Digital Mental Health Innovations in the Face of Climate Change: Navigating a Sustainable Future

With increasing mental health challenges related to existential concerns about climate change, real-time processing of biometric and personal data can have substantial impacts on patient care. For example, AI could support the identification of a mental health risk, provide an early warning, and offer adaptive interventions that provide the needed type and amount of support at the right time. This type of detection and intervention may be particularly beneficial for youths, who are vulnerable to experiencing some of the greatest climate-related mental health burdens. A global study of youths’ climate change concerns noted that 84% of young people reported being at least moderately worried about climate change (2). This approach to individualized care, known as digital phenotyping, has been used to gather data from multiple streams, including social media, smartphones, wearable devices, and electronic health records (4, 6). Social media and smartphone data are rich sources to inform approaches to care for youths.

Of note, young people are seeing more climate change content online, and nearly 70% of Gen Z and 60% of millennial social media users report feeling anxious about the future when viewing online information about climate change (10). Machine learning, deep learning, and natural language processing techniques can be used to identify young people at risk for mental health challenges who share negative emotions or climate-related psychiatric stressors via social media. Passive collection of smartphone data can provide more precise measurements of digital behavior and social signals, such as the times when youths are looking at negative content on social media, changes in social interaction patterns, and activity levels (11). Such data could be combined with ecological momentary assessment data (i.e., repeated measurements of reported stress and well-being) (6), as well as geospatial data indicative of hot spots of potential climate change, thereby creating a digital biomarker of climate-related mental health risk. For example, youths living in areas with more severe weather events may spend more time on their smartphones and may be more likely to engage with climate-related content on social media to validate their lived experience. Digital health platforms that leverage AI algorithms to analyze passive and active data can identify patterns of social media use, mood fluctuations, and stress levels, as well as how these patterns vary among youths who live in areas at high risk for adverse impacts of climate change. In response to the data, just-in-time adaptive mechanisms can then send push notifications with personalized, actionable feedback to promote behavior change (e.g., a suggestion to go for a walk or to spend time in nature to interrupt climate change doomscrolling behaviors, encouragement to connect with peers or to seek support if a severe weather event is approaching, or brief cognitive-behavioral exercises to manage climate anxiety).

Evidence-based interventions that are effective in providing mental health support to disaster-affected communities (e.g., Problem Management Plus [12] or Skills for Life Adjustment and Resilience [13]) could also be integrated into digital technologies to enhance the speed and personalization of care for climate-related distress. One such climate-related digital health intervention is the Text4Hope cognitive-behavioral therapy text messaging intervention. This subscription-based program delivers daily affirming messages at a regular time to encourage mental health and well-being, with proven effectiveness in providing support to individuals during wildfires (7). Integrating AI into this program or similar text-based mental health interventions can enhance their effectiveness by providing real-time support that is tailored to individual needs. For example, rather than having a message sent at the same time each day, push notifications to encourage behavior change could be sent when an individual is engaging in behaviors linked to poor climate-related mental health or before a predicted severe weather event to encourage planning.

AI can also be used to enhance clinical decision making and promote personalized care approaches via electronic health record data. Predictive models could draw on electronic health record data to identify patients with preexisting conditions who live in high-risk areas and who may be more susceptible to climate-related mental health burdens or consequences. In such cases, medical providers could plan to incorporate into patient visits education on the intersection of climate change and mental health, discuss strategies for planning and preparedness, and provide referrals to climate-aware clinicians or support groups as needed. This personalized care approach could also have considerable implications for neonatal and pediatric health, because child learning and cognitive development are particularly vulnerable to climate change impacts, with dire implications for mental and general medical health later in life if these issues are not addressed (1).