Trends and Socioeconomic, Demographic, and Behavioral Determinants of Kidney Disease Prevalence among African Americans in the United States, 1997–2021

Data: National Health Interview Survey

We derived data on kidney disease and selected socioeconomic, demographic, and behavioral characteristics for the entire US population and Black population aged ≥18 from the 1997 to 2000, 2012 to 2018, and 2020 to 2021 National Health Interview Surveys (NHIS).^[13] The NHIS is a national sample household survey that collects data on sociodemographic, behavioral, morbidity, health, and healthcare characteristics through personal household interviews. The information collected in the survey is based on self-reports. The NHIS employs a complex, multistage probability design and is representative of the US civilian non-institutionalized population.^[1,13,14]

We pooled four years of NHIS data from 1997 to 2000 (n = 130,497), seven years of data from 2012 to 2018 (n = 224,638), and two years of data from 2020 to 2021 (n = 60,894) to ensure sufficient sample sizes for analyzing kidney disease patterns among Black Americans and other racial/ethnic groups by various characteristics over time. Kidney disease data were not available for the 2019, 2022, and 2023 NHIS and were available only for the third and fourth quarters of the 2020 survey. While data from all three periods were used to analyze time trends in racial/ethnic and socioeconomic disparities in kidney disease, more in-depth multivariate analyses were conducted for the 2012–2018 period because of a larger sample size and more detailed information on race, ethnicity, SES, and behavioral and health status characteristics.

Dependent or Outcome Variable: Severe Kidney Disease

The dichotomous outcome variable of severe kidney disease was based on the question: “During the past 12 months, have you been told by a doctor or other health professional that you had weak/failing kidneys? Do not include kidney stones, bladder infections, or incontinence.”^[13] During 2012–2018, 194 or 0.086% of the cases with missing data on kidney disease were excluded from the analysis. This resulted in a pooled sample size of 224,444 for the total population and a sample of 29,590 Black adults aged ≥18.

Independent or Predictor Variables: Socioeconomic, Demographic, and Behavioral Characteristics

Based on prior research, we considered the following covariates for analysis: survey year, age, gender, race/ethnicity, nativity/immigrant status, marital status, region of residence, educational attainment, family income/poverty status, occupation, hypertension, diabetes, physical activity, and body mass index (BMI).^{[5,6,8,9,11,15-18]} Race/ethnicity was grouped into five categories: non-Hispanic Whites, non-Hispanic Blacks/African Americans, AIANs, APIs, and Hispanics. Individuals born in one of the 50 US states or Washington, DC, were considered US-born. Individuals born outside the mainland US were considered immigrants or foreign-born. Educational attainment was measured by five categories: <9, 9–11, 12, 13–15, ≥16 years of completed schooling. Income/poverty level, measured as a ratio of annual family income to the federal poverty threshold, was defined by five categories, ranging from <100% to ≥400% of the poverty level. Occupational class consisted of five broad categories: professional and managerial occupations, sales/clerical and technical support occupations, service, craft and repair, laborers, all other occupations, and those outside the labor force.^[13,14]

Survey respondents were considered as having hypertension if they had ever been told by a doctor or other health professional that they had hypertension or high blood pressure.^[13,14] Diagnosed diabetes (yes, no) was a dichotomous variable. Respondents were considered as having diabetes if they had ever been told by a doctor or other health professional that they had diabetes.^[13,14] BMI was classified into three categories: normal weight (BMI < 25), overweight (25≥BMI < 30), and obesity (BMI ≥ 30). Leisure time physical activity, measured by the number of times per week of vigorous, moderate, or light activities of specific duration, consisted of three categories: inactive, engaged in some activity, and regular activity.^[13,14] Physical inactivity included adults who did not engage in any sessions of light or moderate (causing light sweating or a slight to moderate increase in breathing or heart rate) or vigorous (causing heavy sweating or a large increase in breathing or heart rate) leisure-time physical activity of at least 10 min duration or were unable to perform leisure-time physical activity. Regular physical activity included adults who engaged in at least three sessions per week of vigorous leisure-time physical activity lasting at least 20 min in duration or at least five sessions per week of light or moderate leisure-time physical activity lasting at least 30 min in duration.^[1,13,14] All other covariates were measured as shown in Table 1.

Statistical Methods

We used multivariate logistic regression to examine the association between kidney disease prevalence and selected sociodemographic and behavioral characteristics among Black Americans. We used the Chi-square (χ²) statistic to test for compositional differences between Black adults with and without CKD and differences in kidney disease prevalence by education or income/poverty level. We applied t-tests to test for race-specific differences in socioeconomic gradients over time. To account for the complex sample design of the NHIS, we used SUDAAN software to conduct all statistical analyses, including the logistic modeling procedure RLOGIST.^[19]

Descriptive Characteristics of the Pooled Sample

For the large 2012–2018 pooled sample, approximately 35.0% of the Black sample consisted of individuals aged 18–34, and 14.1% of the sample were adults aged ≥65 years [Table 1]. Age-compositional differences differed substantially among Black adults with and without CKD. Only 5.6% of the adults with CKD were aged 18–34, compared with 35.7% of those without CKD. Approximately 45.3% of Black adults with CKD were aged ≥65 years, compared with 13.4% of those without CKD. Compared to adults without CKD, those with CKD were more likely to be US-born, widowed/divorced/separated, Southern residents, had lower educational attainment and income, higher likelihood of being employed in blue-collar occupations, and higher prevalence of hypertension, diabetes, physical inactivity, and obesity.

Disparities in CKD Prevalence and Odds by Socioeconomic, Demographic, and Behavioral Risk Factors

Disparities presented here refer to the large 2012–2018 pooled sample containing detailed information on race, ethnicity, SES, and behavioral and health status covariates. In 2018, 6.0 million adults aged ≥18, including 0.9 million Black adults, reported being diagnosed with severe kidney disease – an increase of 2.1 million total kidney disease patients and 269,538 Black kidney disease patients since 2012. During 2012–2018, the prevalence of severe kidney disease was 2.4% for the Black population aged ≥18 years, compared with 1.8% for non-Hispanic Whites, 3.5% for AIANs, 1.1% for APIs, and 1.7% for Hispanics. Even after controlling for socioeconomic and demographic factors, Blacks had 19% higher odds (odds ratio [OR] = 1.19; 95% confidence interval [CI] = 1.07–1.32) of kidney disease than non-Hispanic Whites and 47% higher odds than APIs (OR = 1.47; 95% CI = 1.32–1.66) [data not shown].

Education, income, occupation, nativity/immigrant status, physical inactivity, BMI, hypertension, and diabetes were strong predictors of kidney disease in Black Americans [Table 2]. The unadjusted prevalence of CKD for Black adults with <9 years of education was 7.9%, 6.2 times higher than that for Black adults with a college degree (≥16 years of education). Consistent income gradients in kidney disease prevalence were found. Black adults with incomes below the poverty level had 2.8 times higher prevalence than those with incomes at ≥400% of the poverty level (3.5% vs. 1.3%). Blacks employed in service, craft/repair, and manual labor occupations had 51 to 89% higher prevalence of CKD than those in professional and managerial occupations. Black immigrants had a CKD prevalence of 1.0%, 63% lower than the prevalence of 2.7% for US-born Blacks. Black adults with hypertension, diabetes, physical inactivity, and obesity had, respectively, 11.1-, 6.8-, 3.4-, and 1.4-times higher prevalence of CKD than those without these behavioral risk factors [Table 2].

In the fully adjusted Model 3 that controlled for various socioeconomic, demographic, and behavioral risk factors, disparities in prevalence and odds of CKD were substantially reduced, compared with unadjusted Model 1 and SES-adjusted Model 2. However, Black immigrants had 53% lower adjusted odds of kidney disease than US-born Black. Blacks with incomes below the poverty level had 2.4 times higher adjusted odds of having CKD than Blacks with incomes at ≥400% of the poverty level. Physical inactivity, diabetes, and hypertension were associated with 1.9, 2.9, and 4.4 times higher adjusted odds of CKD among Black adults, respectively, controlling for other factors [Table 2].

Trends in SES Inequalities in Kidney Disease and Interaction Effects by Race, SES, and Age

Educational and income inequalities in kidney disease prevalence were substantial and consistent for the total population, Blacks, non-Hispanic Whites, and APIs in 1997– 2000, 2012–2018, and 2020–2021, with educational and income gradients persisting over time [Figure 1 and Table 3]. During these time periods between 1997 and 2021, Blacks, non-Hispanic Whites, APIs, and the total US population with <12 years of education had 3–5 times higher risk of kidney disease than their counterparts with a college degree [Figure 1]. The prevalence of severe CKD increased significantly over time, particularly between 2012–2018 and 2020–2021, among Blacks and for the total US population at each education and income level. For example, the prevalence of severe CKD among Blacks with <12 years of education increased from 4.2% in 1997–2000 to 4.4% in 2012–2018 and 6.4% in 2020–2021. The prevalence of severe CKD among Blacks with incomes below the poverty level increased from 3.4% in 1997–2000 to 3.5% in 2012–2018 and 6.4% in 2020– 2021 [Figure 1].

Close

Figure 1:

(a-f) Prevalence (%) of severe kidney disease by race/ethnicity, education, and income/poverty level, United States, 1997–2000, 2012–2018, and 2020–2021. (a) Educational inequalities, 1997–2000, (b) Educational inequalities, 2012–2018, (c) Educational inequalities, 2020–2021, (d) Income/poverty level inequalities, 1997–2000, (e) Income/poverty level inequalities, 2012–2018, (f) Income/poverty level inequalities, 2020–2021 Source: Data derived from the 1997 to 2000, 2012 to 2018, and 2020 to 2021 NHIS. Notes: Chi-square statistics for testing the overall association between education and severe kidney disease prevalence were statistically significant at p < 0.01 for all groups except for American Indians/Alaska Natives in 2012–2018 (p = 0.20). The association between income/poverty level and kidney severe disease prevalence was statistically significant at p < 0.01 for all groups except for American Indians/Alaska Natives (p = 0.02) and Asian/Pacific Islanders (p = 0.38) in 1997–2000 and for American Indians/Alaska Natives (p = 0.09) in 2012–2018. Kidney disease prevalence estimates for American Indians/Alaska Natives were shown for <12 and ≥12 years of education and for <100% and ≥100% of poverty level in 1997–2000 and for <100%, 100-199%, and ≥200% of poverty level in 2012–2018. Kidney disease prevalence estimates for Asian/Pacific Islanders were shown for <200% and ≥200% of the poverty level in 1997–2000.

Export to PPT

To examine possible differential effects of each covariate on kidney disease by age,^[10,20] we conducted a sensitivity analysis and estimated various interaction models of age (18–64, ≥65) with other covariates, for example, age*diabetes and age*hypertension. None of the interactions, except for hypertension and diabetes, were statistically significant, and the socioeconomic, demographic, and behavioral patterns in kidney disease prevalence and odds were generally similar for adults aged 18–64 and those aged ≥65 (data not shown). For statistically significant interaction terms, age*hypertension and age*diabetes, we separately estimated the odds of having CKD by hypertension or diabetes after restricting the sample by age group (18-64 vs. 65+). Black adults aged 18–64 with hypertension had 11.9 times higher unadjusted odds (OR = 11.87; 95% CI = 8.51–16.55) of CKD and 5.4 times higher adjusted odds (OR = 5.40; 95% CI = 3.47–8.18) of CKD than their counterparts without hypertension. Black adults aged ≥65 with hypertension had 3.2 times higher unadjusted odds (OR = 3.17; 95% CI = 1.86–5.39) of CKD and 2.8 times higher adjusted odds (OR = 2.78; 95% CI = 1.66–4.66) of CKD than their counterparts without hypertension.

Black adults aged 18–64 with diabetes had 8.7 times higher unadjusted odds (OR = 8.66; 95% CI = 6.77–11.07) of CKD and 3.2 times higher adjusted odds (OR = 3.15; 95% CI = 2.40–4.12) of CKD than their counterparts without diabetes. Black adults aged ≥65 with diabetes had 2.8 times higher unadjusted odds (OR = 2.81; 95% CI = 2.19–3.61) of CKD and 2.6 times higher adjusted odds (OR = 2.57; 95% CI = 1.98–3.34) of CKD than their counterparts without diabetes.

Limitations of the study

NHIS is a cross-sectional survey; causality cannot be inferred, especially for SES and health-risk factors, and CKD. However, education and racial/ethnic disparities in CKD are less likely to be affected by reverse causality as formal education is generally attained by age 25 by most individuals and is stable over the life course, and CKD is more likely to develop at more advanced ages. Second, our measure of CKD is based on self-report and likely captures stages 3 and 4 of the disease. The self-reported CKD has been shown to be substantially underestimated when compared with clinical measures of CKD, especially among older patients, males, those at early stages of the disease, and those with less access to and use of health services.^[18] Third, NHIS does not collect data on CKD by stage. As a result, we were not able to estimate overall or stage-specific CKD prevalence for Black, other racial/ethnic, and sociodemographic groups. Such information is important to understand the overall kidney disease burden, risk factors, disease progression, and to design better public health interventions, including treatment programs.^[5]

Key Messages

1) During 2012–2018, the prevalence of severe kidney disease (weak and failing kidneys) in the US was 2.4% for the Black population aged ≥18 years, compared with 1.8% for non-Hispanic Whites, 3.5% for AIANs, 1.1% for APIs, and 1.7% for Hispanics. 2) In 2021, 7.2 million US adults aged ≥18, including 0.98 million Black adults, reported being diagnosed with severe kidney disease – an increase of 3.3 million total patients and 303,059 Black patients with severe kidney disease since 2012. 3) Education, income, occupation, nativity/immigrant status, physical inactivity, BMI, hypertension, and diabetes were strong predictors of kidney disease in US Blacks. 4) Controlling for other risk factors, US Blacks aged 18–64 with diabetes and hypertension experienced 3.2- and 5.4-times increased risk of CKD, respectively. 5) Policies aimed at prevention, management, and treatment of kidney disease in the United States need to monitor the socioeconomically disadvantaged segments of the Black population.