Evolution of the US nonmetropolitan mortality disadvantage by sex, state, and year, 1999‐2019

Abstract

Purpose

To examine disparities in nonmetropolitan and metropolitan mortality by state and sex from 1999 to 2019.

Methods

We calculate age‐standardized mortality rates for nonmetropolitan and metropolitan areas by state and sex and compute age‐standardized differences in these rates within each state and relative to the national average. We further estimate the number of excess deaths in nonmetropolitan areas by state. These are deaths that would have been avoided if nonmetropolitan areas had the same age‐specific death rates as metropolitan areas in the same state.

Findings

We document increasing nonmetropolitan mortality disadvantage since 1999 along with significant variation in the magnitude and timing of its emergence by state. Although stagnation in mortality was observed nationally in the mid‐2010s, this was not true in all states or in all metropolitan and nonmetropolitan areas. Additionally, mortality trends became progressively more discordant across and within states. Despite this heterogeneity, we document a steady increase in the number of nonmetropolitan excess deaths from 8,400 in 1999 to 47,000 in 2019, representing 9.0% of all nonmetropolitan deaths.

Conclusions

National‐level mortality trends mask geographic variation by nonmetropolitan and metropolitan areas within and across states. Further research is needed to identify factors that contribute to these divergent patterns.

INTRODUCTION

In the United States, life expectancy is shaped by the complex interplay between social, economic, health policy, and structural factors, some of which are consistent across geography, while others vary regionally, by state, and across rural and urban settings. ¹ , ² , ³ , ⁴ Since the late 20th century, a widening mortality disparity has emerged between nonmetropolitan and metropolitan areas in the United States. ⁵ , ⁶ , ⁷ By 1999, nonmetropolitan areas in the United States had substantially higher age‐standardized mortality rates than metropolitan areas. ⁶ Additionally, although national age‐standardized mortality rates decreased on average by 1.1% annually from 1999 to 2019, mortality rates declined much faster in large metropolitan areas (1.4%) compared to nonmetropolitan areas (0.5%). By 2019, age‐standardized mortality rates in the United States were 25.5% higher in nonmetropolitan areas compared to metropolitan areas.

Increases in the nonmetropolitan‐metropolitan mortality disparity in recent decades have been driven in large part by rising working‐age mortality rates in nonmetropolitan areas. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ While mortality rates decreased in large metropolitan areas among adults aged 24‐64 years between 1999 and 2019, mortality rates increased by 12.1% at these ages in nonmetropolitan areas during this period. ⁶ This concerning trend has generated a substantial body of literature examining the socioeconomic, political, health systems, and structural determinants of early deaths during working ages, with a particular focus on rural areas and the role of “deaths of despair” (eg, deaths from drug and alcohol poisonings and suicide). ¹² , ¹³ , ¹⁴

The distinct social, economic, and health care challenges faced by rural areas have also been implicated in the high prevalence of chronic diseases such as heart disease and diabetes, and unhealthy behaviors related to obesity and smoking. Recent research indicates that these mechanisms may have significantly contributed to recent changes in life expectancy nationally and in rural areas. ¹⁵ , ¹⁶ , ¹⁷ Availability of health care also varies between small towns and large cities. Between 2005 and 2020, 173 rural hospitals have closed, ¹⁸ a fact that some researchers have linked to increased mortality in rural areas. ¹⁹ Disparities in mortality are also evident across different types of rural communities, suggesting that rural health experiences and resulting mortality patterns are highly heterogeneous. ² , ³ , ⁴ , ²⁰ , ²¹ , ²²

At the same time, prior research has also confirmed substantial variation in mortality and longevity across US states. ²³ A growing body of work has called for an effort to “hypothesize upwards” to understand these state‐level differences, ²⁴ or focus on the important role that state policies may play in shaping divergent trends in mortality across states. ²⁵ , ²⁶ Indeed, a wide range of state policies have been tied to mortality, including health insurance coverage and affordability, disability benefits and support, social support for immigrants, worker benefits in the form of minimum wage, sick time, paid leave, and unemployment compensation, gun control, abortion, and criminal justice. ²⁴

Despite these well‐documented inequalities across US states and between nonmetropolitan‐metropolitan areas, few studies have simultaneously examined nonmetropolitan‐metropolitan disparities within and across states. Yet, it is possible that state‐level policies do not have a uniform impact on mortality across the nonmetropolitan‐metropolitan continuum. While recent research has examined the emergence of a nonmetropolitan disadvantage nationally and by Census regions with varying specifications of nonmetropolitan areas, ⁶ , ²¹ , ²⁷ less is known about how nonmetropolitan‐metropolitan disparities in mortality have evolved at the state level over the last 2 decades.

This study contributes to this literature and extends prior studies by examining temporal trends in nonmetropolitan and metropolitan mortality by sex, year, and state between 1999 and 2019. For brevity, we subsequently refer to nonmetropolitan as nonmetro and to metropolitan as metro. We quantify nonmetro‐metro disparities in each state and trace their evolution over time using a variety of metrics that each contribute a different dimension to our understanding of these disparities. Additionally, we compute the number of excess deaths in nonmetro areas that would have been avoided if nonmetro areas within each state experienced the age‐specific mortality rates of the metro areas. These descriptive findings reveal considerable heterogeneity within and across states and caution against making broad generalizations about mortality trends in nonmetro and metro areas.

DATA AND METHODS

We use publicly available death counts and population estimates from 1999 to 2019 by metro status, year, state, sex, and 5‐year age groups (<1, 1‐4, 5‐9, 10‐14, …, 85+) from the National Center for Health Statistics. We extract the data using the Centers for Disease Control and Prevention Wide‐ranging Online Data for Epidemiologic Research (CDC WONDER) tool and construct age‐specific and age‐standardized mortality rates by state for metro counties (large central metros, large fringe metros, and medium and small metros) and all nonmetro counties combined. The metro‐nonmetro codes are based on the 2013 NCHS Urban‐Rural Classification Scheme for counties which uses the Office of Management and Budget's February 2013 delineation of metropolitan statistical areas and micropolitan statistical areas. ²⁸ Delaware, New Jersey, Rhode Island, and the District of Columbia are excluded from the study, including from national totals, because they contain no nonmetro counties according to this classification. All age‐standardized death rates are calculated using the average national age distribution over the study period, 1999‐2019 as a standard. This age distribution is reported in Table S1.

We first compute age‐standardized death rates (ASCDR) separately for nonmetro and metro areas by sex and year, nationally, and by state. Second, we examine the deviations of the state‐specific metro and nonmetro ASCDRs from the national age‐standardized rate. This measure compares state‐ and metro‐specific rates to a common standard which facilitates the identification of areas where rates are particularly low or high. We call this measure area‐specific excess mortality rate. In the equation below, we denote the age‐standardized mortality rate for area $a$, year $t$ with $ASCD{R}_{a,t}$, and the national age‐standardized mortality rates as $ASCD{R}_{NAT,t}$, we have:

$$AreaSpecificExcessMortalityRat{e}_{a,t}=ASCD{R}_{a,t}-ASCD{R}_{NAT,t}$$

Positive area‐specific excess mortality reflects mortality that is higher than the national average, whereas negative area‐specific excess mortality corresponds to mortality that is below the national average. We perform this analysis for 3 years (1999, 2009, and 2019) to describe how the geography of US mortality has changed in the last 2 decades.

Third, we calculate the difference in nonmetro ASCDR and metro ASCDR within each state and call the resulting difference nonmetro excess mortality rate. In the equation below, we denote the nonmetro age‐standardized mortality rate for state $s$, year $t$ with $ASCD{R}_{Nonmetro,s,t}$, and the corresponding metro rates with $ASCD{R}_{Metro,s,t}$, we have:

$$NonmetroExcessMortalityRat{e}_{s,t}=ASCD{R}_{Nonmetro,s,t}-ASCD{R}_{Metro,s,t}$$

This measure captures within‐state trends in nonmetro‐metro mortality differences. Our fourth measure translates the difference in age‐specific death rates between nonmetro and metro areas within states to excess deaths. We calculate excess deaths by sex, year, and state as the difference between deaths that occurred in nonmetro areas and those that would have occurred if the nonmetro areas had the age‐specific death rates of metro areas. We call this measure nonmetro excess deaths.

RESULTS

Trends in metro and nonmetro mortality between 1999 and 2019

Figure 1 shows age‐standardized mortality rates for the United States and by state from 1999 to 2019 for metro and nonmetro areas by sex. During this period, most states experienced continuous mortality declines in both metro and nonmetro areas. In most states, reductions were greater in metro areas compared to nonmetro areas and among males compared to females. Nationally, ASCDRs decreased in both nonmetro and metro areas until approximately 2014, at which time both rates stagnated. In some states, metro and nonmetro mortality rates continued to decline after 2014, whereas in other states, they also leveled off. Figure 1 additionally shows that ASCDRs in nonmetro areas exhibited more variability (measured here as the difference between the 10th and the 90th percentiles) than did metro areas. Rates for males are also more variable than for females. Finally, while variability increased for females in both metro and nonmetro areas, it declined in metro areas for males but remained stable in nonmetro areas (see number at the bottom of each panel in Figure 1).

FIGURE 1.

Nonmetro and metro age‐standardized mortality rates (per 100,000 residents) by sex and state, 1999‐2019. Notes: Each line in the graph represents a state's ASCDR trajectory from 1999 to 2019. The thicker black line represents the average of all states, and the 2 dashed lines represent the 10th and 90th percentiles. The numbers reported at the bottom of each graph are the average difference in age‐standardized mortality rates between the 10th and the 90th quantile in the periods 1999‐2004, 2005‐2009, 2010‐2014, and 2015‐2019. The dashed vertical lines indicate 2014, the year in which we observed a marked slowdown or increase in age‐standardized mortality rates nationally.

Figure S1 presents the mortality trends for each state further showing that the 2014 threshold was not an equally meaningful inflection point in all states. For example, male mortality stabilized or started rising as early as 2011 in nonmetro areas of California, Texas, Virginia, Ohio, Pennsylvania, South Dakota, and Kansas. In several states, for example, Utah, Iowa, West Virginia, and Michigan, ASCDRs in metro areas also stagnated or rose starting in 2011. Conversely, metro areas in California, Texas, Illinois, Georgia, Virginia, and New York did not display a mortality increase and their slowdown in the rate of mortality decline was small compared to the national average.

The evolution of the nonmetro mortality disadvantage

Table 1 reports national‐level nonmetro and metro ASCDRs by sex and year. ASCDRs were higher in nonmetro than metro areas for both males and females throughout the period, except for females in 2000. The difference between female nonmetro and metro ASCDRs grew from 6.0 deaths per 100,000 female residents in 1999 to 56.7 in 2019. The respective figures for males are 38.8 per 100,000 male residents in 1999 and 75.5 in 2019.

TABLE 1.

Age‐standardized mortality rates (deaths per 100,000 residents) for nonmetro and metro areas by sex and year, 1999‐2019.

	Females			Males
	Age‐standardized mortality rates (per 100,000 female residents)			Age‐standardized mortality rates (per 100,000 male residents)
	Metro	Nonmetro	Nonmetro‐metro gap	Metro	Nonmetro	Nonmetro‐metro gap
1999	783.27	789.27	6.00	1,149.53	1,188.37	38.84
2000	784.97	784.85	−0.12	1,134.90	1,167.27	32.38
2001	775.81	782.02	6.21	1,114.59	1,148.06	33.47
2002	774.03	790.31	16.27	1,111.11	1,151.84	40.72
2003	767.44	787.64	20.20	1,090.34	1,134.86	44.52
2004	733.91	758.06	24.14	1,041.90	1,093.47	51.57
2005	739.92	763.70	23.77	1,049.15	1,102.91	53.76
2006	721.21	749.98	28.77	1,016.05	1,071.14	55.09
2007	704.43	742.39	37.95	993.11	1,050.86	57.75
2008	706.83	746.94	40.11	988.33	1,059.93	71.60
2009	681.21	727.08	45.86	960.36	1,027.30	66.93
2010	684.59	726.19	41.59	961.54	1,023.70	62.16
2011	681.29	726.51	45.22	948.71	1,016.95	68.23
2012	674.53	718.26	43.73	936.53	1,006.60	70.07
2013	673.15	724.72	51.57	938.73	1,008.15	69.41
2014	669.33	722.25	52.92	933.90	1,004.24	70.34
2015	681.11	736.81	55.71	944.01	1,023.18	79.18
2016	673.56	727.16	53.59	940.79	1,018.53	77.73
2017	678.62	729.82	51.20	945.12	1,022.15	77.04
2018	668.54	721.67	53.13	936.58	1,016.75	80.17
2019	662.34	719.11	56.77	933.22	1,008.69	75.47

Figure 2A illustrates how the relationships between states’ metro and nonmetro ASCDRs varied over four 5‐year periods by sex between 1999 and 2019. In each period, mortality rates for nonmetro and metro areas were highly correlated across states (Pearson correlation coefficient equal or larger than 0.82). The correlation was similarly strong for females and males but weakened as time progressed (from 0.89 to 0.84 for females and from 0.89 to 0.82 for males) with nonmetro and metro mortality increasingly diverging in many states. Over the entire period, 23 states maintained a nonmetro disadvantage for females and 34 for males, whereas 5 states for females (Colorado, Connecticut, Maryland, Massachusetts, and Nebraska) and 5 for males (Colorado, Connecticut, Massachusetts, Michigan, and Wyoming) had a nonmetro advantage in each period (Table S2). The remaining 19 states for females and 8 states for males experienced 1 or more transitions from nonmetro advantage to disadvantage or vice versa. Most of these transitions involved going from a nonmetro advantage to a disadvantage but some were in the opposite direction (ie, Idaho for males and females and West Virginia for males). Overall, the number of states recording a nonmetro mortality disadvantage increased over time for females (1999‐2004: 28 states; 2005‐2009: 35 states; 2010‐2014: 37 states; 2015‐2019: 39 states) but remained relatively constant for males (1999‐2004: 37 states; 2005‐2009: 42 states; 2010‐2014: 40 states; 2015‐2019: 39 states) (Table S2).

FIGURE 2.

Nonmetro and metro age‐standardized mortality rates (per 100,000 residents, Panel A) and their change (Panel B) by sex, period, and state, 1999‐2019. Notes: Each point in the graph represents a state. In Panel A, the x‐axis measures the metro ASCDR and the y‐axis measures the nonmetro ASCDR. In Panel B, the x‐axis represents change in the ASCDR of metro areas and the y‐axis represents change in the ASCDR of nonmetro areas. The numbers reported on the bottom right corner of each plot in Panel A represent the correlation coefficients of the points in that subplot.

Figure 2B illustrates the relationship between the changes in ASCDRs in nonmetro and metro areas over the same 4 consecutive 5‐year periods by sex. Initially, the widening mortality gap between nonmetro and metro areas nationally was largely the result of faster mortality declines in metro areas than in nonmetro areas. Between 1999‐2004 and 2005‐2009, 41 states for females and 38 states for males experienced faster mortality declines in metro areas compared to nonmetro areas. This pattern became less consistent over time (Table 2). In states where mortality declined between 2010‐2014 and 2015‐2019 in both metro and nonmetro areas, the declines were generally faster in metro areas (10 of 13 states for females; 10 of 12 for males). In states where ASCDRs increased in both metro and nonmetro areas, the increases were generally faster in nonmetro areas in most states (7 of 13 for females; 12 of 15 states for males).

TABLE 2.

State patterns of mortality change in nonmetro and metro areas by sex and period, 1999‐2019.

	Number of states
Pattern of mortality change	Female	Male
From 1999‐2004 to 2005‐2009
Decline in both (faster decrease in metro areas)	41	38
Decline in both (faster decrease in nonmetro areas)	5	9
Decline in metro areas only (increase in nonmetro areas)	1	0
Decline in nonmetro areas only (increase in metro areas)	0	0
Increase in both (faster increase in nonmetro areas)	0	0
Increase in both (faster increase in metro areas)	0	0
From 2005‐2009 to 2010‐2014
Decline in both (faster decrease in metro areas)	37	32
Decline in both (faster decrease in nonmetro areas)	6	15
Decline in metro areas only (increase in nonmetro areas)	3	0
Decline in nonmetro areas only (increase in metro areas)	0	0
Increase in both (faster increase in nonmetro areas)	1	0
Increase in both (faster increase in metro areas)	0	0
From 2010‐2014 to 2015‐2019
Decline in both (faster decrease in metro areas)	10	10
Decline in both (faster decrease in nonmetro areas)	3	2
Decline in metro areas only (increase in nonmetro areas)	13	15
Decline in nonmetro areas only (increase in metro areas)	6	7
Increase in both (faster increase in nonmetro areas)	12	7
Increase in both (faster increase in metro areas)	3	6

Table 2 reports the number of states by direction of mortality changes (increase or decrease) for nonmetro and metro areas by sex and period. It documents how growing heterogeneity in mortality within states (comparing nonmetro areas to metro areas) occurred alongside escalating differences in mortality among states. When comparing mortality rates in 1999‐2004 to 2005‐2009 and rates in 2005‐2009 to 2010‐2014, states experienced similar patterns. All states for males and 43 of 47 for females experienced mortality declines in both metro and nonmetro areas in these 2 periods. However, in the last period from 2010‐2014 to 2015‐2019, the patterns became more variable. In 13 states for females and 12 for males, mortality continued to decline in both metro and nonmetro areas. Most states, in contrast, underwent mortality increases in either or both metro and nonmetro areas. Among them, 13 states for females and 15 for males experienced mortality increases in nonmetro areas but declines in metro areas, and in 15 states for females and 13 for males, mortality increased in both. Finally, a smaller number of states, 6 for females and 7 for males, had mortality increases in metro areas but declines in nonmetro areas.

Geography of US mortality and of the nonmetro mortality disadvantage

Figure 3 compares ASCDRs for nonmetro and metro areas within each state to the national average mortality rate (area‐specific excess mortality rate). Positive rates, indicating higher mortality than the national average, were concentrated in the South in both metro and nonmetro areas in 1999. Many states in the West, the Northeast, and several in the Midwest, had lower mortality than the national average in both metro and nonmetro areas. This same pattern was also visible in 2009 and 2019 but with some notable changes. First, area‐specific excess mortality rates in nonmetro areas in many Southern states increased more than in metro areas. Additionally, the number of states whose nonmetro areas had mortality rates above the national average increased between 1999 and 2019. As a result, the nonmetro mortality disadvantage deepened in some states and emerged in others. States with a nonmetro mortality disadvantage were geographically scattered in 1999, but they became increasingly concentrated in the South and the Pacific regions in 2009 and 2019. In other cases, the lack of nonmetro disadvantage reflected high metro mortality rates rather than low nonmetro mortality rates, for example, West Virginia.

FIGURE 3.

Comparison of state‐level age‐standardized mortality rates (per 100,000 residents) in nonmetro and metro areas to the average national age‐standardized mortality rate in 1999, 2009, and 2019.

Nonmetro excess deaths

Table 3 presents estimates of nonmetro excess deaths by sex and Census division. ²⁹ Nationally, the number of nonmetro excess deaths increased steadily between 1999 and 2019. Among females, nonmetro excess deaths increased from −160 deaths in 1999 (−0.1% of nonmetro deaths) to 14,619 deaths in 2009 (6.7%), and to 21,652 deaths in 2019 (9.3%). For males, nonmetro excess deaths increased from 8,591 deaths in 1999 (3.8%) to 18,175 deaths in 2009 (8.2%), and to 26,079 deaths in 2019 (10.3%).

TABLE 3.

Nonmetro excess deaths, excess death rates (per 100,000 residents), and relative excess mortality by census division, 1999, 2009, and 2019.

	1999			2009			2019
Census division	Excess deaths	Excess death rate	Relative excess (%)	Excess deaths	Excess death rate	Relative excess (%)	Excess deaths	Excess death rate	Relative excess (%)
Female
New England	−66	−8	−0.79	320	36	4.06	209	24	2.37
Middle Atlantic	224	15	1.42	1,121	77	7.99	1,926	139	13.93
East North Central	−1,407	−32	−3.02	1,006	23	2.35	2,862	67	6.30
West North Central	−907	−28	−2.34	681	21	2.01	1,744	54	5.11
South Atlantic	1,601	46	4.50	3,835	103	11.06	5,571	150	14.77
East South Central	−366	−12	−1.12	3,310	106	10.96	4,107	133	12.54
West South Central	154	5	0.45	3,204	103	10.76	3,782	122	12.32
Mountain	294	17	2.28	407	21	2.92	162	8	0.98
Pacific	313	26	3.00	736	56	6.81	1,287	93	10.74
United States	−160	−1	−0.07	14,619	63	6.71	21,652	94	9.34
Male
New England	44	5	0.58	149	17	1.90	182	21	1.89
Middle Atlantic	249	17	1.66	712	48	5.05	1,475	104	9.75
East North Central	−555	−13	−1.25	897	21	2.09	2,549	59	5.20
West North Central	−144	−4	−0.40	1,608	50	5.00	2,501	77	7.05
South Atlantic	4,037	119	11.91	4,916	134	14.09	7,002	192	17.49
East South Central	1,538	53	4.92	4,427	147	14.60	5,258	176	14.83
West South Central	2,126	71	6.57	3,693	118	11.88	5,189	164	15.13
Mountain	726	41	5.02	707	36	4.47	457	22	2.27
Pacific	569	46	4.92	1,067	78	8.72	1,466	104	10.20
United States	8,591	39	3.79	18,175	79	8.21	26,079	113	10.28

Note: Excess deaths are deaths that would be avoided or added if within states nonmetro areas had the same age‐specific mortality rates as metro areas. Relative excess mortality refers to excess deaths divided by the expected deaths in an area (ie, excess nonmetro deaths divided by observed nonmetro deaths). State‐level excess mortality results were aggregated to the Census Division level in this table.

These national numbers, however, hide considerable geographic variation in nonmetro excess deaths. While there were virtually no nonmetro excess deaths nationally in 1999 among females, there were 1,601 excess deaths in the South Atlantic division. In contrast, mortality was higher in metro areas in the East North Central and in the West North Central divisions more than offsetting the excess deaths in the South Atlantic division (Table 3). By 2019, in all divisions, female mortality was higher in nonmetro than metro areas with excess deaths ranging from 5,571 in the South Atlantic division to 262 in the Mountain division. Close to two‐thirds of the excess deaths (62%) occurred in the Southern United States (South Atlantic, East South Central, and West South Central divisions). Among males, mortality in nonmetro areas exceeded metro mortality nationally in 1999, but not in all divisions. Nonmetro areas had lower mortality in the East North Central and in West North Central divisions, but higher mortality in the South Atlantic (4,037 excess deaths), West South Central (2,126), and East South Central divisions (1,538). As was the case with female mortality, nonmetro male excess deaths grew over time across all divisions. In 2019, the highest number of nonmetro excess deaths were found in the South Atlantic, East South Central, and West South Central divisions.

DISCUSSION

Prior research has highlighted states as an important unit for studying geographic inequalities in mortality, ²⁴ , ²⁶ while extensive attention has also been paid to the stagnation of US mortality decline in the mid‐2010s. ⁹ , ³⁰ This study contributes to this literature by demonstrating substantial heterogeneity in these trends by state and between nonmetro and metro areas within states.

While at the national level mortality declines stagnated starting in 2014, we showed that this process began much earlier in some states and varied between metro and nonmetro areas within states. Additionally, while mortality was declining in both nonmetro and metro areas in almost all states up to 2014, this concordance in state‐level trends subsequently disappeared. In 13 states for females and 12 for males, mortality continued to decline in both nonmetro and metro areas until 2019, but in 15 states for females and 13 for males, mortality increased in both. In the remaining states, 19 for females and 22 for males, these patterns were mixed.

As in prior works, ⁵ , ⁶ we document growing nonmetro‐metro disparities in mortality. However, we also show that there was considerable variation in these disparities across states and that this state‐level heterogeneity grew over time. For example, while a nonmetro mortality disadvantage existed nationally in 1999, this pattern did not hold in 19 states for females and 10 states for males. Similarly, while by 2019, we find a nonmetro mortality disadvantage in most states for both females and males, the size of the disparities shows substantial geographic variation, with the largest differences concentrated in the Southern states.

Beyond documenting the heterogeneity of trends in mortality declines across states, we also quantify their consequences for the nonmetro population. We calculate that as a result of the widening nonmetro mortality disadvantage, the number of nonmetro excess deaths grew substantially between 1999 and 2019. By the end of the period, there were more than 47,000 excess deaths in nonmetro areas (21,652 for females and 26,079 for males). These deaths would have been avoided if death rates in nonmetro areas within states had been equal to those in metro areas and corresponded to 9.0% of all deaths in nonmetro areas in 2019. This number is roughly equivalent to all nonmetro deaths from accidents (ICD 10 codes V01‐X59, Y85‐Y86) including drug poisoning, intentional self‐harm (U03, X60‐X84, Y87.0), and chronic liver disease and cirrhosis, associated with alcohol consumption (K70, K73‐K74).

Prior studies have increasingly highlighted the need to examine the macrostructural conditions that shape geographic inequalities in mortality and longevity, identifying a variety of state‐level policies associated with geographic variation in mortality. For example, Montez and colleagues highlighted a variety of broad state policy domains, including tobacco, labor, immigration, civil rights, and the environment, as being key levers for improving life expectancy. Other work has documented the health and longevity benefits associated with specific policies, including earned income tax credit, ³¹ , ³² Medicaid expansion, ³³ , ³⁴ minimum wage, ³⁵ , ³⁶ and immigrant incorporation policies. ³⁷ Yet, very little is known about how state policy may influence nonmetro‐metro disparities within states. Our results suggest that these policies might have differential impacts in nonmetro and metro areas, underscoring this as a promising avenue for future research. Further studies are needed to identify policies that have differential impacts in nonmetro and metro areas and whether targeted interventions would be needed to reduce geographic disparities within states.

Limitations

This study has several limitations. First, we are using a binary classification for nonmetro and metro areas and thus potentially hiding variation within metro and nonmetro areas. Second, we do not distinguish among racial and ethnic groups. Third, we do not distinguish among causes of death. These choices were driven by our goal of examining variation within and across states and finer categorizations would have severely limited the number of states we could include in the analyses.

CONCLUSIONS

Despite these limitations, our analysis provides new insights into mortality trends across metro and nonmetro areas within and across states. The findings highlight the importance of disaggregating national trends by geography.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Supporting Information

Paglino E, Elo IT, Preston SH, Hempstead K, Stokes AC. Evolution of the US nonmetropolitan mortality disadvantage by sex, state, and year, 1999‐2019. J Rural Health. 2025;41:e70040. 10.1111/jrh.70040

DATA AVAILABILITY STATEMENT

Replication materials for the paper are available in an online repository at https://osf.io/4hr7f/.