WE BUILT AN ECONOMY
OF DISTRESS.
Roughly 300 million people live with depression (GBD 2021). One in four adults reports frequent loneliness (Meta-Gallup 2023). American adolescent girls went from 36% reporting persistent sadness in 2011 to 57% in 2021 (CDC YRBSS), the largest single-cohort shift in the survey's history. The platforms that captured those years were not broken when this happened. They were working.
Six figures behind the composite.
One 0-100 score.
Heuristic seed snapshot. Mental-health pressure has been rising for a decade, with a sharp adolescent inflection beginning ~2012 (smartphone + social-media saturation), an additional pandemic shock, and a treatment-access gap that is widest exactly where prevalence is rising fastest. The headline number flattens dramatic differences by age, gender, and region.
Over time.
Persistent sadness, broken out by sex.
CDC Youth Risk Behavior Survey: share of US high schoolers reporting persistent sadness or hopelessness for two+ weeks straight in the past year. The 2011-2021 jump in girls is the largest cohort-level change in the survey's history.
What the score is measuring.
Several traditions reading the same data.
Why are mental-health indicators getting worse, and what does that even mean?
Mood, anxiety, and stress disorders involve genuine neurobiological dysregulation that can be measured, diagnosed, and treated. Rising prevalence is partly real (driven by sleep loss, screen exposure, social-fabric decline, economic precarity, post-pandemic shock) and partly increased detection in populations previously under-served.
Much of what we call 'mental illness' is a normal human response to abnormal social conditions: precarity, loneliness, meaning-loss, oppression, climate dread. Reframing it as individual pathology risks medicating the symptoms while leaving the causes untouched and the pharmaceutical industry over-paid.
“It is no measure of health to be well adjusted to a profoundly sick society.”
The simultaneous, multi-country adolescent mental-health collapse beginning ~2012 closely tracks smartphone + front-facing-camera + algorithmic-feed adoption. Mechanisms: sleep loss, social comparison, fragmented attention, displaced in-person time, exposure to harassment. The case is correlational but increasingly difficult to explain otherwise.
Suffering arises predictably from attachment and from a distracted, reactive mind; what we are watching globally is the predictable outcome of an attention economy designed to maximise both. The interventions that work (present-moment attention, equanimity, compassion practice) are individually low-cost but require cultural support to scale.
Mental health is inseparable from belonging: to land, kin, ancestors, the more-than-human world. Many Indigenous frameworks treat individualised psychiatric categories as a category error; the illness is in the network, not in the person. Recovery requires repair of relationship, not (only) treatment of symptoms.
Many traditions identify the modern condition not as 'mental illness' but as spiritual emptiness: disconnection from God, meaning, ritual, community of obligation. Religious practice, prayer, fasting, and pastoral care have measurable mental-health benefits in observational data; the effect is mediated by community as much as by belief.
From people who have been inside the systems: what helps is consistent relationships, agency, housing, work that has meaning, and time. What hurts is forced treatment, fragmented short-stay care, stigma, and being talked about rather than to. Policy that excludes this perspective tends to be policy that does not work.
What actually moves the indicators, and what is wishful?
Cognitive-behavioural therapies, SSRIs, exercise, and sleep are the best-evidenced individual-level interventions. At population scale, what moves the needle most is workforce expansion, primary-care integration, school programmes, and reducing the access gap in low- and middle-income countries.
Practical, achievable, and increasingly tested at population scale: no smartphones before high school, no social media before 16, phone-free schools, far more unsupervised play. Several jurisdictions (Australia, France, Italy) have adopted versions; early results on adolescent wellbeing are promising.
Mental-health metrics will not improve at scale without addressing the conditions producing distress: housing insecurity, hours-worked, loneliness, climate dread, inequality. Therapy and medication remain necessary but cannot substitute for repair of the social environment.
Recovery rates from severe mental illness are substantially higher in low-income communities with intact extended-family networks than in high-income individualised settings, against the predictions of biomedical models. The protective factor is sustained relational embedding.
Sources, weights, and code are open.
Where every number comes from
The composite index is computed from the signals listed on this page, each backed by one or more named sources. Where the source publishes a public dataset or feed it is linked below; where a signal involves qualitative judgement, the LLM-assisted pass is explicitly marked on the signal card.
- ·CDC BRFSS
- ·CDC YRBSS
- ·Common Sense Media
- ·Data Reportal
- ·Gallup State of the Global Workplace
- ·Jonathan Haidt After Babel research review
- ·Lancet Global Burden of Disease
- ·Lancet Global Mental Health
- ·Meta-Gallup State of Social Connections
- ·OECD
- ·OECD Better Life Index
- ·OECD PISA wellbeing
- ·Our World in Data
- ·Pew Research
- ·WHO Mental Health Atlas
- ·WHO Mortality Database
- ·World Values Survey
Everything is versioned
- → Every hourly snapshot is committed to git with a message naming the signals that moved.
- → A daily snapshot is archived to
data/history-current/for the calibration log. - → Raw scraped article lists are written to
data/raw/so a score is reproducible from its input bundle. - → Signal definitions, weights, and seeded scores all live in plain JSON or TypeScript; anyone can open a PR challenging a value and explain why.
How this pillar is scored.
Methodology & limits
Ten signals, weighted into a single 0-100 score. The number says how much pressure is on the average human mind right now. It does not say who specifically will be harmed; that's a different question and a different research literature.
GBD and the WHO Mental Health Atlas update yearly, so most of this only moves at that cadence. Faster signals (platform changes, the under-16 social-media bans, post-pandemic survey waves) can shift on a monthly horizon.
Mental health is unusually exposed to how we define it. 'Anxiety' and 'depression' have moved across DSM revisions, and prevalence partly tracks detection. The Perspectives section keeps critical-psychiatry and lived-experience readings visible rather than smoothing them out.