Syndromic surveillance is a kind of stethoscope held to the whole population. When someone arrives at an emergency department, the facility forwards a de-identified record — chief complaint, diagnosis codes, age, location — to state, local, and federal health officials. More than eight in ten U.S. emergency departments now feed this stream. It is one of the few near-real-time windows we have into the country's mental health.
But the window has a quirk that shapes everything else on this page. These numbers are not counts. They are rates — specifically, the number of visits tied to a condition for every 100,000 ED visits of any kind. A rate is a fraction, and a fraction has a denominator. When the denominator moves, the rate moves, even if the thing you care about never changed at all.
That distinction is not pedantry. It is the difference between a signal and a mirage — and in the spring of 2020, this dataset produced one of the most instructive mirages in recent public-health data.
In April 2020, everything spiked at once
If a single month saw anxiety, depression, bipolar disorder, schizophrenia, trauma, and suicide attempts all surge together — by anywhere from a fifth to nearly two-thirds — you would be right to be suspicious. Distinct conditions do not rise in lockstep. Denominators do.
When the first pandemic lockdowns emptied waiting rooms, routine and non-urgent ED visits collapsed. People with broken wrists and chest colds stayed home; people in acute mental-health crisis still came. The total number of ED visits — the denominator — fell off a cliff. Mental health's share of what remained shot up, all of it, simultaneously. The chart below is the fingerprint of that collapse.
Strip out the mirage, and the share is drifting down
Once the denominator settled back to normal, a quieter story emerged. Across most conditions, mental health's share of ED visits in 2025 sits below where it was in 2019 — not a crisis acceleration, but a gentle decline.
This does not mean fewer Americans are struggling. A falling share of ED visits can reflect many things: ED volumes recovering, care moving to telehealth and crisis lines, or simply the noise of a maturing data system. It is a measure of where care lands, not of how many people need it. But within that drift, one line refuses to fall.
Suicide attempts wear a young face
The suicide-attempt signal is not evenly spread. When you break the most recent year by age, almost the entire weight of it lands on adolescents.
And unlike the broader drift downward, the adolescent line is not at rest. After the 2020–21 surge and a partial retreat, the youth rate has been climbing again — and the most recent readings approach the pandemic-era peak.
Each condition has its own demographic shape
Mental health is not one signal but seven, and each one shows up at the ED through a different door. Sex is the first divider.
Race and ethnicity draw a third map. The grid below indexes each group's rate against that condition's overall average, so conditions of very different size can be compared side by side. Read across a row to see who is over- or under-represented for that condition.
What this stethoscope can — and can't — hear
How to read these numbers honestly
It measures share, not suffering. Every figure here is a count of mental-health visits divided by all ED visits. It rises when more people in crisis show up or when fewer people show up for everything else. The April 2020 spike was the second kind. Always ask what the denominator is doing.
The ED is a last resort, not a census. These rates capture the moment care reaches the emergency department — crisis acuity, and the absence of somewhere else to go. They say little about the much larger population managing the same conditions in clinics, in therapy, or alone.
Small numbers are hidden on purpose. Monthly counts below ten are suppressed to protect privacy, so the sparsest demographic cells can read as gaps. Rates are also revised as electronic health records catch up, though revisions rarely change the overall direction of a trend.
The categories overlap. A single visit can be flagged for more than one condition, and "Any Mental Health" is broader than the sum of named disorders. Treat each line as its own signal, not as a slice of a fixed pie.
Read with those limits in mind, the surveillance stream is remarkable: a monthly pulse-check on the nation's mental health, drawn from the place people go when there is nowhere left to turn. The signal is real. It just rewards a careful reader.
A note on this topicSuicide and self-harm are difficult subjects. This article treats them as public-health data; if any of it touches something personal, the resources above are there for a reason, and reaching out is a sign of strength, not weakness.