The short version

Claim we investigated

Overdose deaths among people living in King County's subsidized housing have risen faster than the housing since 2021.

Mostly true. Residents' recorded overdose deaths rose 68% and housing beds 25%, unless investigators changed how they record where people lived. If investigators recorded where people lived the same way each year, overdose deaths per housing bed rose 35% from 2021 to 2025. If a recording change explains the housing's rising share, overdose deaths rose 19%, slower than the beds' 25%.

Quote or cite this fact-check
FactHub rates the claim “Overdose deaths among people living in King County's subsidized housing have risen faster than the housing since 2021” Mostly true (moderate confidence): Residents' recorded overdose deaths rose 68% and housing beds 25%, unless investigators changed how they record where people lived. Reviewed November 12, 2026. https://facthub.org/n/article/king-county-subsidized-housing-deaths-since-2021/9

Key points

What the evidence establishes

  • Established. The Medical Examiner recorded 192 overdose deaths among subsidized-housing residents in 2025, up 68% from 114 in 2021; countywide, overdose deaths rose 28%.
  • Established. Federally counted beds in this housing rose 25%, so overdose deaths per 1,000 beds rose from 14.5 to 19.6. Counting rapid rehousing beds too, they still rose 25%.
  • Established. The county hasn't said whether investigators changed how they record where people lived. If that explains the whole shift, overdose deaths rose 19%, slower than the beds.
  • Established. Deaths from all causes among residents rose 23% (309 to 379), about as fast as the beds, so deaths per bed were flat.
  • Established. Overdose deaths among residents peaked at 281 in 2023 and fell to 192 in 2025.
  • Not shown. No true death rate can be computed: the county publishes no count of residents, and its group mixes three kinds of housing.
Important context: Critics of the region's housing-first approach say more people are dying after moving indoors. More residents did die, but not faster than the housing grew: deaths from all causes that the Medical Examiner investigated rose about as fast as the beds (rated False in the common questions). We check the narrower part the evidence can settle: whether overdose deaths in this housing grew faster than the housing itself. The King County Medical Examiner investigates sudden, violent and unexplained deaths, including nearly all overdose deaths, and records where each person lived. One group, called housing service sites, covers permanent supportive housing (apartments with services for people who were homeless), income-restricted apartments and sober recovery homes. We call it subsidized housing. Most of the group's overdose deaths are in permanent supportive housing. We measure the housing by beds, the federal government's yearly count of places in it, because the county publishes no count of residents.

The evidence

The conclusion rests on the combined record. Expand each item to see its role and limitations.

Independence note: This check uses 5 methods across 4 institutions and 1 country or international grouping.
Evidence map
Method Institution / country Year(s) What it shows Key limitation Primary links
Overdose deaths by year among people whose presumed living situation was a housing service site, from the county's fatal overdose dashboard, compared with countywide overdose deaths King County Medical Examiner's Office, Public Health - Seattle & King County United States 2021, 2022, 2023, 2024, 2025 Overdose deaths among people living in subsidized housing were 114 in 2021, 177 in 2022, 281 in 2023, 235 in 2024 and 192 in 2025. That is up 68%, with a peak in 2023; countywide, overdose deaths rose 28%. In 2024 the county's overdose review counted 225 overdose deaths in permanent supportive housing alone, most of the group's 235. Every death count in this article, including the county's supportive-housing tally, comes from the same Medical Examiner records and its judgment of where people lived; only the HUD bed count is independent. FactHub dataset FactHub dataset Public Health - Seattle & King County FactHub dataset
HUD Housing Inventory Count beds for Seattle/King County set beside the Medical Examiner's overdose counts, with and without rapid rehousing beds U.S. Department of Housing and Urban Development, King County Medical Examiner's Office United States 2021, 2025 Permanent supportive and other permanent housing beds counted by the U.S. Department of Housing and Urban Development (HUD) rose 25%, from 7,860 in 2021 to 9,791 in 2025. Overdose deaths per 1,000 beds rose from 14.5 to 19.6 (35%). Adding rapid rehousing beds, which grew fastest, overdose deaths per 1,000 beds still rose 25% (12.6 to 15.8). Beds are a one-night count, not residents. HUD's count leaves out most income-restricted and recovery housing, which the county's group includes, so this compares growth rather than measuring a death rate. Beds also can't show turnover: if more people passed through each bed in a year, deaths per bed could rise without any rise in risk. FactHub dataset FactHub dataset FactHub dataset King County Regional Homelessness Authority
Review of the county's published methods, and comparison of the housing group with deaths among people living in private homes, outdoors, in shelters or in vehicles, and with an unknown living situation, 2021 to 2025 King County Medical Examiner's Office, Public Health - Seattle & King County United States 2021, 2022, 2023, 2024, 2025 Investigators record where each person was living, not where they died, judging from the scene and from witnesses or family rather than from a list of addresses. The county calls this hard because housing programs and locations keep changing, and it has published no change in method. Housing's share of overdose deaths among housed people rose from 21% to 29%; held at 21%, residents' overdose deaths would be up 19%, below the beds. Checks against other groups don't settle it: shelter and street deaths rose, unknowns can explain only part, and natural deaths can't be compared cleanly. No published record shows when the county began recording housing service sites or whether earlier years were recorded the same way; before 2022 its homeless-death reports used a single no-fixed-address field. The county also doesn't say how often a visitor who dies at a housing site is recorded as a resident. Public Health - Seattle & King County FactHub dataset FactHub dataset FactHub dataset FactHub dataset Public Health - Seattle & King County
Counts read by year and manner of death from the Medical Examiner's public annual-report dashboard, with the living-situation filter set to housing service site King County Medical Examiner's Office, Public Health - Seattle & King County United States 2021, 2022, 2023, 2024, 2025 Deaths the Medical Examiner investigated among people living in subsidized housing were 309 in 2021, 389 in 2022, 487 in 2023, 433 in 2024 and 379 in 2025. That is up 23%, about as fast as the beds, so deaths per 1,000 beds were 39.3 and then 38.7. Accidents other than traffic, mostly overdoses, rose from 129 to 200; natural deaths held between 164 and 170. These are only deaths the Medical Examiner investigated; it does not take all natural deaths. The group mixes supportive, income-restricted and recovery housing. FactHub dataset FactHub dataset Public Health - Seattle & King County King County Medical Examiner's Office
Randomized trial evaluated by the Urban Institute Urban Institute United States 2016, 2021 In a Denver trial of supportive housing for frequent jail users, deaths within two years were about the same whether or not people were offered housing. So more deaths where more people are housed doesn't by itself show the housing causes them. The trial was in Denver, ended before fentanyl dominated the drug supply, treated deaths as a secondary outcome, and studied a different group than King County's residents. Urban Institute
1. Recorded overdose deaths: 114 in 2021, 192 in 2025 Overdose deaths by year among people whose presumed living situation was a housing service site, from the county's fatal overdose dashboard, compared with countywide overdose deaths

Overdose deaths among people living in subsidized housing were 114 in 2021, 177 in 2022, 281 in 2023, 235 in 2024 and 192 in 2025. That is up 68%, with a peak in 2023; countywide, overdose deaths rose 28%. In 2024 the county's overdose review counted 225 overdose deaths in permanent supportive housing alone, most of the group's 235.

Limit: Every death count in this article, including the county's supportive-housing tally, comes from the same Medical Examiner records and its judgment of where people lived; only the HUD bed count is independent.

Verify: Change from 2021 to 2025: housing site deaths, county deaths and housing beds Did natural deaths shift toward the housing group the way overdoses did? King County Overdose Fatality Review 2024 Report King County overdose deaths by where people lived: deaths compared with HUD beds

2. Overdose deaths outgrew the housing beds on the recorded counts HUD Housing Inventory Count beds for Seattle/King County set beside the Medical Examiner's overdose counts, with and without rapid rehousing beds

Permanent supportive and other permanent housing beds counted by the U.S. Department of Housing and Urban Development (HUD) rose 25%, from 7,860 in 2021 to 9,791 in 2025. Overdose deaths per 1,000 beds rose from 14.5 to 19.6 (35%). Adding rapid rehousing beds, which grew fastest, overdose deaths per 1,000 beds still rose 25% (12.6 to 15.8).

Limit: Beds are a one-night count, not residents. HUD's count leaves out most income-restricted and recovery housing, which the county's group includes, so this compares growth rather than measuring a death rate. Beds also can't show turnover: if more people passed through each bed in a year, deaths per bed could rise without any rise in risk.

Verify: Change from 2021 to 2025: housing site deaths, county deaths and housing beds Housing site deaths per 1,000 HUD permanent housing beds, 2021-2025 Overdose deaths per bed with rapid rehousing counted Regional services database (resources.json, updated through September 2026)

Overdose deaths in King County subsidized housing rose 68% since 2021; beds rose 25%Percent change from 2021 to 2025 in deaths at King County housing service sites, countywide overdose deaths, other deaths the Medical Examiner investigated, and federally counted permanent housing beds.Overdose deaths in King County subsidized housing rose 68% since 2021; beds rose25%Change from 2021 to 2025facthub.org/n/chart/kcme-housing-deaths-vs-beds-2021-2025Source: King County Medical Examiner's Office and HUDhousing.Note: Beds are a one-night HUD count of permanent supportive and other permanent housing; they are not residents and leave out most income-restricted and recoveryOverdose deaths at housing sites68%Overdose deaths, whole county28%Permanent housing beds (HUD)25%All deaths at housing sites23%Housing-site overdose deaths, worst case19%Other investigated deaths, county-3%No change
Overdose deaths in King County subsidized housing rose 68% since 2021; beds rose 25%Percent change from 2021 to 2025 in deaths at King County housing service sites, countywide overdose deaths, other deaths the Medical Examiner investigated, and federally counted permanent housing beds.Overdose deaths in King Countysubsidized housing rose 68% since2021; beds rose 25%Change from 2021 to 2025Source: King County Medical Examiner's Office and HUDOverdose deaths athousing sites68%Overdose deaths, wholecounty28%Permanent housing beds(HUD)25%All deaths at housing sites23%Housing-site overdosedeaths, worst case19%Other investigated deaths,county-3%No change
Change from 2021 to 2025. Beds are HUD's one-night count of permanent housing beds, not residents. The worst-case bar assumes the housing group's whole rise in share came from a change in how deaths were recorded. Sources: King County Medical Examiner; HUD. View source chart
3. A change in how investigators record housing can't be ruled out Review of the county's published methods, and comparison of the housing group with deaths among people living in private homes, outdoors, in shelters or in vehicles, and with an unknown living situation, 2021 to 2025

Investigators record where each person was living, not where they died, judging from the scene and from witnesses or family rather than from a list of addresses. The county calls this hard because housing programs and locations keep changing, and it has published no change in method. Housing's share of overdose deaths among housed people rose from 21% to 29%; held at 21%, residents' overdose deaths would be up 19%, below the beds. Checks against other groups don't settle it: shelter and street deaths rose, unknowns can explain only part, and natural deaths can't be compared cleanly.

Limit: No published record shows when the county began recording housing service sites or whether earlier years were recorded the same way; before 2022 its homeless-death reports used a single no-fixed-address field. The county also doesn't say how often a visitor who dies at a housing site is recorded as a resident.

Verify: Methods Report: Documentation and Analysis of Fatal Overdose Data Did deaths move from private homes into the housing group? Did deaths move from the shelter group into the housing group? Do fewer 'unknown' living situations explain the rise? Did natural deaths shift toward the housing group the way overdoses did? Report on Deaths Among Presumed Homeless Individuals Investigated by the King County Medical Examiner, 2012-2021

4. Deaths from all causes rose about as fast as the beds Counts read by year and manner of death from the Medical Examiner's public annual-report dashboard, with the living-situation filter set to housing service site

Deaths the Medical Examiner investigated among people living in subsidized housing were 309 in 2021, 389 in 2022, 487 in 2023, 433 in 2024 and 379 in 2025. That is up 23%, about as fast as the beds, so deaths per 1,000 beds were 39.3 and then 38.7. Accidents other than traffic, mostly overdoses, rose from 129 to 200; natural deaths held between 164 and 170.

Limit: These are only deaths the Medical Examiner investigated; it does not take all natural deaths. The group mixes supportive, income-restricted and recovery housing.

Verify: Deaths at housing service sites by year and manner, 2021-2025 Housing site deaths per 1,000 HUD permanent housing beds, 2021-2025 Methods Report: Documentation and Analysis of Fatal Overdose Data KCMEO Annual Report dashboard, Notes tab (last updated October 29, 2026)

King County subsidized housing deaths rose 23% since 2021, all in accidentsDeaths the King County Medical Examiner investigated among people living at housing service sites (supportive, income-restricted and recovery housing), by year and manner of death, 2021 to 2025.{"width":960,"height":540,"plot":{"left":90,"top":106,"right":926,"bottom":462},"yDomain":[0,500],"series":[{"name":"Accident (not traffic)","color":"#0f6e99"},{"name":"Natural","color":"#b6542c"},{"name":"Other","color":"#684b9b"}],"points":[{"series":0,"x":173.6,"y":370.15,"xLabel":"2021","yLabel":"129"},{"series":1,"x":173.6,"y":253.38,"xLabel":"2021","yLabel":"164"},{"series":2,"x":173.6,"y":241.99,"xLabel":"2021","yLabel":"16"},{"series":0,"x":340.8,"y":327.43,"xLabel":"2022","yLabel":"189"},{"series":1,"x":340.8,"y":206.39,"xLabel":"2022","yLabel":"170"},{"series":2,"x":340.8,"y":185.03,"xLabel":"2022","yLabel":"30"},{"series":0,"x":508,"y":259.08,"xLabel":"2023","yLabel":"285"},{"series":1,"x":508,"y":140.18,"xLabel":"2023","yLabel":"167"},{"series":2,"x":508,"y":115.26,"xLabel":"2023","yLabel":"35"},{"series":0,"x":675.2,"y":290.41,"xLabel":"2024","yLabel":"241"},{"series":1,"x":675.2,"y":172.93,"xLabel":"2024","yLabel":"165"},{"series":2,"x":675.2,"y":153.7,"xLabel":"2024","yLabel":"27"},{"series":0,"x":842.4,"y":319.6,"xLabel":"2025","yLabel":"200"},{"series":1,"x":842.4,"y":202.83,"xLabel":"2025","yLabel":"164"},{"series":2,"x":842.4,"y":192.15,"xLabel":"2025","yLabel":"15"}]}King County subsidized housing deaths rose 23% since 2021, all in accidentsDeaths per year investigated by the Medical Examiner, by manner010020030040050020212022202320242025Accident (not traffic)NaturalOther2023 peak: 487facthub.org/n/chart/kcme-housing-site-deaths-by-yearSource: King County Medical Examiner's OfficeNote: Only deaths the Medical Examiner investigated; it does not take all natural deaths.
King County subsidized housing deaths rose 23% since 2021, all in accidentsDeaths the King County Medical Examiner investigated among people living at housing service sites (supportive, income-restricted and recovery housing), by year and manner of death, 2021 to 2025.{"width":400,"height":440,"plot":{"left":90,"top":176,"right":366,"bottom":347},"yDomain":[0,500],"series":[{"name":"Accident (not traffic)","color":"#0f6e99"},{"name":"Natural","color":"#b6542c"},{"name":"Other","color":"#684b9b"}],"points":[{"series":0,"x":117.6,"y":302.88,"xLabel":"2021","yLabel":"129"},{"series":1,"x":117.6,"y":246.79,"xLabel":"2021","yLabel":"164"},{"series":2,"x":117.6,"y":241.32,"xLabel":"2021","yLabel":"16"},{"series":0,"x":172.8,"y":282.36,"xLabel":"2022","yLabel":"189"},{"series":1,"x":172.8,"y":224.22,"xLabel":"2022","yLabel":"170"},{"series":2,"x":172.8,"y":213.96,"xLabel":"2022","yLabel":"30"},{"series":0,"x":228,"y":249.53,"xLabel":"2023","yLabel":"285"},{"series":1,"x":228,"y":192.42,"xLabel":"2023","yLabel":"167"},{"series":2,"x":228,"y":180.45,"xLabel":"2023","yLabel":"35"},{"series":0,"x":283.2,"y":264.58,"xLabel":"2024","yLabel":"241"},{"series":1,"x":283.2,"y":208.15,"xLabel":"2024","yLabel":"165"},{"series":2,"x":283.2,"y":198.91,"xLabel":"2024","yLabel":"27"},{"series":0,"x":338.4,"y":278.6,"xLabel":"2025","yLabel":"200"},{"series":1,"x":338.4,"y":222.51,"xLabel":"2025","yLabel":"164"},{"series":2,"x":338.4,"y":217.38,"xLabel":"2025","yLabel":"15"}]}King County subsidized housingdeaths rose 23% since 2021, all inaccidentsDeaths per year investigated by the Medical Examiner, bymanner010020030040050020212022202320242025AccidentNaturalOther2023 peak: 487Source: King County Medical Examiner's Officeall natural deaths.Note: Only deaths the Medical Examiner investigated; it does not take
Deaths the King County Medical Examiner investigated among people living in subsidized housing, by year and manner of death. Source: King County Medical Examiner. View source chart
5. A randomized trial found no significant difference in deaths Randomized trial evaluated by the Urban Institute

In a Denver trial of supportive housing for frequent jail users, deaths within two years were about the same whether or not people were offered housing. So more deaths where more people are housed doesn't by itself show the housing causes them.

Limit: The trial was in Denver, ended before fentanyl dominated the drug supply, treated deaths as a secondary outcome, and studied a different group than King County's residents.

Verify: Improving Health Care through Housing First: Denver Supportive Housing Social Impact Bond Initiative

Common questions 8

“Deaths the Medical Examiner investigated among King County subsidized-housing residents rose faster than the housing beds, 2021 to 2025.”

Our rating: False

False: those deaths, from all causes, rose 23% from 2021 to 2025, while federally counted housing beds rose 25%. Per 1,000 beds they were 39.3, then 38.7. Only accidental and overdose deaths outgrew the beds.

See the evidence: Deaths from all causes rose about as fast as the beds

“The rise just reflects investigators recording more deaths as subsidized housing.”

Short answer: Can't be ruled out

It can't be ruled out, and it matters: if housing had kept its 2021 share of housed people's overdose deaths, the rise would be 19%, below the beds' 25%. Investigators judge living situation case by case, and the county has published no change in method. The checks we could run don't test the key years. Deaths among people outdoors or in shelters rose 66%, which makes a large move out of that group unlikely but says nothing about private homes. Housing's share of overdose deaths among housed people held steady from 2023 to 2025, but most of its rise came from 2021 to 2023. Natural deaths, which a recording change would also move, can't be compared cleanly because of 2021's COVID-19 surveillance cases.

See the evidence: A change in how investigators record housing can't be ruled out

“Starting in 2021 makes the rise look bigger than it is.”

Short answer: No, it started earlier

No. Residents' overdose deaths were 50 in 2019, so the rise to 192 in 2025 is larger from that start, while beds grew 27%. The county hasn't said whether it recorded housing the same way in 2019. Part of any real rise may come from who moved in. New buildings opened from 2021 to 2023 took in people straight from homelessness, and residents are older and often in poor health. So deaths per bed could rise even if no one fared worse.

See the evidence: Recorded overdose deaths: 114 in 2021, 192 in 2025

“The rise comes from fewer deaths being recorded with an unknown home.”

Short answer: Only part of it

Only part. Overdose deaths with an unknown living situation fell from 36 in 2021 to 3 in 2025. Even if every one of those had been a resident, residents' overdose deaths would still be up 39%, faster than the beds.

See the evidence: A change in how investigators record housing can't be ruled out

“Counting rapid rehousing beds, the housing grew as fast as the deaths.”

Short answer: No, 68% vs 34%

No. With rapid rehousing (short-term rent help in private apartments) added, beds rose 34%, and overdose deaths per 1,000 beds still rose 25%. Rapid rehousing tenants likely count as living in private homes, so adding their beds probably understates the rate.

See the evidence: Overdose deaths outgrew the housing beds on the recorded counts

“The housing group mixes in shelters or other housing that the federal bed count leaves out.”

Short answer: Not enough to change it

Not enough to change the answer. The county puts emergency shelter in its outdoors group, not with housing, though it warns shelters and hotels may be recorded inconsistently. Even if all of the roughly 1,000 tiny-house and emergency-hotel units the region's homeless authority lists as shelter in 2025 were counted as housing, beds would be up about 38%, still well short of the 68% rise in overdose deaths. Income-restricted apartments are in the group but not the bed count; for overdoses that matters little, since the county counted 225 of 2024's 235 in permanent supportive housing.

See the evidence: Recorded overdose deaths: 114 in 2021, 192 in 2025

“Overdose deaths among subsidized-housing residents are still climbing.”

Short answer: No, down since 2023

No. Residents' overdose deaths peaked at 281 in 2023, then fell to 235 in 2024 and 192 in 2025. Deaths from all causes followed the same path, from 487 in 2023 to 379 in 2025.

See the evidence: Recorded overdose deaths: 114 in 2021, 192 in 2025

“Living in subsidized housing is causing these deaths.”

Short answer: Not shown

Not shown. A randomized trial in Denver found no significant difference in deaths with or without supportive housing, and residents' deaths from all causes rose about as fast as the housing. The counts show housing has not kept overdose deaths from rising faster than the housing grew, not that it causes them.

See the evidence: Recorded overdose deaths: 114 in 2021, 192 in 2025

Conclusion

On the county's own counts, overdose deaths among people living in King County's subsidized housing outgrew the housing. The Medical Examiner recorded 114 such deaths in 2021 and 192 in 2025, up 68%, while federally counted beds rose 25%. The result holds with rapid rehousing beds counted, and with the drop in deaths of unknown living situation set aside.

What keeps this from True is how the counts are made. Investigators judge each person's living situation case by case, and the county hasn't said when it began recording these sites or whether 2021 was recorded the same way. Housing's share of overdose deaths among housed people rose from 21% to 29%. If all of that came from a recording change, residents' overdose deaths would be up 19%, slower than the beds. We found no record of a change. But the county's reports through 2021 used a single no-fixed-address field, so its system may have changed right at the start of the period. Our checks can't test 2021 to 2023, when most of the shift happened, and they all use the same Medical Examiner counts; only the bed count is independent.

We rate the claim Mostly true rather than Mixed, a close call. Mixed would mean a significant part of the claim is shown to be wrong, and nothing shows that: reversing the finding would need about 50 of the 56 shifted deaths to be recording changes. The open question is one we can't test, which is why our confidence is moderate, not high.

The broader claim needs care. More residents did die: deaths the Medical Examiner investigated rose from 309 to 379. But they rose 23%, about as fast as the beds, so they didn't outgrow the housing. We can't say why overdose deaths did; some of it may reflect who moved in, people straight from homelessness who face high overdose risk. Resident counts and a statement of method from the county would settle it.

How we checked
  1. We defined the claim as the change from 2021 to 2025 in overdose deaths the King County Medical Examiner recorded among people whose presumed living situation was a housing service site, compared with the change in HUD's count of permanent supportive and other permanent housing beds in Seattle/King County.
  2. We read the county's fatal overdose dashboard and the Medical Examiner's annual-report dashboard by hand on November 7, 11 and 12, 2026, by year and manner of death, and checked that each year's manners add up to the dashboard's total.
  3. We repeated the comparison with rapid rehousing beds added, and with every drop in deaths of unknown living situation counted as a resident.
  4. To test whether recording changed, we read the county's methods report and its overdose fatality reviews, and compared the housing group year by year with deaths among people living in private homes and among people living outdoors, in shelters or in vehicles, for overdoses and for natural deaths.
  5. We used the Denver randomized trial only to judge whether housing causes deaths, not to measure the King County trend, and the county's methods report and overdose reviews only to judge how living situation is recorded.
  6. We recorded each limit beside the evidence it affects; all tables and calculations are in the linked dataset.

Editorial note: This check covers overdose deaths, and deaths from all causes, that the King County Medical Examiner recorded among people living in subsidized housing, 2021 to 2025. It does not rate any housing provider or program, and it does not estimate death rates per resident. No external specialist has reviewed it.

Expert-review status

No external specialist review has been recorded.

Funding and conflicts

Funding: No external funding for this piece.

Ownership: Published by the FactHub Research Team.

Conflicts: None known.

Sources 16

Sources are labeled by kind, with what this fact check takes from each, so readers can check the evidence independently.

  1. FactHub record FactHub dataset. Change from 2021 to 2025: housing site deaths, county deaths and housing beds

    Saved analysis comparing 2021 with 2025.

    • Overdose deaths among housing-site residents rose 68% (114 to 192) from 2021 to 2025; overdose deaths countywide rose 28% (714 to 914).

      Not yet checked

    • HUD permanent supportive and other permanent housing beds in Seattle/King County rose 25%, from 7,860 in 2021 to 9,791 in 2025.

      Not yet checked

  2. FactHub record FactHub dataset. Housing site deaths per 1,000 HUD permanent housing beds, 2021-2025

    Saved analysis dividing deaths by HUD beds.

    • Overdose deaths per 1,000 HUD permanent housing beds rose from 14.5 in 2021 to 19.6 in 2025, and accidental deaths other than traffic from 16.4 to 20.4.

      Not yet checked

    • Qualifies: Deaths from all causes per 1,000 HUD permanent housing beds were 39.3 in 2021 and 38.7 in 2025.

      Not yet checked

  3. FactHub record FactHub dataset. Overdose deaths per bed with rapid rehousing counted

    Saved analysis adding rapid rehousing beds to the bed count.

    • With rapid rehousing beds added, beds rose 34% (9,067 to 12,179) and overdose deaths per 1,000 beds rose 25% (12.6 to 15.8) from 2021 to 2025.

      Not yet checked

  4. FactHub record FactHub dataset. King County overdose deaths by where people lived: deaths compared with HUD beds

    Saved analysis comparing overdose deaths in subsidized housing with HUD beds since 2019.

    • HUD permanent supportive and other permanent housing beds grew 27% (7,718 to 9,791) from 2019 to 2025, while overdose deaths in this housing grew 3.8 times (50 to 192).

      Not yet checked

  5. FactHub record FactHub dataset. Deaths at housing service sites by year and manner, 2021-2025

    Saved analysis of the Medical Examiner's dashboard counts by year and manner of death.

    • Qualifies: Deaths the Medical Examiner investigated among people living at housing service sites were 309 in 2021, 389 in 2022, 487 in 2023, 433 in 2024 and 379 in 2025.

      Not yet checked

    • Qualifies: Accidents other than traffic rose from 129 in 2021 to 200 in 2025; natural deaths were between 164 and 170 each year, and 164 in both 2021 and 2025.

      Not yet checked

  6. FactHub record FactHub dataset. Do fewer 'unknown' living situations explain the rise?

    Saved analysis testing the rise against the drop in deaths recorded with an unknown living situation.

    • Qualifies: Overdose deaths with an unknown living situation fell from 36 to 3 between 2021 and 2025. If every one that left the unknown group was a resident, residents' overdose deaths would still be up 39%; split proportionally, 60%.

      Not yet checked

  7. FactHub record FactHub dataset. Did deaths move from the shelter group into the housing group?

    Saved analysis comparing deaths among people living outdoors, in shelters or in vehicles with deaths at housing sites, 2021 and 2025.

    • Deaths among people living outdoors, in shelters or in vehicles rose 66% (188 to 312) from 2021 to 2025, so that group grew rather than shrank as housing-site deaths rose.

      Not yet checked

  8. FactHub record FactHub dataset. Did deaths move from private homes into the housing group?

    Saved analysis comparing deaths among people living in private homes with deaths at housing sites, 2021 and 2025.

    • Qualifies: Private-home residents' overdose deaths rose 7% (439 to 469) from 2021 to 2025, so housing's share of housed people's overdose deaths rose from 21% to 29%. Had it stayed at 21%, residents' overdose deaths would be up 19% (114 to 136).

      Not yet checked

  9. FactHub record FactHub dataset. Did natural deaths shift toward the housing group the way overdoses did?

    Saved analysis comparing the housing group with private homes, year by year, for natural and overdose deaths.

    • Qualifies: Overdose deaths among housing-site residents were 114 in 2021, 177 in 2022, 281 in 2023, 235 in 2024 and 192 in 2025.

      Not yet checked

    • Qualifies: From 2023 to 2025, housing's overdose deaths per private-home overdose death held near 0.42 and its natural-death ratio fell slightly. For 2021 the natural-death test is unclear, because COVID-19 surveillance cases inflate private-home natural deaths.

      Not yet checked

  10. Government Seattle Office of City Auditor. Overdose and Crime Concentrations audit (July 9, 2024)

    The Seattle city auditor's review of where overdoses and crime concentrate.

    Reliability: An independent city audit office reviewing city and county data.

    • Qualifies: Supportive housing residents are older and high-risk: 95% of one operator's residents were 50 or older.

      What the source says
      among residents of Plymouth Housing's permanent supportive housing (n=146), 95 percent were aged 50 and up
      Footnote 4

      Not yet checked

  11. Government Public Health - Seattle & King County. King County Overdose Fatality Review 2024 Report

    The county's annual review of overdose deaths, with a section on deaths in permanent supportive housing.

    Reliability: Prepared by the county health department with the Medical Examiner; it reports the county's own counts, using a narrower group than the dashboard's housing service sites.

    • The county counted 225 overdose deaths in permanent supportive housing settings in 2024.

      What the source says
      In 2024, there were 225 fatal overdoses in permanent supportive housing settings.
      Page 10, Fatal Overdoses in Permanent Supportive Housing

      Not yet checked

  12. Government Public Health - Seattle & King County. Methods Report: Documentation and Analysis of Fatal Overdose Data

    The county health department's description of how presumed living situation is assigned.

    Reliability: Written by the agency that publishes the dashboards; it documents its own category definitions and their limits.

    • Qualifies: The 'housing service site' category includes permanent supportive housing, income-restricted units and recovery housing.

      What the source says
      Housing service site: This category includes permanent supportive housing, income-restricted units, recovery housing.
      Presumed Living Situation section

      Not yet checked

    • Qualifies: Medical Examiner investigators assign each person's living situation from where they were living at death, based on the scene and on witnesses or next of kin.

      What the source says
      categorize living situation based on where the decedent was living at the time of death
      Presumed Living Situation section

      Not yet checked

    • Qualifies: The county says assigning living situation is complicated by scarce or conflicting information and by housing programs and locations that keep changing.

      What the source says
      complicated by scarcity of information, conflicting information, and dynamic nature of housing service programs and locations
      Presumed Living Situation section

      Not yet checked

    • Qualifies: The county warns that some living situations, such as shelters and hotels, may be inconsistently categorized.

      What the source says
      some housing situations, like staying at a shelter, in a hotel/motel, in an RV, and couch-surfing may be inconsistently categorized
      Presumed Living Situation section

      Not yet checked

    • Qualifies: The county's outdoors group, not its housing service site group, includes emergency shelter.

      What the source says
      includes encampments, vehicles, abandoned buildings, parks, other outdoor locations, emergency shelter
      Presumed Living Situation section

      Not yet checked

  13. Government King County Regional Homelessness Authority. Regional services database (resources.json, updated through September 2026)

    The regional homeless authority's list of shelter and housing sites, with each site's type and number of units.

    Reliability: The public authority that funds and coordinates the region's homeless services; it types each site itself.

    • Qualifies: The authority's database lists 22 tiny house villages with 834 units, and types the Mary Pilgrim Inn (80 units) and The Gateway (120 units) hotels as emergency shelter, not permanent supportive housing.

      What the source says
      "projectType": "tiny house village"
      Entries with projectType "tiny house village" (22 sites), and the Mary Pilgrim Inn and The Gateway entries (projectType "emergency shelter")

      Not yet checked

  14. Government Public Health - Seattle & King County. Report on Deaths Among Presumed Homeless Individuals Investigated by the King County Medical Examiner, 2012-2021

    The county's 2022 report on deaths of people presumed homeless, with a description of the data field it used.

    Reliability: The county's own description of the Medical Examiner's records.

    • Qualifies: For deaths through 2021, the county's homeless-death report drew on a single 'no fixed address' field in the Medical Examiner's database, not the living-situation groups used now.

      What the source says
      individuals recorded as homeless under the 'no fixed address' field in their database
      Methods section, page 2

      Not yet checked

  15. Government King County Medical Examiner's Office. KCMEO Annual Report dashboard, Notes tab (last updated October 29, 2026)

    The dashboard's notes on what its counts include.

    Reliability: The Medical Examiner's own description of its public dashboard; it states the counts cover only deaths it investigated.

    • Qualifies: The Medical Examiner did not take jurisdiction of all natural deaths.

      What the source says
      KCMEO did not take jurisdiction of all natural deaths.
      Notes tab, Overview section

      Not yet checked

  16. Other Urban Institute. Improving Health Care through Housing First: Denver Supportive Housing Social Impact Bond Initiative

    Working paper evaluating a randomized trial that offered supportive housing to frequent jail users in Denver.

    Reliability: The Urban Institute is a nonprofit research organization. The evaluation was funded by the Robert Wood Johnson Foundation; the program itself was funded by social impact bond investors and the city.

    • Qualifies: Deaths within two years were about the same for people offered supportive housing and for the control group.

      What the source says
      Mortality was about the same in both the treatment and control groups (table 5).
      Mortality section and Table 5, p. 13

      Not yet checked

  17. FactHub dataset Deaths the King County Medical Examiner investigated at housing service sites, 2021-2026, by manner of death

    • Tables: context-by-year

      Saved analysis change-2021-2025

    • Tables: deaths-by-year

      Saved analysis deaths-by-year-manner

    • Tables: private-residence-deaths, deaths-by-year

      Saved analysis natural-death-recoding-test

  18. FactHub chart kcme-housing-deaths-vs-beds-2021-2025
  19. FactHub chart kcme-housing-site-deaths-by-year

Update history 9 changes, last

  1. Version 5: changed the claim checked from whether more residents are dying (deaths from all causes rose about as fast as the housing) to whether overdose deaths rose faster than the housing, the finding the evidence supports, and rated it Mostly true because a change in how investigators record where people lived can't be ruled out. Added the county's methods for assigning living situation, its count of overdose deaths in permanent supportive housing, a comparison counting rapid rehousing beds, a year-by-year comparison of natural and overdose deaths with private homes, and a common question on the 2021 starting year. The lead chart is now overdose deaths against bed growth, with a worst-case bar. Moved citations to dataset version 6 and chart versions 4.
  2. Version 6: shortened the title, search title, summary and one key point, and split long sentences.
  3. Version 7: rated the broader claim that more residents are dying than the housing's growth explains (False) as the lead common question (feedback 2516078371000479944-9a2d6d639f4042a505f7e864ae24135e); removed agreement across comparisons as a reason for the rating, since they share the same Medical Examiner counts (feedback 2516078370993708396-22ad042e50db87190bfb77b0feeeb257); softened the shelter-group check and added whether the housing group includes shelters the bed count leaves out, with the county's category definitions and the regional homeless authority's site list (feedback 2516078370995953769-be2ab1d40930164b131db493895fb9fd); noted that the county's recording system may have changed before 2022 and called the rating a close call (feedback 2516078370998245374-1354f4652cb0b5abe94fcdffc9bacaa7). Moved citations to dataset version 7 and chart versions 5.
  4. Version 8: reworded the lead common question so the evidence settles it as worded (feedback 2516078364555363453-fc531b666876c346600b5479247f4bfa); the summary now compares the worst-case rise in deaths with the beds' rise, not deaths per bed (feedback 2516078364560490050-5833a030ef842a3036b9387ea4cd66f7); the conclusion explains why the untestable years still favor Mostly true over Mixed (feedback 2516078364563588415-9aa34128665c80c59607b26bb308a9cb); added that investigators record where people lived rather than where they died, that the county doesn't say how visitors are recorded, and that beds can't show turnover (feedback 2516078364557913587-450a80850067ef65f99b69268421afeb).
  5. Version 9: the lead common question now names deaths the Medical Examiner investigated, so the evidence settles it as worded (feedback 2516078361808359031-e7610441ffbf5618ca05011e832a8ff9); the context and conclusion now say more residents did die, but not faster than the housing grew (feedback 2516078361812796733-274c2a93ad3246a68609f5bd0ed35fa8).
  6. First version.
  7. Version 2: added a common question on whether deaths were recoded from the shelter group into the housing group, answered with Medical Examiner counts for people living outdoors, in shelters or in vehicles (feedback 2516078503403756909-f2dc75ce5c2b308b8cdfcf4fef2e3833); shortened the lead; moved citations to dataset version 4.
  8. Version 3: added a common question on whether deaths moved into the housing group from private homes because subsidized addresses were identified more often, answered with Medical Examiner counts for people living in private homes and a worst-case bound (feedback 2516078450895910267-cb001529d3690faeab6001f13ad2536e); lowered confidence from high to moderate because that shift can't be ruled out; added accidental deaths per 1,000 beds; moved citations to dataset version 5 and chart versions 3.
  9. Version 4: merged the two questions on recoding from other groups into one, and the two questions on what the counts cover into one, leaving eight common questions.

Feedback these versions applied

Cited by

Version details

The exact citing versions, and the version of this article each cites.

  • Article king-county-stay-housed-rates-deaths — version 7 (published) cites version 9

For developers: ClaimReview data

This page includes standardized ClaimReview metadata describing the exact claim, FactHub's finding, publisher, and review dates. You can inspect the record below and confirm that it matches the article you just read.

View ClaimReview JSON 
{"@context":"https://schema.org","@type":"ClaimReview","url":"https://facthub.org/n/article/king-county-subsidized-housing-deaths-since-2021","datePublished":"2026-11-12","dateModified":"2026-11-12","author":{"@type":"Organization","name":"FactHub","url":"https://facthub.org/"},"claimReviewed":"Overdose deaths among people living in King County\u0027s subsidized housing have risen faster than the housing since 2021.","reviewRating":{"@type":"Rating","alternateName":"Mostly true"},"image":"https://facthub.org/node-assets-v2/article/king-county-subsidized-housing-deaths-since-2021/9/westlake/social-1200x630.jpg"}

This metadata helps machines discover the fact-check. It does not prove the conclusion or guarantee special treatment by a search engine.

Challenge this fact-check

Every source above is linked so you can check it yourself. Identify the specific conclusion you dispute and provide verifiable evidence that could change it. Credible challenges deserve review and a visible response.

Submit evidence Send general feedback instead

Technical details: node handle article:king-county-subsidized-housing-deaths-since-2021