Advertisement

The Role of Individual-Level Factors in Rural Mortality Disparities

Open AccessPublished:July 21, 2022DOI:https://doi.org/10.1016/j.focus.2022.100013

      HIGHLIGHTS

      • The gap in early death rates among rural versus urban residents has grown since 1990.
      • The role of modifiable behaviors is poorly understood.
      • Adjusting for alcohol and physical activity attenuated the risk in rural residents.
      • Studies incorporating individual- and population-level data will provide more insight.

      Introduction

      The role of individual risk factors in the rural‒urban mortality disparity is poorly understood. The purpose of this study was to explore the role of individual-level demographics and health behaviors on the association between rural residence and the risk of mortality.

      Methods

      Cancer Prevention Study-II participants provided updated addresses throughout the study period. Rural‒Urban Commuting Area codes were assigned to participants’ geocoded addresses as a time-varying exposure. Cox proportional hazards regression was used to estimate hazard ratios and 95% CIs for mortality associated with Rural‒Urban Commuting Area groups.

      Results

      After adjustment for age and sex, residents of rural areas/small towns had a small but statistically significant elevated risk of all-cause mortality compared with metropolitan residents (hazard ratio=1.04; 95% CI=1.01, 1.06). Adjustment for additional covariates attenuated the association entirely (hazard ratio=0.99; 95% CI=0.97, 1.01). Individually, adjustment for education (hazard ratio=0.99; 95% CI=0.97, 1.01), alcohol use (hazard ratio=1.01; 95% CI=0.99, 1.04), and moderate-to-vigorous intensity aerobic physical activity (hazard ratio=1.00; 95% CI=0.97, 1.02) eliminated the elevated risk.

      Conclusions

      The elevated risk of death for rural compared with that for metropolitan residents appeared to be largely explained by individual-level demographics and health behaviors. If replicated in other subpopulations, these results suggest that modifiable factors may play an important role in reducing the rural mortality disparity.

      Keywords

      INTRODUCTION

      Over the last 4 decades, age-specific mortality has been steadily improving for many demographic subgroups in the U.S.
      • Islami F
      • Ward EM
      • Sung H
      • et al.
      Annual report to the nation on the status of cancer, Part 1: National cancer statistics.
      ,
      GBD 2015 Mortality and Causes of Death Collaborators
      Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015.
      Conversely, the gap in early mortality rates among those living in rural versus urban areas has grown since 1990.
      • Singh GK
      • Siahpush M.
      Widening rural-urban disparities in life expectancy, U.S., 1969–2009.
      Studies using surveillance data show that rural residents experience approximately 135–170 excess deaths per 100,000 residents compared with urban residents.
      • Cosby AG
      • McDoom-Echebiri MM
      • James W
      • Khandekar H
      • Brown W
      • Hanna HL.
      Growth and persistence of place-based mortality in the United States: the rural mortality penalty.
      • Cross SH
      • Califf RM
      • Warraich HJ.
      Rural-urban disparity in mortality in the U.S. from 1999 to 2019.
      • Singh GK
      • Siahpush M.
      Widening rural-urban disparities in all-cause mortality and mortality from major causes of death in the USA, 1969–2009.
      • Monnat SM.
      Trends in U.S. working-age non-Hispanic white mortality: rural-urban and within-rural differences.
      Given that approximately 19% of U.S. adults live in rural areas, understanding this disparity is imperative.

      Census urban and rural classification and urban area criteria. United States Census Bureau. https://www.census.gov/geo/reference/ua/urban-rural-2010.html. Accessed date October 4, 2021;https://www.census.gov/programs-surveys/geography/guidance/geo-areas/urban-rural/2010-urban-rural.html.

      Reasons for the rural mortality disparity in the U.S. are poorly understood but are likely multifaceted. Socioeconomic status may partially explain the disparity because residents of rural areas generally have lower educational attainment than urban residents.
      • Byun SY
      • Meece JL
      • Irvin MJ.
      Rural-nonrural disparities in postsecondary educational attainment revisited.
      In addition, race/ethnicity may compound the rural disparity in mortality because Black residents of rural areas have poorer mortality outcomes than White rural residents.
      • James W
      • Cossman JS.
      Long-term trends in Black and White mortality in the rural United States: evidence of a race-specific rural mortality penalty.
      ,
      • James WL.
      All rural places are not created equal: revisiting the rural mortality penalty in the United States.
      It is also possible that rural‒urban differences in modifiable health behaviors, such as smoking, physical activity, and alcohol consumption, may contribute to the disparity. Adults living in rural areas have a higher prevalence of drinking
      • Matthews KA
      • Croft JB
      • Liu Y
      • et al.
      Health-related behaviors by urban-rural county classification - United States, 2013.
      and tobacco use
      • Lutfiyya MN
      • Shah KK
      • Johnson M
      • et al.
      Adolescent daily cigarette smoking: is rural residency a risk factor?.
      ,
      • Roberts ME
      • Doogan NJ
      • Kurti AN
      • et al.
      Rural tobacco use across the United States: how rural and urban areas differ, broken down by census regions and divisions.
      and a lower prevalence of meeting physical activity guidelines than adults in metropolitan areas.
      • Whitfield GP
      • Carlson SA
      • Ussery EN
      • Fulton JE
      • Galuska DA
      • Petersen R.
      Trends in meeting physical activity guidelines among urban and rural dwelling adults — United States, 2008–2017.
      Nevertheless, the potential role of modifiable risk factors is poorly understood, likely because many studies of rural health are surveillance studies, which lack individual-level data. To address this limitation, we used detailed individual-level data from a large cohort study to explore the role of individual-level demographics (race, educational attainment, marital status, living arrangement), health history (previous comorbidities), and health behaviors (smoking, physical activity, alcohol use, chewing tobacco use, and BMI) on the association between rural residence and mortality risk.

      METHODS

      Study Population

      The Cancer Prevention Study-II Nutrition Cohort (CPS-IINC) is a longitudinal study of cancer incidence and mortality initiated by the American Cancer Society.
      • Calle EE
      • Rodriguez C
      • Jacobs EJ
      • et al.
      The American Cancer Society Cancer Prevention Study II Nutrition Cohort: rationale, study design, and baseline characteristics.
      CPS-IINC was established in 1992 and included >184,000 participants from 21 U.S. states. Beginning in 1997, CPS-IINC participants were mailed biennial surveys to update their health history and health behavior information. Cancer Prevention Study-II is approved by the Emory University IRB.

      Measures

      To assess urban/rural residence, Rural‒Urban Commuting Area (RUCA) codes and census tract‒level codes that integrate measures of population density, urbanization, and daily commuting to identify urban cores and adjacent territories were used to define urban/rural residence. Participants provided a current address in 1997 and subsequently thereafter through mailed biennial surveys. The 10 primary RUCA codes were assigned to participants’ geocoded addresses as a time-varying exposure (with 1990 RUCA codes assigned to 1997–1999 addresses, 2000 codes assigned to 2000–2009 addresses, and 2010 codes assigned to addresses after 2009 to account for area reclassification over time).
      • Byun SY
      • Meece JL
      • Irvin MJ.
      Rural-nonrural disparities in postsecondary educational attainment revisited.
      ,
      • Johnson KM
      • Lichter DT.
      Metropolitan reclassification and the urbanization of rural America.
      If an updated address was unable to be linked to a RUCA code, the previous RUCA code was carried forward. Most (83.4%) participants did not change RUCA codes during the study period. All primary RUCA codes were collapsed into the 3 RUCA groups for analyses: (1) metropolitan (Codes 1–3), (2) micropolitan (Codes 4–6), and (3) rural areas and small towns combined (Codes 7–10).
      • Byun SY
      • Meece JL
      • Irvin MJ.
      Rural-nonrural disparities in postsecondary educational attainment revisited.
      For outcomes, the primary outcome was all-cause mortality (with cardiovascular, cancer, and respiratory mortality analyses included in the appendix only) ascertained through the biennial linkage of the cohort with the National Death Index.
      • Calle EE
      • Terrell DD.
      Utility of the National Death Index for ascertainment of mortality among cancer prevention study II participants.
      Causes of death were classified with ICD-9 codes for deaths occurring from 1992 to 1998, and ICD-10 codes for additional deaths occurring through the end of follow-up (December 2016).

      Statistical Analysis

      Cox proportional hazards regression was used to estimate hazard ratios (HR) and 95% CIs for all-cause mortality associated with RUCA group in models: (1) adjusted for age and sex; (2) further adjusted for individual potential confounders selected a priori; and (3) adjusted for all potential confounders, including race/ethnicity, highest educational attainment, marital status, living arrangement, comorbidity score, aspirin use, occupational dirtiness index, pesticide exposure, average particulate matter exposure, moderate‒vigorous intensity aerobic physical activity (MVPA), BMI, smoking duration and cigarettes/day for current smokers, years since quitting for former smokers, chewing tobacco, and alcohol use (covariates are described in Appendix Table 1, available online).
      Table 1Participant Characteristics by Rural-Urban Commuting Area (RUCA) Group, 1997
      CharacteristicsMetropolitan (n=126,570)Micropolitan (n=14,596)Rural/small town (n=16,861)
      Age at baseline (years), mean (SD)68 (6.24)68.1 (6.29)68.1 (6.3)
      Sex, n (%)
       Female68,334 (54.0)7,750 (53.1)8,916 (52.9)
       Male58,236 (46.0)6,846 (46.9)7,945 (47.1)
      Race, n (%)
       White122,968 (97.2)14,373 (98.5)16,662 (98.8)
       Black1,806 (1.4)123 (0.8)105 (0.6)
       Other1,796 (1.4)100 (0.7)94 (0.6)
      Education, n (%)
       <HS6,578 (5.2)1,320 (9.0)1,956 (11.6)
       HS grad29,923 (23.6)4,522 (31.0)6,096 (36.2)
       Some college36,886 (29.1)4,104 (28.1)4,441 (26.3)
       College grad27,203 (21.5)2,415 (16.5)2,346 (13.9)
       Grad school25,117 (19.8)2,153 (14.8)1,915 (11.4)
       Unknown863 (0.7)82 (0.6)107 (0.6)
      Marital status, n (%)
       Single1,862 (1.5)158 (1.1)152 (0.9)
       Married93,811 (74.1)10,747 (73.6)12,495 (74.1)
       Other30,897 (24.4)3,691 (25.3)4,214 (25)
      Alcohol use, n (%)
       None49,382 (39)6,679 (45.8)8,123 (48.2)
       <1/day45,396 (35.9)4,426 (30.3)4,903 (29.1)
       1/day12,009 (9.5)1,068 (7.3)1,167 (6.9)
       ≥2/day4,892 (3.9)472 (3.2)519 (3.1)
       Missing14,891 (11.8)1,951 (13.4)2,149 (12.7)
      Living arrangement, n (%)
       Alone13,855 (10.9)1,564 (10.7)1,788 (10.6)
       With spouse/family83,412 (65.9)9,409 (64.5)10,829 (64.2)
       Assisted living243 (0.2)29 (0.2)39 (0.2)
       Other/missing29,060 (23)3,594 (24.6)4,205 (24.9)
      Body mass index (kg/m2), n (%)
       18.5 to <2550,314 (39.8)5,124 (35.1)5,662 (33.6)
       25 to <3045,621 (36.0)5,583 (38.3)6,498 (38.5)
       ≥3018,396 (14.5)2,462 (16.9)2,922 (17.3)
       Missing12,239 (9.7)1,427 (9.8)1,779 (10.6)
      Comorbidity score, n (%)
       None42,438 (33.5)4,906 (33.6)5,730 (34.0)
       140,424 (31.9)4,638 (31.8)5,241 (31.1)
       ≥243,708 (34.5)5,052 (34.6)5,890 (34.9)
      Aspirin use, n (%)
       No regular use74,176 (58.6)8,159 (55.9)9,334 (55.4)
       0 to <15 pills/month12,046 (9.5)1,347 (9.2)1,567 (9.3)
       15 to <30 pills/month5,681 (4.5)607 (4.2)678 (4)
       ≥30 pills/month14,509 (11.5)1,741 (11.9)2,121 (12.6)
       Missing20,158 (15.9)2,742 (18.8)3,161 (18.7)
      PM2.5 (quartiles), n (%)
       Q130,127 (23.8)5,437 (37.2)8,373 (49.7)
       Q232,647 (25.8)3,976 (27.2)4,082 (24.2)
       Q331,672 (25.0)3,216 (22.0)2,636 (15.6)
       Q430,748 (24.3)1,869 (12.8)1,674 (9.9)
       Missing1,376 (1.1)98 (0.7)96 (0.6)
      MVPA (MET hours/week), n (%)
       None10,423 (8.2)1,538 (10.5)1,977 (11.7)
       >0 to <7.542,039 (33.2)5,261 (36.0)6,009 (35.6)
       7.5 to <1531,406 (24.8)3,465 (23.7)4,142 (24.6)
       15 to <22.518,402 (14.5)1,892 (13.0)1,994 (11.8)
       ≥22.515,823 (12.5)1,312 (9.0)1,295 (7.7)
       Missing8,477 (6.7)1,128 (7.7)1,444 (8.6)
      Years since quitting smoking, n (%)
       Never smoker55,909 (44.2)6,884 (47.2)8,055 (47.8)
       ≥3024,596 (19.4)2,475 (17.0)2,813 (16.7)
       20 to <3014,853 (11.7)1,539 (10.5)1,734 (10.3)
       10 to <2013,031 (10.3)1,463 (10.0)1,621 (9.6)
       <109,769 (7.7)1,188 (8.1)1,317 (7.8)
       Current smoker7,046 (5.6)882 (6.0)1,079 (6.4)
       Missing1,366 (1.1)165 (1.1)242 (1.4)
      Cigarettes/day, n (%)
       Never smoker55,909 (44.2)6,884 (47.2)8,055 (47.8)
       Former smoker63,548 (50.2)6,816 (46.7)7,708 (45.7)
       <203,871 (3.1)461 (3.2)579 (3.4)
       201,449 (1.1)195 (1.3)252 (1.5)
       >201,513 (1.2)193 (1.3)221 (1.3)
       Missing280 (0.2)47 (0.3)46 (0.3)
      Years smoked, n (%)
       Never55,909 (44.2)6,884 (47.2)8,055 (47.8)
       Former63,548 (50.2)6,816 (46.7)7,708 (45.7)
       <401,768 (1.4)223 (1.5)292 (1.7)
       40 to <503,085 (2.4)377 (2.6)451 (2.7)
       ≥502,124 (1.7)273 (1.9)320 (1.9)
       Missing136 (0.1)23 (0.2)35 (0.2)
      Grad, graduate; HS, high school; MVPA, moderate-to-vigorous intensity aerobic physical activity; PM2.5 = particulate matter with diameter <2.5 µm; RUCA, Rural‒Urban Commuting Area.
      Modification by sex, education (a proxy for SES), race, and MVPA was explored, and a sensitivity analysis, excluding participants with a history of cancer (n=24,698), cardiovascular disease/stroke (n=23,488), or emphysema/chronic bronchitis (n=6,840) at baseline, was conducted.
      All the 160,296 participants who returned the 1997 survey were eligible for inclusion. Participants were excluded if their address could not be geocoded for the entire follow-up period (i.e., match score <60 for all years, n=2,269; excluded versus included participants did not differ in race or education). The remaining 158,027 participants, among whom 76,887 died between 1997 and 2016, were included in analyses.

      RESULTS

      Participants were 53.8% women and 97.5% White with a mean age of 68 (SD=6.3) years (Table 1). More than 10% of participants (n=16,861) lived in a rural area/small town. Rural residents were more likely to report high school as their highest educational attainment, be overweight/obese, and be physically inactive than residents of metropolitan and micropolitan areas. Residents of metropolitan areas had a higher average particulate matter with a diameter <2.5 µm exposure and were more likely to be former smokers than residents of rural and micropolitan areas.
      The risk of all-cause mortality was similar among metropolitan and micropolitan residents (Table 2). After adjustment for age and sex, residents of rural areas had a small but statistically significant elevated risk of all-cause mortality compared with metropolitan residents (HR=1.04; 95% CI=1.01, 1.06). Adjustment for additional covariates attenuated the association entirely (HR=0.99; 95% CI=0.97, 1.01). Individually, adjustment for education (HR=0.99; 95% CI=0.97, 1.01), alcohol use (HR=1.01; 95% CI=0.99, 1.04), and MVPA (HR=1.00; 95% CI=0.97, 1.02) independently eliminated the elevated risk for all-cause mortality for rural residents (Table 2) (relationships with specific causes of death are shown in Appendix Table 2, available online).
      Table 2Association Between Rural‒Urban Commuting Area Group and All-Cause Mortality
      Model covariatesMetropolitan (n = 126,570)Micropolitan (n = 14,596)Rural/small town (n = 16,861)
      N deaths60,5177,5148,856
       Sex and age adjusted1.00 (ref)1.02 (0.99–1.04)1.04 (1.01–1.06)
       Demographics
        + Education1.00 (ref)0.99 (0.96–1.01)0.99 (0.97–1.01)
        + Race1.00 (ref)1.02 (0.99–1.04)1.04 (1.01–1.06)
        + Marital status1.00 (ref)1.02 (0.99–1.04)1.04 (1.01–1.06)
        + Living arrangement1.00 (ref)1.01 (0.99–1.04)1.03 (1.01–1.06)
       Modifiable behaviors
        + BMI1.00 (ref)1.01 (0.99–1.03)1.03 (1.00–1.05)
        + Alcohol use1.00 (ref)1.00 (0.98–1.02)1.01 (0.99–1.04)
        + Cigarettes per day1.00 (ref)1.02 (1.00–1.05)1.04 (1.02–1.07)
        + Smoking duration1.00 (ref)1.02 (1.00–1.05)1.04 (1.02–1.07)
        + Time since quitting smoking1.00 (ref)1.02 (0.99–1.04)1.04 (1.02–1.06)
        + Chewing tobacco1.00 (ref)1.01 (0.99–1.04)1.03 (1.01–1.06)
        + Physical activity1.00 (ref)0.99 (0.97–1.02)1.00 (0.97–1.02)
       Health history
        + Comorbidity score1.00 (ref)1.02 (1.00–1.05)1.04 (1.02–1.06)
        + Aspirin use1.00 (ref)1.01 (0.99–1.04)1.03 (1.00–1.05)
       Environment
        + PM2.51.00 (ref)1.02 (1.00–1.05)1.04 (1.02–1.07)
        + Occupational dirtiness1.00 (ref)1.01 (0.99–1.04)1.02 (1.00–1.05)
        + Pesticide exposure1.00 (ref)1.01 (0.99–1.04)1.03 (1.01–1.05)
       Multivariable adjusted
      All-cause multivariable-adjusted model included age, sex, education, race, marital status, living arrangement, BMI, alcohol use, smoking (cigarettes/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. PM2.5, particulate matter with diameter <2.5 microns.
      1.00 (ref)0.98 (0.96–1.01)0.99 (0.97–1.01)
      Note: Hazard ratios in italics indicate individual covariates that attenuated the elevated age- and sex-adjusted mortality risk for rural/small town residents to null.
      a All-cause multivariable-adjusted model included age, sex, education, race, marital status, living arrangement, BMI, alcohol use, smoking (cigarettes/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure.PM2.5, particulate matter with diameter <2.5 microns.
      Results were largely similar when stratified by sex, race, MVPA, and baseline comorbidity status (Tables 3 and 4). Among participants with higher education, rural residence was associated with an elevated risk of mortality; among those with less formal education, rural residence was associated with a lower risk of mortality (p-interaction=0.007).
      Table 3Association Between Rural‒Urban Commuting Area Group and All-Cause Mortality by Sex, Educational Attainment, Race, and Physical Activity
      MetropolitanMicropolitanRural/small townMetropolitanMicropolitanRural/small town
      MenWomenp-interaction
      N deaths33,8134,2895,03926,7043,2253,817
      Age-adjusted HR1.00 (ref)1.03 (1.00–1.06)1.04 (1.01–1.08)1.00 (ref)1.00 (0.96–1.04)1.03 (0.99–1.06)0.48
      Multivariable adjusted
      Multivariable-adjusted models included age, sex, race, education, marital status, living arrangement, BMI, alcohol use, smoking (cigarettes/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. Hazard ratios in italics indicate that adjustment for individual-level covariates attenuated the elevated mortality risk to null. PM2.5, particulate matter with diameter <2.5 µm.
      1.00 (ref)0.98 (0.94–1.01)0.97 (0.94–1.00)1.00 (ref)0.99 (0.95–1.03)1.01 (0.97–1.05)0.31
      HS graduate or lessSome college or more
      N deaths18,5773,1524,42641,4924,3164,362
      Sex and age adjusted1.00 (ref)0.97 (0.93–1.00)0.96 (0.93–0.99)1.00 (ref)1.01 (0.98–1.05)1.03 (1.00–1.06)0.007
      Multivariable adjusted
      Multivariable-adjusted models included age, sex, race, education, marital status, living arrangement, BMI, alcohol use, smoking (cigarettes/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. Hazard ratios in italics indicate that adjustment for individual-level covariates attenuated the elevated mortality risk to null. PM2.5, particulate matter with diameter <2.5 µm.
      1.00 (ref)0.97 (0.93–1.00)0.97 (0.94–1.01)1.00 (ref)0.99 (0.96–1.03)1.00 (0.97–1.03)0.32
      Non-Hispanic/Latino WhiteAll other racial/ethnic groups
      N deaths58,9007,3908,7521,43610477
      Sex and age adjusted1.00 (ref)1.01 (0.99–1.04)1.04 (1.01–1.06)1.00 (ref)1.18 (0.96–1.45)1.02 (0.79–1.30)0.33
      Multivariable adjusted
      Multivariable-adjusted models included age, sex, race, education, marital status, living arrangement, BMI, alcohol use, smoking (cigarettes/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. Hazard ratios in italics indicate that adjustment for individual-level covariates attenuated the elevated mortality risk to null. PM2.5, particulate matter with diameter <2.5 µm.
      1.00 (ref)0.98 (0.96–1.01)0.99 (0.97–1.02)1.00 (ref)1.14 (0.92–1.40)0.87 (0.68–1.12)MetropolitanMicropolitanRural/small town0.21
      Physically active (7.5 MET-hours/week)Physically active (>0 to <7.5)Physically inactive (0)
      N deaths28,4293,0843,38820,7652,7213,2266,5651,0431,322
      Sex and age adjusted1.00 (ref)1.00 (0.97–1.04)1.01 (0.98–1.05)1.00 (ref)0.99 (0.95–1.03)1.00 (0.96–1.04)1.00 (ref)1.00 (0.94–1.07)0.95 (0.90–1.01)0.48
      Multivariable adjusted
      Multivariable-adjusted models included age, sex, race, education, marital status, living arrangement, BMI, alcohol use, smoking (cigarettes/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. Hazard ratios in italics indicate that adjustment for individual-level covariates attenuated the elevated mortality risk to null. PM2.5, particulate matter with diameter <2.5 µm.
      1.00 (ref)0.98 (0.94–1.02)0.98 (0.94–1.01)1.00 (ref)0.98 (0.94–1.03)1.00 (0.96–1.04)1.00 (ref)1.00 (0.94–1.08)0.98 (0.92–1.04)0.92
      a Multivariable-adjusted models included age, sex, race, education, marital status, living arrangement, BMI, alcohol use, smoking (cigarettes/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. Hazard ratios in italics indicate that adjustment for individual-level covariates attenuated the elevated mortality risk to null.PM2.5, particulate matter with diameter <2.5 µm.
      Table 4Sensitivity Analysis of Association between Rural-Urban Commuting Area Group and Mortality Excluding those with Comorbidities at Baseline
      MetropolitanMicropolitanRural/small town
      All-cause mortality
       N deaths35,7794,3685,175
       Sex and age adjusted HR1.00 (ref)1.00 (0.97, 1.03)1.04 (1.01, 1.07)
       Multivariable adjusted HR
      Multivariable adjusted models included: age, sex, race, education, marital status, living arrangement, BMI, alcohol use, smoking (cig/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. Hazard ratios in italics indicate that adjustment for individual-level covariates attenuated the elevated mortality risk to null. HR, hazard ratio; PM2.5, particulate matter with diameter <2.5 µm.
      1.00 (ref)0.96 (0.93, 1.00)0.99 (0.96, 1.02)
      a Multivariable adjusted models included: age, sex, race, education, marital status, living arrangement, BMI, alcohol use, smoking (cig/day, years since quit, duration), chewing tobacco, physical activity, comorbidity score, aspirin use, PM2.5 exposure, occupational dirtiness index, and pesticide exposure. Hazard ratios in italics indicate that adjustment for individual-level covariates attenuated the elevated mortality risk to null.HR, hazard ratio; PM2.5, particulate matter with diameter <2.5 µm.

      DISCUSSION

      In this study, there was a small but statistically significant elevated risk of death in rural compared with that among metropolitan residents. However, this elevated risk was eliminated after accounting for education, MVPA, and alcohol use. These findings suggest that individual characteristics, including modifiable behaviors, may at least partially explain the rural‒urban mortality disparity.
      The 4% rural‒urban mortality disparity seen in the age-/sex-adjusted model in this study was smaller than the disparity reported in other studies.
      • Cosby AG
      • McDoom-Echebiri MM
      • James W
      • Khandekar H
      • Brown W
      • Hanna HL.
      Growth and persistence of place-based mortality in the United States: the rural mortality penalty.
      • Cross SH
      • Califf RM
      • Warraich HJ.
      Rural-urban disparity in mortality in the U.S. from 1999 to 2019.
      • Singh GK
      • Siahpush M.
      Widening rural-urban disparities in all-cause mortality and mortality from major causes of death in the USA, 1969–2009.
      This may be because the nature and magnitude of the rural mortality disparity vary across the U.S. In fact, certain rural areas have recently seen a decline in mortality rates, whereas others have experienced an increase.
      • James WL.
      All rural places are not created equal: revisiting the rural mortality penalty in the United States.
      For example, relative to high SES and White rural residents, Black and/or lower SES rural residents have higher mortality rates.
      • James W
      • Cossman JS.
      Long-term trends in Black and White mortality in the rural United States: evidence of a race-specific rural mortality penalty.
      ,
      • James WL.
      All rural places are not created equal: revisiting the rural mortality penalty in the United States.
      CPS-IINC had a small percentage of both non-White racial/ethnic groups and low SES participants (although education was used as a proxy), so this study must be replicated in other subpopulations.
      Previous studies suggest that population-level poverty
      • Singh GK
      • Siahpush M.
      Widening rural-urban disparities in all-cause mortality and mortality from major causes of death in the USA, 1969–2009.
      and lack of a college education
      • Cosby AG
      • McDoom-Echebiri MM
      • James W
      • Khandekar H
      • Brown W
      • Hanna HL.
      Growth and persistence of place-based mortality in the United States: the rural mortality penalty.
      partially explain the mortality disparity between urban and rural residents, but studies exploring the potential role of individual-level modifiable behaviors are lacking. The attenuation of the rural mortality risk by MVPA and alcohol in this study suggests that these modifiable behaviors may be contributing to the larger problem. This finding justifies exploring opportunities for behavioral interventions in rural areas. The attenuation of risk with adjustment for education was confirmed in this study. Interestingly, among the highly educated, rural residence was associated with higher mortality; conversely, among those with less formal education, rural residence appeared to be protective. This could be because higher costs of living in urban (than in rural) areas exacerbate the mortality risk associated with lower levels of formal education (and perhaps a lower income). Research incorporating both individual-level and population-level data would likely provide further insight into this disparity.

      Limitations

      A strength of this study is the availability of individual-level data, although lack of income and residential history information before adulthood is a limitation. This study also consists exclusively of an older population and has a small number of non-White racial/ethnic groups and smokers, which may limit generalizability to other rural populations. The study population was also restricted to 21 states at baseline, which could be a limitation; however, as participants moved throughout follow-up, this study ended up including participants living in all the 50 states. The complexity of the impact of race alone and in combination with SES warrants further assessment of generalizability. Another strength of this study is the use of time-varying RUCA codes (updated as participants moved or as communities changed), which are finer in spatial resolution (census tract level) than several other common metrics.
      • Brooks MM
      • Mueller JT
      • Thiede BC.
      County reclassifications and rural-urban mortality disparities in the United States (1970–2018).
      These results are robust because a sensitivity analysis excluding those with a history of chronic disease at baseline produced similar point estimates.

      CONCLUSIONS

      In this study, the elevated risk of death for rural compared with that for metropolitan residents appeared to be largely explained by individual-level demographics and health behaviors. These results suggest that modifiable factors, such as MVPA and alcohol consumption, may play an important role in reducing the rural mortality disparity. Although replication in other subpopulations is necessary, these results underscore the need for regulatory efforts to provide safe opportunities for physical activity and guidance on alcohol consumption to reduce health disparities in rural populations.

      CRediT AUTHOR STATEMENT

      Erika Rees-Punia: Conceptualization; Methodology; Writing – original draft. Emily Deubler: Formal analysis, Writing - review and editing. Alpa V. Patel: Writing − review and editing. W. Ryan Diver: Methodology; Writing – review and editing. James Hodge: Writing - review and editing. Farhad Islami: Writing - review and editing. Minjee Lee: Writing - review and editing. Marjorie L. McCullough: Writing - review and editing. Lauren R. Teras: Methodology; Writing - review and editing.

      ACKNOWLEDGMENTS

      The authors express sincere appreciation to all Cancer Prevention Study-II participants and to each member of the study and biospecimen management group.
      The views expressed in this study are those of the authors and do not necessarily represent those of the American Cancer Society or the American Cancer Society Cancer Action Network. Cancer Prevention Study-II is approved by the Emory University IRB (IRB #00045780).
      The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study-II cohort.
      No financial disclosures were reported by the authors of this paper.

      Appendix. Supplementary materials

      REFERENCES

        • Islami F
        • Ward EM
        • Sung H
        • et al.
        Annual report to the nation on the status of cancer, Part 1: National cancer statistics.
        J Natl Cancer Inst. 2021; 113: 1648-1669https://doi.org/10.1093/jnci/djab131
        • GBD 2015 Mortality and Causes of Death Collaborators
        Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015.
        Lancet. 2016; 388: 1459-1544https://doi.org/10.1016/S0140-6736(16)31012-1
        • Singh GK
        • Siahpush M.
        Widening rural-urban disparities in life expectancy, U.S., 1969–2009.
        Am J Prev Med. 2014; 46: e19-e29https://doi.org/10.1016/j.amepre.2013.10.017
        • Cosby AG
        • McDoom-Echebiri MM
        • James W
        • Khandekar H
        • Brown W
        • Hanna HL.
        Growth and persistence of place-based mortality in the United States: the rural mortality penalty.
        Am J Public Health. 2019; 109: 155-162https://doi.org/10.2105/AJPH.2018.304787
        • Cross SH
        • Califf RM
        • Warraich HJ.
        Rural-urban disparity in mortality in the U.S. from 1999 to 2019.
        J Am Med Assoc. 2021; 325: 2312-2314https://doi.org/10.1001/jama.2021.5334
        • Singh GK
        • Siahpush M.
        Widening rural-urban disparities in all-cause mortality and mortality from major causes of death in the USA, 1969–2009.
        J Urban Health. 2014; 91: 272-292https://doi.org/10.1007/s11524-013-9847-2
        • Monnat SM.
        Trends in U.S. working-age non-Hispanic white mortality: rural-urban and within-rural differences.
        Popul Res Policy Rev. 2020; 39: 805-834https://doi.org/10.1007/s11113-020-09607-6
      1. Census urban and rural classification and urban area criteria. United States Census Bureau. https://www.census.gov/geo/reference/ua/urban-rural-2010.html. Accessed date October 4, 2021;https://www.census.gov/programs-surveys/geography/guidance/geo-areas/urban-rural/2010-urban-rural.html.

        • Byun SY
        • Meece JL
        • Irvin MJ.
        Rural-nonrural disparities in postsecondary educational attainment revisited.
        Am Educ Res J. 2012; 49https://doi.org/10.3102/0002831211416344
        • James W
        • Cossman JS.
        Long-term trends in Black and White mortality in the rural United States: evidence of a race-specific rural mortality penalty.
        J Rural Health. 2017; 33: 21-31https://doi.org/10.1111/jrh.12181
        • James WL.
        All rural places are not created equal: revisiting the rural mortality penalty in the United States.
        Am J Public Health. 2014; 104: 2122-2129https://doi.org/10.2105/AJPH.2014.301989
        • Matthews KA
        • Croft JB
        • Liu Y
        • et al.
        Health-related behaviors by urban-rural county classification - United States, 2013.
        MMWR Surveill Summ. 2017; 66: 1-8https://doi.org/10.15585/mmwr.ss6605a1
        • Lutfiyya MN
        • Shah KK
        • Johnson M
        • et al.
        Adolescent daily cigarette smoking: is rural residency a risk factor?.
        Rural Remote Health. 2008; 8: 875https://doi.org/10.22605/RRH875
        • Roberts ME
        • Doogan NJ
        • Kurti AN
        • et al.
        Rural tobacco use across the United States: how rural and urban areas differ, broken down by census regions and divisions.
        Health Place. 2016; 39: 153-159https://doi.org/10.1016/j.healthplace.2016.04.001
        • Whitfield GP
        • Carlson SA
        • Ussery EN
        • Fulton JE
        • Galuska DA
        • Petersen R.
        Trends in meeting physical activity guidelines among urban and rural dwelling adults — United States, 2008–2017.
        MMWR Morb Mortal Wkly Rep. 2019; 68: 513-518https://doi.org/10.15585/mmwr.mm6823a1
        • Calle EE
        • Rodriguez C
        • Jacobs EJ
        • et al.
        The American Cancer Society Cancer Prevention Study II Nutrition Cohort: rationale, study design, and baseline characteristics.
        Cancer. 2002; 94: 500-511https://doi.org/10.1002/cncr.10197
        • Johnson KM
        • Lichter DT.
        Metropolitan reclassification and the urbanization of rural America.
        Demography. 2020; 57 (/10/01): 1929-1950https://doi.org/10.1007/s13524-020-00912-5
        • Calle EE
        • Terrell DD.
        Utility of the National Death Index for ascertainment of mortality among cancer prevention study II participants.
        Am J Epidemiol. 1993; 137: 235-241https://doi.org/10.1093/oxfordjournals.aje.a116664
        • Brooks MM
        • Mueller JT
        • Thiede BC.
        County reclassifications and rural-urban mortality disparities in the United States (1970–2018).
        Am J Public Health. 2020; 110: 1814-1816https://doi.org/10.2105/AJPH.2020.305895