Key Points
Question
Do disparities in glycemic control associated with race and ethnicity persist over time among adults with private and public insurance and do social, health care, and behavioral or health factors attenuate disparity estimates?
Findings
In this cross-sectional study of 4070 US adults, Hispanic or Latino and non-Hispanic Black individuals had higher odds of poor glycemic control compared with non-Hispanic White individuals, despite high levels of access to care. Social, health care, and behavioral or health factors did not attenuate disparities, which also persisted among adults with private insurance.
Meaning
These findings suggest that having health insurance was insufficient to address disparities in glycemic control associated with race and ethnicity; research is needed to identify barriers preventing Hispanic and Latino and non-Hispanic Black communities from meeting diabetes management goals.
This cross-sectional study describes racial and ethnic disparities in glycemic control among adults with private and public insurance in the US over a 15-year timeframe and evaluates whether social, health care, and behavioral or health status factors attenuate estimates of disparities.
Abstract
Importance
Poor access to care and lack of health insurance are important contributors to disparities in glycemic control. However expanding health insurance coverage may not be enough to fully address the high burden of poor glycemic control for some groups.
Objective
To characterize racial and ethnic disparities in glycemic control among adults with private and public insurance in the US over a 15-year timeframe and to evaluate whether social, health care, and behavioral or health status factors attenuate estimates of disparities.
Design, Setting, and Participants
This cross-sectional study used data from the National Health and Nutrition Examination Survey from 2003 to 2018. Participants included Hispanic or Latino, non-Hispanic Black, and non-Hispanic White adults aged 25 to 80 years with self-reported diabetes and health insurance. Data were analyzed from January 15 to August 23, 2023.
Exposure
Participants self-identified as Hispanic or Latino, non-Hispanic Black, or non-Hispanic White.
Main Outcomes and Measures
The main outcome, poor glycemic control, was defined as glycated hemoglobin A1c (HbA1c) of 7.0% or greater. Information about social (education, food security, and nativity), health care (insurance type, routine place for health care, insurance gap in past year, and use of diabetes medications), and behavioral or health status (years with diabetes, waist circumference, and smoking) factors were collected via questionnaires.
Results
A total of 4070 individuals (weighted mean [SE] age, 61.4 [0.27] years; 1970 [weighted proportion, 49.3%] were women) were included, representing 16 337 362 US adults, including 1146 Hispanic or Latino individuals (weighted proportion, 13.2%), 1196 non-Hispanic Black individuals (weighted proportion, 15.7%), and 1728 non-Hispanic White individuals (weighted proportion, 71.1%). In models adjusted for age, sex, and survey year, Hispanic or Latino and non-Hispanic Black individuals had significantly higher odds of poor glycemic control than non-Hispanic White individuals (Hispanic or Latino: odds ratio [OR], 1.46; 95% CI, 1.16-1.83; Black: OR, 1.28; 95% CI, 1.04-1.57). There was some attenuation after adjustment for social factors, especially food security (Hispanic or Latino: OR, 1.39; 95% CI, 1.08-1.81); Black: OR, 1.39; 95% CI, 1.08-1.81). However, accounting for health care and behavioral or health status factors increased disparities, especially for Hispanic or Latino individuals (OR, 1.63; 95% CI, 1.24-2.16), with racial and ethnic disparities persisting even among those with private insurance (OR, 1.66; 95% CI, 1.10-2.52).
Conclusions and Relevance
In this cross-sectional study of insured adults with diabetes in the US, disparities in poor glycemic control persisted despite adjustment for social, health care, and behavioral factors. Research is needed to identify the barriers contributing to poor control even in populations with access to care.
Introduction
Approximately 37.3 million individuals in the US have diabetes, and it is the seventh leading cause of death and the leading cause of comorbidities, including cardiovascular disease, amputation, and kidney failure, in the US.1 A crucial component of diabetes care is glycemic control, managed through medication, healthy diet, and physical activity.
US trends in glycemic control point to a recent pattern of decline. After some improvement from 1988 to 1994 through 2005 to 2010, the prevalence of glycemic control decreased from 57.4% of adults in 2007 to 2010 to 50.5% of adults in 2015 to 2018, based on data from the National Health and Nutrition Examination Surveys (NHANES).2 Across a similar timeframe, Hispanic or Latino and non-Hispanic Black adults bore a disproportionately higher burden of poor control compared with non-Hispanic White adults.2,3,4
Poor access to care and lack of health insurance are important contributors to disparities in glycemic control.5 Data from the 2020 National Health Interview Survey indicated that close to one-fourth of Hispanic or Latino adults younger than 65 years were uninsured, the highest of any racial or ethnic group.6 Health insurance is a critical component of diabetes care because coverage is associated with earlier detection and better management.7,8 The passage of the 2010 Patient Protection and Affordable Care Act expanded access to health insurance, narrowing gaps in coverage by race and ethnicity, but disparities persist.9
Despite the importance of having health insurance for the successful management of diabetes, there is some evidence that disparities in access to care do not fully account for the high burden of poor control among Hispanic or Latino and non-Hispanic Black individuals.4,10 According to Healthy People 2020,11 a health disparity is a health difference between groups that is unfair. For insured adults with diabetes, financial hardships can make it difficult to adhere to medications to achieve care goals. These and other barriers to accessing nutritious food can interfere with an individual’s ability to make healthy changes to diet.12 Other determinants of poor glycemic control, such as advanced disease at diagnosis, can be impacted by health care–related factors, such as gaps in health insurance coverage or disparities in access to screening.
Identifying the factors that are associated with poor diabetes management is essential for more effective care planning, especially for racial and ethnic groups that bear a disproportionate burden of diabetes-associated morbidity and mortality. Consistent with existing frameworks for addressing health disparities,11,13,14 this study does not focus on the causal status of race or ethnicity. Instead, we apply a descriptive approach to evaluate whether a range of potentially modifiable factors attenuate estimates of disparities in glycemic control. We used nationally representative data to evaluate trends in racial and ethnic disparities in glycemic control among US adults with private and public insurance over a 15-year timeframe. We then evaluated whether a range of social, health care, and behavioral or health status factors attenuated estimates of disparities.
Methods
This cross-sectional study was approved by the National Center for Health Statistics (NCHS) institutional review board approved the surveys, and all participants provided informed consent. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies.
Data came from NHANES, a series of ongoing national surveys conducted by the NCHS. NCHS has conducted surveys in 2-year cycles since 1999. NHANES uses a complex, multistage sample design intended to be nationally representative of the US noninstitutionalized population.15 In the first phase, researchers collected information from household interviews on demographics, socioeconomic indicators, medical history, and health behaviors. In the second phase, participants were administered a physical examination in a mobile examination center.
We pooled 15 years of data from 2003 to 2018 to achieve sufficient sample sizes. We restricted the sample to adults aged 25 to 80 years with diabetes who reported having health insurance, self-reported their race and ethnicity as Hispanic or Latino (combines the Mexican American and Other Hispanic racial and ethnicity categories), non-Hispanic Black, or non-Hispanic White and had complete data on hemoglobin A1c (HbA1c). We did not include participants self-reporting other race (including multiracial) due to small sample size. Participants were classified as having diabetes if they reported having had a physician’s diagnosis of diabetes (other than during pregnancy). HbA1c was measured using whole blood at a central laboratory by a high-performance liquid chromatographic assay and standardized according to the Diabetes Control and Complications method.15 Poor glycemic control was defined by HbA1c levels 7.0% or greater (to convert to proportion of total hemoglobin, multiply by 0.01), consistent with other population-based studies.2,3,10 In sensitivity analyses, we repeated analyses using less stringent targets: 8.0% or greater and 9.0% or greater.
Covariates
Information on other demographic, social, health care, and behavioral or health status factors was collected during the home interview. Social factors included nativity (US-born or non–US-born), education (<high school, high school, some college, or ≥college), and food security, assessed via the US Department of Agriculture 18-item household questionnaire (full food security, marginal food security, low food security, and very low food security).16 Health care–related factors included diabetes medication use (insulin only, oral hypoglycemic agents only, insulin and oral hypoglycemic agents, or none), having a routine place for health care (yes or no), having insurance gaps in the past year (yes or no), and health insurance type. Although NHANES captures various insurance types, we categorized this variable as private, Medicare or Medicaid, and other due to small sample sizes for other insurance types. We considered evaluating what an individual’s health care professional suggested their A1c should be, but these data were not available across all years, and among the years in which these data were collected, missingness was disproportionately high, especially for Hispanic or Latino participants (approximately 33%) and non-Hispanic Black individuals (approximately 26%) compared with non-Hispanic White individuals (approximately 15%). Behavioral or health status factors included years with diabetes, waist circumference, and smoking status (never, former, or current). Waist circumference was measured in centimeters at the midpoint between the bottom of the ribs and the top of the iliac crest. For descriptive purposes, we also dichotomized waist circumference based on established criteria for defining abdominal obesity.17 We chose to evaluate waist circumference over body mass index because several studies have shown that waist circumference is a better indicator of metabolic health.18 Higher waist circumference has also been associated with with increased sedentary activity and diets high in sugar and energy-dense foods.19
Statistical Analysis
All analyses were conducted using Stata statistical software version 14.0 (StataCorp). Survey procedures were used to account for the complex design. Appropriate weights were applied to provide unbiased population estimates and account for survey nonresponse, oversampling, poststratification, and sampling error.20 Weighted proportions and means were used to characterize the analytic sample by race and ethnicity. Differences were evaluated using the t statistic, and 2-sided P < .05 was considered statistically significant. Multiple imputation with chained equations was used to deal with missing data for covariates (approximately 12% missing). Based on previous recommendations, 20 imputations were used.21 Multivariable logistic regression models were used to assess the associations of race and ethnicity with poor glycemic control, defined as HbA1c 7.0% or greater, adjusting for age, sex, and survey year (model 1). We next added the social, health care, and behavioral or health status variables to the model in a sequential manner to evaluate the degree to which estimates of racial and ethnic disparities were attenuated as they were conditioned on each new set of variables. Model 2 further adjusted for nativity, education, and food security. Model 3 further adjusted for diabetes medication use, having a routine place for health care, having insurance gaps in the past year, and insurance type. Model 4 included additional adjustments for years with diabetes, waist circumference, and smoking status. Marginal effects were also calculated for the race and ethnic variables at the sample mean of the covariates. In sensitivity analyses, we evaluated whether disparities persisted by health insurance type (private vs Medicare or Medicaid) using stratified models. We did not stratify on other insurance type due to small sample size. We also reran all models using glycemic targets of 8.0% or greater and 9.0% or greater. Data were analyzed from January 15 to August 23, 2023.
Results
A total of 4070 adults (weighted mean [SE] age, 61.4 [0.27] years; 1970 [weighted proportion, 49.3%] female) were included in this study, representing 16 337 362 insured adults with diagnosed diabetes in the US. Table 1 displays characteristics by race and ethnicity. There were 1146 Hispanic or Latino individuals (weighted proportion, 13.2%), 1196 non-Hispanic Black individuals (weighted proportion, 15.7%), and 1728 non-Hispanic White individuals (weighted proportion, 71.1%). Compared with non-Hispanic White individuals, a higher proportion of Hispanic or Latino and Black individuals had less than a high school education and were less food secure (Table 1). A small proportion of non-Hispanic Black and non-Hispanic White individuals were born outside the US, whereas most Hispanic or Latino individuals were born outside the US (Table 1). Although private insurance was the most common insurance type across racial and ethnic groups, a higher proportion of Hispanic or Latino and non-Hispanic Black individuals had Medicare or Medicaid. Most of the population, regardless of race or ethnicity, had a routine place for health care, but estimates were lower for Hispanic or Latino individuals compared with non-Hispanic White individuals (Table 1). There were no statistically significant differences by race or ethnicity in the proportion of adults with past-year gaps in insurance, in the proportion using diabetes medications, nor in the number of years with diabetes. Hispanic or Latino individuals had the lowest proportion of individuals with high waist circumference, and Hispanic or Latino and non-Hispanic Black individuals had higher proportions of adults who never smoked compared with White individuals (Table 1).
Table 1. Characteristics of Insured Adults Aged 25 to 80 Years With Diagnosed Diabetes by Race and Ethnicitya.
| Characteristic | Individuals, unweighted No. (weighted %) | P value | ||
|---|---|---|---|---|
| Hispanic or Latino (n = 1146) | Non-Hispanic Black (n = 1196) | Non-Hispanic White (n = 1728) | ||
| Age (SE), y | 57.8 (0.62) | 59.5 (0.47) | 62.5 (0.36) | <.001 |
| Sex | ||||
| Male | 571 (49.5) | 578 (42.3) | 951 (52.8) | <.001 |
| Female | 575 (50.5) | 618 (57.7) | 777 (47.2) | |
| Highest educational attainment, unweighted No. (weighted %) | ||||
| <High school | 630 (47.8) | 366 (28.3) | 422 (17.0) | <.001 |
| High school graduate or GED | 201 (19.3) | 290 (25.1) | 464 (26.8) | |
| Some college | 232 (25.0) | 363 (31.5) | 517 (32.7) | |
| ≥College graduate | 80 (7.8) | 172 (15.1) | 325 (23.5) | |
| Food security | ||||
| Full | 670 (57.7) | 766 (63.7) | 1298 (81.7) | <.001 |
| Marginal | 144 (13.9) | 165 (15.0) | 143 (7.6) | |
| Low | 201 (17.7) | 152 (13.3) | 156 (6.6) | |
| Very low | 104 (10.7) | 95 (8.0) | 102 (4.1) | |
| Nativity | ||||
| US-born | 494 (44.2) | 1099 (92.2) | 1664 (96.7) | <.001 |
| Non–US-born | 652 (55.8) | 97 (7.8) | 64 (3.3) | |
| Insurance type | ||||
| Private insurance | 475 (49.2) | 598 (52.6) | 992 (66.8) | <.001 |
| Medicare or Medicaid | 556 (40.8) | 496 (39.0) | 621 (26.2) | |
| Other | 111 (1.0) | 97 (8.4) | 110 (7.0) | |
| Have routine place for health care | 1102 (95.5) | 1179 (98.2) | 1696 (97.8) | .04 |
| Time with no insurance in past year | ||||
| Yes | 250 (21.0) | 168 (15.8) | 200 (15.3) | .20 |
| No | 844 (74.1) | 988 (80.1) | 1479 (82.0) | |
| Missing | 52 (4.9) | 40 (4.0) | 49 (2.8) | |
| Using diabetes medication | ||||
| Insulin only | 128 (11.6) | 192 (15.9) | 254 (14.6) | .29 |
| Oral hypoglycemic agents only | 667 (56.1) | 671 (55.0) | 964 (56.6) | |
| Insulin and oral hypoglycemic agents | 188 (15.5) | 185 (15.7) | 264 (14.6) | |
| No medications | 163 (16.8) | 148 (13.4) | 246 (14.1) | |
| Diabetes duration, mean (SE), y | 11.9 (0.42) | 12.0 (0.33) | 12.3 (0.36 | .77 |
| Waist circumference | ||||
| Mean (SE), cm | 108.0 (0.61) | 111.6 (0.56) | 112.7 (0.50) | <.001 |
| High-riskb | 825 (71.7) | 886 (75.3) | 1342 (79.7) | <.001 |
| Smoking status | ||||
| Current smoker | 135 (13.0) | 220 (18.5) | 251 (13.7) | <.001 |
| Former smoker | 393 (31.8) | 398 (30.2) | 727 (39.7) | |
| Never smoker | 618 (55.2) | 576 (51.3) | 750 (46.6) | |
Abbreviation: GED, general equivalency diploma.
Diabetes diagnosis was based on self-report of a previous diagnosis by a physician or other health professional.
Defined as waist circumference greater than 102 cm for men and greater than 88 cm for women.
Racial and ethnic differences in glycemic control persisted across the 15-year period among insured adults (Figure), including in more recent years (2015-2018), after the implementation of the 2010 Patient Protection and Affordable Care Act. Across all years, the prevalence of poor glycemic control was highest among Hispanic or Latino individuals, followed by non-Hispanic Black and non-Hispanic White individuals.
Figure. Prevalence of Poor Glycemic Control Among Insured Adults With Diabetes by Race and Ethnicity.
Poor glycemic control was defined as glycated hemoglobin A1c 7.0% or greater (to convert to proportion of total hemoglobin, multiply by 0.01).
Table 2 presents results from multivariable regressions combining years 2003 to 2018. In analyses adjusted for age, sex, and survey year (model 1), Hispanic or Latino individuals (odds ratio [OR], 1.46; 95% CI, 1.16-1.83) and non-Hispanic Black individuals (OR, 1.28; 95% CI, 1.04-1.57) had significantly higher odds of poor glycemic control compared with non-Hispanic White individuals. There was some attenuation of estimates after adjusting for the disproportionate burden of low education, immigrant status, and low food security among Hispanic or Latino (OR, 1.39; 95% CI, 1.08-1.81) and non-Hispanic Black (OR, 1.39; 95% CI, 1.08-1.81) individuals (model 2). Food security contributed most to the attenuation (noted when adjusting for each variable individually in sensitivity analyses). Nevertheless, estimates remained statistically significant, especially for Hispanic or Latino individuals. Further adjustment for health care–related factors (model 3) resulted in an even larger magnitude of association. In sensitivity analyses, we evaluated the impact of each health care–related variable on associations to understand which variables were more or less influential in inflating estimates of disparities. Adjusting for the use of diabetes medications contributed most to the widening of disparities, especially for Hispanic or Latino individuals. When we stratified models on medication use, disparities were evident across all strata, but were especially large among adults not using diabetes medication and among individuals using both insulin and oral hypoglycemic agents. Accounting for the lower waist circumference and more favorable smoking status for Hispanic or Latino and non-Hispanic Black individuals, in addition to years with diabetes (model 4), increased estimates of racial and ethnic disparities in probability of poor glycemic control of 62.0% (95% CI, 52.3%-71.6%) for Hispanic or Latino individuals, 56.6% (95% CI, 47.1%-66.0%) for non-Hispanic Black individuals, and 49.9% (95% CI, 40.3%-59.6%) for non-Hispanic White individuals.
Table 2. Multivariable Logistic Regression of Poor Glycemic Control by Race and Ethnicity.
| Characteristic | OR (95% CI)a | |||
|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | |
| Race and ethnicity | ||||
| Hispanic or Latino | 1.45 (1.16-1.82) | 1.39 (1.08-1.81) | 1.58 (1.20-2.09) | 1.63 (1.24-2.16)b |
| Non-Hispanic Black | 1.27 (1.03-1.56) | 1.24 (1.01-1.52) | 1.29 (1.05-1.60) | 1.30 (1.05-1.62)b |
| Non-Hispanic White | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference]b |
| Age, per 1-y increase | 0.99 (0.98-0.99) | 0.99 (0.98-0.99) | 0.99 (0.98-0.99) | 0.99 (0.98-0.99) |
| Sex | ||||
| Male | 1.33 (1.14-1.56) | 1.36 (1.16-1.60) | 1.34 (1.12-1.61) | 1.33 (1.10-1.61) |
| Female | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Nativity | ||||
| Non–US-born | NA | 1.01 (0.78-1.30) | 1.112 (0.84-1.46) | 1.20 (0.91-1.59) |
| US-born | NA | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Education | ||||
| <High school | NA | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| High school graduate or GED | NA | 1.16 (0.92-1.46) | 1.13 (0.88-1.45) | 1.17 (0.92-1.50) |
| Some college | NA | 0.94 (0.72-1.22) | 0.99 (0.73-1.35) | 1.01 (0.75-1.38) |
| ≥College graduate | NA | 0.94 (0.72-1.22) | 0.94 (0.69-1.27) | 1.00 (0.74,1.34) |
| Food security | ||||
| Very low | NA | 1.37 (0.99-1.90) | 1.15 (0.77-1.74) | 1.11 (0.74-1.67) |
| Low | NA | 1.12 (0.83-1.51) | 1.11 (0.81-1.52) | 1.07 (0.78-1.47) |
| Marginal | NA | 1.05 (0.79-1.40) | 0.93 (0.71-1.23) | 0.91 (0.68-1.20) |
| Full | NA | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| Insurance type | ||||
| Medicare or Medicaid | NA | NA | 0.83 (0.67-1.04) | 0.82 (0.66-1.02) |
| Other | NA | NA | 0.72 (0.50-1.04) | 0.71 (0.50-1.01) |
| Private | NA | NA | 1 [Reference] | 1 [Reference] |
| Routine place for health care | ||||
| Yes | NA | NA | 0.55 (0.26-1.16) | 0.56 (0.28-1.13) |
| No | NA | NA | 1 [Reference] | 1 [Reference] |
| Gap in insurance in past 1 y | ||||
| Yes | NA | NA | 1.64 (1.03-2.60) | 1.63 (1.03-2.59) |
| No | NA | NA | 1 [Reference] | 1 [Reference] |
| Missing | NA | NA | 0.97 (0.59-1.58) | 0.99 (0.60-1.64) |
| Diabetes medication | ||||
| Insulin only | NA | NA | 12.89 (8.78-18.94) | 12.00 (8.11-17.76) |
| Oral hypoglycemic agents only | NA | NA | 3.67 (2.70-5.00) | 3.58 (2.60-4.93) |
| Insulin and oral hypoglycemic agents | NA | NA | 16.42 (11.36-23.74) | 14.95 (10.18–21.96) |
| None | NA | NA | 1 [Reference] | 1 [Reference] |
| Waist circumference, per 1-cm increase | NA | NA | NA | 1.01 (1.00-1.02) |
| Diabetes duration, per 1-y increase | NA | NA | NA | 1.01 (1.00-1.02) |
| Smoking status | ||||
| Current smoker | NA | NA | NA | 1.28 (0.96-1.71) |
| Former smoker | NA | NA | NA | 0.91 (0.74-1.11) |
| Never smoker | NA | NA | NA | 1 [Reference] |
Abbreviations: GED, general equivalency diploma; NA, not applicable; OR, odds ratio.
All models include adjustment for survey year.
Corresponds to probabilities of poor glycemic control of 62.0% (95% CI, 52.3%-71.6%) for Hispanic or Latino individuals, 56.6% (95% CI, 47.1%-66.0%) for non-Hispanic Black individuals, and 49.9% (95% CI, 40.3%-59.6%) for non-Hispanic White individuals.
When we stratified models by health insurance type, disparities persisted for Hispanic or Latino individuals with private insurance (OR, 1.66; 95% CI, 1.10-2.52) and non-Hispanic Black individuals with Medicare or Medicaid (OR, 1.39; 95% CI, 1.02-1.88) compared with White individuals (Table 3). All the patterns we reported persisted when we evaluated other thresholds of glycemic control (Table 4).
Table 3. Multivariable Logistic Regression of Poor Glycemic Control by Race and Ethnicity, Stratified by Health Insurance Type.
| Race and ethnicity | OR (95% CI)a | |
|---|---|---|
| Private insurance | Medicare or Medicaid | |
| Hispanic or Latino | 1.66 (1.10-2.52) | 1.42 (0.97-2.09) |
| Non-Hispanic Black | 1.26 (0.96-1.66) | 1.39 (1.02-1.88) |
| Non-Hispanic White | 1 [Reference] | 1 [Reference] |
Abbreviation: OR, odds ratio.
Models adjusted for age, sex, social, health care, and behavioral or health status factors and survey year. Stratified models of other insurance are not shown due to small sample size.
Table 4. Multivariable Logistic Regression of Poor Glycemic Control Across Different HbA1c Cut Points, by Race and Ethnicity.
| Race and ethnicity | OR (95% CI) | |||
|---|---|---|---|---|
| Model 1a | Model 2b | Model 3c | Model 4d | |
| HbA1c >8.0% | ||||
| Hispanic or Latino | 1.80 (1.43-2.27) | 1.70 (1.25-2.31) | 1.93 (1.40-2.66) | 1.95 (1.42-2.70) |
| Non-Hispanic Black | 1.57 (1.26-1.96) | 1.50 (1.21-1.87) | 1.60 (1.27-2.01) | 1.58 (1.25-2.00) |
| Non-Hispanic White | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
| HbA1c >9.0% | ||||
| Race/ethnicity | ||||
| Hispanic or Latino | 2.07 (1.60-2.69) | 2.15 (1.55-2.97) | 2.35 (1.68-3.27) | 2.35 (1.68-3.27) |
| Non-Hispanic Black | 1.92 (1.48-2.50) | 1.89 (1.46-2.44) | 1.98 (1.49-2.62) | 1.93 (1.46-2.55) |
| Non-Hispanic White | 1 [Reference] | 1 [Reference] | 1 [Reference] | 1 [Reference] |
Abbreviations: HbA1c, glycated hemoglobin A1c; OR, odds ratio.
SI conversion factor: To convert HbA1c to proportion of total hemoglobin, multiply by 0.01.
Model 1 adjusts for age, sex, and survey year.
Model 2 adjusts for factors in model 1 plus social factors.
Model 3 adjusts for factors in model 2 plus health care factors.
Model 4 adjusts for factors in model 3 plus behavioral or health status factors.
Discussion
In this cross-sectional study of nationally representative US data from 2003 to 2018, racial and ethnic disparities in poor glycemic control persisted in the US in a population with health insurance, and despite a high prevalence of having a routine health care practitioner. Hispanic and Latino individuals had the highest odds of poor control. There was some attenuation of estimates after adjustment for social factors previously associated with poor chronic disease management,3 particularly food security, but disparities persisted. Adjustment for health care–related factors, especially use of diabetes medications, and for behavioral or health status factors resulted in a widening of estimates of disparities. Similar patterns were found even among adults with private insurance, as well as those with Medicare or Medicaid. These findings suggest that expanding health insurance coverage alone is unlikely to fully address racial and ethnic disparities in glycemic control.
Previous studies have reported somewhat similar patterns. Using earlier NHANES survey years (1988-1994 and 1999-2002), researchers have reported persistence in racial and ethnic disparities in glycemic control after adjusting for access to health care.10,22 A study by Trivedi et al23 used data from Medicare managed-care plans from 1997 to 2003 and found that although there were improvements in several quality of care measures for both Black and White beneficiaries, the magnitude of the racial disparity in glycemic control increased over time.23 Similar to our findings, adjustment for socioeconomic variables did not account for this disparity,10,22,23 although most studies did not account for the range of factors considered in our study.
Building on past research in this area, we used nationally representative data to investigate whether various social, health care, and behavioral or health status factors could attenuate estimates of racial and ethnic disparities in glycemic control in a population of insured adults with diabetes. Somewhat unexpectedly, these factors did not attenuate disparities and, in some instances, resulted in a widening of the racial and ethnic gap. In terms of the social factors we considered, although Hispanic or Latino individuals and non-Hispanic Black individuals were disproportionately burdened by lower levels of educational and less food security than non-Hispanic White individuals, accounting for this uneven distribution resulted in only some attenuation of estimates. Socioeconomic status is a multidimensional construct that is challenging to measure, with a high potential for residual confounding.24 Measures of socioeconomic status, such as education, may be insufficient to capture the socioeconomic barriers to glycemic control. Previous research has also noted the logistical and financial burden of frequent physicians’ appointments and a chronic need for glucose-lowering medications to adequately manage diabetes, all of which pose challenges to low-income populations.23,25 This suggests that these more traditional social measures may be insufficient for capturing all the challenges patients face in managing diabetes.
Accounting for health care–related factors resulted in a widening of racial and ethnic disparities, especially for Hispanic or Latino individuals, and this was largely driven by adjustment for use of diabetes medications. The disparity was especially large among individuals not using medication, and also among those using both insulin and oral hypoglycemic agents. Unmeasured socioeconomic factors could play a role. Cost-related nonadherence to medications is exemplified by delaying therapy or taking smaller or less frequent doses and has been shown to increase the risk of avoidable complications for individuals with chronic diseases, including diabetes,25,26 and to disproportionately impact Hispanic or Latino and non-Hispanic Black individuals.27,28 Unfortunately, we did not have data on cost-related nonadherence or other factors (eg, pharmacy accessibility, length of time using medications, other reasons for nonadherence) that may provide more insight. Future research should examine the reasons for these patterns.
We also observed that accounting for years with diabetes, waist circumference, and smoking status further magnified racial and ethnic disparities in glycemic control. Unlike the adverse patterning seen for the social indicators, Hispanic or Latino and non-Hispanic Black individuals had more favorable profiles than non-Hispanic White individuals regarding smoking status. Moreover, Hispanic or Latino individuals had the lowest proportion of adults with high waist circumference. However, once we adjusted for these theoretically protective factors among Hispanic or Latino and Black individuals, disparities in glycemic control widened. This suggests that finding ways to support the maintenance of these healthier profiles may be one component of a broader strategy to address disparities.
We also demonstrated that disparities persisted regardless of insurance type and were especially large among Hispanic or Latino individuals compared with non-Hispanic White individuals with private health insurance. Although this may seem counterintuitive, private insurance is often accompanied by high deductibles, copays, and coinsurance, which can pose great financial burden. Indeed, high-deductible private insurance plans have been previously associated with forgoing care among lower-income adults with diabetes, which was in turn associated with poor management outcomes.29 A recent analysis of adults with employer-sponsored health insurance also revealed concerning racial and ethnic disparities across a range of chronic disease outcomes, including diabetes.30 These findings suggest that barriers related to quality and processes of care faced by Hispanic or Latino and non-Hispanic Black individuals may be contributing to these patterns.
Limitations
Our study has some limitations. NHANES data are cross-sectional; therefore, it is not possible to assess whether the factors we assessed had any causal association with glycemic control. It is also possible that the cumulative impacts of some factors (eg, poverty, food insecurity, years with health insurance) are more impactful than the 1-time measurement available in NHANES. We also could not distinguish diabetes by type. Treatment guidelines differ by type, and it would have been useful to make this distinction. We also did not have information surrounding treatment adherence, which is known to impact glycemic outcomes. Having this information would allow for improved targeting of diabetes management strategies. We also could not adequately assess the role of health care practitioner glycemic recommendations, since this information was not consistently collected across all the years included in this analysis, and because of the large degree of missingness. However, we noted that missingness disproportionately impacted Hispanic or Latino and non-Hispanic Black individuals. If we assume that missingness correlates with a lack of knowledge of an individual’s HbA1c goals, this highlights a need to understand whether health care practitioners are adequately communicating this information to their patients. Furthermore, although we accounted for the length of time with diabetes, existing evidence has documented racial and ethnic disparities in undiagnosed diabetes.31,32,33 Although there were no racial and ethnic differences in years with diabetes, Hispanic or Latino and non-Hispanic Black adults could have been living with diabetes for more time than non-Hispanic White adults, potentially leading to more severe disease, rendering it more difficult to control.
Conclusions
This cross-sectional study found persistent racial and ethnic disparities in poor glycemic control, even among adults with health insurance. These findings suggest that although improving access to health care remains a critical policy target, health insurance coverage alone was not enough to reduce disparities in diabetes management. Hispanic and Latino and non-Hispanic Black individuals experience a disproportionate burden of diabetes-related morbidity and mortality and associated costs.34 Uncontrolled diabetes has also been linked to more severe consequences from COVID-19.35 The social, health care, and behavioral or health status factors we examined did not attenuate estimates of disparities. Future studies should apply causal frameworks to evaluate the role of other structural barriers contributing to the high burden of poor control among insured Hispanic or Latino and non-Hispanic Black individuals to develop effective interventions.
Data Sharing Statement
References
- 1.Centers for Disease Control and Prevention . National diabetes statistics report. Accessed February 1, 2023. https://www.cdc.gov/diabetes/data/statistics-report/index.html
- 2.Fang M, Wang D, Coresh J, Selvin E. Trends in diabetes treatment and control in U.S. adults, 1999-2018. N Engl J Med. 2021;384(23):2219-2228. doi: 10.1056/NEJMsa2032271 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wang L, Li X, Wang Z, et al. Trends in prevalence of diabetes and control of risk factors in diabetes among US adults, 1999-2018. JAMA. 2021;326(8):1-13. doi: 10.1001/jama.2021.9883 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Selvin E, Parrinello CM, Sacks DB, Coresh J. Trends in prevalence and control of diabetes in the United States, 1988-1994 and 1999-2010. Ann Intern Med. 2014;160(8):517-525. doi: 10.7326/M13-2411 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hill J, Nielsen M, Fox MH. Understanding the social factors that contribute to diabetes: a means to informing health care and social policies for the chronically ill. Perm J. 2013;17(2):67-72. doi: 10.7812/TPP/12-099 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cha AE, Cohen RA. Demographic variation in health insurance coverage: United States, 2020. Natl Health Stat Report. 2022;(169):1-15. [PubMed] [Google Scholar]
- 7.Zhang X, Geiss LS, Cheng YJ, Beckles GL, Gregg EW, Kahn HS. The missed patient with diabetes: how access to health care affects the detection of diabetes. Diabetes Care. 2008;31(9):1748-1753. doi: 10.2337/dc08-0527 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Myerson R, Laiteerapong N. The Affordable Care Act and diabetes diagnosis and care: exploring the potential impacts. Curr Diab Rep. 2016;16(4):27. doi: 10.1007/s11892-016-0712-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chaudry A, Jackson A, Glied SA. Did the Affordable Care Act reduce racial and ethnic disparities in health insurance coverage? The Commonwealth Fund. August 21, 2019. Accessed February 1, 2023. https://www.commonwealthfund.org/publications/issue-briefs/2019/aug/did-ACA-reduce-racial-ethnic-disparities-coverage
- 10.Saydah S, Cowie C, Eberhardt MS, De Rekeneire N, Narayan KM. Race and ethnic differences in glycemic control among adults with diagnosed diabetes in the United States. Ethn Dis. 2007;17(3):529-535. [PubMed] [Google Scholar]
- 11.Advisory committee findings and recommendations. In: The Secretary’s Advisory Committee on National Health Promotion and Disease Prevention Objectives for 2020: Phase I Report. US Department of Health and Human Services; 2008. [Google Scholar]
- 12.Patel MR. Social determinants of poor management of type 2 diabetes among the insured. Curr Diab Rep. 2020;20(11):67. doi: 10.1007/s11892-020-01354-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Institute of Medicine . Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. National Academy Press; 2002. [PMC free article] [PubMed] [Google Scholar]
- 14.Ogunwole SM, Golden SH. Social determinants of health and structural inequities-root causes of diabetes disparities. Diabetes Care. 2021;44(1):11-13. doi: 10.2337/dci20-0060 [DOI] [PubMed] [Google Scholar]
- 15.Centers for Disease Control and Prevention . National Health and Nutrition Examination Surveys. Accessed May 1, 2022. https://www.cdc.gov/nchs/nhanes/index.htm
- 16.Economic Research Service . U.S. household food security survey module: three-stage design, with screeners. Accessed May 12, 2022. https://www.ers.usda.gov/media/8271/hh2012.pdf
- 17.Alberti KG, Eckel RH, Grundy SM, et al. ; International Diabetes Federation Task Force on Epidemiology and Prevention; Hational Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; International Association for the Study of Obesity . Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation. 2009;120(16):1640-1645. doi: 10.1161/CIRCULATIONAHA.109.192644 [DOI] [PubMed] [Google Scholar]
- 18.Ross R, Neeland IJ, Yamashita S, et al. Waist circumference as a vital sign in clinical practice: a consensus statement from the IAS and ICCR Working Group on Visceral Obesity. Nat Rev Endocrinol. 2020;16(3):177-189. doi: 10.1038/s41574-019-0310-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tchernof A, Després JP. Pathophysiology of human visceral obesity: an update. Physiol Rev. 2013;93(1):359-404. doi: 10.1152/physrev.00033.2011 [DOI] [PubMed] [Google Scholar]
- 20.Centers for Disease Control and Prevention . National Health and Nutrition Examination Survey analytic guidelines. Accessed May 1, 2022. https://wwwn.cdc.gov/nchs/nhanes/analyticguidelines.aspx
- 21.Austin PC, White IR, Lee DS, van Buuren S. Missing data in clinical research: a tutorial on multiple imputation. Can J Cardiol. 2021;37(9):1322-1331. doi: 10.1016/j.cjca.2020.11.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Harris MI, Eastman RC, Cowie CC, Flegal KM, Eberhardt MS. Racial and ethnic differences in glycemic control of adults with type 2 diabetes. Diabetes Care. 1999;22(3):403-408. doi: 10.2337/diacare.22.3.403 [DOI] [PubMed] [Google Scholar]
- 23.Trivedi AN, Zaslavsky AM, Schneider EC, Ayanian JZ. Trends in the quality of care and racial disparities in Medicare managed care. N Engl J Med. 2005;353(7):692-700. doi: 10.1056/NEJMsa051207 [DOI] [PubMed] [Google Scholar]
- 24.Kaufman JS, Cooper RS, McGee DL. Socioeconomic status and health in Blacks and Whites: the problem of residual confounding and the resiliency of race. Epidemiology. 1997;8(6):621-628. doi: 10.1097/00001648-199710000-00002 [DOI] [PubMed] [Google Scholar]
- 25.Patel MR, Piette JD, Resnicow K, Kowalski-Dobson T, Heisler M. Social determinants of health, cost-related nonadherence, and cost-reducing behaviors among adults with diabetes: findings from the National Health Interview Survey. Med Care. 2016;54(8):796-803. doi: 10.1097/MLR.0000000000000565 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhang JX, Bhaumik D, Huang ES, Meltzer DO. Change in insurance status and cost-related medication non-adherence among older U.S. adults with diabetes from 2010 to 2014. J Health Med Econ. 2018;4(2):7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Frankenfield DL, Wei II, Anderson KK, Howell BL, Waldo D, Sekscenski E. Prescription medication cost-related non-adherence among Medicare CAHPS respondents: disparity by Hispanic ethnicity. J Health Care Poor Underserved. 2010;21(2):518-543. doi: 10.1353/hpu.0.0314 [DOI] [PubMed] [Google Scholar]
- 28.Tseng CW, Tierney EF, Gerzoff RB, et al. Race/ethnicity and economic differences in cost-related medication underuse among insured adults with diabetes: the Translating Research Into Action for Diabetes Study. Diabetes Care. 2008;31(2):261-266. doi: 10.2337/dc07-1341 [DOI] [PubMed] [Google Scholar]
- 29.Wharam JF, Zhang F, Eggleston EM, Lu CY, Soumerai SB, Ross-Degnan D. Effect of high-deductible insurance on high-acuity outcomes in diabetes: a Natural Experiment for Translation in Diabetes (NEXT-D) Study. Diabetes Care. 2018;41(5):940-948. doi: 10.2337/dc17-1183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Morgan Health . Health disparities in employer-sponsored insurance. Accessed March 15, 2023. https://www.jpmorganchase.com/content/dam/jpmc/jpmorgan-chase-and-co/who-we-are/our-business/documents/jpmc-morgan-health-norc-report-ada.pdf
- 31.Selvin E, Wang D, Lee AK, Bergenstal RM, Coresh J. Identifying trends in undiagnosed diabetes in U.S. adults by using a confirmatory definition: a cross-sectional study. Ann Intern Med. 2017;167(11):769-776. doi: 10.7326/M17-1272 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Menke A, Casagrande S, Geiss L, Cowie CC. Prevalence of and trends in diabetes among adults in the United States, 1988-2012. JAMA. 2015;314(10):1021-1029. doi: 10.1001/jama.2015.10029 [DOI] [PubMed] [Google Scholar]
- 33.Gupta K, Obeidat L, Kakar TS, et al. Trends in undiagnosed diabetes mellitus among United States adults: cross-sectional analyses from NHANES 2011-2020. Am J Cardiol. 2022;175:184-185. doi: 10.1016/j.amjcard.2022.04.032 [DOI] [PubMed] [Google Scholar]
- 34.Golden SH, Brown A, Cauley JA, et al. Health disparities in endocrine disorders: biological, clinical, and nonclinical factors—an Endocrine Society scientific statement. J Clin Endocrinol Metab. 2012;97(9):E1579-E1639. doi: 10.1210/jc.2012-2043 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wong R, Hall M, Vaddavalli R, et al. ; N3C Consortium . Glycemic control and clinical outcomes in U.S. patients with COVID-19: data from the National COVID Cohort Collaborative (N3C) Database. Diabetes Care. 2022;45(5):1099-1106. doi: 10.2337/dc21-2186 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Sharing Statement