September 18-19, 2006 – Denver, Colorado Sponsored by the U.S. Department of Housing and Urban...
-
Upload
christian-peters -
Category
Documents
-
view
216 -
download
3
Transcript of September 18-19, 2006 – Denver, Colorado Sponsored by the U.S. Department of Housing and Urban...
September 18-19, 2006 – Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development
Understanding the Extent and Nature of Homelessness at the Local Level:
Implementing More Efficient and Accurate Point-in-Time Counts
Dr. Martha Burt, Urban InstituteMary Joel Holin, Abt Associates Inc.
Karen Booth, Baltimore Homeless Services, Inc.Jay Bainbridge, NYC Department of Homeless Services
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 2
Overview
• Learning objectives
• Conducting point-in-time counts – the basics
• Guidance for integrating HMIS into your count– Baltimore City case study
• Karen Booth – Office of Homeless Services, Baltimore, MD
• Techniques for assessing the accuracy of counts– New York City case study
• Jay Bainbridge – New York City Department of Homeless Services
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 3
Learning Objectives
• To increase understanding of the key components of unsheltered and sheltered point-in-time counts and be prepared to increase the efficiency of counts in your own community.
• To learn how HMIS can assist with tracking bed inventories, determining unduplicated counts, and conducting a street and shelter count.
• To identify strategies to cope with challenges of gathering point-in-time data and using HMIS for point-in-time counts.
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 4
What Does It Mean to “Count” Homeless People?
• A point-in-time count enumerates the sheltered and unsheltered homeless population in your community
• A “count” = to collect information about1. Enumeration data on the number of homeless persons2. Descriptive information on those counted
• Demographic• Service use• Needs
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 5
Why Count Homeless People?
• To raise public awareness
• For planning and program development– Understand characteristics and needs – Develop programs based on need– Access resources for services and housing
• To measure progress in eliminating homelessness and to ensure accountability
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 6
Raising Public Awareness
• Conducting a count draws attention to the issue of homelessness– Possible opportunity for media attention– Especially useful in rural and suburban areas where
homeless people are not typically visible– Opportunity for community discussion
• Opportunity to educate the public and local government officials about homelessness– Who becomes homeless and why– Service and housing needs
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 7
Planning and Program Development
• A point-in-time count/survey is a planning tool– Identifies characteristics and needs – Guides decisions about program development and
resource allocation– Helps to quantify needed resources
• Point-in-time counts provide information that helps secure resources for homeless services– McKinney-Vento grant application – Other grant applications or private support– Justify requests for local or state government funds
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 8
Progress and Accountability
• Point-in-time counts can help measure progress in addressing homelessness over time– At the CoC and neighborhood levels
• Help answer the question: Do available services meet the existing needs?– Feeds into planning process and thoughtful program
development
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 9
Whom to Count and Gather Information About
• According to HUD, point-in-time counts must be: • “statistically reliable, unduplicated counts or estimates of
homeless persons in sheltered and unsheltered locations at a one-day point in time” (SuperNOFA 2006)
• Conducted at least every other year and during the last seven days of January
• At minimum, count homeless persons according to HUD definitions and guidelines
• Information is reported in the Population and Subpopulations Chart in the McKinney-Vento application (Exhibit 1, Chart K)
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 10
HUD Guidelines and Reporting Requirements
• Count sheltered and unsheltered adults, children, and unaccompanied youth – Unsheltered homeless people reside in a places not
meant for human habitation, such as cars, parks, sidewalks, abandoned buildings, on the street
– Sheltered homeless people reside in emergency shelter or a transitional housing (include hotel or motel vouchers for homeless people)
• Count the number of individuals, number of families, and number of persons in families
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 11
HUD Guidelines and Reporting Requirements
• Number of sheltered and unsheltered chronically homeless people
• Six other subpopulation categories required for sheltered and optional for unsheltered persons
Part 2: Homeless Subpopulations
Sheltered Unsheltered Total
a. Chronically Homeless (For sheltered, list persons in emergency shelter only)
b. Severely Mentally Ill * c. Chronic Substance Abuse * d. Veterans * e. Persons with HIV/AIDS * f. Victims of Domestic Violence * g. Unaccompanied Youth (Under 18) *
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 12
HUD Guidelines and Reporting Requirements
• Housing Inventory Charts– Number of beds and units– By program type and individual project– Emergency shelter, transitional housing, AND
permanent supportive housing
• Update inventory every year
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 13
Methods for Counting Homeless Persons
• Probably need a ‘menu’ approach, using a combination of where, when, how to count
• Unsheltered Counts (Street or public places counts) – Where to count
• Complete coverage• Known locations• Non-shelter services
– When to count• Night designated for the count: Last seven days in January• Length of data collection
– A ‘blitz’ count, 24 hours or less– More than 24 hours
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 14
Methods for Counting Homeless People
• Unsheltered Counts (cont.) – How to count unsheltered homeless people
• Simple count with observation• Count plus interviews• Service-based count (includes interviews) • Probability sampling
– Used by large cities and requires statistical expertise
• What methods are best for your community?• Urban / Suburban• Suburban / Rural
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 15
Methods for Counting Homeless People
• Sheltered Counts– Already know where and when to count– How to count
• Homeless Management Information System– Extract client-level count and/or subpopulation data
• Provider report or survey– Aggregate count of people in program– Subpopulation information or estimate for entire program
• Client-level survey using standardized instrument– Interview each client or a sample of clients– Program staff complete survey based on case records or knowledge of client
• Always collect count information from the provider, even if client interviews are completed
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 16
Challenges for Rural CoCs and Colonias
• Covering the territory
• Differentiating poorly housed from homeless people• The continuum of residential instability• Finding allies by including some high-risk-of-
homelessness populations—assuming you differentiate when you report
• Special populations (migrant workers, border crossers, “snowbirds”)
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 17
Benefits of Using HMIS for Point-in-Time Counts
• Requires fewer resources over time than a non-HMIS PIT count
• Helps avoid duplicate counting
• Provides in-depth subpopulation data on persons who are counted without repetitive interviews
• Reinforces the value of the HMIS and contributes to year-round HMIS participation and data quality
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 18
How Can HMIS Be Usedfor a Point-in-Time Count?• HMIS is a tool for gathering information about people
who use services– More useful for sheltered count
• Communities will need to continue street or service-based counts even if outreach workers are using HMIS
• Potential uses for unsheltered counts– Provides an opportunity to populate the HMIS – Helpful with de-duplicating street or service-based
counts
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 19
How Can HMIS Be Usedfor a Point-in-Time Count?
• Provides two components: 1. Count 2. Subpopulation information
• Typically used in combination with other data collection techniques
• Key considerations:– Data quality, data quality, data quality!– Make sure information is gathered or extrapolated for
each provider
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 20
How Can HMIS Be Usedfor a Point-in-Time Count?
• Examples
1. HMIS for count + subpopulation information
2. HMIS for participating providers + paper survey for non-participating providers
3. HMIS for count + paper survey for subpopulation information
4. HMIS for subpopulation information + paper survey for count
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 21
How Can HMIS Be Usedfor a Point-in-Time Count?
• Examples (cont.)
5. HMIS for participating providers + extrapolation for non-participating providers
6. Have participating providers generate paper surveys using HMIS-canned reports
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 22
Methods for Collecting Housing Inventory Information• HMIS bed management (possibly supplemented by
surveys for non-HMIS providers)
• Survey providers– During point-in-time population count– Send the last Housing Inventory Chart for providers to
update
• Include instructions on how to count seasonal beds, overflow beds, family units, vouchers
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 23
Key Considerations for Using HMIS
• Start planning early
– 4 to 6 months prior to count
– Leave adequate time to assess data quality, improve data quality, and decide whether and how to use HMIS
– Create a plan to gather sheltered data from non-participating providers
– Refer to Guidance on Counting Sheltered Homeless People for details
September 18-19, 2006 – Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
Karen BoothInformation Systems Coordinator
Baltimore Homeless ServicesBaltimore City Health Department
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 25
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
• Background:
– Conducted in January 2005
– BHS worked with the Center for Poverty Solutions
– PIT consisted of 3 components:• Street count/interviews• Shelter count (HMIS)• Shelter surveys
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 26
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
• Street Count/Surveys:
– BHS staff and volunteers– Service-based approach– Count ran from 6:30am until 9:00pm– Counted individuals on the street and engaged some to
conduct in-depth interviews
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 27
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
• Shelter Survey:
– BHS staff and volunteers– Went to emergency and transitional programs– Engaged clients and conducted in-depth interviews
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 28
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
• Shelter Count:
– BHS attempted to use the HMIS to count the number of clients housed in both emergency and transitional programs on January 30, 2005
– Gaps Analysis report from HMIS• Number of individuals and individuals in families in shelter• Number of families and average family size• Number of chronically homeless• Other “special population” characteristics
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 29
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
• What we found:– Numbers in the Gaps analysis did not match what BHS
believed to be the utilization rate for all shelters for that night
• What it meant:– Programs were not entering and exiting clients as
diligently as they should have been
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 30
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
• What we found:– Excess number of volunteers were sent to shelters that
were using the HMIS
• What it meant:– Need to re-think how we allocate valuable resources for
the next PIT. Need to put volunteers out on the street and in shelters that do not use the HMIS.
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 31
Using HMIS for a Point-in-Time Count: Baltimore City’s Experience
• What we found:– Other special characteristics (chronic, substance abuse,
etc.) numbers were not as high as expected
• What it meant:– Shelter staff needed more PIT-specific training and more
advance notice about the PIT and what information would be collected
AND– BHS staff needed to employ more rigorous data quality
measures
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 32
Using HMIS for a Point-in-Time Count: Lessons Learned
• It’s not just about coverage
– Just because an agency’s beds are accurately reflected in your system, does not mean their data is 100% accurate
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 33
Using HMIS for a Point-in-Time Count: Lessons Learned
• Verify data quality and accuracy prior to PIT date
– Work with providers to verify:• Number of beds and/or units represented on HMIS• Staff enter data on a timely basis• Staff are aware of importance of data quality and accuracy
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 34
Using HMIS for a Point-in-Time Count: Lessons Learned
• Conduct PIT/HMIS training with provider staff
– Review data that will be used for PIT– Train on data quality and accuracy
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 35
Using HMIS for a Point-in-Time Count: Recommendations
• Recommendations:
– PIT staff and HMIS staff should work collaboratively to plan
– Providers should also be involved in planning
– Review previously used PIT instruments, HMIS data fields, current reports, etc.
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 36
Using HMIS for a Point-in-Time Count: Recommendations
• Recommendations (cont’d):
– Look at coverage on HMIS• Parameters - Which programs participate in your HMIS and
which do not?This will tell you how to allocate your “manpower”
• What percentage of emergency beds and transitional beds/units are represented on your HMIS?
This will help to inform your decision about only using your HMIS or supplementing
September 18-19, 2006 – Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development
New York City HMIS Count: Use of Sampling and Plant-Capture
for a Point-in-Time Estimate
Jay Bainbridge, Ph.D.Assistant Commissioner
Policy and Planning DivisionNew York City Department of Homeless Services
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 38
NYC HMIS Point-in-Time Count
Sheltered• DHS shelters• Drop-in centers• Private shelters • Other CoC shelters
Unsheltered• Streets• Subway trains• Subway stations• Parks• Transportation hubs• Hard-to-reach areas
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 39
NYC HMIS Sheltered Count
Prepare for the count
Use existing databases
Survey programs without HMIS
Estimate subpopulation counts
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 40
NYC HMIS Unsheltered CountHomeless Outreach Population Estimate (HOPE)
Counts within public places Uses a sampling strategy
Relies on volunteers
Builds in a plant-capture method for quality assurance
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 41
NYC HOPE – Sampling Frame
Divide the City into small geographic areas Treat subways as their own entity
Sort the areas into those with many unsheltered individuals and those with few, based on:• Previous year’s data• Experience and knowledge of experts
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 42
NYC HOPE – Sampling Strategy
Cover all areas designated as having many homeless people
Draw a stratified random sample from the remaining areas
Stratify by borough and subway
Determine the appropriate sample sizes, based on:• Standard deviations• Confidence level (95 percent)• Precision (differs by borough)
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 43
Assign 3-5 volunteers per team
Select a team leader who has:• Outreach experience,• Detailed local knowledge, or • HOPE experience
Cover 1-4 areas/subway stations
NYC HOPE – Survey Teams
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 44
Administer quickly• If sleeping, consider person homeless• If awake, ask a series of 4-7 questions
Determine homeless status
Eliminate double-counting
Make available in both English and Spanish
NYC HOPE – Survey Instrument
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 45
Assign 70-150 volunteers per training site
Have a leader with several helpers manage each site
Train volunteers from 11-12 p.m.
Survey from 12:05 – 4:00 a.m.
Take security measures
Allow for contingencies
NYC HOPE – The Night of the Survey
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 46
Survey everyone in the assigned areas
Cover the assigned areas only once
Approach everyone respectfully
Follow rules to determine homeless status
Call on outreach if a person wishes to come inside
Follow protocol for emergency weather conditions
Stay safe
NYC HOPE – Volunteer Training
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 47
Increasing HOPE Accuracy through a Plant-Capture Study
Goal: Determine if volunteers successfully found and counted the visible homeless in their assigned study areas
Approximation: Deploy decoys and see if they are counted
Adjust the census count to correct for the estimated number of uncounted decoys
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 48
Stratify allocation of plants: by borough + density, according to City’s sampling-fractions
Randomly distribute according to above formula –150 plants to cover 75 sites
Assign plants in pairs (for safety) to precise positions after advance reconnaissance of sites
Instruct plants to stay awake, behave appropriately, and turn over stickers when approached by HOPE survey teams
Plant-Capture Design
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 49
Plant-Capture Benefits
Enables CoC to better quantify the unsheltered
Helps provide visibility to the count
Motivates volunteers and enhances overall success
Does not interfere with general operations
Lends credence with independent evaluation
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 50
Plant-Capture Challenges
Managing logistics
Ensuring adequate coverage by decoys• Having more plants would be better, but at a cost
Recruiting decoys who look and act like stereotypical street-dwellers
Determining who was “missed”
September 18-19, 2006 - Denver, Colorado Sponsored by the U.S. Department of Housing and Urban Development 51
Lessons Learned
Good News• Focus on data and targets is leading to fewer homeless
The Count is Better Every Year• More volunteers showed and more decoys were counted
It’s not just about the numbers
Important Work Ahead