Trends And Patterns Of Disparities in Diabetes And Chronic Kidney Disease Mortality Among US Counties, 1980–2014

Abstract Introduction: Diabetes and chronic kidney diseases are associated with a large health burden in the USA and globally. 

Objective: To estimate age-standardized mortality rates by county from diabetes mellitus and chronic kidney disease.

Design and setting: Validated small area estimation models were applied to de-identified death records from the National Center for Health Statistics (NCHS) and population counts from the census bureau, NCHS, and the Human Mortality Database to estimate county-level mortality rates from 1980 to 2014 from diabetes mellitus and chronic kidney disease (CKD). 

Exposures: County of residence. Main outcomes and measures: Age-standardized mortality rates by county, year, sex, and cause. Results: Between 1980 and 2014, 2,067,805 deaths due to diabetes were recorded in the USA. The mortality rate due to diabetes increased by 33.6% (95% UI: 26.5%–41.3%) between 1980 and 2000 and then declined by 26.4% (95% UI: 22.8%–30.0%) between 2000 and 2014. Counties with very high mortality rates were found along the southern half of the Mississippi River and in parts of South and North Dakota, while very low rates were observed in central Colorado, and select counties in the Midwest, California, and southern Florida. A total of 1,659,045 deaths due to CKD were recorded between 1980 and 2014 (477,332 due to diabetes mellitus, 1,056,150 due to hypertension, 122,795 due to glomerulonephritis, and 2,768 due to other causes). CKD mortality varied among counties with very low mortality rates observed in central Colorado as well as some counties in southern Florida, California, and Great Plains states. High mortality rates from CKD were observed in counties throughout much of the Deep South, and a cluster of counties with particularly high rates was observed around the Mississippi River.

Conclusions and Relevance: This study found large inequalities in diabetes and CKD mortality among US counties. The findings provide insights into the root causes of this variation and call for improvements in risk factors, access to medical care, and quality of medical care. 

Keywords: Diabetes, Chronic kidney disease, Disparities, Mortality

Supportive Service Of Wecistanche-The largest cistanche exporter in the China:

Email:wallence.suen@wecistanche.com 

Whatsapp/Tel:+86 15292862950

Shop For More Specifications Details:

https://www.xjcistanche.com/cistanche-shop

Introduction 

Diabetes mellitus accounted for 77.7 thousand deaths (2.6% of all deaths) and 4.46 million disability-adjusted life years (DALYs) in 2019 in the USA [1, 2]. Diabetes prevalence has increased rapidly in the USA in recent decades, reaching 11.8% in 2019 [2, 3]. Diabetes is associated with several diseases including chronic kidney disease (CKD). CKD was the 6th leading cause of death in 2019, accounting for 3.6% of all deaths [2]. In 1990, CKD, which is preventable by adequate medical care, was the 14th leading cause of death, accounting for 1.5% of all deaths [2].

The prevalence of diabetes has increased despite previous calls for action. Recent data from the Behavioral Risk Factor Surveillance System (BRFSS), a large state-based surveillance system, show that the self-reported prevalence of diagnosed diabetes in 2020 was 10.6% among adults aged 18 or older. West Virginia had the highest prevalence (15.7%) and Alabama 14.8%, while the District of Columbia (7.5%) and Colorado (7.6%) had the lowest rates [4].

Obesity increased in all states from 1990 to 2020 [4]. Obesity is a major risk factor for type 2 diabetes, and there is a signifcant association between weight gain and diabetes incidence [5–7]. The prevalence of obesity it is likely to continue to rise in the years ahead unless effective interventions are implemented. Furthermore, diabetes is associated with a high medical cost [8]. Behavioral and metabolic risk factors such as poor diet and lack of physical activity are also risk factors for type 2 diabetes [9, 10]. Therefore, diabetes is expected to increase rapidly in the coming decades due to the aging and growth of the US population, poor diet, obesity, and low physical activity [11, 12].

Methods

The methods used for this analysis were previously reported in detail elsewhere and are described briefly here [15]. This research received institutional review board approval from the University of Washington. Informed consent was not required because the study used de-identified data and was retrospective.

Data 

This analysis used de-identified death records from the National Center for Health Statistics (NCHS) [16] and population counts from the Census Bureau [17], NCHS [18–20], and the Human Mortality Database [21]. Deaths and population were tabulated by county, age group (0, 1–4, 5–9, …, 75–79, and ≥80), sex, year, and cause. County-level information on levels of education, income, race/ethnicity, Native American reservations, and population density derived from data provided by the Census Bureau and NCHS was utilized as covariates (more detail on these data sources is available in Additional file 1: eTable S1 in the supplement). These variables were selected based on data availability because we expect that these variables are likely to be predictive of mortality across a range of causes. In a small number of cases, counties were combined to ensure historically stable units of analysis (Additional file 1: eTable S2).

Cause list and garbage redistribution

The study used the cause list developed for the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) [22]. Tis cause list is arranged hierarchically in four levels, and within each level, the list is exhaustive and mutually exclusive. Additional file 1: eTable  S3 in the supplement lists all causes in the GBD cause list and the ICD9 and ICD10 codes corresponding to each cause. The focus of this study was on diabetes mellitus and chronic kidney disease. Chronic kidney disease was also subdivided into chronic kidney disease due to diabetes mellitus, chronic kidney disease due to hypertension, chronic kidney disease due to glomerulonephritis, and chronic kidney disease due to other causes. Although the focus of this study was diabetes mellitus and chronic kidney disease, all causes of death in the GBD cause list were analyzed concurrently.

Previous studies have documented the existence of insufficiently specific or implausible causes of death used in death registration data that may lead to misleading geographic and temporal patterns [23]. Algorithms developed for the GBD were used to reallocate deaths and assign one of these "garbage codes" to plausible alternatives [22]. First, plausible target causes were assigned to each garbage code or group of garbage codes. Second, deaths were reassigned to specified target codes according to proportions derived in one of four ways: (1) published literature or expert opinion; (2) regression models; (3) according to the proportions initially observed among targets; and (4) for HIV/AIDS specifically, by comparison with years before HIV/AIDS became widespread.

Small area models 

The number of deaths observed in a given county, year, age, and sex is typically small, and the directly observed mortality rates at this level are often highly unstable. We use a small area estimation model to stabilize these estimates by "borrowing strength" across counties, periods, age groups, and external information (covariates). Tis model has been previously validated and shown to perform well even for counties with relatively small populations. [15]

Bayesian, spatially explicit mixed effects regression models were estimated for each cause in the GBD hierarchy, separately for males and females. Te model for each cause was specified as:

where Dj,t,a, Pj,t,a and mj,t,a are the number of deaths, the population, and the underlying mortality rate, respectively, for county j, year t, and age group a. The model for mj,t,a contained six components: an intercept (β0), fixed covariate effects (β1), random age-time effects (γ1,a,t), random spatial effects (γ2,j), random space–time effects (γ3,j, and γ4,j,t), and random space-age effects (γ5,j and γ6,j,a). Te model incorporated seven covariates: the proportion of the adult population who graduated high school, the proportion of the population that is Hispanic, the proportion of the population that is Black, the proportion of the population that is a race other than Black or White, the proportion of a county that is contained within a state or federal Native American reservation, the median household income, and the population density. γ 1, γ 2, γ 3, and γ 5 were assumed to follow conditional autoregressive distributions, which allow for smoothing over adjacent age groups and years (γ 1) or counties (γ 2, γ 3, and γ 5). [24, 25]. γ 4 and γ 6 were assumed to follow independent mean-zero normal distributions.

Models were ft using the Template Model Builder Package [26] in the R version 3.2.4 statistical software (R Foundation for Statistical Computing). One thousand draws of mj,t,a were taken from the approximated posterior distribution. Tese draws were raked [27] (i.e., scaled along multiple dimensions) to ensure consistency between levels of the cause hierarchy and to ensure consistency with national estimates from the GBD  [22]. After raking, age-standardized mortality rates were calculated using the US 2010 census population as the standard, and years of life lost (YLLs) were calculated for each age group by multiplying the mortality rate by population by life expectancy at the average age at death in the reference life table used in the GBD [22] and then summing across all ages. Point estimates were calculated from the mean of all draws, and 95% uncertainty intervals were calculated from the 2.5th and 97.5th percentiles. Changes over time were considered statistically signifcant if the posterior probability of an increase (or decrease) was ≥ 95%. Code for fitting the small area models is available from the authors upon request.

Results

Deaths, years of life lost, and age-standardized mortality rates at the national level, and the distribution of age-standardized mortality rates at the county level by cause in 2014 are presented in Table 1. Chronic kidney disease had the highest mortality rate in 2014 (22.4 [95% UI: 21.4– 23.3] deaths per 100,000 population) followed by diabetes (19.7 [95% UI: 18.9–20.6] deaths per 100,000 population).

Diabetes 

Between 1980 and 2014, 2,067,805 deaths due to diabetes were recorded in the USA. The mortality rate due to diabetes increased by 33.6% (95% UI: 26.5–41.3%) between 1980 and 2000 and then declined by 26.4% (95% UI: 22.8–30.0%) between 2000 and 2014. Te age-standardized mortality rate from diabetes was 20.1 (95% UI: 19.2–21.0), 22.4 (95% UI: 21.8–23.1), 26.8 (95% UI: 26.1– 27.6), and 19.7 (95% UI: 18.9–20.6) deaths per 100,000 population in 1980, 1990, 2000, and 2014, respectively. Counties with very high mortality rates were found along the southern half of the Mississippi River and in parts of South and North Dakota (Fig. 1). On the other hand, counties with very low rates were observed in central Colorado, and select counties in the Midwest, California, and southern Florida. Among counties, the lowest estimated mortality rate in 2014 was observed in Summit County, Colorado (2.4 [95% UI: 1.9–2.8] deaths per 100,000 population), while the highest was observed in Oglala Lakota County, South Dakota (118.7 [95% UI: 106.2–132.1] deaths per 100,000 population). Most countries experienced an increase in the mortality rate due to diabetes between 1980 and 2014 (65.9%; statistically significant in 51.6%), but counties where the mortality rate declined were found in most states. Similar geographic patterns were observed for males and females in 2014 (Additional fle 1: eFigs. S1 and S2); however, the mortality rate from diabetes increased by 12.5% (95% UI: 5.4– 20.2%) among males and decreased by 14.5% (95% UI: 7.0–22.7%) among females between 1980 and 2014.

Chronic kidney disease

A total of 1,659,045 deaths due to chronic kidney disease were recorded between 1980 and 2014 (477,332 due to diabetes mellitus, 1,056,150 due to hypertension, 122,795 due to glomerulonephritis, and 2768 due to other causes). Chronic kidney disease mortality varied among counties (Fig. 2). Very low mortality rates were observed in central Colorado as well as some counties in southern Florida, California, and Great Plains states. On the other hand, high mortality rates were observed in counties throughout much of the Deep South and a cluster of counties with particularly high rates was observed around the Mississippi river. Te lowest estimated mortality rate in 2014 was observed in Summit County, Colorado (4.9 [95% UI: 4.4–5.6] deaths per 100,000 population), while the highest was observed in East Carroll Parish, Louisiana (70.2 [95% UI: 64.4–76.2] deaths per 100,000 population). Nationally, the mortality rate due to CKD was relatively stable between 1980 and 1990 (percent change:−2.4% [95% UI:−6.4 to 1.7%]) but increased by 50.1% (95% UI: 43.3–57.1%) between 1990 and 2014. CKD mortality rates increased in most counties between 1980 and 2014 (97.3%; statistically signifcant in 94.3%). Counties with the largest increases were located primarily in western Oregon, Iowa and Minnesota, southern Illinois, and parts of Texas, Tennessee, Kentucky, and West Virginia. Similar geographic patterns were observed for males and females in 2014 (Additional fle  1: eFigs. S3 and S4), although females experienced a larger relative increase in the CKD mortality rate between 1980 and 2014 than men (56.8% [95% UI: 44.8–69.0%] vs. 31.5% [95% UI: 21.1–41.8%]).

Chronic kidney disease by underlying cause 

When CKD deaths were examined separately by underlying cause (i.e., diabetes, hypertension, glomerulonephritis, or other factors), several geographic patterns emerged. Counties with very high mortality rates from CKD due to diabetes were found along the Mississippi River, near the border of West Virginia and Kentucky, in southern Texas, in New Mexico, and in North and South Dakota (Fig. 3). Nationally, mortality rates from CKD due to diabetes mellitus declined by 5.0% (95% UI: 1.0–9.1%) between 1980 and 2000 but increased by 73.8% (95% UI: 64.2–88.6%) between 2000 and 2014. Almost all counties experienced an increase in the mortality rate from CKD due to diabetes mellitus between 1980 and 2014 (97.8%; statistically signifcant in 96.0%). Counties with the largest increases in the mortality rate were located predominantly in west-coast states, Texas, and the northern Midwest. Conversely, counties where mortality declined were located primarily in eastern states and Alaska. Similar geographic patterns were observed for mortality rates among males and females in 2014 (Additional fle 1: eFigs. S5 and S6), but the mortality rate was higher overall among males compared to females (10.7 [95% UI: 9.8–12.0] vs. 7.5 [95% UI: 6.8–8.5] deaths per 100,000 population in 2014).

Counties with high CKD mortality rates due to hypertension were mainly concentrated in the Southeast with the exception of Florida (Fig. 4). Clusters of counties with particularly high mortality rates were present around the Mississippi river in Mississippi, Louisiana, and Arkansas as well as in Georgia and parts of South Carolina. National mortality rates were relatively stable between 1980 and 1990 (percent change: −2.2% [95% UI: −6.2 to 2.1%]), but increased by 44.0% (95% UI: 34.4–52.5%) between 1990 and 2014. Counties with relatively high rates of increase in the mortality rate between 1980 and 2014 were found primarily in more central states, from Texas in the south to Minnesota in the north, and West Virginia in the east. Similar geographic and time trends were observed among males and females (Additional fle  1: eFigs. S7 and S8), although males experienced higher mortality rates in 2014 compared to females (13.1 [95% UI: 11.8–14.1] vs. 10.2 [95% UI: 9.3–11.1] deaths per 100,000).

Mortality rates from CKD due to glomerulonephritis varied widely among counties (Fig. 5). Clusters of counties with particularly high mortality rates were observed in the south along the Mississippi River, parts of South and North Carolina, in some counties in North and South Dakota. Nationally, mortality rates were relatively stable from 1980 to 2000 (percent change: −1.0% [95% UI: −5.9 to 4.4%]) but increased by 16.9% (95% UI: 7.8–24.8%) from 2000 to 2014, and counties with particularly large increases were found on the West Coast, in the Midwest, and in Maine. Similar patterns were observed for males and females in 2014 (Additional file 1: eFigs. S9 and S10).

CKD mortality due to other causes had very different geographic patterns compared to CKD mortality overall (Fig. 6). Very high rates were observed in counties in Alaska and parts of Montana, North and South Dakota, and Maine. Nationally, mortality rates increased by 32.7% (95% UI: 22.9–42.9%) between 1980 and 2014, with clusters of high increases in Maine, northern Great Plains states, and Alaska. However, some counties, primarily in southern and eastern states and in California, experienced a decline in mortality over this period. Similar patterns were observed for males and females in 2014 (Additional file 1: eFigs. S11 and S12), although mortality rates were higher overall among males compared to females in 2014 (0.21 [95% UI: 0.16–0.28] vs. 0.14 [95% UI: 0.11–0.19] deaths per 100,000 population).

Supportive Service Of Wecistanche-The largest cistanche exporter in the China:

Email:wallence.suen@wecistanche.com 

Whatsapp/Tel:+86 15292862950

Shop For More Specifications Details:

https://www.xjcistanche.com/cistanche-shop