Next Article in Journal
Trends in Medicare Utilization and Reimbursement for Traumatic Brain Injury: 2003–2021
Previous Article in Journal
Acute Care Rehabilitation Services Utilization and Post-Acute Discharge Destination among Adults with Traumatic Brain Injury: The Moderating Effect of Functional and Physical Performance at Discharge
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Trauma Care
Volume 4
Issue 4
10.3390/traumacare4040023
traumacare-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Spatial Cluster Analysis of the Social Determinants of Health and Fatal Crashes Involving US Geriatric and Non-Geriatric Road Users

by
Oluwaseun Adeyemi
1,*,
Charles DiMaggio
2,3,
Corita Grudzen
4,
Sanjit Konda
5,
Erin Rogers
3,
Saul Blecker
3,6,
JohnRoss Rizzo
7 and
Joshua Chodosh
3,6,8
1
Ronald O. Perelman Department of Emergency Medicine, New York University Grossman School of Medicine, New York, NY 10016, USA
2
Department of Surgery, New York University Grossman School of Medicine, New York, NY 10016, USA
3
Department of Population Health, New York University Grossman School of Medicine, New York, NY 10016, USA
4
Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY 10065, USA
5
Department of Orthopedic and Trauma, New York University Grossman School of Medicine, New York, NY 10003, USA
6
Department of Medicine, New York University School of Medicine, New York, NY 10016, USA
7
Department of Rehabilitation Medicine, New York University School of Medicine, New York, NY 10016, USA
8
Medicine Service, Veterans Affairs, New York Harbor Healthcare System, New York, NY 10101, USA
*
Author to whom correspondence should be addressed.
Trauma Care 2024, 4(4), 266-281; https://doi.org/10.3390/traumacare4040023
Submission received: 10 July 2024 / Revised: 5 October 2024 / Accepted: 9 October 2024 / Published: 17 October 2024
Browse Figures
Versions Notes

Abstract

:
Social determinants of health (SDoH) are nonmedical factors impacting health outcomes. We evaluated the relationship between the county-level measure of SDoH and county-level fatal crash counts among geriatric and non-geriatric road users. We pooled data from the Fatality Analysis Reporting System and limited our analyses to the 3108 contiguous US counties. The outcome measures were county-level fatal crash counts involving (1) geriatric (65 years and older), (2) non-geriatric, and (3) all road users. The predictor variable was the multidimensional deprivation index (MDI), a composite measure of SDoH, measured as a three-level categorical variable defined as very highly deprived, highly deprived, and average-to-low deprived. We performed a Bayesian spatial Poisson regression analysis using integrated nested Laplace approximations and reported the adjusted crash fatality rate ratios (plus 95% credible intervals (CrI)). The median (Q1, Q3) standardized mortality rate ratios among geriatric and non-geriatric road users were 1.3 (0.6, 2.5) and 1.6 (0.9, 2.7), respectively. Counties classified as very highly deprived had 23% (95% CrI: 1.10–1.38) and 20% (95% CI: 1.08–1.32) increased geriatric and non-geriatric fatality crash rate ratios. In conclusion, improving county-level SDoH may reduce the county-level fatal rate ratios equally among geriatric and non-geriatric road users.

1. Introduction

Social determinants of health are nonmedical conditions in the environment where people are born, grow, live, work, and age that affect people’s quality of life and health outcomes [1,2,3]. These conditions are commonly grouped into five interrelated domains: (1) economic stability, (2) education access and quality, (3) healthcare access and quality, (4) neighborhood and built environment, and (5) community and social context [3]. Each of these wide-ranging conditions has been associated with individual and community-level health disparity and inequities and may serve as predictors of increased susceptibility to acute and chronic disease morbidity and mortality [4,5,6,7]. The multidimensional deprivation index (MDI) is a composite measure that captures six dimensions: standards of living, health, education, economic security, housing quality, and neighborhood quality [8]. It has been used as a composite measure of SDoH [9,10].
Fatal crash injuries represent a domain of injury whose occurrence may be partly explained by SDoH. Earlier studies have reported that low-income areas experience worse injury severity compared to high-income areas [11,12], and the higher the educational level, the lower the risk of fatal and nonfatal crash injuries [13,14]. Also, fatal crash injury rates are significantly higher in rural areas compared to urban areas [15]; road users who live in rural areas experience longer emergency medical service response times and are more likely to experience fatal crash injuries without resuscitation [16,17,18]. The built (such as intersections, highways, ramps, and work zones) and natural (such as rain, fog, and snow) road environments are associated with fatal crash injuries, and the impact of the built and natural road environment on fatal crash injury differs in rurality/urbanicity [19,20].
Road users, aged 65 years and older, represent a unique population that is at greater risk of worse injury outcomes when compared to younger adults with the same injury severity [21,22]. It is estimated that every day, 20 older adults sustain fatal crash injuries in the US, and fatal crash injury is the second leading cause of preventable death among US older adults [23]. Given the higher risk of injury outcomes among older adults, the effect of SDoH may affect older adult road users differently compared to other road users. Specifically, the possibility exists that when SDoH is measured as a singular measure, areas deprived of adequate health may be at increased risk of higher fatal crash injuries, and the risk will be further accentuated among older adult road users. With older adults disproportionately residing in rural and micropolitan areas, the possibility remains that older adult road users in highly deprived rural and micropolitan areas will be at increased risk of fatal crash injuries.
Indeed, the literature on geriatric fatal crash injuries is sparse, and very few studies assess the relationship between SDoH and crash injuries [19,24,25,26]. Yet, SDoH may provide insight into how the conditions surrounding the place we live influence the occurrence of crash injuries. A comprehensive understanding of the interplay between SDoH and fatal crash injury among all road users, including older adults and other adult road users, is crucial for an increased understanding of how to reduce the risk associated with preventable crash injury deaths. For this study, we first identified counties and the distribution of high levels of health deprivation as well as fatal crash rates involving geriatric and non-geriatric road users. We hypothesized that there would be a heterogeneous distribution of SDoH and fatal crash rates across the US. Secondly, we assessed the relationship between SDoH and fatal crash counts across all road users and separately across the geriatric and non-geriatric populations. We hypothesized that county-level measures of SDoH would be associated with county-level fatal crash counts among the geriatric and non-geriatric populations. Thirdly, we hypothesized that the effect size of the association between SDoH and geriatric and non-geriatric fatal crash counts will be greater in rural areas compared to urban areas. Lastly, we identified counties with excessively high fatal crash rate ratios after adjusting for measures of social determinants of health and other county-level variables.

2. Methods

2.1. Study Design and Population

For this ecological study, we employed a cross-sectional design to assess the association between county-level measures of fatal crash counts and SDoH. Our unit of analysis was at the county level.

2.2. Outcome Variable

The outcome variable is county-level fatal crash counts. We extracted the 2018 to 2020 data on fatal crash counts from the fatality analysis reporting system (FARS), one of the national crash databases of the National Highway Traffic Safety Administration [27,28]. The FARS dataset is a census of crash events during which one or more persons died either at the scene or within 30 days of the index crash event [27]. The FARS data consists of multiple linkable files that provide information on the accident, person, vehicle, and risk factor-related information. Data in the FARS database is provided either at the individual or crash event levels.
For this study, we defined three sets of county-level fatal crash deaths: across all road users, geriatric (aged 65 years and older), and non-geriatric road users. We defined these three sets of county-level fatal crash counts in three steps. First, from the person file of the FARS data, we created three datasets, restricting each by age (dataset 1: all ages, dataset 2: 65 years and older, dataset 3: less than 65 years). Next, we restricted each dataset to persons whose injury status was recorded as deceased. Lastly, we aggregated the death counts in each dataset by county using the five-digit federal information processing system (FIPS) code. We limited the counties to the 3108 contiguous US counties, excluding Alaska and the Virgin Islands.

2.3. Predictor Variable

The predictor variable is the multidimensional deprivation index (MDI). The MDI is a measure of SDoH developed by the US Census Bureau [10]. It consists of six dimensions: standard of living, health, education, economic security, housing quality, and neighborhood quality (Report, A.C.S., Multidimensional Deprivation in the United States: 2017. 2019) [8]. Each of the six dimensions is measured at the individual level and has specific defining criteria [8]. An individual is considered deprived if at least two criteria are met. The ACS published the national and county-level deprivation index [10], constructed from individual-level measures using the Alkire–Foster method [29]. Additionally, the ACS-defined MDI was defined as a five-level categorical variable defined as county deprivation: very low (i.e., less than half of the national rate), low (ranged from just above half of the national rate and below the lower limit of the 90% confidence interval), at national level (within 90% confidence interval of the national rate), high (above the upper limit of the 90% confidence interval and just below twice the national rate), and very high (at least two times the national rate) [10]. For this study, we defined MDI as a continuous measure and a three-level categorical variable: very highly deprived, highly deprived, and average-to-low deprived.

2.4. Control Variables

We selected county-level sociodemographic and crash characteristics based on prior literature. The county-level sociodemographic and health characteristics included proportions of males, Blacks, Hispanics, the poor, those with bachelor’s degrees, and excessive alcohol intake and emergency department utilization rate. These variables, excluding excessive alcohol intake and emergency department utilization rate, were extracted from the American Community Survey [30]. Excessive alcohol intake, defined as the proportion of adults with self-reported binge or heavy drinking, was extracted from the County Health Rankings and Roadmaps [31,32]. The emergency department utilization per 1000 Medicare beneficiaries was extracted from the Healthcare Cost and Utilization Project data of the Agency for Healthcare Research and Quality [33]. The county-level crash characteristics include the proportions of fatal crash events that occurred at night, during the rush hour period, and the proportion of crash victims screened for alcohol and drugs. Night driving was defined as crash events occurring between 12 midnight and 6 am. Rush hour driving was defined as crash events occurring between 6 and 9 am and 3 and 7 pm [34]. Persons with positive alcohol use were defined as those with a positive blood alcohol level of 0.08% or higher or whose alcohol use status was either self-reported or officer-reported. Persons with positive drug use were defined as those with positive drug tests or whose drug status was either self-reported or officer-reported. All crash characteristics were extracted from FARS.

2.5. Stratification

We stratified the counties by rurality/urbanicity using the 2010 rural–urban commuting area (RUCA) codes [35]. The RUCA code is a 10-level classification based on density, urbanization, and daily commuting. We defined rurality/urbanicity in three levels: metropolitan–urban (RUCA codes 1–3), micropolitan–urban (RUCA codes 4–7), and rural and small towns (RUCA codes 8–10).

2.6. Spatial Weight Matrix

Using the Euclidean distance across the 3108 contiguous US counties, we created a spatial weight matrix to map the spatial relationship across the counties. We used the inverse-weighted interpolation, which defines weight as a decreasing function of distance, to generate the spatial weight and account for the presence of spatial autocorrelation in crash occurrences [36].

2.7. Data Analysis

We reported three county-level fatal crash injury metrics: counts, case fatality rates (CFR), and standardized mortality ratio (SMR) across all road users and among geriatric and non-geriatric road users in metropolitan–urban, micropolitan–urban, and rural areas. Counts represent the number of deaths per count. CFR represents the percent of the number of crash deaths divided by the number of people involved in the fatal crash event. SMR represents the ratio of observed deaths divided by the expected number of deaths. The expected count (also explained below) is defined as the county population multiplied by the total fatal counts across all counties divided by the total population across all counties. We reported the frequency distributions, mean (and standard deviation (SD)), and median (and first and third quartiles (Q1, Q3)) for all the predictor and control variables, as appropriate. We visualized the distribution of the county-level crude fatality rates and MDI values on choropleths. We assessed the construct validity of the MDI across the crash, health, and population characteristics using correlation analysis. The MDI was considered to correlate with variables across the domains of SDoH if the correlation coefficient was ≥0.3. For the regression models, variables with correlation coefficients of ≥0.3 were deemed to exhibit substantial correlation with MDI and were removed from the models.
Since our outcome variable is a count measure, we performed model testing to compare the Poisson, the negative binomial regression, and the zero-inflated Poisson and negative binomial regression models using a spatial Bayesian paradigm with non-informative priors. The offset variable was the county-level expected count, defined as county population multiplied by the ratio of the sum of fatal counts and the sum of the population across the 3108 counties. Consistent with our definition of the outcome variable, we generated three different county population sizes (all ages, geriatric, and non-geriatric) and computed county-level expected counts for all road users and geriatric and non-geriatric road users. We reported the Akaike information criteria and Bayesian information criteria.
After performing model fitting and selecting the Poisson model, we performed a spatial Bayesian Poisson regression analysis. For the actual fatal crash counts Y s i , we assumed a Poisson distribution in the data. The mortality rate ratios { λ ( s i ) ,   i = 1 ,   … ,   n } over the n = 3108 counties were modeled by county-level K independent variables X k s i , k = 1 , … , K , a set of regression coefficients β k , k = 0,1 , … , K , spatial random effects terms { ν ( s i ) ,   i = 1 ,   … ,   n } , and error components { ε ( s i ) ,   i = 1 ,   … ,   n } that follow independent and identical zero-mean Gaussian distributions with variance parameter σ 2 . The regression equation takes the form:
l o g ( λ ( s i ) ) = β 0 + ∑ k = 1 K β k X k s i + ν ( s i ) + ε ( s i ) ,   i = 1 ,   … ,   n .
Thereafter, we reported the unadjusted and adjusted mortality rate ratios (incidence rate ratios) and reported the 95% credible interval. For each model, we computed the mortality rate ratios for all road users as well as geriatric and non-geriatric road users. We computed the predicted rate ratios for each county and visualized the distribution on choropleths. Additionally, we performed stratified analysis, and for each of the three outcome measures we generated mortality rate ratios by rurality/urbanicity.

3. Results

Across the 3108 counties, the median (Q1, Q3) MDI rate was 13.0 (9.4, 18.1) (Table 1). The proportion of counties classified as highly and very highly deprived were 9% and 26%, respectively. The mean (SD) crash case fatality rate across all age groups was 44.7% (15.1), with the rates lower among the geriatric population (9.1% (8.8)) compared to the non-geriatric population (35.6% (1.0)). The median (Q1, Q3) county SMR across all age groups was 1.6 (1.0, 2.7), with the ratio marginally lower among the geriatric (1.3 (0.6, 2.5)) population compared to the non-geriatric population (1.6 (0.9, 2.7)). The mean county proportion of males was 50% (2.2), while the median county proportions of non-Hispanic Blacks and Hispanics were 2.6% and 4.5%, respectively. The mean county proportion of the population classified as poor was 14.5%, while those with bachelor’s degree was 19.6%. Also, the mean county proportions of alcohol intake and ED utilization rates were 0.2% and 5.6%, respectively. Approximately a quarter of crashes occurred at night and 43% occured during the rush hour period. Furthermore, mean alcohol and drug screening rates across the counties were 43% and 39%, respectively. Counties with very high deprivation indices were found in South Dakota, Arizona, New Mexico, and in some southern region states—Texas, Louisiana, Mississippi, Alabama, Georgia, South Carolina, and Kentucky (Figure 1A). County clusters of high deprivation indices were similarly identified in South Dakota, Arizona, New Mexico, and in all southern US states (Figure 1B). Counties with standardized mortality rates of 10 or higher were identified in Montana, North and South Dakota, Wyoming, Colorado, New Mexico, Texas, Nevada, California, and Oregon (Figure 2A). This pattern of high standardized mortality rate was consistent among geriatric and non-geriatric road users (Figure 2B,C).
We evaluated the construct validity of the MDI across the five constructs of the SDoH (Table 2). In the economic stability domain, MDI exhibited a statistically significant strong positive correlation with the county proportions of those classified as poor (r = 0.72, p < 0.01). In the education access and quality domain, MDI exhibited a statistically significant moderate negative correlation with county proportions of those with bachelor’s degrees (r = −0.38, p < 0.01). In the healthcare access and quality domain, MDI exhibited a moderate positive correlation with county-level ED utilization rate (r = 0.38, p < 0.01). In the neighborhood and built environment, MDI exhibited a statistically significant mild positive correlation with rurality (r = 0.14; p < 0.01). In the community and social context domain, MDI exhibited a statistically significant moderate positive correlation with county proportions of Blacks (r = 0.49, p < 0.01) and a moderate negative correlation with county proportions of excessive alcohol intake (r = −0.48, p < 0.01). The result of the model testing showed that the spatial Poisson regression was the most parsimonious model compared to the negative binomial, zero-inflated negative binomial, and zero-inflated Poisson models (Table 3). Across the entire population and separately among the geriatric and non-geriatric populations, the spatial Poisson model had the lowest AIC and DIC.
Table 4 summarizes the univariate (unadjusted) association between crash fatality counts and the predictor and control variables. Across all age groups, a unit increase in MDI was associated with a 1% (95% CrI: 1.01–1.01) increase in crash fatality rates. Also, those living in very highly deprived and highly deprived counties had 33% (95% CrI: 1.20–1.48) and 25% (95% CrI: 1.17–1.32) increased crash mortality rate ratio compared to those living in counties with average to low MDI. This pattern of association was consistent among geriatric and non-geriatric road users. In the adjusted model (Table 5), those living in very highly deprived and highly deprived counties had 20% (95% CrI: 1.09–1.32) and 18% (95% CrI: 1.12–1.25) increased crash mortality rate ratio compared to those living in counties with average to low MDI. This pattern of association was consistent among geriatric and non-geriatric road users. When the model was stratified by rurality/urbanicity, older adult road users living in very highly deprived and highly deprived rural counties had 46% (95% CrI: 1.16–1.83) and 28% (95% CrI: 1.09–1.49) increased crash mortality rate ratio compared to those living in rural counties with average to low MDI. Also, older adult road users living in very highly deprived and highly deprived non-metropolitan urban counties had 44% (95% CrI: 1.21–1.71) and 26% (95% CrI: 1.13–1.42) increased crash mortality rate ratio compared to those living in non-metropolitan urban counties with average to low MDI. Additionally, older adult road users living in highly deprived metropolitan urban counties had 19% (95% CrI: 1.09–1.31) increased crash mortality rate ratio compared to those living in metropolitan urban counties with average to low MDI.
Several states in the West and South of the US had four-fold crash mortality rate ratios (Figure 3A). Among geriatric road users, states with several counties with a four-fold fatal mortality rate ratio were found in Oklahoma, Kansas, Texas, Utah, Nevada, California, and Oregon (Figure 3B). Among non-geriatric road users, states with higher than four-fold mortality rate ratio were found in almost all states in the West and South US regions as well as in North and South Dakota, Nebraska, and Kansas states (Figure 3C).

4. Discussion

Earlier studies have reported that social determinants of health are associated with cardiovascular [9,37,38], diabetes [39,40], cancer [41,42], and mental health-related morbidities and mortality [43,44]. Our study adds to the extant literature by reporting that social determinants of health are associated with fatal crash injuries among geriatric and non-geriatric road users. Our study validated that the multidimensional deprivation index is a composite measure that can be used to rank social determinants of health at the county level, and we report that about a third of US counties had high to very high multidimensional deprivation indices. Rural–urban disparities exist in both the multidimensional deprivation index and fatal crash counts, CFR, and SMR across all populations and among the geriatric and non-geriatric populations. Counties with high and very high deprivation indices have significantly elevated fatal crash rate ratios, and the ratios are not substantially different among geriatric and non-geriatric road users. Across all age groups, counties with high and very high deprivation indices have significantly elevated fatal crash rate ratios, but the elevated fatal crash rate ratios remained elevated in rural and micropolitan–urban counties among the geriatric population and elevated in micropolitan–urban and metropolitan–urban counties among the non-geriatric population. Lastly, we identified county hotspots of high deprivation indices as well as counties with more than four-fold increased fatality rate ratios in Texas, Oklahoma, Nevada, and Utah.
Although geriatric road users have higher case fatality rates and standardized mortality rates compared to non-geriatric road users, county-level measures of social determinants of health affect geriatric and non-geriatric road users equally. The elevated geriatric case fatality rate and standardized mortality rates may be a reflection of several factors such as the presence of co-morbidities [45,46,47,48], frailty [49,50,51,52,53,54], and under-appreciation and under-triage of the injury severity in the geriatric trauma population [55,56,57,58,59]. However, the comparable effect of the multidimensional deprivation index on fatal crash injuries of geriatric and non-geriatric road users underscores the extent to which social determinants of health affect all road users, irrespective of age. This finding indicates that interventions targeting social determinants could be beneficial for all road users, regardless of age. Efforts to improve socioeconomic conditions, enhance educational opportunities, ensure equitable access to healthcare services, and create safe and supportive neighborhoods will, therefore, have a positive effect on the health and safety outcomes of both older adults and younger individuals. These findings emphasize the need for comprehensive approaches that address social determinants of health as a means of promoting road safety and well-being for all age groups.
We report that geriatric road users in rural counties have significantly higher fatality rate ratios, while geriatric road users in metropolitan–urban counties have no significantly elevated crash fatality rate ratio. The significantly higher fatality rate ratios observed in rural counties suggest that there may be specific challenges unique to rural areas. Earlier studies have reported that residents in rural areas have longer emergency medical service (EMS) response times [16,17,18], limited access to level I or II trauma centers [60,61,62], and a higher proportion of deaths at the crash scene. In contrast, the lack of a significantly elevated crash fatality rate ratio among geriatric road users in metropolitan–urban counties suggests that factors such as better access to emergency medical services and more developed transportation networks, including helicopter EMS access, may contribute to improved injury outcomes for older adults in these areas. Addressing social determinants of health in rural areas may reduce fatal crash occurrences, especially among geriatric road users. Strategies to improve access to healthcare services, enhance transportation infrastructure, and create age-friendly environments can play a vital role in promoting road safety and protecting the well-being of geriatric road users. There is a need for collaborative efforts among policymakers, healthcare professionals, transportation agencies, and community stakeholders to develop and implement comprehensive interventions that address social determinants of health and improve road safety outcomes in rural areas.
Among non-geriatric road users, micropolitan–urban and metropolitan–urban counties with high and very high deprivation indices have significantly higher fatality rate ratios, but this pattern of association was not observed in rural counties. This observed finding may suggest the disproportionate distribution of non-geriatric road users in metropolitan urban areas and the impact of risky driving behaviors associated with young and middle-aged drivers. Earlier studies have reported that non-geriatric road users are more likely to drive while under the influence of drugs and alcohol [63,64,65] and engage in phone-related distracted driving [66,67,68,69]. These unique characteristics may account for the observed elevated fatal crash rate ratio in non-geriatric road users.
We identified fatality rate ratios in several counties in Texas, Oklahoma, Nevada, and Utah. These states also have county hotspots with very high multidimensional deprivation indices. Earlier studies have reported that these states have a high proportion of counties with prolonged EMS response time, county hotspots of rush hour-related fatal crash rates, and fatal crash rates due to non-use of seatbelts [16,19,70,71]. Identification of counties with high crash fatality rates and with high multidimensional deprivation indices allows for focused interventions. The use of spatial models and cluster analyses for fatal crash injury assessment can complement current efforts to address predictors of fatal crash injuries using the Motor Vehicle Prioritizing Interventions and Cost Calculator for States (MV PICCS) [72,73]. Using the MV PICCS to identify which specific domain of social determinants of health will be the most cost-effective in reducing fatal crash injury at the state level and using spatial and spatiotemporal models to identify counties within the states that should receive the highest priority may be an effective and efficient way of addressing the social determinants of fatal crash injuries at the state and county levels.
We used the spatial Bayesian paradigm with non-informative priors to provide a robust framework for analyzing the spatial distribution of fatal crashes related to social determinants of health. This methodology captures spatial dependencies, allowing us to account for the relationships between observations in geographically close areas, which is crucial since nearby locations can influence each other’s outcomes [74,75]. By utilizing non-informative priors, we minimized bias and ensured that our results were predominantly driven by the observed data, promoting objectivity in our analysis [76,77]. Additionally, the spatial model facilitates the identification of spatial patterns and trends, helping to quantify how geographic factors impact crash mortality rates [78,79]. This approach enhances our ability to make accurate crash fatality predictions and evaluate regional differences, ultimately providing valuable insights for targeted interventions in specific counties.
An important variable that our model did not directly capture is the distance to trauma centers, which could significantly impact fatal crash outcomes. While this variable was unavailable in the FARS dataset, distance to care is a significant predictor of crash survival rates [80,81], particularly in rural and underserved areas [82,83]. Studies have shown that longer travel times to trauma centers are linked to higher mortality following severe injuries [17,84]. Distance to trauma centers represents a measure of healthcare access—one of the five components of SDoH. Addressing SDoH, including access to timely trauma care, could, therefore, help reduce fatal crash rates especially in highly deprived counties.
This study has its limitations. The ecological nature of the study makes it impossible to make causal inferences. We pooled our data from FARS, a data repository that relies on state-level reporting of crash events. Data entry and processing errors, as well as inconsistent reporting of crash and crash-related events across the US, cannot be eliminated. The MDI was computed from multiple data sources. Data errors intrinsic to each of the data are inherently transferred into the MDI computation and misclassification bias cannot be eliminated. Despite these limitations, this study is one of the few studies that assessed the association between social determinants of health and fatal crash rates among geriatric and non-geriatric road users. By identifying counties with high rates of fatal crash injury occurrences as those with a high level of deprivation, this study provides information that can guide targeted interventions to counties and states that are in greatest need. Additionally, our study provides information that can inform policy and resource allocation for improving the social determinants of health and preventing fatal crash occurrences in the presence of other competing community needs.

5. Conclusions

In summary, the more deprived a county is, the more likely its residents involved in crash injuries are to die. Also, despite older adults being less frequent road users, geriatric road users have higher case fatality rates and standardized mortality rates. County-level deprivation measures of the SDoH are equally associated with geriatric and non-geriatric crash-related fatal rate ratios. This raises the question for future studies on whether the effect stems from the county or ZIP code where one is born, resides, or experiences crashes. Irrespective of the potential predictor of county-level mortality, policies and interventions that improve county-level SDoH may reduce the county-level fatal rate ratios equally among geriatric and non-geriatric road users.

Author Contributions

Conceptualization, O.A.; methodology, O.A. and C.D.; software, O.A. and J.C.; formal analysis, O.A.; data curation, O.A.; writing—original draft preparation, O.A.; writing—review and editing, O.A., C.D., C.G., S.K., E.R., S.B., J.R. and J.C.; visualization, O.A.; supervision, C.D. and J.C.; project administration, C.D. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Ethical review and approval were waived for this study due to the study using publicly available data.

Informed Consent Statement

Not applicable since the data used for this study is publicly available de-identified data.

Data Availability Statement

Data is available at https://doi.org/10.6084/m9.figshare.26218550.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Braveman, P.; Egerter, S.; Williams, D.R. The social determinants of health: Coming of age. Annu. Rev. Public Health 2011, 32, 381–398. [Google Scholar] [CrossRef]
  2. Center for Disease Control and Prevention. Social Determinants of Health: Know What Affects Health. Available online: https://www.cdc.gov/about/priorities/why-is-addressing-sdoh-important.html (accessed on 9 July 2024).
  3. Healthy People. Social Determinants of Health. Available online: https://health.gov/healthypeople/priority-areas/social-determinants-health (accessed on 9 July 2024).
  4. Stajduhar, K.I.; Mollison, A.; Giesbrecht, M.; McNeil, R.; Pauly, B.; Reimer-Kirkham, S.; Dosani, N.; Wallace, B.; Showler, G.; Meagher, C.; et al. “Just too busy living in the moment and surviving”: Barriers to accessing health care for structurally vulnerable populations at end-of-life. BMC Palliat. Care 2019, 18, 11. [Google Scholar] [CrossRef]
  5. Bailey, V. Rural Hospitals Facing Risk of Closure Due to Financial Stress. Revcycle Intelligence. 5 May 2022. Available online: https://www.techtarget.com/revcyclemanagement/news/366600881/Rural-Hospitals-Facing-Risk-of-Closure-Due-to-Financial-Stress (accessed on 9 July 2024).
  6. Kaufman, B.G.; Thomas, S.R.; Randolph, R.K.; Perry, J.R.; Thompson, K.W.; Holmes, G.M.; Pink, G.H. The Rising Rate of Rural Hospital Closures. J. Rural. Health 2016, 32, 35–43. [Google Scholar] [CrossRef]
  7. Miller, K.E.M.; James, H.J.; Holmes, G.M.; Van Houtven, C.H. The effect of rural hospital closures on emergency medical service response and transport times. Health Serv. Res. 2020, 55, 288–300. [Google Scholar] [CrossRef]
  8. Glassman, B. Multidimensional Deprivation in the United States: 2017. In American Community Survey Reports; U.S. Census Bureau: Washington, DC, USA, 2019. Available online: https://www.census.gov/content/dam/Census/library/publications/2019/demo/acs-40.pdf (accessed on 9 July 2024).
  9. Flynn, A.; Vaughan, A.; Casper, M.; Glassman, B. Differences in Heart Disease Death Rates and Trends by County-Level Deprivation, 2010–2019. In American Community Survey Reports; U.S. Census Bureau: Washington, DC, USA, 2021. Available online: https://www.census.gov/content/dam/Census/library/working-papers/2021/demo/sehsd-wp2021-19.pdf (accessed on 9 July 2024).
  10. Glassman, B. Producing County-Level MDI Rates Using Public Use Data: 2010 to 2019. In American Community Survey Reports; U.S. Census Bureau: Washington, DC, USA, 2022. Available online: https://www.census.gov/content/dam/Census/library/working-papers/2022/demo/sehsd-wp2022-19.pdf (accessed on 9 July 2024).
  11. Dumbaugh, E.; Li, Y.; Saha, D.; Marshall, W. Why do lower-income areas experience worse road safety outcomes? Examining the role of the built environment in Orange County, Florida. Transp. Res. Interdiscip. Perspect. 2022, 16, 100696. [Google Scholar] [CrossRef]
  12. Morency, P.; Gauvin, L.; Plante, C.; Fournier, M.; Morency, C. Neighborhood social inequalities in road traffic injuries: The influence of traffic volume and road design. Am. J. Public Health 2012, 102, 1112–1119. [Google Scholar] [CrossRef]
  13. Saha, D.; Alluri, P.; Gan, A.; Wu, W. Spatial analysis of macro-level bicycle crashes using the class of conditional autoregressive models. Accid. Anal. Prev. 2018, 118, 166–177. [Google Scholar] [CrossRef]
  14. Sami, A.; Moafian, G.; Najafi, A.; Aghabeigi, M.R.; Yamini, N.; Heydari, S.T.; Lankarani, K.B. Educational level and age as contributing factors to road traffic accidents. Chin. J. Traumatol. 2013, 16, 281–285. [Google Scholar] [CrossRef]
  15. Zwerling, C.; Peek-Asa, C.; Whitten, P.S.; Choi, S.W.; Sprince, N.L.; Jones, M.P. Fatal motor vehicle crashes in rural and urban areas: Decomposing rates into contributing factors. Inj. Prev. 2005, 11, 24–28. [Google Scholar] [CrossRef]
  16. Adeyemi, O.J.; Paul, R.; Arif, A. An assessment of the rural-urban differences in the crash response time and county-level crash fatalities in the United States. J. Rural. Health 2021, 38, 999–1010. [Google Scholar] [CrossRef]
  17. Adeyemi, O.J.; Paul, R.; DiMaggio, C.; Delmelle, E.; Arif, A. The association of crash response times and deaths at the crash scene: A cross-sectional analysis using the 2019 National Emergency Medical Service Information System. J. Rural. Health 2022, 38, 1011–1024. [Google Scholar] [CrossRef]
  18. Byrne, J.P.; Mann, N.C.; Dai, M.; Mason, S.A.; Karanicolas, P.; Rizoli, S.; Nathens, A.B. Association Between Emergency Medical Service Response Time and Motor Vehicle Crash Mortality in the United States. JAMA Surg. 2019, 154, 286–293. [Google Scholar] [CrossRef]
  19. Adeyemi, O.; Paul, R.; Delmelle, E.; DiMaggio, C.; Arif, A. Road environment characteristics and fatal crash injury during the rush and non-rush hour periods in the U.S: Model testing and cluster analysis. Spat. Spatio-Temporal Epidemiol. 2023, 44, 100562. [Google Scholar] [CrossRef]
  20. Wu, B.; Wang, X.; Liu, T.; Dong, N.; Li, Y. Exploring Factors Contributing to Crash Injury Severity in the Rural-Urban Fringe of the Central City. J. Adv. Transp. 2021, 2021, 8453465. [Google Scholar] [CrossRef]
  21. Cunha-Diniz, F.; Taveira-Gomes, T.; Santos, A.; Teixeira, J.M.; Magalhães, T. Are There Any Differences in Road Traffic Injury Outcomes between Older and Younger Adults? Setting the Grounds for Posttraumatic Senior Personal Injury Assessment Guidelines. J. Clin. Med. 2023, 12, 2353. [Google Scholar] [CrossRef]
  22. Gopinath, B.; Harris, I.A.; Nicholas, M.; Casey, P.; Blyth, F.; Maher, C.G.; Cameron, I.D. A comparison of health outcomes in older versus younger adults following a road traffic crash injury: A cohort study. PLoS ONE 2015, 10, e0122732. [Google Scholar] [CrossRef]
  23. Centers for Disease Control and Prevention. Older Adult Drivers. Available online: https://www.cdc.gov/older-adult-drivers/about/index.html (accessed on 9 July 2024).
  24. Adeyemi, O. An Assessment of the Knowledge, Attitude, and Practice of Phone Use While Driving and Crash Outcomes Among Drivers in Oyo State, Nigeria. Int. Res. Rev. 2022, 11, 25–44. [Google Scholar]
  25. Adeyemi, O. Alcohol- and Drug-Associated Injury Outcomes Among Older Adults Involved in Car Crashes. Innov. Aging 2021, 5, 127. [Google Scholar] [CrossRef]
  26. Saeednejad, M.; Sadeghian, F.; Fayaz, M.; Rafael, D.; Atlasi, R.; Kazemzadeh Houjaghan, A.; Abedi Kichi, R.; Asgardoon, M.H.; Zabihi Mahmoudabadi, H.; Salamati, Z.; et al. Association of Social Determinants of Health and Road Traffic Deaths: A Systematic Review. Bull. Emerg. Trauma 2020, 8, 211–217. [Google Scholar] [CrossRef]
  27. National Highway Traffic Safety Administration. Fatality Analysis Reporting System Analytical User’s Manual 1975–2015; National Highway Traffic Safety Administration: Washington, DC, USA, 2016. Available online: https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/813556 (accessed on 9 July 2024).
  28. National Highway Traffic Safety Administration. Fatalities and Fatality Rates by State, 1994–2018—State: USA. Available online: https://www-fars.nhtsa.dot.gov/States/StatesFatalitiesFatalityRates.aspx (accessed on 9 July 2024).
  29. Alkire, S.; Foster, J. Counting and multidimensional poverty measurement. J. Public Econ. 2011, 95, 476–487. [Google Scholar] [CrossRef]
  30. United States Census Bureau. American Community Survey Data. Available online: https://www.census.gov/programs-surveys/acs/data.html (accessed on 9 July 2024).
  31. County Health Rankings and Roadmaps. Ranking Data & Documentation. Available online: https://www.countyhealthrankings.org/explore-health-rankings/rankings-data-documentation (accessed on 9 July 2024).
  32. University of Wisconsin Population Health Institute. 2020 County Health Rankings Key Findings Reports; County Health Ranking and Roadmaps: Madison, WI, USA, 2020; Available online: https://www.countyhealthrankings.org/reports/2020-county-health-rankings-key-findings-report (accessed on 9 July 2024).
  33. Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project. Available online: https://hcup-us.ahrq.gov/databases.jsp (accessed on 9 July 2024).
  34. Adeyemi, O.J.; Arif, A.A.; Paul, R. Exploring the relationship of rush hour period and fatal and non-fatal crash injuries in the U.S.: A systematic review and meta-analysis. Accid. Anal. Prev. 2021, 163, 106462. [Google Scholar] [CrossRef]
  35. Economic Research Services. Rural-Urban Commuting Area Codes; U.S. Department of Agriculture: Washington, DC, USA, 2019. Available online: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes/ (accessed on 9 July 2024).
  36. Ke, W.; Cheng, H.P.; Yan, D.; Lin, C. The application of cluster analysis and inverse distance-weighted interpolation to appraising the water quality of Three Forks Lake. Procedia Environ. Sci. 2011, 10, 2511–2517. [Google Scholar] [CrossRef]
  37. Powell-Wiley, T.M.; Baumer, Y.; Baah, F.O.; Baez, A.S.; Farmer, N.; Mahlobo, C.T.; Pita, M.A.; Potharaju, K.A.; Tamura, K.; Wallen, G.R. Social Determinants of Cardiovascular Disease. Circ. Res. 2022, 130, 782–799. [Google Scholar] [CrossRef]
  38. Kimenai, D.M.; Pirondini, L.; Gregson, J.; Prieto, D.; Pocock, S.J.; Perel, P.; Hamilton, T.; Welsh, P.; Campbell, A.; Porteous, D.J.; et al. Socioeconomic Deprivation: An Important, Largely Unrecognized Risk Factor in Primary Prevention of Cardiovascular Disease. Circulation 2022, 146, 240–248. [Google Scholar] [CrossRef]
  39. Butler, A.M. Social Determinants of Health and Racial/Ethnic Disparities in Type 2 Diabetes in Youth. Curr. Diabetes Rep. 2017, 17, 60. [Google Scholar] [CrossRef]
  40. Walker, R.J.; Strom Williams, J.; Egede, L.E. Influence of Race, Ethnicity and Social Determinants of Health on Diabetes Outcomes. Am. J. Med. Sci. 2016, 351, 366–373. [Google Scholar] [CrossRef]
  41. Coughlin, S.S. Social determinants of breast cancer risk, stage, and survival. Breast Cancer Res. Treat. 2019, 177, 537–548. [Google Scholar] [CrossRef]
  42. Coughlin, S.S. Social determinants of colorectal cancer risk, stage, and survival: A systematic review. Int. J. Color. Dis. 2020, 35, 985–995. [Google Scholar] [CrossRef]
  43. Alegría, M.; NeMoyer, A.; Falgàs Bagué, I.; Wang, Y.; Alvarez, K. Social Determinants of Mental Health: Where We Are and Where We Need to Go. Curr. Psychiatry Rep. 2018, 20, 95. [Google Scholar] [CrossRef]
  44. Allen, J.; Balfour, R.; Bell, R.; Marmot, M. Social determinants of mental health. Int. Rev. Psychiatry 2014, 26, 392–407. [Google Scholar] [CrossRef]
  45. Salive, M.E. Multimorbidity in older adults. Epidemiol. Rev. 2013, 35, 75–83. [Google Scholar] [CrossRef]
  46. Divo, M.J.; Martinez, C.H.; Mannino, D.M. Ageing and the epidemiology of multimorbidity. Eur. Respir. J. 2014, 44, 1055–1068. [Google Scholar] [CrossRef]
  47. Shigemura, N.; Toyoda, Y. Elderly patients with multiple comorbidities: Insights from the bedside to the bench and programmatic directions for this new challenge in lung transplantation. Transpl. Int. 2020, 33, 347–355. [Google Scholar] [CrossRef]
  48. Davis, J.W.; Chung, R.; Juarez, D.T. Prevalence of comorbid conditions with aging among patients with diabetes and cardiovascular disease. Hawaii Med. J. 2011, 70, 209–213. Available online: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3215980/ (accessed on 9 July 2024).
  49. Adeyemi, O.; Grudzen, C.; DiMaggio, C.; Wittman, I.; Velez-Rosborough, A.; Arcila-Mesa, M.; Cuthel, A.; Poracky, H.; Meyman, P.; Chodosh, J. Pre-Injury Frailty and Clinical Care Trajectory of Geriatric Trauma Patients: A Retrospective Cohort Analysis of A Large Level I US Trauma Center. medRxiv 2023. [Google Scholar] [CrossRef]
  50. Bandeen-Roche, K.; Seplaki, C.L.; Huang, J.; Buta, B.; Kalyani, R.R.; Varadhan, R.; Xue, Q.-L.; Walston, J.D.; Kasper, J.D. Frailty in Older Adults: A Nationally Representative Profile in the United States. J. Gerontol. Ser. A 2015, 70, 1427–1434. [Google Scholar] [CrossRef]
  51. Chen, X.; Mao, G.; Leng, S.X. Frailty syndrome: An overview. Clin. Interv. Aging 2014, 9, 433–441. [Google Scholar] [CrossRef]
  52. Curtis, E.; Romanowski, K.; Sen, S.; Hill, A.; Cocanour, C. Frailty score on admission predicts mortality and discharge disposition in elderly trauma patients over the age of 65 y. J. Surg. Res. 2018, 230, 13–19. [Google Scholar] [CrossRef]
  53. Fehlmann, C.A.; Nickel, C.H.; Cino, E.; Al-Najjar, Z.; Langlois, N.; Eagles, D. Frailty assessment in emergency medicine using the Clinical Frailty Scale: A scoping review. Intern. Emerg. Med. 2022, 17, 2407–2418. [Google Scholar] [CrossRef]
  54. Lin, H.S.; Watts, J.N.; Peel, N.M.; Hubbard, R.E. Frailty and post-operative outcomes in older surgical patients: A systematic review. BMC Geriatr. 2016, 16, 157. [Google Scholar] [CrossRef]
  55. Alshibani, A.; Alharbi, M.; Conroy, S. Under-triage of older trauma patients in prehospital care: A systematic review. Eur. Geriatr. Med. 2021, 12, 903–919. [Google Scholar] [CrossRef]
  56. Anantha, R.V.; Painter, M.D.; Diaz-Garelli, F.; Nunn, A.M.; Miller, P.R.; Chang, M.C.; Jason Hoth, J. Undertriage Despite Use of Geriatric-Specific Trauma Team Activation Guidelines: Who Are We Missing? Am. Surg. 2020, 87, 419–426. [Google Scholar] [CrossRef]
  57. Chang, D.C.; Bass, R.R.; Cornwell, E.E.; Mackenzie, E.J. Undertriage of elderly trauma patients to state-designated trauma centers. Arch. Surg. 2008, 143, 776–781; discussion 782. [Google Scholar] [CrossRef]
  58. Chu, I.; Vaca, F.; Stratton, S.; Chakravarthy, B.; Hoonpongsimanont, W.; Lotfipour, S. Geriatric Trauma Care: Challenges Facing Emergency Medical Services. Available online: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2860422/pdf/cjem8_2p0051.pdf (accessed on 9 July 2024).
  59. Horst, M.A.; Morgan, M.E.; Vernon, T.M.; Bradburn, E.H.; Cook, A.D.; Shtayyeh, T.; D’Andrea, L.; Rogers, F.B. The geriatric trauma patient: A neglected individual in a mature trauma system. J. Trauma Acute Care Surg. 2020, 89, 192–198. [Google Scholar] [CrossRef]
  60. Dave, U.; Gosine, B.; Palaniappan, A. An Overview of Trauma Center Levels and Disparities in Rural Trauma Care. Reconstr. Rev. 2020, 10, 23–25. [Google Scholar] [CrossRef]
  61. Hsia, R.; Shen, Y.C. Possible geographical barriers to trauma center access for vulnerable patients in the United States: An analysis of urban and rural communities. Arch. Surg. 2011, 146, 46–52. [Google Scholar] [CrossRef]
  62. Newgard, C.D.; Fu, R.; Bulger, E.; Hedges, J.R.; Mann, N.C.; Wright, D.A.; Lehrfeld, D.P.; Shields, C.; Hoskins, G.; Warden, C.; et al. Evaluation of Rural vs Urban Trauma Patients Served by 9-1-1 Emergency Medical Services. JAMA Surg. 2017, 152, 11–18. [Google Scholar] [CrossRef]
  63. Adeyemi, O.J.; Paul, R.; DiMaggio, C.; Delmelle, E.; Arif, A. Rush Hour-Related Road Crashes: Assessing the Social and Environmental Determinants of Fatal and Non-Fatal Road Crash Events. Ph.D. Thesis, The University of North Carolina at Charlotte, Ann Arbor, MI, USA, 2021. Available online: https://ninercommons.charlotte.edu/islandora/object/etd%3A2929/datastream/PDF/download/citation.pdf (accessed on 9 July 2024).
  64. Li, G.; Brady, J.E.; Chen, Q. Drug use and fatal motor vehicle crashes: A case-control study. Accid. Anal. Prev. 2013, 60, 205–210. [Google Scholar] [CrossRef]
  65. Martin, J.-L.; Gadegbeku, B.; Wu, D.; Viallon, V.; Laumon, B. Cannabis, alcohol and fatal road accidents. PLoS ONE 2017, 12, e0187320. [Google Scholar] [CrossRef]
  66. Bergmark, R.W.; Gliklich, E.; Guo, R.; Gliklich, R.E. Texting while driving: The development and validation of the distracted driving survey and risk score among young adults. Inj. Epidemiol. 2016, 3, 7. [Google Scholar] [CrossRef]
  67. Boccardi, V.; Paolisso, G. Distracted driving and crash risk. N. Engl. J. Med. 2014, 370, 1565. [Google Scholar] [CrossRef]
  68. Coben, J.H.; Zhu, M. Keeping an eye on distracted driving. JAMA J. Am. Med. Assoc. 2013, 309, 877–878. [Google Scholar] [CrossRef]
  69. Foss, R.D.; Goodwin, A.H. Distracted driver behaviors and distracting conditions among adolescent drivers: Findings from a naturalistic driving study. J. Adolesc. Health 2014, 54, S50–S60. [Google Scholar] [CrossRef]
  70. Adeyemi, O.; Paul, R.; Arif, A. Spatial Cluster Analysis of Fatal Road Accidents From Non-Use of Seat Belts Among Older Drivers. Innov. Aging 2020, 4, 113–114. [Google Scholar] [CrossRef]
  71. Adeyemi, O.; Rajib, P.; Arif, A. Rush Hour Accidents: Assessing the Relationship between Road Environmental Determinants and County-Level Fatal Road Accidents Patterns in the United States. Available online: https://apha.confex.com/apha/2020/meetingapp.cgi/Paper/483798 (accessed on 9 July 2024).
  72. Center for Disease Control and Prevention. Motor Vehicle Prioritizing Interventions and Cost Calculator for States (MV PICCS). Available online: https://wisqars.cdc.gov/mvpiccs/ (accessed on 9 July 2024).
  73. Ecola, L.; Ringel, J.S.; Connor, K.; Powell, D.; Jackson, C.P.; Ng, P.; Miller, C. Costs and Effectiveness of Interventions to Reduce Motor Vehicle-Related Injuries and Deaths: Supplement to Tool Documentation. RAND Health Q. 2018, 8, 9. Available online: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6183771/ (accessed on 9 July 2024).
  74. Mitra, S. Spatial Autocorrelation and Bayesian Spatial Statistical Method for Analyzing Intersections Prone to Injury Crashes. Transp. Res. Rec. 2009, 2136, 92–100. [Google Scholar] [CrossRef]
  75. Wang, Y.; Chen, X.; Xue, F. A Review of Bayesian Spatiotemporal Models in Spatial Epidemiology. ISPRS Int. J. Geo-Inf. 2024, 13, 97. [Google Scholar] [CrossRef]
  76. Wheeler, M.W. An investigation of non-informative priors for Bayesian dose-response modeling. Regul. Toxicol. Pharmacol. 2023, 141, 105389. [Google Scholar] [CrossRef]
  77. van de Schoot, R.; Kaplan, D.; Denissen, J.; Asendorpf, J.B.; Neyer, F.J.; van Aken, M.A.G. A gentle introduction to Bayesian analysis: Applications to developmental research. Child Dev. 2014, 85, 842–860. [Google Scholar] [CrossRef]
  78. Alam, M.S.; Tabassum, N.J. Spatial pattern identification and crash severity analysis of road traffic crash hot spots in Ohio. Heliyon 2023, 9, e16303. [Google Scholar] [CrossRef]
  79. Liu, J.; Das, S.; Zhan, F.B.; Khan, M.N. Spatial analysis of geographical disparities in pedestrian safety. Transp. Policy 2024, 156, 164–181. [Google Scholar] [CrossRef]
  80. Hu, W.; Dong, Q.; Huang, B. Correlations between road crash mortality rate and distance to trauma centers. J. Transp. Health 2017, 6, 50–59. [Google Scholar] [CrossRef]
  81. Wiratama, B.S.; Chen, P.L.; Chao, C.J.; Wang, M.H.; Saleh, W.; Lin, H.A.; Pai, C.W. Effect of Distance to Trauma Centre, Trauma Centre Level, and Trauma Centre Region on Fatal Injuries among Motorcyclists in Taiwan. Int. J. Environ. Res. Public Health 2021, 18, 2998. [Google Scholar] [CrossRef]
  82. Jarman, M.P.; Castillo, R.C.; Carlini, A.R.; Kodadek, L.M.; Haider, A.H.J.S. Rural risk: Geographic disparities in trauma mortality. Surgery 2016, 160, 1551–1559. [Google Scholar] [CrossRef]
  83. Brown, J.B.; Rosengart, M.R.; Billiar, T.R.; Peitzman, A.B.; Sperry, J.L. Distance matters: Effect of geographic trauma system resource organization on fatal motor vehicle collisions. J. Trauma Acute Care Surg. 2017, 83, 111–118. [Google Scholar] [CrossRef]
  84. Jarman, M.P.; Curriero, F.C.; Haut, E.R.; Pollack Porter, K.; Castillo, R.C. Associations of Distance to Trauma Care, Community Income, and Neighborhood Median Age With Rates of Injury Mortality. JAMA Surg. 2018, 153, 535–543. [Google Scholar] [CrossRef]
Traumacare 04 00023 g001
Figure 1. (A) US county distribution of the multidimensional deprivation index. (B) Cluster identification of multidimensional deprivation index across US counties.
Figure 1. (A) US county distribution of the multidimensional deprivation index. (B) Cluster identification of multidimensional deprivation index across US counties.
Traumacare 04 00023 g001
Traumacare 04 00023 g002
Figure 2. County-level distribution of the motor vehicle crash standardized mortality rates among (A) road users of all ages, (B) geriatric road users, and (C) non-geriatric road users.
Figure 2. County-level distribution of the motor vehicle crash standardized mortality rates among (A) road users of all ages, (B) geriatric road users, and (C) non-geriatric road users.
Traumacare 04 00023 g002
Traumacare 04 00023 g003
Figure 3. County-level distribution of adjusted crash mortality rates among (A) road users of all ages, (B) geriatric road users, and (C) non-geriatric road users.
Figure 3. County-level distribution of adjusted crash mortality rates among (A) road users of all ages, (B) geriatric road users, and (C) non-geriatric road users.
Traumacare 04 00023 g003
Table 1. County-level frequency distribution and summary statistics of multidimensional deprivation index, crash, and sociodemographic characteristics by rurality/urbanicity (N = 3108).
Table 1. County-level frequency distribution and summary statistics of multidimensional deprivation index, crash, and sociodemographic characteristics by rurality/urbanicity (N = 3108).
VariableAll (N = 3108)Rural + Small Towns (n = 627 (20.2%))Micropolitan–Urban (n = 1321 (42.5%))Metropolitan–Urban (n = 1160 (37.3%))p-Value
Multidimensional Deprivation Index
MDI Rate (Median (Q1, Q3))13.0 (9.4, 18.1)13.4 (9.7, 19.0)14.9 (10.8, 19.3)11.4 (8.1, 15.5)<0.001 ##
Deprivation Index (n (%))
  Very Highly Deprived283 (9.1)125 (11.4)109 (11.8)49 (4.5)<0.001 ###
  Highly Deprived806 (25.9)290 (26.5)302 (32.6)214 (19.7)
  Average-to-Low Deprived2019 (65.0)681 (62.1)516 (55.7)822 (75.8)
Fatal Crash Measures
  Fatal Counts (Geriatric) (Median (Q1, Q3))3.0 (1.0, 6.0)1.0 (0.0, 2.0)3.0 (1.0, 5.0)6.0 (3.0, 11.0)<0.001 ##
  Fatal Counts (Non-Geriatric) (Median (Q1, Q3))11.0 (5.0, 23.0)5.0 (2.0, 8.0)13.0 (7.0, 20.0)25.0 (11.0, 50.0)<0.001 ##
  Fatal Counts (All Ages) (Median (Q1, Q3))14.0 (6.0, 28.0)6.0 (3.0, 11.0)16.0 (9.0, 25.0)31.0 (15.0, 62.0)<0.001 ##
  Case Fatality Rate (Geriatric) (Mean (SD))9.1 (8.8)10.3 (12.3)8.8 (6.8)8.2 (5.7)<0.001 #
  Case Fatality Rate (Non-Geriatric) (Mean (SD))35.6 (1.0)38.8 (19.6)35.1 (11.5)32.9 (10.5)<0.001 #
  Case Fatality Rate (All Ages) (Mean (SD))44.7 (15.1)49.1 (19.0)43.9 (12.6)41.1 (11.2)<0.001 #
  SMR (Geriatric) (Median (Q1, Q3))1.3 (0.6, 2.5)1.7 (0.0, 3.5)1.5 (0.9, 2.5)1.1 (0.6, 1.7)<0.001 ##
  SMR (Non-Geriatric) (Median (Q1, Q3))1.6 (0.9, 2.7)2.3 (1.3, 3.7)1.8 (1.2, 2.6)1.1 (0.7, 1.7)<0.001 ##
  SMR (All Ages) (Median (Q1, Q3))1.6 (1.0, 2.7)2.3 (1.4, 3.6)1.8 (1.2, 2.6)1.1 (0.7, 1.7)<0.001 ##
Sociodemographic Characteristics
  Male Proportion (Mean (SD))50.1 (2.2)50.6 (2.6)50.1 (2.2)49.5 (1.6)<0.001 #
  Black Proportion (Median (Q1, Q3))2.6 (0.9, 11.1)1.0 (0.6, 3.5)2.8 (1.1, 11.4)6.4 (1.9, 15.8)<0.001 ##
  Hispanic Proportion (Median (Q1, Q3))4.5 (2.5, 10.2)3.5 (2.1, 7.3)4.4 (2.5, 10.6)5.8 (3.1, 12.1)<0.001 ##
  Poverty Proportion (Mean (SD))14.5 (5.8)15.6 (6.3)15.8 (5.6)12.2 (4.7)<0.001 #
  Bachelor’s Degree Proportion (Median (Q1, Q3))19.6 (15.3, 26.0)18.1 (14.3, 22.2)17.5 (14.3, 22.1)25.2 (18.7, 33.5)<0.001 ##
  Excessive Alcohol Intake Proportion (Mean (SD))0.2 (0.1)0.2 (0.1)0.2 (0.1)0.2 (0.1)<0.001 #
  ED Utilization Rate (Mean (SD))5.6 (1.1)5.5 (1.3)6.0 (1.0)5.4 (0.9)<0.001 #
Crash Event Characteristics
  Night Fatal Crash Proportion (Mean (SD))24.8 (17.3)21.5 (21.7)24.5 (14.4)28.4 (13.6)<0.001 #
  Rush Hour Fatal Event Proportion (Mean (SD))43.3 (20.9)43.0 (28.2)44.5 (16.8)42.7 (14.2)0.112
  Alcohol Screening Proportion (Mean (SD))43.3 (20.2)46.1 (25.2)42.7 (17.5)40.9 (16.2)<0.001 #
  Drug Screening Proportion (Mean (SD))38.7 (22.9)38.7 (26.9)38.8 (21.3)38.8 (19.7)0.995
SMR: standardized mortality rate.; CFR: case fatality rate; ED: emergency department; #: one-way ANOVA performed; ##: Kruskal–Wallis test performed; ###: Chi-square test performed.
Table 2. Pairwise correlation coefficient assessing the relationship between county-level multidimensional deprivation index, sociodemographic characteristics, and fatal crash events.
Table 2. Pairwise correlation coefficient assessing the relationship between county-level multidimensional deprivation index, sociodemographic characteristics, and fatal crash events.
IDVariables1234567891011121314
1MDI Rate
2Male Proportion−0.03
3Black Proportion0.49 **−0.14 **
4Hispanic Proportion0.21 **0.15 **−0.09 **
5Poverty Proportion0.72 **0.09 **0.48 **0.07 **
6Bachelor’s Degree Prop.−0.38 **−0.19 **−0.08 **−0.01−0.48 **
7Night Fatal Crash Prop.0.07 **−0.04 *0.16 **0.11 **0.07 **0.06 **
8Rush Hour Fatal Crash0.010.01−0.06 **−0.010.00−0.05 **−0.48 **
9Excessive Alcohol Intake −0.48 **0.17 **−0.36 **0.00−0.54 **0.43 **0.00−0.02
10Alcohol Screening Prop.−0.13 **0.06 **−0.15 **−0.20 **−0.03 **0.010.010.000.10 **
11Drug Screening Proportion−0.11 **0.06 **−0.14 **−0.12 **−0.020.04 *0.020.010.11 **0.75 **
12ED Utilization Rate0.38 **−0.030.33 **0.010.46 **−0.42 **0.05 **0.03−0.25 **−0.06 **0.00
13Rurality/Urbanity0.14 **0.21 **−0.15 **−0.06 **0.24 **−0.36 **−0.15 **0.04 *−0.17 **0.06 **−0.030.03
14Geriatric CFR−0.06 **0.04 *−0.07 **−0.08 **−0.03−0.06 **−0.16 **0.02−0.010.06 **0.07 **−0.030.10 **
MDI: Multidimensional deprivation index; **: significant at <0.01; * significant at <0.05.
Table 3. Summary of model testing.
Table 3. Summary of model testing.
Equation Y = β 0 5 + β 1 V H D + β 2 M D 2 + β 3 C o v a r i a t e s + E x p   E x p e c t e d   C o u n t s + 1   C o u n t y   I D ) + 1 S p a t i a l   W e i g h t   M a t r i x )
CategoriesModelsDICWAIC
Geriatric Road UsersPoisson12,773.0812,797.04
Negative Binomial13,078.2213,097.60
Zero-inflated Poisson12,799.5312,837.25
Zero-inflated Negative Binomial13,099.4713,141.36
Non-Geriatric Road UsersPoisson18,272.9318,148.86
Negative Binomial18,862.4118,670.81
Zero-inflated Poisson18,306.9318,226.43
Zero-inflated Negative Binomial18,791.4418,766.62
All Road UsersPoisson19,123.4918,972.55
Negative Binomial19,596.3019,490.90
Zero-inflated Poisson19,126.4418,972.53
Zero-inflated Negative Binomial19,418.8219,319.98
Y represents Y1 (fatal counts of geriatric road users) or Y2 (fatal counts of non-geriatric road users) or Y3 (fatal counts of all road users); VHD: dummy of very highly deprived vs. average-to-low Deprived; HD: dummy of highly deprived vs. average-to-low deprived; Exp: exposure; DIC: deviance information criteria; WAIC: Watanabe–Akaike information criteria; expected counts computed as: [Sum (Fatal counts)/Sum (County Population)] × County Population.
Table 4. Univariate association between county-level crash fatality and county multidimensional deprivation index, socioeconomic and crash characteristics.
Table 4. Univariate association between county-level crash fatality and county multidimensional deprivation index, socioeconomic and crash characteristics.
VariablesGeriatric Road UsersNon-Geriatric Road UsersAll Age Groups
Incidence Rate Ratio
(95% CrI)		Incidence Rate Ratio
(95% CrI)		Incidence Rate Ratio
(95% CrI)
MDI Rate (Median (Q1, Q3))1.01 (1.01–1.02)1.01 (1.01–1.02)1.01 (1.01–1.01)
Deprivation Index (n (%))
  Very Highly Deprived1.34 (1.17–1.52)1.35 (1.20–1.50)1.33 (1.20–1.48)
  Highly Deprived1.21 (1.13–1.30)1.26 (1.18–1.34)1.25 (1.17–1.32)
  Average-to-Low DeprivedRef.Ref.Ref.
Sociodemographic Characteristics
  Male Proportion1.07 (1.06–1.09)1.05 (1.04–1.06)1.05 (1.04–1.06)
  Black Proportion0.99 (0.99–0.99)0.99 (0.99–1.00)0.99 (0.99–1.00)
  Hispanic Proportion0.99 (0.99–0.99)0.99 (0.98–0.99)0.99 (0.99–0.99)
  Poverty Proportion 1.03 (1.02–1.03)1.03 (1.02–1.03)1.02 (1.02–1.03)
  Bachelor’s Degree Proportion0.97 (0.97–0.97)0.96 (0.96–0.97)0.96 (0.96–0.97)
  Excessive Alcohol Intake Prop.0.00 (0.00–0.00)0.00 (0.00–0.00)0.00 (0.00–0.00)
  ED Utilization Rate1.10 (1.07–1.14)1.10 (1.07–1.13)1.09 (1.06–1.11)
  Rural + Small Towns1.81 (1.67–1.96)2.06 (1.94–2.20)1.97 (1.86–2.10)
  Micropolitan–Urban1.41 (1.32–1.50)1.52 (1.45–1.60)1.47 (1.40–1.54)
  Metropolitan–UrbanRef.Ref.Ref.
Crash Event Characteristics
  Night Fatal Crash Proportion0.99 (0.99–0.99)1.00 (0.99–1.00)0.99 (0.99–1.00)
  Rush Hour Fatal Event Prop.1.00 (1.00–1.01)1.00 (1.00–1.00)1.00 (1.00–1.00)
  Alcohol Screening Proportion0.99 (0.99–1.00)1.00 (1.00–1.00)1.00 (1.00–1.00)
  Drug Screening Proportion0.99 (0.99–1.00)1.00 (1.00–1.00)1.00 (1.00–1.00)
County Metrics (Coefficients)
  County: Random Component11.93 (8.11–18.99)7.58 (6.67–9.70)9.15 (7.13–11.71)
  County: Spatial Component1.65 (1.27–2.07)1.40 (1.14–1.69)1.43 (1.18–1.71)
MDI: multidimensional deprivation index. Significant results in bold fonts. Ref. = reference category for categorical variables.
Table 5. Incidence rate ratios of crash fatalities involving older adults, teenagers, and all age groups across US counties with varying levels of deprivation (N = 3108).
Table 5. Incidence rate ratios of crash fatalities involving older adults, teenagers, and all age groups across US counties with varying levels of deprivation (N = 3108).
Road UsersVariableAll Areas (95% CrI)Rural + Small Towns (95% CrI)Non-Metropolitan Urban (95% CrI)Metropolitan Urban (95% CrI)
Geriatric Road Users (65+ years)Deprivation Index
  Very Highly Deprived1.23 (1.10–1.38)1.46 (1.16–1.83)1.44 (1.21–1.71)1.17 (0.98–1.41)
  Highly Deprived1.15 (1.08–1.24)1.28 (1.09–1.49)1.26 (1.13–1.42)1.19 (1.09–1.31)
  Average-to-Low DeprivedRef.Ref.Ref.Ref.
Non-Geriatric Road Users Deprivation Index
  Very Highly Deprived1.20 (1.08–1.32)1.20 (0.97–1.47)1.28 (1.10–1.48)1.25 (1.05–1.48)
  Highly Deprived1.19 (1.12–1.26)1.13 (0.99–1.29)1.15 (1.05–1.27)1.34 (1.23–1.47)
  Average-to-Low DeprivedRef.
All Age GroupsDeprivation Index
  Very Highly Deprived1.20 (1.09–1.32)1.22 (1.01–1.48)1.29 (1.12–1.48)1.23 (1.04–1.45)
  Highly Deprived1.18 (1.12–1.25)1.15 (1.02–1.30)1.16 (1.06–1.27)1.32 (1.21–1.44)
  Average-to-Low DeprivedRef.Ref.Ref.Ref.
Model adjusted for male proportion, Hispanic proportion, the proportion of fatal crash events at night and during the rush hour period, the proportion of alcohol screening behavior at fatal crash events, and rurality/urbanicity. Rurality/urbanicity excluded from the rural, non-metropolitan–urban, and metropolitan–urban models since the variable was used for stratification. Significant results in bold fonts. Ref. = reference category for categorical variables.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Adeyemi, O.; DiMaggio, C.; Grudzen, C.; Konda, S.; Rogers, E.; Blecker, S.; Rizzo, J.; Chodosh, J. Spatial Cluster Analysis of the Social Determinants of Health and Fatal Crashes Involving US Geriatric and Non-Geriatric Road Users. Trauma Care 2024, 4, 266-281. https://doi.org/10.3390/traumacare4040023

AMA Style

Adeyemi O, DiMaggio C, Grudzen C, Konda S, Rogers E, Blecker S, Rizzo J, Chodosh J. Spatial Cluster Analysis of the Social Determinants of Health and Fatal Crashes Involving US Geriatric and Non-Geriatric Road Users. Trauma Care. 2024; 4(4):266-281. https://doi.org/10.3390/traumacare4040023

Chicago/Turabian Style

Adeyemi, Oluwaseun, Charles DiMaggio, Corita Grudzen, Sanjit Konda, Erin Rogers, Saul Blecker, JohnRoss Rizzo, and Joshua Chodosh. 2024. "Spatial Cluster Analysis of the Social Determinants of Health and Fatal Crashes Involving US Geriatric and Non-Geriatric Road Users" Trauma Care 4, no. 4: 266-281. https://doi.org/10.3390/traumacare4040023

APA Style

Adeyemi, O., DiMaggio, C., Grudzen, C., Konda, S., Rogers, E., Blecker, S., Rizzo, J., & Chodosh, J. (2024). Spatial Cluster Analysis of the Social Determinants of Health and Fatal Crashes Involving US Geriatric and Non-Geriatric Road Users. Trauma Care, 4(4), 266-281. https://doi.org/10.3390/traumacare4040023

Article Metrics

Back to TopTop