558,922 37,650
psychiatric beds · 1955 psychiatric beds · today
One Hundred Years of The Untreated
America closed its psychiatric hospitals. It never built what was supposed to replace them.
ENTER THE RECORD →
One Hundred Years of

The Untreated

Data sources, confidence levels, and methodological limitations.
Report   06 / 10
Confidence   3-tier system
Confidence system

Three tiers of confidence

Every claim in this report is assigned a confidence level. The data speaks, but it speaks with varying degrees of certainty.
HIGH CONFIDENCE
Verified
Multiple independent sources confirm. Government statistics with consistent methodology. Directly measurable.
• 1955 peak: 558,922 beds (SAMHSA, APA, TAC all agree)
• BJS incarceration rates (official census data)
• SMI in jails/prisons (DOJ surveys + TAC analysis)
CANDIDATE
Estimated
Derived from credible studies but with known methodological limitations. May involve extrapolation or modeling assumptions.
• Cost figures (vary by state; national averages)
• 16x encounter risk (TAC 2015 study; methodology debated)
• Homelessness counts (PIT methodology undercounts)
SPECULATIVE
Inferred
Reasonable inference from available data but not directly measured. Included for completeness with explicit uncertainty flags.
• Youth pipeline flow numbers (multiple data sources combined)
• Historical bed counts pre-1940 (limited records)
Data sources

Primary sources

Psychiatric beds. Historical data from Grob (1994), The Mad Among Us; Torrey et al. (2010); Treatment Advocacy Center reports (2012, 2016, 2020); SAMHSA Uniform Reporting System. The 1955 peak of 558,922 is a consensus figure across multiple independent sources.

Incarceration. Bureau of Justice Statistics, National Prisoner Statistics (NPS-1), 1925-2024. State and federal prisoner counts; rates computed against Census Bureau population estimates.

Cost data. SAMHSA cost-effectiveness studies; Vera Institute of Justice jail cost analyses; Treatment Advocacy Center economic impact reports; Congressional Budget Office estimates.

Fatal force. Washington Post Fatal Force database (2015-present); Treatment Advocacy Center (2015), "Overlooked in the Undercounted."

Homelessness. HUD Annual Homeless Assessment Report (AHAR); Point-in-Time (PIT) counts.

Interpolation method

Between anchor data points, this report uses linear interpolation. All anchor points are sourced; interpolated years are marked as derived. The interpolation does not add information — it connects known points for visual continuity.

Known limitations

• Bed counts. Pre-1940 records are incomplete. State-by-state reporting was inconsistent before SAMHSA standardization.

• SMI prevalence. Definitions of "serious mental illness" vary across surveys and decades. SAMHSA's current definition may not match historical classifications.

• Causal claims. Correlation between bed closures and incarceration increases is well-documented but does not prove direct causation. Multiple factors contributed to mass incarceration, including the War on Drugs and mandatory minimum sentencing.

• Cost comparisons. Per-person costs vary widely by state and facility. National averages obscure significant regional variation.

• Homelessness data. PIT counts are known to undercount. The actual number of people experiencing homelessness with SMI is likely higher than reported.

What this report does not argue

This report does not advocate for re-institutionalization. It does not argue that the old state hospital system was humane or effective. It documents a specific claim: that America dismantled one system of care without building an adequate replacement, and the consequences of that failure are measurable in beds, cells, sidewalks, and morgues.