LSA & SPM DQ Guide for ICA

This guide is designed to help ICA staff complete the necessary steps of internal data preparation for the LSA and SPM.

Table of Contents

What is the LSA?
What is the SPM?
Timeline
What projects does this affect?
General Notes?
What reports are needed?

What is the LSA?

The Longitudinal Systems Analysis (LSA) is a federal report submitted to HUD that informs the Annual Homelessness Assessment Report to Congress (for more, see this page on the HUD Exchange).

The Longitudinal System Analysis (LSA) replaces the previous data submission for the AHAR – Annual Homeless Assessment Report to Congress.

The Annual Homeless Assessment Report (AHAR) is a HUD report to the U.S. Congress that provides nationwide estimates of homelessness, including information about the demographic characteristics of homeless persons, service use patterns, and the capacity to house homeless persons. The report is based on Homeless Management Information Systems (HMIS) data about persons who experience homelessness during a 12-month period, point-in-time counts of people experiencing homelessness on one day in January, and data about the inventory of shelter and housing available in a community. See more on the HUD Exchange here.

The LSA submission consists of 10 .csv files submitted to HUD via HDX 2.0. These files contain project-level data on funding source, geography, and inventory data contained within the provider profile in HMIS. They also contain slightly aggregated, yet highly-specific, data on clients and households served during the fiscal year by projects participating in HMIS.

What data is pulled into the LSA?

Project Descriptor Data

• Project type, operating dates, organization information
• CoC and location information
• Funding Information
• Bed Inventories

Enrollment Data

• Entry/Exit dates (nights of stay)
• Prior Living Situation
• Relationship to Head of Household
• Disability Status
• Fleeing Domestic Violence Status
• Chronic Status
• Housing Move-In Date (if applicable)
• Exit Destination

Client and Household Demographics

• Age
• Gender
• Race/Ethnicity
• Veteran Status

LSA Data Cohorts

Client Cohorts

Only data from valid enrollments are included. A valid enrollment has:

• Entry Date
• Exit date must be at least one day later than the Entry date, if present
• Personal ID
• Project ID
• Household ID
• Valid Relationshipt to HoH
• Can only have 1 HoH per household
• HoH must be present for duration of project stay
• Enrollment CoC associated with HoH's Enrollment ID = Report CoC

Note on housing move-in dates. If a housing move-in date falls on the same day as the Exit date, it is counted as:

• Housed for RRH
• Not housed for PSH

Active Cohort

Clients or Households served in continuum ES, SH, TH, RRH, and PSH projects during the report period.

Data includes:

• Client-level demographic data
• Enrollment-level data (EE dates, living situation, disabling condition)
• Adults and HoH: enrollment data for at least 3 years prior to client's last active date in the report period (Relevant to chronic homelessness, and for HoH system path)
• For clients served with HoHs on enrollments before report period (even if not active during report period), date of birth and EE dates are used to determine household type What is the SPM?

Exit Cohorts

Exit cohorts are clients who exited from a continuum SO, ES, SH, TH, RRH, or PSH project during 3 different time period and were not enrolled in any continuum ES, SH, TH, RRH, or PSH project in the 14 days after exit.

Exit Cohort time periods

• Exit Cohort 0: exited in first 6 months of report period
• Exit Cohort 1: exited in the year beginning 1 year prior to report start
• Exit Cohort 2: exited in the year beginning 2 years prior to report start

Data includes:

• Client- and enrollment-level data for the HoH
• Enrollment-level data for active enrollments
• For clients served with HoHs on enrollments, date of birth (for household type) and disabling condition, etc (for population identifiers)

Point in Time Cohorts

Data from clients with bed nights on 4 nights in the report period

• October 31
• January 31
• April 30
• July 31

What is the SPM?

The System Performance Measures (SPM) is a federal report submitted to HUD that examines the performance of the local coordinated system of homeless assistance programs (for more, see this page on the HUD Exchange). 

What data is pulled into the SPMs?

Enrollment Data

• Entry/Exit dates (nights of stay)
• Prior Living Situation
• Exit Destination
• Income for CoC-funded projects

Timeline

HUD's timeline

HUD has stated (via HMIS Lead call) that the LSA will likely open in October/November 2023 to start the submission process, but have not released an updated reported calendar. The LSA will likely be due in January 2023.

When do I have complete data cleanup?

TBD exactly.

Kaitlyn will be working with CoC teams to ensure successful DQ cleanup.

What projects does this affect?

The LSA and SPM pull data from specific continuum projects,  regardless of funding source. The chart below indicates whether your specific project type participates in each. 

General Notes

Maintaining notes for LSA flags

In the LSA Airtable, there is a table, "Project Notes", that lists all projects for all CoCs, along with basic information, and a space for you to add notes. This is where you can note any programmatic oddities, or unique circumstances, or DQ issues that can't be resolved, so that we can more efficiently respond to LSA flags.

Report Period

The SPM pulls data only for the Fiscal Year (FY). The LSA pulls PDDE data only for the FY, and pulls client data for the last 3 FYs, focusing only on exit information for the 2 prior FYs. For these internal reports, focus on the FY only.

The FY runs from 10/1 - 9/30. 

So, for FY2023, the dates are: 10/1/2022- 9/30/2023.

Continuum = Yes

Only HMIS providers that collect client-level information should have the following:

• Continuum = yes
  • Any other provider (e.g., a level 2 agency, or placeholder level 3 project) with continuum=yes will get erroneously pulled into the report causing issues
• Funding sources
• CoC codes
• Bed and Unit inventories

If one of your agencies has a level 3 or 4 provider that acts as a placeholder (i.e., does not have client data tied to it), make sure that provider has Continuum = No.

What reports are needed?

Required

• HUD's EVA DQ Tool
• Utilization
• Inventory vs. Enrollment Report

HUD'S EVA DQ Tool

Report Guide

See here for the internal report tipsheet

DQ Impact on LSA

This covers everything from PDDEs, to CoC-wide errors and warnings, to project-level errors and warnings.

Utilization Report

Report Guide

See here for the internal report tipsheet

Other than the folder path, there is no difference between the Missouri and Wyoming version of the report.

DQ Impact on LSA

There are many LSA flags on utilization, based on HH types, and PIT dates.

Inventory vs. Enrollment Report

Report Guide

See here for the internal report tipsheet

Other than the folder path, there is no difference between the Missouri and Wyoming version of the report.

DQ Impact on LSA

There are LSA flags on utilization regarding funding type (e.g., no vets served), as well as issues like clients enrolled outside the operation dates.

Strongly Recommended

• PDDE DQ Report
• Enrollment Overlap Report
• HoH and Client Location DQ
• HMIS BUI vs HIC

Why are these not required?

Because EVA covers most (if not all) of the information contained in the PDDE, Enrollment Overlap, and HoH and Client Location reports. However, it can still be helpful to run these reports to get detail information so you can verify not only is the data not causing a HUD error, but that it also matches what you know is true about the

PDDE DQ Report

Report Guide

See here for the internal report tipsheet

Other than the folder path, there is no difference between the Missouri and Wyoming version of the report.

DQ Impact on LSA

Some PDDE errors will prohibit the LSA from exporting from ServicePoint (e.g., no funding sources or invalid funding sources), others will result in LSA flags.

Enrollment Overlap Report

Report Guide

See here for the internal report tipsheet

Other than the folder path, there is no difference between the Missouri and Wyoming version of the report.

Why run this report?

Enrollment overlaps, when in error, indicate data entry issues and should be resolved. Some represent impossibilities based on the definitions of the different project types. They also can lead to issues with counts of clients served where there are discrepancies between project-level counts and CoC-wide counts.

See here for HUD's guidance on enrollment overlap.

DQ Impact on LSA

The LSA flags these discrepancies. Eliminating impossible overlaps will mitigate these flags. 

HoH and Client Location DQ

Report Guide

See here for the Missouri report tipsheet

See here for the Wyoming report tipsheet (HoH only; WY doesn't have the Client Location issue)

Why run this report?

Agencies will also be running this report to check their data quality on relationship to head of household errors and client location errors ( Missouri Data Preparation Guide, Wyoming Data Preparation Guide).ICA CoC teams should also run this report, at the CoC level, to review any remaining DQ errors and any errors that are a result of visibility issues.

Missouri-specific

If the Client Location (CoC) issue is due entirely to a visibility issue, you can make appropriate updates in the system as long as the change is documented.

DQ Impact on LSA

• Flag if there is >1 or <1 HoH in an enrollment
• Flag if there is a mismatch between Client Location (CoC) and provider CoC

DQ Impact on SPM

• Clients will not pull into a CoC's SPM reports if their Client Location (CoC) does not match the CoC you are running for.

HMIS BUI vs HIC

Note - if HMIS was updated to match HIC data in Spring 2023, this is not necessary. However, it could be a good check.

Report Guide

See here for the tipsheet

DQ Impact on LSA

A whole host of LSA flags about mismatches between HIC and HMIS

Copyright 2023 Institute for Community Alliances. All Rights Reserved.