Those sample sizes weren’t particularly impressive so I went looking for more info, and according to this review the relationship between creativity and mental health depends on how creativity is conceptualized:

[…] When creativity is conceptualized or operationalized as dispositional, the association is negative, whereas it is positive when creativity is treated as a strategy (e.g., as an intervention method or regulation activity). Indeed, some studies have provided direct support for this idea. For example, Acar et al. (35) found that the approaches to measure creativity are what account for the variation in the association, with a stronger association occurring when creativity is measured by instruments focusing on creative activity and behavior than by those looking at divergent thinking tasks. Nevertheless, we also acknowledge that some alternative explanations cannot be excluded without rigorously controlled studies. For example, Paek et al. (63) conducted a meta-analysis using 89 studies to examine the overall relationships between the most common psychopathologies and little-c creativity and revealed that the overall mean effect size was not different from zero but varied, with effect sizes ranging from −0.97 to 0.95, and 54% of the total effect sizes being below zero and 44.4% of the total effect sizes being above zero. These results actually confirmed the paradoxical association between creativity and mental health. Additionally, their moderator analyses showed that effect sizes varied by the assessment of both psychopathology and creativity as well as by level of intelligence.

So it looks like creativity as a personality trait is negatively correlated with mental health, while being asked to think creatively or to complete creative tasks leads to a positive correlation.