evidoria

← Back to browse

Good practice Imported

LA County's Homelessness Prevention Unit — A UCLA California Policy Lab Model Ranks 90,000 People to Target $26 Million in Aid

United States of America · Los Angeles · See the United States of America profile · See the Los Angeles profile

Evidence: Quasi-experimental Top 24% 73/100 · Ask Evidence Copilot about this practice

Since 2021, Los Angeles County's Homelessness Prevention Unit has used a California Policy Lab model built on 580 administrative-data points to rank around 90,000 people by homelessness risk, targeting $26 million in ARPA-funded aid at the households most likely to lose housing.

$26 million
Program funding
~90,000
People ranked by risk model
3.5x
Model accuracy vs. random selection
-71%
Reduced need for street outreach/shelter within 18 months (pilot) (May 2022-Feb 2023)
86% of 700+
People served, remaining stably housed (as of Apr 2024)
LA County's Homelessness Prevention Unit — A UCLA California Policy Lab Model Ranks 90,000 People to Target $26 Million in Aid

Details

Maturity
Scaling
Promoter
Los Angeles County Department of Health Services & California Policy Lab at UCLA
Period
2021-2026
Keywords
homelessness prevention, social services, predictive analytics, public health

Context

Los Angeles County's Department of Health Services launched its Homelessness Prevention Unit in 2021, funded with $26 million from the American Rescue Plan Act, to find people at imminent risk of losing housing before they become homeless.

Activities

The California Policy Lab at UCLA built a predictive model using 580 administrative-data factors from health, mental-health, social-services, criminal-justice and hospital records to rank roughly 90,000 people by risk; a dedicated outreach team then contacts and enrolls prospective participants, offering financial assistance and case management.

Results

The model is reported as 3.5 times more accurate than random selection; in a May 2022-February 2023 pilot, enrolled participants were 71% less likely to need street outreach or shelter within 18 months than similar high-risk people not enrolled. By April 2024 over 700 people had been served, with 86% remaining stably housed.

Conclusions

A 2024 fairness review found no significant performance differences across race, ethnicity or gender (slightly more accurate for Black residents). A more rigorous RCT was underway as of 2025-2026 reporting; county-wide homelessness fell 4% in the same period, but that trend reflects many programs, not the Unit alone.

Implementation
Indicative cost
High (€500k–€5M)
Time to results
Long (> 3 years)
Staffing & skills
LA County Department of Health Services, California Policy Lab at UCLA (predictive model), outreach and case-management team (grew from 4 to 20 staff)

Conditions for success

  • 580-factor administrative-data model with periodic fairness review
  • multi-channel outreach (phone, letter, email) that raised enrolment 67%

Common failure modes

  • county-wide homelessness trend cannot be attributed to the Unit alone
  • rigorous RCT results not yet available at time of reporting

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful