Growing evidence suggests a concerning decline in the mental health of lesbian, gay, bisexual, and transgender (LGBT) individuals (Grzenda et al., 2021, Kuroki, 2021). While previous research has established a link between sexual orientation and gender identity (SOGI) and poorer mental health outcomes, the evolving significance of this relationship—specifically whether SOGI has been becoming more important in predicting poor mental health—has not been examined. This study aimed to investigate the changing predictive importance of SOGI on mental health outcomes using a large, population-based dataset. Specifically, we sought to quantify the relative importance of SOGI among a broader range of socioeconomic and demographic factors in predicting mental health outcomes and to identify potential shifts in the predictive power of SOGI over time.

To achieve these objectives, we used two popular ensemble methods as well as traditional logistic regression. Ensemble methods combine the predictions of multiple individual models to produce a more accurate and stable result. They are particularly effective at capturing complex, non-linear relationships and interactions among a large number of predictors without the need for manual specification. These methods are increasingly being employed in mental health research (Shatte et al., 2019, Le Glaz et al., 2021, Garriga et al., 2022).

Our findings suggest a significant shift in the factors that contribute to mental health disparities. The rising predictive power of LGBT identity goes beyond merely confirming the existence of a mental health gap; it indicates that LGBT identity is becoming an increasingly central part of an individual’s risk profile for mental distress relative to other well-established factors like income, education, and marital status. From a theoretical standpoint, this finding may be interpreted as a reflection of escalating social and political stressors disproportionately affecting the LGBT community. The increasing political discourse and legislative actions targeting this group over the past decade may have amplified their psychological vulnerability, making their identity a more powerful predictor of mental distress than in previous years.

From a practical and public health perspective, the change in variable importance is a critical signal for policymakers and healthcare professionals. Given the growing predictive power of SOGI, public health efforts should increasingly prioritize resources and interventions tailored to the mental health needs of the LGBT community. Furthermore, clinicians should be increasingly mindful of a patient's SOGI as a key risk factor for mental health challenges. Our findings provide empirical support for the routine collection of SOGI data in clinical settings to better screen and provide targeted care. In sum, the change in variable importance is a tangible reflection of a worsening public health crisis for the LGBT community and underscores the urgency for focused, long-term interventions.