Racial and Socioeconomic Inequity in the Spatial Distribution of LGBTQ Human Services: an Exploratory Analysis of LGBTQ Services in Chicago

LGBTQ people of color and low-socioeconomic status face a heavy burden of health, economic, and social disparities. Understanding the factors that hinder or facilitate efforts to address these disparities is critical to achieving equity and improving social welfare. This study explores one potential factor: the spatial distribution of LGBTQ human services. A spatially referenced dataset of 193 LGBTQ human service sites in the Chicago area was compiled between November 2015 and January 2016. Sites were geocoded and mapped in ArcGIS along with data from the 2011–2015 American Community Survey. Geospatial techniques were used to assess racial and socioeconomic patterns in service distribution. Analyses revealed that sites were disproportionately located in majority non-Hispanic/Latino White (71.0%) and upper-income (63.7%) block groups. Additionally, analyses revealed that Black/African American, Hispanic/Latino, and poor to low-income residents of Chicago disproportionately live in LGBTQ service deserts. These patterns in the distribution of LGBTQ services may present barriers to addressing disparities by making services less accessible for racial/ethnic minorities and lower-income individuals. System- and structural-level interventions are needed to reduce spatial inequities in the distribution of LGBTQ human services in Chicago in order to facilitate efforts to address LGBTQ disparities among racial/ethnic minorities and individuals with lower socioeconomic status.

Despite the particular challenges facing LGBTQ people of color and of low SES, previous literature suggests that many LGBTQ advocacy and human service organizations prioritize White, middle-to upper-class subgroups, often to the detriment of their more marginalized counterparts (DeFilippis, 2016;Mananzala & Spade, 2008;Ward, 2008;Cohen, 1999). Several scholars have argued that this prioritization stems from pervasive social perceptions of gay identity as White and welloff, as well as political strategies that value assimilation and conformity over liberation and redistributive justice (DeFilippis, 2016;Hollibaugh & Weiss, 2015;Mananzala & Spade, 2008;Ward, 2008;Berube, 2001;Cohen, 1999). Within human service settings, this dynamic of centering whiteness and affluence may manifest through disparities in organizational staffing (Smith, 2015). For instance, a 2015 survey of major LGBTQ and HIV/AIDS service providers in Chicago found that among organizations reporting staff and client demographics, 89% had majority White staff despite 78% serving majority non-White clientele (Smith, 2015). Some LGBTQ organizations' priorities and rhetoric have also reflected a prioritization of White, affluent populations. In an ethnographic study of three queer organizations in Los Angeles, Ward (2008) demonstrated that LGBTQ organizations primarily prioritized White and affluent populations by enforcing norms of professionalism that gave these stakeholders more influence over the governance of the organization. These LGBTQ organizations also used racial and socioeconomic diversity as a marketing tactic to attract funders and public support while still maintaining a White-normative, corporate culture that failed to meaningfully incorporate the language, customs, and priorities of racial/ethnic minority and low-SES groups (Ward, 2008). Finally, these organizations required marginalized groups to compete for resources by presenting their identities as most Bat-risk^and in need of funding, creating tensions within and between these groups (Ward, 2008). Overall, these dynamics may make LGBTQ human services less accessible and responsive to the needs of racial/ethnic minority or low-SES groups, particularly groups of LGBTQ, low-SES individuals of color who face intersecting forms of sexual, gender, race, and class marginalization (Irazábal & Huerta, 2016). Consequently, these patterns of prioritization may present barriers to addressing disparities in health, economics, and other areas of social welfare among these populations.
The spatial distribution of LGBTQ human services is another dynamic that may contribute to inequitable racial/ethnic and socioeconomic prioritization, as well as impede efforts to address disparities in health and social outcomes. As sociospatial dialectic theory suggests, spatial and social dynamics are mutually formative and inter-reactive (Soja, 1980). In this sense, the spatial organization of LGBTQ human services may be expected to both reflect and perpetuate broader social dynamics of racial/ethnic and socioeconomic prioritization in LGBTQ organizations. More specifically, inequity in spatial access to these services across race and class lines may manifest by mirroring other inequities seen in LGBTQ organizations; in turn, these inequities in access may hinder the ability of these service providers to reach higher need populations and address disparities (Wilson & Yoshikawa, 2007). Additionally, racial and socioeconomic inequity in LGBTQ services' spatial context (i.e., the sociodemographics of areas in which services are situated) may also manifest as a reflection of social dynamics of prioritization; dynamics such as neighborhood racial and class exclusion, which are related to a service's spatial context, may reinforce social inequity by presenting barriers to individuals utilizing services and receiving quality care (Doan, 2015). In line with this theory, some scholars have noted that LGBTQ services are often located in or near gayborhoods (neighborhoods in urban settings with a preponderance of LGBTQ residents and establishments), which are generally predominantly White and affluent (Doan, 2015;Wilson & Yoshikawa, 2007;Mills et al., 2001). Although these gayborhoods have a greater number of LGBTQ residents, previous research suggests that LGBTQ individuals of color are less likely to live in these neighborhoods than their White peers and instead are more likely to live in racial enclaves (Mills et al., 2001). Additionally, recent work suggests that LGBTQ people are decreasingly living in gayborhoods over time as social acceptance has increased (Ghaziani, 2014). For these reasons, it is increasingly important to examine spatial access and spatial context of LGBTQ services across a city's entire landscape rather than solely focusing on gayborhoods.
Despite the role the spatial distribution of LGBTQ human services may play in efforts to address disparities, this factor has remained underexplored in the literature. However, a number of previous studies have examined the spatial distribution of HIV services. For example, a study by Kaukinen and Fulcher (2006) found that neighborhoods in Toronto with a high proportion of immigrants, Black Canadians, and socioeconomically disadvantaged populations had lower levels of access to HIV-related services, despite these populations facing a disproportionately high burden of HIV. Similarly, a study of HIV prevention services in Chicago found that the areas in which young Black men who have sex with men (MSM) typically reside tend to have low densities of HIV prevention services, even though young Black MSM face some of the highest rates of HIV transmission (Pierce, Miller, Morales, & Forney, 2007). Lastly, a zip code-level analysis of Miami-Dade County found that there was a lack of Ryan White HIV/ AIDS service providers in a predominantly Latino immigrant part of the county, despite areas within that region having a high incidence of HIV/AIDS (Ganapati, Ganapati, De La Rosa, & Rojas, 2010).
However, HIV services represent only one type of human service need for LGBTQ communities, and to our knowledge, no prior research has explored whether similar spatial patterns exist with other types of services. Additionally, not all HIV/ AIDS services are specifically tailored to LGBTQ communities. Thus, additional research examining the spatial distributions of a comprehensive range of LGBTQ human services is sorely needed. This exploratory study aims to address this gap in the extant literature by using geographic information system (GIS) methods to explore the spatial distribution of LGBTQ human services in the Chicago area. Specifically, this study examines racial and socioeconomic patterns in both spatial context and spatial access related to LGBTQ human services in and near the city of Chicago. For the purpose of this study, spatial context refers to the sociodemographic makeup of services' surrounding areas, while spatial accessibility refers to the proximity of services to various sociodemographic groups.

Data
A spatially referenced dataset of LGBTQ human services sites was compiled and verified between November 2015 and January 2016, which was the length of time required to collect and validate organizational data. One member of the research team initially identified potential sites for inclusion in the sample using two online, publically accessible directories of LGBTQ services in the Chicago area (Center on Halsted, 2015;Adler University LGBTQ Mental Health & Inclusion Center, 2015). The first directory was maintained by staff at Center on Halsted, an LGBTQ community center in Chicago. The second was maintained by staff at Adler University, though representatives of service providers had the ability to submit and update listings to this directory themselves. Both directories were comprehensive, each containing over a hundred listings and including a wide range of programs and organizations; some listed organizations were relatively small, such as a transgender support group that served around a dozen individuals, while others were large multi-site health providers that offered wrap-around services to thousands of patients with a range of identities. To begin compiling the dataset, a member of the research team extracted the following data for each listing: organization name, address, and types of LGBTQ-focused services offered. Services were classified into seven types: (1) recreational and arts programs (e.g., youth groups, mindfulness and yoga, dance workshops), (2) social support (e.g., support groups, mentorship, identity-based discussion groups), (3) medical services (e.g., HIV/STI testing and treatment, primary care, transition-related hormones and surgeries), (4) mental health (e.g., therapy, counseling, psychiatry, crisis intervention), (5) housing and basic needs (e.g., shelter, food, hygiene supplies, clothing), (6) career development and employment (e.g., GED programs, resume support, career counseling), and (7) legal services (e.g., consultation, adoption-and marriage-related services, assistance changing name/gender marker). Verification of the extracted data then proceeded in two steps. First, a member of the research team cross-checked the data with information listed on organizations' websites when available. Next, a member of the research team contacted organizations via phone or site visits to verify the data collected from the directories and websites. Three contact attempts were made for this step. When discrepancies in the data provided by these sources occurred, data from phone calls and visits were used if available, followed by website data and, lastly, by directory data. This process was used because organizational representatives were considered to be a more reliable source of up-to-date data than websites while websites were considered more reliable than the directories. Sites were determined to be non-operational if the website or an organizational representative stated they had closed. Sites were also classified as non-operational if their website was offline or had not been updated in the past 6 months and a successful contact could not be made. For multi-site organizations, address and service data were collected for each individual site. Inclusion criteria for the analytic dataset were as follows: located in Chicago or within 3 miles of the city border, currently operational, and offer not-forprofit services targeted or tailored to LGBTQ communities in at least one of the seven categories listed previously. Sites outside the analytic region were excluded because they could not be easily accessed by residents of the city of Chicago. Forprofit sites and sites with programming outside of the seven service categories were excluded because they were generally businesses, political organizations, or activist groups rather than human service providers.
Additional data utilized in analyses included the following: racial/ethnic and household income estimate data from the 2011-2015 American Community Survey (ACS), retrieved from Social Explorer (US Census Bureau, 2016a); block group shapefiles retrieved from TIGER/Line ® (US Census Bureau, 2016b); and street and boundary shapefiles retrieved from the City of Chicago Data Portal and Cook County Data Catalog (City of Chicago, 2016;Cook County Government, 2016). For the analysis of spatial context, block groups were categorized by majority racial/ethnic group and median household income according to the ACS. Block groups were categorized as majority White if over 50% of residents in the ACS were categorized as White alone (non-Hispanic/Latino), majority Black/African American if over 50% were Black or African American alone (non-Hispanic/Latino), majority Asian if over 50% were Asian alone (non-Hispanic/Latino), majority Hispanic/Latino if over 50% were Hispanic/Latino, and no majority racial/ethnic group if none of these groups made up over 50% of the area's residents. Median income was used to categorize block groups as poor ($24,260 or less), lower-income ($24,261-$38,817), middle-income ($38,818-$58,225), and upper-income ($58,226 or more). These block group-level income categories were calculated based on Chicago's 2015 median household income ($48,522), with poor defined as a median household income below 50% of Chicago's, lower-income as a median household income between 50 and 80% of Chicago's, middleincome as a median household income between 80 and 120% of Chicago's, and upper-income as a median household income above 120% of Chicago's (US Census Bureau, 2016a). All income data is presented in 2015 inflationadjusted dollars. Analysis of spatial access relied on individual-level racial/ethnic income data rather than aggregated block-group data. Individual-level racial/ethnic groups for analysis included White (non-Hispanic/Latino), Black/ African American (non-Hispanic/Latino), Asian (non-Hispanic/Latino), and Hispanic/Latino. Individual-level income groups for analysis included poor to low-income and middle-to upper-income. Residents were defined as poor to low-income if their households earned less than 200% of the poverty threshold (e.g., under $23,540 for a single adult, under $31,860 for two adults, under $40,180 for a family of four) and middle-to upper-income if they earned more than 200% of the poverty threshold. These income classifications were based on the 2015 Federal Poverty Guidelines from the US Department of Health and Human Services (2015). Again, all income data is presented in 2015 inflation-adjusted dollars.

Geovisualization and Analysis of Spatial Context
Service sites were first geocoded in ArcGIS 10.3 (Redlands, CA) using address data. This provided a geographic depiction of the service sites, allowing the research team to characterize their spatial distribution within the urban landscape. To examine the spatial context of these services (i.e., the sociodemographic makeup of services' surrounding areas), the geocoded sites were overlaid onto the block group-level racial/ethnic and median household income data from the 2011-2015 ACS, producing two maps. Block groups were utilized because analyses of metropolitan areas across the US have found these units to be more sensitive to local patterns of segregation than other spatial units such as census tracts (Wong, 2004). Next, ArcGIS's intersect tool was used to count the number of service sites located in block groups of each racial/ethnic and household income category. Onesample z tests of proportions were then conducted to examine whether LGBTQ services were proportionately or disproportionately located in block groups of each racial/ethnic or household income category. The null hypothesis for these z tests was that LGBTQ services in Chicago are distributed proportionally by block group-level race/ethnicity and household income (i.e., the percentage of services in block groups of a given racial/ethnic or income category is equivalent to the percentage of block groups of that category in or near Chicago). For this analysis, ArcGIS's intersect tool was used to calculate the percentage of block groups in Chicago or within 3 miles of the city border that fell into each racial/ ethnic or income category. Similar methods relying on z tests of proportions have been employed in other geographic studies (Swak, Hangen, & Northrup, 2014;Zandbergen & Chakraborty, 2006).

Geovisualization and Analysis of Spatial Access
To examine spatial access (i.e., the proximity of services to various populations), service areas and deserts for the sites were constructed using ArcGIS's network analyst tool and street data from the City of Chicago Data Portal. For this analysis, routes of equal distance are drawn beginning from the site location and moving along street vectors until the set distance has been reached. Next, endpoints of these routes are connected to produce polygons, which represent the service areas. ArcGIS's intersect tool can then be used to construct service desert polygons (areas outside the service area) by removing the service area polygons from a polygon of the city. This use of network analysis is a common procedure for constructing service areas and deserts in spatial investigations (Curtin, 2007). For this analysis, areas within a 3-mile driving distance of a site were defined as being within that site's service area. This 3-mile distance has been used as a standard measure of access in a number of studies examining the service areas of various urban amenities, though most other studies use a 3-mile radius rather than driving distance (Hertel & Sprague, 2007;Guagliardo, 2004;Mehta, Baiman, & Persky, 2004). This study utilized driving distance because it more realistically reflects the path an individual would need to travel in order to reach a service, taking into account geographic barriers such as street layout, rivers, parks, and deadends (Eberhart, Share, Shpaner, & Brady, 2014). Next, maps were produced depicting the service areas and deserts (areas outside a 3-mile driving distance) for each of the seven service types. A composite map overlaying these service areas was also produced to visually represent varying levels of service access across the city. To examine demographic differences in access, the service desert polygons for each of the seven service types and the composite were overlaid with block grouplevel racial/ethnic and household income data from the 2011-2015 ACS. Next, ArcGIS's intersect tool was used to calculate estimated proportions of Chicago residents living in each service desert by racial/ethnic and income groups. For block groups completely contained within a given desert, all block group residents were counted as living within that desert. For block groups partially overlapping with a given desert, the percentage of overlapping area was calculated and this percentage was multiplied by the numbers of residents within the block group to produce estimated counts of those living within the desert. To produce estimated proportions of Chicago residents living within the deserts by race/ethnicity and income, counts for all the intersecting block groups were then summed and divided by the total number of Chicago residents for each racial/ethnic and income group.
Given that the protocol and procedures did not include human subjects for research and relied on publically available data, this exploratory study was deemed exempt from institutional review board review.

Summary of LGBTQ Service Sites
Identification of LGBTQ services through the online directories resulted in an initial list of 326 potential sites, of which 62 were excluded from the dataset because they did not offer services targeted or tailored to LGBTQ communities in at least one of the seven categories (recreational and arts programs, social support, medical, mental health, housing and basic needs, career development and employment, and legal). Another 41 service sites were excluded because they were not located in Chicago or within 3 miles of the city border, and 19 sites were excluded because their LGBTQ services were for profit. Lastly, 11 sites were excluded because they were deemed not operational at the time of data collection. This yielded a final analytic dataset of 193 service sites. Of the 193 services sites, 66.3% offered recreational and arts programs, 35.8% offered social support, 19.2% offered medical services, 18.1% offered mental health services, 13.0% offered housing and basic needs, 13.0% offered career development and employment services, and 10.4% offered legal services (see Table 1).

Spatial Context of LGBTQ Service Sites
Geocoding revealed that service sites were predominantly located in three main areas (see Figs. 1 and 2). First, a high proportion of services were located in Chicago's northeastern neighborhoods along Lake Michigan. This area includes Chicago's two gayborhoods, East Lakeview (also colloquially known as Boystown) and Andersonville. Second, there were a significant number of sites in the Loop, Chicago's downtown center which is located midway along the city's lakeshore. Third, a number of services were located in or near Chicago's Hyde Park neighborhood, an area in the mid-South Side along Lake Michigan where the University of Chicago and University of Chicago Medical Center are located. Two bordering suburbs also housed a number of service sites. These included Evanston, a suburb to the north of the city that is home to Northwestern University, and Oak Park, a suburb to the west of the city.

Spatial Accessibility of LGBTQ Service Sites
Mapping service areas and deserts revealed varying levels of LGBTQ service access across Chicago's landscape (see Fig. 3). Access was highest in the eastern part of the city along the lakeshore, with much of this region within the service areas for all seven service types. Access was lowest on the far south and southwest parts of the city, with much of these regions within the service deserts for all seven service types. Access was also low on the far west and northwest parts of the city, with much of these regions only within the service areas for one to three service types.
Demographic analysis of LGBTQ service deserts revealed racial/ethnic disparities in access to LGBTQ services and particularly low levels of access for Black/African American and Hispanic/Latino residents (see Table 4). A substantially higher proportion of Chicago's Hispanic/Latino and Black/African American residents live in LGBTQ service deserts than non-Hispanic/Latino White and Asian residents. Overall, an estimated 28.7% of Black/African American and 31.9% of Hispanic/Latino residents live in complete service deserts (areas more than a 3-mile driving distance from any type of LGBTQ service site), compared to an estimated 9.9% of White and 7.7% of Asian residents. Following this pattern, racial/ethnic disparities in access of more than 5% were also seen across all seven service types, though it was most pronounced among legal and career development and employment services. An estimated 87.2% of Black/African American and 79.1% of Hispanic/Latino residents live in LGBTQ legal service deserts compared to 43.7% of White and 46.7% of Asian residents. Similarly, an estimated 72.5% of Black/African American and 69.0% of Hispanic/Latino residents live in LGBTQ career development and employment service deserts compared to 43.8% of White and 31.1% of Asian residents. Although levels of racial/ethnic disparity in access varied by type of service, access for Black/African American and Hispanic/Latino residents was low for all types of services. All seven of the service desert types included between 36.8 and 87.2% of the city's Black/African and Hispanic/Latino populations.
Demographic analysis of LGBTQ service deserts also revealed socioeconomic disparities in access to LGBTQ services and particularly low levels of access for poor to low-     Upper-income block groups are defined as those with a median household income more than 120% of Chicago's median household income ($48,522) income residents, though socioeconomic disparities in access were not as pronounced as racial/ethnic disparities (see Table 4). A higher proportion of Chicago's poor to lowincome residents live in LGBTQ service deserts than middle-to upper-income residents. Overall, an estimated 26.8% of poor to low-income residents live in complete service deserts compared to an estimated 18.9% of middle-to upper-income residents. Following this pattern, socioeconomic disparities in access of more than 5% were also seen for six of the seven service types: recreational and arts programs, social support, mental health, housing and basic needs, career development and employment, and legal. The pattern was most pronounced among legal services and recreational and arts programs. An estimated 77.6% of poor to low-income residents live in LGBTQ legal service deserts compared to 60.9% of middle-to upper-income residents. Similarly, an estimated 35.3% of poor to low-income residents live in LGBTQ recreational and arts programs service deserts Fig. 3 Chicago LGBTQ service areas and deserts. Service areas are defined as the area within a 3-mile driving distance of services. Service deserts are defined as the area outside a 3-mile driving distance of services compared to 23.4% of middle-to upper-income residents. The pattern of socioeconomic disparity in access was not seen for medical services, for which there were fairly comparable proportions of poor to low-income (47.6%) and middle-to upperincome (45.1%) residents living in service deserts. Although levels of socioeconomic disparity in access varied by type of service desert, access for poor to low-income residents was low for all types of services. All seven of the service desert types included between 35.3 and 77.6% of the city's poor to low-income populations.

Discussion
Overall, the findings of this exploratory study suggest that the spatial distribution of LGBTQ services in Chicago may be racially and socioeconomically inequitable. These inequitable patterns have the potential to present barriers to addressing disparities in social, health, and economic outcomes among racial minority and low-SES subsets of the LGBTQ community. First, our analysis of spatial context found that nearly three quarters of LGBTQ service sites in Chicago were located in majority White block groups and nearly two thirds were located in upper-income block groups (median household income ≥ $58,266). For both race and income, this was nearly twice the expected proportion assuming the sites had been evenly distributed across block groups. Furthermore, analysis of spatial access suggests that Hispanic/Latino and Black/African American residents are roughly three to four times more likely to live in LGBTQ service deserts compared to White and Asian residents. Approximately 30% of the Chicago's Black/African American and Hispanic/Latino residents live in a complete LGBTQ service desert, putting them more than a 3-mile driving distance from a single LGBTQ service site, compared to under 10% of the city's White and Asian residents. Similarly, poor to low-income residents were nearly one and a half times more likely to live in LGBTQ service deserts compared to middle-to upper-income residents; complete service deserts contain a little over a quarter of poor to low-income residents but under a fifth of middle-to upper-income residents. Given the dialectic nature of social and spatial relationships (Soja, 1980), it is possible that these patterns of inequity are, at least in part, a spatial manifestation of LGBTQ service providers' broader prioritization of White, affluent populations (DeFilippis, 2016; Mananzala & Spade, 2008;Ward, 2008;Cohen, 1999). However, it is also possible that these inequities are further reinforced by more mundane concerns such as organizations' facility needs or connections to networks of other service providers. Future research should explore the factors that influence LGBTQ organizations' decision-making around location and contribute to spatial inequity. Furthermore, the patterns of spatial inequity in access seen in this study also have meaningful implications for addressing disparities among the LGBTQ population. Although our analysis examined access to LGBTQ services for the entire population of Chicago due to a lack of existing data on LGBTQ residency patterns, it is likely similar patterns of inequity exist for the LGBTQ subpopulation of the city. Previous research suggests that LGBTQ people of color are less likely to live in gayborhoods and more likely to live in racial enclaves, many of which were located in service deserts in this study (Mills et al., 2001). Furthermore, the literature suggests that LGBTQ individuals are becoming more evenly distributed across the city landscape over time as social acceptance increases (Ghaziani, 2014), and LGBTQ adolescents may already be more evenly distributed due to generally residing with their families. Thus, the racial and socioeconomic inequity in the spatial accessibility of LGBTQ services seen in this study may present many LGBTQ individuals of color and low SES with longer travel, which could impede efforts to address disparities. Indeed, previous studies have demonstrated that increased distance and travel time to health services were associated with worse health outcomes, including HIV outcomes (Leibowitz & Taylor, 2007;Bauermeister et al., 2015;Ridgway, Almirol, Schmitt, Schuble, & Schnedier, 2017). Furthermore, previous studies have found that proximity to mental health care was associated with higher rates of antidepressant adherence, lower rates of hospitalization, and fewer missed appointments (Gonzalez, Williams, Hitchcock Noël, & Lee, 2005;Fortney, Owen, & Clothier, 1999;Campbell, Staley, & Matas, 1991). Thus, the inequity in LGBTQ medical and mental health service access seen in this study may present barriers to addressing disparities in health outcomes documented in other research (CDC, 2017a, b;Frost et al., 2016;James et al., 2016;Bouris et al., 2015;Habarta et al., 2015;Bostwick, Boyd, et al., 2014a;Denning & DiNenno, 2010;Meyer, 2008).
Although the impact of proximity to health services has been previously explored in the literature, the impact of proximity to other types of services is less established. Nevertheless, given that long travel can impose a significant barrier to service utilization, the racial and socioeconomic inequity in access seen across other service types examined in this study may impede efforts to address disparities in social and economic outcomes among the LGBTQ population (Allard, 2009). The large racial and socioeconomic disparities in access to legal services as well as career development and employment services are particularly striking, given that Black/African American, Hispanic/Latino, and low-SES LGBTQ people face high rates of criminal-legal system involvement and disparities in employment and wages (Meyer et al., 2017;James et al., 2016;Lambda Legal, 2015;Douglas & Steinberger, 2014;Kastanis & Wilson, 2014;Mogul et al., 2011). Future research should examine the relationship between outcomes among the LGBTQ population and proximity to LGBTQ non-medical services. Although longer travel likely serves a deterrent to seeking and accessing services for most LGBTQ individuals, it is possible that some prefer to travel outside their residential areas or access non-LGBTQ-specific services in their neighborhoods due to confidentiality concerns (Beach et al., 2018). Future research exploring this dynamic would be valuable as it could allow more complex measures of access to be developed that take into account traveling to maintain confidentiality. Additionally, future research should examine whether service utilization and outcomes among the LGBTQ population differ based on proximity to LGBTQ services versus proximity to comparable services that are not tailored to LGBTQ populations.
Inequitable racial and socioeconomic patterns in the spatial context of LGBTQ services also have implications for addressing disparities among the LGBTQ community, as these organizations' surrounding environments may impact the experiences of individuals utilizing their services. Due to the low numbers of LGBTQ services in communities of color and poor to middle-income areas, racial/ethnic minority and low-SES LGBTQ residents in Chicago may often need to travel to majority White, affluent neighborhoods to utilize needed services. Indeed, previous research has suggested that many young gay and bisexual men of color are spatially polygamous, meaning that they regularly travel outside their residential neighborhoods (Duncan, Kapadia, & Halkitis, 2014). Seeking LGBTQ services may be one potential reason these individuals leave their neighborhoods, a possibility that warrants further consideration in future studies of mobility. In navigating services located in majority White, upper-income areas, LGBTQ individuals of color and low SES may also have to negotiate unfamiliar environments and experience profiling, negative interactions, and violence from police and residents, which can serve as barriers to obtaining needed services (Johnson, St. Vil, Gilbert, Goodman, & Arroyo Johnson, 2019;Levesque, Harris, & Russell, 2013;Stewart, Baumer, Brunson, & Simon, 2009;Shannon et al., 2008). For instance, a recent community-participatory study explored the experiences of LGBTQ youth of color accessing services in Boystown, one of two majority White, upper-income gayborhoods located in northeastern Chicago where there is a preponderance of LGBTQ services. This study found that residents of Boystown often treated LGBTQ youth of color negatively by making exclusionary comments, spreading rumors, and policing their behavior (Felner, Dudley, & Ramirez-Valles, 2018).
Additionally, as individuals of color and low-SES travel to majority White, affluent neighborhoods in pursuit of services, they may be placed into the crossfire of larger neighborhoodlevel racial and socioeconomic conflicts. For instance, in 2011, a neighborhood conflict erupted in Boystown after a high-profile instance of violence (Felner et al., 2018;Greene, 2014;Daniel-McCarter, 2012). In reaction to the incident, a group of neighborhood residents started a Facebook page entitled, BTake Back Boystown,^which called for heightened policing and the exclusion of people of color, homeless individuals, and transgender women from the neighborhood (Felner et al., 2018;Greene, 2014;Daniel-McCarter, 2012). Consequently, racial minority and low-SES individuals traveling to Boystown in pursuit of services found themselves at the center of a cross-community debate, having to lay claim to the neighborhood in order to maintain access to needed services (Felner et al., 2018;Greene, 2014;Daniel-McCarter, 2012). Overall, these dynamics may negatively impact the experiences of racial minority and low-SES individuals seeking LGBTQ services and discourage service seeking and utilization among these groups (Felner et al., 2018). These conflicts may also expose LGBTQ individuals of color and low SES to heightened minority stress, which can have negative impacts on their well-being (Meyer, 2013). The impact of these conflicts may be particularly great for LGBTQ individuals who are both of color and low SES, as they face multiple, intersecting forms of marginalization (Irazábal & Huerta, 2016;Meyer, 2013). Ultimately, racial and socioeconomic conflicts at the neighborhood level may be exacerbated by inequitable patterns in the spatial context of services, thus presenting a barrier to addressing disparities. Future studies should explore how patterns in the spatial context of LGBTQ services directly impact neighborhood conflict as well as service utilization and client experiences among racial minority and low-SES groups.
The racial and socioeconomic patterns in the spatial context of LGBTQ services may also influence the priorities and practices of LGBTQ service providers in ways that have negative repercussions for racial minority and low-SES groups. Human service organizations generally cater to the interests and concerns of a number of competing stakeholders, including surrounding neighborhood residents and community groups (Van Puyvelde, Caers, Du Bois, & Jegers, 2012). Because of inequitable patterns in the distribution of services, neighborhood stakeholders of LGBTQ service organizations in Chicago are, on the whole, Whiter and wealthier than the general city population. Thus, LGBTQ organizations may disproportionately tailor their services and messaging to White, middle-to upperclass residents, potentially contributing to dynamics of white normativity and corporatization (Ward, 2008).
LGBTQ service organizations may also find themselves having to negotiate neighborhood-level racial and socioeconomic conflicts due to their spatial context, impacting the quality of care they provide to clients. For instance, many LGBTQ human service organizations in Boystown were implicated in the BTake Back Boystown^debate, as some residents called for these organizations' closure based on the belief that they were attracting crime (Daniel-McCarter, 2012). Consequently, many of these organizations were compelled to respond to residents' concerns by increasing surveillance and instituting stricter disciplinary policies such as banning clients for small infractions (Felner et al., 2018;Rosenberg, 2017). These approaches may negatively impact racial minority and low-SES subsets of the LGBTQ community by contributing to criminalization of these populations, diverting organizational resources from service provision to surveillance, and creating an unwelcoming environment (Felner et al., 2018;Rosenberg, 2017). Furthermore, these dynamics may also restrict racial minority and low-SES subsets of the LGBTQ community from utilizing services, impeding efforts to address disparities. Future research should examine how the spatial context of LGBTQ human service organizations influences organizational priorities and practices and the potential impact this has on efforts to address disparities.
The findings of this exploratory study also have a number of implications for interventions in community and human service planning. First and foremost, these data suggest that substantially greater investment in LGBTQ services may be needed in Black/African American, Hispanic/Latino, and poor to lower-income communities, particularly in LGBTQ legal, career and employment, and mental health services. Mobile interventions offering portable and convenient services that travel to sites across the city may be one possibility for expanding LGBTQ service access, as services targeted to LGBTQ populations could be integrated into programs offering HIV/STI testing and counseling and other sexual and reproductive health services in mobile clinics (Ellen, Bonu, Arruda, Ward, & Vogel, 2003;O'Connor, Pastdaughter, Gatson Grindel, Taveira, & Steinberg, 1998;Henrickson, 1990;Stefansson, Webb, Hebert, Masinter, & Gilliam, 2018). In addressing issues of LGBTQ service access for these groups, interventions may also be needed to improve transportation access (e.g., providing transit passes, small transit stipends, or shuttles, taxis, and rideshares), as previous research suggests that this may improve access and encourage greater service utilization (Fraze, Lewis, Rodriguez, & Fisher, 2016;George et al., 2009;Friedmann, D'Aunno, Jin, & Alexander, 2000). Although intervention should primarily focus on addressing racial and socioeconomic inequity in the distribution of services and expanding access, interventions may also be needed within White, upper-income areas that already have a preponderance of LGBTQ services. Community actors in these areas (e.g., organizations, residents, local political leaders) should engage in community-wide initiatives designed to foster cohesion, build solidarity, and mitigate conflicts between racial/ethnic groups, socioeconomic groups, and the residents and non-residents utilizing LGBTQ services in the neighborhood. Possible neighborhood-level initiatives could include educational programming, public events, arts initiatives, leadership programs, community driven-planning and development, community organizing and mobilization, and coalition-building (Glenn et al., 2018;Shen et al., 2017;Irazábal & Huerta, 2016;Molitor, Rossi, Branton, & Field, 2011;Jeannotte, 2010;Eisen, 1994).
Despite the insights this exploratory study provides into the spatial distribution of LGBTQ services in Chicago, it is not without limitations. First, this study identified LGBTQ human service sites using two comprehensive online directories, so it is possible service providers that are less connected to the networks which maintain the directories were not included in the analytic sample. Given that the distribution of services listed in the directories may differ from those that were not listed, the patterns of spatial inequity seen in our analysis may be somewhat under-or overstated. Nevertheless, the directories we used were fairly comprehensive resulting in a robust analytic sample of nearly 200 sites. Furthermore, although we used a three-step process to collect and verify data for the service sites, it is possible that some of the data provided were inaccurate or outdated resulting in misclassification of the sites. Our study also utilized service categories that were relatively broad (e.g., medical included primary care, HIV/STI testing, and transition-related care), and many sites in the dataset only serve specific subsets of the LGBTQ community (e.g., youth, MSM, and transgender populations). Furthermore, this analysis did not examine the quality of care, funding, or capacity of organizations, which may also impact access and utilization. It may be valuable for future studies to categorize LGBTQ services more specifically and to examine service areas according to the subsets of the LGBTQ population targeted. Follow-up studies may also assess and account for quality of care and organization size.
Additionally, due to the exploratory nature of the study and limitations of the dataset, our analysis relied on descriptive statistics and z tests. Larger follow-up studies should incorporate regression analyses and account for spatial autocorrelation. Our analysis of spatial context also relied on block groups, which, like any analysis utilizing one type of geographic unit, is subject to the modifiable areal unit problem, a potential source of statistical bias resulting from how boundaries are drawn (Fotheringham & Wong, 1991). Furthermore, construction of service areas and deserts in this study was based on a 3-mile driving distance. However, many individuals accessing services are likely walking or using public transit. Moreover, appropriate service area sizes may differ by the types of services offered. Future research may incorporate more complex models that utilize multiple forms of transportation in measuring access. Although this study utilized race and income data from the ACS, similar data on sexual orientation and gender identity are not available. Thus, while residency patterns differ between LGBTQ people and their heterosexual, cisgender counterparts, this was not a factor that could be incorporated into analyses. The addition of questions assessing sexual orientation and gender identity to the ACS would be valuable and allow for more robust analyses that consider LGBTQ residency patterns. Given previous research demonstrating spatial polygamy among LGBTQ populations, future research should incorporate mobility as well as residential data from LGBTQ populations, potentially through the use of global positioning system (GPS) technology (Duncan, Kapadia, et al., 2014;Duncan, Chaix, et al., 2018). Finally, because urban landscapes differ from city to city, the results of this study are not generalizable to other cities or areas. Nevertheless, this exploratory study highlights the potential role that spatial inequity in LGBTQ services may play in impeding efforts to address disparities, suggesting that research examining the distribution of these services in other cities and areas would be valuable.

Conclusion
Overall, this exploratory study suggests that the Chicago's LGBTQ human services are inequitably distributed across the city's geography. Although racial minority and low-SES subsets of the LGBTQ community face heavy burdens of health, economic, and social disparities, LGBTQ human service sites were predominantly located in majority White and upper-income areas. Additionally, Black/African American, Hispanic/Latino, and poor to lower-income individuals had lower levels of spatial access compared to their respective counterparts (White, Asian, and middle-to upper-income residents). Due to neighborhood dynamics, the inequitable patterns in the spatial context of these services may impact the provision and utilization of these services in ways that impede efforts to address disparities. Additionally, inequitable patterns in spatial access may restrict racial minority and low-SES members of the LGBTQ community from consistently reaching needed services and thus present an additional barrier to addressing disparities. Future system-level interventions in Chicago are needed to address inequity in the spatial distribution of LGBTQ services, expand LGBTQ service access for racial minority and low-SES groups, and address exclusionary neighborhood dynamics in areas with a high density of LGBTQ services. Additionally, future research should examine how the spatial context of LGBTQ services impacts provision, utilization, and client experiences, as well as how spatial access to LGBTQ services impacts outcomes among racial minority and low-SES populations. Future studies examining the spatial distribution of LGBTQ services in other cities and areas would also be valuable in order to better understand the distribution of LGBTQ services and improve efforts to address health, economic, and social disparities nationally.