Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
An accessibility analysis of the homeless populations' potential access to healthcare facilities in the Los Angeles Continuum of Care
(USC Thesis Other)
An accessibility analysis of the homeless populations' potential access to healthcare facilities in the Los Angeles Continuum of Care
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
An Accessibility Analysis of the Homeless Populations’ Potential Access to Healthcare Facilities
in the Los Angeles Continuum of Care
by
Erin Barr
A Thesis Presented to the
Faculty of the USC Dornsife College of Letters, Arts, and Sciences
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Master of Science
(Geographic Information Science and Technology)
May 2020
Copyright © 2020 by Erin Barr
All rights reserved
iii
Acknowledgements
With great appreciation for the professors who have helped me throughout my undergraduate
and graduate career, the faculty and staff of the Spatial Sciences Institute, and the University of
Southern California for all it has given me.
iv
Table of Contents
Acknowledgements ...................................................................................................................... iii
List of Figures .............................................................................................................................. vii
List of Tables ................................................................................................................................ ix
List of Abbreviations .................................................................................................................... x
Abstract ......................................................................................................................................... xi
Chapter 1 Introduction ................................................................................................................ 1
1.1 Definitions ...................................................................................................................... 2
1.1.1 Homelessness definitions ............................................................................................ 2
1.1.2 CoC and PIT definitions ............................................................................................. 2
1.1.3 Hospital and medical center definitions ...................................................................... 3
1.2 CoC Funding Allocation ............................................................................................... 3
1.3 Study Area ..................................................................................................................... 4
1.4 Hospitals and Medical Centers .................................................................................... 5
1.5 Data-driven homelessness studies ................................................................................ 6
1.6 Motivation ...................................................................................................................... 8
Chapter 2 Related Work ............................................................................................................ 10
2.1 Homelessness in the United States ................................................................................... 10
2.1.1 State estimates .............................................................................................................. 11
2.1.2 CoC estimates .............................................................................................................. 13
2.2 Homelessness in the Los Angeles CoC ............................................................................ 14
2.2.1 LA CoC efforts ............................................................................................................ 14
v
2.2.2 LA CoC homeless count methodology ........................................................................ 15
2.2.3 Homelessness in LA by the numbers ........................................................................... 17
2.3 Access and utilization of medical care ............................................................................ 18
2.3.1 Accessibility ................................................................................................................. 19
2.3.2 Walkability ................................................................................................................... 20
2.4 2SFCA ................................................................................................................................ 21
2.4.1 2SFCA methodology ................................................................................................... 21
2.4.2 Enhanced 2SFCA methodology ................................................................................... 22
2.5 Summary of Related Works ............................................................................................. 23
Chapter 3 Methods ..................................................................................................................... 25
3.1 Data .................................................................................................................................... 26
3.1.1 Homeless Count 2019 Results by Census Tract dataset .............................................. 27
3.1.2 Los Angeles Hospitals and Medical Centers (2019) shapefile .................................... 28
3.1.3 Los Angeles Census Tracts (2010) shapefile ............................................................... 28
3.1.4 Los Angeles CoC shapefile (2018) .............................................................................. 29
3.2 Visualization of homelessness in Los Angeles ................................................................ 29
3.2.1 Homeless population counts choropleth map .............................................................. 29
3.2.2 Homeless population counts hot spot analysis map ..................................................... 30
3.3 Two-step floating catchment accessibility methodology ............................................... 31
3.4 2SFCA Methodology Caveats .......................................................................................... 32
3.4.1 Walkable distance determination ................................................................................. 32
3.4.2 Census tract population by areal extent ....................................................................... 33
3.5 Overview of 2SFCA methodology ................................................................................... 34
vi
3.6 2SFCA data preparation .................................................................................................. 34
3.7 First catchment .................................................................................................................. 35
3.8 Second catchment .............................................................................................................. 36
3.9 Distance to nearest hospital .............................................................................................. 37
Chapter 4 Results ........................................................................................................................ 38
4.1 Homeless population distribution .................................................................................... 38
4.1.1 Homeless population counts choropleth map .............................................................. 38
4.1.2 Homeless population counts hot spot analysis ............................................................. 40
4.2 2SFCA and Accessibility Index ....................................................................................... 41
4.3 Hospital accessibility ......................................................................................................... 46
4.4 Distance to nearest hospital .............................................................................................. 48
Chapter 5 Discussion and Conclusion ....................................................................................... 50
5.1 Analysis Discussion ........................................................................................................... 52
5.1.1 Homeless population distribution ................................................................................ 52
5.1.2 Accessibility analysis by Service Area ........................................................................ 57
5.1.3 2SFCA Accessibility analysis ...................................................................................... 58
5.1.4 Accessibility by Hospital ............................................................................................. 60
5.1.4 Distance measurements to hospitals ............................................................................ 61
5.2 Limitations ......................................................................................................................... 63
5.3 Future Research ................................................................................................................ 64
5.4 Final Conclusions .............................................................................................................. 66
References .................................................................................................................................... 68
vii
List of Figures
Figure 1. Los Angeles CoC boundary ............................................................................................. 5
Figure2. Los Angeles CoC boundary with census tracts ................................................................ 5
Figure 3. Hospitals and medical centers in the LA CoC ................................................................. 6
Figure 4 Homelessness by household type and sheltered status (2018) ....................................... 11
Figure 5 Estimates of homeless people by state (2018) ................................................................ 12
Figure 6 California homeless statistics (2018) .............................................................................. 12
Figure 7 States with highest rates of unsheltered homeless people (2018) .................................. 13
Figure 8. LAHSA and USC homeless source data ...................................................................... 16
Figure 9. Import and update shapefiles ......................................................................................... 35
Figure 10. First catchment ............................................................................................................ 36
Figure 11. Second catchment ........................................................................................................ 37
Figure 12. Choropleth map of total homeless population (# of individuals)
by LA CoC census tract ................................................................................................................ 39
Figure 13. Total homeless population (# of individuals) normalized
by total population of census tract ................................................................................................ 40
Figure 14. Hot spot analysis of homeless population by census tract .......................................... 41
Figure 15. Hospital service areas by accessibility index .............................................................. 43
Figure 16. Downtown hospital service areas by accessibility index ............................................ 44
Figure 17. Census tracts by 2SFCA accessibility index ............................................................... 45
Figure 18. Downtown LA census tracts by 2SFCA accessibility index ....................................... 46
Figure 19. Distance from census centroid to nearest hospital ...................................................... 49
Figure 20. Downtown Los Angeles by census tract with homeless population counts ................ 53
viii
Figure 21. Downtown Los Angeles with Skid Row boundaries ................................................... 54
Figure 22. Hot spot analysis of downtown LA and Santa Monica ............................................... 55
Figure 23. Regions of LA ............................................................................................................. 58
Figure 24. 2SFCA accessibility in Downtown Los Angeles ....................................................... 60
ix
List of Tables
Table 1. Major city CoCs with highest homeless population (2018) ........................................... 14
Source: AHAR 2018 ...................................................................................................................... 14
Table 2. 2019 Greater Los Angeles Point-In-Time Count conducted in January 2019 ................ 17
Source: LAHSA ............................................................................................................................. 17
Table 3. Datasets used by name, description, and source ............................................................. 27
Table 4. Curated fields of LAHSA Homeless Count 2019 by Census Results
by Census Tract ............................................................................................................................. 30
Table 5. Accessibility index for hospital service areas ................................................................. 42
Table 6. 2SFCA Accessibility Index ............................................................................................ 45
Table 7. Top ten hospitals by the total number of homeless individuals
within a 1-mile range .................................................................................................................... 47
Table 8. Distance measurements ................................................................................................... 48
Table 9. Total homeless population .............................................................................................. 48
x
List of Abbreviations
AHAR Annual Homeless Assessment Report
CA California
CD Council District
CoC Continuums of Care
CONUS Continental United States
DHHS Department of Health and Human Services
HMIS Homeless Management Information System
HUD Department of Housing and Urban Development
LA Los Angeles
LAHSA Los Angeles Homeless Services Authority
PIT Point-in -Time
USC University of Southern California
USICH US Interagency Council on Homelessness
VA U.S. Department of Veteran Affairs
2SFCA Two-step floating catchment area
xi
Abstract
Los Angeles has a homelessness crisis. The city has long struggled to meet the needs of the
growing homeless population, and the problem continues to amplify as the most recent 2019
Point-In-Time (PIT) Count shows an increase in homelessness. The Department of Housing and
Urban Development (HUD) Continuum of Care (CoC) federal grant program establishes
regional or local planning bodies to coordinate housing and services funding for homeless people
in an effort to promote an integrated system of care. As a local planning body, CoCs address the
issues their local communities and have the potential to affect positive change. Access to
healthcare is one such issue facing homeless populations that the LA CoC could better address
using spatial analysis, namely where homeless populations reside in the CoC boundaries relative
to established hospitals and medical facilities.
This project used a geographic information system (GIS) to assess the state of
homelessness in the Los Angeles CoC as of June 2019. A population distribution and density
analysis was conducted, indicating that homeless populations tended to be larger and more
concentrated in the census tracts comprising downtown Los Angeles and Santa Monica. To
determine the degree to which homeless individuals can access hospitals and medical facilities,
an accessibility analysis was conducted using a modified two-step floating catchment area
(2SFCA) methodology. The 2SFCA accessibility index indicated that census tracts within the
downtown area had homeless populations within a 1-mile distance of at least one hospital as
opposed to more rural tracts that tended to lack any access. However, access to medical facilities
within a walkable distance varied in the downtown census tracts. Recommendations for funding
allocation, the establishment of transportation initiatives, and additional medical facilities to
improve access were made.
1
Chapter 1 Introduction
Homelessness is a national issue that requires both federal and local government attention.
Federal spending on homelessness is distributed via programs and grants through a number of
agencies including HUD, the Department of Veteran Affairs (VA), Department of Education,
and Department of Health and Human Services (DHHS). One of the longest standing federal
programs addressing the issue of homelessness is the HUD CoC grant program which uses
funding to coordinate homelessness planning and response at the local level (Tinoco 2019). The
HUD CoC federal grant program establishes and funds local and regional level administrative
services that provide housing and essential services, including healthcare. This program is a
prime example of an effort that has had and can continue to have a positive impact in reducing
homelessness and improving the lives of those who have fallen into homelessness.
Within the state of California, homelessness is particularly prevalent in several cities and
counties, one of which being Los Angeles. The LA CoC is the local planning body tasked with
coordinating housing and service funding for homeless people in Los Angeles. The CoC can
have a significant impact on the state of homelessness in LA if equipped with the proper data. A
Greater Los Angeles PIT Count was conducted in January 2019, providing data on where
homeless populations are residing at the census tract level. Accounting for the census tracts
within the LA CoC administrative boundaries, an accessibility analysis of medical facilities was
conducted to provide LA CoC administrators and those concerned about homeless individuals’
access to medical care facilities a spatial study of accessibility.
2
1.1 Definitions
1.1.1 Homelessness definitions
The definitions of key terms throughout this study were based on definitions provided by
HUD (Henry et al. 2018). Individuals considered “homeless” were defined those who lacked a
fixed, regular, and adequate nighttime residence. An “individual” refers to a person who is not
part of a family with children during an episode of homeless, such as single adults or
unaccompanied youth. Homeless families and children are accounted for as individuals in the
total population counts of the PIT. Homeless individuals were considered “sheltered” if they
were staying in emergency shelters, transitional housing programs, or safe havens. “Unsheltered”
homelessness referred to people whose primary nighttime location was a public or private place
not designated for or ordinarily used as a regular sleeping accommodation. Examples of such
locations given were streets, vehicles, and parks (Henry et al. 2018).
1.1.2 CoC and PIT definitions
HUD defines a Continuum of Care as a regional or local planning body responsible for
coordinating the full range of homelessness services in a geographic area. The area may cover a
city, county, metropolitan area, or an entire state. The jurisdictions vary from state to state and
based on funding allocation. CoCs are generally composed of nonprofit service providers and
local government agencies, including health and human services, public housing agencies, and
other stakeholders (Center for Evidence-Based Solutions to Homelessness 2017). As part of their
commitment to the region they service and in order to meet funding requirements, CoCs are
tasked with conducting counts and surveys of the homeless population in their territory. Biannual
3
PIT counts of the homeless population is one such effort that CoCs conduct (National Alliance
2010).
Point-in-Time Counts are used in HUD homeless counts. HUD outlines a PIT as an
unduplicated one-night estimate of both sheltered and unsheltered homeless populations. One-
night counts are conducted by CoCs nationwide and occur during the last week in January of
each year (Henry et al. 2018). The specific methodology for the LA CoC PIT will be discussed
in the Methods Chapter.
1.1.3 Hospital and medical center definitions
The difference between a hospital and medical center depends on naming and branding by
healthcare providers. Medical centers and hospitals both offer a variety of services, including but
not limited to emergency services, primary care, and numerous specialties. Many consumers
differentiate the two based on their name. But contradictory to consumer opinion, there is no
functional difference between hospitals and medical centers (Rivkin & Bauman 2011). Thus
throughout this document, the labels are used interchangeably.
1.2 CoC Funding Allocation
HUD has developed specific guidelines as to how CoCs can acquire funding and what
information must be provided to advocate for additional funding. Funding is distributed after
considering the results of two community planning efforts that CoC administrators must prepare,
the Consolidated Plans and Continuum of Care Plans.
The Consolidated Plans outlines the framework for the CoC to identify housing, homeless,
community, and economic development needs and resources to develop a strategic plan that
4
meets those needs. This strategy lays out a three to five-year plan to implement the proposed
efforts that requires the funding requested. Homeless population size and level of community
need are crucial factors considered when allocating funds.
The second documentation required by HUD is the Continuum of Care Plan which details the
housing and services proposed to meet the needs of the homeless as they move toward stable
housing and maximum self-sufficiency. The CoC Plan focuses on providing actionable steps to
end homelessness and prevent a return to homelessness. A majority of HUD’s homeless
assistance funds are awarded based on the CoC Plan (NCHV 2019).
Identifying local funding priorities and areas of need is a critical part of this plan. The current
project seeks to identify areas within the LA CoC with large homeless populations and where
access to healthcare is limited to pinpoint areas, medical facilities, and populations with
demonstrated need where additional funding could prove beneficial.
1.3 Study Area
The state of California has a total population of 39.6 million, with 10.1 million, roughly 25
percent, of those individuals live in Los Angeles County (U.S. Census Bureau 2018). The
County is approximately 4,084 sq. mi (California Department of Finance 2018). The City of Los
Angeles is the largest city out of the 88 cities contained within LA County. The LA CoC shares
boundaries with the county of Los Angeles, encompassing the same neighborhoods and census
tracts. Pasadena, Glendale, and Long Beach are exceptions, as each have their own CoCs (Figure
1). In total, 2,161 census tracts compose the LA CoC (Figure 2).
5
1.4 Hospitals and Medical Centers
Los Angeles County maintains a database of officially recognized and accredited hospitals
and medical facilities within Los Angeles. From this data source, 147 hospitals and medical
facilities fell completely within the boundaries of the LA CoC (Figure 3). These facilities were
considered when determining which hospitals and medical facilities are accessible to those
residing in the LA CoC. Facilities falling outside the boundary were not included even if they
were close to the border as they were not in the jurisdiction of the LA CoC, therefore, CoC funds
would not be applied to increase access to these facilities.
Figure 2. Los Angeles CoC with census tracts
Figure 1. Los Angeles CoC boundary
6
1.5 Data-driven homelessness studies
This study compares the locations of homeless populations with medical facilities and uses
accessibility analysis to determine if the current resources are physically accessible to this
subsection of the Los Angeles population. The results of this analysis are intended to be of use to
LA CoC administrators when drafting the next CoC Plan for funding justification. Identifying
areas and hospitals within the CoC that have large populations and/or low accessibility helps to
inform where resources like transportation services and additional medical facilities should be
stood up to improve medical care accessibility for homeless people.
Los Angeles commonly makes headlines for its homeless crisis. However, Los Angeles’
efforts to address homelessness have increased over the years in response to the issue (Oreskes
Figure 3. Hospitals and medical centers in the LA CoC
7
2019). Los Angeles supports a variety of collaborative partnerships and projects, engages with
the community, and pursues funding at various levels of government. Data-driven spatial
analysis can help us understand what homelessness looks like in Los Angeles. This enables
decisionmakers and interest groups to determine how to intervene most effectively.
An incredibly valuable, but not fully utilized source of data is the most recent 2019 PIT Count
conducted for the LA CoC. Part of a HUD-mandated yearly PIT Count, the count creates an
estimate of the number of homeless individuals by location, including demographic
characteristics of sheltered and unsheltered individuals. This project leverages the statistical,
demographic, and spatial data collected to provide an evaluation of the accessibility of LA
medical facilities to the homeless population. Considering homeless populations when
deliberating the allocation of funds for medical services or the establishment of a service like a
shuttle transport program or new medical resource enables more informed decision-making on
the part of program administrators. Having specific medical facilities and regions in their CoC in
mind in terms of where more funding is required also provides greater justification to HUD in
the LA CoC’s plan documentation required for grants.
Identifying where homeless populations are located and how their location determines access
to healthcare would benefit from the application of GIS. GIS helps visualize a phenomenon
spatially and highlight the crucial role location plays. This project contributes to existing
literature by analyzing the most up-to-date data within the LA CoC. The focus on accessibility to
healthcare has historically been less of a focus in the LA CoC relative to housing, so it is the
hope that program administrators utilizing this data would gain a geospatial assessment of
healthcare accessibility they may not currently have access to. Measuring the walkability of these
medical facilities is also less common within assessments of hospital accessibility, typically
8
considering motor transportation, but is nonetheless important and should be considered for
individuals who do not have access to motor transportation (Vale et al. 2016).
1.6 Motivation
This project provides data-driven geographic analysis of the homeless population in the Los
Angeles CoC and the degree to which the homeless can access to healthcare facilities.
Ultimately, the intent of the project is to provide a survey of homelessness in the LA CoC and
determine hospital access within the CoC based on walkability to a hospital from where
individuals dwell. PIT counts are the primary means of data collection on homeless populations
and the results of these studies are often used as indicators of government programs’
effectiveness and used to justify resources allocation (Grumdahl 2019).
While PIT counts have obvious limitations as people are not stagnant points, the prevalence of
this type of data and the current acceptance of this method by government organizations made it
well-suited for the current study that aims to provide analysis of use to government bodies. In
this initial analysis, areas with significant homeless populations and poor accessibility index
ratings were identified to recommend where the LA CoC should add resources. This study
utilized the most recent LA CoC PIT count which occurred January 2019, thus providing up-to-
date information to LA CoC administrators and other interested parties.
The motivation for this project was to assist the LA CoC and other organizations, agencies,
and individuals trying to better serve homeless populations. Identifying areas within the LA CoC
with large homeless populations and/or low access to medical resources could be included in the
next CoC Plan report submitted to HUD for funding justification. It is the hope that conducting
this research will address healthcare accessibility, a major issue facing homeless individuals and
9
affecting their quality of life, and will encourage further research into this topic (Khandor et al.
2011). Most importantly, this project aims to draw conclusions to benefit homeless individuals
who lack access to medical facilities and help provide them with the resources they need.
10
Chapter 2 Related Work
This project analyzed 2019 PIT data collected by LAHSA to provide insight into healthcare
accessibility for LA’s homeless population. Studies pertaining to homelessness throughout
California and Los Angeles are referenced for context. The HUD CoC program is discussed and
the efforts of the Los Angeles CoC are included to demonstrate past and current efforts and foci
of the program. The inclusion of details about the HUD program includes recommendations so
that they assist in current efforts and can be further implemented by the CoC in future
interventions. The existing body of accessibility analysis research is discussed and adapted to the
particular challenges of homelessness.
2.1 Homelessness in the United States
The most recent HUD 2018 Annual Homeless Assessment Report (AHAR), presented to
Congress, provides an annual evaluation of homelessness and PIT estimates throughout the
country. PIT counts were collected for the 398 CoCs, which cover nearly the entire United
States. One-night counts were conducted during the last 10 days of January 2018 to provide an
estimation of the number of people experiencing homelessness. The first part of the study
provides the results of the PIT estimates of sheltered and unsheltered homelessness on a single
night.
The 2018 HUD AHAR study had several key findings that provide a snapshot of
homelessness (2018). On any given night in 2018, roughly 553,000 people experienced
homelessness in the United States. Those included in this estimate were individuals staying in
sheltered locations (emergency shelters or transitional housing programs) which comprised 65%
of the total. The remaining 35% of the total homeless population were residing in unsheltered
11
locations such as on the street, in cars, in abandoned
buildings, and in other locations deemed not suitable
for human habitation (Figure 4). The results of this
PIT count indicate a modest increase (0.3 percent
increase in total population and 2 percent increase in
unsheltered populations) in homelessness for the
second year in a row (Henry et al. 2018).
2.1.1 State estimates
When looking at homelessness across the country, the distribution of the population is not
evenly divided between states, as certain states account for a disproportionate number of
homeless individuals. Over half of all people experiencing homelessness were in one of the
following five states: California, New York, Florida, Texas, or Washington (Figure 5). Of those
states, California has the largest homeless population. 129,972 individuals were homeless in
California according to the results of the national PIT homeless count conducted by the United
States Interagency Council on Homelessness in 2018 (Figure 6) (USICH, 2018). Of California’s
homeless population, CA has the greatest percentage of unsheltered homeless with well over half
of its homeless population being unsheltered (Figure 7).
Figure 4. Homelessness by household type and
sheltered status (2018)
Source: AHAR 2018
Figure 1 Homelessness by household type and
12
Figure 5. Estimates of homeless people by state (2018)
Source: AHAR 2018
Figure 2 Estimates of homeless people by state (2018)
Figure 6. California homeless statistics (2018)
Source: USICH 2018
Figure 3 California homeless statistics (2018)
13
2.1.2 CoC estimates
In the PIT counts, it was found that over half of all unsheltered homeless people are in CoCs
that include the nation’s 50 largest cities. The Los Angeles City and County CoC have the
second largest homeless population behind New York City (Table 1). At the time of this count,
more than 1 in 5 people experiencing homelessness were living in New York or Los Angeles
(Henry et al. 2018). Los Angeles had one of the highest rates of unsheltered homeless with 75
percent of the Los Angeles homeless population being unsheltered. The fact that a majority of
Los Angeles’ homeless population is unsheltered is uncharacteristic of urban CoCs, which tend
to have the highest percentage of sheltered people. The unsheltered nature of LA’s homeless
population presents increased risks and needs that the planning body is tasked to address.
Figure 7. States with highest rates of unsheltered homeless people (2018)
Source: AHAR 2018
Figure 4 States with highest rates of unsheltered homeless people (2018)
14
2.2 Homelessness in the Los Angeles CoC
2.2.1 LA CoC efforts
The Los Angeles CoC receives the most funding from HUD within the state of California,
totaling $123,707,061 for LA City & County CoC during FY18 (Homeless Assistance Award
Report 2018). The Los Angeles Homeless Services Authority (LAHSA) is the lead agency for
the LA CoC, coordinating and managing more than $243 million annually in federal, state,
county, and city funds (LAHSA, 2018). As a CoC, a local Homeless Management Information
System (HMIS) is maintained to collect and report data on the characteristics of those
comprising the LA homeless population. The CoC program’s service use pattern for the
resources is also collected and measured. Reports such as those compiled by the CoC help justify
current and future funding and facilitate community-wide awareness of homelessness (LAHSA
2019).
In June 2019, Los Angeles officials released the results of the 2019 Greater Los Angeles
Homeless Count. The trend was dramatic. Homelessness increased in LA County by 12 percent
Table 1. Major city CoCs with highest homeless population
(2018)
Source: AHAR 2018
15
and in the City of Los Angeles by 16 percent from 2018 to 2019. Based on the most recent data,
LA County’s homeless population sits at approximately 58,936 in the county and 36,300 in the
city (LAHSA 2019). But while the numbers may not reflect it, government initiatives taken by
the city of Los Angeles made great strides in addressing the issue. LA has increased program
funding to provide more affordable housing and access to essential services (Cowan 2019).
2.2.2 LA CoC homeless count methodology
LAHSA partnered with USC’s School of Social Work and the Leonard D. Schaeffer
Center for Health Policy & Economics to design, implement, and analyze the Greater Los
Angeles Homeless Count. The study provides PIT estimates of the homeless population in the
LA CoC geographic area, fulfilling HUD’s requirement for CoCs that an annual count and
demographic characteristics estimates occur.
To generate count results, various sources of data were used (Figure 8). Estimates of the
homeless population were extrapolated from data obtained via street counts of unsheltered
people, a demographic survey of unsheltered adults, a youth count and survey, administrative
data from the HMIS, and the MyOrg data collection system as it pertains to sheltered individuals.
LAHSA conducted street counts by visually counting and recording people experiencing
unsheltered homelessness, including those dwelling in cars, vans, RVs, tents, and makeshift
shelters. These were conducted in all 2,160 census tracts of the LA CoC and was collected
during the last 10 days of January 2019 for temporal consistency.
Demographic surveys were conducted on a sample of homeless adults. These surveys
were used to estimate the characteristics of unsheltered homeless adults across the CoC and to
determine number of people living in makeshift shelters captured in the Street Count. LAHSA
expressed concerns about the accuracy of representation of homelessness at a large geographic
16
scale. This prompted USC to run preliminary estimates of sample
sizes for different levels of standard error and precision at
multiple geographic levels. For sample selection, the USC
methodology implemented a two-stage stratified random sample.
A decision was made to take council districts (CDs) into account
when defining geographic strata. A reported 5 percent margin
error was used to calculate a target sample size for surveys and
the prior year’s average population estimates were used to define
the final sample size per census tract. A shelter count was
conducted to provide the raw number of homeless individuals
living in various shelter types. For the shelter count, LAHSA
conducted a PIT that had a 100 percent enumeration of all shelters
in the CoC, for which LAHSA asserted there was no sampling or
sampling error.
HMIS data was used to estimate subpopulations
(individual v. family, type of shelter, etc.) and demographic
estimates. Demographics collected included household
composition, veteran status, gender, and age. Complete HMIS
records with demographic characteristics were used to derive the
distribution of demographic and subpopulation characteristics for the sheltered homeless
population. Eliminating collection redundancies and screening for shelter used, demographic
characteristics, and subpopulations was generated by estimating the proportion within the HMIS
data for each type of shelter and household type (Henwood et al. 2018).
Figure 8. LAHSA and USC
homeless source data
Figure 5 LAHSA and USC
17
2.2.3 Homelessness in LA by the numbers
In the LA CoC, there was a total homeless population of 58,936 people in 2019, indicating a
12 percent increase in the total homeless population in LA County from the PIT count of 2018.
This percentage change was deemed significant by LAHSA according to a significance test at the
95 percent confidence interval (LAHSA Count 2019). In addition to Los Angeles, Orange
County, Ventura, San Bernardino, and Kern counties all reported significant increase in their
respective homeless populations (Oreskes & Smith 2019). Veterans accounted for 7 percent of
the total homeless population. Men were the majority gender represented, with 67 percent of the
population. With respect to age, the vast majority (85 percent) were 25 years of age or older
(Table 2).
Population
1
Sheltered Unsheltered Total
Prevalence of
Homeless Pop. (%)
Percent Change
2018 ‐ 2019
Significant
Difference
2018 ‐ 2019
2
All Persons 14,722 44,214 58,936 100% +12% Yes
Individuals (Those not in family units) 7,590 42,481 50,071 85% +13% Yes
Chronically Homeless 1,517 14,337 15,854 27% +17% Yes
Veterans 965 2,874 3,839 7% +1% No
Unaccompanied Minors (Under 18) 21 45 66 0.1% +5% No
Family Members (Those in family units) 7,111 1,688 8,799 15% +6% Yes
Children in Families (Under 18) 4,322 892 5,214 9% +6% Yes
Chronically Homeless 474 200 674 1% +31% No
Veterans 53 59 112 0.2% +19% Yes
All Veterans 982 2,896 3,878 7% ‐0% No
Chronically Homeless Veterans 92 1,208 1,300 2% ‐15% No
Male 7,940 31,408 39,348 67% +11% Yes
Female 6,634 11,697 18,331 31% +13% Yes
Transgender 125 932 1,057 2% +14% No
Gender Non‐Conforming 23 177 200 0.3% +14% No
Under 18 4,343 937 5,280 9% +6% Yes
18 ‐ 24 1,511 2,124 3,635 6% +17% No
25 and Over 8,868 41,153 50,021 85% +12% Yes
Individuals (Those not in family units) 1,517 14,337 15,854 27% +17% Yes
Family Members (Those in family units) 474 200 674 1% +31% No
Total Chronically Homeless Persons 1,991 14,537 16,528 28% +17% Yes
Health/Disability Indicator
3
Sheltered Unsheltered Total
Prevalence in Over 18
Homeless Pop. (%)
Percent Change
2018 ‐ 2019
Significant
Difference
2018 ‐ 2019
2
Substance Use Disorder 859 6,977 7,836 15% +10% No
HIV/AIDS 315 991 1,306 2% +76% Yes
Serious Mental Illness 2,278 11,392 13,670 25% +7% No
Domestic/Intimate Partner Violence Sheltered Unsheltered Total
Prevalence in Over 18
Homeless Pop. (%)
Percent Change
2018 ‐ 2019
Significant
Difference
2018 ‐ 2019
2
Homeless Due to Fleeing Domestic/Intimate
Partner Violence 327 2,784 3,111 6% +1% No
Household Composition
2019 Greater Los Angeles Homeless Count ‐ Data Summary
Los Angeles County
1
All Persons
Notes:
1. The Los Angeles County Data Summary includes Long Beach, Pasadena, and Glendale.
2. Significance tested at the 95% confidence interval.
3. Health/Disability indicators are not mutually exclusive (a person may report more than one). Numbers will not add up to 100%.
Prepared by Los Angeles Homeless Services Authority (June 2019).
Data from 2019 Greater Los Angeles Point‐In‐Time Count conducted in January 2019. Visit http://www.lahsa.org/homeless‐count/ to view dashboards.
Domestic/Intimate Partner Violence
Health and Disability
Chronically Homeless
Age
Gender
Veterans
Table 2. 2019 Greater Los Angeles Point-In-Time Count conducted in January 2019
Source: LAHSA
18
2.3 Access and utilization of medical care
Much of the focus for the LA CoC and other organizations working to address LA’s homeless
crisis is on the provision of housing. A lack of stable housing is the primary issue faced by
homeless individuals and is rightly prioritized (Sadowski et al. 2009). The phenomenon of the
prioritization of homeless individuals’ needs has long been recognized by researchers.
Competing priorities refers to the phenomenon that basic needs such as the need for food,
shelter, and safety tends to be prioritized over needs such as healthcare based on perceived
necessity and importance to daily life. A study by Gelberg et al. (1997) offered empirical support
for the phenomenon, recognizing the nonfinancial barriers to utilization of health services by
homeless individuals. Housing also has a direct effect on other critical life-sustaining necessities
like access and utilization to medical care.
LA CoC services have mirrored this prioritization, with the largest funding grants going to
housing projects in FY19 (FY2019 LA CoC Project Priority List). In addition to funding for the
housing issue, issue of access to medical care could be better demonstrated through an
accessibility analysis to provide the documentation and identification of specific locations that
would benefit from additional resources.
Kushel et al. 2011 found various factors associated with health care utilization by homeless
people. The authors found that homeless individuals experience high rates of physical illness,
mental illness, substance abuse, and early mortality. However, despite having a higher burden of
illness, homeless people have fewer encounters with ambulatory care than non-homeless
individuals. Ambulatory care refers to medical services performed as an outpatient without
admission to a hospital, including specialty clinics and urgent care clinics (Heinrich 2017). The
19
authors associate this occurrence with the fact that health care directly competes with more
immediate needs, such as obtaining shelter and food. The authors found that, given the
opportunity, homeless individuals are willing to obtain health care for chronic conditions if they
believe such care is important. Access was found to be a critical component of the decision to
seek medical care. Of those surveyed in the study, one fourth of respondents reported that at
some point in the past year they needed medical care that they had not been able to receive
(Kushel et al. 2001).
2.3.1 Accessibility
Defining “accessibility” is necessary to determine the usability of healthcare clinics for
homeless populations. Various factors contribute to accessibility, with unique considerations in
determining accessibility for homeless adults. Access to health care is influenced by a multiple
phenomenon, including the availability of health services in the area, the number of people living
in the area, the population’s health status, and its socio-economic standing (Chan et al. 2014).
Chan et al. (2014) looked at accessibility and community integration among homeless
individuals. The researchers were motivated by the recognition there had been few empirical
investigations into the proximity of community features on resource use and integration. GIS was
used to examine how accessibility and proximity to community features related to the types of
locations homeless people were able to access within the community. Overall, the authors
concluded that the ability to navigate and use community resources was associated with better
accessibility and feeling a part of the community.
A relevant concept from the Chan et al. study is the difference between “potential
accessibility,” which centers on probable utilization of services, as compared to “revealed
accessibility,” which documents the actual use of services (Chan et al. 2014). Determining the
20
potential accessibility of clinics considers the homeless population within a determined
geographic and travel time span and assumed various methods of travel, including foot, public
transit, and/or motor vehicle. Of those deemed to be able to use the resources, the percentage of
individuals in that area who utilize the services can be determined from statistical data. Spatial
accessibility is a factor considered for these two types of accessibility. The measurement of
distance and time, as well as the importance of spatial separation between supply and demand as
a barrier or facilitator of use are also key considerations (Wang 2012). These types of
accessibility were considered when determining the accessibility of healthcare facilities in this
thesis.
Potential accessibility is considered in this thesis as the entire homeless population within
a determined walking distance from a hospital is considered to have access to the medical
facility. Potential accessibility is more inclusive of a possibly larger group of individuals as it
may overestimate the ability of all within a given range to actually access the facility. For the
purposes of this study to supply the LA CoC with estimations of need according to area and
hospital, it was determined that being more inclusive would be beneficial to advocate for funding
in terms of all people the additional resources could potentially help.
2.3.2 Walkability
Yang and Diez-Roux et al. 2012 conducted a study to determine the acceptable walking
distance (typically determined to be 0.25 mi) in U.S. research studies based on distance and
purpose of movement. The researchers found an inverse correlation between longer walking
distances and socioeconomic status. Of their nationally representative sample, the distribution of
walking trips by distance had the highest frequency for 1 mile and nearly one-fifth of the sample
21
walked more than a mile (Yang & Diez-Roux et al. 2012). Considering the findings of this study,
it could be inferred that he homeless population faces similar challenges to those of low
socioeconomic status, with lack of funds for motor transportation and only occasion public
transport options. The distance to a destination in “walking distance” for a homeless individual
could subsequently be a longer distance than would typically be considered accessible.
2.4 2SFCA
Determining place-based accessibility via modeling has long been used to study the
accessibility of a location with respect to its intended audience. The two-step floating catchment
area model determines the accessibility of a location according to the density and/or proximity to
surrounding target locations (Neutens et al. 2010). While 2SFCA was originally used to study
healthcare accessibility (Luo and Wang 2003), the approach has been applied to a wider range of
accessibility studies (Chen & Jia 2019).
2.4.1 2SFCA methodology
Luo and Wang’s 2SFCA method is a form of the gravity model, which considers accessibility
to be mediated by distance decay and the interactions between supply and demand. The authors’
methodology integrates spatial and non-spatial factors that affect accessibility to provide a more
accurate representation of the phenomenon. As Wang (2012) stresses, both spatial and non-
spatial factors must be accounted for to develop a complete picture of health care accessibility.
Utilizing GIS analysis, spatial access emphasizes the importance of spatial separation between
the supply (the medical facility) and demand (the population). Non-spatial factors affect the
spatial component, as demographics and socioeconomic status relate to location and influence
how an individual and/or group engages with space.
22
This method was originally developed to evaluate spatial inequity of health care services,
measuring the cost (distance, time, money, etc.) associated with getting to a medical facility
based on one’s location. Since the authors first implemented the methodology, its application has
been seen in the field of urban planning and other fields modeling accessibility spatially. Several
researchers have also modified the methodology to improve accuracy and account for important
factors in their studies (Yang et al. 2006).
2.4.2 Enhanced 2SFCA methodology
Luo & Qi (2009) introduced an 2SFCA method in which weights were applied to differentiate
travel time zones to account for distance decay. The consideration of distance decay and factors
that affect accessibility measurements are commonly considered in current 2SFCA
methodologies, including the present thesis. McGrail and Humphreys (2009), amongst others,
adjusted the 2SFCA method to study rural areas. The authors of this study attempted to rectify
two shortcomings of the 2SFCA, namely the use of only one catchment size for all populations
and the assumption that proximity is undifferentiated within a catchment. Rural communities are
the subject of the study, as they are often characterized by poorer health status and increased
problems of accessing health services compared to cities (Humphreys and Solarsh 2008).
Distance barriers and diminished local availability of health care services are identified as
common accessibility issues in these areas.
McGrail and Humphreys’ 2SFCA approach of tailoring the method to their subject matter
provides an example for how to customize the 2SFCA. This study also emphasized the
importance of considering one’s area and topic of focus, and how a unique subject impacts how
the study should be conducted. Identifying the weaknesses of the 2SFCA method also enables
improvements to be made. Rural areas face similar disparities as homeless individuals, such as
23
limited resources within close proximity and a higher cost (time, distance, etc.) associated with
gaining access. Thus this study serves as a good example for the current project when designing
the 2SFCA methodology to account for the constraints on homeless individuals’ mobility and
financial resources.
Studies using 2SFCA studies typically look at health care accessibility for a geographic
region’s population. Variations exist with respect to methodology, subject, and type of provider,
but using this technique for the study of homeless accessibility is less common. This gap in
application of the methodology may be due in part to the difficulty of collecting estimates on this
population and the lack of spatial analysis on the topic of homeless healthcare accessibility.
Defining access in terms of walkability is one such modification that must be made, assuming a
lack of motorized transport. Accounting for the nonspatial attributes of homelessness, spatial
analysis via the 2SFCA has the potential to produce valuable results that researchers could
leverage by using GIS.
2.5 Summary of Related Works
A literature review of the studies conducted in the United States to determine homeless
counts provided the necessary context and justification for choosing the LA CoC as a location
where the study of homelessness is timely and relevant. Understanding the methodology of prior
research offered guidance for the current project. This background research established the
credibility of this current thesis by verifying that the data used was collected and disseminated in
a scientifically rigorous manner. The studies referenced influenced the methodology of the
current study and helped to validate the results. From the PITs conducted, a picture of
24
homelessness in LA established an understanding of the population being considered by the
current study.
Prior studies of accessibility provided guidelines when measuring health care accessibility as
it applies to homeless populations. From this study, the 2SFCA methodology emerged as the
preeminent technique. The original and adapted version of this methodology serve as examples
for the modified 2SFCA used in the current study, though it was tailored to the study of homeless
access to health care facilities.
25
Chapter 3 Methods
The objective of this study was to determine regions in the Los Angeles CoC that have hospitals
accessible to the homeless population, and where there are concentrations of homeless
individuals with limited or no access to medical facilities. Using the most recent LAHSA
Homeless Count 2019, concentrations of homelessness in the LA CoC were determined at the
census tract level. Studies of homelessness often use census tracts as the spatial scale due to the
prevalence of data available at this scale and the utility of studies at this scale to policymakers to
understand larger areas they serve (HRI, 2014). A series of analytical steps using ArcGIS were
conducted to determine population densities to evaluate the status of their accessibility. The
distances between the homeless populations and the nearest hospital were also calculated.
To conduct a healthcare accessibility analysis, a 2SFCA method was used to determine
accessibility for homeless populations within a determined area surrounding the medical
facilities. The accessibility index defines accessibility as the ratio of total homeless individuals
within service range of healthcare facilities, considering the total bed count of these medical
facilities per the total homeless population. As the ratio approaches 0, the higher the demand
based on hospital bed availability per the homeless population within range. Areas of greater
service need and were access was not available were identified to highlight were additional
funding could improve access by providing transportation services or additional resources.
The service area distances considered were 0.25, 0.5, and 1-mile walking distances.
These distances were chosen as 0.25 miles is typically viewed as an accessible distance to walk.
1-mile was determined to be more inclusive of homeless populations with potential access as
homeless individuals typically have to walk longer distances to access resources. 0.5 miles
served as a natural in-between distance to include more an additional measurement.
26
The homeless population counts by census tract were distributed by areal extent to
account for the coverage of the service areas and to provide an estimated distribution of the
population outside the centroid of the census tract. Population distribution by areal extent
accounts for the population of homeless individuals that are in the intersection of the two feature
classes (the census tract layer with population data and the service areas). This percentage of
coverage was than applied to the total population number to determine the number of individuals
falling within the service area.
Distances between the centroid of the census tract and the healthcare facility were
measured to determine the distance via walkable routes to the nearest facility. The centroid of the
census tract was used as exact points for the homeless populations were not available and the
center point served as a uniform method for distance measurement. Euclidean distance was used
because GIS tools available could not provide walkable distance for the number of hospitals
being concerned.
3.1 Data
The data used for this project was collected and provided by various government
organizations at the federal and local level. The data came as shapefiles and Excel sheets that
were geocoded to produce shapefiles in ArcGIS pro (Table 3).
27
3.1.1 Homeless Count 2019 Results by Census Tract dataset
The LAHSA 2019 Greater Los Angeles Homeless Count provides Point-In-Time
estimates of the homeless population in the LA CoC geographic area as defined by HUD.
Estimates of the homeless population were extrapolated from data obtained by a street count of
the unsheltered and sheltered populations and further refined by demographic categorization
(Henwood et al. 2018). LAHSA collected population counts and accompanying demographic
data for the 2,160 census tracts in the LA CoC that are accounted for in the current study.
The dataset provided numerous subcategories of the homeless population identified in the
CoC census tracts (various living situations, shelter status, etc.). The primary fields of interest for
the current study was the tract code, community name, and total homeless population
(accounting for sheltered and unsheltered individuals). The data was provided as an Excel sheet
which was cleaned to remove additional fields.
Dataset Name Description Source
Homeless Count
2019 Results by
Census Tract
Captures a Point-In-Time (PIT) estimate of unsheltered,
emergency sheltered, transitional housing, safe haven,
and total homeless population. Collected at the census
tract level on the last 10 days of 2019 from volunteer
counts and surveys. Available as an Excel dataset.
Los Angeles Homeless
Services Authority
(LAHSA)
Los Angeles
Hospitals and
Medical Centers
(2019)
Dataset of hospitals and medical centers in Los Angeles
County as a shapefile. Contains various attributes such
as site location, service type, total bed count, and
contact information.
Location Management
System (LMS) County of
Los Angeles GIS Program
Los Angeles
County Census
Tracts (2010)
Data was downloaded from the Census Bureau website
and clipped to LA County boundary. Census data from
the 2010 census updated to the 2012 Census Geography
Update.
Los Angeles County GIS
Data Portal and U.S.
Census Bureau
Los Angeles CoC
Boundaries
shapefile (2018)
Geographic boundaries for HUD’s CoC areas by Year.
Shapefile for California CoC provided with the Los
Angeles CoC (CA-600) selected.
U.S. Department of
Housing and Urban
Planning (HUD)
Table 3. Datasets used by name, description, and source
28
3.1.2 Los Angeles Hospitals and Medical Centers (2019) shapefile
The LA Hospitals and Medical Centers shapefile was obtained from the Location
Management System (LMS) via ArcGIS Hub. LMS is the County of Los Angeles GIS program
that maintains a comprehensive geographic database of locations countywide. Data on the
location as well as descriptions and contact information for each medical facility was provided.
The dataset contains all hospitals and medical facilities in the county of Los Angeles. Of the
complete dataset, 147 hospitals and medical facilities that fell completely within the CoC
boundaries were selected to create a layer in ArcGIS. Attributes of particular interest included
the location, services provided, care category, name of facility, and total bed count. The total bed
count was used in the 2SFCA to determine the ratio of total beds per total homeless population in
the service area. The facilities were presented as points and the symbology was changed to red
crosses to indicate a medical facility.
3.1.3 Los Angeles Census Tracts (2010) shapefile
The census tract data for Los Angeles County was provided by the Los Angeles County
GIS Data Portal using U.S. Census Bureau 2010 census data. Though the census data was
collected in the latest 2010 census, the boundaries of the census tracts were updated to the 2012
Census Geography Update. Therefore, the CT10 (census tract ID number) field reflected the
2012 Census update. The shapefile was projected in NAD 1983 StatePlane California V FIPS
0405 Feet. The boundary shapefiles were the primary data of interest as opposed to the
demographic data provided as the Homeless Count PIT demographic data was the population
data of interest.
29
3.1.4 Los Angeles CoC shapefile (2018)
The shapefile for the Los Angeles CoC came from HUD’s Continuum of Care GIS Tools.
Since HUD provides competitive funding for homeless services through a CoC structure, all
research submitted to acquire funding at the CoC level must use the geographic boundaries and
related data provided by HUD (HUD Exchange 2018). This dataset was selected to ensure this
project could be used as support for funding and meet the standards of HUD. The shapefile
contains the geographic boundaries for all CoCs within the state of California projected in GCS
North American 1983. Los Angeles is CoC number 600 within the state of California. CoC
number CA-600 was selected by attribute and placed into a separate layer to focus the study area
to Los Angeles.
3.2 Visualization of homelessness in Los Angeles
Various ArcGIS symbologies were used to visualize the spatial relationship between
homelessness and location within Los Angeles. Choropleth maps were created to visualize the
total homeless populations by census tract. A hot spot analysis map was prepared to display areas
of concentration of homeless populations and analyze the density distribution of the homeless
population in LA CoC.
3.2.1 Homeless population counts choropleth map
Homelessness counts at the census tract level were used to map the CoC’s homeless
population. The most recent LAHSA Homeless Count 2019 Results by Census Tract dataset was
used for population data. The dataset was curated to only include data fields of interest, namely
the tract number and the total population of sheltered and unsheltered homeless people (Table 4).
30
This dataset was imported into ArcGIS Pro as a table. The U.S. Census Tracts shapefile was
imported into ArcGIS for the census tract polygons. The Join tool was used to merge the
LAHSA data with the census tracts polygons based on the tract ID attribute. The updated census
tract layer with the homeless data was then spatially joined with the Los Angeles CoC shapefile
(2018) to only include census tracts completely within the CoC. The projection for both layers
was set to NAD 1983 StatePlane California V FIPS 0405 Feet.
The census tracts that returned a Null value when merged were deleted in the attribute
table as this indicated the census tract was not included in the CoC. The symbology of the CoC
census tract layer was set to Graduated Colors with the total homeless population field selected, a
Normal Breaks (Jenks) method, and five classes with a teal blue (small population) to bright pink
(large population) color scheme. A choropleth map of the total homeless population normalized
by the total population in the census tract was also included.
3.2.2 Homeless population counts hot spot analysis map
To create a hot spot analysis of where homeless populations reside in LA CoC, the CoC
census tract layer with the LAHSA homeless data was used. To create points that could be used
Column Name Description
tract 2010 US Census Tract Code
Year Year Tract was Counted
City City Name
Community_Name Community Name
totUnsheltPeople Total population of unsheltered homeless persons
totSheltPeople Total population of sheltered homeless persons
totPeople Total population of sheltered and unsheltered homeless persons
Table 4. Curated fields of LAHSA Homeless Count 2019 by Census Results
by Census Tract
31
to make the hot spot map, the centroids of the census tracts were used as the location of the
homeless population. This method was used as the exact location of the homeless individuals
were not recorded in the count. The assumption was made that the population aggregates at the
center of the census tract, so the XY Table to Point Data Management tool was used with the
fields X_Center (longitude) and Y_Center (latitude). Using the LA Census Tracts data, the
centroids of the census tract polygons were found to create the
Los_Angeles_Census_Tract_Centroid layer by using the XY Table to Point tool with the
X_Center and Y_Center fields.
The ArcGIS Optimized Hot Spot Analysis tool was used to create the hot spot map. The
total homeless population field of the centroid layer was analyzed to determine where the largest
concentration of homeless people was. The results are represented as the centroids color graded
according to accessed hot spot to cold spot values. The symbology of the hot spot analysis map
was a gradation from cold spot (blue) to hot spot (red).
3.3 Two-step floating catchment accessibility methodology
A 2SFCA methodology was used to measure healthcare accessibility for the homeless
population in Los Angeles CoC. When conducting analysis for Los Angeles CoC, Glendale,
Pasadena, and Long Beach were excluded as each has its own CoC. Catalina and San Clemente
Island were removed to contain the analysis to the Continental United States (CONUS). Two
areas of coverage, or catchments, were created in this multi-step process and layered over each
other, “floated”, to produce representations of spatial accessibility. The methodology used in the
current study is adapted from the Vo et al. (2015) 2SFCA process. The two steps included the
following:
32
Step 1: Given a 0.25, 0.5, and 1-mile distance (catchment) from a healthcare provider,
sum up the total homeless population that the provider can reach within that distance.
Determine a provider-to-population ratio. In the case of each medical facility, the provider
ratio is the total bed count of the hospital to the total homeless population reached.
Step 2: Obtain the previously computed provider-to-population ratio of each healthcare 1-
mile service area. Compute accessibility index of the census tract by summing up all
provider-to-population ratios of the hospital service areas that fall within the boundaries of
the census tract .
3.4 2SFCA Methodology Caveats
Modifications were made to the Vo et al. (2015) 2SFCA methodology. Analysis was
informed by other practitioners’ methodologies, and adaptations were made to accommodate the
unique considerations when looking at homeless accessibility. ArcGIS Pro was used for analysis.
3.4.1 Walkable distance determination
When determining the catchment size, a service area of 1-mile was chosen to represent the
area surrounding the hospital considered to be within “walking distance.” The distance of 1-mile
was selected based on existing literature and from considering the nature of the study population.
Drawing on studies of walkability such as that conducted by Yang & Diez-Roux et al. 2012, it
was determined that 1 mile was an acceptable distance to classify as accessible considering the
transient lifestyle of homeless individuals.
33
A 1-mile service area around the healthcare facility was used to determine the homeless
population within range of the facility. To account for the standard 0.25-mile distance, the
service areas were broken up into 0.25, 0.5, and the 1-mile outer limit to allow for additional
inferences to be made about accessibility based on the varying distances. The service areas were
created using ArcGIS Network Analyst to only account for walkable paths, assuming homeless
individuals would lack motor transportation.
3.4.2 Census tract population by areal extent
To determine the number of homeless individuals within a given service area, population
by areal extent of the service area was used. The ArcGIS tool “Tabulate Intersection” was used
weight census tract population by areal extent that was covered by the 3-ring (0.25, 0.5, 1 mile)
service areas to allocate the population. The areal extent was determined to be the percentage of
each census tract within a given service area. Within the Tabulate Intersection tool, the Service
Areas were selected as the Input Zone Feature, the Input Class Features was the
Census_Tracts_2010 layer and the Sum Field was the total population to add the populations
from the census tracts that fall within the same service area.
This tool provided the percentage of the census tract within a service area. That
percentage was then multiplied by the total population of the census tract to get the areal extent
of the population that falls within the census tract. For service areas that covered multiple census
tracts, the Dissolve tool was used to sum the populations from the various census tracts to obtain
the total population in the service area. This methodology allowed for a good estimate for the
number of homeless individuals potentially within the range of the service areas based on census
tract area and potential dispersion. This approach clearly takes liberty in allocating the
34
population when precise locational data for the individuals is not known, however, this approach
emerged as the most even-handed method to determine potential population within the service
areas.
3.5 Overview of 2SFCA methodology
The following 2SFCA method leveraged the LAHSA Homeless Count 2019 by Census Tract
and Los Angeles Hospitals and Medical Centers (2019) datasets in conjunction with the Los
Angeles Census Tracts (2010) and Los Angeles CoC (2018) shapefiles. The first catchment, with
a 1-mile service area from a healthcare provider, sums the total homeless population that the
provider can reach within that distance to determine a provider to population ratio. In the case of
each medical facility, the provider ratio is total bed count of the hospital to the total homeless
population reached. The second catchment is produced from the census tract total population
based on areal extent to obtain the previously computed provider-to-population ratio of each
hospital that resides within the service areas. The provider-to-population ratios for hospitals’
service areas that were in the same census were summered together to determine accessibility at
the census tract level.
3.6 2SFCA data preparation
First, the data and shapefiles were loaded, cleaned, and modified in ArcGIS Pro (Figure 9).
The LAHSA Homeless Count 2019 Results by Census Tract excel sheet was added to ArcGIS as
a table and joined to the Los_Angeles_Census_Tracts (2010) shapefile by tract ID. The
LA_Census_Tracts shapefile was then spatially joined with the Los_Angeles_CoC Shapefile
(2018) to exclude census tracts not in the CoC. The results layer was named
35
Los_Angeles_CoC_Census_Tracts. Lastly, the Los Angeles Hospitals and Medical Centers
(2019) shapefile was added and hospital locations outside the LA CoC were excluded
(Los_Angeles_CoC_Hospital_and_Medical_Centers layer). All layers came projected or were
changed to the NAD 1983 StatePlane California V FIPS 0405 Feet projection.
3.7 First catchment
For the first catchment, 1-mile walking route service areas (broken up into 0.25, 0.5, and
1-mile) around the Los_Angeles_CoC_Hospital_and_Medical_Centers points were created using
the ArcGIS Network Analyst Service Area tool. The next steps were to calculate the provider-to-
population ratio. First, the population within each service area had to be determined. This was
done by using the Tabulate Intersection tool to find the population by areal extent of the census
tract covered by the service area. Once the percentages were determined, the total population of
the census tract was multiplied by the percentage covered. For service areas that covered
multiple census tracts, the populations from each census tract were added together. This process
determined the total homeless population within the service areas. The resulting output feature
Figure 9. Import and update shapefiles stage
36
class was the LA_First Catchment. To create the provider-to-population ratio, a new field was
added to the LA_First_Catchment layer. The provider-to-population ratio was calculated by
using the Total Bed Count divided by the homeless population count (Figure 10).
3.8 Second catchment
For the second catchment, the accessibility index for each census tract was computed by
summing up all provider-to-population ratios from the 1-mile service areas of the healthcare
facilities that were within the tract (Figure 11). For the provider-to-population field, the Merge
Rule was set to “Sum” to obtain the spatial accessibility index calculation. Populations within the
census tracts that did not fall within a service area were added into the population figure for the
tract. The provider-to-population ratios were joined to the LA_CoC_Census_Tract_Buffer
shapefile. The Buffer shapefile was then joined to the LA_CoC_Census_Tract layer according to
shared GEOID. The resulting output is the LA_2SFCA layer with the data and ratios calculated
Figure 10. First catchment
37
being incorporated into the census tract polygons. The symbology was then changed to
Graduated Colors for the Provider-to-Population field to represent the differences in accessibility
as a gradation of colors.
3.9 Distance to nearest hospital
The distance from the homeless populations of each census tract to the nearest hospital
was measured as another determinate of accessibility. The location of the homeless population
was represented as the census tract centroid. The distance between the centroid to the nearest
hospital was measured using the ArcGIS Near tool which measures the distance between input
features. In the Near tool, the LA_Census_Tracts_Centroid layer was used as the “Input
Features” and the Los_Angeles_CoC_Hospital_and_Medical_Centers was set as the “Near
Features.” From this tool, the NEAR_DIST (near distance) was calculated which represents the
Euclidean distance from the centroid point to the hospital point that is the shortest distance. The
NEAR_FID field identified the medical center ID number of the closest hospital.
Figure 11. Second catchment
38
Chapter 4 Results
This chapter presents the results of the homeless population distribution analysis and the 2SFCA
methodology for accessibility analysis. Hospital accessibility and distance measurements from
the nearest medical facilities to the census tracts’ populations are also included. The results are
presented in a collection of maps and tables.
4.1 Homeless population distribution
4.1.1 Homeless population counts choropleth map
The Los Angeles CoC on CONUS is comprised of 2,161 census tracts. These census
tracts were weighted based on total number (shelter and unsheltered) of homeless individuals
within the tract. The largest population, and the only community in the highest tier between 707-
3,180, was in Skid Row with 3,180 homeless individuals. Thirteen communities comprised the
next tier of 300 to 706. 61 communities were in the third tier with populations of 111 to 259.
With homeless populations between 34 and 110, there were 272 communities. Lastly, with
homeless populations of 33 people or less, there were 1,810 communities with 375 communities
having no recorded homeless individuals. These values were compared in a choropleth map with
graduated colors representing total homeless population. The color scale ranged from teal blue
(0-32), medium blue (33-109), dark blue (110-258), purple (259-705), and pink (706-3179). The
resulting choropleth map is displayed in Figure 12.
39
The same color scheme was used in the results were normalized according to total
population of the census tracts (Figure 13). The results of the normalized choropleth were the
percentages of the homeless population relative to the total population, ranging from 0-6.2%.
Figure 12. Choropleth map of total homeless population (# of individuals)
by LA CoC census tract
Figure 6 Choropleth map of total homeless population (# of individuals) by
40
4.1.2 Homeless population counts hot spot analysis
An optimized hot spot analysis map was created to illustrate which census tracts had the
largest homeless populations and where populations tended to cluster geographically. The
density of census tracts is represented using a color scale from blue to red indicating cold spots,
statistically irrelevant zones, and hot spots (Figure 14). The census tracts with the largest
homeless populations in the densest areas in the Los Angeles CoC are observed in the downtown
Los Angeles and Santa Monica areas. In general, the southern region of the LA CoC is denser
Figure 13. Total homeless population (# of individuals) normalized by
total population of census tract
41
compared to the sparser northern region of the CoC, with the exception of Hi Vista in the upper
northeast corner of the CoC.
4.2 2SFCA and Accessibility Index
Using the 2SFCA method, measurements of the accessibility of the hospitals and medical
facilities to homeless populations were made for the LA CoC. The accessibility index measured
accessibility as a ratio of the total number of beds per hospital to the total population of homeless
individuals within 0.25, 0.5, and 1-mile service areas. The population counts from the census
Figure 14. Hot spot analysis of homeless population by census tract
42
tracts within the service areas were summed together to represent the total potential homeless
population the hospitals can service. Access was determined as to whether the total homeless
population within the service area could potentially be cared for by a hospital based on their total
bed count.
To look at access within each service area, an accessibility index was established for the
services areas of the hospitals. The ratio of the accessibility index was the total number of beds
per hospital to the total homeless population within the service areas surrounding each hospital
The scale was broken down into five Natural Breaks ranges (from greatest demand to least) as
well as areas of no access (Table 5). The larger the homeless population within the service area
relative to the number of beds at the facility, the smaller the accessibility index value. This step
allows for accessibility to be observed at the hospital level.
Each hospital service area within the LA CoC was coded by color according to the
accessibility index (Figure 15 & 16). 3 service areas had an accessibility index of less than 1,
signifying the homeless population of the area was greater than the number of available beds. 95
service areas had an accessibility index greatest than 1 and less than or equal to 39. 42 service
areas had an accessibility index of 40-131. 30 service areas had an accessibility index between
Accessibility Index (Hospital)
0.1-39
40-131
132-321
322-560
560-1035
0 (No Access)
Table 5. Accessibility index for hospital service
areas
43
132-321. 17 service areas had an accessibility index between 322-560. 8 service areas had an
accessibility index between 560-1035, indicating least demand. 24 service areas were determined
to have no population therefore no accessibility index value was assigned to them (Figure 16 &
17).
Figure 15. Hospital service areas by accessibility index
44
For the 2SFCA, the accessibility index measured the ratio of number of beds from each
hospital per census tract to the total homeless individuals. Census tracts with no homeless
individuals within a 1-mile service area of a hospital had a value of 0, or no access. Populations
of one individual or more within a 1-mile walking distance to a hospital had a value on the
accessibility index. The smaller the accessibility index value, the larger the total homeless
population relative to the number of beds at all the hospitals within each census tract. The scale
was broken down into five Natural Breaks ranges (from greatest demand to least) as well as areas
of no access (Table 6).
Figure 16. Downtown hospital service areas by accessibility index
45
Accessibility index values were assigned to census tracts with homeless populations that
were in a 1-mile walking distance of hospitals within the census tract. The accessibility
values assigned to each census tract and the distribution of the indexes according to census
tract are observed in Figures 17 & 18.
2SFCA Accessibility Index
0.1-71
72-185
186-361
362-668
669-1765
0 (No Access)
Table 6. 2SFCA Accessibility Index
Figure 17. Census tracts by 2SFCA accessibility index
46
4.3 Hospital accessibility
From the 2SFCA methodology used, the potential accessibility of each hospital was
determined by the total number of homeless individuals the facility could service within its 0.25,
0.5, and 1-mile service area. In total, 147 hospitals and medical facilities in the LA CoC were
included. 100 of these facilities had homeless individuals within a 1-mile service area of the
location. The accessibility index with the same Natural Breaks classification was used to
measure the accessibility of the hospitals.
The U.S. Department of Veteran Affairs (VA) Los Angeles Ambulatory Care Center was
the hospital with the greatest homeless population within a 1-mile diameter, with 5,264 people.
Figure 18. Downtown LA census tracts by 2SFCA accessibility
index
47
The top ten hospitals and medical facilities with the largest potential population reach were:
Good Samaritan Hospital (2299), Saint Vincent Medical Center (2052), Los Angeles Orthopedic
Hospital (1617), Queenscare Family Clinics in Echo Park (1569), California Hospital Medical
Center (1567), and Shriners Hospitals for Children (1562), Southern California Hospital at
Hollywood (1,389), Filipino-American Service Group, Inc. Community Wellness Center (1059),
Silver Lake Medical Center (1047) (Table 7).
Table 7. Top ten hospitals by the total number of homeless individuals within a 1-mile range
48
4.4 Distance to nearest hospital
The distances between the census tract centroids to the nearest hospitals were measured.
The distances were divided into mileage ranges from 0 to 25 miles (Table 8). The total homeless
population layer was used as a base layer to compare the distance with the total homeless
population (Table 9). The centroids were color-coded according to distance from the nearest
hospital (Figure 19).
913 census tracts were within a 1-mile distance from the nearest hospital. 506
communities were within 2 miles away from the nearest hospital. Within a range of 2-5 miles
from the nearest hospital, there were 321 census tracts. 41 communities were 5-10 miles away
from the nearest hospital. 16 tracts were the furthest away with a distance of 10-25 miles from
the population location to the nearest hospital.
Distance to Nearest Hospital (mi)
0-1
1.01-2
2.01-5
5.01-10
10.01-25
Total Homeless Population
0-33
34-110
111-259
300-706
707-3180
Table 8. Distance measurements Table 9. Total homeless population
49
Figure 19. Distance from census centroid to nearest hospital
50
Chapter 5 Discussion and Conclusion
This study’s primary objectives were to analyze the distribution of the homeless population in
the Los Angeles CoC and conduct an accessibility analysis of the hospitals within it. These
objectives were met using a methodology modeled on previous research pertaining to homeless
initiatives. This project focused on creating a population distribution and hospital accessibility
analysis.
To better understand the distribution of the homeless population, the most recent PIT
count data collected by the City of Los Angeles was examined at the census tract level and
visualized using various symbologies in ArcGIS. Observing the distribution of homeless
individuals by census tracts helped to identify areas within the CoC that may be in more need of
CoC resources based on number of individuals and current access. Analyzing the population
distribution provides insights into homeless living patterns that could inform future studies.
The 2SFCA method was chosen to study accessibility by census tract according to the
number of beds from hospitals within the census tract to homeless within a 1-mile walking
distance. The 2SFCA method, originally developed to study healthcare accessibility, was
modified for the purposes of this study to look at a population, homeless individuals, that
presented unique accessibility challenges necessitating the tailoring of the method to the project.
Access was determined to be a ratio of the number of beds at a hospital per the number of
homeless individuals within the various services areas. The maximum walking distance of 1 mile
was chosen considering the limited means of transportation homeless individuals tend to have
and prior research into what distance is deemed “accessible.” By using the 2SFCA method, this
study created an accessibility index categorizing the census tracts and service areas surrounding
medical facilities by potential access accounting for the bed count of each hospital.
51
This study found that for individual hospitals, potential access was lesser in the
downtown area as hospitals had a greater homeless population to potentially service relative to
their bed availability. The demand was greatest in the downtown area, as determined by the
larger homeless population, creating a potential issue of the hospitals’ ability to accommodate
these populations in addition to non-homeless individuals.
The 2SFCA accessibility index determined that access by census tract varied in the
downtown region as a result of the varying sizes and number of hospitals in each census tract.
Census tracts in northeast LA and south LA tended to lack accessibility entirely with few to no
hospitals in walking distance. It is recommended that additional funding be allocated to select
hospitals in the downtown area with potentially high burden to service homeless individuals and
that transport routes are established in various geographically areas in the CoC to enable better
medical care access.
Distance measurements between census tracts and their nearest hospital were calculated
to provide a general operating picture of what hospital distribution in Los Angeles looks like. A
concentration of hospitals was observed in the metropolitan area while available medical
resources were sparser in the more rural areas of Los Angeles. Providing LA CoC administrators
with this information identifies census tracts that generally lack easy access to medical facilities.
Providing this data could potentially prompt the establishment of transportation services for
homeless people or the empirical data to support any requests to the city, local government, or
private companies for the addition of more medical facilities in the area.
The following chapter discusses the results of the study, the potential implications of the
findings, and future research recommendations for the LA CoC administrators and stakeholders.
Areas identified with large homeless populations and low access to medical facilities are
52
recommended to be highlighted in the LA CoC’s future CoC Plan for funding justification. The
core intent of this project was to provide these parties with spatial analysis of homelessness
within their boundaries using the most up-to-date PIT data available. The results will ideally be
used to inform distribution of funding and services to areas identified as having greater need in
terms of population size and/or accessibility.
5.1 Analysis Discussion
The homeless population distribution by census tract and 2SFCA accessibility analysis
provided insights into the state of homelessness and access to healthcare within the LA CoC as
of the 2019 PIT study.
5.1.1 Homeless population distribution
Understanding where homeless populations reside at the census tract level enables
analysis of where service needs are and where accessibility to medical facilities is an issue. The
homeless population choropleth maps showing the census tracts according to total homeless
population show a concentration of the population in the downtown area (Figure 20). This result
was not surprising as larger citizen populations tend to be in the metropolitan area of large cities.
The area within the LA CoC that emerged with the largest homeless population by a substantial
amount (3180 followed by 705) was Skid Row.
53
Skid Row is an area a vast majority of Los Angelinos and those familiar with the
homeless crisis in America associate with homelessness due to the sheer size of the homeless
population and the dire humanitarian situation there. Within the 0.4 square mile zone of Skid
Row, approximately 3% of the county’s entire homeless population resides (Figure 21)
(CRA/LA 2005). Historically, Skid Row has been an area where homeless shelters established
locations to service the homeless population, primarily starting efforts in 1981 due to the
recession and rise of unemployment. Throughout the following decades, the number of homeless
Figure 20. Downtown Los Angeles by census tract with homeless
population counts
Figure 18. Downtown Los Angeles by census tract with homeless
54
individuals steadily increased and the need for shelter and services outgrew the capacity of local
shelters. People resorted to sleeping on the streets and in public areas surrounding the shelters
and in the greater Skid Row area, thus contributing to the establishment of Skid Row as a hub of
homelessness in LA (Flaming and Blasi 2019).
In totality, the area of Skid Row has been the focus of many LA homeless initiatives and
efforts to provide shelter and aid. The results of this study further supports the need in the area
and identifies Skid Row as an area of high priority for service providers. Due to the large amount
of individuals in the area and the burden placed on local homeless service providers, it is
recommended that the LA CoC continue to provide funds to Skid Row services and possibly add
Figure 21. Downtown Los Angeles with Skid Row boundaries
Source: Community Redevelopment Agency of the City of Los Angeles
55
transportation services that could take people to outside service providers to reduce the burden
within Skid Row.
A cluster of high density, large population census tracts were observed in the downtown
area of the CoC as well as Santa Monica (Figure 22).
Santa Monica has been an area of Los Angeles that has fluctuated with the size of its
homeless population throughout the years. In 2019, the city witnesses a 3 percent increase in its
homeless population to a total of 985 people. The City Council has implement a funding strategy
to address its homeless issue and has seen the benefits of efforts to provide low-income housing
Figure 22. Hot spot analysis of downtown LA and Santa Monica
56
and provide medical services for mental and physical health issues (Pauker et al. 2019).
Additional LA CoC resources dedicated to Santa Monica could prove valuable to their efforts
and have high impact results as is an area of concern due to its larger homeless population. Santa
Monica’s homeless infrastructure is not as developed as Skid Row’s, so establishing additional
medical service provider sites in the city would be valuable to enable access to medical
professionals.
Overall, the distribution of homelessness was consistent with what would be expected
with larger populations in the more metropolitan downtown part of the city and smaller to no
individuals in more rural areas. Skid Row, Santa Monica, Brentwood, and Hollywood should be
prioritized in terms of providing services to the greatest number of individuals with Skid Row
and Santa Monica being areas of the highest density of homeless individuals to more efficiently
consolidate efforts. This concentration of the population in the downtown area offers the LA
CoC administrative bodies the opportunity to reach a significant portion of the homeless
population in a fairly compact area. Infrastructure and means to access resources, such as transit
and established service locations, tend to be in the central downtown area, potentially enabling
more immediate and efficient resource distribution and allocation.
The other region with a large homeless population which was more surprising was the
Palmdale/Lancaster area in the northeast corner of the LA CoC boundary. Specifically, Hi Vista
in the upper left corner of the CoC boundaries with a homeless population of 309 individuals.
This area may be a lesser known area that could benefit from more attention and funding from
the CoC. These cities do comprise larger areas and have high population counts at the census
tract level which may account for the larger homeless population. Further inquiry into the
57
homeless population in this area and the regional causes and contributions to the issue of
homelessness in the area would be beneficial on the part of local officials.
5.1.2 Accessibility analysis by Service Area
The resulting maps of accessibility by service area provided visualizations of how many
people were within three varying distances from hospitals throughout the LA CoC and where
there wasn’t potential access. As observed previously, resources such as medical facilities tend to
aggregate in more metropolitan areas, as seen with the concentration of service areas in the
downtown LA region. The service areas within the downtown area also tended to have lower
accessibility indexes indicating a high demand from the population able to be serviced relative to
the number of beds at the hospitals. This aligns with the observation that the homeless population
of the LA CoC tended to be greater in the downtown area, resulting in more potential burden on
hospitals in that area. This result also matched existing literature such as the 2018 AHAR study
that found that the median percentage of the population living in urban areas among major city
CoCs (like Los Angeles) was 70 percent (Henry et al., 2018). A greater number of walkable
paths were found in the more developed, metropolitan communities, enabling greater access to
hospitals and providing individuals more options of service within a shorter distance.
Hospital bed counts depended on the type of hospital. All hospitals were scored on the
accessibility index, however, this did not account for the fact that some hospitals were specialty
clinics or provided services that not all people in the area would necessarily need or qualify for.
No observable pattern was identified regarding if accessibility varied according to the various
distances from the medical facilities. The results of this accessibility analysis are more useful to
CoC administrators to evaluate individual hospitals in terms of where the burden for homeless
58
care potentially falls and which hospitals may qualify for CoC support funding or in what areas
transportation options should be established to take people to less busy hospitals.
5.1.3 2SFCA Accessibility analysis
The results of the 2SFCA
accessibility analysis showed potential
access to hospitals within each census
tract. Census tracts scored, thus having
homeless population within walking
distance of a hospital, tended to be in
the downtown and southern region of
the LA CoC. Figure 23 provides
regional names in the LA CoC
referenced.
Looking at the accessibility
values of the census tracts relative to
each other, the census tracts of
Antelope Valley had no access to
hospitals. The communities of Antelope Valley also had the longest distances to the nearest
hospital, with most being in the 20 mile or greater distance range from a hospital. Additional
medical facilities and transportation options to connect individuals to medical resources would
improve accessibility, but for fewer people as the homeless population in the area is relatively
small.
Figure 23. Regions of LA
Source: The Los Angeles Times’ Mapping LA Project
59
The communities in the San Fernando Valley were generally shorter on beds relative to
the other regions of LA. Compared to Antelope Valley, there were a greater number of hospitals
in the region, however, there was also a larger homeless population in this region as well.
Increasing medical resources available in the San Fernando Valley could alleviate some of the
burden the medical resources in the area face and would increasingly encounter should use by
homeless individuals increased. The addition of ambulatory services catering to homeless
individuals could provide an opportunity to provide resources tailored to common afflictions
those who live on the streets face. CoC funding to an additional site could be proposed as there is
demonstrated need and demand in the region.
The Westside and San Gabriel Valley had medical resources in place and generally
varied in accessibility as homeless population greatly varied by census tract. Additional funding
from the CoC or any other government or private entity could provide the support hospitals
would welcome to provide care to all their patients.
In terms of access, the census tracts in the downtown region varied in accessibility. The
census tracts composing Boyle Heights had high accessibility values with attributed to the
multiple high capacity hospitals within the area providing potential service even though the
population here was large. Other areas such as Lincoln heights had large homeless populations
but less hospitals in the area and service area overlap in the census tract to provide access (Figure
24). The results of this accessibility index map indicate that while homeless populations tend to
be higher in the downtown region, potential access to medical care is not evenly disbursed and
not universally greater in the downtown area.
60
5.1.4 Accessibility by Hospital
When considering access to hospitals, type of hospital is an important consideration as
not all hospitals and medical centers offer the same services and accessibility to certain services
may differ from access to simply the nearest facility. The 147 hospitals and medical centers were
ranked by population within service range, however, the actual accessibility to target patients or
by specialty of care effects accessibility. For instance, the hospital with the largest population
within a 1-mile walking distance was a VA clinic with 5264 people within the service areas
surrounding the facility. Many of the people included within this accessibility measurement may
Figure 24. 2SFCA accessibility in Downtown Los Angeles
61
not be able to utilize the resources of this clinic or the facility may not offer the type of care
needed.
For type of services provided, only 20 of the hospitals had emergency services which is a
service that homeless individuals tend to rely on due to cost limitations of receiving preventative
care and other financial access considerations (Baggett and O’Connell 2016). Support to
emergency service providers could enable greater support on behalf of the medical facilities to
accommodate a greater number of patients. However, this study offers limited insights into
access based on hospital type but recommends future studies with access to more detailed
hospital data consider this aspect of healthcare accessibility.
The hospitals with the greatest homeless populations in range, less accessibility according
to amount of potential burden, and general service options were the Good Samaritan Hospital,
Saint Vincent, and California Hospital Medical Center. It is recommended these hospitals be
considered for funding allocation when applying for funds and crafting the next LA CoC Plan to
encourage partnership between the facilities and the CoC and enable the ability of these hospitals
to provide services.
5.1.4 Distance measurements to hospitals
The intent on calculating distance measurements from the census tracts to their nearest
hospital was to give LA CoC administrators a sense of which locations might have a general
issue with hospital accessibility for their entire population, not necessarily just the homeless
population. Systemic accessibility issues could influence the issues facing homeless individuals,
so providing administrators with a more complete operating picture could enable more informed
62
decision-making and provide opportunities to partner with local government and community
efforts to improve conditions for all citizens.
Falling in line with expectations based on prior research and urban design, census tracts
in the downtown region tended to have shorter distances to the nearest hospital. This result was
attributed to the concentration of hospitals within the downtown area and the smaller area of the
census tracts in the area. Larger census tracts and more rural locations experienced greater
distances to reach the nearest hospital. The areas in which the nearest hospital was in the double
digit range for mileage are identified as high priority communities to service, regardless of
homeless population within the area as this indicates that all citizens of that region are without
potential walking access, including homeless individuals.
West Antelope Valley, Hi Vista, and Juniper Hills were notable as the distances for these
communities were in the 20 mile plus range. For these locations, the establishment of a clinic
would be preferable. Hi Vista (most northeast census tract in the LA CoC) is identified as a
community of high priority with a large homeless population of 309 individuals and a distance of
22 miles to the nearest hospital (Figure 16). Additional resource allocation and funding would be
well-spent to increase accessibility through the establishment of local healthcare services and/or
a transportation service to a hospital. Within the city, distances from hospitals were typically
within a zero to two-mile distance. This indicates that the need for transportation may be less
than a more rural location. However, considering the large populations and proximity of the
hospitals, a transportation system may be more easily established and have more effect. The
establishment of a shuttle circuit that pickups in various communities and services different types
of hospitals has the potential to improve access for more people.
63
5.2 Limitations
This project achieved its objective of providing an analysis of homeless accessibility
using the data gathered in the LAHSA 2019 PIT study. There are, however, limitations in the use
and accuracy of the study created by the data used, methodology implemented, and resources
available.
A primary limitation of the results of this study is the generalizations made by the need to
aggregate homeless populations to the census tract level. The methodology and the results of the
PIT count data used limits the spatial accuracy of the individuals’ locations to the census tract so
no further refinement in location was possible. To account for population distribution throughout
a census tract, population by areal extent was used to allocate populations. However, this method
does not divide people based actual location. For the purposes of this study, estimation at the
census tract level was sufficiently accurate to give CoC administrators and stakeholders an
operational picture of homeless population distribution and hospital accessibility at the
community level. However, should more detailed and specific resource allocation requirements
arise necessitating greater accuracy for population locations, this study is not suited for such
analysis.
The 2SFCA methodology used also has some drawbacks. The service area modelling for
the 2SFCA methodology was limited to walkable routes, which was an overgeneralization of the
transportation options of homeless individuals. It was assumed that homeless individuals would
only be able to walk to hospitals and the distances they would walk would be 1 mile or less. This
is clearly a generality made that potentially excludes individuals who have motor transportation
(public or private) available or who are able and do walk more than this limited distance to reach
their destinations. Conversely, some individuals may not be able to walk such distances and
64
could only access needed resources via transportation which is not considered in the current
study.
For measurements of distance to the nearest hospital, method used to determine the
distance was not exact due to the use of Euclidean distance. The distance between the population
centroid and the hospital point was measured as a straight line. This is problematic as it does not
account for geographic barriers or actual roads or walkways that would be used to more
accurately represent distance. Using the census tract centroid as the “location” of the homeless
population of a census tract is also problematic as the entire homeless population in a community
does not reside in the center of the community.
5.3 Future Research
This project demonstrates the insights that can be garnered from PIT data when spatial
analysis is leveraged. Applying the 2SFCA methodology to the study of homeless healthcare
accessibility is also relatively new in the variations of this method. Further research and
improvements in the data used, methodology implemented, and analytical tools leveraged would
only improve the accuracy of the results and provide higher fidelity spatial analysis for those
looking to use the results for funding justification or other administrative purposes.
PIT data at a finer spatial skill would greatly improve the positional accuracy of the
individuals and would result in more accurate population distribution analysis. In addition to the
study of homeless healthcare accessibility, higher fidelity PIT data would benefit the study of
homelessness by enabling a greater sense of where people are and what demographic
characteristics they have to better tailor services to these populations.
65
Creating various types of services areas surrounding the hospitals and medical centers
would be valuable to see how accessibility differs according to transportation type. Further data
collection and research would need to be conducted to gain a better understanding of homeless
transportation availability and options. Looking at how LA CoC services, like shelter and
transportation, are distributed would also be interesting to see how government efforts effect
accessibility and if there are gaps in current services that create accessibility.
Conducting accessibility analysis based on type of medical services offered would
provide more tailored medical care accessibility analysis and may provide more accurate
measurements of accuracy. Gaining a better understanding of accessibility based on medical care
type would also inform where gaps in certain types of care exist and could provide homeless
services the opportunities to establish locations that offer this type of care or motivate hospitals
in the area to add this type of care.
For distance measurements between individuals and hospitals, more precise locations for
the individuals would greatly benefit this analysis as opposed to mass generalizations such as
that the entire population of a community being located at the center of a census tract. This
generalization was made due to lack of more precise location data for the homeless individuals,
however, with more accurate data, improved distance measurements could be made. Measuring
distance in non-Euclidean terms would also improve the accuracy of the distance measurements
and provide a more realistic determination of separation between individuals and hospitals. More
tailored distance analysis to a smaller region and/or fewer hospitals would allow for more
accurate distance measurements.
66
5.4 Final Conclusions
This project achieved several of the stated goals and objectives. Through geospatial
analysis, conclusions were drawn about the distribution of homelessness throughout the LA CoC
and the accessibility of hospitals in the area. It was determined that communities in downtown
LA had a higher density, large-population census tracts and the largest homeless populations.
Specific communities with large homeless populations were identified as areas LA CoC
administrators might consider allocating more resources to. Access to hospitals and medical
centers was also considered, resulting in the conclusion that there was a concentration of
hospitals in the downtown area which had greater accessibility for larger populations as opposed
to more rural locations. These hospitals were also closer in terms of distance to communities.
Accessibility according to census tract showed more variation, reaffirming that tailored
responses to homelessness need to be created and adopted in communities to best meet the needs
of their people.
A primary intent was to demonstrate the utility of geospatial analysis through the creation
of several maps for the visualization of homeless data and analysis of large datasets otherwise
not easily translated to actionable information. Visualizing data in terms of location helped to
transform the data into a form useful and more readily understandable to stakeholders such as
CoC administrators and city planners. The development of a relatively simple 2SFCA
methodology was intended to provide an approach to spatial analysis that city GIS specialists and
individuals with access to GIS platforms could replicate with different data. The hope is that this
project will encourage local governments, non-profits, and any organization collecting data on a
topic of concern to use spatial analysis to provide an added layer of clarity into a situation.
67
The project was designed to be of value to LA CoC administrators in their efforts to
allocate resources and request funding. All data and methods used were ensured to be compliant
with HUD’s data standards so that the results could be used as justification for budget and
resource allocation. Using the most recent PIT count data collected by the CoC, this project
provided timely analysis that remains relevant to the situation LA faces with homelessness today.
This project contributed valuable analysis of the homelessness crisis that Los Angeles
faces. Positioning the LA CoC to best advocate for increased funding and more targeted resource
allocation would help to relieve some of the burdens associated with homelessness and get
individuals the medical care they need. Spatial analysis as applied to homelessness is relatively
limited, so studies like this should continue to be conducted to build a greater body of literature
and improve our understanding of the issue.
68
References
Baggett, Travis P., and J. J. O'Connell. 2016."Healthcare of homeless persons in the United
States." UptoDate. Waltham, MA: UpToDate.
California Department of Finance. 2018. “City of Los Angeles Statistics.” County of Los
Angeles. https://www.lacounty.gov/government/geography-statistics/statistics/#1481130319359-
d19956e6-ad46.
Center for Evidence based Solutions to Homelessness. 2017. “Continuum of Care.” Center for
Evidence based Solutions to Homelessness,
http://www.evidenceonhomelessness.com/factsheet/continuum-of-care/.
Chen, Xiang, and Jia, Pengfei. 2019. “A Comparative Analysis of Accessibility Measures by the
Two-Step Floating Catchment Area (2SFCA) Method.” International Journal of Geographical
Information Science 33, no. 9: 1739–1758.
Community Redevelopment Agency of the City of Los Angeles (CRA/LA). 2005. “Los Angeles’
Skid Row”
Cowan, Jill. 2019. “Homeless Populations Are Surging in Los Angeles. Here's Why.” The New
York Times. The New York Times.
Flaming, Daniel, and Gary Blasi. 2019. “Los Angeles: Why Tens of Thousands of People Sleep
Rough.” BBC News. BBC.
Gelberg, L, Gallagher, T C, Andersen, R M, Koegel, P. 1997. Competing priorities as a barrier to
medical care among homeless adults in Los Angeles. American journal of public health.
1997;87(2):217-220. doi:10.2105/AJPH.87.2.217
Grumdahl, Eric. 2019. “More Than a Number: Point-in-Time Counts Are Crucial Data.”
Substance Abuse and Mental Health Services Administration (SAMHSA). U.S. Department of
Health and Human Services. https://www.samhsa.gov/homelessness-programs-resources/hpr-
resources/point-time-counts-are-crucial-data.
Hawthorne, Timothy L., and Kwan, Mei-Po. 2012. “Using GIS and Perceived Distance to
Understand the Unequal Geographies of Healthcare in Lower-income Urban
Neighborhoods.” Geographical Journal 178, no. 1: 18–30.
Heinrich, Anna. 2017 “What Is Ambulatory Care? Learning More About the Future of
Healthcare.” What is Ambulatory Care? Learning More About the Future of Healthcare |
Rasmussen College. Rasmussen College.
Henry, M., Mahathey, A., Morrill, T., Robinson, A., Shivji, A., and Watt, R. 2018. "The 2018
Annual Homeless Assessment Report (AHAR) to Congress." The U.S. Department of Housing
and Urban Development Office of Community Planning and Development. 1-80.
69
Henwood, F. B., Lahey, A. J., Petering, R., Byrnes, K., Gascue, L., Kwack, S., and St. Clair,P.
2018. "2018 Los Angeles Continuum of Care Homeless Count Methodology Report." The
University of Southern California. 1-17.
Homelessness Research Institute (HRI). 2014. “A Research Agenda for Ending Homelessness.”
Homelessness Research Institute.
Department of Housing and Urban Development (HUD) Exchange. 2018. “Continuum of Care
GIS Tools.” Department of Housing and Urban Development.
Department of Housing and Urban Development (HUD). Fiscal Year 2018. Continuum of Care
Competition Homeless Assistance Award Report. CA-600.
https://www.hudexchange.info/onecpd/assets/File/2018-california-coc-grants.pdf
Humphreys, J. S., & Solarsh, G. 2008. At-risk populations: rural. In H. K. Heggenhougen, & S.
Quah (Eds.), The international encyclopedia of public health. London: Academic Press.
Institute of Medicine (US) Committee on Health Care for Homeless People. 1988.
Homelessness, Health, and Human Needs. Washington (DC): National Academies Press (US); 4,
Access to Health Care Services for Homeless People.
Khandor, Erika, Kate Mason, Catharine Chambers, Kate Rossiter, Laura Cowan, and Stephen W.
Hwang. 2011. "Access to primary health care among homeless adults in Toronto, Canada: results
from the Street Health survey." Open Medicine 5, no. 2: e94.
Kushel, Margot B., Eric Vittinghoff, and Jennifer S. Haas. 2001."Factors associated with the
health care utilization of homeless persons." Jama 285, no. 2: 200-206.
LAHSA. 2018. Los Angeles Homeless Services Authority. “LA Awarded More Than $109
Million For HUD Continuum Of Care Programs.” https://www.lahsa.org/news?article=336-la-
awarded-more-than-109-million-for-hud-continuum-of-care-programs.
LAHSA. 2019. “Los Angeles Continuum of Care About.” LAHSA. https://www.lahsa.org/coc/.
Los Angeles Housing Services Authority (LAHSA). 2019. “Los Angeles Continuum of Care.”
https://www.lahsa.org/coc/.
Luo, W. and Wang, F. 2003. Measures of spatial accessibility to health care in a GIS
environment: synthesis and a case study in the Chicago region. Environment and Planning B:
Planning and Design, 30 (6), 865–884.
Luo, W. and Qi, Y., 2009. An enhanced two-step floating catchment area (E2SFCA) method for
measuring spatial accessibility to primary care physicians. Health & Place, 15 (4), 1100–1107.
doi:10.1016/j.healthplace.2009.06.002
70
McGrail, Matthew R. and John S. Humphreys. 2009. "Measuring spatial accessibility to primary
care in rural areas: improving the effectiveness of the two-step floating catchment area
method." Applied Geography 29, no. 4: 533-541.
National Alliance to End Homelessness. 2010. “What Is a Continuum of Care?” National
Alliance to End Homelessness. https://endhomelessness.org/resource/what-is-a-continuum-of-
care/.
National Coalition for Homeless Veterans (NCHV). 2019. “HUD’s Continuum of Care: How
Funding is Allocated.” PDF file.
http://www.nchv.org/images/uploads/HUD_CoC_brochure_9.14_.04_.pdf
Neutens, T., et al. 2010. Equity of urban service delivery: a comparison of different accessibility
measures. Environment & Planning A, 42 (7), 1613–1635.
Oreskes, Ben. 2019. “California Has the Most Homeless People of Any State. But L.A. Is Still a
National Model.” Los Angeles Times. Los Angeles Times.
https://www.latimes.com/california/story/2019-08-01/california-homeless-people-housing-
national-model-conference.
Oreskes, Benjamin, and Doug Smith. 2019. “Homelessness Jumps 12% in L.A. County and 16%
in the City; Officials 'Stunned'.” Los Angeles Times. Los Angeles Times.
https://www.latimes.com/local/lanow/la-me-ln-homeless-count-encampment-affordable-housing-
2019-results-20190604-story.html.
Pauker, Madeleine, Mike Fulper, J. Jennings, Matthew Hall, Brennon Dixson, and David Pisarra.
2019. “Santa Monica Homeless Population Grows by 3 Percent.” Santa Monica Daily Press.
Santa Monica Daily Press.
Rivkin, Steve, and Bauman, Sandra. 2011. “Hospital” vs. “Medical Center” Survery. Rivkin &
Associates and Bauman Research & Consulting. https://baumanresearch.com/consumers-prefer-
hospitals-over-medical-centers/.
Sadowski, Laura S., Romina A. Kee, Tyler J. VanderWeele, and David Buchanan. 2009. "Effect
of a housing and case management program on emergency department visits and hospitalizations
among chronically ill homeless adults: a randomized trial." Jama 301, no. 17: 1771-1778.
Tudor Morar, Radu Radoslav, Luiza Cecilia Spiridon, and Lidia Păcurar. 2014. “Assessing
Pedestrian Accessibility to Green Space Using Gis.” Transylvanian Review of Administrative
Sciences 10, no. 42: 116–139.
Tinoco, Matt. 2019. “What Is The Federal Government Doing To End Homelessness In LA?”
LAist. https://laist.com/2019/08/19/federal_government_assist_la_homeless.php.
U.S. Census Bureau. 2017. American Community Survey 1-Year Estimates
71
U.S. Census Bureau. 2018. “U.S. Census Bureau QuickFacts: Los Angeles County, California;
California.” Census Bureau QuickFacts.
https://www.census.gov/quickfacts/fact/table/losangelescountycalifornia,CA/PST045218.
US Interagency Council on Homelessness (USICH). 2018. “California Homelessness Statistics.”
Homeless in California Statistics 2018. Homeless Estimation by State | US Interagency Council
on Homelessness.
Vale, David S., Miguel Saraiva, and Mauro Pereira. 2016. "Active accessibility: A review of
operational measures of walking and cycling accessibility." Journal of transport and land use 9,
no. 1: 209-235.
Wang, Fahui. 2012. “Measurement, Optimization, and Impact of Health Care Accessibility: A
Methodological Review.” Annals of the Association of American Geographers 102, no. 5: 1104–
1112.
Yang, Duck-Hye, Robert Goerge, and Ross Mullner. 2006. "Comparing GIS-based methods of
measuring spatial accessibility to health services." Journal of medical systems30, no. 1: 23-32.
Yang, Yong and Ana V. Diez-Roux. 2012. "Walking distance by trip purpose and population
subgroups." American journal of preventive medicine 43, no. 1: 11-19.
Abstract (if available)
Abstract
Los Angeles has a homelessness crisis. The city has long struggled to meet the needs of the growing homeless population, and the problem continues to amplify as the most recent 2019 Point-In-Time (PIT) Count shows an increase in homelessness. The Department of Housing and Urban Development (HUD) Continuum of Care (CoC) federal grant program establishes regional or local planning bodies to coordinate housing and services funding for homeless people in an effort to promote an integrated system of care. As a local planning body, CoCs address the issues their local communities and have the potential to affect positive change. Access to healthcare is one such issue facing homeless populations that the LA CoC could better address using spatial analysis, namely where homeless populations reside in the CoC boundaries relative to established hospitals and medical facilities. ❧ This project used a geographic information system (GIS) to assess the state of homelessness in the Los Angeles CoC as of June 2019. A population distribution and density analysis was conducted, indicating that homeless populations tended to be larger and more concentrated in the census tracts comprising downtown Los Angeles and Santa Monica. To determine the degree to which homeless individuals can access hospitals and medical facilities, an accessibility analysis was conducted using a modified two-step floating catchment area (2SFCA) methodology. The 2SFCA accessibility index indicated that census tracts within the downtown area had homeless populations within a 1-mile distance of at least one hospital as opposed to more rural tracts that tended to lack any access. However, access to medical facilities within a walkable distance varied in the downtown census tracts. Recommendations for funding allocation, the establishment of transportation initiatives, and additional medical facilities to improve access were made.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Providing a new low-cost primary care facility for under-served communities: a site suitability analysis for Service Planning Area 6 in Los Angeles County, California
PDF
A spatial analysis of veteran healthcare accessibility
PDF
Spatial analysis of veteran access to healthcare in Los Angeles County
PDF
Reducing maternal mortality by improving medical facility accessibility: a methodology demonstrated for the Democratic Republic of the Congo
PDF
Spatial analysis of vision services of Kaiser Permanente members
PDF
Analysis of park accessibility in Redan, Georgia Web GIS application
PDF
The use of site suitability analysis to model changes in beach geomorphology due to coastal structures
PDF
An analysis of racial disparity in the distribution of alcohol licenses and retailers in Orange County, California
PDF
Suitability analysis for wave energy farms off the coast of Southern California: an integrated site selection methodology
PDF
Assessing homeless accessibility to community resources in the city of Los Angeles
PDF
Enriching the Demographic Survey sampling for the Los Angeles County Annual Homeless Count with spatial statistics
PDF
Web-based relational spatiotemporal geodatabase of glaciers in California
PDF
Access to active play parks for youth segments in Alexandria, Virginia
PDF
Network accessibility and population change: historical analysis of transportation in Tennessee, 1830-2010
PDF
Exploring the impact of the natural gas leak on home values in Porter Ranch, California
PDF
Use of least-cost path analysis to identify potential movement corridors for jaguars across the US-Mexico border
PDF
Site suitability analysis for implementing tidal energy technology in southern California
PDF
Urban areas and avian diversity: using citizen collected data to explore green spaces
PDF
Suitable California Opportunity Zone (COZ) locations for affordable housing development in the cities of Bakersfield, Los Angeles, and Palmdale
PDF
Using aerial imagery to assess tropical forest cover surrounding restoration sites in Costa Rica
Asset Metadata
Creator
Barr, Erin
(author)
Core Title
An accessibility analysis of the homeless populations' potential access to healthcare facilities in the Los Angeles Continuum of Care
School
College of Letters, Arts and Sciences
Degree
Master of Science
Degree Program
Geographic Information Science and Technology
Publication Date
02/14/2020
Defense Date
01/09/2020
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
2SFCA,GIS,Homelessness,Los Angeles,OAI-PMH Harvest
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Bernstein, Jennifer (
committee chair
), Vos, Robert (
committee member
), Wu, An-Min (
committee member
)
Creator Email
erinbarr.alumni@usc.edu,erinbarr@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-270870
Unique identifier
UC11675390
Identifier
etd-BarrErin-8171.pdf (filename),usctheses-c89-270870 (legacy record id)
Legacy Identifier
etd-BarrErin-8171.pdf
Dmrecord
270870
Document Type
Thesis
Rights
Barr, Erin
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
2SFCA
GIS