So you’ve been using HIFIS for a while. You’re doing okay, but you know your community could have better data quality. And you even have some ideas of what the gaps are, things like: caseworkers forgetting to close off their case files; shelters forgetting to record last known housing before a book-in; and staff generally not remembering to add source of income.
You know where you want to be. You know where you are now. You just don’t know how to get there. So what should you do?
Articulate your expectations
Do you have data quality expectations clearly spelled out in policy? Do you have any documentation that says “this data point must be entered within 24 hours” or “our expectation is 90% completion in this other data point”?
If you don’t have data quality standards, you’re not clearly signaling to your stakeholders (a) that data quality is important in general, and (b) which elements are the most important. Your service providers could rightly point out that they didn’t realize it was so important to enter some data point, or they could say that they had other priorities, or they weren’t aware of what the requirements were. And it trickles down. If management doesn’t have a target they need to achieve, they won’t emphasize it for front line staff.
By the way, if you hold the funding strings, you absolutely can and should put data quality expectations into your contracts, next funding cycle. (In the United States, each CE equivalent (they call them COCs down there) are federally mandated to have a data quality plan, and data quality is definitely a condition of funding. You can check out some of the resources they have on HUD Exchange for ideas.)
Now, articulating your data quality expectations does not immediately fix anything. But it sets you up for long-term success and signals that your community cares about data quality. Think about it as laying down some foundations.
Double down on your training
What does HIFIS training look like in your community? In our experience, it varies quite a bit from community to community. In some places it’s all done centrally and rigorously; there’s only one trainer and everyone receives training in exactly the same way and they only get access to HIFIS once training is complete and signed off on. In other places, it’s a decentralized model. A supervisor at a shelter will bring new hires up to speed, and it’s on-the-job training so they might not follow an official curriculum. The supervisor might only tell the staff what they remember and forget about some policies or how to handle edge cases. Then those new hires might go on to train the next round of new hires six months later. It’s a game of broken telephone.
We’re not saying that there’s only one right way to do HIFIS training, but if you don’t train your staff well, they’re not going to be entering data into HIFIS the way that you want them to.
Look at training in your own community. Are you confident that all your HIFIS users are well trained? If not, you should probably revisit how you train your HIFIS users. For some ideas, you can watch my Training for Front-Line Workers webinar from #CAEH24. Here are some things you can think about:
It might be time to upgrade your training mode. If you currently have people watching pre-recorded videos, consider upgrading to a LMS system. Or add an evaluation component to the end of your training - “now that you’ve completed your training, let’s do a practice assignment.”
If you rely on a decentralized model in which there are many trainers, come up with a one-page curriculum. Have a list of topics that must get covered, with checkboxes, and require your trainers to check all the boxes and send it back to you.
Consider implementing a refresher training requirement, so everyone’s knowledge is up-to-date if you have changed your HIFIS requirements or environment over time. Here’s a question: in 4.0.60.1, a new mandatory Racial Identity field was added to the Client Vitals. Have all of your users received training since you updated your software to 4.0.60.1 or higher, or have some current HIFIS users been most recently trained on 4.0.59 or lower?
Training doesn’t immediately fix a data quality problem, but it does address the cause of the problem, which is that staff were not entering in data correctly in the first place. You will almost certainly need to address your training in some way to fix your data quality.
For this specific problem that you have, though, you can think about your training solution as being general or specific. For example, if your data quality is generally bad, then generally improving your training will generally improve your data quality.
However, if you have a specific problem, like shelter workers not recording a client’s previous housing at the time of intake, then you need a more specific, targeted training solution. You could make a one-page job aid for shelter workers with bullet points saying the required steps to go through at intake. You could review your training materials for shelter workers and make sure it covers information in the way you want it covered. You could put together a mini training course for shelter workers that covers recording previous housing at the time of intake and 1-2 other things you want to highlight for shelter workers, like how they should handle family book-ins or when a client says they’ll be back but doesn’t show.
So, make sure staff are getting trained to do exactly what you need them to do, and then…
Do a blitz
Some communities have very successfully done all-hands-on-deck style blitzes. Here’s how that works: you pick a specific thing to focus on, like Housing History. You set a goal, or create a contest, or otherwise “gamify” the data entry. Something like “the caseworker that adds the most Housing History records this month gets a pizza party for their team” or “let’s try to get to less than 20% of clients in the Coordinated Access module with Unknown housing status by the end of the month” or “each Housing History record created this month enters you into a draw for a Tim Horton’s gift card. The more Housing History records you enter, the more chances you have to win.”
The more you can frame it like a fun thing to do, like a race or a challenge, the more average front-line staff may want to participate just for fun. If you offer small incentives like a pizza party or a $10 gift card, you could have staff competing with each other to see who can do the best job of data entry. Provide progress updates - “Carly’s team is in the lead but Andy’s team is close behind!” or bring a current Unknown Housing Status % to the Coordinated Access table or make it easy for each person to see how many times their name is in the draw. Remind people of upcoming deadlines but assure them that there’s still time to make an impact. Announce successes and winners to all of your system stakeholders, so people are constantly reminded that they, too, could have won a gift card this week.
This can be a great way to bring awareness to a data quality issue and build short-term enthusiasm and urgency. However, it doesn’t work that well on a sustained basis. If you only rely on this mechanism, it might work for some people, but a lot of staff will check out (they didn’t win anything last month, so they won’t try as hard this month). So this is a good method if you want to close a large backlog and you want a major push, but not great for ongoing improvement.
Give people reminders
People don’t really love constant nagging, like an email every week that says “remember to close your case files!” But there are ways that you can remind staff about something important without being annoying.
Dufferin County uses the HIFIS banner to remind staff about something that’s currently important to the community. They change the message and the colour of the banner every couple of weeks so it’s always fresh and doesn’t fade into the background.
Fort McMurray hands out tons of “I ❤️ Housing History” stickers to front line staff, which are (a) super cute, and (b) a good reminder about this very important data element.
HICC periodically sends out a HIFIS Newsletter. The most recent one had the following topics: information about upcoming HIFIS releases, an update about reports and Power BI, a note about upcoming office hours, a case study, a celebration of success, and then a reminder of how you can do a specific task in HIFIS.
In Iqaluit, they use Bulletins to remind staff about how to set up user accounts, which is a multi-step process that’s not always easy to remember (and they don’t do it frequently), in an attempt to streamline the hiring process.
Reminders can be a powerful tool to nudge people to remember their training. It can be gentle or heavy-handed, and let’s be honest, sometimes you just need to nag someone over and over until they finally do what you need them to do.
Make it visible
If you have a data quality problem you want staff to be addressing, it’s really important for them to be able to access the information they need to accomplish that goal. For example, if you tell them to “close your open cases!” but they don’t have a way to get list of their cases that are still open, they can’t do what you’re asking them to do. If you tell them to “make sure all clients have an income record” but they can’t easily see which clients have income records and which don’t, some are going to get missed.
We can’t customize what data is displayed on different screens in HIFIS, but we can customize the reports that exist, and of those reports, which staff have access to which reports.
Some communities do their staff a disservice by not giving them access to enough useful reports. Then front line staff tend to think of reporting as a manager’s job. They think of HIFIS as a black box that information goes into but they can’t see where it goes or what it’s for. I believe that there are tons of reports that would be really helpful for front line staff, like: the Activity Journal report which many service providers are using to communicate shift notes from one shift to the next; or the My Calendar report which shows things like upcoming move-ins; or the Client Profile and Client Service History which can give you an overview of a single client you’re working with; or the Simple Case Management List which gives you a list of your open and recently closed cases; or the My Client Dashboard developed in Waterloo to meet the needs of their front line staff.
The point is that if you give front line staff the ability to pull relevant data out of HIFIS, they are now given a set of tools. These tools can be used to allow them to be proactive, find gaps, and improve their own data, rather than relying on someone else to tell them what the problems are.
In general, staff will like HIFIS better if they are able to get something useful out of it. Sometimes, that means developing a report that doesn’t directly help with data quality, coordinated access, or KPIs, but makes your front line staff happy. When your front line staff are happy with HIFIS, they’re more likely to be using more, which indirectly means you’ll have better data quality.
Make sure it matters
We presented at the CAEH a few years back about Collecting the Right Data, and in it we talked about data collection as a balancing act - don’t collect enough data and you won’t know anything about your system, but ask for too much and it’s burdensome.
Our number one take-away is just to make sure that every piece of data you’re collecting matters. Don’t ask about a client’s source of income just so you can put it on an end-of-year report, ask about it if there’s something actionable you’re going to do with that information (e.g. apply for social assistance, make a budget, conduct a housing search) or don’t ask about it at all. Don’t ask whether a client’s been recently discharged from foster care unless they get prioritized higher on the BNL if they have been or you have a program just for that group you can refer them to. Don’t ask if they’re fleeing domestic violence just to collect a bunch of information before handing their file over to a caseworker, ask them if you have a protocol in place that would be triggered if they say yes.
If there’s a data quality gap that you want to address, make sure that it’s worth addressing. Let’s say only 25% of your clients have ID recorded in HIFIS… is that a problem? Why? What specific thing is impacted by a low completeness here? If you want to do an ID collection blitz, can you explain to staff why this is the major thing we should be focusing on this month? How is that specific data point impacting your community’s ability to address and reduce homelessness?
Conclusion
There’s no magic wand you can wave to improve your data quality overnight. There are a lot of ingredients that go into having quality data, and you need to tackle the problem from many different angles to get the best result.
It’s probably best to think of data quality improvement as an ongoing project, just like self-improvement is also an ongoing project. You can work out for a while and get stronger, but if you stop going to the gym, you’re going to lose some of your muscle mass. You’re never going to be “done.”
You can, however, reach a kind of equilibrium where you can be relatively confident that if you run a report in HIFIS it’ll be pretty accurate. Because you know that your HIFIS users are all trained in the same way. Because you know that all of your service providers have a shared understanding of the importance of data quality. Because you don’t have any major gaps or backlogs. Because your front line staff see HIFIS as a useful tool and not a waste of time.
Then, maybe, you can relax a little.
Comments