Mental health, climate change and digital psychiatry

How might we study mental health (including the impacts of climate change on mental and emotional wellbeing) and scale up mental health care using mobile apps?

 

Mental health, climate change and digital psychiatry

How might we study and support mental health (especially when impacted by climate change) using digital technologies?

INITIAL CURIOSITY

After volunteering on a suicide prevention hotline and studying climate change for years, I became curious about the relationship between environmental change and mental health. Of course, there's research about how green spaces improve people's mood. There are studies emerging about how natural disasters—such as drought in Australia and flooding in England—cause anxiety and depression. But beyond the "spectacle": what about the "slow violence" that accumulates and afflicts over time?

Slow violence – a violence that occurs gradually and out of sight, a violence of delayed destruction that is dispersed across time and space, an attritional violence typically not viewed as violence at all. It shapes our inattention to calamities that are slow and long lasting, calamities that patiently dispense their devastation while remaining outside our flickering attention spans – and outside the purview of a spectacle-driven corporate media.
— Rob Nixon, "Slow Violence and the Environmentalism of the Poor"

How does the gradual loss of biodiversity, the growing distance to the nearest spring, the single degree of warming or cooling affect people's mental and emotional wellbeing? What are the environmental determinants of mental health? How can people be supported in their emotional responses to climate change in locations where mental health care doesn’t exist?

ACTION

1. I learned that environmental determinants of mental health are inextricable from social, cultural, political, economic and historical factors. I emailed a bunch of people whose work on the intersection between climate change and mental health I admired, and was grateful when Dr. Ashlee Cunsolo-Willox and Dr. Susan Chatwood Skyped with me and when Dr. Helen Berry corresponded via email. Through these conversations, I tried to unpack the environmental determinants of mental health and wellbeing, such as the importance of land for Indigenous communities in the Circumpolar North. When climate change alters ice quality, it affects people's ability to travel and socialize, intergenerational knowledge transfer, food security and more, all of which affect mental health. For example, since time immemorial elders have used seal hunting to teach youth how to be patient, comfortable with silence and attuned to their surroundings. Without the ability to perform this ritual, both elders and youth may be distressed by the prevention of cultural transmission and detachment from the environment.

I was struck not only by the impacts of climate change on Indigenous mental health, but also by meta-narratives regarding climate change in the Arctic. Dr. Kyle Powys Whyte describes the recolonization done by environmental allies through either romanticizing Indigenous lifestyles ("perfect harmony with the environment") or erasing them ("climate change affects us all"), the former promoting colonialism and the latter ignoring the colonial context. I was familiar with issues of race within the environmental movement: green imperialism, fabricated dilemmas about a just transition versus a fast transition (as if the oppression of people is extricable from the exploitation of the planet) and the whiteness of climate warriors. I was reminded of James Moore's argument that climate change is not "anthropogenic," a term that overlooks the fact that climate change was caused by a select few while others powerfully resisted or had no say at all. Moore believes it is actually "capitalogenic." (Naomi Klein also discusses this at length in This Changes Everything: Capitalism versus the Climate.) While often ignored in ecological assessments, the political and economic drivers of climate change are hugely impactful in shaping which populations experience climate change first and worst—and thus who suffers the mental health effects.

I didn't know what to do with all this information besides write something. It became a very long paper for a fairly short Planetary Health assignment (condensed blog post version here). Around this time, I was also accepted into a program that would allow me to design a research project in northeastern Madagascar. I decided to focus on exactly this question: what does environmental change make people feel? 

MentalHealth_determinants

2. Because not enough is known about mental health—such as idioms of distress and drivers of resiliency—in Madagascar, the starting point for understanding climate change impacts on mental health is exploring the nature of mental health itself. Due to gaps in knowledge about attitudes toward or experiences of mental health in northeastern Madagascar, I began my research upstream from my original research question. I turned to the field of psychiatric epidemiology to consider ways of understanding and assessing mental health. After a conversation with Dr. Karestan Koenan, I broke my topic down into idioms of distress, manifestations of mental health problems, impacts of mental health problems, sources of mental distress and resiliency. My conceptualization of mental health indicators were drawn from the Patient Health Questionnaire (-9, -15, -SADS), World Health Organization's Composite International Diagnostic Interview and Generalized Anxiety Disorder-7.

Ultimately, Dr. Chris Golden and I conducted interviews and focus groups in two communities, Marofototra and Antaravato. People described the concept of "miasa loha," or "working of the mind," which seemed to encompass symptoms of depression and anxiety. They confirmed that mental sickness exists and can be seen in people walking around the market place.

Kids have been warned about their future life. The worst is that they might not know anything about rosewood, for example. They might ask, ‘What does the leaf of that tree look like?’ because they’ve never seen it. Maybe they can see it in a book or on the Internet.
— An interviewee, when asked how climate change makes him feel

While we don't have enough material to draw conclusions, it was clear that people experienced negative emotional and somatic responses to climate change and that there was virtually no mental health care available. Madagascar has 0.06 psychiatrists per 100,000 people.

3. Digital mental health interventions are not a substitution for human care, but mobile mental health apps are an option for scaling up mental health care in communities where it doesn’t currently exist. The largest barrier to the field of digital psychiatry is usability, so mobile mental health apps can be designed to best fit the needs, values and lifestyles of people with serious mental illness. While working on a video for the World Bank and World Health Organization's first-ever conference on global mental health care, I learned that mental health will be the biggest burden of disease by 2030. As a society, we’re abysmally underprepared to treat it in general, much less address the specific impacts of climate change. How can mental health care be scaled up worldwide when mental health care is currently unavailable for the majority of the world's population?

In parallel with my preparations for research in Madagascar, I began exploring opportunities for digital diagnoses and interventions for mental health. Over a phone interview, Dr. John Torous at Beth Israel Deaconess Medical Center framed the potential of digital psychiatry. Of course, digital mental health treatments never can or should replace human care. They are, however, useful supplements. Whereas therapy sessions occur once a day or week or month, a mobile mental health app is able to capture more momentary states of mental wellbeing, providing more information about a patient's "illness trajectory" (a concept introduced to me by Dr. Maia Jacobs). For example, an app used by someone with manic-depressive disorder could collect passive data, such as location and number of phone calls made. A doctor could look at the data and extrapolate their mental state if the person didn't leave home for a week and then made 40 phone calls in one day. Other benefits of digital mental health interventions are the lowered barrier for entry, affordability, accessibility in places where mental health care may not otherwise exist and anonymity where there may be a stigma around mental health. On the other hand, the perils of digital psychiatry include apps or wearables that invade people's privacy or don't work, creating situations in which patients and doctors don't know which to trust. The design of mobile mental health apps is particularly important. If someone experiences paranoia as a symptom of their schizophrenia, how can a mobile mental health app help them track their voices without triggering fears that their phone is out to get them? I think participatory design is critical here (and everywhere).

Despite their potential and early evidence suggesting their efficacy, mobile mental health apps suffer from low uptake and low sustained use. In short, people aren't using mobile mental health apps, even if they could benefit from them. The ability for digital psychiatry to scale up mental health care worldwide thus hinges upon an understanding of users: what drives usability and engageability for mobile mental health apps?

While I had hoped to (finally) put my computer science skills to good use and build something, there wouldn't be much of a point if I didn't know how to design a service that would be taken up by people. I set out to explore this usability dilemma, with lots of support from Dr. Torous and Dr. Joe Firth. During a lit review of 40 papers (reduced from an original search return of 925), I synthesized how different studies conceptualize and evaluate "user engagement indicators" in mobile mental health apps, including feasibility, acceptability, usability and engageability. I found that there's no standardized method for evaluating usability and engageability in mobile mental health apps, indicating that no one knows exactly what they are. Studies use a range of qualitative and quantitative criteria to all somehow come to the conclusion that...their app is usable. 40/40 studies (100%!) confirmed the usability of their app despite wide ranges in actual use metrics and feedback from participants.

Our meta-analysis resulted in this first-author, peer-reviewed paper, published in the journal of the American Psychiatric Association.

MentalHealth_usability

4. Passive sensing, such as social media data, can also be used to take the pulse of people’s experiences of certain events — from disease outbreaks to extreme weather events — in aggregate. While working for the International Management Institute, I proposed a project to the Data Science for Social Good Foundation that would analyze Twitter data to characterize people’s experiences of Cyclone Amphan. Cyclone Amphan made landfall in West Bengal, India on May 20, 2020. With wind speeds gusting up to 200 kilometers per hour, it damaged nearly 3 million houses, 18,000 square kilometers of agricultural lands, and 449,000 electric poles, leaving 18 million affected people in its wake. Cyclone Amphan was also the costliest cyclone in the history of the North Indian Ocean.

I liaised between five Solve for Good data scientist volunteers (Ancil Crayton, Joao Fonseca, Kanav Mehra, Jared Ross, and Marcelo Sandoval-Castaneda) and IWMI staff, with a lot of support from Rachel von Gnechten and Simon Langan, to scope and complete this project, which would serve as proof-of-concept for IWMI. Can Twitter content help trace how collective knowledge of extreme weather events is produced, characterize people’s experiences, and/or supplement on-the-ground relief efforts?

The data scientist volunteers then took the lead. We first collected around 470,000 Tweets using the Twitter API. They pre-processed the data and began extracting features, including sentiment analysis, point-of-view extraction, zero-shot text classification, and user and Tweet embeddings. Jared led the development of a web dashboard (currently for internal use only), which enables users to apply filters to search through a database of the Tweets and explore several data visualizations based on our analyses. A work-in-progress paper was accepted by the “Tackling Machine Learning with Climate Change” workshop at NeurIPS 2020 and the “Artificial Intelligence for Social Good” workshop at IJCAI 2020. We’re still completing the analysis, so to be continued…

5. Tbd. Looking forward to exploring further!

NEXT QUESTIONS

I believe that digital psychiatry is an opportunity to improve access to mental health care worldwide, which can also help study and address emotional responses to climate change in particular. But there’s clearly a lot to be done beforehand:

  • When mental health is hyper-localized and culturally context-specific, is it possible for mobile mental health apps to scale? If so, how can they be designed to do so?

  • The effects of climate change are temporally and spatially distributed. How can the idea of “illness trajectories” be applied here, to conceptualize the mental health care that might be necessary to treat people experiencing different effects of climate change at different times?

  • How can my future research on these topics be participatory and community-based?


This details my “body of work” regarding mental health, as it intersects with my environmental studies, foray into user research, and experience as a mental health counselor. Many thanks Samaritans, Room 13, Professor Sean Cubitt, Dr. Ruth Goldstein, Dr. Sam Myers, Dr. Christopher Golden, Dr. Maia Jacobs, Dr. Krzysztof Gajos, Dr. John Torous, Dr. Joe Firth, and Dr. Simon Langan for their support of a meandering path!