Spatial accessibility for deploying mobile lung cancer screenings in a rural state using two-step floating catchment spatial analysis in Oklahoma, United States 2025

Abstract

Objective

Lung cancer remains the leading cause of cancer death in the United States. Mobile screening units have reduced access barriers. To assess lung cancer screening spatial access, we utilized a two-step floating catchment analysis (2SFCA) and an enhanced two-step floating catchment analysis (E2SFCA) to identify areas for mobile screening deployment.

Methods

A spatial analysis was conducted using 2019–2023 census block group population data (ages 50–80) and a validated list of 165 lung cancer screening facilities in Oklahoma collected between December 2024 – May 2025, United States. Block groups were categorized into four spatial access groups. Analyses incorporated rural-urban continuum area codes. The 2SFCA used 30-min drive-time catchments, while the E2SFCA applied gravity-weighted 10-, 20-, and 30-min catchments.

Results

Across both methods, 41 % of eligible Oklahomans resided in low or no spatial access areas, covering over 60 % of the state's land. Small rural areas had the highest proportion of residents (up to 67 %) with limited or no spatial access. Urban areas showed better spatial access, but up to 19 % of eligible residents still lacked 30-min access.

Conclusions

Large sections of Oklahoma, particularly rural, lack spatial access to lung cancer screening. The 2SFCA and E2SFCA methods effectively identified underserved regions.

Highlights

• •Over 177,000 Oklahomans were eligible for lung cancer screening each year, 2019–23.
• •Among those eligible 41 % lived in areas with low or no screening spatial access.
• •Rural northwest and southeast Oklahoma had the poorest screening spatial access.
• •In Urban areas, 1 in 5 block groups lacked nearby lung cancer screening.

1. Introduction

Lung cancer is the leading cause of cancer death in the United States, with an estimated 125,070 deaths in 2024, and is the third most common cancer, with an estimated 234,580 cases (Siegel et al., 2024). In 2011, the National Lung Cancer Screening Trial reported that annual low-dose computed tomography (LDCT) screening in high-risk individuals resulted in a 20 % relative decrease in lung cancer mortality (Adams et al., 2023). In 2021, the United States Preventive Services Task Force updated its lung cancer screening recommendations to adults aged 50 to 80 years with a 20 pack-year smoking history who currently smoke or have quit within the past 15 years (Potter et al., 2021). LDCT screening rates are low in the United States and Oklahoma, with estimates of 16 % and 9 % of those eligible, respectively, in 2022 (American Lung Association, 2025). Historically, mobile mammography screening vans have played a crucial role in enhancing access to mammography breast cancer screening, particularly for underserved and rural communities (Vang et al., 2018). They bring life-saving screening services directly to areas where healthcare facilities may be scarce or difficult to access. The mobile vehicle approach in lung cancer screening may help reduce disparities by reaching populations that face barriers, such as limited spatial access and inadequate transportation (Boudreau et al., 2021; Carter-Harris et al., 2018; Haddad et al., 2020; Odahowski et al., 2019).

Both spatial and non-spatial factors influence accessibility to screening services. Non-spatial factors include service acceptability, cost, perceived quality, socioeconomic status, and social barriers (Carter-Harris et al., 2018; Miller et al., 2019; Wang et al., 2019). Spatial accessibility encompasses elements such as travel distance, traffic congestion, road conditions, and construction. Several methodologies exist to assess spatial accessibility, including population-to-provider ratios, simple buffer zones (e.g., Euclidean distance or drive-time catchments), and floating catchment area methods (McGrail, 2012). The population-to-provider ratio, for instance, is commonly employed by the Health Resources and Services Administration to identify Health Professional Shortage Areas and Medically Underserved Areas. However, this method relies on administrative boundaries, which often fail to reflect human mobility patterns. Previous studies have examined spatial access to the American College of Radiology Lung Cancer Screening Registry, finding that 5 % of the population lacks access to screening within 40 miles, as determined by a road network analysis of the county and census tract population centroid (Sahar et al., 2021; Sahar et al., 2022). The Floating Catchment Area method assesses spatial accessibility by measuring provider-to-population ratios within a set radius, ignoring administrative boundaries.

This method is grounded in the fundamental economic principles of supply and demand, where supply refers to the availability of healthcare services, and demand reflects the population's need for those services. For healthcare delivery to be effective, facilities must be located near areas of demand. Although the conceptual framework of balancing supply and demand is straightforward, operationalizing this relationship in spatial analyses can be complex, requiring consideration of geographic, demographic, and infrastructural factors.

This project aims to utilize two-step floating catchment analysis (2SFCA) and enhanced two-step floating catchment analysis (E2SFCA) to determine the spatial accessibility of fixed lung cancer screening locations in Oklahoma (McGrail, 2012; Gao et al., 2021; Luo and Qi, 2009; Kiran et al., 2020; Langford et al., 2016; Luo, 2004). These approaches have been used to determine the spatial accessibility of primary care physicians (McGrail, 2012; McGrail and Humphreys, 2009), hospitals (Gao et al., 2021; Kiani et al., 2021), medical clinics (Vadrevu and Kanjilal, 2016), pharmacies (Zhou et al., 2021), fire stations (Kiran et al., 2020), residential care facilities (Ni et al., 2015), and mammography facilities (Eberth et al., 2014; Maheswaran et al., 2006). Our goal was to identify and plan the optimal deployment locations for mobile LDCT, targeting areas in the state that lack ready spatial access. A secondary goal was to compare the results of the two methods.

2. Methods

2.1. Study area and data sources

The study area is the state of Oklahoma in the United States. Oklahoma had an estimated population of 4,053,824 in 2024.

For supply or lung cancer screening locations, between December 2024 through May 2025, we developed a list that compiled all hospitals in Oklahoma performing lung cancer screening. We screened each of them online to determine if they advertised lung cancer screening services. We then called each of them to confirm that they provided lung cancer screening. We did the same process for radiology centers in Oklahoma. We chose not to use the American College of Radiology Lung Cancer Screening Locator Tool, as we were aware of many locations not listed in the tool. Of the 565 hospitals and radiology centers we contacted, we found 165 (Fig. 1) compared to the 18 listed on the locator tool. Because they are not accessible to all populations, we performed the analysis excluding those located in Indian Health Service, Tribal Health, Urban Indian clinics (ITUs), Veterans Administration, and Department of Defense facilities. We included only lung cancer screening locations within the state of Oklahoma.

Fig. 1.

Lung cancer screening locations, Oklahoma, United States, 2025.

For the demand or population data, we used the block group geography, the lowest geographic unit that includes population by age groups. Based on recommended ages for lung cancer screening,27 we obtained the population aged 50 to 80 from the National Historical Geographic Information System estimates for 2019–2023. We multiplied that by the county estimates for 2022 of cigarette smoking, as reported in the 2025 county health rankings. We then used the statewide estimates for the prevalence of people who formerly smoked cigarettes (27.2 %). Finally, we used the estimates of those eligible for screening among people who currently or formerly smoked cigarettes for Oklahoma from the 2022 Behavioral Risk Factor Surveillance System (32.6 %).

Block group polygons were converted into centroid points to represent population demand within the service area. A 30-min drive-time catchment was used for the 2SFCA method, while 10, 20, and 30-min drive-time service areas were applied in the E2SFCA analysis. For analytical clarity, block groups with spatial access to lung cancer screening facilities were categorized into tertiles representing low, moderate, and high spatial access. An additional category was created for areas with zero spatial access. This classification resulted in four distinct spatial access groups: no spatial access, low spatial access (lowest tertile), moderate spatial access (middle tertile), and high spatial access (highest tertile).

To measure rurality, the United States Department of Agriculture, Economic Research Service's 2020 census tract level Rural-Urban Continuum Area (RUCA) codes were used. Urban refers to census tracts in the RUCA 1–3 group. To analyze different levels of rurality in Oklahoma, RUCA 4–6 census tracts were classified as large rural and RUCA 7–10 as small rural.

For the geocoding and service area development of lung cancer screening facilities, we used the ESRI Street Map for the 3rd quarter of 2024 in ArcGIS Pro (v. 3.4.2, ESRI Inc.). Our analysis utilized the tool in ArcGIS Pro, and the classic and enhanced 2SFCA tools developed by Hashtarkhani.28 All analyses were performed using ArcGIS Pro.2.2 Statistical Analysis.

We conducted a 2SFCA and E2SFCA analysis to calculate spatial accessibility scores for each block group in Oklahoma. The 2SFCA method (classic) assesses spatial accessibility by accounting for both service supply and demand, within a 30-min drive-time catchment area. We first evaluated service availability at each provider's location and the surrounding population, determining the supply-to-demand ratio. The following formula represents the supply-to-demand ratio for each location.

$$\raisebox{1ex}{$R_{j=S_j}$}\!\left/ \!\raisebox{-1ex}{${\sum }_{k\in d_{\mathit{kj}\le d_0}}{P}_k$}\right.$$

$R_j$= provider-to-population ratio at location $j.$

$S_j$= supply at location $j$

$P_k$= population at location $k$within the catchment

$d_{\mathit{kj}}$= travel distance or time between population $k$and provider $j$

$d_0$= maximum catchment size (e.g., 30-min drive)

We then identified spatially accessible providers within each population catchment area and summed their supply-to-demand ratios to derive the final spatial accessibility score. The following formula represents the provider's ratios for each location.

$$A_{i=\sum _{j\in d_{\mathit{ij}\le d_0}}{R}_j}$$

A_i = accessibility score for population location $i$

$R_j$= ratio from formula above

$d_{\mathit{ij}}$= travel distance or time between population $i$and provider $j$

The E2SFCA adjusted the spatial accessibility scores by applying distance-based weights. We applied Hashtarkhani's (Hashtarkhani et al., 2024) 2024 distance decay weights: 1.0 for the 10-min zone, 0.68 for the 20-min zone, and 0.22 for the 30-min zone.

2.2. Human subject review

This study utilized publicly available data, and the University of Oklahoma Health Sciences Institutional Review Board determined that it did not constitute human subjects' research.

3. Results

An estimated 177,474 people who either currently or formerly smoked cigarettes were eligible for lung cancer screening each year from 2019 to 2023 (Table 1). The mean population of the block groups was 279 (SD 156.4), ranging from 0 to 1056. Oklahoma has an area of 181,038 km², 3190 of which is water.

Table 1.

Number of block groups, population of estimated people eligible for lung cancer screening, and square kilometers of land or water by high, moderate, low, and zero spatial access to lung cancer screening using a two-step floating catchment analysis and enhanced two-step floating catchment analysis: Oklahoma, United States, 2025.

	Block Groups		Eligible Population		Square Kilometers
Spatial Access	2SFCA	E2SFCA	2SFCA	E2SFCA	2SFCA	E2SFCA
High	1037	1035	46,724	45,930	47,764	20,545
Moderate	1052	1033	58,159	59,057	21,014	19,942
Low	1021	1035	57,121	56,669	46,298	68,827
Zero	264	271	15,470	15,818	65,961	71,724
Total	3374		177,474		181,038

2SFCA = Two-Step Floating Catchment Analysis.

E2SFCA = Enhanced Two-Step Floating Catchment Analysis.

3.1. Two-step floating catchment analysis (2SFCA)

Among those eligible for lung cancer screening, 15,470 lived in block groups without spatial access to lung cancer screening centers, and 57,121 lived in block groups with low spatial access (Table 1). The population is well dispersed, with 26 % residing in high spatial access areas, 33 % living in moderate spatial access areas, 32 % in low spatial access areas, and 9 % living in areas with zero spatial access to lung cancer screening. This 9 % of the population lived in 36 % of the land area. If we consider both zero and low spatial access, we see that 41 % of the eligible population lived on 62 % of Oklahoma's land area.

Urban areas were most likely to have high (68 %) or moderate (78 %) spatial access to LDCT (Table 2). While urban blocks were less likely to have zero spatial access, one in five (21 %) block groups lacked spatial access to lung cancer screening. Moreover, 24 % of eligible people lived in urban block groups with no spatial access to lung cancer screening. In large rural areas, over half (58 %) of the block groups and 56 % of eligible people had low or no spatial access to lung cancer screening. In small rural areas, almost half of the block groups (47 %) and 44 % of eligible people had no spatial access to lung cancer screening. Interestingly, 20 % of block groups and 22 % of people in small rural areas had high spatial access to lung cancer screening.

Table 2.

Number of block groups, number of people eligible for lung cancer screening, and geographic area by high, moderate, low, and zero spatial access using a two-step floating catchment analysis and enhanced two-step floating catchment analysis for urban, large, and small rural areas: Oklahoma, United States, 2025.

	2SFCA				E2SFCA
	High	Moderate	Low	Zero	High	Moderate	Low	Zero
	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
Block Groups
Urban	694 (66.9)	825 (78.4)	503 (49.3)	51 (19.3)	900 (87.0)	836 (80.9)	286 (27.6)	51 (18.8)
Large Rural	170 (16.4)	139 (13.2)	314 (30.8)	36 (13.6)	29 (2.8)	100 (9.7)	488 (47.2)	42 (15.5)
Small Rural	173 (16.7)	88 (8.4)	204 (20.0)	177 (67.1)	106 (10.2)	97 (9.4)	261 (25.2)	178 (65.7)
Total	1,037	1,052	1,021	264	1,035	1,033	1,035	271
Population in Block Groups
Urban	28,475 (60.9)	44,934 (77.3)	29,336 (51.4)	3,366 (21.8)	38,177 (83.1)	47,430 (80.3)	17,138 (30.2)	3,366 (21.3)
Large Rural	8,712 (18.6)	7,845 (13.5)	16,232 (28.4)	2,166 (14.0)	1,785 (3.9)	5,839 (9.9)	24,906 (43.9)	2,425 (15.3)
Small Rural	9,537 (20.4)	5,380 (9.3)	11,553 (20.2)	9,938 (64.2)	5,968 (13.0)	5,788 (9.8)	14,625 (25.8)	10,027 (63.4)
Total	46,724	58,159	57,121	15,470	45,930	59,057	56,669	15,818
Square Kilometers
Urban	5,307 (11.1)	6,856 (32.7)	14,161 (30.6)	9,306 (14.1)	3,879 (18.9)	6,688 (33.5)	15,756 (22.9)	9,306 (13.0)
Large Rural	15,329 (32.1)	4,758 (22.7)	12,468 (26.9)	10,854 (16.5)	2,163 (10.5)	4,322 (21.7)	21,831 (31.7)	15,093 (21.0)
Small Rural	27,129 (56.8)	9,400 (44.7)	19,670 (42.5)	45,801 (69.4)	14,503 (70.6)	8,933 (44.8)	31,239 (45.4)	47,326 (66.0)
Total	47,765⁎	21,014	46,298	65,961	20,545	19,943⁎	68,827	71,725⁎

2SFCA = Two-Step Floating Catchment Analysis.

E2SFCA = Enhanced Two-Step Floating Catchment Analysis.

Percentages represent column totals.

Percentage totals may not add to 100 % due to rounding errors.

^⁎Totals may not equal the total on Table 1 due to rounding errors.

Relatively large and disparate geographic areas in Oklahoma lacked spatial access to lung cancer screening (Fig. 2); specifically, large sections of the sparsely populated Northwestern and Southeastern Oklahoma had no access to LDCT. Considering low spatial access, there are areas in urban central Oklahoma, particularly in and around the major cities of Oklahoma City and Tulsa, where more than half of the population resides (see the supplement for population density in Oklahoma). There are several significant areas in western Oklahoma with high spatial accessibility located near towns with rural hospitals.

Fig. 2.

A two-step floating catchment analysis of fixed lung cancer screening locations for block groups by high, moderate, low, and zero spatial access, Oklahoma, United States, 2025.

3.2. Enhanced two-step floating catchment analysis (E2SFCA)

Using E2SFCA, we do not see a significant difference in the number of eligible individuals for lung cancer screening. Just 348 more, 15,818, lived in the 271 block groups without spatial access to lung cancer screening, and slightly fewer individuals (56,669) lived in low spatial access block groups (Table 1, Table 2). This, zero spatial access, represented 9 % of the population living in 40 % of the land area. If we consider both zero and low spatial access, we see that 41 % of the eligible population lived on 78 % of Oklahoma's land.

Urban areas were most likely to have high (67 %) or moderate (78 %) spatial access to lung cancer screening (Table 2). Urban blocks were less likely to have zero spatial access; however, one in five (19 %) block groups lacked spatial access to lung cancer screening. Moreover, 19 % of eligible people lived in urban block groups with no spatial access to lung cancer screening. In large rural areas, nearly half (44 %) of the block groups had low or no spatial access to lung cancer screening. In small rural areas, two in three of the block groups (67 %) and 64 % of eligible people had no spatial access to lung cancer screening. Only 7 % of block groups and 8 % of people in small rural areas had high spatial access to lung cancer screening. When it comes to land areas, a high proportion (71 %) of high spatial access areas and 66 % of zero spatial access were found in small rural areas.

Relatively large and disparate geographic areas in Oklahoma lack spatial access to lung cancer screening (Fig. 3); specifically, large sections of Northwestern and Southeastern Oklahoma have no spatial access to these services. Considering low spatial access, there were large areas in central Oklahoma, particularly in and around the major cities of Oklahoma City and Tulsa. There were pockets of high spatial access throughout Oklahoma.

Fig. 3.

An enhanced two-step floating catchment analysis of fixed lung cancer screening locations for block groups by high, moderate, low, and zero spatial access, Oklahoma, United States, 2025.

3.3. Classic compared to enhanced

One slight difference between the results of the 2SFCA and E2SFCA was the increase in small rural regions within high spatial access areas in the enhanced model (Table 2). The proportion of the population of small rural regions within high spatial access areas decreased from 20 % to 13 %. Additionally, the percentage of people in small rural regions within low spatial access areas changed from 20 % for the classic methods to 25 % for the enhanced method. Moreover, in high spatial access areas, the proportion of the population in urban regions increased from 61 % to 83 %. In contrast, in areas with low spatial access, the proportion decreased from 51 % to 30 % among the urban population. There was also a slight difference in land area. In areas with high spatial access, the proportion of small rural regions increased from 53 % to 71 %. Most changes occurred in northwestern Oklahoma, where high spatial access block groups declined (Fig. 2, Fig. 3).

4. Discussion

Previous studies have examined spatial access to the LDCT, finding that 5 % of the population lacks access to screening within a 40-mile radius, and that rural areas were disproportionately affected (Sahar et al., 2021; Sahar et al., 2022; Atkins et al., 2017; Rohatgi et al., 2020). Our findings revealed that even more significant disparities exist in one particular rural state, despite more facilities offering LCDT than previously known. While urban areas had better spatial access, 41 % of eligible individuals lived in areas with low or no spatial access, covering up to 77 % of the state's land. Rural regions, especially in the northwest and southeast, were most underserved. We were also able to pinpoint the exact locations of these sites, enabling us to plan our LDCT mobile screening unit accordingly.

Given that the E2SFCA uses gravity weighting of the 10, 20, and 30-min rings, we would expect more areas to have low spatial access to lung cancer screening facilities. The proportion of urban areas increased, not surprisingly; however, a quarter of the urban population still lacks spatial access to lung cancer screening within a 30-min service area. The E2SFCA found more high spatial access areas in urban areas than the 2SFCA. This is not surprising given that the third ring (the 20–30 min service area) was weighted at just 22 %. This reflects reality more accurately, as it accounts for the decreased likelihood of people having lung screening further out in the service area where they reside. On the other hand, we know that rural populations drive further for services (Chan et al., 2006), suggesting that perhaps we should consider rural and urban areas with different distances.

This study used the 2SFCA and E2SFCA to evaluate the spatial accessibility of lung cancer screening facilities in Oklahoma. This study employed a descriptive approach, aiming to identify target areas with limited spatial access to these facilities. This will enable us to plan the optimal deployment locations for mobile lung cancer screening services. We observed that a large proportion of small rural areas had either low or no spatial access to lung cancer screening. Conversely, there are still small rural areas of high spatial access. It is essential to note that people with high spatial access do not necessarily use the service due to non-spatial access issues, such as acceptability, cost, perceived quality of services, socio-economic barriers to services, language issues, and lack of knowledge about screening (Miller et al., 2019; US Preventive Services Task Force, 2024; Trentham-Dietz et al., 2024; Alexandraki and Mooradian, 2010).

The use of lung cancer screening among the AI/AN population in Oklahoma is difficult to untangle. For this study, we excluded facilities that were ITUs, which are accessible only to AI/AN populations. We attempted to account for AI/AN populations through an analysis excluding 50 % of the AI/AN eligible population (see supplement). Findings were not substantially different from those described in this study.

4.1. Strengths and limitations

Our study has several strengths, including the relatively small geographic units of the block groups used in this study, following the recommendation of previous works in the area of LDCT spatial access (Sahar et al., 2021). Another strength is that we contacted all hospitals and radiology centers in Oklahoma to inquire about their LDCT. Thus, we believe that our list of facilities is more comprehensive than previous studies (Sahar et al., 2021; Sahar et al., 2022). Our study also has some limitations. We were unable to compensate for edge effects and account for services that were spatially accessible but outside the state's boundaries. We do, however, know that many facilities do not accept out-of-state insurance. Thus, we do not consider this a significant weakness. Additionally, traffic congestion, road construction, and public transportation lead to uncertainty about travel time (Chen et al., 2020). For this study, we did not consider travel time uncertainty. While it is an essential factor, it tends to be more critical in urban areas, where there is a higher likelihood of traffic congestion and increased chances of delay in travel time. We assessed LDCT through online review and phone calls to each facility (a strength); however, we did not assess whether the program offered comprehensive services, such as tobacco cessation support, tracking, and follow-up (Slatore et al., 2021). Because this study is ecological, we cannot infer results to individuals. In addition, our study is susceptible to the Modifiable Area Unit Problem because our study unit was the block group, and our results may vary depending on the administrative unit chosen. We recognize that we might come to different conclusions if we had used, for example, census tracts. An additional limitation is that we do not account for areas of high need that also have high uninsured rates, resulting in the inability to obtain the scan, despite a clear need. Finally, because it is challenging to obtain estimates of former cigarette smokers and pack-year history, we had to use state estimates for these variables; we believe that both variables are likely to differ geographically.

5. Conclusions

Lung cancer is the leading cause of cancer death in the United States, with significant disparities in spatial access to services, particularly in underserved and rural areas (Atkins et al., 2017; Rohatgi et al., 2020), despite early detection improving survival rates (Adams et al., 2023). Many areas were shown to have limited spatial accessibility. Findings emphasize the need for the targeted deployment of mobile lung cancer screening units to address geographic and demographic disparities, thereby ensuring equitable spatial access to lung cancer screening. Additionally, this study offers a method for other countries, states, counties, or health regions.

CRediT authorship contribution statement

Janis E. Campbell: Writing – review & editing, Writing – original draft, Visualization, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ayesha B. Sambo: Writing – review & editing, Conceptualization. Nubwa St. James: Writing – review & editing, Data curation. Kelly Willingham: Writing – review & editing, Resources, Data curation. Hayley Warren: Writing – review & editing, Data curation. Jerson Penaflor: Writing – review & editing, Data curation. Mark P. Doescher: Writing – review & editing, Conceptualization.

Funding

Research is supported by National Institute of General Medical Sciences, Grant/Award Number: U54GM104938 awarded to the University of Oklahoma Health Sciences Center and the National Cancer Institute, Cancer Center Support Grant P30CA225520, awarded to the University of Oklahoma Stephenson Cancer Center and used the Stephenson Cancer Center Biostatistics and Research Design Shared Resource. The content is solely the authors' responsibility and does not necessarily represent the official views of the National Institutes of Health or the Stephenson Cancer Center.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.