how the system is countedsnapshot taken 2026-09-15

How many Americans have a mental illness?

Two federal agencies answer two different questions. SAMHSA asks a sample of the population, every year, whether they were depressed, distressed, suicidal or using substances, and publishes the share who said yes, by age. AHRQ counts the people who were admitted to a hospital for it, what the stay cost and how long it lasted. This page puts both on the same screen, as time series, with each agency's own limits printed above its numbers. It mirrors the explorer in the veisund app, because a chart inside an app cannot be linked, quoted or checked, and a page can.

Four rules that govern every number below. First, nothing is estimated or interpolated here: every value is as published by the source named under its chart, and a number the sources do not publish is listed in the final section as missing rather than approximated. Second, the survey lines start in 2021 and not before, because SAMHSA redesigned the survey that year and says its earlier estimates are not comparable; a chart that ran a line back to 2008 would be publishing something false. Third, the survey counts people living in households, dorms and shelters. It does not reach people in hospitals, prisons, nursing homes or on the street, so the groups most likely to be severely ill are the ones it cannot see. Fourth, every value is plain text in the table under its chart, not only pixels in a picture.

Depression, as the survey counts it

A "major depressive episode" in NSDUH means the respondent answered a fixed set of interview questions in a way that matches the DSM criteria for an episode in the past year. It is not a diagnosis and it is not a screening score; the questionnaire page linked at the bottom explains how those three things differ. Adolescents are asked a separate set of questions and are charted separately.

Major depressive episode in the past year, adults, by age

Read this before the numbers

  • Major depressive episode in the past year, adults 18 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks the 2021-2025 change for adults as a statistically significant decrease.
  • Major depressive episode in the past year, adults 18 to 25. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. The young-adult decline (19.3 to 14.2) is the largest movement in the mental-health section; SAMHSA marks it as a decrease.
  • Major depressive episode in the past year, adults 26 to 49. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks 2021-2025 as no significant change.
  • Major depressive episode in the past year, adults 50 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks 2021-2025 as no significant change.

Major depressive episode in the past year, adults, by age

Four age groups from the same survey. The 18 to 25 line is the one that moved most.

0%10%20%30%20212022202320242025
all adults 18+ages 18 to 25ages 26 to 49ages 50 and older
% of people in the age group, past year · measurement type: survey self-report, past-year, structured interview items · SAMHSA, Key Substance Use and Mental Health Indicators in the United States: Results from the 2025 National Survey on Drug Use and Health (Annual National Report), figure tables, percentages 2021-2025 · retrieved
show the numbers
Major depressive episode in the past year, adults, by age, 2021 to 2025, in % of people in the age group, past year. all adults 18+: from 8.5% in 2021 to 7.4% in 2025; ages 18 to 25: from 19.3% in 2021 to 14.2% in 2025; ages 26 to 49: from 9.6% in 2021 to 9.1% in 2025; ages 50 and older: from 4.5% in 2021 to 4% in 2025. The exact values are in the table under the chart.
yearall adults 18+ages 18 to 25ages 26 to 49ages 50 and older
20218.5%19.3%9.6%4.5%
20228.8%20.1%9.7%4.6%
20238.5%17.5%10.2%4.5%
20248.2%15.9%10%4.4%
20257.4%14.2%9.1%4%
Major depressive episode in the past year, adults, by age, 2021 to 2025, in % of people in the age group, past year. all adults 18+: from 8.5% in 2021 to 7.4% in 2025; ages 18 to 25: from 19.3% in 2021 to 14.2% in 2025; ages 26 to 49: from 9.6% in 2021 to 9.1% in 2025; ages 50 and older: from 4.5% in 2021 to 4% in 2025. The exact values are in the table under the chart.

methodology: all adults 18+ · methodology: ages 18 to 25 · methodology: ages 26 to 49 · methodology: ages 50 and older

Major depressive episode in the past year, adolescents 12 to 17

Read this before the numbers

  • Major depressive episode in the past year, adolescents 12 to 17. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. Read from the Figure 54 table (MDE row); 2025 is 15.1% or 3.7 million adolescents. Of adolescents with a past year MDE in 2025, SAMHSA reports 57.7% (2.1 million) received mental health treatment; treatment questions changed in 2025, so no treatment trend is served (see absences).
  • Major depressive episode with severe impairment in the past year, adolescents 12 to 17. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. Figure 54 table, severe-impairment row; 2025 is 11.0% or 2.7 million, which SAMHSA reports as 73.8% of adolescents with an MDE.

Major depressive episode in the past year, adolescents 12 to 17

The whole episode measure and the subset with severe impairment, from the adolescent questionnaire.

0%10%20%30%20212022202320242025
any major depressive episodewith severe impairment
% of adolescents 12 to 17, past year · measurement type: survey self-report, past-year, structured interview items · SAMHSA, Key Substance Use and Mental Health Indicators in the United States: Results from the 2025 National Survey on Drug Use and Health (Annual National Report), figure tables, percentages 2021-2025 · retrieved
show the numbers
Major depressive episode in the past year, adolescents 12 to 17, 2021 to 2025, in % of adolescents 12 to 17, past year. any major depressive episode: from 20.8% in 2021 to 15.1% in 2025; with severe impairment: from 15.2% in 2021 to 11% in 2025. The exact values are in the table under the chart.
yearany major depressive episodewith severe impairment
202120.8%15.2%
202219.5%14.6%
202318.1%13.5%
202415.4%11.3%
202515.1%11%
Major depressive episode in the past year, adolescents 12 to 17, 2021 to 2025, in % of adolescents 12 to 17, past year. any major depressive episode: from 20.8% in 2021 to 15.1% in 2025; with severe impairment: from 15.2% in 2021 to 11% in 2025. The exact values are in the table under the chart.

methodology: any major depressive episode · methodology: with severe impairment

Any mental illness and serious mental illness

These two are not asked directly. SAMHSA estimates them from a short distress and disability scale using a statistical model calibrated on a subsample who received clinical interviews. "Serious" means the model predicts serious functional impairment. Both are adult-only measures.

Any mental illness and serious mental illness in the past year, adults

Read this before the numbers

  • Any mental illness in the past year, adults 18 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. AMI is a model-based estimate of any diagnosable mental disorder in the past year, not a count of diagnoses. SAMHSA marks 2021-2025 as no significant change for all adults.
  • Any mental illness in the past year, adults 18 to 25. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks the young-adult 2021-2025 change as a decrease.
  • Serious mental illness in the past year, adults 18 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SMI is the subset of AMI with serious functional impairment, model-estimated. SAMHSA marks 2021-2025 as a decrease.
  • Serious mental illness in the past year, adults 18 to 25. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks the young-adult 2021-2025 change as a decrease.

Any mental illness and serious mental illness in the past year, adults

All adults and the 18 to 25 group, for both measures.

0%10%20%30%40%20212022202320242025
any mental illness, adults 18+any mental illness, ages 18 to 25serious mental illness, adults 18+serious mental illness, ages 18 to 25
% of people in the age group, past year · measurement type: model-estimated from a distress and impairment scale, calibrated to clinical interviews · SAMHSA, Key Substance Use and Mental Health Indicators in the United States: Results from the 2025 National Survey on Drug Use and Health (Annual National Report), figure tables, percentages 2021-2025 · retrieved
show the numbers
Any mental illness and serious mental illness in the past year, adults, 2021 to 2025, in % of people in the age group, past year. any mental illness, adults 18+: from 21% in 2021 to 20.6% in 2025; any mental illness, ages 18 to 25: from 33.2% in 2021 to 28.6% in 2025; serious mental illness, adults 18+: from 7.9% in 2021 to 6.9% in 2025; serious mental illness, ages 18 to 25: from 16.8% in 2021 to 12.1% in 2025. The exact values are in the table under the chart.
yearany mental illness, adults 18+any mental illness, ages 18 to 25serious mental illness, adults 18+serious mental illness, ages 18 to 25
202121%33.2%7.9%16.8%
202221.3%35.2%8.4%17.6%
202320.9%32.7%7.9%15.3%
202421.5%31.9%7.8%14.1%
202520.6%28.6%6.9%12.1%
Any mental illness and serious mental illness in the past year, adults, 2021 to 2025, in % of people in the age group, past year. any mental illness, adults 18+: from 21% in 2021 to 20.6% in 2025; any mental illness, ages 18 to 25: from 33.2% in 2021 to 28.6% in 2025; serious mental illness, adults 18+: from 7.9% in 2021 to 6.9% in 2025; serious mental illness, ages 18 to 25: from 16.8% in 2021 to 12.1% in 2025. The exact values are in the table under the chart.

methodology: any mental illness, adults 18+ · methodology: any mental illness, ages 18 to 25 · methodology: serious mental illness, adults 18+ · methodology: serious mental illness, ages 18 to 25

Suicidal thoughts and attempts

Asked directly, of adults only, about the past twelve months. The two age lines run in opposite directions over these five years, which is the kind of thing a single national average hides.

Serious thoughts of suicide and suicide attempts in the past year, adults

Read this before the numbers

  • Serious thoughts of suicide in the past year, adults 18 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks all adults 2021-2025 as no significant change; the age groups move in opposite directions (see the 18 to 25 and 50 or older series).
  • Serious thoughts of suicide in the past year, adults 18 to 25. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks 2021-2025 as a decrease.
  • Serious thoughts of suicide in the past year, adults 50 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks 2021-2025 as an INCREASE (2.0 to 2.9), the opposite direction from young adults.
  • Attempted suicide in the past year, adults 18 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks 2021-2025 as an increase (0.7 to 0.9); attempts among adults 50 or older rose from 0.2 to 0.5. Small percentages of a large population: 0.9% of adults is roughly 2.3 million people.

Serious thoughts of suicide and suicide attempts in the past year, adults

Thoughts for three age groups; attempts for all adults.

0%5%10%15%20212022202320242025
serious thoughts, adults 18+serious thoughts, ages 18 to 25serious thoughts, ages 50 and olderattempted suicide, adults 18+
% of people in the age group, past year · measurement type: survey self-report, past-year, structured interview items · SAMHSA, Key Substance Use and Mental Health Indicators in the United States: Results from the 2025 National Survey on Drug Use and Health (Annual National Report), figure tables, percentages 2021-2025 · retrieved
show the numbers
Serious thoughts of suicide and suicide attempts in the past year, adults, 2021 to 2025, in % of people in the age group, past year. serious thoughts, adults 18+: from 4.9% in 2021 to 5.3% in 2025; serious thoughts, ages 18 to 25: from 13.4% in 2021 to 10.8% in 2025; serious thoughts, ages 50 and older: from 2% in 2021 to 2.9% in 2025; attempted suicide, adults 18+: from 0.7% in 2021 to 0.9% in 2025. The exact values are in the table under the chart.
yearserious thoughts, adults 18+serious thoughts, ages 18 to 25serious thoughts, ages 50 and olderattempted suicide, adults 18+
20214.9%13.4%2%0.7%
20225.2%13.6%2.4%0.6%
20235%12.2%2.1%0.6%
20245.5%12.6%2.9%0.8%
20255.3%10.8%2.9%0.9%
Serious thoughts of suicide and suicide attempts in the past year, adults, 2021 to 2025, in % of people in the age group, past year. serious thoughts, adults 18+: from 4.9% in 2021 to 5.3% in 2025; serious thoughts, ages 18 to 25: from 13.4% in 2021 to 10.8% in 2025; serious thoughts, ages 50 and older: from 2% in 2021 to 2.9% in 2025; attempted suicide, adults 18+: from 0.7% in 2021 to 0.9% in 2025. The exact values are in the table under the chart.

methodology: serious thoughts, adults 18+ · methodology: serious thoughts, ages 18 to 25 · methodology: serious thoughts, ages 50 and older · methodology: attempted suicide, adults 18+

Substance use

The same survey asks about alcohol and drugs. "Substance use disorder" is the DSM-5 definition applied to interview answers; "misuse" of a prescription drug means use without a prescription of one’s own, or in a way a prescriber did not direct. Everyone 12 and older is in the denominator.

Substance use disorder, prescription misuse and marijuana use, people 12 and older

Read this before the numbers

  • Substance use disorder in the past year, people 12 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SUD is assessed against DSM-5 criteria from questionnaire items. SAMHSA marks 2021-2025 as a decrease.
  • Marijuana use in the past month, people 12 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks 2021-2025 as an increase for people 12 or older and for adults 26 or older, and a decrease among adolescents.
  • Prescription opioid misuse in the past year, people 12 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. Misuse means use without a prescription of one's own or other than as directed. SAMHSA marks 2021-2025 as a decrease.
  • Prescription stimulant misuse in the past year, people 12 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. Flat at 1.4-1.5% overall; among adults 18 to 25 it fell from 4.1 to 2.7, which SAMHSA marks as a decrease.
  • Opioid use disorder in the past year, people 12 or older. NSDUH redesigned its methods in 2021; SAMHSA publishes comparable estimates for 2021 onward only and re-weighted 2021, so no honest line extends left of 2021. Self-reported survey measure, not a diagnosis: MDE is assessed against DSM criteria from questionnaire items, AMI and SMI are model-based predictions from those items. SAMHSA marks 2021-2025 as a decrease (1.9 to 1.4).

Substance use disorder, prescription misuse and marijuana use, people 12 and older

Five measures on one axis. Marijuana is past month; the rest are past year.

0%5%10%15%20%20212022202320242025
any substance use disorder, past yearmarijuana use, past monthprescription opioid misuse, past yearprescription stimulant misuse, past yearopioid use disorder, past year
% of people in the age group, past year · measurement type: survey self-report, past-year, structured interview items · SAMHSA, Key Substance Use and Mental Health Indicators in the United States: Results from the 2025 National Survey on Drug Use and Health (Annual National Report), figure tables, percentages 2021-2025 · retrieved
show the numbers
Substance use disorder, prescription misuse and marijuana use, people 12 and older, 2021 to 2025, in % of people in the age group, past year. any substance use disorder, past year: from 16.7% in 2021 to 15.3% in 2025; marijuana use, past month: from 13.2% in 2021 to 15.1% in 2025; prescription opioid misuse, past year: from 3% in 2021 to 2.4% in 2025; prescription stimulant misuse, past year: from 1.4% in 2021 to 1.4% in 2025; opioid use disorder, past year: from 1.9% in 2021 to 1.4% in 2025. The exact values are in the table under the chart.
yearany substance use disorder, past yearmarijuana use, past monthprescription opioid misuse, past yearprescription stimulant misuse, past yearopioid use disorder, past year
202116.7%13.2%3%1.4%1.9%
202217.3%15%2.8%1.5%2%
202317.1%15.4%2.9%1.4%1.8%
202416.8%15.4%2.6%1.4%1.7%
202515.3%15.1%2.4%1.4%1.4%
Substance use disorder, prescription misuse and marijuana use, people 12 and older, 2021 to 2025, in % of people in the age group, past year. any substance use disorder, past year: from 16.7% in 2021 to 15.3% in 2025; marijuana use, past month: from 13.2% in 2021 to 15.1% in 2025; prescription opioid misuse, past year: from 3% in 2021 to 2.4% in 2025; prescription stimulant misuse, past year: from 1.4% in 2021 to 1.4% in 2025; opioid use disorder, past year: from 1.9% in 2021 to 1.4% in 2025. The exact values are in the table under the chart.

methodology: any substance use disorder, past year · methodology: marijuana use, past month · methodology: prescription opioid misuse, past year · methodology: prescription stimulant misuse, past year · methodology: opioid use disorder, past year

Who ends up in a hospital bed

A different agency and a different kind of number. AHRQ’s Healthcare Cost and Utilization Project does not ask anyone anything; it counts discharge records from community hospitals. A stay is counted by its principal diagnosis. Costs are what the hospital spent, not what it billed, in the dollars of the year, not adjusted for inflation.

Mental health and substance use hospital stays per 100,000 people, all ages

Read this before the numbers

  • Mental health and substance use hospital stays per 100,000 people, all ages. Hospital discharges, not people; a person admitted twice is two stays. The mental health/substance use type is assigned by the principal diagnosis (MDC 19 or 20 from late 2015; a CCS-based definition before), so the 2015 ICD-10 transition is a definition change that AHRQ documents in the file's notes. The rate fell from 697 in 2010 to 514 in 2023 while the all-cause rate fell less; whether that is fewer episodes or fewer beds is not something a discharge count can say.

Mental health and substance use hospital stays per 100,000 people, all ages

Stays with a mental health or substance use principal diagnosis, community hospitals, 2002 to 2023.

020040060080020022004200620082010201220142016201820202022
stays per 100,000 population
inpatient stays per 100,000 population · measurement type: hospital discharge records, principal diagnosis · AHRQ, HCUP Fast Stats, National Trends in Inpatient Stays (data years through 2023; file dated 12 December 2025), Nationwide Inpatient Sample · retrieved
show the numbers
Mental health and substance use hospital stays per 100,000 people, all ages, 2002 to 2023, in inpatient stays per 100,000 population. stays per 100,000 population: from 645.9 in 2002 to 513.9 in 2023. The exact values are in the table under the chart.
yearstays per 100,000 population
2002645.9
2003648.4
2004697.2
2005626.7
2006643.4
2007648.6
2008645.1
2009679.8
2010697
2011677.4
2012663.6
2013648.3
2014654.1
2015664.8
2016591.7
2017599.2
2018594.4
2019577.2
2020525.9
2021529.1
2022495.7
2023513.9
Mental health and substance use hospital stays per 100,000 people, all ages, 2002 to 2023, in inpatient stays per 100,000 population. stays per 100,000 population: from 645.9 in 2002 to 513.9 in 2023. The exact values are in the table under the chart.

methodology: stays per 100,000 population

What a hospital stay costs the hospital, mental health against everything

Read this before the numbers

  • Cost per mental health or substance use hospital stay, nominal dollars. Hospital discharges, not people; a person admitted twice is two stays. The mental health/substance use type is assigned by the principal diagnosis (MDC 19 or 20 from late 2015; a CCS-based definition before), so the 2015 ICD-10 transition is a definition change that AHRQ documents in the file's notes. Nominal dollars; AHRQ also publishes an inflation-adjusted series ($7,564 in 2023 versus $10,320 nominal). Cost is what the hospital spent, estimated from cost-to-charge ratios, not what was billed.
  • Cost per hospital stay, all causes, nominal dollars. Served beside the mental-health line for scale: in 2023 the average stay cost $17,697 and the average mental health/substance use stay $10,320, at 5.1 versus 7.5 days.

What a hospital stay costs the hospital, mental health against everything

Nominal dollars per stay, 2000 to 2023. Cost, not charges; not inflation-adjusted.

$0$5,000$10,000$15,000$20,000200020022004200620082010201220142016201820202022
mental health or substance use stayall stays, all causes
$ per stay, nominal (hospital cost, not charges) · measurement type: hospital discharge records, cost-to-charge converted · AHRQ, HCUP Fast Stats, National Trends in Inpatient Stays (data years through 2023; file dated 12 December 2025), Nationwide Inpatient Sample · retrieved
show the numbers
What a hospital stay costs the hospital, mental health against everything, 2000 to 2023, in $ per stay, nominal (hospital cost, not charges). mental health or substance use stay: from $4,435.3 in 2000 to $10,319.9 in 2023; all stays, all causes: from $6,072.8 in 2000 to $17,696.7 in 2023. The exact values are in the table under the chart.
yearmental health or substance use stayall stays, all causes
2000$4,435.3$6,072.8
2001$4,690.9$6,333.1
2002$5,842.9$6,889.2
2003$5,083.2$7,276.9
2004$5,667$7,613.6
2005$5,677$8,035.7
2006$5,770.1$8,305.2
2007$6,054.1$8,652.8
2008$5,890$9,106
2009$5,843.5$9,111.2
2010$5,824.9$9,680.9
2011$6,438.7$9,973.1
2012$6,388.3$10,355.4
2013$7,193.4$10,730
2014$6,667.9$10,888.9
2015$6,829$11,259.4
2016$6,614.1$11,727.9
2017$6,817.4$12,182.7
2018$6,973.5$12,511.8
2019$7,417.2$13,438.1
2020$8,331.6$14,912.6
2021$8,835.4$15,875.3
2022$9,757.8$16,720.2
2023$10,319.9$17,696.7
What a hospital stay costs the hospital, mental health against everything, 2000 to 2023, in $ per stay, nominal (hospital cost, not charges). mental health or substance use stay: from $4,435.3 in 2000 to $10,319.9 in 2023; all stays, all causes: from $6,072.8 in 2000 to $17,696.7 in 2023. The exact values are in the table under the chart.

methodology: mental health or substance use stay · methodology: all stays, all causes

How long a mental health or substance use stay lasts

Read this before the numbers

  • Average length of a mental health or substance use hospital stay. Hospital discharges, not people; a person admitted twice is two stays. The mental health/substance use type is assigned by the principal diagnosis (MDC 19 or 20 from late 2015; a CCS-based definition before), so the 2015 ICD-10 transition is a definition change that AHRQ documents in the file's notes. The all-cause mean was 5.1 days in 2023.

How long a mental health or substance use stay lasts

Mean days per stay, 1994 to 2023.

0510151996200020042008201220162020
mean days per stay
mean days per stay · measurement type: hospital discharge records · AHRQ, HCUP Fast Stats, National Trends in Inpatient Stays (data years through 2023; file dated 12 December 2025), Nationwide Inpatient Sample · retrieved
show the numbers
How long a mental health or substance use stay lasts, 1994 to 2023, in mean days per stay. mean days per stay: from 10.1 in 1994 to 7.5 in 2023. The exact values are in the table under the chart.
yearmean days per stay
199410.1
19959.3
19968.5
19978
19987.5
19997.6
20007
20017.3
20027.6
20037.1
20047.2
20057.3
20067.1
20077.2
20087.2
20097
20106.9
20117.2
20126.8
20136.9
20146.9
20156.9
20166.9
20177
20187
20197
20207.2
20217.3
20227.6
20237.5
How long a mental health or substance use stay lasts, 1994 to 2023, in mean days per stay. mean days per stay: from 10.1 in 1994 to 7.5 in 2023. The exact values are in the table under the chart.

methodology: mean days per stay

Children hospitalised for depression, per 100,000 children

Read this before the numbers

  • Hospital stays for mood disorders per 100,000 children, ages 0 to 17 (CCS definition, 2004-2015). ENDS IN 2015 BY DESIGN. Through 2015 AHRQ grouped children's stays under the ICD-9 CCS category 'Mood disorders'; from 2016 the ICD-10 CCSR category 'Depressive disorders' is narrower. The two series must not be joined; the 2016 drop is partly the definition. Maternal and neonatal stays excluded.
  • Hospital stays for depressive disorders per 100,000 children, ages 0 to 17 (CCSR definition, 2016-2023). STARTS IN 2016 BY DESIGN (see the 2004-2015 series). Depressive disorders were the number one non-birth reason for a child's hospital stay in 2020, 2021 and 2023 and number two in 2019 and 2022, ahead of respiratory failure, bronchitis and asthma. Maternal and neonatal stays excluded.

Children hospitalised for depression, per 100,000 children

Two series, not one. Through 2015 AHRQ counted "mood disorders" under ICD-9; from 2016 it counts the narrower "depressive disorders" under ICD-10. The band marks where the ruler changed and no line crosses it.

definition changed: do not read across0501001502002004200620082010201220142016201820202022
mood disorders (ICD-9 CCS), 2004 to 2015depressive disorders (ICD-10 CCSR), 2016 to 2023
inpatient stays per 100,000 children 0-17 · measurement type: hospital discharge records, principal diagnosis, ages 0 to 17 · AHRQ, HCUP Fast Stats, U.S. National Top Diagnoses (data years 2004-2023; file dated 12 December 2025), maternal and neonatal stays excluded · retrieved
show the numbers
Children hospitalised for depression, per 100,000 children, 2004 to 2023, in inpatient stays per 100,000 children 0-17. mood disorders (ICD-9 CCS), 2004 to 2015: from 110 in 2004 to 157.2 in 2015; depressive disorders (ICD-10 CCSR), 2016 to 2023: from 104.2 in 2016 to 121.3 in 2023. The exact values are in the table under the chart.
yearmood disorders (ICD-9 CCS), 2004 to 2015depressive disorders (ICD-10 CCSR), 2016 to 2023
2004110
200596.9
200681.8
2007105.9
2008107.1
2009112.2
2010155.1
2011143.3
2012134.4
2013146.3
2014158.2
2015157.2
2016104.2
2017122.4
2018124.3
2019124.7
2020110.4
2021133.6
2022119.1
2023121.3
Children hospitalised for depression, per 100,000 children, 2004 to 2023, in inpatient stays per 100,000 children 0-17. mood disorders (ICD-9 CCS), 2004 to 2015: from 110 in 2004 to 157.2 in 2015; depressive disorders (ICD-10 CCSR), 2016 to 2023: from 104.2 in 2016 to 121.3 in 2023. The exact values are in the table under the chart.

methodology: mood disorders (ICD-9 CCS), 2004 to 2015 · methodology: depressive disorders (ICD-10 CCSR), 2016 to 2023

Young adults hospitalised for depression and for psychosis, per 100,000 adults 18 to 44

Read this before the numbers

  • Hospital stays for depressive disorders per 100,000 adults, ages 18 to 44 (2016-2023). CCSR definition from 2016; the pre-2016 'Mood disorders' category is not comparable and is not served for this age group. Third most common non-birth reason for a stay in 2023, behind septicemia and schizophrenia spectrum disorders.
  • Hospital stays for schizophrenia spectrum and other psychotic disorders per 100,000 adults, ages 18 to 44 (2016-2023). Second most common non-birth reason for a hospital stay among adults 18 to 44 in 2023 (206,005 stays), behind septicemia. Discharges, not people.

Young adults hospitalised for depression and for psychosis, per 100,000 adults 18 to 44

Two principal-diagnosis groups from the same file, 2016 to 2023.

010020030020162017201820192020202120222023
depressive disordersschizophrenia spectrum and other psychotic disorders
inpatient stays per 100,000 adults 18-44 · measurement type: hospital discharge records, principal diagnosis, ages 18 to 44 · AHRQ, HCUP Fast Stats, U.S. National Top Diagnoses (data years 2004-2023; file dated 12 December 2025), maternal and neonatal stays excluded · retrieved
show the numbers
Young adults hospitalised for depression and for psychosis, per 100,000 adults 18 to 44, 2016 to 2023, in inpatient stays per 100,000 adults 18-44. depressive disorders: from 202.3 in 2016 to 162.7 in 2023; schizophrenia spectrum and other psychotic disorders: from 183.3 in 2016 to 173 in 2023. The exact values are in the table under the chart.
yeardepressive disordersschizophrenia spectrum and other psychotic disorders
2016202.3183.3
2017214.7186.4
2018208.7189.8
2019202.7186.6
2020172.2192
2021168.8193.5
2022156.3181.7
2023162.7173
Young adults hospitalised for depression and for psychosis, per 100,000 adults 18 to 44, 2016 to 2023, in inpatient stays per 100,000 adults 18-44. depressive disorders: from 202.3 in 2016 to 162.7 in 2023; schizophrenia spectrum and other psychotic disorders: from 183.3 in 2016 to 173 in 2023. The exact values are in the table under the chart.

methodology: depressive disorders · methodology: schizophrenia spectrum and other psychotic disorders

What is not here, and why

The API this page mirrors refuses to serve some numbers people expect to find, and says why. The refusals are part of the dataset and are reproduced in full.

  • Mental health prevalence before 2021. NSDUH changed its data collection and weighting in 2021 (web and in-person modes, re-weighted 2021 estimates). SAMHSA publishes comparable trend estimates for 2021 onward only and says earlier estimates should not be compared. A line drawn through 2019 and 2021 would show the survey redesign, not the population, so none is served.
  • Share of people receiving mental health treatment, over time. NSDUH rewrote its treatment questions for 2025, and SAMHSA states that fewer than four years of comparable data are available, so it publishes 2025 only: 65.9% of adults with a past year major depressive episode and 57.7% of adolescents with one received mental health treatment. A single year is a number, not a series, and is carried in the caveats rather than plotted.
  • National antidepressant use and spending by year. AHRQ's MEPS household survey holds this series, but its public data tool serves it only through interactive Tableau dashboards and the legacy summary tables refused an unauthenticated read on the retrieval date. The library will carry it when the underlying MEPS tables can be read and cited directly; a number transcribed from a chart image is not a citation.
  • Who pays for mental health hospital stays, nationally. HCUP publishes payer breakdowns for mental health and substance use stays by STATE, for the states that participate in a given year (43 states in 2023). Summed across reporting states, Medicaid was the expected payer for roughly 37% of adult mental health and substance use inpatient stays and 38% of such emergency visits in 2023, private insurance about a quarter, self-pay 8% of stays and 16% of ED visits. That is a sum over participating states, not a national estimate, so it is stated here and not plotted.

Questions people ask

Short answers, each a number already on this page with its limit attached.

How many Americans have a mental illness?

By the federal survey (NSDUH), 20.6% of adults 18 and older had any mental illness in the past year in 2025, about one in five, and the figure has stayed between 20.6% and 21.5% every year since 2021. Serious mental illness, the subset with serious functional impairment, was 6.9% of adults in 2025, down from 7.9% in 2021. Both are model estimates from a distress and disability scale, not counts of diagnoses, and the survey does not reach people in hospitals, prisons, nursing homes or on the street.

What percentage of adults had a major depressive episode in the past year?

7.4% of adults 18 and older in 2025, down from 8.5% in 2021. By age: 14.2% of 18 to 25 year olds, 9.1% of 26 to 49 year olds, and 4.0% of people 50 and older. A major depressive episode in NSDUH means interview answers that matched the diagnostic criteria; it is not a clinician diagnosis and not a screening score.

Is teen depression going up or down?

Down in the federal survey. Among adolescents aged 12 to 17, a past-year major depressive episode fell from 20.8% in 2021 to 15.1% in 2025, and an episode with severe impairment fell from 15.2% to 11.0%. SAMHSA gives no cause. In 2025, 57.7% of adolescents with an episode received any mental health treatment.

What percentage of adults have suicidal thoughts?

5.3% of adults 18 and older had serious thoughts of suicide in the past year in 2025, unchanged from 2021 nationally. Under that average, adults 18 to 25 fell from 13.4% to 10.8% while adults 50 and older rose from 2.0% to 2.9%. Suicide plans rose from 1.4% to 1.8% of adults and attempts from 0.7% to 0.9%. If you are in crisis in the US, call or text 988.

Why do the survey numbers only start in 2021?

SAMHSA redesigned NSDUH in 2021, moving from in-person interviews to a mix of web and in-person, and states that estimates from 2021 onward are not comparable with 2020 or earlier. This page and the app explorer draw no line across that break; earlier years are listed as a deliberate absence rather than approximated.

How many people are hospitalised for mental health reasons, and what does a stay cost?

AHRQ counted 1,719,031 community hospital stays with a mental health or substance use principal diagnosis in 2023, a rate of 513.9 per 100,000 people, down from 697.0 per 100,000 in 2010. The hospital's cost per stay was $10,320 in nominal 2023 dollars ($7,564 inflation-adjusted in AHRQ's file) against $17,697 for all stays, and the average mental health stay lasted 7.5 days against 5.1 for all causes. Cost means what the hospital spent, not what was billed or paid.

Who pays for mental health hospital stays?

Across the 43 states reporting to AHRQ for 2023, Medicaid was the expected payer for about 37% of adult mental health or substance use stays, private insurance about 24%, Medicare (65 and older) about 11%, and self-pay or no charge about 8%. These shares are our arithmetic on AHRQ's state file and are not a national estimate.

The same people, from other angles

The survey counts symptoms and distress; a different federal survey counts prescriptions, and the two do not move together. That count is on how many adults take psychiatric medication. The nine-question form a primary care office uses to produce a depression score, and why the score is not the same thing as the survey's episode measure, is on who wrote the PHQ-9 and who gets paid on your score. What the hospital stays and everything else cost, in aggregate, is on where the money goes.

Method, in full

This page renders a pinned snapshot of /api/explorer/prevalence, taken 2026-09-15, the same payload the veisund app explorer renders. The API serves each series as a pinned, hand-checked snapshot of its named source rather than proxying the source agencies live; the retrieval date printed under each chart is the date the values were taken from that source. The survey series are read from the figure tables of SAMHSA's 2025 NSDUH Annual National Report; the hospital series are read from AHRQ's HCUP Fast Stats data files (National Trends in Inpatient Stays and U.S. National Top Diagnoses, data years through 2023). Each chart links its source's methodology page. Titles, units, caveats, source lines and the absences list are reproduced verbatim from the payload. Nothing is edited, estimated, deflated, interpolated or extrapolated here, and the page draws no conclusions the sources do not state.