Adults in Alameda County's Homelessness System

Section 1

Who Is in the System

This analysis draws on Alameda County HMIS data covering adults ages 18–59 who accessed homelessness services between January 2023 and April 2026. The dataset captures 16,420 total program enrollments across 5,222 unique individuals — including disability status, chronic homelessness designation, income, and health coverage at intake.

The profile of this population is not one of people on the edge who need a short bridge. Eighty-six percent have a documented disability. Nearly six in ten were already classified as chronically homeless the moment they entered the system. These characteristics point toward a need for long-term, wrap-around support rather than short-term crisis intervention.

Key insight

2,451 individuals (47%) are still active in the system with no exit date recorded as of April 2026. Monthly entries rose from ~470/month in early 2024 to 600–660/month by mid-2025. Demand is not stabilizing.

Metric	Figure
Total unique individuals	5,222
Total program enrollments	16,420
Avg program touchpoints per person	3.1
Still active (no exit recorded)	2,451 (47%)
Have a disability at intake	~4,487 (86%)
Chronically homeless at entry	~3,072 (59%)
No income at intake	~1,027

Section 2

How Long People Stay

Length of stay is measured from each person's first recorded program entry to their last exit — or to April 2026 if still active. This captures the full arc of each person's journey, including all cycling between programs.

Time in system — unique individuals

From first entry to last exit (or April 2026 if still active)

Median months by program type

Per individual enrollment (not total system time)

Note: Simultaneous enrollment in multiple programs is common. These figures reflect individual program duration, not total time in system. A PSH median of 15 months reflects stable occupancy, not churn.

2,569 people — nearly half — have been in the system for more than a year

Of those, 1,353 have been involved for two years or more. These are not short-term crisis situations. They are chronic conditions being managed, incompletely, by a system built for emergencies.

Section 3

Concurrent vs. Sequential Enrollment

The average person touches 3.1 programs. But are they moving from one to the next, or stacking multiple programs at once? The answer is mostly both.

Pattern	Clients	Share
Only concurrent (multiple programs at the same time)	1,099	33%
Only sequential (one ends, then another begins)	323	10%
Both patterns at different points	1,859	57%

The dominant pattern is layered, not linear. Most people are enrolled in multiple programs simultaneously — someone in Emergency Shelter may also be enrolled in Coordinated Entry and receiving Services Only case management at the same time. The system creates layers, not clean handoffs.

Cost implication

Concurrent enrollment means the county is often funding multiple programs for the same person at the same time. When one person is simultaneously in Emergency Shelter, Coordinated Entry, and Services Only, the county is paying three program costs for one unhoused individual.

Section 4

How People Move Between Programs

When people do move sequentially from one program to the next, the flows reveal a system cycling on itself rather than moving people toward stability. The diagram below shows sequential transitions — one program ends, another begins.

Program flow diagram showing churn cluster on left and stable exits on right

Interactive transition flow

Sequential program-to-program moves · Sources on left · Destinations on right · Hover ribbons for detail

Node sub-labels show count and % of all transitions shown

The churn loop

The top two transitions alone — Services Only → Coordinated Entry (837) and back again (393) — account for 1,230 moves. These are not steps toward housing. They are the same loop, repeated.

From	To	Count
Services only	Coordinated entry	837
Coordinated entry	Services only	393
Services only	Emergency shelter	322
Emergency shelter	Services only	290
Street outreach	Services only	267

The narrow exit

Across the entire dataset, only 233 sequential transitions ended in PSH. Coordinated Entry — the program explicitly designed to connect people to permanent housing — sent 393 people back to Services Only and 159 to Emergency Shelter for every 37 it moved to PSH.

The gateway doesn't open

This is not a flow problem. It is a capacity problem. There are not enough PSH units for Coordinated Entry to do its job.

Section 5

Client Journeys: Three Groups

Looking beyond transitions to individual clients reveals three distinct populations in the data. Each tells a different story about the system's effectiveness.

Group A 635 12.2% of all clients

Entered the system through a non-housing program and later transitioned into PSH or Rapid Re-Housing. These are the clients who successfully navigated the system to reach stable housing.

Group B 678 13.0% of all clients

Were already in PSH or RRH as their first recorded enrollment — stably housed when the dataset begins. 61% are still in PH with no exit recorded. Only 3% returned to homelessness after exiting PH.

Group C 3,909 74.9% of all clients

Never reached PSH or RRH during the observation window. Of these, 3,081 exited the system entirely without a housing placement. Their destinations are unknown — the data does not include exit destination fields.

The 25% vs 12.2% distinction matters

An earlier figure of "25% reached housing" combined Groups A and B. Group B clients were already stably housed when the data begins — they didn't transition into housing, they were already there. The 12.2% (Group A) is the honest measure of the system moving people from homelessness into housing during the observation window.

Group A: How they got there

Of the 635 clients who transitioned into housing, 376 (59%) used Coordinated Entry as a gateway at some point before reaching housing. Services Only was the most common starting point.

Group A — journey to housing (635 clients)

First program entered · gateway used · housing destination

Node labels show client count and % of Group A total

Group B: What happens to already-housed clients

The 678 clients who entered the dataset already in PSH or RRH are mostly a stable population.

Still in PH — no exit recorded412 (61%)

Exited PH, no return to homelessness304 (45%)

Exited PH, returned to homeless programs23 (3%)

PSH works — for those who get it

255 of the 678 access other programs (services, case management) during their PH period — appropriate concurrent enrollment. The 3% return rate is evidence that stable housing holds. The problem is scale, not model.

Group C: The unknown 74.9%

3,909 clients never reached PSH or RRH. Of the 3,081 who exited the system, their destinations are unknown because this dataset does not include exit destination fields. They may have resolved housing independently, moved out of county, or left while still homeless. The 828 still active are enrolled in non-housing programs with no exit recorded.

Critically: 346 of Group C clients did have a PSH or RRH enrollment at some point — they reached housing, then later re-entered non-housing programs. These represent a small but important population of housing exits followed by system re-engagement.

Section 6

Time to Permanent Housing

Two ways to measure time-to-housing produce very different pictures. The first (Group A only) shows how long it took the clients who succeeded. The second (all clients, survival curve) shows the full population — including the 75% who never reached housing at all.

Among clients who reached housing (Group A)

Pathway	N	Median	Mean	90th pct
Any PH (PSH or RRH)	635	10 mo	12.1 mo	25 mo
PSH only	459	13 mo	14.0 mo	26 mo
RRH only	252	6 mo	8.7 mo	20 mo

MH-positive clients wait 1–2 months longer on every pathway. The difference is real but modest. One quarter of PSH clients waited 21+ months. One in ten waited 26+ months — nearly two and a half years from first contact to stable housing.

Important: these are the success cases only

These medians describe only the 635 clients who made it. The full population picture — including those who never reached housing — looks very different. See the survival curve below.

Survival curve: probability of not yet being housed

All 5,222 clients included · Clients who never reached housing shown as the flat tail · Based on Kaplan-Meier method

All clients MH-positive Non-MH — dashed = Non-MH · flat tail = never housed

The curve flattens where housing placements stop. The gap between the final value and 0% represents clients never housed during the observation window. MH-positive clients (orange) reach housing at a higher rate than non-MH (green), consistent with PSH targeting people with disabilities.

The honest picture: ~88% of clients never reached stable housing

At 36 months from first entry, the survival curve for all clients settles at ~79% still unhoused. The flat tail is not a slow decline — it's the structural reality of a system with far more need than PSH capacity.

Section 7

Story Angles for SHCLA

The data supports several narratives depending on audience and purpose. Each is grounded directly in the analysis above.

Angle 1

The Revolving Door

The average adult touches 3.1 programs, and 57% experience both concurrent enrollment and sequential cycling at different points. The county is funding the same person across multiple programs simultaneously, month after month. PSH breaks that pattern: 81% of PSH residents are stably housed with no exits recorded. Only 3% of Group B clients ever returned to homelessness.

Angle 2

The Long Tab

1,353 adults have been in the system 2+ years. Emergency shelter runs ~$60–100/night. Concurrent enrollment means the county often pays three program costs for one unhoused individual. PSH at a fixed per-unit monthly cost is not an additional expense — it is a substitution for a more expensive, less effective status quo.

Angle 3

The Gateway Doesn't Open

Coordinated Entry moved 37 clients to PSH. It sent 393 back to Services Only and 159 to Emergency Shelter. That is a capacity problem, not a process problem. There are not enough PSH units for CE to do its job. Every person who cycles back after a CE assessment is a person the system assessed, flagged, and had nowhere to send.

Angle 4

Already Here

5,222 adults. 86% with a disability. 59% chronically homeless at intake. 2,451 still active today. The question is not whether this population needs long-term support — the data answers that. The question is what kind of support produces the best outcome per dollar. The survival curve shows what happens without it: 88% never reach housing.

Angle 5

Follow One Person

A fictionalized composite: an adult with a disability enters Coordinated Entry, moves to Emergency Shelter for 2 months, gets referred to Services Only (running concurrently with a new CE enrollment), waits on a PSH list, and lands in stable housing 14 months after first contact — touching 3+ programs, generating 3+ program costs. Month 15: still housed. One program. Stable.

Angle 6

Demand Is Growing

Monthly entries jumped from ~470/month in early 2024 to 600–660 by mid-2025. The system is absorbing more people, not fewer. At a moment when need is rising, the county's highest-return investment is expanding what is already proven to work — models that combine stable housing, supportive services, and community programming for adults who would otherwise remain in costly, high-churn emergency settings.

Data note: Source: Alameda County HMIS via UC Berkeley data request. Dataset covers adults ages 18–59, January 2023–April 2026. 621 enrollments share identical client ID, program start month, and project type — potential duplicate records. The figure of 5,222 unique individuals is based on distinct Client IDs and is considered conservative. Exit destination fields are not present in this extract, meaning the outcomes of the 3,081 "exited without housing" clients cannot be confirmed from this data alone.