There are significant differences in rates of diagnosed and undiagnosed diabetes between subgroups of Hispanic and Asian Americans, a federal government study finds.
Hispanics and Asians represent 23% of the U.S. population and are expected to account for 38% by 2060.
And, these groups may be at higher risk for type 2 diabetes due to genetic, lifestyle and environmental factors, the researchers said.
The U.S. Centers for Disease Control and Prevention (CDC) study analyzed federal health survey data for 2011 through 2016. Among Hispanics, 25% of Mexicans, 22% of Puerto Ricans, 21% of Cuban/Dominicans, 19% of Central Americans, and 12% of South Americans were living with diagnosed or undiagnosed diabetes, the study found.
Among Asians, the percentages were 23% for South Asians, 22% for Southeast Asians, and 14% for East Asians, according to findings published Dec. 20 in JAMA Internal Medicine.
As in previous research, this study found that Asians tend to have a lower body mass index (BMI) than other racial/ethnic groups. BMI is an estimate of body fat based on weight and height. The higher the BMI, the greater the risk of type 2 diabetes.
“This landmark diabetes survey provides essential data that will better inform public health efforts to reach more Americans with tailored, effective prevention and treatment strategies,” CDC director Dr. Robert Redfield said in an agency news release.
“This defined data on the prevalence of diabetes among Hispanic and Asian demographic groups can help health care providers and patients reduce the risk for type 2 diabetes,” he added.
Ann Albright, director of CDC’s division of diabetes translation, said the findings establish a baseline for future estimates and highlight differences in the diabetes burden among Hispanic and non-Hispanic Asian subgroups.
“These data also provide insights that allow us to reach groups at higher risk and provide opportunities to strengthen diabetes detection and type 2 diabetes prevention and care in these groups,” she said in the news release.
In the United States, the rise in prevalence of diabetes, especially type 2 diabetes (T2D), has been particularly marked within racial and ethnic minority populations and among older individuals.1,2
This increase in the number of individuals developing diabetes has added a significant financial burden to health care systems with current annual global costs estimated to be US $673 billion and these are projected to rise to US $802 billion by 2040.3
In the United States, diabetes care already accounts for one of every four dollars spent on health care.4
Both old age and diabetes are important risk factors for functional decline and disability in older people with the potential to add significant costs to health care systems.5
As a corollary, being uninsured or a Medicaid recipient presents formidable challenges to improving cost-effective outcomes for people with diabetes.6,7
Currently, in the United States, more than 29 million people are uninsured, with substantial inequalities in access to health care along economic, gender, racial, and ethnic lines.8 Previous studies have documented that racial and ethnic minority groups also receive low quality of health care, including preventative health services, compared with their white counterparts9–11 and that racial and ethnic minority groups have higher rates of diabetes-related complications.12–17
More recently, there is evidence that for Medicare beneficiaries in managed care plans (including Medicare Managed Care and Medicare Fee-for-Service plans) as well as for Veterans Administration populations, there has been progress in reducing disparities in diabetes care.18,19
However, it remains to be determined whether similar improvements have been seen in other US populations based on race and ethnicity and what impact they have had on the economic costs of diabetes.
While old age and diabetes place a disproportionate burden on racial and ethnic minorities compared with non-Hispanic whites (nHWs), Hispanics (who comprise 18% of the US population) are almost twice as likely to develop T2D as nHWs and experience high rates of poorly controlled diabetes and related complications.20–23
Therefore, the aim of this study was to identify differences in diabetes prevalence, health care resource utilization, and cost for Medicare beneficiaries among Hispanics compared with nHWs with both T2D and type 1 diabetes (T1D).
Results
A total of 1,397,933 individuals who met the eligibility criteria were included in the study. Racial and ethnic distribution of the sample showed that 1.7% were Hispanic, 1.9% were Asian, 7.8% were black, 86.1% were nHWs, and 2.5% were other.
Prevalence of T1D and T2D
The prevalence of both T1D and T2D was significantly higher in Hispanics compared with nHWs (p<0.0006 and p<0.0001, respectively) (Table 1).
Table 1.
Prevalence of Type 1 and Type 2 Diabetes in Individuals Stratified by Age, Racial and Ethnic Group, and Gender from the 2012 to 2013 Medicare 5% Sample
| Hispanic (H) (%) | Non-Hispanic Whites (nHWs) (%) | p-Value | |
|---|---|---|---|
| Age (years) | |||
| Type 1 diabetes | |||
| 65–69 | 1.8 | 1.1 | 0.0000 |
| 70–74 | 2.9 | 1.4 | 0.0000 |
| 75–79 | 4.2 | 2.7 | 0.0000 |
| 80–84 | 5.1 | 2.4 | 0.0000 |
| 85–90 | 4.5 | 2.2 | 0.0000 |
| >90 | 4.7 | 1.6 | 0.0000 |
| Total | 3.4 | 1.8 | 0.0000 |
| Type 2 diabetes | |||
| 65–69 | 25.8 | 17.0 | 0.0000 |
| 70–74 | 31.0 | 20.6 | 0.0000 |
| 75–79 | 38.8 | 27.5 | 0.0000 |
| 80–84 | 40.3 | 26.2 | 0.0000 |
| 85–90 | 39.2 | 24.2 | 0.0000 |
| >90 | 34.9 | 19.6 | 0.0000 |
| Total | 33.4 | 21.9 | 0.0000 |
| Sex | |||
| Type 1 diabetes | |||
| Male | 39.2 | 46.6 | 0.0089 |
| Female | 60.8 | 53.4 | 0.0012 |
| Type 2 diabetes | |||
| Male | 39.6 | 47.1 | 0.0000 |
| Female | 60.4 | 52.9 | 0.0000 |
There were significant differences by race and ethnicity for gender distribution, with higher rates of T1D and T2D in women than men in both Hispanic and nHW populations, a difference that is more skewed in Hispanics than nHWs.
Specifically, 60.8% of Hispanics with T1D are female, while 53.4% of nHWs with T1D are female (p=0.0012). Likewise, 60.4% of Hispanics with T2D are female, while 52.9% of nHWs with T2D are female (p=0.0000).
The prevalence of both T1D and T2D in various age groups also differed significantly between Hispanics and nHWs (p<0.0001) (Table 1).
The prevalence of both T1D and T2D increased with age among nHWs, with the highest rate noted in individuals aged 75–79 years, after which the prevalence gradually declined for both T1D and T2D. However, in Hispanics, the highest prevalence of T1D and T2D was observed in the age group of 80–84 years.
Costs PMPM
The overall allowed and paid costs PMPM for individuals with and without diabetes and Hispanics versus nHWs are shown in Table 2 (unadjusted and adjusted).
Overall, both the allowed and paid adjusted and unadjusted costs PMPM were two- to fourfold higher for individuals with diabetes compared with those without (p<0.0001).
Table 2.
Unadjusted and Adjusted Average (Mean) Costs (Allowed and Paid) Per Member Per Month by Type of Diabetes and Racial and Ethnic Group (Hispanics Compared with Non-Hispanic Whites) from the 2012 to 2013 Medicare 5% Sample
| Type of diabetes | Average amount per member per month | Difference: Hispanics vs. nHWs | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Race | 2012 | 2013 | 2012 | 2013 | 2012–2013a | 2012–2013a | |||
| Allowed ($) | Paid ($) | Allowed ($) | Paid ($) | Allowed | Paid | ||||
| pa | |||||||||
| Unadjusted | |||||||||
| No diabetes | Hispanic | 461 | 424 | 399 | 367 | 442 | 383 | <0.0001 | <0.0001 |
| No diabetes | nHW | 557 | 548 | 477 | 468 | 552 | 472 | ||
| T1D | Hispanic | 2708 | 2747 | 2354 | 2381 | 2726 | 2366 | <0.0001 | <0.0001 |
| T1D | nHW | 2223 | 2191 | 1928 | 1890 | 2208 | 1910 | ||
| T2D | Hispanic | 1390 | 1431 | 1211 | 1249 | 1410 | 1230 | <0.0001 | <0.0001 |
| T2D | nHW | 1233 | 1241 | 1067 | 1071 | 1237 | 1069 | ||
| Adjusted | |||||||||
| No diabetes | Hispanic | 608 | 643 | 530 | 565 | 626 | 548 | <0.0001 | <0.0001 |
| No diabetes | nHW | 528 | 616 | 451 | 529 | 572 | 490 | ||
| T1D | Hispanic | 2138 | 2292 | 1869 | 1999 | 2214 | 1933 | <0.0001 | <0.0001 |
| T1D | nHW | 2176 | 2479 | 1886 | 2148 | 2320 | 2010 | ||
| T2D | Hispanic | 1170 | 1468 | 1020 | 1294 | 1320 | 1158 | <0.0001 | <0.0001 |
| T2D | nHW | 1214 | 1387 | 1050 | 1203 | 1300 | 1127 |
a2012–2013 average.
T1D, type 1 diabetes; T2D, type 2 diabetes; nHW, non-Hispanic white.
For individuals with T2D, unadjusted and adjusted allowed and paid costs PMPM for 2012–2013 combined were higher for Hispanics compared with their nHW counterparts (both p<0.0001). For individuals with T1D, unadjusted costs were higher for Hispanics, but after adjusting for potential confounders, costs were higher for nHWs.
Discussion
In the United States, Hispanics represent the fastest growing ethnic and minority group and are known to have an excess burden of diabetes and associated complications compared with the general population.27 Meanwhile, the number of older people developing diabetes has also increased exponentially over recent years.28
It is also known that older people with diabetes have significantly more comorbidities, such as myocardial infarction, stroke, peripheral arterial disease, and renal impairment, compared with older people without diabetes.2
To assess the burden of diabetes among seniors by race and ethnicity in the United States, we analyzed the Medicare 5% sample file by race and ethnicity for both T1D and T2D for the years 2012–2013.
Although nHWs accounted for the majority of this population, the prevalence of T1D and T2D was higher for Hispanics for both types of diabetes. Our economic findings related to this higher prevalence show that a diagnosis of diabetes adds significantly to the already increased costs for older individuals and Hispanics in the US population. Additionally, Hispanics tended to use more acute care resources, also, resulting in longer hospit
al stays. This highlights an important disparity in the use of health care resources across the two populations and may provide an opportunity for new approaches toward more cost-effective use of health care resources.
The health economic burden for Hispanic seniors with diabetes reported here is notable, but unsurprising. In the United States, ∼1 in 9 adults has diabetes; however, the prevalence of both diagnosed and undiagnosed diabetes is nearly twice as high among Hispanic adults of Mexican origin than nHW adults.27
Furthermore, Hispanics are also disproportionately impacted by complications of diabetes.28 The reasons for the excess burden of diabetes in Hispanic adults include confounding factors such as acculturation, age, and socioeconomic status in addition to much-studied biological risk factors such as obesity and high blood pressure.29–32
In the Medicare population, age per se also appears to add significantly to the health economic burden for individuals with diabetes beyond the costs of common microvascular and macrovascular complications.33
This may be relevant given our finding that in this cohort, the average age of Hispanics with diabetes was significantly higher compared with nHWs and could have added to the cost burden. As a corollary, some of the differences seen here could be attributed to the age bias of the sample and may not hold up when applied to the entire US population.
At all ages, rates of screening for complications of diabetes have been reported to be lower for Hispanics.31 Surprisingly, in this older population, we found that rates of screening for diabetes-related complications, HbA1c, and lipid assessments were performed more often in Hispanic seniors with diabetes than their nHW counterparts.
One possible explanation for this finding could be increased proactivity among clinicians once this population is diagnosed with diabetes, and perhaps not reflective of screening and prevention rates before diagnosis.
Indeed, the positive effects of insurance coverage on health outcomes for adults with long-term conditions such as diabetes include greater use of health services and improved health outcomes, including disease control.34
Thus, these findings are consistent with recent research indicating care improvement after diagnosis and suggest that access to Medicare further enables screening for complications as well as advanced treatment.
Fewer studies have been conducted on the complications of diabetes, and these generally suggest that the rates of complications of diabetes vary by type and by racial and ethnic minority groups.35–38
Consistent with the present study, another has reported that rates of early- and end-stage kidney disease are up to 2.5 times higher among Hispanics compared with nHWs.39 Despite a higher prevalence of obesity and risk factors for cardiovascular disease in Hispanics, some studies have suggested that Hispanics may have a lower risk of cardiovascular-related mortality (the Hispanic paradox).40 However, as this is not a consistent finding, cardiovascular disease in the Hispanic population needs further study.41–42
Limitations
There are a number of important limitations to this study. The Medicare 5% sample applies only to seniors aged 65 years and above and therefore the results may not be applicable to younger age groups with diabetes. The sample size of Hispanic seniors with diabetes was also relatively small, limiting the applicability of findings.
Furthermore, the T1D cohort is relatively small and the dataset does not detail the criteria for diagnosis of T1D versus other forms of diabetes. For the purposes of this study, CMS-approved ICD-9 diagnosis codes for T1D and T2D were used, which could have resulted in coding misclassifications.
In addition, given the negative impact of a diagnosis of T1D at a young age on long-term survival,43 this cohort is unlikely to be representative of T1D in general. The Medicare 5% File does not contain outpatient prescription drug data, with only payments for inpatient medications included in the file.
There are also differences in the distribution of racial and ethnic groups between the national census and the Medicare 5% File. As per the national census, Hispanics comprised almost 8% of the population aged ≥65 years; however, in the Medicare 5% File, Hispanics comprised only 1.7% of all racial and ethnic groups. This could be due to varying definitions of race and ethnicity among the US census versus Medicare.
Conclusion
Analysis of data from the Medicare 5% File showed that a diagnosis of diabetes in older Americans adds significantly to the cost of health care, and the prevalence of diabetes appears to be higher for Hispanics than for nHWs.
Resource utilization, including hospital admissions and length of stay, was higher in Hispanics with diabetes, resulting in higher health care costs for this growing minority. While the disproportionate use of health care resources could be attributed to a variety of factors, the resulting spending for the health care system underlines the need for new approaches to diabetes care.
Considering the high burden of chronic disease in Hispanics and the exponential growth in older US and Hispanic populations, this call for action to our health care providers is even more imperative.
More information: The American Academy of Family Physicians has more on diabetes.
