Everyone wants to have confidence in their HMIS data quality.
No one wants to lie awake at night wondering if the client who was determined ineligible for services was wrongly turned away because of a possible intake error. No one wants to second-guess their decision to build a drop-in center at a particular location because they’re going off of mere “hunches” about patterns in their data without actually seeing these patterns. And no Continuum of Care wants to lose government funding because their HMIS data painted an inaccurate, incomplete picture of homelessness in their community.
HUD defines data quality as referring to the reliability and validity of client-level data collected in the HMIS. It’s measured by the extent to which the client data in the system reflects actual information in the real world. With good data quality, communities can “tell the story” of the population experiencing homelessness.
It’s clear that good data quality is important to ending homelessness. But achieving this standard can be challenging due to factors such as quality of HMIS software, truthfulness of the client, question and answer interpretation, staff training, language differences, and more.
We’ve pulled together our staff expertise and dug through various resources to compile a list of ideas you can put into action to ensure better quality of HMIS data.
At the end of this article, you’ll also see an invitation to join us for a webinar in which we’ll discuss the top 3 common data entry issues regarding veteran homelessness and how to identify errors within SSVF data quality reports.
But first, let’s look at the list of data quality ideas.
1. Develop a community-level HMIS data quality plan.
First and foremost, the best way to ensure good data quality throughout your CoC is to have a community-level data quality plan. A data quality plan is a set of policies and procedures that facilitates the ability of the CoC to achieve complete, accurate, and timely client-level data. It lays out data quality goals, the steps necessary to measure progress toward those goals, and the roles and responsibilities for making sure HMIS data is reliable and valid.
Many of the below ideas are typically addressed in a data quality plan, but not every community has a clear, comprehensive plan in place. These ideas may be helpful to refining or updating your community’s data quality plan, or overall best practices, if need be.
2. Enter HMIS data within 24 hours of intake.
Accuracy of data largely depends on timeliness, particularly if the collection of data doesn’t happen directly within the HMIS. As you enter HMIS data, you may be relying on handwritten notes or your own recall of a case management session, service transaction, program entry or exit date, etc.
Set a goal to transfer data from notes or memory into the HMIS within 24 hours of intake, increasing the chances the data will be correct. This also ensures data is entered as close to real-time as possible, making it accessible when needed.
3. Create and maintain by-name lists.
Many communities are recognizing the need to develop, maintain, and use a by-name list (BNL), a continually updated snapshot of all individuals experiencing homelessness. A BNL can include categories such as Veteran status, chronic status, active/inactive status, homeless/housed status, and more.
According to Community Solutions, because a by-name list allows communities to see each homeless person by name, it’s easier to track the percentage of clients that have incomplete assessment data. As a result, this can facilitate your efforts to continuously improve any gaps in data quality.
4. Have a process for verifying and documenting eligibility.
For subpopulations such as homeless veterans or the chronically homeless, good data quality is key for correct prioritization and provision of specific benefits. It’s important to make sure a process is in place for verifying and documenting eligibility.
For homeless veterans, this can be an issue with false self-reports and lack of efficient data sharing between VA and non-VA service providers. For individuals who are chronically homeless and whose history of homelessness has occurred across different communities and/or states, there is difficulty with the lack of communication between in-state and out-of-state providers.
Collaboration and HMIS customization serve to streamline this verification and documentation process. You can learn more in our article on by-name lists and eligibility.
5. Establish clear, consistent definitions and interpretations.
Consistency is crucial to good data quality. For example, there could be inconsistent interpretation of the exact meaning of a field, such as “disability.” Two people with the same condition might provide completely different answers in response to whether they have a disability, leading to inconsistent data.
To avoid this, you must fully understand the meaning of the field and query further to elicit the most accurate response from the client. There needs to be established definitions and interpretations of questions, answers, and data entry processes, including which HMIS fields require completion.
Following HUD’s HMIS Data Standards is also a large part of maintaining consistency. Additionally, having an HMIS that has display logic and integrated tooltip explanations also helps with achieving consistent interpretations and processes.
6. Make sure your client understands the question.
Similarly, errors in data collection occur when the client misunderstands the question. A common example of this is misunderstanding what is meant by the “Residence Prior to Project Entry” question. The client may give you a response referring to where they lived for years prior to becoming homeless as opposed to the place they stayed for one night prior to shelter entry.
Even the question “What is your name?” may elicit two different responses—the client’s legal name one day, and their nickname in a follow-up interview. Where there is any room for misunderstanding, always elaborate and explain the question to make sure the most accurate data is collected.
7. Accommodate language differences.
Language differences during a case management session can lead to poor quality of data if procedures aren’t in place to accommodate those differences. Service providers should identify common languages spoken by clients within their community (other than English). Then have a copy of questions and answers in those languages available for clients to read along if they wish.
8. Be in a quiet and private intake space.
Speaking of language barriers, you may sometimes hear a client’s response incorrectly, especially when working with clients who have strong accents or language barriers. Create an intake space that is quiet and private so that you can hear clearly and feel comfortable following up on sensitive questions to make sure you understand the response.
9. Take precautions to avoid duplicate entries.
It’s easy to accidentally create duplicate records for the same client. Typos in social security number, date of birth, and misspellings of names are common intake errors that can lead to duplicate client records. However, these are easy to avoid by following the simple rule of always confirming the response.
In regards to spelling, even common names like “James” can sometimes be spelled “Jaymes.” For unusual spellings like this, it’s a good idea to circle or highlight the name to help ensure data entry staff make note of it and don’t try to correct it to usual spelling.
10. Be proactive in your monitoring plan.
Being proactive is key. Don’t wait until there are problems before you start paying attention to data quality. Make a plan to monitor data quality regularly so that you aren’t scrambling to identify and correct data errors right before reports are due. Establishing standards and corresponding plans to monitor these standards is a large part of the community-level data quality plan.
For example, one point of a monitoring plan could be to print a client detail report from the HMIS every month to double-check that the list of current and exited clients is accurate. You could also run weekly or biweekly data quality reports to identify missing/refused/don’t know responses in advance of report deadlines.
11. Use positive reinforcement and incentives.
Typically, the final part of a data quality plan is a protocol for enforcement measures and incentives. For example, you could give quarterly or annual recognition awards to HMIS case managers who have substantially improved data quality.
For HMIS participating agencies who consistently meet data quality standards, you could award extra bonus points on their NOFA application, a discount off their HMIS participation fees, or first priority status for bonus projects, RFPs, and special funding availability. By describing incentives, you can reinforce the importance of good data quality throughout your CoC.
“To meet the HMIS goal of presenting accurate and consistent information on homelessness, it is critical that an HMIS have the best possible representation of reality as it relates to homeless people and the programs that serve them. Specifically, it should be our goal to record the most accurate, consistent and timely information in order to draw reasonable conclusions about the extent of homelessness and the impact of homeless services.”
Good data quality is a critical part of ending homelessness. Having a timely, accurate, and complete understanding of homelessness within your community affects your ability to provide comprehensive care to clients, measure impact, and receive funding.
While there are challenges in achieving a high level of HMIS data quality, we know following the ideas listed in this article are sure steps to attaining this standard within your community.