Preventing homelessness with help from a computer model

When her phone rang in February, Mashawn Cross was skeptical of the soft voice offering help on the end of the line.

“You said you’re doing what? And who are you with?” that The 52-year-old remembered the statement.

Cross, who doesn’t work because of her bad back and knees, made ends meet on about $200 a month in aid and whatever she could make from recycled bottles and cans. Her gas and electric bills chewed up her checks. She’d been in and out of the emergency room, her doctor said she might need a colostomy bag, and depression plagued her day in and day out.

Kourtni Gouché listened and began to help. The LA County clerk helped Cross get household items so she could save money and cover her utility bills. She offered to get her a new bed to soothe her aching back. She began associating Cross with programs to relieve her depression and stop smoking, something Cross had long wanted but struggled to do.

“I feel like I have a friend here,” Cross said, occasionally tearing up as she cheered up the case officer who had repeatedly advocated for her. In her South LA duplex apartment, over the hum of a fan, she suddenly remembered a question she’d forgotten to ask Gouché during their regular conversations.

“How did you even come up with my name?” Cross asked.

The answer is an unusual mobilization of data analysis to try to prevent homelessness before it starts.

Cross is part of a rare attempt by LA County to unify predictive modeling – a tool for predicting events by tracking patterns in current and historical data – with the deeply personal work of homelessness prevention.

Workers visit woman in her house.

Cross, left, with Vanderford and Gouché.

(Robert Gauthier/Los Angeles Times)

The county found Cross and numerous other people through a prediction tool developed by UCLA researchers that pulls data from eight LA ​​County agencies to help outreach workers focus their attention and support on people they think they may be are most at risk of losing their homes.

LA County is struggling to keep up with the number of people who become homeless each year, even as it ramps up efforts to help people find housing. Figuring out who to help is crucial as millions of residents appear vulnerable yet avoid homelessness, said Janey Rountree, founding director of California Policy Law at UCLA.

“You would never have enough money to prevent everyone who seems at risk,” Rountree said. “You really need a different strategy to figure out who’s actually going to be homeless if they don’t get help right away.”

Researchers have found it surprisingly difficult to guess who will slide into homelessness and who will avoid it: In a report published three years ago, the California Policy Lab and the University of Chicago Poverty Lab said that a decent prediction would require at least 50 factors — and that the best models would need “somewhere between 150 and 200”.

The prediction model now in use in LA County uses an algorithm that the UCLA team says contains about 500 features.

It pulls data from eight county agencies to determine exactly who needs help and examines a wide range of data across county systems: who ended up in the ER? Who was booked in prison. Who suffered a psychiatric crisis that resulted in hospitalization. Who received cash assistance or food benefits — and who provided a county office as a “home address” for such programs, an indicator he was often homeless at the time.

Rountree and her team discovered when they began running such models to identify who was most at risk that “they were not the people who enrolled in typical homelessness prevention programs.” In 2020, the UCLA team found that few of the people identified by its predictive modeling — fewer than two dozen in two years — received services specifically designed to prevent homelessness under Measure H, a voter-approved sales tax of LA County.

So the county decided to call them. In July, the newly formed Homeless Prevention Unit began reaching out to people classified as most at risk under the prediction model, by making unannounced calls to residents like Cross. The UCLA analysis is conducted with data stripped of identifying details, which the county matches with names and information to find potential customers.

The work is being conducted at the Housing for Health Division of the LA County Department of Health Services, which focuses on homeless and vulnerable patients. Program participants may receive financial assistance and referrals to other services to support their general health and housing needs. Single adults receive grants of up to $4,000 or $6,000, and families receive an allowance based on size; the program analyzes whether higher amounts make a difference in the results.

For Anthony Padilla Cordova, it was “perfect timing” when the call came in February. “I just didn’t know where I was going.”

Padilla, 29, had been released from state prison during the pandemic and was trying to stay away from alcohol and drugs. He eventually got into rehab after being arrested again for his drinking and moved from there to a sober dorm, he said. But his life still felt precarious.

Many of his roommates who were struggling with their own addictions seemed “still stuck in the prison mindset,” he said, and Padilla worried that if he lost his temper and started fighting someone, he’d be kicked out. Padilla had started working as a prep cook and dishwasher, but he hadn’t regained his driver’s license, which could mean hours on the bus to get all the way to Irvine for jobs.

Fabian Barajas said he could help. The caseworker arranged to pay for a breathalyzer that had to be installed in a vehicle and activated before Padilla could drive legally again. He got gift cards to cover everyday expenses like groceries. He got him clothes and shoes for work.

And “if I feel like I’ve lost hope … I have Fabian to call,” Padilla said.

Case managers work with each participant for four months, but can extend this period by up to two months more if necessary. As Padilla nears the end of the four-month program, he now resides in his own studio in MacArthur Park. He stayed sober. He’s got his driver’s license back. And he saved enough money for a down payment on a car, Barajas said.

“Things are looking good – as long as you can stay consistent and sober,” Barajas told him when they met on a recent Tuesday.

Padilla reflected on the help he had received from the county and other programs. “If I didn’t have these resources,” he said, “I would probably be homeless.”

Approximately 150 people have completed or are currently participating in the LA County program. It started with $3 million in funding — half from Measure H, half from the Conrad N. Hilton Foundation — and quickly raised nearly $14 million in federal money through the American Rescue Plan, which will allow the program to run through 2024, said Dana Vanderford, associate director for homelessness prevention. Its staff, once “a small and powerful team of seven,” will soon grow to 28, including 16 case managers, Vanderford said.

So far, about 90% of participants have retained their housing during the program, Vanderford said. Rountree and her team are still evaluating how such results compare to similar people who didn’t participate in the program, which will help determine if it was successful.

However, based on past patterns and analysis, it is estimated that if the program had not existed and “the world were exactly the same,” the first 54 people who entered the program would have had a 33 percent risk of becoming homeless. as was done at the time of the analysis, Rountree said.

Still: “The world has probably changed. So we need to do the full causal study,” she said.

At first glance, Cross may seem like a surprising candidate for homelessness prevention. She has a housing credit for her unit in South LA. She does not oppose an eviction or rent increase. But Cross has been homeless before — and without the help of Gouché and the program, she worries about the choices she’d have to make between eating and paying the electric bills.

“If she didn’t come, I would be sitting here and saying, ‘Do I have to do this, or do I have to do this?’ ‘ said Cross. Now, “I don’t need to stress if I’m getting enough recycling this month.”

In the living room of their house, next to a TV table with elephant figures that brought good luck, postcards from a trip to Catalina Island and an open Bible to the Book of Psalms, she and Gouché discussed their next steps.

They talked about getting her a new bed and pillows to relieve her back pain; the fear that Cross said was “nothing to play with”; her discomfort with the many pills she had been prescribed which she feared might leave her with another addiction.

Cross also wanted to quit smoking, saying it was costing her money she couldn’t afford. Gouché offered to have her phone numbers for smoking cessation programs, as well as additional resources that would help her with mental health and substance use issues.

Cross nodded. “You can try everything at least once,” she says. “If it works, I’ll continue.”

Anger and depression would have sent her back to cigarettes before, she said.

Gouché said, “The big part is your awareness — you just said it — that that’s essentially how you deal with difficult situations.”

The programs could help build on that awareness and help her find other ways to deal with it, the case officer said.

“You’re definitely on the right track,” she said to Cross. Preventing homelessness with help from a computer model

Alley Einstein is an automatic aggregator of the all world’s media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials, please contact us by email – The content will be deleted within 24 hours.

Related Articles

Back to top button