University of Southern California Dissertations and Theses

Incomplete data & insufficient methods: transgender population health research in the US

(USC Thesis Other)

Incomplete data & insufficient methods: transgender population health research in the US

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

INCOMPLETE DATA & INSUFFICIENT METHODS:

TRANSGENDER POPULATION HEALTH RESEARCH IN THE US

by

Avery Rose Everhart

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(POPULATION, HEALTH AND PLACE)

August 2022

Copyright © 2022 Avery Rose Everhart
ii
Dedication

This dissertation is dedicated to every trans and gender-diverse community around the world in
hopes that this work will uplift us and enable us to live healthier, fuller lives.

iii
Acknowledgements
I want to thank my co-chairs first and foremost, Drs. John Wilson and Laura Ferguson,
for their steadfast support of me in completing this dissertation. I also want to thank Dr. Juan De
Lara with whom I’ve had the pleasure of working since I joined USC and before joining the PHP
program. His belief in me empowered and enabled me to make a lot of my most ambitious and
inspired moves as a scholar. And of course, I want to express my gratitude to my remaining
committee members, Dr. Jo Olson-Kennedy and Dr. Arjee Restar, for seeing the potential in my
work and committing to seeing me through this process. I also want to thank my colleagues in
the PHP program (in no particular order) Bita Minaravesh, Leo Lerner, Emily Serman, Shea
Gilliam, Lois Park, Kate Vavra-Musser, Sarah van Norden, and Li Yi.
Throughout my time at USC, I’ve had the privilege of working with other trans experts
from around the world. I am deeply indebted to not only my intellectual communities in the
Transgender Professional Association for Transgender Health, and the Center for Applied
Transgender Studies, but also for their unwavering support. Being able to lean on you all
throughout this has been nothing short of lifesaving. In particular I want to thank Dr. TJ Billard,
Dr. Elle Lett, Dr. Clair Kronk, Dr. Eartha Mae Guthman, (soon to be Dr.) Noah Adams, (soon to
be) Dr. Erique Zhang, and myriad others. I also want to thank the steering committee for our
conference “Converging Crises” we held in 2021. Your willingness to work on a rail-thin budget
and put together an event attended by over 500 clinicians, activists, advocates, and researchers
was inspirational. I look forward to continuing to work together.
Beyond this group of folks, I want to thank Drs. Kristi Gamarel and Oliver Haimson for
their unwavering support during my job market year. I don’t know how I would have gotten
iv
through to this point without you. I can’t wait to work together in the future and grow our
friendships and professional collaborations.
In addition to my scholarly support network and my trans and queer community, I have
had the rare experience of an unconditionally loving and supportive family. To my parents,
Teresa and Dennis, I thank you for raising me to never forget where I’m from, who I am, and
what I’m on Earth to do. To my siblings by blood and marriage, I am deeply indebted to you for
your support and friendship. Kathryn, Jesse, Max and Libby you’ve each played an integral part
in my becoming who I am and getting to this point. I learned from each of your examples and
I’m as proud as can be to call myself an Everhart.
Last, but certainly not least, I have to thank my best friend Sean Waldbillig, and my
partner Matt Gray Brush. Between the two of you, you’ve seen me through this degree from start
to finish. I’m forever grateful for your friendship, Sean. And to Matthew, I can’t even begin to
express what your constant support, love, and kindness have meant to me. You have become the
skeleton key to every door that I need opening to make my life not only livable, but joyful. I
thank you for your grace, your patience, and for sharing in my tenacity and ambition.

v
Table of Contents
Dedication ....................................................................................................................................... ii
Acknowledgements ........................................................................................................................ iii
List of Tables ............................................................................................................................... viii
List of Figures ................................................................................................................................ ix
Abbreviations ................................................................................................................................. xi
Abstract ......................................................................................................................................... xii
Chapter 1 Introduction .................................................................................................................... 1
1.1. Transgender Life Post-Tipping Point..................................................................................1
1.2. Outline of Chapters & Research Motivations .....................................................................2
Chapter 2 Construction and Validation of a Spatial Database of Providers of Transgender
Hormone Therapy in the US.................................................................................................... 5
2.1. Introduction .........................................................................................................................5
2.1.1. Study Design ..............................................................................................................7
2.2. Methods...............................................................................................................................8
2.2.1. Database Construction ...............................................................................................8
2.2.2. Validation .................................................................................................................11
2.3. Results ...............................................................................................................................14
2.4. Discussion .........................................................................................................................19
2.4.1. Interpretation of Results ...........................................................................................19
2.4.2. Limitations ...............................................................................................................23
2.4.3. Recommendations ....................................................................................................25
Chapter 3 Contingent Communities, Defective Demographics: Spatial Analysis Reveals
Limitations to Understanding Transgender Life in the US ................................................... 26
3.1. Introduction .......................................................................................................................26
3.1.1. Public Health Surveillance Systems and Population Surveys .................................28
vi
3.1.2. Transgender Communities and the Small Numbers Issue .......................................30
3.1.3. Making Transgender Count .....................................................................................32
3.2. Data and Methods .............................................................................................................33
3.2.1. Behavioral Risk Factor Surveillance System ...........................................................33
3.2.2. Household Pulse Survey ..........................................................................................34
3.2.3. Differential Protocols for Collecting and Ascertaining Gender Identity Data ........35
3.2.4. Comparison of Weighting Procedures .....................................................................37
3.3. Results ...............................................................................................................................38
3.3.1. Identifying Trans Subpopulations within the General Population ...........................38
3.3.2. Geographic Representativeness ...............................................................................40
3.3.3. Ethno-Racial Diversity within Trans Population .....................................................45
3.4. Discussion .........................................................................................................................46
3.5. Conclusion ........................................................................................................................51
Chapter 4 Measuring Geographic Access to Transgender Hormone Therapy in Texas: A
Three-step Floating Catchment Area Analysis...................................................................... 52
4.1. Introduction .......................................................................................................................52
4.2. Data Sources .....................................................................................................................55
4.2.1. Demographic Data ...................................................................................................55
4.2.2. Care Shortages, Rurality & Urbanicity ....................................................................56
4.3. Methods.............................................................................................................................58
4.4. Results ...............................................................................................................................60
4.5. Discussion and Conclusion ...............................................................................................69
Chapter 5 Conclusion and Future Work ....................................................................................... 75
5.1. Future Work ......................................................................................................................78
References ..................................................................................................................................... 85
Appendix A Supplementary Table from Chapter 2 .................................................................... 101
vii
Appendix B Supplement to Chapter 3 ........................................................................................ 103

viii
List of Tables
Table 2.1 Stratification by self-reported travel distance of USTS subsample who reported
current GAHT use and current access to trans-related care provider .................................... 16
Table 2.2 Comparison of derived distance to nearest provider and self-reported travel
distance stratified by USTS question 11.5 categories ........................................................... 16
Table 3.1 Comparison of unweighted and weighted rates of identification as trans across the
BRFSS 2014-2020 and HPS pooled data .............................................................................. 40
Table 3.2 Comparison of transgender identification rates in select states and metropolitan
areas in HPS pooled data ....................................................................................................... 44
Table 3.3 Comparison of ethnoracial breakdown of general population versus trans
subpopulation in the BRFSS ................................................................................................. 45
Table 3.4 Comparison of ethnoracial breakdown of general population versus trans
subpopulation in the HPS ...................................................................................................... 46
Table 4.1 Number and percentage of census tracts in study area by geographic area
stratified by SPAI for GAHT access ..................................................................................... 61
Table A.1. Number of respondents in the USTS subsample, the number of GAHT facilities,
and the ratio used in Figure 2.4 of facilities per 1,000 respondents stratified by state ....... 101
Table B.1. Comparison of unweighted and weighted rates of identification as trans across
the BRFSS and HPS pooled data with the addition of the “none of these” category
from the HPS ....................................................................................................................... 103

ix

List of Figures
Figure 2.1 Process for construction of the GAHT provider database ........................................... 10
Figure 2.2 Construction of USTS subsample for testing with provider database ......................... 12
Figure 2.3 Question 11.5 from the USTS questionnaire ............................................................... 13
Figure 2.4 Distribution of facilities with providers who offer GAHT to trans patients per
1,000 respondents in the USTS subsample visualized by state ............................................. 15
Figure 2.5 Number of USTS respondents who may have accessed providers not included in
database expressed as a percentage of the total USTS respondents from the subsample
by state ................................................................................................................................... 18
Figure 3.1 Unweighted versus weighted rates of transgender identification from year
to year in BRFSS data (2014-2020) ...................................................................................... 39
Figure 3.2 Number of years the optional BRFSS SOGI module was used by each state.
Those states with cross-hatching have no data because they have never used the SOGI
module ................................................................................................................................... 41
Figure 3.3 State level rates of transgend identification pooled across all seven waves of data
collection for the BRFSS ....................................................................................................... 42
Figure 3.4 Rates of identification as trans from the HPS pooled at state level ............................. 43
Figure 3.5 Percent change from BRFSS to HPS in rates of identification as transgender by
state ........................................................................................................................................ 44
Figure 4.1 Spatial Accessibility Index (SPAI) values for GAHT access across all census
tracts in Texas. Grey represents care shortages defined as SPAI values less than
1/3500 and whited out tracts are censored ............................................................................ 63
Figure 4.2 Spatial Accessibility Index Ratio (SPAR) values for access to GAHT across all
Texas census tracts except for the extreme values. The classes are divided into three
quantiles with the lightest shade representing the quantile with the lowest access to
GAHT and the darkest representing the quantile with the highest. Those areas in white
are censored. .......................................................................................................................... 64
Figure 4.3 Bivariate map of access to GAHT by urbanicity. The same threshold for high
GAHT access as in Figure 4.1 is used and low urbanicity is defined as only those
areas classified as rural (e.g., those with rural-urban continuum codes of 8 or higher). ....... 65
Figure 4.4 Bivariate map illustrating the distribution of MUAs and the distribution of
spatial access to GAHT. MUAs were dichotomous and “low GAHT access” was
x
defined as SPAI of less than the physician to population ratio discussed above of
1/3500 and “high GAHT access” was any SPAI above that threshold. ................................ 67
Figure 4.5 Hot spot analysis showing clusters of high and low access to GAHT for
transgender people ................................................................................................................. 68
Figure 4.6 Bivariate map showing MUA status and hot or cold spot according to
Getis-Ord Gi* analysis of spatial access to GAHT for transgender people .......................... 70

xi
Abbreviations
3SFCA 3-step floating catchment area
AAAQ Availability, Accessibility, Acceptability & Quality
AFAB Assigned female at birth
AMAB Assigned male at birth
ASAB Assigned sex at birth
BRFSS Behavioral Risk Factor Surveillance System
FCA Floating Catchment Area
GAHT Gender-affirming hormone therapy
GI Gender identity
GIS Geographic Information Systems
HPS Household Pulse Survey
ICESCR International Covenant on Economic, Social & Cultural Rights
LGBTQ Lesbian, Gay, Bisexual, Transgender & Queer
MSA Metropolitan Statistical Area
MUA Medically underserved area
TGD Transgender and gender diverse
SOGI Sexual orientation & gender identity
USTS US Transgender Survey

xii
Abstract
Few studies have analyzed geographic access to transgender-specific medical care, and
none have done so at a scale larger than a single city. However, this is perhaps due to a lack of
reliable and readily available data on transgender populations. To investigate access to trans-
specific care, this dissertation focuses on gender-affirming hormone therapy (GAHT) because it
is routine care and entails at least annual visits. Given that the provision of GAHT is not tracked
by public health entities the way other kinds of medical care are, the first aim of this dissertation
is to determine where this care is provided in the US. Chapter 2 features the construction of a
spatial, national database of GAHT providers in the US as well as a test of its comprehensiveness
using the US Transgender Survey data. Chapter 3 offers an exploratory spatial analysis of rates
of transgender identification across the US comparing the most often cited data source, the
Behavioral Risk Factor Surveillance System, with a new source, the Household Pulse Survey.
Ultimately this chapter illustrates that existing estimates of the size and demographics of the US
trans population are likely severe undercounts. Chapter 4 builds on chapters 2 and 3 to quantify
geographic access to GAHT using a three-step floating catchment area method (3SFCA). It
deploys population estimates from chapter 3 to estimate the distribution of trans populations at
the census tract level in Texas. Next, it uses the 3SFCA method to measure access for these tract-
level trans subpopulations to GAHT providers derived from the GAHT provider database
detailed in chapter 2. It also offers recommendations for planning health services and concludes
that patterns of geographic access to transgender-specific medical care do not follow known
patterns of access to primary care for the general population. The final dissertation chapter
provides a synthesis and some policy recommendations to address the data gaps that served as
the impetus for these studies.

1

Chapter 1 Introduction
This dissertation focuses on one specific aspect of transgender-specific medical care:
gender-affirming hormone therapy. Thus, it is crucial to begin with a note on language.
Transgender is a broad category of identities and can be defined as a name for those people
whose current gender identities differ from the sex they were assigned at birth (Stryker 2008).
Throughout this dissertation we use transgender interchangeably with the shortened form “trans,”
(Stryker, Currah, and Moore 2008). However, the definition we offer in this introduction is not
universal and each of the data sets we draw from has its own distinct definition of transgender. In
these instances, we defer to the definition used in the data collection and dissemination, but we
generalize to the broadest possible and most inclusive definition of trans people. Given that this
work focuses on gender-affirming hormone therapy (GAHT), it is also important to note that not
all people who identify as trans, or identify with the term, medically transition. When it is
pertinent, we delineate and specify whether we are referring to all trans people or only to those
trans people who may have accessed GAHT in the past or the present or may do so in the future.
Ultimately this dissertation’s focus on GAHT is a reflection of the existing gaps in the scientific
literature and the introductory sections of each substantive chapter highlight those gaps.
1.1. Transgender Life Post-Tipping Point
When Laverne Cox, a Black trans woman of Orange is the New Black fame, appeared on
the cover for TIME magazine, the related story declared that we had reached a “transgender
tipping point,” (Steinmetz 2014). While the tipping point may have represented a pivotal
moment in media representation for trans communities in the US, and arguably the world over,
many of the issues that Steinmetz named in the article, from difficulties in changing gender
markers on identification documents to limited healthcare access for youth and adults alike,

2

remain central concerns for trans people. In the first four months of 2022 alone, there have been
over 200 bills put forth in state legislatures attempting to outlaw, or in some cases criminalize,
what amounts to the material support and affirmation of trans people’s lived realities. It is in this
deeply polarized political landscape that this dissertation takes up the topic of transgender
healthcare. And while the analyses offered in subsequent chapters proceed without a direct
empirical engagement with legislative attacks on transgender lives, it is impossible to consider
this work outside of that context. Thus, we begin the dissertation with this brief mention of the
complex and harrowing sociolegal environment in which trans people continue to toil to be able
to live their truths.
1.2. Outline of Chapters & Research Motivations
What follows is a set of three studies that seek to address iterative research questions that
remain largely unanswered, or at least underexplored within the scientific literature on
transgender communities. The first study (chapter 2) asks where transgender-specific medical
care, and more specifically GAHT, is provided in the US. Myriad community-facing resources
exist that are designed to point trans people seeking GAHT in the appropriate direction to find
friendly and knowledgeable providers; however, this kind of healthcare is not tracked by state
and federal data repositories that pinpoint where medical care facilities are located and what kind
of care they offer. Thus, we created a spatial database of facilities where GAHT is offered and
tested its comprehensiveness against the largest data set of all trans-identified respondents: the
US Transgender Survey (James et al. 2016). Through this first study, we produced a research
data set that illustrates the supply of GAHT care in the US.
Having established the supply side of the equation, we turned to the demand in the
second study (chapter 3). Existing estimates of the size of the US transgender population vary

3

widely depending on the source or context, but the most often cited estimate in the health
sciences literature stems from a single report (Flores et al. 2016), which is itself from a single
data source: the Behavioral Risk Factor Surveillance System (CDC and BRFSS 2015). In this
study we undertook an exploratory spatial demographic analysis by comparing the BRFSS data
with a new data source: the Household Pulse Survey (US Census Bureau 2021a). Through this
analysis we demonstrated that there is strong evidence that existing population estimates
circulating in the scientific literature may be severe undercounts. And while not all trans people
transition using GAHT, the population estimates produced from this analysis offer the demand to
complement the supply from Chapter 2.
Finally, in chapter 4 we put these pieces, the spatial database of GAHT providers and the
spatially explicit estimates of the rate of trans identification, together. The primary research
motivation in this third and final study was to understand the geographic distribution of
accessible GAHT care across the state of Texas using the best available methods from the fields
of medical geography and spatial epidemiology. To do this, we used a three-step floating
catchment area (3SFCA) that quantifies the spatial accessibility of target populations to
healthcare facilities within an established drive time window (Wan, Zou, and Sternberg 2012).
The case study in chapter 4 focuses on the state of Texas because of its spread of rural counties
interspersed with large, metropolitan areas. The results demonstrate how an improvement in the
available data for researching access to healthcare for trans people can produce insights that are
much more actionable and precise than in previous studies without a spatially explicit lens.
Ultimately, these three studies serve two central goals. The first goal is to deploy the best
available data in order to illustrate where geographic disparities in access to trans-specific
healthcare may exist. The Texas case study serves as an instructive model that can be replicated

4

in other geographic contexts and at any scale for which data exists. This first goal is more
practical in that it can be more immediately translated into operationalizable information used by
health services planners, federal or state governmental agencies, and other researchers.
The second goal is more theoretical. Each of these studies also serves as an illustration of
the limitations of the existing data available on transgender life that we rely upon to produce the
actionable insights needed to improve trans people’s lives. Kevin Guyan’s recent book Queer
Data: Using Gender, Sex and Sexuality Data for Action makes the convincing argument that
unless queer and trans communities are counted, meaning quite literally enumerated, then we do
not count (Guyan 2022). By drawing on both the empirical and theoretical branches of
scholarship on data and queer communities he suggests that “queer data emerges as the
production of tension between categories and anti-categories, assimilation and difference,
intrinsic qualities and social constructs, and individuals and populations,” (Guyan 2022, p. 14).
Given the timing of its publication, it is likely that we were working on our separate projects at
the same time; however, we still claim his work as an inspiration for this dissertation and hope
that it can both effect meaningful, material change and simultaneously insist that we refuse the
folding of data on trans communities and lives into the machinations of population surveillance
for surveillance’s sake. At its best, data can be a tool for what Lisa Bowleg calls intersectionality
praxis (Bowleg 2021), that is, the theoretical push of intersectionality to move beyond
documentation of disparities or poor health outcomes and life experiences and toward the
development of solutions informed by both the theory and practice of liberation. While neither
intersectionality, nor queer theory feature in the empirical core of this dissertation, scholarship
like Bowleg’s and Guyan’s have pushed us to think critically about the politics of data, from
collection to analysis to dissemination, even as we both wield and produce it.

5

Chapter 2 Construction and Validation of a Spatial Database of Providers of Transgender
Hormone Therapy in the US
The scientific literature on transgender health has established that trans people face many
barriers to accessing healthcare generally and especially trans-specific medical care. However,
no studies to date have sought to identify where trans-specific medical care is provided. This
chapter outlines the method for the construction of a spatial database of trans hormone therapy
providers across the US and includes an analysis of its efficacy and comprehensiveness using
data from the largest survey of trans respondents conducted to date.
2.1. Introduction
Transgender people, that is those whose gender identity differs from the sex they were assigned
at birth (Stryker 2008), often face barriers in accessing healthcare in the US (Gonzales and
Henning‐Smith 2017; Safer et al. 2016; Arrowsmith 2017). These barriers include lack of or
inadequate insurance coverage for transgender-specific healthcare (Bakko and Kattari 2020;
2021), and anticipated discrimination due to being trans (Kcomt 2020; Poteat 2013; Romanelli
and Lindsey 2020; Rodriguez, Agardh, and Asamoah 2018a; Rood et al. 2016) which is often
worse among trans people of color (Macapagal, Bhatia, and Greene 2016; Kattari et al. 2015;
Goldenberg et al. 2019). Notably these barriers lead to differential access to services across states
(White Hughto et al. 2016; Kachen and Pharr 2020; O’Bryan et al. 2020; Goldenberg et al. 2020;
Zaliznyak et al. 2021), and could be driven by a lack of emphasis on cultural competency in
working with transgender patients in medical education across disciplines (Arora et al. 2020; K.
A. Clark, White Hughto, and Pachankis 2017; Jabson, Mitchell, and Doty 2016; Khalili, Leung,
and Diamant 2015; Korpaisarn and Safer 2018; Moe, Perera-Diltz, and Sparkman-Key 2018;
Rowan et al. 2019; Safer 2013; Snelgrove et al. 2012; Turban et al. 2018); although, the situation

6

is improving (Eriksson and Safer 2016; Redfern and Sinclair 2014; Ali, Fleisher, and Erickson
2016; Bristol, Kostelec, and MacDonald 2018). Even further, the US Transgender Survey
(USTS), the largest nationally representative sample of transgender people found that as many as
33% of all respondents who had seen a medical provider within the last year reported some kind
of negative experience due to being transgender when accessing care (James et al. 2016).
1
These
systemic and structural barriers to care lead to health inequities and poor health outcomes (Zou et
al. 2018; Cicero et al. 2020; Downing and Przedworski 2018), which are exacerbated by
structural racism (Lett et al. 2020; Lett, Dowshen, and Baker 2020).
Geographic information systems (GIS) have been taken up more and more within
biomedical and health services research (Wang 2020). Yet, despite a growing body of literature
on the experiences of transgender and gender diverse (TGD) people in accessing healthcare, very
little research has been conducted to discern where trans-specific medical care is offered, even in
a resource rich country like the US. While the US has myriad health surveillance systems, and
registries at various geographic granularities to identify what kinds of care are available and
where, none yet exists which describes the full extent of transgender-specific medical care.
While this poses problems for trans people seeking care first and foremost, it also poses a
significant problem to researchers who seek to understand patterns of access to care to improve
them for trans communities. The focus on barriers to care, discrimination, avoidance, and denial
has overshadowed the need for a better understanding of availability of trans-specific medical
services.

1
This measure is skewed by the demographic weights that the USTS uses to render the data
more like larger patterns according to the US Census Bureau along lines of age, education, and
race. These weights also attempt to ‘correct’ for a lack of racial diversity within the sample, more
than 80% of whom self-identified as white. We use it here as a citation, but the measure would
look different if calculated as an unweighted statistic.

7

2.1.1. Study Design
This study aims to address the lack of a comprehensive registry of transgender medical service
providers by outlining the construction of a spatial database of facilities where one or more
providers offer gender-affirming hormone therapy (GAHT) to transgender patients.
2
There is a
precedent for constructing these kinds of spatial databases for healthcare facilities (Maina et al.
2019; Noor et al. 2009; Linard et al. 2010; MacRury et al. 2016), and the need for such databases
has been demonstrated for other populations with access issues, especially to address rural
disparities in access to care (A. Baker 2011). From recent work on health information seeking
behavior among transgender people (Augustaitis et al. 2021; Haimson and Veinot 2020), we
know the barriers to accessing trans-specific medical care begin with the reality that many do not
even know what services are available to them. Before they even access care to experience these
persistent patterns of discrimination, transgender care seekers often do not know where to begin
looking for providers. Therefore, this study aims to close this gap. Ultimately, we ask where this
kind of care is provided and how can it be found using available resources and existing
methodologies. While dissemination of this information to the community of people seeking this
care is important, the goal in this chapter is constructing the database for research purposes. In
the next section, we outline the method for constructing the database and how we tested its
efficacy. The best available means for determining the efficacy of the database was to use the
2015 US Transgender Survey, the largest sample ever gathered of exclusively transgender-
identified respondents.

2
Gender-affirming hormone therapy may also be referred to as hormone replacement therapy,
hormone therapy, or even cross-sex hormone therapy. This is the provision of exogenous sex-
linked hormones concordant with the patient’s true gender (e.g., a woman who was assigned
male at birth, a trans woman, would take estrogen and/or anti-androgens to convert from
testosterone dominance to estrogen dominance).

8

2.2. Methods
2.2.1. Database Construction
Geographic information was pulled from existing databases and provider resources first. There
were two databases which were already spatially enabled: Erin’s Informed Consent HRT Map
and Trans* in the South produced by the Campaign for Southern Equality. Each of these data
sets were pulled as KML files then converted to shapefiles for analysis in ArcGIS Pro. These
files were then reverse geocoded to produce address information in keeping with the format of
most of the other data and for later verification. Other databases which were not yet spatially
enabled included Out Care Health, the Gender Infinity Resource Locator, and Planned
Parenthood. Additionally, once the Trans in the South report was re-released in 2021, the list of
providers was updated, but was only available in tabular form on their website. Notably, Out
Care Health and Planned Parenthood required some internal searching because they were not
trans-specific resources. Each of these web resources were searched using only the word
‘transgender.’ Out Care Health had an interactive provider list that could be filtered by search
term and 268 relevant providers were included when filtered by ‘transgender.’ Planned
Parenthood had a resource list containing links to all the facilities that provide GAHT to trans
patients for each state. The remaining aspatial resources contained provider information in
tabular format including varying degrees of specific geographic information (full address, city
and state, ZIP code, etc.). Data from these sources were scraped and then added to the database.
Once these data from the above six sources were amalgamated, they were verified
manually. This entailed removal of duplicates, and manual confirmation of providers using
inclusion and exclusion criteria. Only those facilities which had publicly available information
on the internet that indicated they provide GAHT to transgender patients were kept. In this
process, any facilities that had demographic exclusions were removed. These included facilities

9

which were either: 1) part of the Veterans Affairs health system and therefore required veteran
status to receive care, 2) adolescent or pediatric facilities that had age restrictions for accessing
care, 3) health centers located on college or university campuses that only provided care to
registered students, 4) HIV-specific programs that only provided care to patients already
undergoing HIV treatment, 5) women’s and/or feminist healthcare facilities which offered
GAHT to trans people who were assigned female at birth (AFAB) seeking testosterone, but not
to trans people assigned male at birth (AMAB) seeking estrogen, or 6) providers or facilities
which exclusively offered telehealth. These exclusions were made because other researchers with
data on trans communities may not have information related to veteran status, eligibility for
pediatric or adolescent facilities, student status, or HIV status, or even their assigned sex at birth.
Further, information about eligibility criteria for pediatric and adolescent facilities were not
always clear and the USTS data we used for validation only had respondents who were 18 years
old or older.
During the verification process, further resources and providers were found via
snowballing. By searching for the information pulled from the data sources, local and standalone
resources that included such providers were identified. When this happened, further providers
and facilities were added if they were not already included in the database, and they met the
inclusion criteria. Further providers were identified via verification as well. When searching for
websites and other materials, if providers showed up that worked within larger health systems, or
the healthcare facility or program had a website, any other providers who also advertised
offering GAHT to transgender patients and met the inclusion criteria were added as well. Figure
2.1 explains this process in further detail.

10

Figure 2.1. Process for construction of the GAHT provider database

11

2.2.2. Validation
After the data set of GAHT providers was geocoded and finalized with correct location
information, an analysis was run to test its efficacy. As mentioned in the study design, we tested
the efficacy of the database using the USTS, access to which was provided by the Institutional
Review Board of the University of Southern California. The USTS is the largest sample of
transgender people ever collected with over 27,000 observations. Using the USTS data, we
constructed a subsample of respondents who reported accessing a trans care provider (see Figure
2.2). The relevant question in the USTS asked respondents to select a categorical response based
on approximate travel distance (see Figure 2.3); however, there were no further data collected on
mode of transportation. These responses were geocoded as points by self-reported ZIP code of
residence using ArcGIS Pro and its address locator. For ZIP codes with multiple responses, the
same location for each response was used to control error in calculation distance with
consistency. The number of responses per ZIP code ranged from 1 to as many as 57; however,
most ZIP codes had only one respondent.
To determine the completeness of the provider database, we measured the availability of
providers from the database, paying particular attention to respondents who reported traveling a
shorter distance than the distance our analysis revealed between their ZIP code of residence and
the nearest provider from the database. To accomplish this, we elected to use buffers, which are
essentially circles with a radius of a specified distance in every direction from a central point. In
this case, we used the ZIP code centroids as centers and constructed buffers of distances that
corresponded to the high end of each category from the above-mentioned categorical question
about travel distance (i.e., 10, 25, 50, 75, and 100 miles, respectively). Respondents who
indicated traveling further than 100 miles were included in the analysis; although, the buffers
must have specified radii, and there was no upper limit for the “more than 100 miles” option.

12

Figure 2.2. Construction of USTS subsample for testing with provider database

13

Figure 2.3. Question 11.5 from the USTS questionnaire

Further, these buffers were constructed with geodesic distances, meaning that straight lines were
drawn for the radii irrespective of the presence of road networks or other means of transportation
across the Earth’s surface. This method of measuring travel distance has been tested against
other methods, for example, network distance which uses road networks to measure travel
distance more explicitly and when working on the national scale, the differences are negligible
(Boscoe, Henry, and Zdeb 2012). Finally, we calculated ratios of available facilities that offer
GAHT to the number of respondents by state. These were calculated to look for regional or state
level patterns, including possible disparities, in availability of care relative to the respondents.
We also conducted brief secondary analyses to ensure that our analysis of the
comprehensiveness of the database accounted for the possibility that some of these unknown
providers could be facilities that were excluded from the final database for their demographic
restrictions. These two analyses focused on veterans and respondents who reported accessing
GAHT before age 18 and were between 18 and 24 years old at the time of the survey. Veterans
would be eligible to access care at Veterans Affairs facilities and young adults may be eligible to
access adolescent or pediatric care. The results for these secondary analyses are included
alongside the results of the overall analysis.

14

2.3. Results
The combination of geocoding existing community resources and snowballing from the
verification process resulted in the discovery of 913 unique facilities where GAHT is offered.
Figure 2.4 illustrates the distribution of these facilities across the country per 1,000 respondents
from the USTS subsample. This ratio does not represent the more common physician to patient
ratios, such as those calculated as part of floating catchment area methods for quantifying access
to healthcare (Bryant and Delamater 2019; Luo 2004; Luo and Qi 2009; McGrail 2012; Tao,
Cheng, and Liu 2020; Wan, Zou, and Sternberg 2012); however, it does reveal patterns of access
to care that can be explored with the use of the provider database. Notably, the central and
southeastern parts of the country along with Alaska and Hawai’i boast higher proportions of
facilities to respondents than others, such as the west coast, or northeast. Montana also stands out
because it had only 39 respondents and nine facilities which makes its ratio of facilities to
respondents appear inflated alongside the other states with much higher response rates to the
USTS. For these results in tabular form, see Appendix A.
The majority (70.91%) of respondents reported traveling 25 miles or less to access
transgender-specific medical care (Table 2.1). Within this majority, 90% reported a travel
distance to care that matched the derived distance to at least one nearby facility from the
database). As the reported travel distance increases, so does the likelihood that there is a provider
within that reported travel range included in the database. However, the other 10% of
respondents within this majority reported a travel distance that indicates they accessed a provider
that was not included the database. For example, if they reported traveling less than 10 miles to
access GAHT, but the buffer analysis did not capture a nearby provider from the GAHT database

15

Figure 2.4. Distribution of facilities with providers who offer GAHT to trans patients per 1,000
respondents in the USTS subsample visualized by state.

until the 50-mile buffer, then that result may point to a gap in the database’s coverage.
Ultimately, this kind of result is useful for measuring the completeness of the database in terms
of how many of the healthcare facilities in the US that offer GAHT are included, and how many
may be missed. The overall sample analysis revealed as many as 1,082 respondents whose
reported travel distance suggests they accessed a provider not included in the database, making
up 7.51% of the total sample. The percentages by category range from 1.27% to 12.4% for
accessing unknown providers. Table 2.2 shows the number of USTS respondents with their
nearest provider from the database in each buffer range stratified by their self-reported travel
distance as well as the number and percent of respondents in each category that may have
accessed unknown providers.

16

Table 2.1. Stratification by self-reported travel distance of USTS subsample who reported
current GAHT use and current access to trans-related care provider
Reported Distance Traveled to GAHT Provider among USTS
subsample
N %
Less than 10 miles 6375 44.24%
10 – 25 miles 3843 26.67%
25 – 50 miles 1957 13.58%
50 – 75 miles 866 6.01%
75 – 100 miles 468 3.25%
Over 100 miles 900 6.25%

Table 2.2. Comparison of derived distance to nearest provider and self-reported travel distance
stratified by USTS question 11.5 categories
N = 14,409 USTS Survey – Self-Reported Travel Distance Categories
<10 mi
10 – 25
mi
25 – 50
mi
50 – 75
mi
75 –
100 mi
>100
mi
Total
No. of
respondents
with at least
one facility
from database
within range
<10
mi
5,584 2,938 1,036 357 182 442 10,097
10 –
25
mi
394 675 614 207 82 122 2,094
25 –
50
mi
241 153 264 224 97 142 1,121
50 –
75
mi
92 55 30 67 81 109 434
75 –
100
mi
35 15 13 6 19 57 145
>100
mi
29 7 0 5 7
28

76
Total 6,375 3,843 1,957 866 468 900 14,409
No. of
respondents
who may have
accessed
providers not
included in the
database

791
(12.4%)
230
(5.98%)
43
(2.20%)
11
(1.27%)
7
(1.50%)
0
(0%)
1,082
(7.51%)

17

Two secondary analyses were performed to test whether the excluded facilities in the first
phase of database construction may account for some of these results. First, we identified the
veterans or active-duty military personnel within the USTS subsample (n=1645) and geocoded
the Veterans Affairs (VA) healthcare facilities advertised through the VA website. There were 23
veterans who may have accessed a VA facility that fell within the travel distance reported where
no other provider from the final database without VA facilities could be found. These 23 account
for only 0.16% of the overall subsample (n=14,409) and 2.13% of all respondents who may have
accessed an unknown provider (1,082). Second, to assess possible missingness due to the
exclusion of adolescent medical programs with age limits, we identified 455 respondents from
the subsample who reported accessing either hormones or puberty blockers before age 18. Of
these, only 24 may have accessed a facility that was not in the database. Within that 24, only 19
were 24 years of age or older and therefore are possibly able to continue accessing their provider
from when they began to medically transition before being of majority age. These 19 represent
only 0.132% of the overall subsample and, like the veterans, account for a negligible amount of
the potential missingness (1.76%) in the final database. Notably, each of these 19 reported ZIP
codes of residence that were within 5 miles of the nearest facility from the final database which
could possibly be attributed to the use of centroids within ZIP code tabulation areas rather than
precise residential locations.
Some pertinent geographic patterns emerge when the results of the buffer analysis
measuring database completeness are mapped. Figure 2.5 shows the distribution of the
percentage of respondents per state who may have accessed providers unknown to the database,
with Wyoming as a clear outlier. While Wyoming was not among the states with the lowest ratio

18

of facilities to respondents, the buffer analysis revealed more than 39% of respondents residing
in the state may have accessed providers not included in the database. Moreover, South Dakota
(30%), West Virginia (25%), and North Dakota (21.74%) were within a similar range indicating
that in these four states in particular the GAHT provider database is less comprehensive. In
contrast, the District of Columbia had no respondents accessing care outside of the district, and
Rhode Island, Minnesota, and Washington all had ≤ 2% of respondents accessing care outside of
the state. This pattern could be due to the relative population distributions of these states.

Figure 2.5. Number of USTS respondents who may have accessed providers not included in
database expressed as a percentage of the total USTS respondents from the subsample by state.

19

Regardless of explanation it is noteworthy that there are no glaring regional disparities, despite
existing discourses related to the socially or politically repressive environments in the Southeast
or Midwest.
2.4. Discussion
2.4.1. Interpretation of Results
There are multiple possible explanations for these results given both the collection of this
specific data in the USTS as a categorical rather than numeric variable, and the differences
between perceived and actual distances. First, self-reported travel distances tend to differ
significantly from real distances between points on the Earth’s surface, regardless of the method
of transportation or method of measurement of these distances (Parthasarathi, Levinson, and
Hochmair 2013; Witlox 2007; Xianyu, Rasouli, and Timmermans 2017). Further, it is possible
that respondents selected an inaccurate category given the nature of the question, which had
overlaps in categories (see Figure 2), and its length, which was 324 question items organized into
32 distinct sections. Finally, it is possible that either the provision of transgender specific GAHT
has increased in availability between the time of the survey and the time during which data were
collected, or perhaps respondents traveled further than was necessary for their GAHT care. The
contributing factors and potential systemic drivers of the decision among trans people to possibly
travel further than is necessary for their GAHT care is of particular interest and should be studied
further. Future work that studies these factors could incorporate the database to complement their
work and measure the differences between nearest available facilities and the facilities trans
people actually access.
Alternatively, it is possible that the facilities at which USTS respondents accessed their
GAHT care were initially included via scraping the data from existing resources and later

20

excluded based on our criteria. Two specific kinds of healthcare facilities may be pertinent to
understanding the availability of GAHT care but were not included in the database because of
demographic restrictions, meaning that the care provided at these facilities are limited to specific
populations rather than available to anyone of legal age to consent to care. First are facilities that
are part of the Veterans Affairs health system which were excluded based on the requirement
that care seekers have veteran status. Second were student health centers on college and
university campuses which required student status to access care. We conducted a secondary
analysis for both veterans and respondents aged 18-24, but these subsets of the USTS sample did
not account for a significant amount of missingness (see Appendix B). Ultimately, excluded
providers from the VA system and other facilities with demographic restrictions for accessing
GAHT were not included because veterans are not required to access care via the VA system,
those individuals who begin GAHT as minors are not required to continue with their pediatric or
adolescent care providers upon becoming majority age, nor are college students required to
access care at their educational institution. It is impossible to determine which of these is the
more likely scenario either based on existing literature or the information available from the
provider database and the USTS data. What it does point to is the possibility that patterns of
accessing care may be more complex than the current literature assumes. More research is
needed to understand the driving factors behind decisions to travel further than may be necessary
to access care. Given the established barriers to healthcare for trans people, it stands to reason
that trans community members would informally refer one another to providers and be willing to
travel to access competent and conscientious care. While these care-seeking behaviors are
beyond the scope of this dissertation, the dataset upon which this chapter focuses may be useful
to researchers and health services planners seeking to understand facilitators to accessing care,

21

thus shifting the focus away from just barriers. Further, from a provider perspective, it is
important to note that there are myriad reasons why clinicians may not advertise offering these
services including stigma, ostracization within their communities, and ongoing efforts to legally
limit or even criminalize the provision of this kind of care.
Moreover, capturing the experiences of transgender patients within health systems proves
challenging. While this chapter focuses on the construction of a national, spatial database of
GAHT providers, it is important to note the ongoing challenges that researchers studying
transgender populations face. Researchers and clinicians with lived experience have been writing
about the utility of two-step methods for capturing gender identity (Tate, Ledbetter, and Youssef
2013; Deutsch et al. 2013; Deutsch and Buchholz 2015), but it remains uncommon to include
such measures in electronic health records (EHR), despite evidence that it would be acceptable
for a general population to capture gender identity beneficial for trans populations (Bosse et al.
2018; Kronk et al. 2021; Maragh-Bass 2019; Maragh-Bass et al. 2017). There has been some
success in recent years in Argentina to meaningfully incorporate gender identity into EHRs
(Cassarino, Correa, Minoletti, Jauregui, et al. 2020; Cassarino, Correa, Minoletti, Botto, et al.
2020) and even into public health surveillance systems (Levi et al. 2019), but thus far this has not
been taken up in the US.
3
Even when EHRs may capture self-reported gender, or include
preferred names and pronouns, the laboratories where blood and other samples are tested may be
inadequately equipped to capture that information (Goldstein, Corneil, and Greene 2017), which

3
The authors are aware that there is an ongoing effort to incorporate sexual orientation and
gender identity measures into county-wide data collection protocols within Los Angeles County,
but it is in the design and community feedback stage. Implementation does not yet have a
publicly available timeline.

22

could lead to incorrect interpretation of results that influence care provision (Humble et al.
2019).
The sociolegal environment for transgender people in the US is repressive. In the first
four months of 2022 alone over 200 bills have been introduced to limit access to trans-specific
medical care, participation in sports, and full participation in American life. This environment is
further complicated by the health system in this country. And in fact, some of the above bills
attempt to criminalize the provision of medically necessary, gender-affirming care. In this
environment and in an already complex and woefully inadequate health system, transgender
people often rely on informal networks to access essential medicines, not unlike in other
countries (Arístegui et al. 2017; Tan et al. 2020). Additionally, the ongoing COVID-19 pandemic
enabled a surge of telehealth care provision in the US where it often was not covered by
insurance until it became unsafe to congregate indoors. Even though the chief concern of this
chapter and its analysis was to outline the method for the construction of a provider database, it
is important to note that telehealth and self-care interventions, including informal non-licensed
access to exogenous hormones, are well beyond the scope of this work.
These factors, limitations in available information on where to access medical care, the
inherent limitations of the best available and largest datasets, a repressive sociolegal
environment, and a difficult to navigate health system pose significant challenges to existing
methodology. Working on transgender issues in any arena often poses challenges to both
individuals accessing transgender healthcare and researchers studying it. More specifically for
research, existing methods for measuring population health outcomes assume dichotomously
opposed and essentialist categories of sex and gender (Lagos 2018), and trans communities are
regarded as ‘hard to reach’ and ‘vulnerable’ populations (Reback et al. 2015; Katz et al. 2020).

23

In keeping with similar research, this study also illustrates the challenges faced when conducting
research on transgender healthcare. However, what we have done differently is embrace the need
for innovation and mixed methods to address a pressing research question. Working on
transgender issues can push us past existing methodologies and expected outcomes to rethink the
assumptions inherent in our approaches and theories.
2.4.2. Limitations
The authors are not aware of any studies that effectively measure the availability of GAHT for
transgender people in the US, though there are some that measure availability of surgical care
(Canner et al. 2018; Nolan, Kuhner, and Dy 2019; Terris-Feldman et al. 2020). In fact, the study
that is the focus of this chapter demonstrates that it is necessary to rely upon community-
generated resources to attempt to measure where and how much care is available to transgender
communities. The challenges of this study were due in no small part to the disparate nature of the
health system in the US; however, it is common to pull from publicly available datasets to
discern where healthcare facilities are located and what kinds of care they provide. What proved
impossible was relying on those datasets alone to determine where transgender-specific medical
care is provided. Moreover, there are potential pitfalls to advertising the provision of
transgender-specific medical care in conventional care databases, given the current political
climate and ongoing legislative attacks on transgender communities. However, doing so could
increase awareness of available services, and thereby help to proliferate access to care.
First, it is important to note that we constructed this database over a period from Fall
2019 to Spring 2021. The landscape for health services for trans people is constantly changing as
evidenced by the fact that so many facilities initially included in these databases could not be
verified with publicly available information. There could very easily be temporal mismatch

24

between the database and the USTS data used to validate the efficacy of the database in
predicting availability of care. Further, it may not be possible to create a comprehensive database
because of how complex the health system is in the US. The goal of this study was to illustrate
the current landscape of GAHT providers to the best of our ability. Given the focus on barriers to
accessing care and what we know about health information seeking behavior in general and for
trans people specifically, it stands to reason that the scientific literature may suggest that there is
far less availability of services than exists.
In addition, the USTS data had clear limitations in terms of its generalizability that are
beyond the scope of this chapter. A lack of racial and ethnic diversity is troubling given how
frequently the USTS data is used, and this topic has received some attention recently (Lett and
Everhart 2021). However, one immediately relevant limitation is in the design of the question
used for the subsample creation in the analysis. Self-reported distances are often inaccurate and
less useful than empirical measures or ground truth. Further, the stratification of categories was
uneven which limited the interpretability of the results. The buffers aligned with the upper limits
to account for the overlap between categories of self-reported distance in the USTS question;
however, this overlap may have influenced responses which cannot be corrected. Ultimately the
largest assumption that this analysis makes is that a respondent would travel to the nearest
available GAHT provider and that their self-reported travel distance reflects that. The second
largest assumption is that respondents answered correctly, by which we mean accurately in terms
of real distance across the Earth’s surface and in terms of accuracy to their own perception. It is
important to note that this chapter does not infer or conclude anything about travel patterns to
accessing transgender-specific medical care. Rather, the analysis provided is a preliminary

25

measure of potential efficacy of a spatial database of GAHT providers constructed from the
mixed-methods approach outlined above.
2.4.3. Recommendations
Future work should consider the drivers of availability of care to complement the existing data
on both health information seeking and on drivers of accessibility to care for trans communities.
For example, work that surveys providers about their motivations for offering GAHT, or services
in general, to trans people would be a significant contribution to our understanding of how the
health system enables and discourages care seeking among trans people. Methods for data
collection on other types of parallel services, notably abortion and reproductive healthcare, exist
and are robust. In 2014, the Guttmacher Institute began conducting an ‘abortion provider census’
in the US that could easily be extrapolated to enumerate providers of transgender medical care
(Jones, Witwer, and Jerman 2017). While the method is sound, the cost of such a means of data
collection is prohibitive, especially given the landscape of funding for trans specific research.
Our methodology for database construction may also prove useful to researchers working on
other types of services whether they are politicized, like abortion and HIV specific care, or
specialized, like audiology or pain medicine. Ultimately, further work on understanding where
care is provided to trans people would be a vital contribution to the literature.

26

Chapter 3 Contingent Communities, Defective Demographics:
Spatial Analysis Reveals Limitations to Understanding Transgender Life in the US
This chapter interrogates the most often cited estimate of the size of the transgender population
in the US and compares it to more recent data from a new source. This chapter demonstrates,
largely through descriptive and exploratory analysis, that there are strong reasons to doubt the
accuracy of existing trans population estimates and that geographically representative sampling
is needed to better understand the demographics of the US trans population.
3.1. Introduction
The intersection of geography and population health has expanded in tandem with the rise of the
use of public health surveillance systems and population level surveys. Researchers have
demonstrated that the spatial turn in demographic and health research has coincided with not
only increases in availability of spatially explicit data, but also the ongoing advancement in
spatial statistics and geographic methods for drawing insight into population level patterns
(Richardson et al. 2013). Beyond the health arena, methods from within geographic information
science (GIS) have been adapted to study macro-level patterns in residential segregation (Catney
and Lloyd 2020; Roberto and Korver-Glenn 2021), and when conducted in combination with
qualitative empirical work, like ethnography (Carney 2021), rich insights can be gleaned. The
larger field of spatial demography has undergone a deep transformation as well, and researchers
within the field have illustrated the necessity of geographic thinking and spatial analysis in
analyzing and describing population level phenomena (Raymer, Willekens, and Rogers 2019). In
a recent critical reflection on the field over the past decade, experts identified the need for further
innovation in method, more meaningful inclusion of researchers working in the Global South,

27

and a recognition of the importance of localized versus global models for understanding identity,
and therefore, the categorization upon which demographic research relies (Matthews et al. 2021).
However, much of the discussion at these intersections relies upon two core assumptions,
which this chapter seeks to address through its case study of the US transgender population. The
first is data availability. Most methods, from geographically weighted regression to Bayesian
spatiotemporal models, and even less complex analytical methods such as measures of spatial
autocorrelation rely on the assumption of a large amount of data. While some methods, such as
spatial microsimulation (Tanton 2017), have emerged to synthesize data and fill gaps in available
data, the tools used by spatial demography to draw meaningful conclusions about demographic
patterns require a baseline level of knowledge that we do not yet have for transgender
populations. The second assumption is data quality. These methods enable researchers to analyze
and describe demographic patterns because the data used to do so are reliable. This is not
universally the case, especially in less resource-rich countries; however, spatial demographic
analysis requires high quality, reliable data to be of use.
In this chapter, we use descriptive analysis with the best available data on transgender
populations in the US to bolster existing understandings of American transgender life. By testing
whether what has been established in the literature on transgender communities holds true when
comparing datasets, we aim to offer a firm foundation upon which more robust analyses at this
intersection of spatial demography and population health can be conducted with transgender
communities. Indeed, we argue that this focus on transgender communities will reveal what
foundational data and insights are needed to better understand any community that faces the
small numbers issue.

28

3.1.1. Public Health Surveillance Systems and Population Surveys
The use of public health surveillance systems data in the scientific literature has dramatically
risen in recent years, as has their deployment in population level research. While initially
designed to understand patterns of infectious disease spread, they have been expanded to
incorporate multiple determinants of health, chronic disease outcomes, mental health, and
injuries. Moreover, the combined use of historical and contemporaneous data along with
significant improvements in computational capacity, statistical and other methodologies, and
heightened attention to targeted health interventions have driven this increase in the use of public
health surveillance systems data (Groseclose and Buckeridge 2017).
Despite these advances, it remains difficult to pose specific research questions about
subpopulations that account for small portions of the total population. Given that these systems,
and other population level surveys, are designed to be generalizable and represent the entire
population, those populations and communities which are smaller in size may appear less
frequently in the data. The smaller the numbers of a community or subpopulation, the less likely
they are to be represented in population level surveys and surveillance systems. Population
researchers have explored the effects of this small numbers issue on Native Hawai’ian and
Pacific Islanders in the US (Galinsky et al. 2019), including changes in self-identified race and
ethnicity over time (Liebler et al. 2017), the centering of Black-White comparisons in measures
of segregation (Fowler et al. 2016), and the necessity of disaggregated Native Hawaiian and
Pacific Islander data for health equity (Kauh, Read, and Scheitler 2021; Kaholokula et al. 2019).
Even within larger populations, the terminology used, and the method of data collection
can affect the quality of the insights gleaned from disaggregation. Another case that population
health researchers have studied is the confusion surrounding Hispanic & Latine terminology
(López and Míguez Bóveda 2021). A systematic review of the literature on Latine health

29

revealed that in spite of the relatively large Latine population in the US, the extant literature
rarely incorporated meaningful disaggregation along lines of country of origin or heritage, the
difference between Hispanic and Latine, and other issues of importance to Latine communities
(Alcántara et al. 2021).
Identifying these issues of disaggregation and lack of representation in population level
surveys and surveillance systems is only possible because of enumeration efforts in the decennial
census which provide a baseline against which sampling efforts are compared. Importantly, the
limitations of past and present questions on race and ethnicity have been critiqued given the
importance of such data for understanding, and ultimately redressing, structural inequalities
(Strmic-Pawl, Jackson, and Garner 2018). Indeed, the tradition of critical demography has
thoroughly unpacked these issues of representation in data and the logic of identity as static that
underpins quantitative sociological analysis (Horton 1999; Massey 1999; Zuberi 2001; Zuberi
and Bonilla-Silva 2008).
While these issues related to data disaggregation, sampling, and representation of
ethnoracially minoritized groups cannot be conflated with the issues facing gender and sexual
minorities, this tradition, and ongoing efforts to improve data collection, analysis and insight
prove instructive. Therefore, this chapter expands on the small numbers issue by turning to the
US transgender population as a case study. Specifically, this chapter explores the insights that
can be drawn from comparing data sources with differing sampling methods and protocols for
collecting sexual orientation and gender identity (SOGI) data to illustrate how the small numbers
issue affects our understanding of geographic variation, ethnoracial diversity, and estimates of
proportionality of the transgender population in the US. Having established the importance of
spatial analysis to population and health research, and of accurate data on “small” populations

30

like ethnoracially minoritized communities, we move now to transgender communities in the US
as a case study for highlighting both.
3.1.2. Transgender Communities and the Small Numbers Issue
Transgender people, that is those whose current gender identity differs from the sex they were
assigned at birth (Stryker 2008), are said to represent 0.6% of the US population, or roughly 1.4
million Americans (Flores et al. 2016). While this constitutes a relatively small portion of the
general population, transgender communities have unique needs that highlight the limitations of
the legal system such as differential processes for changing legal names and gender markers
federally and in different states (NCTE n.d.), with inequitable practices across states (NCTE
2021), and with yet another set of processes for changes in birth certificates compared to identity
documents (NCTE 2020). Even further, gender-affirming healthcare, which is healthcare related
to gender transition such as the use of GAHT or surgical interventions, has been deeply
politicized. Only 24 states and the District of Columbia have legislation that expressly prohibits
insurance companies from excluding transgender healthcare coverage (Movement Advancement
Project 2022). In 2021, Arkansas passed House Bill 1570, otherwise known as the Save
Adolescents from Experimentation (SAFE) Act or Act 626, which not only banned gender-
affirming healthcare for transgender people under 18, but also banned the use of public funds for
gender-affirming care, prohibited insurance from covering gender-affirming care, and made the
provision of treatment in violation of this ban subject to professional sanctions and the possibility
of legal action.
4
Beyond this legislation, lawmakers in dozens of other states, as of the time of
writing, have put forward similar bills attempting to ban gender-affirming care for trans youth,

4
Advocates for trans youth were able to win a stay on this legislation in Arkansas, and legal
battles to overturn these laws or reduce the harm they cause, especially in Alabama and Florida,
are ongoing

31

trans youth’s participation in sports, and any use of public funds for gender-affirming healthcare.
Despite the relatively small size of the population, trans communities have faced a barrage of
threats to not only their ability to access healthcare, but also their civic participation.
Transgender people face barriers to accessing healthcare beyond even this push to
legislate away their ability to access healthcare free from discrimination on the basis of their
gender identity. A lack of adequate insurance coverage (Stroumsa et al. 2020; Learmonth et al.
2018), lower self-rated mental health and more frequent distress (Crissman et al. 2019), and the
stigma associated with transgender identity (White Hughto, Reisner, and Pachankis 2015) are
among the barriers that drive health disparities between trans communities and the general
cisgender population (J. Feldman et al. 2016). Moreover, when compared to their cisgender
counterparts, transgender people experience higher rates of depression (Sinnard, Raines, and
Budge 2016; Witcomb et al. 2018; Reisner et al. 2016), anxiety (Borgogna et al. 2019; Sinnard,
Raines, and Budge 2016; Dickey, Reisner, and Juntunen 2015), and suicidality (Coulter et al.
2015; Perez-Brumer et al. 2017). Transgender communities have also been found to be more
ethno-racially diverse than the general population (Meyer et al. 2017). However, Indigenous,
Black, and other trans people of color face compounding issues due to systemic racism
(Goldenberg et al. 2019; Crosby, Salazar, and Hill 2016; Lett et al. 2020; Lett, Dowshen, and
Baker 2020; Howard et al. 2019) which remain understudied. Regardless of the size of the
population, these overlapping and complex issues are reason enough to justify structural
interventions to improve the life chances, and health and well-being of trans communities in the
US. Yet, much of the extant literature relies on estimates of the size of transgender populations
that stem from a single source: the Behavioral Risk Factor Surveillance System. The scale of the
problems facing trans communities may in fact be much larger because the calculated size of

32

transgender populations may be underestimated. Therefore, this chapter interrogates the existing
population estimate of 0.6% of Americans identifying as transgender and compares it to more
recent data from a new source: the Household Pulse Survey. In addition, we investigate the
hypothesis that the US transgender population is more ethnoracially diverse than the general
population (K. E. Baker 2019; J. L. Feldman et al. 2021; Meyer et al. 2017). Finally, we analyze
the differences in rates of transgender identification in some of most populous metropolitan areas
of the US compared to the states in which they are located.
3.1.3. Making Transgender Count
The decennial census has yet to ask any questions related to sexual orientation or gender
identity (SOGI), and only instructs enumerators to ask whether a respondent is male or female.
Thus, no attempts at enumerating the entire transgender population have been made in the US. In
contrast, India and Nepal have had third gender categories beyond male and female since their
respective 2011 censuses (Mandal, Debnath, and Sil 2020; Knight, Flores, and Nezhad 2015), the
United Kingdom 2021 census included voluntary questions for those 16 years of age or older
about whether respondents identified with their assigned sex at birth and their sexual orientation
(Barton 2021),
5
and Canada incorporated questions about assigned sex at birth and current
gender identity in its 2021 census (Statistics Canada 2021). Importantly, gathering data on
lesbian, gay, bisexual, and transgender (LGBT) populations is part of the Healthy People 2030
agenda from the US Department of Health and Human Services’ Office of Disease Prevention
and Health Promotion (“Goal: Improve the Health, Safety, and Well-Being of Lesbian, Gay,

5
Notably this census only included England and Wales and therefore was not a full census of the
United Kingdom. Scotland is planning its own census in 2022.

33

Bisexual, and Transgender People.” n.d.). Nonetheless, the current picture of transgender life in
the US is incomplete due to data limitations.
3.2. Data and Methods
Data used in this study are from two sources: the Behavioral Risk Factor Surveillance System
(BRFSS), and the Household Pulse Survey (HPS). The results compare the two in terms of their
respective protocols for gender identity data collection, the overall rates of identification as
transgender, the ethnoracial diversity of trans subsamples compared to their cisgender
counterparts, and the geographic distribution of transgender identification rates. The analysis is
exploratory and descriptive with the goal of highlighting potential differences in these results
based on the chosen data source. A secondary goal is discerning the extent to which each source
aligns with existing demographic estimates of the size and ethnoracial diversity of the
transgender population in the US.
3.2.1. Behavioral Risk Factor Surveillance System
The BRFSS is a national health survey of US adults collected annually by each state and
territory’s public health department. It deploys random digit dialing for both cell phones and
landlines based on probability sampling to recruit participants. In 2014 an optional module for
collecting SOGI data was adopted in some states and has been used by an increasing number of
states and territories every year since. Pooling data across all seven years, from 2014-2020, of
the BRFSS in which the SOGI protocol has been used by some states allows for a total sample of
1,457,371 respondents. This may allow for richer insight than even the most robust of
convenience samples given the breadth of data collected and the means of sampling. Each year
of the sample is weighted separately before being pooled to accurately reflect the sampling
response rates and representativeness of a given year.

34

3.2.2. Household Pulse Survey
The HPS is an online survey studying the impacts of the ongoing coronavirus pandemic on
households nationwide. It is conducted by the US Census Bureau in collaboration with 13 other
federal agencies. The HPS collects information related to childcare, education, employment,
energy use, food security, health, housing, household spending, Child Tax Credit payments, and
attitudes toward vaccination. The Census Bureau randomly selected addresses to invite to
participate and created a generalizable sample using similar weighting procedures to the
American Community Survey and the decennial census. Data collection has been conducted in
week-long phases since April of 2020 and in July 2021 as part of Phase 3.2 a SOGI protocol was
introduced. Phases 3.2 and 3.3 are the only phases thus far to have included a SOGI protocol and
data were pooled across all seven collection waves thus far. Each collection wave is
independently sampled, and each household was interviewed once; thus, pooling across all seven
waves renders a sample of 446,000 independent observations. The weighting procedure for the
HPS data relies on the same weighting procedures that the Census Bureau uses for the decennial
census and its other data collection efforts. All results are reported with person-level weights,
rather than household level weights.
Data from both sources were cleaned, weighted, and analyzed in SPSS 27. Comparisons
across geography were conducted by geocoding both the BRFSS and HPS data at state level, and
again at the metropolitan statistical area (MSA) level in HPS, using ArcGIS Pro 2.9. BRFSS and
HPS data are compared here in terms of state-level proportions of the population identified as
transgender. BRFSS data had more complex categorization for race and ethnicity which were
recoded for direct comparison to the HPS which only reported on five categories: White non-
Hispanic, Black non-Hispanic, Asian non-Hispanic, Other non-Hispanic, and Hispanic.

35

3.2.3. Differential Protocols for Collecting and Ascertaining Gender Identity Data
From 2014 to 2018 the gender identity component of the BRFSS asked respondents only where
they identify as transgender, and if they responded affirmatively, they were asked whether they
identify as male-to-female (e.g., transgender women), female-to-male (e.g., transgender men), or
as gender-nonconforming. In 2019 an additional optional protocol was introduced alongside the
existing SOGI module which asked about assigned sex at birth. This module enabled those
respondents who may not identify with the word transgender but who do have a discordant
current gender identity (GI) and assigned sex at birth (ASAB) to be identified. Six states
(Hawai’i, Louisiana, Minnesota, New York, Utah, and Vermont) used this module in 2019 in
addition to the SOGI module, enabling a comparison of rates of self-identification as transgender
and discordant GI and ASAB, and one state (Pennsylvania) only used the ASAB protocol
without the use of the SOGI module. In 2020, the most recent iteration of the BRFSS, 11 states
(California, Georgia, Hawai’i, Iowa, Louisiana, Minnesota, New Mexico, New York, Ohio,
Utah, and Vermont) used the ASAB protocol in combination with the SOGI module.
Notably, the 2014 and 2015 BRFSS questionnaires explicitly state that interviewers
should only ask the respondent’s sex if necessary. Thus, sex was based on interviewer
assessment of the respondent based only on their speaking voice. In 2016 and 2017, interviewers
were instructed to ask the respondent “Are you male or female?”; however, the sex field was
automatically populated from the enumeration data used to create samples. The questionnaire
for 2018 instructed interviewers to explicitly ask all respondents; however, some states chose to
ask, “Are you male or female?” while other choses to ask, “What was your sex at birth?”.
Finally, in 2019 and 2020, interviewers were instructed to terminate the interview and thank the
respondent for their time if they refused to answer or said they didn’t know when asked “are you

36

male or female?”. Despite the inclusion of a question about ASAB in these years for some states,
all completed responses are coded as either male or female according to this sex question.
Unlike the BRFSS, the HPS SOGI protocol is mandatory and is used across all states and
territories. Additionally, the SOGI protocol is standardized to include both an ASAB and an
explicit trans self-identification question. Their protocol closely follows what is known as the
“two step method,” which has been identified as a gold standard in the existing literature. As part
of the demographics section, respondents are first asked “what sex were you assigned at birth, on
your original birth certificate?” and given the option of either male or female. Then respondents
are asked “do you currently describe yourself as male, female, or transgender?” and given the
options of 1) male 2) female 3) transgender or 4) none of these. For this chapter, only those
respondents who reported a discordant GI and ASAB, and those who explicitly self-identified as
transgender were included in the trans sample. The ‘none of these’ category was included in the
general population, but not as part of the trans subpopulation. This decision was made to respect
that these respondents disidentified with any of the three other available options for gender
identities (male, female, or transgender). However, censoring these data would lead to inflated
rates of identification as trans, so they were included in the general population, but not included
in the transgender subsample. Given that the survey is internet-based rather than conducted by a
live interview like the BRFSS, the survey flags any discordance between ASAB and current GI
and asks respondents to confirm their responses. Additionally, respondents were able to skip this
question if other demographic information was given. Thus, we can infer that respondents’ self-
identification is accurate.

37

3.2.4. Comparison of Weighting Procedures
Generally, the BRFSS uses a two-step weighting procedure to render its sample generalizable to
the whole population. First, the responses are assigned design weights based on geographic
strata, the number of adults in a respondent’s household, and the overall number of records from
which samples are drawn. According to the CDC, these design weights are meant to account for
potential sample overlap for those households that use both landlines and cellphones along with
the variable population density within each state. Then a process called iterative proportional
fitting, or raking, is used to fit the data across eight margins overall: age group by sex, race and
ethnicity, education level, marital status, tenure, gender by race and ethnicity, age by race and
ethnicity, and phone ownership. For some larger geographic regions, data are fitted by a further
four margins: region, region by age, region by gender, and region by race and ethnicity. These
regions are determined on a state-by-state basis according to population density and may vary
from year to year. Finally, if more than 500 responses are collected in each county, then they are
further raked by county, and county by age category (CDC and BRFSS 2015; 2016; 2017). From
2017 onward, this same process is used with an additional iteration within the raking process to
fit the data to population controls, drawn from American Community Survey 5-year population
estimates, according to sixteen margins. The highest priority margin is always sex by age group,
meaning that the data are weighted to most closely resemble the general population according to
sex and age group (CDC and BRFSS 2018; 2019b; 2019a; 2021).
Similarly, the HPS uses the iterative proportional fitting, or raking, procedure to finalize
its weights. However, given that the HPS was designed to measure both personal and household
patterns of issues American face during the COVID-19 pandemic, the process resulted in weights
for both person and household level insights. Given that the HPS is a federally administered
survey, the sampling is done at the national level to gather a nationally representative sample of

38

the US population. For all but the 11 smallest states, sample sizes were limited to a 3%
coefficient of variation for an estimated 40% of the population in all sample areas, and smaller
states were sampled to a 3.5% coefficient of variation. Additionally, sample sizes were adjusted
for an assumed 9% nonresponse rate. After adjusting for nonresponse, weights are adjusted by
the occupied household unit ratio, converted to person level weights based on the number of
adults reported to live in a household, and finally raked. After raking, person-level coverage
ratios are compared by sex (which is defined as sex assigned at birth), age, race, ethnicity, and
education level. And like many population surveys, the crude coverage ratios reveal that
respondents were more likely to be female, 40+ years old, white, non-Hispanic, and college
educated (US Census Bureau 2021a; 2021b; 2021c; 2021d; 2021e; 2021f; 2021g).
3.3. Results
3.3.1. Identifying Trans Subpopulations within the General Population
BRFSS and HPS respondents were labeled as trans if they either self-identified as transgender, or
if they reported a discordant ASAB and current GI. In the pooled, unweighted BRFSS data, 681
trans respondents were identified via discordant ASAB and GI, and a further 6,005 were
identified through self-identification as trans. This made for a weighted sample of 4,780,753
trans respondents to the BRFSS across the pooled data set. When pooled and weighted across all
seven years of data at the national level, the transgender subpopulation makes up 0.55% of the
general population. However, when comparing from year to year, there is a general linear trend
upward which is consistent across the crude and weighted data (Figure 3.1).

39

Figure 3.1. Unweighted versus weighted rates of transgender identification from year to year in
BRFSS data (2014-2020)

In the full, unweighted sample of the Household Pulse Survey, 1,323 respondents self-
identified as transgender, an additional 1,088 respondents reported an ASAB that was discordant
with their current GI, and a further 4,757 respondents identified as “none of these,” which
entailed explicitly not identifying as transgender, but also not identifying with their ASAB. The
‘none of these’ category is included in the general population but is not included in the trans
subpopulation. This makes for a raw sample of 2,411 trans respondents which constituted 0.54%
of the total sample. When weighted, the trans subsample increases to 16,330,852, or 0.95% of
the total sample (Table 3.1). The rates of cisgender, transgender, and none of these can be found
in Appendix A. Overall, the HPS data resulted in a higher proportion of trans respondents, even
when compared just to the BRFSS 2020 data which had the highest rates of identification as
trans.

0.00%
0.10%
0.20%
0.30%
0.40%
0.50%
0.60%
0.70%
2014 2015 2016 2017 2018 2019 2020
Unweighted % Weighted %

40

Table 3.1. Comparison of unweighted and weighted rates of identification as trans across the
BRFSS 2014-2020 and HPS pooled data
3.3.2. Geographic Representativeness
The SOGI module in the BRFSS is optional and a total of 43 states have used it at least once
between 2014 and 2020. The seven states which never utilized the SOGI module are Alabama,
Maine, Nebraska, New Hampshire, North Dakota, Oregon, and South Dakota; the District of
Columbia has also never utilized the SOGI module. Only five states, Minnesota, New York,
Ohio, Virginia, and Wisconsin, used the optional SOGI module all seven years since its
introduction in 2014. Figure 3.2 depicts this pattern of cumulative years of SOGI data collection
in the BRFSS below.
From the available data, rates of transgender identification vary from state to state. They
range from as little as 0.26% in Wyoming to as much as 1.1% in Utah with most states ranging
from 0.26% to 0.75%. The highest rates are in Louisiana (0.94%), North Carolina (0.87%), and
Utah. These states stand out given that they are bordered by states with much lower rates of trans
identification (Figure 3.3).
In contrast, the HPS rates of identification as trans are higher overall. They range from
0.42% in Nebraska to 1.49% in Texas. Only six states, including Nebraska, have rates less than
0.75%, which is the opposite of BRFSS where most states are below that rate. Five states stand
out as having the highest rates, all of which exceed 1.25%. These five are Tennessee and
Wyoming, both at 1.31%, West Virginia (1.36%), Utah (1.42%), and Texas (1.49%). Notably,
Unweighted Weighted
Source % Trans % Trans
BRFSS 0.464 0.551
HPS 0.543 0.949

41

Figure 3.2. Number of years the optional BRFSS SOGI module was used by each state. Those
states with cross-hatching have no data because they have never used the SOGI module.

each of these five are in one of two census regions: the South, and the West. Additionally, each
of these states, except Texas, border one of the states with the lowest rates (Figure 3.4).
Because the BRFSS and the HPS both contain state level data with transgender
subsamples, state-level rates of identification as transgender can be mapped. Overall, the BRFSS
shows lower rates for most states with one notable exception. North Carolina had a higher rate of
trans identification in the BRFSS (0.87%) than in the HPS (0.69%). The percent change between
the HPS and BRFSS rates were taken and then mapped to demonstrate this. The BRFSS were
used as original rates and HPS used as new rates; therefore, positive percent change indicates an

42

Figure 3.3. State levels rates of transgender identification pooled across all seven waves of data
collection for the BRFSS

increase in HPS rates of trans identification from BRFSS and negative percent change indicates a
decrease (Figure 3.5).
Finally, the HPS included not only the state of residence for respondents, but also
identified which respondents lived in the 15 largest Metropolitan Statistical Areas (MSAs). Of
these 15 MSAs, 10 are located entirely within a single state and they are Atlanta-Sandy Springs-
Roswell (Georgia), Dallas-Fort Worth-Arlington (Texas), Detroit-Warren-Dearborn (Michigan),
Houston-Baytown-Sugar Land (Texas), Los Angeles-Long Beach-Anaheim (California), Miami-
Fort Lauderdale-West Palm Beach (Florida), Phoenix-Mesa-Scottsdale (Arizona),

43

Figure 3.4. Rates of identification as trans from the HPS pooled at state level

Riverside-San Bernardino-Ontario (California), San Francisco-Oakland-Fremont (California),
and Seattle-Tacoma-Bellevue (Washington). Table 3.2 depicts the rates of trans identification at
the state-level inclusive of the MSAs, in each MSA, and at state level without the MSAs.
Arizona, California, Florida, and Michigan’s rate increases when the MSAs are censored out
whereas Texas and Washington’s rates drop when MSAs are censored. Georgia remains the most
consistent whether the Atlanta MSA is included or not. Of the MSAs, Seattle has by far the
highest rate of trans identification at 1.69%, and the lowest is found in the Miami area at 0.77%.

44

Figure 3.5. Percent change from BRFSS to HPS in rates of identification as transgender by state

Table 3.2. Comparison of transgender identification rates in select states and metropolitan areas
in HPS pooled data
State % Trans MSA % Trans
% Trans
without MSA(s)
Arizona 0.895 Phoenix 1.017 1.080
California 0.992
Los Angeles 0.862
1.212 Riverside 0.791
San Francisco 0.785
Florida 0.904 Miami 0.772 0.949
Georgia 1.314 Atlanta 1.298 1.312
Michigan 1.062 Detroit 0.882 1.2
Texas 0.915
Dallas 1.117
0.715
Houston 1.024
Washington 1.212 Seattle 1.691 0.606

45

3.3.3. Ethno-Racial Diversity within Trans Population
Results from the pooled samples of both the BRFSS and HPS trans samples align with existing
literature on ethno-racial diversity in trans populations. The BRFSS contains more complex,
disaggregated data on race and ethnicity, but results here were recoded to match the race and
ethnicity data from the HPS for comparison. Table 3.3 depicts the ethnoracial breakdown of the
general population compared to the trans population in the BRFSS data. Table 3.4 depicts the
same with the HPS data. Across both data sets, the percentage of the trans population that
identifies as White non-Hispanic is much lower than the percentage of the general population
that identifies as White non-Hispanic. Whether comparing the weighted or unweighted data,
trans populations are less likely to identify as White non-Hispanic and are more likely to identify
as Hispanic or as a race other than White, Black, or Asian, with one notable exception that, upon
weighting, the proportion of the trans population that identifies as Black non-Hispanic and Asian
non-Hispanic was lower than in the general population in the HPS data.

Table 3.3. Comparison of ethnoracial breakdown of general population versus trans
subpopulation in the BRFSS
Unweighted Weighted
Race
% General
Population
Trans
Total
% Trans
Population
% General
Population
Trans
Total
% Trans
Population
White non-
Hispanic
75.93 4,316 64.55 62.81 2,590,873 54.19
Black non-
Hispanic
7.61 613 9.17 11.72 615,270 12.87
Asian non-
Hispanic
2.83 272 4.07 5.08 292,2478 6.12
Other non-
Hispanic
6.63 747 11.17 4.61 326,713 6.83
Hispanic 7.0 738 11.04 15.78 955,420 19.99

46

Table 3.4. Comparison of ethnoracial breakdown of general population versus trans
subpopulation in the HPS
Unweighted Weighted
Race
% General
Population
Trans
Total
% Trans
Population
% General
Population
Trans
Total
% Trans
Population
White non-
Hispanic
75.93 4,316 64.55 62.43 8,424,081 51.58
Black non-
Hispanic
7.61 613 9.17 11.26 1,415,773 8.67
Asian non-
Hispanic
2.83 272 4.07 5.42 361,565 2.21
Other non-
Hispanic
6.63 747 11.17 3.72 1,233,490 7.55
Hispanic 7.0 738 11.04 17.18 4,895,944 29.98

3.4. Discussion
Both state-level and national trans population estimates vary significantly across the BRFSS and
HPS data. The main distinctions between the two data sources are their sampling method, survey
delivery means, and notably, their protocols for collecting gender identity data. The BRFSS data
does show a linear trend upward from year to year for both unweighted and weighted data, which
implies that the rate of identification as trans is increasing over time. In fact, the 2020 BRFSS
national level unweighted rate of trans identification (0.54%) is the same as the overall rate of
identification as trans in the 2021 HPS data. Yet, the weighting procedure for HPS raises that
overall rate to 0.95% when pooled across all states and all seven data waves. This difference in
weighting is not simply due to procedures given that both data sets use raking to finalize their
weights and use ACS 5-year estimates to generate initial design weights. Thus, the difference
may be attributable to the difference in survey delivery and design.
Principally, the decision to make SOGI modules optional in the BRFSS may have led to
undercounting trans respondents. The integration of SOGI protocols not as a module, but as a
mandatory demographic question at the outset of the survey seems to have led to a dramatic

47

increase in rates of trans identification in the HPS. At the same time, given how small the
numbers are for trans subsamples in both data sets, the population level estimates drawn from
either are likely unstable and unreliable, especially when disaggregated, whether by race and
ethnicity, or by geography. This inherently limits the insights that can be drawn from either
dataset as to the intersectional experiences or intracommunity differences within US trans
populations. Moreover, both data sets rely on sex for sampling, stratification, and weighting.
While the HPS asks about sex assigned at birth, it is coded as sex for the purposes of weighing
samples and the same is done for BRFSS even in those states which include modules on SOGI
and the ASAB question. While some demographic work on trans populations argues that trans
people should be compared to their ASAB (e.g. transgender women compared with cisgender
men because both were assigned male at birth) (Lagos 2018), it is unlikely that trans respondents
would find it acceptable to report their trans status if they knew they would be counted with their
ASAB instead of aggregated with their current GI. Ultimately, neither data set was designed to
be able to draw these kinds of conclusions; however, the BRFSS has been used as the primary
source for the most common estimates of the size of the trans population in the US (Flores,
Brown, and Herman 2016; Flores et al. 2016; Herman et al. 2017). More recently, the HPS has
begun to be taken up in the same way (Conron and O’Neill 2021).
Another key difference is that the HPS included the option to self-identify as “none of
these” as opposed to male, female, or transgender. This data deserves further exploration as the
rates of identification as “none of these” far exceed both the combined, computed rate of
identification as trans, and the separate rates of discordant ASAB/GI and explicit self-
identification as trans. Given that the survey required respondents to verify their responses to
both ASAB and GI when they did not match, it can be assumed that anyone identifying with

48

“none of these” expressly did not wish to be counted as transgender, female, or male. While this
could be construed as an endorsement of non-binary identity, it would be presumptuous to
ascribe that category to these respondents as it would be to label them as transgender or
cisgender. Given that there are no population level studies that consider potential differences
between the transgender and non-binary populations, it might be fruitful to further delve into the
“none of these” subpopulation. While it is beyond the scope of this chapter to incorporate this
category, it merits further research.
Prior studies using BRFSS data have established the limitations of the protocols used to
ascertain transgender subsamples (Lagos 2018; Lett and Everhart 2021; Tordoff, Andrasik, and
Hajat 2019). Indeed, the efficacy of different protocols for ascertaining gender identity have been
well studied (Bauer et al. 2017; Cahill et al. 2014; Lombardi and Banik 2016; Reisner et al.
2015; Tate, Ledbetter, and Youssef 2013; Tordoff et al. 2019). Additionally, both surveys rely on
household sampling meaning anyone without a permanent residence, such as people
experiencing homelessness, are not counted. The BRFSS and HPS also only interview adults,
meaning youth are not counted in the SOGI modules each survey employs. Given rates of
homelessness among transgender people in the US (Homelessness Research Institute 2020),
which may be even higher among youth due to family rejection (Keuroghlian, Shtasel, and
Bassuk 2014), it is plausible that our estimates undercount transgender people due to the
sampling methods in each survey. While both data sets have pros and cons depending on what
researchers, or lay people, wish to investigate, both have serious limitations to generalizability.
This is due in large part to the fact that the enumeration process does not include any protocols
for ascertaining gender identity either as separate from or as part of sex ascertainment. Without
enumerated data, we have no baseline against which to compare a sample’s representativeness or

49

the generalizability of findings. Given that trans populations have been shown to differ from the
general population demographically, it is difficult or perhaps even misguided to offer
generalizable findings from population surveys, especially those that do not deploy validated and
accepted measures of gender identity. To return to the data sources upon which this analysis
relies, it is unnecessary, and would perhaps be misleading, to holistically rate one data set as
better, or more accurate than the other. However, the notable differences in data collection
methods clearly produce empirically different results regarding the size of the trans population in
the US.
This chapter highlights a significant data gap. Namely, trans communities may have
different demographic patterns from the general population, and that little is known about how
these patterns vary geographically. The significance of this gap in data cannot be overstated.
Indeed, the existing literature that establishes demographic and epidemiologic patterns among
trans communities relies upon the population estimates from the BRFSS that this chapter
interrogates. While the challenges that trans communities across the US face would be
significant regardless of the size of the population they impact, we cannot adequately determine
how many people are affected and therefore how best to respond without sound and reliable data
on the size and demographics of trans communities.
Ultimately, transgender communities in the US, and indeed globally, face unique
challenges, especially when navigating the health and legal systems. These challenges have been
the subject of much of the scientific literature on transgender life. However, this chapter has
highlighted the need for a focus on the size, demographics, and geographic scale and spread of
trans populations in order to contextualize these challenges. Research that illustrates challenges
faced cannot be the driving force behind the creation of targeted interventions or service

50

planning. Scholarship in transgender population health has taken for granted that we do not yet
know enough of the basics, i.e., the demographics, upon which all insights into population health
are built. Future research should therefore focus on enumeration as well as investigation of the
spatial patterns that shape transgender life in the US. The goals that the federal government has
outlined for incorporation of SOGI protocols into federal data collection by 2030 may come too
late if bills continue to pass into law that attempt to legislate trans people out of existence.
Therefore, we call for increased attention to the specific needs of trans communities in the hopes
of addressing their unique needs, which will necessarily lead to insight into how overlapping
health, legal, and other social systems can be improved to better address the needs of all.
While the challenges facing the trans community in the clinics and courts may be unique,
the larger issue of small numbers and reliable demographic data is common among many
minoritized communities. This chapter drew from the literature on data gaps affecting Native
Hawai’ian and Pacific Islander communities, as well as the disaggregation of demographic data
within the larger category defined as Hispanic in the US. This small numbers issue is ubiquitous
and could be addressed with innovative methods like spatial microsimulation or Bayesian
statistics to create synthetic data and model potential population change. On the surface it seems
that the baseline data against which these models could be tested exists and is reliable for Native
Hawai’ian/Pacific Islander and Latinx communities; however, the research conveyed in this
study suggests that perhaps even those baselines are not as reliable as they may seem. Better and
more inclusive enumeration procedures are needed if we are to be able to respond to the ever-
changing dynamics and demographics in the US population. While protocols for measuring race
and ethnicity are beyond the specific scope of this study, they should necessarily be investigated
as part and parcel of the kinds of data reform that we call for in this chapter. By adopting an

51

intersectional lens, demographers and population researchers from all disciplines can begin to
make the connections between SOGI, race/ethnicity and spatial data and push for better
procedures and protocols that will making all minoritized communities’ count.
3.5. Conclusion
In this chapter, we compared two sources, the BRFSS and HPS, and their respective rates of
trans identification, both overall and at the state level. While the BRFSS showed a linear trend
upward in trans identification rates from 2014 to 2020, the HPS rates were notably higher overall
and at state level for all but one state. Both sources illustrated that the US trans population is
more ethnoracially diverse than the general population, which has been suggested in the previous
literature. Interestingly, the HPS data showed some noticeable differences in trans identification
rates when comparing state level to metropolitan level data that suggests urbanicity among trans
populations may vary. More research on urbanicity among trans communities is needed to
understand this particular demographic aspect of the US trans population. Overall, while further
research is needed on trans communities in general, we argue here that better practices for data
collection must be adopted to produce truly meaningful data, especially the inclusion of SOGI
protocols in more federal level surveys and the decennial census enumeration. We suggest that
attention paid to trans communities in the US can reveal how the small numbers issue affects
other minoritized communities. Ultimately, we conclude that the adoption of an intersectional
lens enables demographers and population researchers to not only identify but come up with
solutions for the small numbers issues facing trans communities and thereby better address the
needs of all minoritized communities.

52

Chapter 4 Measuring Geographic Access to Transgender Hormone Therapy in Texas:
A Three-step Floating Catchment Area Analysis
To date no studies have analyzed access to transgender-specific medical care with spatially
explicit methods. In this chapter, a three-step floating catchment area analysis is used with Texas
as a case study to illustrate the health services planning insights that can emerge from spatial
analyses. Additionally, patterns of access to gender-affirming hormone therapy are compared to
patterns of access to primary care and stratified by urbanicity.
4.1. Introduction
Estimates of the size of the transgender population in the US range from just under 1 million
adults (Meerwijk and Sevelius 2017) to as many as 1.4 million adults (Flores et al. 2016).
Despite the relatively small size of the transgender population, they face significant barriers to
accessing healthcare (Gonzales and Henning‐Smith 2017), including primary care (Edmiston et
al. 2016; Vermeir, Jackson, and Marshall 2018; Nisly et al. 2018; Kattari et al. 2021) as well as
gender-affirming care (Puckett et al. 2018; White Hughto et al. 2017; Warner and Mehta 2021).
Additionally, the complicated health system and different protocols for insurance coverage have
limited the availability of and access to care for trans communities (Bakko and Kattari 2021;
2020; Learmonth et al. 2018; Stroumsa et al. 2020). Each of these issues is exacerbated by
exposure to systemic racism for trans people of color (Lett et al. 2020; Lett, Dowshen, and Baker
2020; Lett et al. 2022; Goldenberg et al. 2019; Bukowski et al. 2018). The confluence of factors
shaping the health services landscape for trans people has had the downstream effects of
increased rates of poor health outcomes in trans communities (Cicero et al. 2020). While
research on barriers to access, stigma, and poor health outcomes among trans people proliferates,
less attention has been paid to how access to trans-specific healthcare is determined

53

geographically. This chapter addresses that gap by offering a spatially explicit analysis of access
to transgender healthcare.
Public health has begun to recognize the utility of geographic information science (GIS),
spatial analysis, and geographic thinking (Wang 2020). Indeed, the spatial turn in health research
has fomented interest in investigating how space, place and time not only factor into
understanding patterns of health outcomes, but shape population health in many ways
(Richardson et al. 2013). A series of methods known as the floating catchment area (FCA)
approach have been developed specifically to quantify spatial access to healthcare services. The
foundational method evolved out of gravity-based models and introduces a second step
deploying GIS to measure the ratio of physicians to population (Luo and Wang 2003; Luo 2004).
This method, known as the two-step floating catchment area (2SFCA), has since been elaborated
upon to account for distance decay (Luo and Qi 2009; McGrail 2012), incorporate Kernel
Density estimation (Dai and Wang 2011), introduce network and raster data as alternatives to
vectors representing civic and administrative geographic boundaries (Delamater et al. 2012), and
adapted to account for regions with established suboptimal availability of health services
(Delamater 2013). Finally, the method was expanded to incorporate a third step that calculates
selection weights between every possible combination of population origin and healthcare
facility destination, which became the three-step floating catchment area method (3SFCA) (Wan,
Zou, and Sternberg 2012).
Thus far, the FCA methods series has primarily been used to measure access to primary
care, as is the case in most examples cited above. However, it can also be applied to other kinds
of care to better understand the health services landscape from a spatially explicit perspective. To
that end, this study analyzes the spatial accessibility of one specific branch of transgender

54

medical care: gender-affirming hormone therapy (GAHT). More specifically, we begin to
identify potentially underserved areas and whether they align with known underserved areas for
other kinds of care or may be associated with urbanicity. Given that rural areas face difficulties
in retaining primary care and other physicians (McGrail et al. 2017), and that physician supply is
associated with both worse health outcomes (Macinko, Starfield, and Shi 2007), and higher
mortality rates than their urban counterparts (Basu et al. 2019), we begin to explore whether
access to trans healthcare follows similar patterns. Direct hypothesis testing for association
between urbanicity or rurality and access to GAHT is beyond the scope of this current study with
its focus on applying the gold standard method for measuring spatial access to care; however, an
exploratory descriptive analysis is included to discern what kind of further research is needed.
GAHT can be defined as the use of exogenous hormones for the purposes of gender
transition. Notably, GAHT provision is conducted across medical specialties, including primary
care. Thus, tracking its provision to specific healthcare providers and healthcare facilities is
difficult using publicly available data. Moreover, as alluded to above, insurance coverage for
GAHT varies based on factors such as insurance provider protocols, state-level policies, and
even the specific diagnostic and billing codes used. While these particular factors are beyond the
scope of this analysis, they are worth future investigation. In fact, it is likely that the complex
nature of tracking GAHT provision through a geographic lens has contributed to the persistence
of our knowledge gap in spatial access to trans-specific healthcare. In the remainder of the
chapter, we outline the data sources we use for both transgender population estimates, and the
location of GAHT providers. We then introduce our case study, Texas, as well as elaborate on
the 3SFCA method we use to measure access within the state. Finally, we share results and

55

discuss implications both from the perspective of health services planning and through the lens
of transgender population health in the US.
4.2. Data Sources
4.2.1. Demographic Data
In order to derive population estimates for the trans population, we relied upon the Household
Pulse Survey (HPS). The HPS was introduced in 2020 in response to the COVID-19 pandemic
by the Census Bureau in collaboration with eleven other federal agencies. It is a randomly
sampled, internet-based survey used to discern household level needs in the US during the
ongoing pandemic including issues related to housing, employment, childcare, and others. The
data are reported by state and even further by metropolitan statistical area (MSA), but only for
the 15 largest MSAs. In Phase 3.2 in Summer of 2021 a mandatory protocol for measuring
sexual orientation and gender identity (SOGI) was introduced to the survey. Since then, there
have been seven waves of data collection, each lasting a week about every six weeks. To derive
estimates of the size and spatial distribution of the trans population in the US, we pooled data
across these seven waves. For this study, we focus on Texas and calculated a rate of
identification as trans for the Dallas-Fort Worth-Arlington MSA, the Houston-The Woodlands-
Sugar Land MSA, and the rest of the state of Texas (with both MSAs censored). We then pulled
Census 2020 data for adult population counts at the census tract level for all of Texas and
calculated estimates of trans population rates using the respective rates derived from the HPS.
Texas was chosen because it is the 2
nd
largest state in terms of land mass while also being
the 2
nd
most populous. Additionally, it has a noteworthy mix of rural and urban areas with the
majority of urban counties lying in the eastern part of the state and most rural counties in the
west. Texas also has Medically Underserved Areas, areas with a measurable shortage of primary

56

care which we discuss in the next subsection, in all but one of its 254 counties according to the
Health Resources & Services Administration. Finally, transgender communities have been the
subject of ongoing political and legal controversies in the state, with the current governor and
attorney general directing Child Protective Services to investigate the parents of transgender
youth for child abuse for allowing their children to access GAHT or puberty blockers. Our focus
is only on adults, in order to limit, as much as possible, potential confounding from the existing
limitations to the provision of GAHT that youth face currently. However, while the method we
employ cannot measure the sociopolitical or legal drivers of access to care directly, it is likely
that these factors influence the availability and accessibility of transgender-specific healthcare in
the region. Therefore, Texas is an instructive case study for deploying our spatial method for
measuring access to care and will be helpful for understanding what further factors driving
access to care should be investigated both in Texas and elsewhere.
To complement the population estimates we used a spatial database of gender-affirming
hormone therapy (GAHT) providers of the first author’s creation. The entire data set was used to
account for the possibility of GAHT providers within the drive-time cutoff, but outside of the
state of Texas. Finally, to calculate travel times, we used Esri’s StreetMap Premium Road
Network data through ArcGIS Pro.
4.2.2. Care Shortages, Rurality & Urbanicity
A “care shortage” has been defined for primary care as a physician to population ratio of less
than 1:3500 (Ricketts and Holmes 2007), and this ratio was used in the development of the
3SFCA method (Wan, Zou, and Sternberg 2012). Additionally, the Health Resources & Services
Administration (HRSA) designates certain geographic areas as Medically Underserved
Populations (MUAs) based on reports from state-level Primary Care Officers and a combination

57

of demographic, geographic, and health data (HRSA 2021). However, both the ratio used in the
3SFCA method development and the MUA designation from HRSA measure access to primary
care, specifically. To account for this, we use a designation from the Association of American
Medical Colleges (AAMC) and their report on physician specialties. The biannual special reports
indicate the physician to population ratio by specialty, which are calculated using census data
and the American Medical Association’s (AMA) physician master file. Their most recent report
in 2020 used 2019 data from the AMA physician Masterfile and 2020 census data. The most
closely related specialty to GAHT provision for transgender people is endocrinology and
according to the AAMC 2020 report there were 41,460 people for every one endocrinologist
currently active in the US (AAMC 2020). We use this physician to population ratio for a
secondary measure of low spatial access to GAHT to complement the primary care figure from
the original 3SFCA method.
To illustrate the potential relationship between MUAs and spatial access to GAHT in
Texas, we used bivariate mapping. We pulled components of these MUAs from HRSA’s data
portal and included those which had not been petitioned to be removed. The final count included
811 designated MUAs, a combination of entire counties and clusters of neighboring census tracts
within areas not otherwise considered underserved. Of the 6,422 census tracts in the case study
area, 2,596 tracts (40.42%) fell within one of the 811 MUAs. We labeled all the tracts with an
MUA as a dichotomous variable based on this and used it to illustrate the patterns in our results.
We also conducted a descriptive analysis of the correlation between urbanicity and spatial
access to GAHT. To classify census tracts within the study area by urbanicity we used the
county-level designations on the rural-urban continuum code developed by the US Department
of Agriculture (USDA 2020). These codes are recalculated every ten years, and the most recent

58

was from 2013. These codes range from one, the most urbanized metropolitan areas, to nine, the
most rural. We classified those counties scoring 1-3 as urban, 4-7 as nonmetropolitan, and 8-9 as
rural. We used bivariate mapping again to illustrate the potential relationship between urbanicity
and spatial GAHT access within Texas.
4.3. Methods
We used the three-step floating catchment area (3SFCA) method to quantify the spatial
accessibility to GAHT providers for population centers (Wan, Zou, and Sternberg 2012). To
accomplish this, we used the Origin-Destination (OD) Matrix function of Network Analysis in
ArcGIS Pro to calculate every possible combination of census tract and GAHT provider within
the drive-time window of 120 minutes from the census tract origin. Each census tract was
converted to a simple centroid, rather than a population-weighted centroid because the estimated
trans population per census tract was too low to meaningfully derive a population-weighted
centroid. Every GAHT provider location was already represented as a point. Upon creating this
OD Matrix, we followed the three steps as outlined by Wan et al. They are reproduced with our
own explanation below.

Step 1:
𝐺 𝑖𝑗
=
𝑇 𝑖𝑗
∑ 𝑇 𝑖𝑘 𝑘 ∈{𝐷𝑖𝑠𝑡 (𝑖 ,𝑘 )<𝑑 0
}

This step calculates selection weights, 𝐺 𝑖𝑗
, for each combination of population center, i, and
healthcare facility (HCF), j. The individual weight, 𝑇 𝑖𝑗
, is assigned based on the sub-zone within
the cutoff distance, 𝑑 0
, according to the Gaussian distribution determined by the impedance

59

coefficient. The denominator represents the sum of all individual weights, 𝑇 𝑖𝑘
, for any HCF, k,
that is within the overall cutoff distance from the population center. Thus, if only one HCF is
within the drive-time cutoff, which is 120 minutes in our study, then the selection weight is 1.
However, the selection weight will decrease as the number of HCFs within that drive-time
window increases. The sub-zones divide into 15 minutes, 30 minutes, 60 minutes, and cut off at
120 minutes.

Step 2:
𝑅 𝑗 =
𝑆 𝑗 ∑ ∑ 𝐺 𝑘𝑗
𝑃 𝑘 𝑊 𝑟 𝑘 ∈𝐷 𝑟 𝑟 =1,2,3,4

=
𝑆 𝑗 ∑ 𝐺 𝑘𝑗
𝑃 𝑘 𝑊 1 𝑘 ∈𝐷 1
+ ∑ 𝐺 𝑘𝑗
𝑃 𝑘 𝑊 2 𝑘 ∈𝐷 2
+ ∑ 𝐺 𝑘𝑗
𝑃 𝑘 𝑊 3 𝑘 ∈𝐷 3
+ ∑ 𝐺 𝑘𝑗
𝑃 𝑘 𝑊 4 𝑘 ∈𝐷 4

In step 2, a physician-population ratio, 𝑅 𝑗 , is calculated for each HCF, j. The numerator
represents the medical capacity, 𝑆 𝑗 , or number of GAHT providers at that service location. The
individual selection weight, 𝐺 𝑘𝑗
, is taken from step 1 for each population center and HCF
combination. The total population of a given population center, k, is represented by 𝑃 𝑘 , and 𝑊 𝑟 ,
the Gaussian weight for the respective sub-zone, r, within the catchment in which the HCF is
located. Thus, the product of each selection weight, target population, and sub-zone weight is
calculated for each population center within the 60-minute drive time window of the facility is
taken for each sub-zone within the catchment. The denominator is therefore the sum of these
products for each of the four sub-zones within the catchment. The quotient of the medical
capacity and this sum of products represents the physician-population ratio for each HCF.

60

Step 3:
𝐴 𝑖 𝐹 = ∑ ∑ 𝐺 𝑖𝑗
𝑅 𝑗 𝑊 𝑟 𝑗 ∈𝐷 𝑟 𝑟 =1,2,3,4

= ∑ 𝐺 𝑖𝑗
𝑅 𝑗 𝑊 1
𝑗 ∈𝐷 1
+ ∑ 𝐺 𝑖𝑗
𝑅 𝑗 𝑊 2
𝑗 ∈𝐷 2
+ ∑ 𝐺 𝑖𝑗
𝑅 𝑗 𝑊 3
𝑗 ∈𝐷 3
+ ∑ 𝐺 𝑖𝑗
𝑅 𝑗 𝑊 4
𝑗 ∈𝐷 4

Step 3 results in the Spatial Accessibility Index (SPAI), 𝐴 𝑖 𝐹 , for each population center. First,
each selection weight, 𝐺 𝑖𝑗
, is multiplied by the physician-population ratio, 𝑅 𝑗 , and the assigned
Gaussian weight for each sub-zone. Then the sum of all products in each respective sub-zone is
calculated and finally added together to compute the overall SPAI for the population center.
Finally, the average SPAI is taken for all population centers and each population’s SPAI is
divided by the average to calculate the Spatial Accessibility Ratio (SPAR) for each population
center. The SPAR has been shown to be robust against possible skewing introduced by variation
in the impedance coefficient chosen and thus it is used here as well. To look at patterns of the
distribution of spatial access, we also conducted a hot spot analysis using Getis-Ord Gi*
statistic., which is a measure of spatial autocorrelation that identifies clusters of statistically
significantly related values of a given variable based on proximity to other areas with similar
values.
4.4. Results
We used both the SPAI and SPAR measures to analyze the distribution of spatial access to
GAHT across Texas. Table 4.1 shows the distribution of census tracts that fell into shortage
areas according to the endocrinology physician to population ratio threshold, shortage areas
according to the primary care threshold, and those tracts with the best spatial access. In Figure

61

4.1 below, the worst shortage areas are the lightest gray, the shortage areas only according to the
primary care threshold are the next darkest, and the darkest shade represented those tracts with
the best access; censored areas due to trans population estimates of less than 1 are in white.
These care shortage areas also include tracts where the SPAI was zero, a number that reflects
that no GAHT providers were located within the 120 minutes cutoff used in the 3SFCA method.

Table 4.1 Number and percentage of census tracts in study area by geographic area stratified by
SPAI for GAHT access
Geography No. Tracts with SPAI
< Endocrinology
Ratio
No. Tracts with SPAI
< Primary Care Ratio
No. Tracts with SPAI
> Primary Care Ratio
Dallas-Fort Worth-
Arlington MSA
184 (10.75%) 1,110 (64.87%) 417 (24.38%)
Houston-The
Woodlands-Sugar
Land MSA
182 (11.55%) 890 (56.51%) 503 (31.94%)
Texas without Dallas
or Houston
429 (13.68%) 1,702 (54.27%) 1,005 (32.05%)
Texas Overall 795 (12.38%) 3,702 (57.64%) 1,925 (29.98%)

While initially Texas appears to have a large swath of shortage areas defined by the SPAI
metric, when normalized by the average SPAI to produce the SPAR (Wan, Zou, and Sternberg
2012), there are zero shortage areas. The SPAR values range from 0.00111 up to 146.4; however,
the mean is 1.0, the median value is 3.7 and the standard deviation is 3.7. The Kurtosis for the
distribution is quite high at 967.4 indicating some extreme values and right-tailed skewness.
Only four census tracts break 100 in their SPAR measures, and all but three of those are in the
Houston-The Woodlands-Sugar Land MSA. When the 24 tracts whose SPAR value is greater
than 12.1 (e.g., 3 standard deviations above the mean), the mean drops to 0.85, the median
decreases to 0.47, and the standard deviation decreases to 1.0. The removal of these 24 tracts also

62

lowers the Kurtosis to 14.2. Figure 4.2 shows the distribution of SPAR values after the removal
of the extreme values divided into three quartiles with the lightest shades showing lowest spatial
access and the darkest showing the highest spatial access to GAHT.
Many of the tracts with the highest spatial access to GAHT lie outside of the two largest
metropolitan areas, Dallas and Houston. Those tracts within the metropolitan areas tend to have
higher population density overall which extends to the estimated trans population as well.
Suburban, or perhaps even rural trans people may have similar access to urban communities of
trans people if they live in areas where multiple providers are available within the 120-minute
drive-time range. We investigate this further with a descriptive analysis of urbanicity and GAHT
access (Figure 4.3).
High spatial access to GAHT was defined here as those tracts with an SPAI measure
above the 1/3500 ratio for primary care used in the original 3SFCA method. The threshold we
chose for low urbanicity was only those counties considered to be rural, meaning those that were
either completely rural or had less 2,500 residents if adjacent to a metropolitan area. Given this
threshold for rurality versus urbanicity, there is a large amount of variation in access to GAHT
across Texas’s 191 rural counties. Additionally, there is clear variation in access to GAHT even
within major metropolitan and other highly urbanized areas.

63

Figure 4.1. Spatial Accessibility Index (SPAI) values for GAHT access across all census tracts in Texas. Grey represents care
shortages defined as SPAI values less than 1/3500 and whited out tracts are censored.

64

Figure 4.2. Spatial Accessibility Ratio (SPAR) values for access to GAHT across all Texas census tracts except for the extreme
values. The classes are divided into three quantiles with the lightest shade representing the quantile with the lowest access and the
darkest representing the quantile with the highest. Those areas in white are censored.

65

Figure 4.3 Bivariate map of access to GAHT by urbanicity. The same threshold for high GAHT access as in Figure 4.1 is used and
low urbanicity is defined as only those areas classified as rural (e.g., those with rural-urban continuum codes of 8 or higher).

66

To understand how this may align with established MUAs, we repeated this bivariate
mapping schema with MUAs (Figure 4.4). Like the map comparing GAHT access by urbanicity,
Figure 4.4 shows observable heterogeneity in access to GAHT based on established MUAs for
primary care. There is significant variation within both the Dallas and Houston MSAs and across
the many rural counties in central Texas.
In light of the results of our descriptive analysis of the distribution of MUAs and areas
with high and low access to GAHT, we chose to conduct a hot-spot analysis using Getis-Ord Gi*
to account for clustering and identify potential areas in need of more providers (Figure 4.5). The
Houston metropolitan area had hot spots of high access to GAHT with no cold spots. In contrast,
the Dallas metropolitan area had hot spots to the west in Weatherford, but cold spots in the Fort
Worth and Lewisville areas to the west of Dallas and in eastern Dallas. The larger Austin area
also had significant cold spots while San Antonio and the surrounding area to the south had
significant hot spots of high access to GAHT.
There were 42 counties overall that had statistically significant cold spots for GAHT
access. Fourteen of the most metropolitan counties in Texas had statistically significant cold
spots: Caldwell, Comal, Dallas, Denton, Galveston, Guadalupe, Hays, Hunt, Johnson, Kaufman,
Rockwall, Tarrant, Travis, and Williamson. Five of the second most metropolitan counties
(Cameron, Lynn, McLennan, Potter, & Randall), and nine of the third most metropolitan
counties (Archer, Clay, Ector, Gregg, Midland, Rusk, Smith, Tom, Green, & Wichita) also had
statistically significant cold spots. Twelve nonmetropolitan counties had statistically significant
cold spots: Andrews, Burnet, Castro, Cherokee, Fannin, Fayette, Henderson, Hill, Lamar,
Swisher, Van Zandt, and Willacy. Finally, only two completely rural counties have statistically
significant cold spots: Blanco and Schleicher.

67

Figure 4.4. Bivariate map illustrating the distribution of MUAs and the distribution of spatial access to GAHT. MUAs were
dichotomous, and “low GAHT access” was defined as SPAI of less than the physician to population ratio discussed above of
1:3500 (0.0002857) and “high GAHT access” was any SPAI above that threshold.

68

Figure 4.5. Hot spot analysis showing clusters of high and low access to GAHT for transgender people

69

Beyond this county level breakdown of statistically significant cold spots, we illustrate
the distribution of the statistically significant hot and cold spots for GAHT access overlaid by
MUA status. In Figure 4.6 below, the lighter areas in both blue and red represent tracts that fall
within MUAs according to HRSA data. Both blues, light and dark, represent statistically
significant cold spots for GAHT, meaning those areas with the lowest access according to SPAI
measures. Both reds, light and dark, represent the areas with highest access to GAHT, those who
were in statistically significant hot spot clusters. All of the tracts that are not in MUAs and are
hot spots for GAHT access are in the Houston and San Antonio areas. The largest clusters of
tracts with cold spots for GAHT access not in otherwise MUAs are in Austin, Fort Worth, and
the eastern part of Dallas and its surrounding area.
4.5. Discussion and Conclusion
Focusing specifically on the results from the SPAI measure shows significant service shortages,
as illustrated above. These shortages were revealed even with twice as long of a drive time
window than used in the original paper outlining the method (Wan, Zou, and Sternberg 2012). In
contrast, the SPAR measure erases all shortage areas, but also introduced some extreme values.
While the SPAR measure was introduced to correct for potential overestimation of service gaps,
it is unlikely that this correction is necessary in the case of access to GAHT the way it was with
primary care, for which it was designed. With implications for research in other areas of
specialty healthcare, this significant shift in results points to a potential limitation of the SPAR
measure when applied to a kind of medical care that is less commonly provided, and likely less
commonly accessed, than primary care even though the overall methodology is sound.
The methodologies for identifying MUAs and the 3SFCA for measuring spatial access
are different but share a similar goal: identifying potential care shortages. Building on the results

70

Figure 4.6. Bivariate map showing MUA status and hot or cold spot according to Getis-Ord Gi* analysis of spatial access to GAHT
for transgender people

71

of the 3SFCA, the hot spot analysis using the Getis-Ord Gi* revealed clustering patterns
of high and low access to GAHT that warrant further investigation from a health services
planning perspective. There was both alignment and mismatch between the distribution of
MUAs and spatial access to GAHT. First, some areas with statistically significantly low access
to GAHT within the greater Dallas metropolitan area, especially Fort Worth, were in areas that
were not otherwise designated as MUAs for primary care. Second, some areas within the greater
Houston area, such as Sugar Land to the south, had statistically significantly high spatial access
to GAHT despite being designated MUAs for primary care (Figure 4.6). These results
demonstrate that patterns of spatial access to transgender-specific healthcare do not necessarily
follow the same trends as primary care. Moreover, urbanicity and rurality did not necessarily
align with higher or lower access to GAHT, respectively (Figure 3). Where existing studies of
access to care have established that rural communities face disparities, our findings do not
demonstrate a direct link between rurality and lower access to care. While this could be due in
part to the larger drive time window we used of two hours, this choice we made in our analysis
would not explain the lack of strong association between rurality and access to GAHT on its
own. Below we begin to flesh out drivers of access to GAHT that go beyond the scope of our
analysis, but which we have reason to hypothesize significantly impact access to care for trans
people.
These shortage areas and the patterns of misalignment between MUAs and access to
GAHT which our study identified represent the complex interplay of the sociopolitical climate
for transgender people as well as infrastructural issues within the health system and where
healthcare is provided. Indeed, the local and state level policy context, including ordinances on
non-discrimination or whether conversion therapy is banned, likely influences the availability of

72

GAHT and transgender people’s willingness or ability to access it. These kinds of policies that
enable or hinder the provision of and access to GAHT and other medical care for trans people
also impact the quality of life for trans communities. The legal and policy environment, while
beyond the scope of our empirical analysis in this chapter, should be studied in future work.
Moreover, the current literature on transgender-specific healthcare has identified a
persistent lack of medical education regarding trans-specific issues (Safer 2013; Arora et al.
2020; Korpaisarn and Safer 2018), and even stigma toward trans people among their would be
providers (Goldenberg et al. 2019; Ali, Fleisher, and Erickson 2016; Bristol, Kostelec, and
MacDonald 2018; Jabson, Mitchell, and Doty 2016; Rowan et al. 2019; Turban et al. 2018).
Addressing this gap in education may lead to an increase in clinicians’ willingness to provide
GAHT, and indeed more inclusive and culturally competent care for all kinds of trans patients.
However, proliferating access to GAHT would not necessarily address the continuing concerns
about access to primary care. In addition to educational initiatives that would increase the supply
of knowledgeable providers able to offer trans-specific medical care like GAHT, state and local
health authorities and other governing bodies should incentivize cultural competency training to
make primary care more accessible and inclusive for trans people. In the context of our case
study, Texas, where resources are limited and almost every county has at least one MUA,
targeted interventions are crucial. Our analysis offers results that can help target areas where
trans-specific care provision is in low supply, like in Fort Worth and east Dallas. Even further,
efforts to address the existing care shortages for primary care in areas like Waco, the suburbs to
the southeast of Dallas, and the outskirts of Austin should also involve training providers to work
with trans communities and provide transgender-specific healthcare to be most effective.
However, while the alignment, and just as often the misalignment, of areas with poor spatial

73

access to GAHT and areas designated as MUAs according to HRSA can be useful for crafting
geographically specific interventions, our results suggest that the MUA is likely a limited metric
for understanding trans-specific care and, perhaps even more so, trans people’s access to
healthcare generally. Future research should consider whether the methodology for identifying
MUAs is relevant to the study of access to healthcare beyond just primary care for the general
population. Even further, the 3SFCA treats the entire population in a defined geographic area as
a singular unit and assumes each person would have the same access to the same facilities.
Our analysis was limited to the spatial dimensions of access to GAHT. However, it has
also been established that trans people seeking GAHT will sometimes access hormones without
the supervision of a medical provider (Hernandez, Santos, and Wilson 2020; Rotondi et al. 2013;
Clark et al. 2018; Metastasio et al. 2018). In addition, the recent rise of telemedicine has begun
to change the means of accessing hormones for trans people (Asaad et al. 2020; Hamnvik et al.
2020). These developments are not surprising considering the discrimination, and anticipation of
it, that transgender people often face when accessing healthcare of any kind (James et al. 2016;
Bradford et al. 2013; Kattari et al. 2015; Kcomt 2020; Rodriguez, Agardh, and Asamoah 2018b;
Romanelli and Lindsey 2020; Miller and Grollman 2015). Thus, further work in multiple
contexts to destigmatize trans identities, and indeed trans people, is needed to improve access to
care for trans communities. Moreover, as telemedicine expands, especially with companies like
Folx and Plume that specialize in transgender and broader LGBT healthcare emerging, access to
care could further expand if this format for delivery of care is popularized and made widely
available.
Ultimately, this chapter measures a downstream problem, access to care, which reflects
numerous upstream issues. These upstream issues, such as the specific concerns for transgender

74

healthcare mentioned in this section and in the introduction, are beyond the scope of our analysis
in this chapter. We are unaware of any models that exist which quantify the specific access to
care concerns that transgender communities have. Future work should incorporate other barriers
and facilitators of accessing trans-specific medical care like GAHT. The 3SFCA could be
adapted to incorporate other concerns about access to medical care, especially from a health and
human rights perspective, which would also make the model more useful for measuring access
holistically, as well as for measuring access to other kinds of care and in other geographic
contexts. Specifically, drawing on the Availability, Accessibility, Acceptability and Quality
(AAAQ) framework from general comment 14 on the right to health within the International
Covenant on Economic, Social and Cultural Rights could be quite useful. Experts in health and
human rights have suggested that the AAAQ framework can inform all aspects of the health
system from policy to service planning to provision of care (Gruskin, Bogecho, and Ferguson
2010). Incorporating these principles into the 3SFCA approach would offer a more holistic
measure and understanding of access to healthcare not only for trans people seeking GAHT, but
for other kinds of care from abortion and HIV treatment to primary care. For this reason, we call
for an intersectional, praxis-oriented approach to research that seeks to move past documenting
disparities in favor of proposing solutions to them (Bowleg 2021). While this study begins to
propose possible solutions, the core issue we have identified is that the dearth of data on
transgender healthcare access necessitates a reliance on estimates and synthetic data. In this way,
we contribute by way of beginning to identify the potential spatial dimensions to the established
health disparities that transgender communities face. More and better data is needed to be able to
adequately identify, measure, and understand these spatial dimensions in access to transgender
healthcare so that better interventions can be designed to eliminate the gaps in services.

75

Chapter 5 Conclusion and Future Work
This dissertation sought to address the dearth of publicly available data on transgender
health and demonstrate the kind of insights that can be drawn from analyses that use spatially
explicit data. Given this focus, the work presented here can be seen as an answer to Jen Jack
Gieseking call for “good enough GIS,” (Gieseking 2018). This took shape as a focus on
producing meaningful resources while also acknowledging the limitations of the available data.
For example, the USTS data used in chapter 2 in its raw form contains less than 14% responses
from trans people who did not exclusively identify as white. Additionally, the earliest attempt at
creating the database from chapter 2 relied on web scraping and web crawling to pull data on
healthcare facilities that provide GAHT to trans people. However, the fact that there is no
consistent, standardized language for discussing this kind of care and no existing gold standard
database against which to measure our database’s efficacy, seeing those computational methods
through would have produced a test of the methods’ limitations rather than as comprehensive of
a database as possible. For this reason, it made more sense from the “good enough GIS”
perspective to focus on amalgamating community facing resources and validating them through
publicly available information as a means to model how users themselves might seek out GAHT
providers online. This will also enable a more useful user facing resource in the future even
though this first iteration of the database was designed strictly for research purposes. However,
such an effort will be limited by the willingness, and indeed the legality, of providers to advertise
that they offer these services to patients. The current hostile sociolegal and political environment
in the US not only affects trans people themselves, but also the clinicians who provider their
care. A future public-facing version of this database will need to account for this limitation and
also consider the ethical concerns in growing the database. While its current iteration only

76

includes those healthcare facilities and providers which publicly advertise that they offer GAHT
to trans patients, there are likely many more clinicians who provide this care, but are unable to
safely advertise their services. The construction of the database in and of itself is important
because it is a first of its kind that can be used for research and serve as a foundation for a user
facing resource and that is the “good enough” that Gieseking outlines as paramount.
Relatedly, Chapter 3 demonstrated that the existing hypothesis that the US trans
population is more ethnoracially diverse than the general population seems to hold true when
tested against a new data set. Moreover, the studies that produce estimates of the size of the US
trans population often tout that the source upon which they rely, the BRFSS, is nationally
representative; however, we unequivocally demonstrate that while the survey overall may be
nationally representative, it is far from it when it comes to the trans subpopulation. We also
demonstrated that making the gender identity questions not only mandatory, but actually the only
set of questions that ascertain sex and gender for all respondents, likely accounted for, at least in
part, the significant jump in rates of trans identification for the HPS from the BRFSS. In this
way, our descriptive and exploratory spatial analysis serves as a necessary first step toward more
advanced, inferential studies that can test hypotheses that emanate from our work. Without this
study to demonstrate that other data sources exist which give us reason to believe the BRFSS
presents an undercount of the US trans population, others studying trans people would have to
rely on conjecture. This contribution serves a different purpose than something like an inferential
comparison of self rated health across the two data sets because such an analysis would treat
them as reasonably similar enough to draw such comparisons. In fact, we demonstrate that they
are perhaps more different than alike when it comes to the trans population.

77

Finally, in chapter 4 we offered a first of its kind analysis that measures the potential
geographic accessibility of GAHT in Texas with spatially explicit methods. This study builds on
chapters 2 and 3 because it deploys both the database produced in chapter 2 and the population
estimates derived from chapter 3, and also because it moves from data collection and exploratory
analysis to an inferential study. At the same time that it is a first of its kind study with important
results, we ultimately deploy estimates at census tract level that were derived at the state or MSA
level and thereby assume that trans people’s patterns of residence and mobility follow the
general population completely. While it could be argued that we have no reason to believe that
trans people’s geographic and demographic patterns are any different from the general
population, by the same token, we could argue we have no reason to believe they are the same.
In fact, our work to this point suggests that trans communities’ unique experiences warrant
further specific investigation in order to test the hypothesis that there are significant differences
in terms of community demographics, patterns of residence and mobility, and health outcomes
between trans people and the general population. There are more advanced techniques that could
have been used to model potential patterns of spatial access to GAHT than what we used in
chapter 4. Namely spatial microsimulation could have created synthetic data at the same or
smaller geographic scale where demographic characteristics could also more closely follow the
established demographic patterns in the US trans population. However, from the perspective of
good enough GIS, and really from a perspective that prioritizes actionable insights over and
against advanced methods for the sake of using advanced methods, the more layers of predictive
modeling or simulation that are built into data, the more assumptions we make about a
population about which very little has been firmly established.

78

In this way, this dissertation, when taken collectively, models what can be done with the
best available data to predict and model spatial and demographic patterns while staying close to
the data as it appears in its sources. And perhaps most importantly, the studies in this dissertation
underscore the profound need for more and better data on transgender life, in all its forms, not
only those which have either been made the most legible or have been allowed to take up the
most space in data, and indeed the popular imagination. While advocacy work on behalf of trans
communities, and indeed by trans communities, will continue regardless of the conclusions this
dissertation draws, we suggest that improving not only the processes and protocols for data
collection with trans people, but also the policies that enable, or even mandate, such data
collection will necessarily make for better and more informed advocacy. With this in mind, we
sketch out avenues for future work that build off of these studies in what remains of the
conclusion.
5.1. Future Work
To that end, the next phase of the work begun in chapter 2 will be two-fold. First, I
intend to create a user-facing resource out of the database I made that can both incorporate user
feedback on providers and facilities and be designed as an accessible and easy to use interface.
The database that came out of the work outlined in chapter 2 was designed for research, but it
could be expanded in this way to become a useful tool for trans people seeking care in an
increasingly complex and difficult to navigate health system. It could also be replicated for other
countries and globalized. Second, I aim to follow the Guttmacher Institute’s model of conducting
a census of care providers. This census could be used to not only expand the database, but also to
better understand providers’ motivations for serving trans clientele and their attitudes toward
how the health system could be improved. Taken together, the expansion of this work begun in

79

the dissertation could provide immeasurable benefits to trans communities not only in the US but
around the world who consistently face stigma, attacks on their rights, and difficulty in accessing
healthcare.
Building on this theme, chapter 3 offered a spatially explicit analysis of trans population
estimates in the US. While the results point to the likelihood of existing undercounts, they are
limited in that they both emanate from data sets not designed to be about the US trans
population. Further, both protocols for ascertaining gender identity have limited acceptability
among trans communities due to the ever-changing nature of language among trans people and
the overall lack of trans people with lived expertise involved in designing such protocols. Thus,
future work emerging from this study should be both empirically driven and advocacy-focused. I
aim to collaborate with other trans experts to design a study that will test the feasibility of a new
set of protocols for ascertaining gender identity that can, as much as is possible, incorporate all
the unique ways that trans and gender diverse people self-identify. This proposed future research
will build off of a proposed theory from Florence Ashley which they name “gender modality,”
(Ashley 2022). Ashley argues that the term gender modality “refers to how a person’s gender
identity stands in relation to their gender assigned at birth,” (Ashley 2022, 22). In my own
qualitative recoding of the write-in data for gender identity in the USTS I found that just under
900 respondents refused the 25 available gender identity options in favor of writing in either man
or woman despite the inclusion of trans man and trans woman in the 25 available options (Kronk
et al. 2021). Adapting Ashley’s concept of gender modality in data collection would enable new
kinds of study designs in which we need not rely solely on assigned sex at birth or self-
identification as transgender (as a separate category from male or female) to ascertain non-
binary, transgender, and other diverse gender modalities. Dylan Felt, a research scientist with

80

expertise in trans population health based at Northwestern University, has forthcoming work in
which she adapted gender modality into demographic data collection to ascertain cisgender,
transgender, and non-binary identities and the potential overlaps between them. Future work
should build off of Felt’s study by incorporating cognitive debriefing and focus groups to ensure
that the feasibility for statistical testing that gender modality introduces is matched by
acceptability of the protocols among trans communities. In building out this research agenda, we
would have data that can be used to advocate for incorporation of the latest and best
demographic data collection protocols not only for research, but for population level data
collection by state and federal entities.
In contrast to the above, a recent report from a committee on “Measuring Sex, Gender
Identity, and Sexual Orientation for the National Institutes of Health” convened by the National
Academies of Sciences, Engineering and Medicine (NASEM) offered a multidisciplinary
perspective on how to collect data that they argue is both acceptable to LGBTQ communities and
feasible for government entities (Committee on Measuring Sex, Gender Identity, and Sexual
Orientation et al. 2022). However, the report suggests protocols for ascertaining gender identity
that are almost identical to those first proposed by Charlotte Tate almost a decade prior (Tate,
Ledbetter, and Youssef 2013). An update is needed and my analysis in chapter 3 demonstrates
the need for not only new protocols, but spatially explicit sampling and analysis to understand
not only how trans people can meaningfully participate in large scale data collection, but also
how their identities, issues, and concerns vary geographically.
The analysis in chapter 4 produced the most actionable insights into where access to
GAHT can be improved. The 3SFCA demonstrated that in some areas which are also
underserved from a primary health care perspective, like Austin, Fort Worth, and east Dallas,

81

access to GAHT could be improved with targeted interventions to increase the provision of
GAHT, for example by incentivizing education in both trans-friendly primary care and the
provision of GAHT for providers working to address existing primary care gaps for the general
population in those areas. In contrast, some places had significantly worse access to GAHT in
areas that were not otherwise medically underserved. The interventions here are more specific to
transgender healthcare and would need to be targeted directly perhaps by cross-training existing
providers working in the area. Overall, the results of my analysis suggest that patterns of access
to trans-specific healthcare do not neatly follow the same geographic patterns as access to
primary care for the general population. This insight is perhaps surprising from a health services
planning perspective, but less so from the vantage point of those working in the trans health
space. Future work will entail replicating this analysis at a larger scale to understand regional
patterns and potential geographic disparities across the US. Producing the same kind of
geographically specific insights that the 3SFCA offers will be crucial in understanding where
different kinds of interventions are needed to address gaps in access to transgender healthcare.
Ideally, funding opportunities can be sought to incentivize the educational interventions
mentioned above that could proliferate access to GAHT as part and parcel of proliferating access
to primary care for everyone.
Ultimately, new methods are needed to better understand the patterns of geographic
access to trans healthcare that the substantive portion of this dissertation ends with. While the
3SFCA works well for illustrating spatial access in terms of travel time, from a health and human
rights perspective it is limited to only one pillar of what constitutes good access to healthcare.
For example, General Comment no. 14 on the right to the highest attainable standard of health
from Article 12 of the International Covenant on Economic, Social and Cultural Rights

82

(ICESCR) outlines the principles of Availability, Accessibility, Acceptability, and Quality, or
AAAQ (UN Committee on Economic, Social, and Cultural Rights 2000). Sofia Gruskin, Dina
Bogecho, and Laura Ferguson elaborate on the AAAQ principles and offer a framework for
implementing and assessing them in both health policies and at the health systems level
(Gruskin, Bogecho, and Ferguson 2010). Within this framework, the physical availability of
services is only one small piece and indeed is the lowest rung of the ladder that leads toward
realizing the right to health. Future work should incorporate a health and human rights
perspective in order to expand upon quantitative methods like the 3SFCA and meaningfully
incorporate other aspects of accessibility, as well as acceptability and quality. More specifically,
the consent model a healthcare facility uses for transgender healthcare can be a determining
factor in not only how trans people access services, but also whether they access them at all.
Adapting the 3SFCA method to incorporate not only medical capacity, which in the method
entails only the number of providers working at a given facility, but also things like fee models,
consent models, insurance and co-payment schemas among other factors would be a welcome
addition to both the fields of medical geography and health and human rights. Moreover, access
to information, including translated materials and interpretation services for visits with providers,
could be incorporated into the existing quantitative method as a means to elaborate upon the
3SFCA approach. Other aspects from the broadest sense of the right to health, such as the rights
to participation, non-discrimination, and accountability, which are perhaps more difficult to
quantify and empirically analyze are paramount to a more holistic understanding, and therefore
analysis of whether healthcare facilities and providers are actually protecting and promoting the
right to health for trans clientele. Indeed, this elaboration of the 3SFCA approach would have
much wider application than only transgender-specific medical care, especially in countries like

83

the US that have only ostensibly recognized the right to health by signing the ICESCR, but not
ratifying it and meaningfully incorporating it into its legal system.
Moreover, novel results beget novel theories. While this dissertation stays close to the
data, I am inspired by Donna Haraway’s call for us to “stay with the trouble,” and her insistence
on sym-poiesis (making-with) over and against auto-poiesis (self-making) (Haraway 2016) . For
this work, staying with the trouble entails wrestling with the contradictions inherent to
quantitative analysis that attempts to consolidate complex life experiences into compact sets of
data to find emerging patterns. Rather than insisting that the contradictions be resolved, staying
with the trouble means refusing to resolve them because they reflect the complex interplay of
trans people’s lived realities, the promise and limits of empirical analysis, and my own
positionality as a trans woman invested in the health and rights of my gender diverse community.
For example, Chapter 4 includes some unexamined results wherein border towns and majority
Latinx areas had ostensibly better access to GAHT compared to the remainder of Texas despite
the compounding structural inequalities that necessarily limit access to all kinds of healthcare,
regardless of what the results might show. My future work will require a more in-depth
examination of this phenomenon, and others, beyond the existing theoretical explanations that
quantitative geographical analysis offers such as edge effects, long drive time windows, or the
limitations of available data. Quantitative analysis, from my perspective, is useful as a
foundation for understanding population level patterns, or even how these macro level patterns
differ across space and place, and across demographics. Qualitative empirical analysis, then, is
the logical next step wherein rich, narrative driven data can be collected to give deeper context to
those macro level patterns. Theory at its best scaffolds on both the quantitative and qualitative
analytical modes to make meaning so that data and statistics make sense. While this dissertation

84

is limited in its theoretical conclusions, my own background is not. As a humanist cum social
and spatial scientist, this dissertation work and my doctoral training have prepared me to bridge
the gaps between these modes of analysis and thinking. Even though I cannot accomplish this
alone, I will take advantage of my multidisciplinary training and network of collaborators to
push the scientific and theoretical literature. The analyses in this dissertation have become a
strong foundation on which I intend to build as I launch my post-doctorate career.
In conclusion, this dissertation identified and sought to address a significant gap in the
scientific literature on transgender population health. And at the same time that it is an
accomplishment of which I am quite proud, it also highlighted precisely how much more work
needs to be done. My hope is that this dissertation will be useful to my colleagues and my
community as they work to improve access to healthcare for trans people and advocate for health
equity for trans people, and indeed everyone else as well. While this work represents the
culmination of my time as a doctoral student, it also represents the first step onto my path as a
credentialed researcher who can now, hopefully, work to affect the kind of change my
communities need.

85

References
AAMC. 2020. “Physician Specialty Data Report 2020.” Association of American Medical
Colleges. https://www.aamc.org/what-we-do/mission-areas/health-care/workforce-
studies/interactive-data/number-people-active-physician-specialty-2019.
Alcántara, Carmela, Shakira F. Suglia, Irene Perez Ibarra, A. Louise Falzon, Elliot McCullough,
Talha Alvi, and Leopoldo J. Cabassa. 2021. “Disaggregation of Latina/o Child and Adult
Health Data: A Systematic Review of Public Health Surveillance Surveys in the United
States.” Population Research and Policy Review 40 (1): 61–79.
https://doi.org/10.1007/s11113-020-09633-4.
Ali, Nareesa, William Fleisher, and Julie Erickson. 2016. “Psychiatrists’ and Psychiatry
Residents’ Attitudes Toward Transgender People.” Academic Psychiatry 40 (2): 268–73.
https://doi.org/10.1007/s40596-015-0308-y.
Arístegui, Inés, Pablo D. Radusky, Virginia Zalazar, Marcela Romero, Jessica Schwartz, and
Omar Sued. 2017. “Impact of the Gender Identity Law in Argentinean Transgender
Women.” International Journal of Transgenderism 18 (4): 446–56.
https://doi.org/10.1080/15532739.2017.1314796.
Arora, Maansi, Kaete Walker, Judy Luu, Robbert J. Duvivier, Tinashe Dune, and Katie Wynne.
2020. “Education of the Medical Profession to Facilitate Delivery of Transgender Health
Care in an Australian Health District.” Australian Journal of Primary Health 26 (1): 17.
https://doi.org/10.1071/PY19102.
Arrowsmith, Laura. 2017. “‘Go Back To California’: When Providers Fail Transgender
Patients.” Health Affairs 36 (9): 1679–82. https://doi.org/10.1377/hlthaff.2017.0738.
Asaad, Malke, Aashish Rajesh, Krishna Vyas, and Shane D. Morrison. 2020. “Telemedicine in
Transgender Care: A Twenty-First–Century Beckoning.” Plastic & Reconstructive
Surgery 146 (1): 108e–9. https://doi.org/10.1097/PRS.0000000000006935.
Ashley, Florence. 2022. “‘Trans’ Is My Gender Modality.” In 978-0190092726, 2nd ed., 22–23.
Oxford University Press.
Augustaitis, Laima, Leland A. Merrill, Kristi E Gamarel, and Oliver L. Haimson. 2021. “Online
Transgender Health Information Seeking: Facilitators, Barriers, and Future Directions.”
In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems,
1–14. Yokohama Japan: ACM. https://doi.org/10.1145/3411764.3445091.
Baker, Annie. 2011. “The Importance of Creating a Healthcare Provider Reference Database for
Rural and Underserved Populations.” Perspectives in Health Information Management /
AHIMA, American Health Information Management Association 8 (Fall): 1e.
Baker, Kellan E. 2019. “Findings From the Behavioral Risk Factor Surveillance System on
Health-Related Quality of Life Among US Transgender Adults, 2014-2017.” JAMA
Internal Medicine 179 (8): 1141. https://doi.org/10.1001/jamainternmed.2018.7931.
Bakko, Matthew, and Shanna K. Kattari. 2020. “Transgender-Related Insurance Denials as
Barriers to Transgender Healthcare: Differences in Experience by Insurance Type.”
Journal of General Internal Medicine 35 (6): 1693–1700. https://doi.org/10.1007/s11606-
020-05724-2.
———. 2021. “Differential Access to Transgender Inclusive Insurance and Healthcare in the
United States: Challenges to Health across the Life Course.” Journal of Aging & Social
Policy 33 (1): 67–81. https://doi.org/10.1080/08959420.2019.1632681.

86

Barton, Cassie. 2021. “Preparing for the 2021 Census (England and Wales).” House of
Commons Library. https://researchbriefings.files.parliament.uk/documents/CBP-
8531/CBP-8531.pdf.
Basu, Sanjay, Seth A. Berkowitz, Robert L. Phillips, Asaf Bitton, Bruce E. Landon, and Russell
S. Phillips. 2019. “Association of Primary Care Physician Supply With Population
Mortality in the United States, 2005-2015.” JAMA Internal Medicine 179 (4): 506.
https://doi.org/10.1001/jamainternmed.2018.7624.
Bauer, Greta R., Jessica Braimoh, Ayden I. Scheim, and Christoffer Dharma. 2017.
“Transgender-Inclusive Measures of Sex/Gender for Population Surveys: Mixed-
Methods Evaluation and Recommendations.” PloS One 12 (5): e0178043.
https://doi.org/10.1371/journal.pone.0178043.
Borgogna, Nicholas C., Ryon C. McDermott, Stephen L. Aita, and Matthew M. Kridel. 2019.
“Anxiety and Depression across Gender and Sexual Minorities: Implications for
Transgender, Gender Nonconforming, Pansexual, Demisexual, Asexual, Queer, and
Questioning Individuals.” Psychology of Sexual Orientation and Gender Diversity 6 (1):
54–63. https://doi.org/10.1037/sgd0000306.
Boscoe, Francis P., Kevin A. Henry, and Michael S. Zdeb. 2012. “A Nationwide Comparison of
Driving Distance Versus Straight-Line Distance to Hospitals.” The Professional
Geographer : The Journal of the Association of American Geographers 64 (2): 188–96.
https://doi.org/10.1080/00330124.2011.583586.
Bosse, Jordon D., Raeann G. Leblanc, Kasey Jackman, and Ragnhildur I. Bjarnadottir. 2018.
“Benefits of Implementing and Improving Collection of Sexual Orientation and Gender
Identity Data in Electronic Health Records.” CIN: Computers, Informatics, Nursing 36
(6): 267–74. https://doi.org/10.1097/CIN.0000000000000417.
Bowleg, Lisa. 2021. “Evolving Intersectionality Within Public Health: From Analysis to
Action.” American Journal of Public Health 111 (1): 88–90.
https://doi.org/10.2105/AJPH.2020.306031.
Bradford, Judith, Sari L. Reisner, Julie A. Honnold, and Jessica Xavier. 2013. “Experiences of
Transgender-Related Discrimination and Implications for Health: Results From the
Virginia Transgender Health Initiative Study.” American Journal of Public Health 103
(10): 1820–29. https://doi.org/10.2105/AJPH.2012.300796.
Bristol, Sarah, Teresa Kostelec, and Ryan MacDonald. 2018. “Improving Emergency Health
Care Workers’ Knowledge, Competency, and Attitudes Toward Lesbian, Gay, Bisexual,
and Transgender Patients Through Interdisciplinary Cultural Competency Training.”
Journal of Emergency Nursing 44 (6): 632–39. https://doi.org/10.1016/j.jen.2018.03.013.
Bryant, James, and Paul L. Delamater. 2019. “Examination of Spatial Accessibility at Micro- and
Macro-Levels Using the Enhanced Two-Step Floating Catchment Area (E2SFCA)
Method.” Annals of GIS 25 (3): 219–29.
https://doi.org/10.1080/19475683.2019.1641553.
Bukowski, Leigh A., Cristian J. Chandler, Stephanie L. Creasy, Derrick D. Matthews, Mackey
R. Friedman, and Ronald D. Stall. 2018. “Characterizing the HIV Care Continuum and
Identifying Barriers and Facilitators to HIV Diagnosis and Viral Suppression Among
Black Transgender Women in the United States:” JAIDS Journal of Acquired Immune
Deficiency Syndromes 79 (4): 413–20. https://doi.org/10.1097/QAI.0000000000001831.
Cahill, Sean, Robbie Singal, Chris Grasso, Dana King, Kenneth Mayer, Kellan Baker, and
Harvey Makadon. 2014. “Do Ask, Do Tell: High Levels of Acceptability by Patients of

87

Routine Collection of Sexual Orientation and Gender Identity Data in Four Diverse
American Community Health Centers.” Edited by Garrett Prestage. PLoS ONE 9 (9):
e107104. https://doi.org/10.1371/journal.pone.0107104.
Canner, Joseph K., Omar Harfouch, Lisa M. Kodadek, Danielle Pelaez, Devin Coon, Anaeze C.
Offodile II, Adil H. Haider, and Brandyn D. Lau. 2018. “Temporal Trends in Gender-
Affirming Surgery Among Transgender Patients in the United States.” JAMA Surgery
153 (7): 609–16. https://doi.org/10.1001/jamasurg.2017.6231.
Carney, Nikita. 2021. “Constructing Race and Ethnicity: ‘It Has to Do with Where You Are.’”
Spatial Demography 9 (3): 375–96. https://doi.org/10.1007/s40980-021-00087-6.
Cassarino, Melanie, Eunice Correa, Sebastián Minoletti, Amalia Botto, Romina Rapisarda, María
Florencia Grande Ratti, and Daniel Luna. 2020. “A Path Towards Inclusion:
Transdisciplinary Experience for the Inclusion of Self-Perceived Gender in an
Information System.” Studies in Health Technology and Informatics 270 (June): 901–5.
https://doi.org/10.3233/SHTI200292.
Cassarino, Melanie, Eunice Correa, Sebastián Minoletti, Ignacio Jauregui, María Florencia
Grande Ratti, Carlos Otero, and Daniel Luna. 2020. “A Path to Inclusion: Design and
Prototype of Transgender Identity in an Electronic Health Record.” Studies in Health
Technology and Informatics 270 (June): 1181–82. https://doi.org/10.3233/SHTI200352.
Catney, Gemma, and Christopher D. Lloyd. 2020. “Population Grids for Analysing Long-Term
Change in Ethnic Diversity and Segregation.” Spatial Demography 8 (3): 215–49.
https://doi.org/10.1007/s40980-020-00071-6.
CDC, and BRFSS. 2015. “Behavioral Risk Factor Surveillance System - Overview: BRFSS
2014.” Centers for Disease Control and Prevention.
https://www.cdc.gov/brfss/annual_data/2014/pdf/Overview_2014.pdf.
———. 2016. “Behavioral Risk Factor Surveillance System - Overview: BRFSS 2015.” Centers
for Disease Control and Prevention.
https://www.cdc.gov/brfss/annual_data/2015/pdf/overview_2015.pdf.
———. 2017. “Behavioral Risk Factor Surveillance System - Overview: BRFSS 2016.” Centers
for Disease Control and Prevention.
https://www.cdc.gov/brfss/annual_data/2016/pdf/overview_2016.pdf.
———. 2018. “Behavioral Risk Factor Surveillance System - Overview: BRFSS 2017.” Centers
for Disease Control and Prevention.
https://www.cdc.gov/brfss/annual_data/2017/pdf/overview-2017-508.pdf.
———. 2019a. “Behavioral Risk Factor Surveillance System - Overview: BRFSS 2019.”
Centers for Disease Control and Prevention.
https://www.cdc.gov/brfss/annual_data/2019/pdf/overview-2019-508.pdf.
———. 2019b. “The Behavioral Risk Factor Surveillance System - Overview: BRFSS 2018.”
Centers for Disease Control and Prevention.
https://www.cdc.gov/brfss/annual_data/2018/pdf/overview-2018-508.pdf.
———. 2021. “Behavioral Risk Factor Surveillance System - Overview: BRFSS 2020.” Centers
for Disease Control and Prevention.
https://www.cdc.gov/brfss/annual_data/2020/pdf/overview-2020-508.pdf.
Cicero, Ethan C., Sari L. Reisner, Elizabeth I. Merwin, Janice C. Humphreys, and Susan G.
Silva. 2020. “The Health Status of Transgender and Gender Nonbinary Adults in the
United States.” Edited by Jonathan Garcia. PLOS ONE 15 (2): e0228765.
https://doi.org/10.1371/journal.pone.0228765.

88

Clark, Kirsty A., Jaclyn M. White Hughto, and John E. Pachankis. 2017. “‘What’s the Right
Thing to Do?’ Correctional Healthcare Providers’ Knowledge, Attitudes and Experiences
Caring for Transgender Inmates.” Social Science & Medicine 193 (November): 80–89.
https://doi.org/10.1016/j.socscimed.2017.09.052.
Clark, Kirsty, Jesse B. Fletcher, Ian W. Holloway, and Cathy J. Reback. 2018. “Structural
Inequities and Social Networks Impact Hormone Use and Misuse Among Transgender
Women in Los Angeles County.” Archives of Sexual Behavior 47 (4): 953–62.
https://doi.org/10.1007/s10508-017-1143-x.
Committee on Measuring Sex, Gender Identity, and Sexual Orientation, Committee on National
Statistics, Division of Behavioral and Social Sciences and Education, and National
Academies of Sciences, Engineering, and Medicine. 2022. Measuring Sex, Gender
Identity, and Sexual Orientation. Edited by Tara Becker. Washington, D.C.: National
Academies Press. https://doi.org/10.17226/26424.
Conron, Kerith J, and Kathryn O’Neill. 2021. “Food Insufficiency Among Transgender Adults
During the COVID-19 Pandemic.” Los Angeles, CA: The Williams Institute.
https://williamsinstitute.law.ucla.edu/wp-content/uploads/Trans-Food-Insufficiency-Dec-
2021.pdf.
Coulter, Robert W.S., John R. Blosnich, Leigh A. Bukowski, A.L. Herrick, Daniel E. Siconolfi,
and Ron D. Stall. 2015. “Differences in Alcohol Use and Alcohol-Related Problems
between Transgender- and Nontransgender-Identified Young Adults.” Drug and Alcohol
Dependence 154 (September): 251–59. https://doi.org/10.1016/j.drugalcdep.2015.07.006.
Crissman, Halley P., Daphna Stroumsa, Emily K. Kobernik, and Mitchell B. Berger. 2019.
“Gender and Frequent Mental Distress: Comparing Transgender and Non-Transgender
Individuals’ Self-Rated Mental Health.” Journal of Women’s Health 28 (2): 143–51.
https://doi.org/10.1089/jwh.2018.7411.
Crosby, Richard A., Laura F. Salazar, and Brandon J. Hill. 2016. “Gender Affirmation and
Resiliency Among Black Transgender Women With and Without HIV Infection.”
Transgender Health 1 (1): 86–93. https://doi.org/10.1089/trgh.2016.0005.
Dai, Dajun, and Fahui Wang. 2011. “Geographic Disparities in Accessibility to Food Stores in
Southwest Mississippi.” Environment and Planning B: Planning and Design 38 (4): 659–
77. https://doi.org/10.1068/b36149.
Delamater, Paul L. 2013. “Spatial Accessibility in Suboptimally Configured Health Care
Systems: A Modified Two-Step Floating Catchment Area (M2SFCA) Metric.” Health &
Place 24 (November): 30–43. https://doi.org/10.1016/j.healthplace.2013.07.012.
Delamater, Paul L, Joseph P Messina, Ashton M Shortridge, and Sue C Grady. 2012. “Measuring
Geographic Access to Health Care: Raster and Network-Based Methods.” International
Journal of Health Geographics 11 (1): 15. https://doi.org/10.1186/1476-072X-11-15.
Deutsch, Madeline B., and David Buchholz. 2015. “Electronic Health Records and Transgender
Patients—Practical Recommendations for the Collection of Gender Identity Data.”
Journal of General Internal Medicine 30 (6): 843–47. https://doi.org/10.1007/s11606-
014-3148-7.
Deutsch, Madeline B., Jamison Green, JoAnne Keatley, Gal Mayer, Jennifer Hastings,
Alexandra M. Hall, and World Professional Association for Transgender Health EMR
Working Group. 2013. “Electronic Medical Records and the Transgender Patient:
Recommendations from the World Professional Association for Transgender Health

89

EMR Working Group.” Journal of the American Medical Informatics Association:
JAMIA 20 (4): 700–703. https://doi.org/10.1136/amiajnl-2012-001472.
Dickey, Lore M., Sari L. Reisner, and Cindy Lee Juntunen. 2015. “Non-Suicidal Self-Injury in a
Large Online Sample of Transgender Adults.” Professional Psychology: Research and
Practice 46 (1): 3–11. https://doi.org/10.1037/a0038803.
Downing, Janelle M., and Julia M. Przedworski. 2018. “Health of Transgender Adults in the
U.S., 2014–2016.” American Journal of Preventive Medicine 55 (3): 336–44.
https://doi.org/10.1016/j.amepre.2018.04.045.
Edmiston, E. Kale, Cameron A. Donald, Alice Rose Sattler, J. Klint Peebles, Jesse M. Ehrenfeld,
and Kristen Laurel Eckstrand. 2016. “Opportunities and Gaps in Primary Care
Preventative Health Services for Transgender Patients: A Systematic Review.”
Transgender Health 1 (1): 216–30. https://doi.org/10.1089/trgh.2016.0019.
Eriksson, Sven E.S., and Joshua D. Safer. 2016. “Evidence-Based Curricular Content Improves
Student Knowledge and Changes Attitudes towards Transgender Medicine.” Endocrine
Practice 22 (7): 837–41. https://doi.org/10.4158/EP151141.OR.
Feldman, Jamie, George R. Brown, Madeline B. Deutsch, Wylie Hembree, Walter Meyer, Heino
F.L. Meyer-Bahlburg, Vin Tangpricha, Guy T’Sjoen, and Joshua D. Safer. 2016.
“Priorities for Transgender Medical and Healthcare Research.” Current Opinion in
Endocrinology, Diabetes & Obesity 23 (2): 180–87.
https://doi.org/10.1097/MED.0000000000000231.
Feldman, Jamie L., Winston Ekaprasetia Luhur, Jody L. Herman, Tonia Poteat, and Ilan H.
Meyer. 2021. “Health and Health Care Access in the US Transgender Population Health
(TransPop) Survey.” Andrology 9 (6): 1707–18. https://doi.org/10.1111/andr.13052.
Flores, Andrew R, Taylor N. T. Brown, and Jody L. Herman. 2016. “Race and Ethnicity of
Adults Who Identify as Transgender in the United States.” Los Angeles, CA: The
Williams Institute. https://williamsinstitute.law.ucla.edu/wp-content/uploads/Race-
Ethnicity-Trans-Adults-US-Oct-2016.pdf.
Flores, Andrew R, Jody L Herman, Gary J Gates, and Taylor N T Brown. 2016. “How Many
Adults Identify as Transgender in the United States?” Los Angeles, CA: The Williams
Institute. https://williamsinstitute.law.ucla.edu/wp-content/uploads/Trans-Adults-US-
Aug-2016.pdf.
Fowler, Christopher, Barrett Lee, Stephen Matthews, Christopher S Fowler, Barrett A Lee, and
Stephen A Matthews. 2016. “The Contributions of Places to Metropolitan Ethnoracial
Diversity and Segregation: Decomposing Change Across Space and Time.” Demography
(Springer Nature) 53 (6): 1955–77. https://doi.org/10.1007/s13524-016-0517-3.
Galinsky, Adena M., Catherine Simile, Carla E. Zelaya, Tina Norris, and Sela V. Panapasa.
2019. “Surveying Strategies for Hard-to-Survey Populations: Lessons From the Native
Hawaiian and Pacific Islander National Health Interview Survey.” American Journal of
Public Health 109 (10): 1384–91. https://doi.org/10.2105/AJPH.2019.305217.
Gieseking, Jen Jack. 2018. “Operating Anew: Queering GIS with Good Enough Software.” The
Canadian Geographer / Le Géographe Canadien 62 (1): 55–66.
https://doi.org/10.1111/cag.12397.
“Goal: Improve the Health, Safety, and Well-Being of Lesbian, Gay, Bisexual, and Transgender
People.” n.d. Government website. Healthy People 2030. Accessed January 26, 2022.
https://health.gov/healthypeople/objectives-and-data/browse-objectives/lgbt.

90

Goldenberg, Tamar, Laura Jadwin-Cakmak, Elliot Popoff, Sari L. Reisner, Bré A. Campbell, and
Gary W. Harper. 2019. “Stigma, Gender Affirmation, and Primary Healthcare Use
Among Black Transgender Youth.” Journal of Adolescent Health 65 (4): 483–90.
https://doi.org/10.1016/j.jadohealth.2019.04.029.
Goldenberg, Tamar, Sari L. Reisner, Gary W. Harper, Kristi E. Gamarel, and Rob Stephenson.
2020. “State Policies and Healthcare Use Among Transgender People in the U.S.”
American Journal of Preventive Medicine 59 (2): 247–59.
https://doi.org/10.1016/j.amepre.2020.01.030.
Goldstein, Zil, Trevor A Corneil, and Dina N Greene. 2017. “When Gender Identity Doesn’t
Equal Sex Recorded at Birth: The Role of the Laboratory in Providing Effective
Healthcare to the Transgender Community.” Clinical Chemistry 63 (8): 1342–52.
https://doi.org/10.1373/clinchem.2016.258780.
Gonzales, Gilbert, and Carrie Henning‐Smith. 2017. “Barriers to Care Among Transgender and
Gender Nonconforming Adults.” The Milbank Quarterly 95 (4): 726–48.
https://doi.org/10.1111/1468-0009.12297.
Groseclose, Samuel L., and David L. Buckeridge. 2017. “Public Health Surveillance Systems:
Recent Advances in Their Use and Evaluation.” Annual Review of Public Health 38 (1):
57–79. https://doi.org/10.1146/annurev-publhealth-031816-044348.
Gruskin, Sofia, Dina Bogecho, and Laura Ferguson. 2010. “‘Rights-Based Approaches’ to
Health Policies and Programs: Articulations, Ambiguities, and Assessment.” Journal of
Public Health Policy 31 (2): 129–45. https://doi.org/10.1057/jphp.2010.7.
Guyan, Kevin. 2022. Queer Data: Using Gender, Sex and Sexuality Data for Action.
Bloomsbury Studies in Digital Cultures. London New York Oxford New Delhi Sydney:
Bloomsbury.
Haimson, Oliver L., and Tiffany C. Veinot. 2020. “Coming Out to Doctors, Coming Out to
‘Everyone’: Understanding the Average Sequence of Transgender Identity Disclosures
Using Social Media Data.” Transgender Health 5 (3): 158–65.
https://doi.org/10.1089/trgh.2019.0045.
Hamnvik, Ole-Petter R., Shailesh Agarwal, Christopher G. AhnAllen, Anna L. Goldman, and
Sari L. Reisner. 2020. “Telemedicine and Inequities in Health Care Access: The Example
of Transgender Health.” Transgender Health, December, trgh.2020.0122.
https://doi.org/10.1089/trgh.2020.0122.
Haraway, Donna Jeanne. 2016. Staying with the Trouble: Making Kin in the Chthulucene.
Experimental Futures: Technological Lives, Scientific Arts, Anthropological Voices.
Durham: Duke University Press.
Herman, Jody L., Andrew R Flores, Taylor N. T. Brown, Bianca D.M. Wilson, and Kerith J
Conron. 2017. “Age of Individuals Who Identify as Transgender in the United States.”
Los Angeles, CA: The Williams Institute. http://williamsinstitute.law.ucla.edu/wp-
content/uploads/Age-Trans-Individuals-Jan-2017.pdf.
Hernandez, Christopher J., Glenn-Milo Santos, and Erin C. Wilson. 2020. “Association of
Documentation of Legal Residency Status with Nonprescribed Hormone Use Among
Hispanic/Latina Trans Women in San Francisco.” Health Equity 4 (1): 263–71.
https://doi.org/10.1089/heq.2019.0104.
Homelessness Research Institute. 2020. “Transgender Homeless Adults & Unsheltered
Homelessness: What the Data Tell Us.” National Alliance to End Homelessness.

91

https://endhomelessness.org/wp-content/uploads/2020/07/Trans-Homelessness-Brief-
July-2020.pdf.
Horton, Hayward Derrick. 1999. “Critical Demography: The Paradigm of the Future?”
Sociological Forum 14 (3): 363–67.
Howard, Susanna D., Kevin L. Lee, Aviva G. Nathan, Hannah C. Wenger, Marshall H. Chin, and
Scott C. Cook. 2019. “Healthcare Experiences of Transgender People of Color.” Journal
of General Internal Medicine 34 (10): 2068–74. https://doi.org/10.1007/s11606-019-
05179-0.
HRSA. 2021. “What Is a Shortage Designation?” Health Resources & Services Administration.
February 2021. https://bhw.hrsa.gov/workforce-shortage-areas/shortage-designation.
Humble, Robert M, Katherine L Imborek, Nicole Nisly, Dina N Greene, and Matthew D
Krasowski. 2019. “Common Hormone Therapies Used to Care for Transgender Patients
Influence Laboratory Results.” The Journal of Applied Laboratory Medicine 3 (5): 799–
814. https://doi.org/10.1373/jalm.2018.027078.
Jabson, Jennifer M., Jason W. Mitchell, and S. Benjamin Doty. 2016. “Associations between
Non-Discrimination and Training Policies and Physicians’ Attitudes and Knowledge
about Sexual and Gender Minority Patients: A Comparison of Physicians from Two
Hospitals.” BMC Public Health 16 (1): 256. https://doi.org/10.1186/s12889-016-2927-y.
James, Sandy E., Jody L. Herman, Susan Rankin, Mara Keisling, Lisa Mottet, and Ma’ayan
Anafi. 2016. “The Report of the 2015 United States Transgender Survey.” National
Center for Transgender Equality.
Jones, Rachel K., Elizabeth Witwer, and Jenna Jerman. 2017. “Abortion Incidence and Service
Availability the United States.” Guttmatcher Institute.
https://doi.org/10.1363/2019.30760.
Kachen, Axenya, and Jennifer R. Pharr. 2020. “Health Care Access and Utilization by
Transgender Populations: A United States Transgender Survey Study.” Transgender
Health 5 (3): 141–48. https://doi.org/10.1089/trgh.2020.0017.
Kaholokula, Joseph Keawe’aimoku, this link will open in a new window Link to external site,
Scott K. Okamoto, and Barbara W. K. Yee. 2019. “Special Issue Introduction: Advancing
Native Hawaiian and Other Pacific Islander Health.” Asian American Journal of
Psychology 10 (3): 197–205. http://dx.doi.org/10.1037/aap0000167.
Kattari, Shanna K., Jarrod Call, Brendon T. Holloway, Leonardo Kattari, and Kristie L. Seelman.
2021. “Exploring the Experiences of Transgender and Gender Diverse Adults in
Accessing a Trans Knowledgeable Primary Care Physician.” International Journal of
Environmental Research and Public Health 18 (24): 13057.
https://doi.org/10.3390/ijerph182413057.
Kattari, Shanna K., N. Eugene Walls, Darren L. Whitfield, and Lisa Langenderfer-Magruder.
2015. “Racial and Ethnic Differences in Experiences of Discrimination in Accessing
Health Services Among Transgender People in the United States.” International Journal
of Transgenderism 16 (2): 68–79. https://doi.org/10.1080/15532739.2015.1064336.
Katz, Amy S., Billie-Jo Hardy, Michelle Firestone, Aisha Lofters, and Melody E. Morton-
Ninomiya. 2020. “Vagueness, Power and Public Health: Use of ‘vulnerable‘ in Public
Health Literature.” Critical Public Health 30 (5): 601–11.
https://doi.org/10.1080/09581596.2019.1656800.

92

Kauh, Tina J., Jen’nan Ghazal Read, and A. J. Scheitler. 2021. “The Critical Role of
Racial/Ethnic Data Disaggregation for Health Equity.” Population Research and Policy
Review 40 (1): 1–7. https://doi.org/10.1007/s11113-020-09631-6.
Kcomt, Luisa. 2020. “Healthcare Avoidance Due to Anticipated Discrimination among
Transgender People: A Call to Create Trans-Affirmative Environments.” Population
Health, 8.
Keuroghlian, Alex S., Derri Shtasel, and Ellen L. Bassuk. 2014. “Out on the Street: A Public
Health and Policy Agenda for Lesbian, Gay, Bisexual, and Transgender Youth Who Are
Homeless.” American Journal of Orthopsychiatry 84 (1): 66–72.
https://doi.org/10.1037/h0098852.
Khalili, Joshua, Lucinda B. Leung, and Allison L. Diamant. 2015. “Finding the Perfect Doctor:
Identifying Lesbian, Gay, Bisexual, and Transgender–Competent Physicians.” American
Journal of Public Health 105 (6): 1114–19. https://doi.org/10.2105/AJPH.2014.302448.
Knight, K. G., A. R. Flores, and S. J. Nezhad. 2015. “Surveying Nepal’s Third Gender:
Development, Implementation, and Analysis.” TSQ: Transgender Studies Quarterly 2
(1): 101–22. https://doi.org/10.1215/23289252-2848904.
Korpaisarn, Sira, and Joshua D. Safer. 2018. “Gaps in Transgender Medical Education among
Healthcare Providers: A Major Barrier to Care for Transgender Persons.” Reviews in
Endocrine and Metabolic Disorders 19 (3): 271–75. https://doi.org/10.1007/s11154-018-
9452-5.
Kronk, Clair A, Avery R Everhart, Florence Ashley, Hale M Thompson, Theodore E Schall,
Teddy G Goetz, Laurel Hiatt, et al. 2021. “Transgender Data Collection in the Electronic
Health Record: Current Concepts and Issues.” Journal of the American Medical
Informatics Association, no. ocab136 (September).
https://doi.org/10.1093/jamia/ocab136.
Lagos, Danya. 2018. “Looking at Population Health Beyond ‘Male’ and ‘Female’: Implications
of Transgender Identity and Gender Nonconformity for Population Health.” Demography
55 (6): 2097–2117. https://doi.org/10.1007/s13524-018-0714-3.
Learmonth, Claire, Rebekah Viloria, Cei Lambert, Hilary Goldhammer, and Alex S.
Keuroghlian. 2018. “Barriers to Insurance Coverage for Transgender Patients.” American
Journal of Obstetrics and Gynecology 219 (3): 272.e1-272.e4.
https://doi.org/10.1016/j.ajog.2018.04.046.
Lett, Elle, Matthew P. Abrams, Arya Gold, Farrah-Amoy Fullerton, and Avery Everhart. 2022.
“Ethnoracial Inequities in Access to Gender-Affirming Mental Health Care and
Psychological Distress among Transgender Adults.” Social Psychiatry and Psychiatric
Epidemiology, February. https://doi.org/10.1007/s00127-022-02246-6.
Lett, Elle, Emmanuella Ngozi Asabor, Sourik Beltrán, and Nadia Dowshen. 2020.
“Characterizing Health Inequities for the U.S. Transgender Hispanic Population Using
the Behavioral Risk Factor Surveillance System.” Transgender Health, December,
trgh.2020.0095. https://doi.org/10.1089/trgh.2020.0095.
Lett, Elle, Nadia L. Dowshen, and Kellan E. Baker. 2020. “Intersectionality and Health
Inequities for Gender Minority Blacks in the U.S.” American Journal of Preventive
Medicine 59 (5): 639–47. https://doi.org/10.1016/j.amepre.2020.04.013.
Lett, Elle, and Avery Everhart. 2021. “Considerations for Transgender Population Health
Research Based on US National Surveys.” Annals of Epidemiology, October,
S1047279721003173. https://doi.org/10.1016/j.annepidem.2021.10.009.

93

Levi, Denise, María Victoria Vazquez, María Victoria Giussi, S. Esteban, and Analía Baum.
2019. “Representation of the Transgender Population in Electronic Health Records:
Implementation Strategy in the Public Health Care System of Buenos Aires City.”
Studies in Health Technology and Informatics 264 (August): 1131–34.
https://doi.org/10.3233/SHTI190402.
Liebler, Carolyn, Sonya Porter, Leticia Fernandez, James Noon, Sharon Ennis, Carolyn A
Liebler, Sonya R Porter, Leticia E Fernandez, James M Noon, and Sharon R Ennis. 2017.
“America’s Churning Races: Race and Ethnicity Response Changes Between Census
2000 and the 2010 Census.” Demography (Springer Nature) 54 (1): 259–84.
https://doi.org/10.1007/s13524-016-0544-0.
Linard, Catherine, Victor A Alegana, Abdisalan M Noor, Robert W Snow, and Andrew J Tatem.
2010. “A High Resolution Spatial Population Database of Somalia for Disease Risk
Mapping.” International Journal of Health Geographics 9 (1): 45.
https://doi.org/10.1186/1476-072X-9-45.
Lombardi, Emilia, and Swagata Banik. 2016. “The Utility of the Two-Step Gender Measure
Within Trans and Cis Populations.” Sexuality Research and Social Policy 13 (3): 288–96.
https://doi.org/10.1007/s13178-016-0220-6.
López, Ártemis, and Carla Míguez Bóveda. 2021. “Ti, Eu, Ele e a Linguaxe Non Binaria Na
Tradución Entre Inglés e Castelán.” Viceversa. Revista Galega de Tradución, no. 21
(April): 253–61. https://doi.org/10.35869/viceversa.v0i21.3469.
Luo, Wei. 2004. “Using a GIS-Based Floating Catchment Method to Assess Areas with Shortage
of Physicians.” Health & Place 10 (1): 1–11. https://doi.org/10.1016/S1353-
8292(02)00067-9.
Luo, Wei, and Yi Qi. 2009. “An Enhanced Two-Step Floating Catchment Area (E2SFCA)
Method for Measuring Spatial Accessibility to Primary Care Physicians.” Health & Place
15 (4): 1100–1107. https://doi.org/10.1016/j.healthplace.2009.06.002.
Luo, Wei, and Fahui Wang. 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): 865–84. https://doi.org/10.1068/b29120.
Macapagal, Kathryn, Ramona Bhatia, and George J. Greene. 2016. “Differences in Healthcare
Access, Use, and Experiences Within a Community Sample of Racially Diverse Lesbian,
Gay, Bisexual, Transgender, and Questioning Emerging Adults.” LGBT Health 3 (6):
434–42. https://doi.org/10.1089/lgbt.2015.0124.
Macinko, James, Barbara Starfield, and Leiyu Shi. 2007. “Quantifying the Health Benefits of
Primary Care Physician Supply in the United States.” International Journal of Health
Services 37 (1): 111–26. https://doi.org/10.2190/3431-G6T7-37M8-P224.
MacRury, Sandra, Jim Finlayson, Susan Hussey-Wilson, and Samantha Holden. 2016.
“Development of a Pseudo/Anonymised Primary Care Research Database: Proof-of-
Concept Study.” Health Informatics Journal 22 (2): 113–19.
https://doi.org/10.1177/1460458214535118.
Maina, Joseph, Paul O. Ouma, Peter M. Macharia, Victor A. Alegana, Benard Mitto, Ibrahima
Socé Fall, Abdisalan M. Noor, Robert W. Snow, and Emelda A. Okiro. 2019. “A Spatial
Database of Health Facilities Managed by the Public Health Sector in Sub Saharan
Africa.” Scientific Data 6 (1): 134. https://doi.org/10.1038/s41597-019-0142-2.

94

Mandal, Chaitali, Paramita Debnath, and Apyayee Sil. 2020. “‘Other’ Gender in India: An
Analysis of 2011 Census Data.” Preprint. SocArXiv.
https://doi.org/10.31235/osf.io/j48qy.
Maragh-Bass, Allysha C. 2019. “The ‘Ask’ Is the ‘Answer’: Implementing Routine
Documentation of Sexual Orientation and Gender Identity in Health Care.” American
Journal of Public Health 109 (8): 1071–73. https://doi.org/10.2105/AJPH.2019.305192.
Maragh-Bass, Allysha C., Maya Torain, Rachel Adler, Anju Ranjit, Eric Schneider, Ryan Y.
Shields, Lisa M. Kodadek, et al. 2017. “Is It Okay To Ask: Transgender Patient
Perspectives on Sexual Orientation and Gender Identity Collection in Healthcare.” Edited
by Zachary F. Meisel. Academic Emergency Medicine 24 (6): 655–67.
https://doi.org/10.1111/acem.13182.
Massey, Douglas S. 1999. “What Critical Demography Means to Me.” Sociological Forum 14
(3): 525–28.
Matthews, Stephen A., Laura Stiberman, James Raymer, Tse-Chuan Yang, Ezra Gayawan,
Sayambhu Saita, Sai Thein Than Tun, et al. 2021. “Looking Back, Looking Forward:
Progress and Prospect for Spatial Demography.” Spatial Demography 9 (1): 1–29.
https://doi.org/10.1007/s40980-021-00084-9.
McGrail, Matthew R. 2012. “Spatial Accessibility of Primary Health Care Utilising the Two Step
Floating Catchment Area Method: An Assessment of Recent Improvements.”
International Journal of Health Geographics 11 (1): 50. https://doi.org/10.1186/1476-
072X-11-50.
McGrail, Matthew R., Peter M. Wingrove, Stephen M. Petterson, and Andrew W. Bazemore.
2017. “Mobility of US Rural Primary Care Physicians During 2000–2014.” The Annals of
Family Medicine 15 (4): 322–28. https://doi.org/10.1370/afm.2096.
Meerwijk, Esther L., and Jae M. Sevelius. 2017. “Transgender Population Size in the United
States: A Meta-Regression of Population-Based Probability Samples.” American Journal
of Public Health 107 (2): e1–8. https://doi.org/10.2105/AJPH.2016.303578.
Metastasio, Antonio, Attilio Negri, Giovanni Martinotti, and Ornella Corazza. 2018.
“Transitioning Bodies. The Case of Self-Prescribing Sexual Hormones in Gender
Affirmation in Individuals Attending Psychiatric Services.” Brain Sciences 8 (5): 88.
https://doi.org/10.3390/brainsci8050088.
Meyer, Ilan H., Taylor N. T. Brown, Jody L. Herman, Sari L. Reisner, and Walter O. Bockting.
2017. “Demographic Characteristics and Health Status of Transgender Adults in Select
US Regions: Behavioral Risk Factor Surveillance System, 2014.” American Journal of
Public Health 107 (4): 582–89. https://doi.org/10.2105/AJPH.2016.303648.
Miller, Lisa R., and Eric Anthony Grollman. 2015. “The Social Costs of Gender Nonconformity
for Transgender Adults: Implications for Discrimination and Health.” Sociological
Forum 30 (3): 809–31.
Moe, Jeffry, Dilani Perera-Diltz, and Narketta Sparkman-Key. 2018. “Exploring the
Relationships among Practitioners’ Consultation Skills, Competence to Counsel Lesbian,
Gay, and Bisexual Clients, and Attitudes toward Transgender People.” Journal of Gay &
Lesbian Social Services 30 (4): 374–92. https://doi.org/10.1080/10538720.2018.1516170.
Movement Advancement Project. 2022. “Equality Maps: Healthcare Laws and Policies.”
Movement Advancement Project. 2022. https://www.lgbtmap.org/equality-
maps/healthcare_laws_and_policies.

95

NCTE. 2020. “Summary of State Birth Certificate Gender Change Laws.” National Center for
Transgender Equality.
https://transequality.org/sites/default/files/images/Summary%20of%20State%20Birth%2
0Certificate%20Laws%20Apr%2028%202020.pdf.
———. 2021. “How Trans Friendly Is the Driver’s License Gender Marker Change Policy in
Your State?” National Center for Transgender Equality.
https://transequality.org/sites/default/files/Drivers%20License%20Grades%20July%2020
21a_0.pdf.
———. n.d. “Identification Documents and Transgender People: An Overview of the Name and
Gender Marker Change Process in the United States.” National Center for Transgender
Equality. Accessed January 27, 2022.
https://transequality.org/sites/default/files/docs/resources/ID-Documents-Overview.pdf.
Nisly, Nicole L., Katherine L. Imborek, Michelle L. Miller, Susan D. Kaliszewski, Rachel M.
Williams, and Matthew D. Krasowski. 2018. “Unique Primary Care Needs of
Transgender and Gender Non-Binary People.” Clinical Obstetrics & Gynecology 61 (4):
674–86. https://doi.org/10.1097/GRF.0000000000000404.
Nolan, Ian T., Christopher J. Kuhner, and Geolani W. Dy. 2019. “Demographic and Temporal
Trends in Transgender Identities and Gender Confirming Surgery.” Translational
Andrology and Urology 8 (3): 184–90. https://doi.org/10.21037/tau.2019.04.09.
Noor, Abdisalan M, Victor A Alegana, Peter W Gething, and Robert W Snow. 2009. “A Spatial
National Health Facility Database for Public Health Sector Planning in Kenya in 2008.”
International Journal of Health Geographics 8 (1): 13. https://doi.org/10.1186/1476-
072X-8-13.
O’Bryan, Jane, Melissa Scribani, Kimberly Leon, Nancy Tallman, Christopher Wolf-Gould,
Carolyn Wolf-Gould, and Anne Gadomski. 2020. “Health-Related Quality of Life among
Transgender and Gender Expansive Youth at a Rural Gender Wellness Clinic.” Quality of
Life Research 29 (6): 1597–1607. https://doi.org/10.1007/s11136-020-02430-8.
Parthasarathi, Pavithra, David Levinson, and Hartwig Hochmair. 2013. “Network Structure and
Travel Time Perception.” Edited by Bin Jiang. PLoS ONE 8 (10): e77718.
https://doi.org/10.1371/journal.pone.0077718.
Perez-Brumer, Amaya, Jack K. Day, Stephen T. Russell, and Mark L. Hatzenbuehler. 2017.
“Prevalence and Correlates of Suicidal Ideation Among Transgender Youth in California:
Findings From a Representative, Population-Based Sample of High School Students.”
Journal of the American Academy of Child & Adolescent Psychiatry 56 (9): 739–46.
https://doi.org/10.1016/j.jaac.2017.06.010.
Poteat, Tonia. 2013. “Managing Uncertainty: A Grounded Theory of Stigma in Transgender
Health Care Encounters.” Social Science, 8.
Puckett, Jae A., Peter Cleary, Kinton Rossman, Brian Mustanski, and Michael E. Newcomb.
2018. “Barriers to Gender-Affirming Care for Transgender and Gender Nonconforming
Individuals.” Sexuality Research and Social Policy 15 (1): 48–59.
https://doi.org/10.1007/s13178-017-0295-8.
Raymer, James, Frans Willekens, and Andrei Rogers. 2019. “Spatial Demography: A Unifying
Core and Agenda for Further Research.” Population, Space and Place 25 (4): e2179.
https://doi.org/10.1002/psp.2179.
Reback, Cathy J., Dahlia Ferlito, Kimberly A. Kisler, and Jesse B. Fletcher. 2015. “Recruiting,
Linking, and Retaining High-Risk Transgender Women Into HIV Prevention and Care

96

Services: An Overview of Barriers, Strategies, and Lessons Learned.” International
Journal of Transgenderism 16 (4): 209–21.
https://doi.org/10.1080/15532739.2015.1081085.
Redfern, Jan S., and Bill Sinclair. 2014. “Improving Health Care Encounters and
Communication with Transgender Patients.” Journal of Communication in Healthcare 7
(1): 25–40. https://doi.org/10.1179/1753807614Y.0000000045.
Reisner, Sari L., Kerith J. Conron, Scout, Kellan Baker, J. L. Herman, Emilia Lombardi, E. A.
Greytak, Alison M. Gill, and A. K. Matthews. 2015. “‘Counting’ Transgender and
Gender-Nonconforming Adults in Health Research: Recommendations from the Gender
Identity in US Surveillance Group.” TSQ: Transgender Studies Quarterly 2 (1): 34–57.
https://doi.org/10.1215/23289252-2848877.
Reisner, Sari L., Sabra L. Katz-Wise, Allegra R. Gordon, Heather L. Corliss, and S. Bryn Austin.
2016. “Social Epidemiology of Depression and Anxiety by Gender Identity.” Journal of
Adolescent Health 59 (2): 203–8. https://doi.org/10.1016/j.jadohealth.2016.04.006.
Richardson, Douglas B., Nora D. Volkow, Mei-Po Kwan, Robert M. Kaplan, Michael
Goodchild, and Richard Croyle. 2013. “Spatial Turn in Health Research.” Science 339:
1390–92.
Ricketts, Thomas C., and George M. Holmes. 2007. “Mortality and Physician Supply: Does
Region Hold the Key to the Paradox?” Health Services Research 42 (6 Pt 1): 2233–51;
discussion 2294-2323. https://doi.org/10.1111/j.1475-6773.2007.00728.x.
Roberto, Elizabeth, and Elizabeth Korver-Glenn. 2021. “The Spatial Structure and Local
Experience of Residential Segregation.” Spatial Demography 9 (3): 277–307.
https://doi.org/10.1007/s40980-021-00086-7.
Rodriguez, Amanda, Anette Agardh, and Benedict Oppong Asamoah. 2018a. “Self-Reported
Discrimination in Health-Care Settings Based on Recognizability as Transgender: A
Cross-Sectional Study Among Transgender U.S. Citizens.” Archives of Sexual Behavior
47 (4): 973–85. https://doi.org/10.1007/s10508-017-1028-z.
———. 2018b. “Self-Reported Discrimination in Health-Care Settings Based on Recognizability
as Transgender: A Cross-Sectional Study Among Transgender U.S. Citizens.” Archives
of Sexual Behavior 47 (4): 973–85. https://doi.org/10.1007/s10508-017-1028-z.
Romanelli, Meghan, and Michael A. Lindsey. 2020. “Patterns of Healthcare Discrimination
Among Transgender Help-Seekers.” American Journal of Preventive Medicine 58 (4):
e123–31. https://doi.org/10.1016/j.amepre.2019.11.002.
Rood, Brian A., Sari L. Reisner, Francisco I. Surace, Jae A. Puckett, Meredith R. Maroney, and
David W. Pantalone. 2016. “Expecting Rejection: Understanding the Minority Stress
Experiences of Transgender and Gender-Nonconforming Individuals.” Transgender
Health 1 (1): 151–64. https://doi.org/10.1089/trgh.2016.0012.
Rotondi, Nooshin Khobzi, Greta R. Bauer, Kyle Scanlon, Matthias Kaay, Robb Travers, and
Anna Travers. 2013. “Nonprescribed Hormone Use and Self-Performed Surgeries: ‘Do-
It-Yourself’ Transitions in Transgender Communities in Ontario, Canada.” American
Journal of Public Health 103 (10): 1830–36. https://doi.org/10.2105/AJPH.2013.301348.
Rowan, Shon P., Christa L. Lilly, Robert E. Shapiro, Kacie M. Kidd, Rebecca M. Elmo, Robin
A. Altobello, and Manuel C. Vallejo. 2019. “Knowledge and Attitudes of Health Care
Providers Toward Transgender Patients Within a Rural Tertiary Care Center.”
Transgender Health 4 (1): 24–34. https://doi.org/10.1089/trgh.2018.0050.

97

Safer, Joshua D. 2013. “Transgender Medical Research, Provider Education, and Patient Access
Are Overdue.” Endocrine Practice 19 (4): 575–76. https://doi.org/10.4158/EP13013.ED.
Safer, Joshua D., Eli Coleman, Jamie Feldman, Robert Garofalo, Wylie Hembree, Asa Radix,
and Jae Sevelius. 2016. “Barriers to Healthcare for Transgender Individuals:” Current
Opinion in Endocrinology & Diabetes and Obesity 23 (2): 168–71.
https://doi.org/10.1097/MED.0000000000000227.
Sinnard, Morgan T., Christopher R. Raines, and Stephanie L. Budge. 2016. “The Association
Between Geographic Location and Anxiety and Depression in Transgender Individuals:
An Exploratory Study of an Online Sample.” Transgender Health 1 (1): 181–86.
https://doi.org/10.1089/trgh.2016.0020.
Snelgrove, John W, Amanda M Jasudavisius, Bradley W Rowe, Evan M Head, and Greta R
Bauer. 2012. “‘Completely out-at-Sea’ with ‘Two-Gender Medicine’: A Qualitative
Analysis of Physician-Side Barriers to Providing Healthcare for Transgender Patients.”
BMC Health Services Research 12 (1): 110. https://doi.org/10.1186/1472-6963-12-110.
Statistics Canada. 2021. “2021 Census: 3A.” Government website. April 28, 2021.
https://www.statcan.gc.ca/en/statistical-programs/instrument/3901_Q5_V4.
Steinmetz, Katy. 2014. “The Transgender Tipping Point.” TIME, June 9, 2014.
Strmic-Pawl, Hephzibah V., Brandon A. Jackson, and Steve Garner. 2018. “Race Counts: Racial
and Ethnic Data on the U.S. Census and the Implications for Tracking Inequality.”
Sociology of Race and Ethnicity 4 (1): 1–13. https://doi.org/10.1177/2332649217742869.
Stroumsa, Daphna, Halley P. Crissman, Vanessa K. Dalton, Giselle Kolenic, and Caroline R.
Richardson. 2020. “Insurance Coverage and Use of Hormones Among Transgender
Respondents to a National Survey.” The Annals of Family Medicine 18 (6): 528–34.
https://doi.org/10.1370/afm.2586.
Stryker, Susan. 2008. Transgender History. Seal Studies. Berkeley, CA: Seal Press : Distributed
by Publishers Group West.
Stryker, Susan, Paisley Currah, and Lisa Jean Moore. 2008. “Introduction: Trans-, Trans, or
Transgender?” Women’s Studies Quarterly 36 (3/4): 11–22.
Tan, Kyle K.H., Sonja J. Ellis, Johanna M. Schmidt, Jack L. Byrne, and Jaimie F. Veale. 2020.
“Mental Health Inequities among Transgender People in Aotearoa New Zealand:
Findings from the Counting Ourselves Survey.” International Journal of Environmental
Research and Public Health 17 (8): 2862. https://doi.org/10.3390/ijerph17082862.
Tanton, Robert. 2017. “Spatial Microsimulation: Developments and Potential Future Directions.”
International Journal of Microsimulation 11 (1): 143–61.
https://doi.org/10.34196/ijm.00176.
Tao, Zhuolin, Yang Cheng, and Jixiang Liu. 2020. “Hierarchical Two-Step Floating Catchment
Area (2SFCA) Method: Measuring the Spatial Accessibility to Hierarchical Healthcare
Facilities in Shenzhen, China.” International Journal for Equity in Health 19 (1): 164.
https://doi.org/10.1186/s12939-020-01280-7.
Tate, Charlotte Chuck, Jay N. Ledbetter, and Cris P. Youssef. 2013. “A Two-Question Method
for Assessing Gender Categories in the Social and Medical Sciences.” The Journal of Sex
Research 50 (8): 767–76. https://doi.org/10.1080/00224499.2012.690110.
Terris-Feldman, Alexandra, Alessandra Chen, Grace Poudrier, and Maurice Garcia. 2020. “How
Accessible Is Genital Gender-Affirming Surgery for Transgender Patients With
Commercial and Public Health Insurance in the United States? Results of a Patient-

98

Modeled Search for Services and a Survey of Providers.” Sexual Medicine 8 (4): 664–72.
https://doi.org/10.1016/j.esxm.2020.08.005.
Tordoff, Diana, Michele Andrasik, and Anjum Hajat. 2019. “Misclassification of Sex Assigned
at Birth in the Behavioral Risk Factor Surveillance System and Transgender
Reproductive Health: A Quantitative Bias Analysis.” Epidemiology 30 (5): 669–78.
https://doi.org/10.1097/EDE.0000000000001046.
Tordoff, Diana M., Jennifer Morgan, Julia C. Dombrowski, Matthew R. Golden, and Lindley A.
Barbee. 2019. “Increased Ascertainment of Transgender and Non-Binary Patients Using a
2-Step Versus 1-Step Gender Identity Intake Question in an STD Clinic Setting.”
Sexually Transmitted Diseases 46 (4): 254–59.
https://doi.org/10.1097/OLQ.0000000000000952.
Turban, Jack L, Joel Winer, Susan Boulware, Timothy VanDeusen, and John Encandela. 2018.
“Knowledge and Attitudes toward Transgender Health.” The Clinical Teacher 15 (3):
203–7. https://doi.org/10.1111/tct.12738.
UN Committee on Economic, Social, and Cultural Rights. 2000. “General Comment No. 14
(2000), The Right to the Highest Attainable Standard of Health (Article 12 of the
International Covenant on Economic, Social and Cultural Rights).” Geneva: United
Nations.
US Census Bureau. 2021a. “Source of the Data and Accuracy of the Estimates for the Household
Pulse Survey - Phase 3.2 - Week 34.” United States Census Bureau.
https://www2.census.gov/programs-surveys/demo/technical-documentation/hhp/Phase3-
2_Source_and_Accuracy_Week%2034.pdf.
———. 2021b. “Source of the Data and Accuracy of the Estimates for the Household Pulse
Survey - Phase 3.2 - Week 35.” United States Census Bureau.
https://www2.census.gov/programs-surveys/demo/technical-documentation/hhp/Phase3-
2_Source_and_Accuracy_Week%2035.pdf.
———. 2021c. “Source of the Data and Accuracy of the Estimates for the Household Pulse
Survey - Phase 3.2 - Week 36.” United States Census Bureau.
https://www2.census.gov/programs-surveys/demo/technical-documentation/hhp/Phase3-
2_Source_and_Accuracy_Week%2036.pdf.
———. 2021d. “Source of the Data and Accuracy of the Estimates for the Household Pulse
Survey - Phase 3.2 - Week 37.” United States Census Bureau.
https://www2.census.gov/programs-surveys/demo/technical-documentation/hhp/Phase3-
2_Source_and_Accuracy_Week%2037.pdf.
———. 2021e. “Source of the Data and Accuracy of the Estimates for the Household Pulse
Survey - Phase 3.2 - Week 38.” United States Census Bureau.
https://www2.census.gov/programs-surveys/demo/technical-documentation/hhp/Phase3-
2_Source_and_Accuracy_Week%2037.pdf.
———. 2021f. “Source of the Data and Accuracy of the Estimates for the Household Pulse
Survey - Phase 3.2 - Week 39.” United States Census Bureau.
https://www2.census.gov/programs-surveys/demo/technical-documentation/hhp/Phase3-
2_Source_and_Accuracy_Week39.pdf.
———. 2021g. “Source of the Data and Accuracy of the Estimates for the Household Pulse
Survey - Phase 3.3 - Week 40.” United States Census Bureau.
https://www2.census.gov/programs-surveys/demo/technical-documentation/hhp/Phase3-
3_Source_and_Accuracy_Week40.pdf.

99

USDA. 2020. “Rural Urban Continuum Codes.” USDA Economic Research Service: US
Department of Agriculture. December 10, 2020. https://www.ers.usda.gov/data-
products/rural-urban-continuum-codes.aspx.
Vermeir, Ella, Lois A. Jackson, and Emily Gard Marshall. 2018. “Barriers to Primary and
Emergency Healthcare for Trans Adults.” Culture, Health & Sexuality 20 (2): 232–46.
https://doi.org/10.1080/13691058.2017.1338757.
Wan, Neng, Bin Zou, and Troy Sternberg. 2012. “A Three-Step Floating Catchment Area
Method for Analyzing Spatial Access to Health Services.” International Journal of
Geographical Information Science 26 (6): 1073–89.
https://doi.org/10.1080/13658816.2011.624987.
Wang, Fahui. 2020. “Why Public Health Needs GIS: A Methodological Overview.” Annals of
GIS 26 (1): 1–12. https://doi.org/10.1080/19475683.2019.1702099.
Warner, David Michael, and Arunab Harish Mehta. 2021. “Identifying and Addressing Barriers
to Transgender Healthcare: Where We Are and What We Need to Do About It.” Journal
of General Internal Medicine 36 (11): 3559–61. https://doi.org/10.1007/s11606-021-
07001-2.
White Hughto, Jaclyn M., Gabriel R. Murchison, Kirsty Clark, John E. Pachankis, and Sari L.
Reisner. 2016. “Geographic and Individual Differences in Healthcare Access for U.S.
Transgender Adults: A Multilevel Analysis.” LGBT Health 3 (6): 424–33.
https://doi.org/10.1089/lgbt.2016.0044.
White Hughto, Jaclyn M., Sari L. Reisner, and John E. Pachankis. 2015. “Transgender Stigma
and Health: A Critical Review of Stigma Determinants, Mechanisms, and Interventions.”
Social Science & Medicine 147 (December): 222–31.
https://doi.org/10.1016/j.socscimed.2015.11.010.
White Hughto, Jaclyn M., Adam J. Rose, John E. Pachankis, and Sari L. Reisner. 2017. “Barriers
to Gender Transition-Related Healthcare: Identifying Underserved Transgender Adults in
Massachusetts.” Transgender Health 2 (1): 107–18.
https://doi.org/10.1089/trgh.2017.0014.
Witcomb, Gemma L., Walter Pierre Bouman, Laurence Claes, Nicola Brewin, John R. Crawford,
and Jon Arcelus. 2018. “Levels of Depression in Transgender People and Its Predictors:
Results of a Large Matched Control Study with Transgender People Accessing Clinical
Services.” Journal of Affective Disorders 235 (August): 308–15.
https://doi.org/10.1016/j.jad.2018.02.051.
Witlox, Frank. 2007. “Evaluating the Reliability of Reported Distance Data in Urban Travel
Behaviour Analysis.” Journal of Transport Geography 15 (3): 172–83.
https://doi.org/10.1016/j.jtrangeo.2006.02.012.
Xianyu, Jianchuan, Soora Rasouli, and Harry Timmermans. 2017. “Analysis of Variability in
Multi-Day GPS Imputed Activity-Travel Diaries Using Multi-Dimensional Sequence
Alignment and Panel Effects Regression Models.” Transportation 44 (3): 533–53.
https://doi.org/10.1007/s11116-015-9666-2.
Zaliznyak, Michael, Eric E. Jung, Catherine Bresee, and Maurice M. Garcia. 2021. “Which U.S.
States’ Medicaid Programs Provide Coverage for Gender-Affirming Hormone Therapy
and Gender-Affirming Genital Surgery for Transgender Patients?: A State-by-State
Review, and a Study Detailing the Patient Experience to Confirm Coverage of Services.”
The Journal of Sexual Medicine 18 (2): 410–22.
https://doi.org/10.1016/j.jsxm.2020.11.016.

100

Zou, Yuanshu, Rhonda Szczesniak, Alexis Teeters, Lee Ann E. Conard, and Daniel H.
Grossoehme. 2018. “Documenting an Epidemic of Suffering: Low Health-Related
Quality of Life among Transgender Youth.” Quality of Life Research 27 (8): 2107–15.
https://doi.org/10.1007/s11136-018-1839-y.
Zuberi, Tukufu. 2001. Thicker than Blood: How Racial Statistics Lie. Minneapolis: University of
Minnesota Press.
Zuberi, Tukufu, and Eduardo Bonilla-Silva, eds. 2008. White Logic, White Methods: Racism and
Methodology. Lanham: Rowman & Littlefield Publishers.

101

Appendix A Supplementary Table from Chapter 2
Table A1. Number of respondents in the USTS subsample, the number of GAHT facilities, and
the ratio used in Figure 2.4 of facilities per 1,000 respondents stratified by state

State Total
Respondents
# GAHT
Facilities
Facilities per 1,000
respondents
Alabama 76 9 118
Alaska 47 11 234
Arizona 266 14 53
Arkansas 106 7 66
California 2000 131 66
Colorado 386 19 49
Connecticut 141 7 50
Delaware 43 4 93
District of
Columbia
134 9 67
Florida 472 47 100
Georgia 272 15 55
Hawaii 35 8 229
Idaho 76 8 105
Illinois 583 40 69
Indiana 207 8 39
Iowa 116 15 129
Kansas 97 13 134
Kentucky 127 6 47
Louisiana 127 13 102
Maine 91 12 132
Maryland 308 16 52
Massachusetts 724 32 44
Michigan 406 17 42
Minnesota 376 28 74
Mississippi 36 3 83
Missouri 241 12 50
Montana 39 9 231
Nebraska 75 5 67
Nevada 112 7 62
New Hampshire 115 10 87
New Jersey 273 3 11
New Mexico 125 6 48
New York 1017 48 47
North Carolina 353 40 113

102

North Dakota 23 2 87
Ohio 460 27 59
Oklahoma 99 8 81
Oregon 683 25 37
Pennsylvania 632 28 44
Rhode Island 66 15 227
South Carolina 100 9 90
South Dakota 20 2 100
Tennessee 188 14 74
Texas 695 45 65
Utah 144 4 28
Vermont 94 14 149
Virginia 338 16 47
Washington 943 59 63
West Virginia 28 2 71
Wisconsin 271 10 37
Wyoming 23 1 43

103

Appendix B Supplement to Chapter 3
Table B1. Comparison of unweighted and weighted rates of identification as trans across the
BRFSS and HPS pooled data with the addition of the “none of these” category from the HPS
Unweighted Weighted
Source % Trans % None of
These
%
Combined
%
Trans
% None of
These
%
Combined
BRFSS 0.464 0.551
HPS 0.543 1.072 1.615 0.949 1.631 2.58

These data from HPS reveal that there was a substantial subsample of respondents who neither
identified as transgender, nor with their assigned sex at birth. In the main part of the chapter, we
include those who replied “none of these” as part of the general population due to their explicit
rejection of the category transgender. However, the format of the question implies “transgender”
is a gender identity unto itself, like male or female. Transgender is often used as a modifier, an
adjective, and not as an identity unto itself. This is evidenced by those who identified their
gender as one “opposite” to their assigned sex at birth rather than choosing transgender as an
option. Given that there were roughly twice as many respondents who chose “none of these” as
their gender, further research into the collection of gender identity data at the federal level is
needed. There are adaptations of the existing protocol the HPS used that would be able to
account for variation in gender identity within the larger category of transgender. Future
population level survey research should employ these protocols.

Abstract (if available)

Abstract Few studies have analyzed geographic access to transgender-specific medical care, and none have done so at a scale larger than a single city. However, this is perhaps due to a lack of reliable and readily available data on transgender populations. To investigate access to trans-specific care, this dissertation focuses on gender-affirming hormone therapy (GAHT) because it is routine care and entails at least annual visits. Given that the provision of GAHT is not tracked by public health entities the way other kinds of medical care are, the first aim of this dissertation is to determine where this care is provided in the US. Chapter 2 features the construction of a spatial, national database of GAHT providers in the US as well as a test of its comprehensiveness using the US Transgender Survey data. Chapter 3 offers an exploratory spatial analysis of rates of transgender identification across the US comparing the most often cited data source, the Behavioral Risk Factor Surveillance System, with a new source, the Household Pulse Survey. Ultimately this chapter illustrates that existing estimates of the size and demographics of the US trans population are likely severe undercounts. Chapter 4 builds on chapters 2 and 3 to quantify geographic access to GAHT using a three-step floating catchment area method (3SFCA). It deploys population estimates from chapter 3 to estimate the distribution of trans populations at the census tract level in Texas. Next, it uses the 3SFCA method to measure access for these tract-level trans subpopulations to GAHT providers derived from the GAHT provider database detailed in chapter 2. It also offers recommendations for planning health services and concludes that patterns of geographic access to transgender-specific medical care do not follow known patterns of access to primary care for the general population. The final dissertation chapter provides a synthesis and some policy recommendations to address the data gaps that served as the impetus for these studies.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Defining neighborhood for health research in Arizona

PDF

Modeling patient access to point-of-care diagnostic resources in a healthcare small-world network in rural Isaan, Thailand

PDF

Estimating populations at risk in data-poor environments: a geographically disaggregated analysis of Boko Haram terrorism 2009-2014

PDF

Characterizing self-reported spatial and provider barriers to maternal health care utilization in Malawi

PDF

To what extent can wicked problems of health and place be addressed through clumsy solutions? Three experiments in reconstituting networks of spatial knowledge production

PDF

Understanding the role of population and place in the dynamics of seasonal influenza outbreaks in the United States

PDF

Homeless encampments and access to Water, Sanitation, and Hygiene (WaSH) services in Los Angeles, CA

PDF

Enriching the Demographic Survey sampling for the Los Angeles County Annual Homeless Count with spatial statistics

PDF

Unsettled lives: Medicine, culture, and transgender subjects

PDF

Integrating spatial visualization to improve public health understanding and communication

PDF

A spatial analysis of veteran healthcare accessibility

PDF

Mapping future population impacts caused by sea level rise in Huntington Beach and Newport Beach: comparing the cadastral-based dasymetric system to past dasymetric mapping methods

PDF

Personal PM2.5 exposure during pregnancy in an environmental health disparities population

PDF

Use of GIS for analysis of community health worker patient registries from Chongwe district, a rural low-resource setting, in Lusaka Province, Zambia

PDF

Spatiotemporal studies of out-of-hospital cardiac arrests and bystander cardiopulmonary resuscitation in Los Angeles

PDF

Application of geospatial methods to cancer surveillance data to improve cancer prevention and control

PDF

Transgender patients’ perceptions of healthcare: A study of gender minority stress and resilience factors in predicting healthcare behavioral intentions

PDF

Health care utilization and spending of the U.S. aging population

PDF

Population substructure and its impact on genome-wide association studies with admixed populations

PDF

Spatial analysis of vision services of Kaiser Permanente members

Asset Metadata

Creator Everhart, Avery Rose (author)

Core Title Incomplete data & insufficient methods: transgender population health research in the US

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Health and Place,Population

Degree Conferral Date 2022-08

Publication Date 07/20/2022

Defense Date 05/18/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag critical data studies,critical GIS,demography,Geographic Information Science,health geography,medical geography,OAI-PMH Harvest,population health,spatial demography,spatial epidemiology,transgender,transgender health,transgender studies

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ferguson, Laura (committee chair), Wilson, John (committee chair), De Lara, Juan (committee member), Olson-Kennedy, Johanna (committee member), Restar, Arjee (committee member)

Creator Email avery.everhart@inbox.com,everhart@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111373253

Unique identifier UC111373253

Legacy Identifier etd-EverhartAv-10874

Document Type Dissertation

Format application/pdf (imt)

Rights Everhart, Avery Rose

Type texts

Source 20220720-usctheses-batch-957 (batch), 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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

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

Repository Email cisadmin@lib.usc.edu