Measuring a Functional End to Homelessness

It is no easy task to measure an end to something as fluid and dynamic as homelessness. I am not writing this blog to critique approaches that have been suggested by others. Instead my focus is to add my voice on considerations and approaches that communities may want to contemplate to truly declare “Functional Zero”.

Focus on outcomes, not outputs. 

An output measures the volume of an activity. An outcome is the measurement of what difference any of it makes. If you are focused on outcomes, the system changes WHILE the program participants are getting housed. With veterans I would expect, for example, different outcomes related to transitional housing and GPD programs than what they currently are because of the focus on permanent housing. With chronic homelessness, for example, I would expect a change in community policing and engagement with people living outdoors. These are just a couple of examples amongst many that are possible.

The difference of any program focused on housing should be not just that people get housed (really important) but also that people stay housed (though maybe not in the same place they were first housed). Housing retention rates (preferably by acuity) are a good measure, especially if we can track changes in acuity post-housing.

Don’t blame the service participant; examine the service provider.

Noting an offer of service is good insofar as it tells us something about whether people are engaged by outreach or other service providers. What it does not tell us is if the proper service is being offered. I would suggest that we need to know HOW service providers are adjusting their services and service offers to best meet the needs of people that are in need of services. Think of it this way:

The manager of the restaurant asks his server, “Did you go to their table?”

And the server replies, “Many times.”

The manager asks, “Did they get what they came here for?”

And the server replies, “I offered them drinks each time and they kept saying ‘no’.”

So the manager inquires, “Did you offer to make them the drink of their choice? Did you offer them the dinner menu with water? Did you see if they were here just for dessert.”

And the server replies, “You just said we needed to check in on our tables. You didn’t say we had to cater what we offer to give them what they need.”

Measure fidelity to a specific type of evidence-informed support model.

If your community has an ACT program to assist with service provision, then you can measure fidelity to the ACT model. If your community is using an ICM program to assist with service provision, the you can measure fidelity to an ICM model. These are well established and will tell you if service providers are doing what is necessary to provide supports necessary to keep people housed in alignment with the main currents of thought and practice, as evaluated and published.

The assumption here is that an evidence-informed model of support is being used in your community to end homelessness. If it is not, then that probably raises bigger flags – especially if you are trying to measure an end to homelessness.

Focus on what you know you know, while respecting there are some things that you don’t know you don’t know.

A presumption and preference of knowing each person or family experiencing homelessness by name is a very good thing. But there is a huge (flawed) assumption in that, which is there is a mechanism to be all-knowing in real time.

There are things we know we know. There are things we know we do not know. There are things we do not know that we do not know.

To the best that your resources allow and to which people engage with you, you may have a voluntary list of persons that want services from you. That is a great thing. Do not assume, however, that it is 100% representational of what is happening in your community.

Measure the things you control, and not the things you do not.

Outcome measurement should be limited to the impact of resources that are directly within your control to influence. For example, considered a VA funded program that requires the person to be anything other than dishonorably discharged to be eligible. Measuring an end to veteran homelessness, therefore, may want to include people that were, say, dishonorably discharged, but the VA does not control those resources. Therefore, it is impossible and inappropriate for the VA to prefer or compel a non-VA resource to help out a veteran that was dishonorably discharged. Consider my restaurant example again:

“I sent the couple to Murphy’s next door because they did not meet our dress code” says the maitre’d to the manager.

“Very well,” says the manager, “but please go next door and make sure Murphy’s served them well.”

“But we do not own or manage Murphy’s,” says the maitre’d.

“Yes,” says the manager, “but they should serve them because we cannot.”

And what would we expect? That Murphy’s served the couple? That they would give up a table that could have served other patrons because the restaurant next door – that does not own or manage them – sent the couple over? Is that in the other restaurant’s best interest?

Decrease variation. Always.

If we are going to measure an end to homelessness there will be local circumstances that influence efforts (for example: vacancy rates, fair market rent, etc.). However, the more we suggest there can be variation rather than standardizing based upon certain characteristics, the more flawed the measurement. If we are going to benchmark, what prevents stating something like, “In markets where the vacancy rate is 2.5% or below and the fair market rent for a one-bedroom unit is $800 or above, the average length of homelessness should not exceed x.”

If it is politically driven in picking the date of when homelessness has been ended or declaring functional zero for veterans happens on or near Veterans Day, call bullshit. 

The great thing about measurement driven by data is that it is independent of political interference. There is absolutely no need to figure out when an end to homelessness fits into a Mayor’s “messaging cycle”. Functional zero happens when functional zero happens. Not when it is convenient.

Which leads me to Veterans Day proclamations. Mark my words: this will be the year when dozens of communities declare an end to homelessness amongst veterans on or near November 11. Just too big of a convenience. It is hogging the spotlight on a day when the media spotlight is conveniently focused on the issues of veterans.

Measure the right things. In the right way. Do it at the right time. Measure it with integrity. If the lives of people experiencing homelessness were important enough to you to house and support them in the first place, surely they are important enough to measure the progress of that goal in a way that respects them.