I had the opportunity to review a recent study from St. Louis colleagues Kube, Das and Fowler called Allocating Interventions Based on Predicted Outcomes: A Case Study on Homelessness Services.
Here’s the gist:
The authors used predictive modelling to give a sense of which program types (Permanent Supportive Housing, Intensive Case Management, Rapid Rehousing, etc.) would be best matched to which population group to get best housing outcomes and reduce returns to homelessness using 2007-2014 data from the Homelessness Management Information System (HMIS). (To put it in a fancy way, the data was analyzed using Bayesian Additive Regression Trees to determine if a household would have re-entered the homeless system if assigned to a different intervention.)
Here’s what they found:
Their model found improvements that could be made to get better systems-level results. By improving program-client matching, re-entries into homelessness could be reduced by 15% compared to the status quo. There were some pros and cons to this approach as well, but overall, interesting but not mind-blowing.
However, here’s something that was surprising (at least to me):
The model showed that the highest-needs clients (homelessness chronicity, higher substance use, very low income, etc.) would be the ones who would benefit the most from … PREVENTION program matching!
The data suggested that matching higher-needs clients experiencing chronic homelessness to prevention interventions would get best long-term outcomes compared to other interventions.
That’s a big deal and here’s why: we normally advise the prevention is a lower-intensity, shorter term intervention usually matched for lower acuity, short-term homeless clients. We tend to refer high acuity, chronically homeless clients into Permanent Supportive Housing or Housing First Intensive Case Management.
So what would Chronic Homelessness Prevention programming look like?
Communities like St. John’s, Medicine Hat, and Hamilton are experimenting with their current slate of Housing First programs to develop ways that support higher needs individuals without waiting for them to hit the shelter door.
This raises interesting questions:
- What would eligibility criteria and referral processes look like with that in mind?
- How would the program work with shelters, supportive housing or Coordinated Access?
- What’s the right level and expertise in case management/support staff?
- What’s the right length of stay and intensity level of support?
- What does client success look like?
- How does the program fit in the broader social safety net?
If you have ideas or if you have a chronic homelessness prevention program, get in touch – we’d love to hear about it!