/books (Topic 7), promotion of mental health memoirs (Topic 11), and mental health support groups (Topic 12). Lastly, the service providers theme primarily included information about mental health clinics (Topic 8). Notably, Topic 14 was about animal rights and was deemed irrelevant to our study and therefore excluded from further interpretations.

Figure 1 summarized the results of the regression analysis of all Spanish-language Twitter topics. Positive sentiment was statistically significant for the following topics: Topic 1 self-care and well-being (β = 0.052, p < .001), Topic 5 mental health advocacy (β = 0.005, p < .05), Topic 6 world mental health day (β = 0.068, p < .001), Topic 7 referrals to blogs/books (β = 0.019, p < .001), Topic 8 mental health clinics (β = 0.034, p < .001), Topic 12 mental health support groups (β = 0.011, p < .001), Topic 13 mindfulness self-care (β = 0.074, p < .001), and Topic 15 fostering resilience (β = 0.013, p < .001). Conversely, negative sentiment was statistically significant for Topic 2 intimate partnerships and mental health (β=-0.128, p < .001), Topic 3 astrology and well-being (β=-0.015, p < .001), Topic 4 youth mental health (β=-0.124, p < .001), Topic 9 welfare advocacy for mental illness (β=-0.027, p < .001), and Topic 11 promotion of memoirs (β=-0.013, p < .001). No significant sentiment was observed in Topic 10 mental health services advocacy.

Sentiment-Based Topical Variations in Spanish-Language Twitter Sample, Twitter Academic API Search from 09/19/22 − 10/10/22. NOTE: Every topic is visualized using a horizontal line with a central dot, representing the uncertainty level and coefficient of the sentiment on that topic. When the line does not intersect the central axis (0), it suggests a significant divergence in the topic between positive and negative sentiments. In contrast, if the line intersects the central axis, it implies no substantial difference in the topic between these two sentiments

The English-language tweets comprising the 15 resulting topics were classified into six overarching and distinct themes: awareness, self-care, lived experience, service providers, political activism, and marketing. See Table 3 for high-probability words and example tweets for each of the 15 topics. Within the awareness theme, topics included PTSD and suicide awareness (Topic 6), World Mental Health Day (Topic 7), and comorbidity related to trauma, pain, substance abuse, and cardiovascular disease (Topic 12). The self-care theme encompassed topics such as faith and wellbeing (Topic 3), mental health self-care (Topic 4), mindfulness self-care (Topic 11), workplace mental health (Topic 14), and fostering resilience (Topic 15). Under the lived experience theme, there were topics related to referrals to blogs/books (Topic 8) and the promotion of mental health memoirs (Topic 13). The service providers theme focused on supporting mental health service providers (Topic 2). The political activism theme included topics on mental illness labels for protest (Topic 5) and promotion of white papers (Topic 9). Lastly, the marketing theme comprised topics on mental illness labels for marketing (Topic 1) and complementary and alternative medicine (Topic 10).

Regression analysis revealed that the following Topics exhibited a statistically significant presence of positive emotions tweets (see Fig. 2): Topic 2 supporting mental health service providers (β = 0.019, p < .001), Topic 3 faith and wellbeing (β = 0.070, p < .001), Topic 4 mental health self-care (β = 0.029, p < .001), Topic 7 World Mental Health Day (β = 0.070, p < .001), Topic 8 referrals to blogs/books (β = 0.012, p < .001), Topic 9 promotion of white papers (β = 0.003, p < .001), Topic 10 complementary and alternative medicine (β = 0.020, p < .001), Topic 11 mindfulness self-care (β = 0.065, p < .001), Topic 13 promotion of mental health memoirs (β = 0.007, p < .001), Topic 14 workplace mental health (β = 0.010, p < .001), and Topic 15 fostering resilience (β = 0.016, p < .001). In contrast, the following Topics were significantly associated with negative sentiment tweets (see Fig. 2): Topic 1 mental illness labels for marketing (β=-0.056, p < .001), Topic 5 mental illness labels for protest (β=-0.056, p < .001), Topic 6 PTSD and suicide awareness (β=-0.068, p < .001), and Topic 12 trauma, pain, substance abuse, and cardiovascular disease (β=-0.141, p < .001).

Sentiment-Based Topical Variations in English-Language Twitter Sample, Twitter Academic API Search from 09/19/22 − 10/10/22. NOTE: Every topic is visualized using a horizontal line with a central dot, representing the uncertainty level and coefficient of the sentiment on that topic. When the line does not intersect the central axis (0), it suggests a significant divergence in the topic between positive and negative sentiments. In contrast, if the line intersects the central axis, it implies no substantial difference in the topic between these two sentiments