Preventing homelessness: the tough job of predicting who is at real risk

When the unknown number popped up on her phone, Jocelyn Escanuela was in the checkout line at Walmart. She still can’t explain why she picked up and then listened to a cold-caller’s pitch that sounded a lot like a scam.

She had been selected to receive a grant of $6,000, the caller told her. And she would have a personal assistant to help her get her through her “crisis.”

How did they even know she was in a crisis?

It turned out the caller was legitimate. She was from the Homelessness Prevention Unit, an experimental Los Angeles County program that is testing whether it is feasible to stop homelessness before it starts — one person at a time — by picking them out of mountains of data.

Escanuela’s crisis was detected not by a person but a predictive statistical model that was developed to solve a conundrum that has made homelessness prevention a tantalizing but underused strategy.

Despite sound evidence that services such as eviction defense and financial assistance can prevent people from becoming homeless, it’s impossible to know after the fact whether any given person would have become homeless without the help. Research has shown that only a small percentage would. The elusive goal of prevention is to identify that small percentage.

“With limited prevention resources to work with, there are real consequences to not getting them to the people who need them most,” said Steve Berg, chief policy officer for the National Alliance to End Homelessness, which has historically frowned on costly prevention programs.

But Berg said “it would be good news if these emerging technologies turned out to be effective at predicting who’s most likely to become homeless if they don’t get help.”

Attaining that elusive precision will be increasingly important as both the city’s ULA “mansion tax” and the countywide Measure A sales tax begin to direct millions of dollars into homelessness prevention.

The model that picked Escanuela as high risk is being tested to see how effective it is.

It was created by the California Policy Lab at UCLA, a research institute that has access to data from county agencies such as the departments of health and social services, which interact with people at their most vulnerable. The Policy Lab sifts through all that data, evaluating some 500 markers to generate a list of individuals and families that its model predicts to be at high risk of becoming homeless. It turns that list over to the Homelessness Prevention Unit and its Housing Stabilization Team.

“We meet people when they have just gotten out of the hospital, we meet people when they have just lost a job,” said Dana Rae Vanderford, associate director of the Homelessness Prevention Unit. “We meet people when they have lost a family member who was the sole provider. We meet people as they are receiving verbal eviction warnings from their landlord.”

The Homelessness Prevention Unit analysts randomly work their way through the names on the high-risk list to come up with two groups of candidates. Half will be offered intervention — a cash stipend and a case manager for four months. The other half will receive nothing and never know they were chosen, but will be monitored through any contacts with county or homeless agencies they make.

Escanuela landed in the target group — the fortunate half — of the random clinical trial.

The holy grail of prevention would be a model that could pinpoint those who would become homeless, and avoid spending money on those who never would.

In a 2023 report, Notre Dame University’s Lab for Economic Opportunities found that people served by a Santa Clara County prevention program were nearly 80% less likely than a control group to become homeless after receiving services.

That’s not as impressive as it sounds because only 4.1% of those who got no help became homeless, suggesting that a lot of money was invested in people who wouldn’t have become homeless without the help.

“Prediction is possible, even if it’s not great,” said Vanderbilt University research professor Beth Shinn, who studied New York City’s Homebase prevention program.

Her research found that a model did moderately better than outreach workers at predicting.

“Even with the way the city was doing it, it was cost-effective and moderately successful,” Shinn said.

Those studies involved people who sought prevention services. The Policy Lab and Homelessness Prevention Unit are taking the next step, using predictive analytics to find people who have not sought out services.

Early findings are promising. In the data used to construct the model, about 47,000 people receiving county services, 24% of those predicted to be at high risk actually became homeless compared with only 7% of the whole sample.

It’s also proved effective at finding people who are likely to become chronically homeless.

“Our clients are living with really high levels of risk,” Vanderford said. “They have complex health and mental health conditions. They are meeting us at a real moment of crisis. The timing with which we reach out to our clients seems magical to me.”

Full results of the trial will not be final until 2027 after a sufficient number of people have been tracked for 18 months after completing the four-month program.

The Homelessness Prevention Unit was created with funding from the American Rescue Plan Act supplemented by county funds. It has about 250 active clients and, with a turnover of four to six months, can handle 750 a year. About 90% retained housing or found new homes, Vanderford said.

It’s labor-intensive work. Four analysts go through the raw lists randomly screening out ineligible candidates. Because there is a delay before the Policy Lab obtains the county data, many on the final list are already homeless, proving the predictions accurate.

“There is a real challenge in getting in touch with people,” Vanderford said. “Phones go off. A client may be hospitalized or in jail. Clients might be mistrustful of getting this call out of the blue that sounds a little too good to be true. Voicemails go unresponded to.”

“I never answer calls like that,” Escanuela said. “I don’t know what compelled me to answer.”

Neither Escanuela nor Vanderford know what specific factors placed her on the high-risk list except that she was accessing county services.

But the call was timely. She and her family were in a protracted eviction battle, and she feared becoming homeless again.

As a child, she said, she had a long spell of spending days in a park and nights in a church and later lived at Union Rescue Mission downtown and in its Hope Gardens shelter in Sylmar.

“I didn’t want to fall back into that, especially as an adult,” she said.

Once enrolled, Escanuela got a call from Chris Schuchert, one of the program’s 20 contract case managers. Over the four months of the program and a two-month extension, they communicated by phone and text. He helped in various ways, from getting her a debit card for groceries to addressing her emotional needs.

“Chris was able to find me a therapist,” Escanuela said. “I was going through so much at the time I just needed someone to talk to.”

The program’s case managers handle clients’ expenses, making direct payments to vendors and landlords. The case manager has to make a requisition, and that often follows a negotiation with the client. Status items such as $250 shoes are talked down to the $50 model, but an $800 bed for better sleep could be approved, Schuchert said.

Case managers make referrals to health and mental health agencies. And when clients are thousands of dollars in arrears in rent, they can refer them to Stay Housed LA and other groups that help distressed tenants.

In Escanuela’s case, that wasn’t necessary because the landlord had stopped accepting rent during the eviction process. Schuchert thought it would be best for her to leave the apartment she shared with her mother and brother to avoid an eviction on her record.

Today, she lives with her mother in a pleasant apartment in Pomona. She had saved enough money to pay move-in costs and buy equipment for a home business offering eyelash services.

Business is good enough, she said, that she’s paying her own way.