Crisis data reveals the mental health impact of COVID-19

Executive overview

Mental health crises don't lag physical ones — they mirror and then amplify them. Crisis Text Line's real-time data exposed two distinct waves: first a surge in anxiety tracking COVID hotspots, then a second wave of grief, domestic violence, and substance abuse as quarantine set in.

Real-time crisis data is a public health asset that policy-makers are missing.

Nancy Lublin, CEO of Crisis Text Line, shows how a distributed volunteer model can scale rapidly without cutting corners — and why the absence of a national social services coordination system costs lives.

Two waves of crisis volume

• Wave one: anxiety spiked in exact proportion to COVID-positive states — top 15 COVID states matched top 15 states mentioning it
• Wave two: grief from losses; domestic violence, child abuse, and substance abuse as quarantine trapped people at home
• The word "home" rose from 6% to 14% of conversations; that 14% showed a 2X increase in abuse
• Bad words early: "intense anxiety," "freaked out," "panic"
• Good signals: "mom" shifted from a pain point to a source of solace; "courage" appeared; short time horizons ("tonight," "tomorrow") helped de-escalate

Who is struggling and who isn't

• Normally 53% of volume comes from under-17s — that group is now doing better (no school stress, no relationship drama)
• Hardest hit: 18–34 year-olds — back home, isolated, financially stressed, this was supposed to be their moment
• Fastest-growing segment: 44–55 year-olds — sandwiched between elderly parents and financial anxiety
• Volume peaked at 6,000 conversations per day in the US; settled around 4,500; all-time highs hit in Canada and the UK

Scaling a volunteer network under pressure

• Crisis Text Line operates like a distributed marketplace — no control over supply or demand
• Volunteers apply online, pass a background check, complete 30 hours of remote training; only 30–32% pass
• When volume surged 50–100%, existing volunteers stepped up with higher "lift" (more conversations per person)
• Satisfaction (NPS) rose from 85% to 88% during the surge — Starbucks benchmarks at 78%
• Training standards were never relaxed; quality and wait times both improved
• The model is a distributed network of "empathy MVPs" turning their own stress into care for strangers

The broken social services infrastructure

• No equivalent of OpenTable exists for social services — organizations know where resources are, but not real-time availability
• Crisis counselors send texters to hospitals or shelters with no way to know if beds, rape kits, or spots are available
• The same fragmentation is now visible in COVID resource allocation: tests, ventilators, hospital beds
• Nancy's ask: even an API or an Excel spreadsheet — any shared, live feed of service availability
• The data that system would generate (where rape kits run out, where shelter beds are short) would itself be a policy tool

Nonprofit funding under strain

• The nonprofit business model — "I do great work, please give me money" — was already broken before COVID
• Spring fundraiser dinners, a lifeline for many organizations, are now impossible
• Crisis Text Line's model: a small number of high-net-worth donors plus passive inbound giving driven by NPS
• Organizations without diversified funding are in acute danger

Advice for leaders

• Kindness is compatible with hard decisions — layoffs and furloughs done with care are remembered
• Peer connection matters: reach out to other founders who understand the pressure
• Shrink the horizon: "what are you doing tonight to stay strong?" works better than grappling with the whole crisis
• Asian texters doubled from 5% to 10% due to harassment; Asian volunteer applications jumped from 10% to 27% — turning pain into action