Abstract
Background:
Racial and ethnic minority groups have a higher prevalence of diabetes, increased risk for adverse complications, and worse health outcomes compared to Non-Hispanic Whites. Evidence suggests they also have higher healthcare expenses associated with diabetes care. Therefore, the objective of this study was to assess racial and ethnic differences in out-of-pocket (OOP) costs among a nationally representative sample of adults with diabetes.
Methods:
Cross-sectional study of 17,702 adults (aged ≥18 years) with diabetes from years 2002–2011 in the Medical Expenditure Panel Survey Household Component. The outcome was OOP expenditures, and the primary predictor was race/ethnicity. Descriptive statistics summarized the sample population. Unadjusted mean values were computed to compare OOP expenses overtime. A two-part model was used to estimate adjusted incremental OOP expenses.
Results:
For the overall sample, OOP expenditures decreased significantly over time. In addition, compared to NHWs, racial and ethnic minority groups had significantly lower OOP costs per year when adjusted for sociodemographic characteristics, comorbid conditions, and time. NHBs paid $481 less than NHWs; Hispanics paid $591 less than NHWs; and individuals in the ‘Other’ racial/ethnic category paid nearly $645 less compared to NHWs (p < 0.001).
Conclusions:
Racial/ethnic minority patients with diabetes had significantly less OOP expenses compared to NHWs, possibly due to differences in healthcare utilization. OOP expenses decreased significantly over time for all racial and ethnic groups. Additional research is needed to understand the factors associated with differences in OOP expenditures among racial groups.
Keywords: Race/ethnicity, Disparities, Expenditures, Adults, Diabetes
INTRODUCTION
The prevalence of diabetes mellitus in the United States has tripled since the 1980s, currently affecting over 30 million Americans.1 The number of Americans with diabetes is expected to increase, as up to one-third of patients with prediabetes have a greater risk of developing diabetes within five years.2,3 Diabetes is estimated to cost $245 billion annually secondary to total medical expenditures and lost productivity and wages related to the diagnosis.1 The average person with diabetes pays $12,180 US dollars in comparison to a person without diabetes paying $5058 in unadjusted average direct expenditures for medical care.4 Over 60% of these costs are due to prescription drugs needed to maintain blood glucose levels within the appropriate range and inpatient stays secondary to diabetes exacerbated complications.5,6 Despite previous research showing an overall 5% reduction in out-of-pocket (OOP) costs between 2002 and 2011 for diabetes care, almost 25% of patients with diabetes still have a high OOP burden.3,7 The need for comprehensive treatment plans requiring frequent access for management combined with the risk of complications requiring hospital admissions are believed to contribute to this exacerbation in expenses.7,8
Racial and ethnic minority groups historically have a higher prevalence of diabetes, adverse complications, and worse health outcomes such as glycemic control compared to Non-Hispanic Whites (NHW).2,5 For example, approximately 13% of both Hispanics and Non-Hispanic Blacks (NHB) and 16% of American Indians/Alaska Natives have diabetes compared to only 8% of NHWs.1 Additionally, NHBs are 2.6 times more likely to have end-stage renal disease and higher mortality despite receiving recommended care in comparison to NHW.9,10 Because diabetes is such a prevalent condition, especially in minority populations, it is important to understand the OOP costs associated with both short- and long-term diabetes treatment and management.11 In general, evidence shows that people diagnosed with diabetes accumulate medical-related costs 2—3 times more in a lifetime than people with similar lifestyles who do not have diabetes.12,13 Earlier research suggests disparities in socioeconomic status and quality of care for diabetes patients may contribute to this economic divide.13–16 In addition, it has been found that NHBs and Hispanics with type 2 diabetes have lower costs than NHWs due to less access to care and less access to prescription insurance coverage.17
Prior research shows a correlation between glycemic control and costs for diabetes patients, such that better glycemic control is associated with lower costs.18 Cost models suggest that by maintaining better glycemic control, patients with diabetes can benefit by saving nearly $400 on OOP costs within one or two years.19 Given higher costs for care in patients with diabetes and the higher prevalence of diabetes in racial/ethnic minorities, the objective of this study was to assess racial and ethnic differences in OOP expenditures among a nationally representative sample of adults with diabetes with the hypothesis that significant differences in OOP expenses will be observed among different racial and ethnic population groups.
METHODS
Data source and sample
The study used data from the Medical Expenditure Panel Survey Household Component (MEPS-HC) from 2002 to 2011, a survey cosponsored by Agency for Healthcare Research and Quality (AHRQ) and the National Center for Health Statistics. We used full-year consolidated data from MEPS-HC to examine the association between race/ethnicity and OOP expenditures among adults (aged ≥18 years) with diabetes. MEPS-HC is a nationally representative survey that collects detailed information on healthcare utilization and expenditures, health insurance, and health status, as well as a wide variety of social, demographic and economic characteristics for the U.S. civilian, non-institutionalized population. In this retrospective study, we identified 17,702 adults (weighted sample of 17,857,174) with self-reported diabetes from MEPS-HC.20–22 The AHRQ validates MEPS as a self-reported instrument by administering many quality assurance procedures like validation of an interviewer’s work and comparing MEPS numbers with other data source numbers like the Census Bureau and National Health Interview Survey.22
To ensure sufficient sample size and statistical power, we pooled 10 years of MEPS data and computed mean estimates per year from the pooled data. Our study accounts for the sampling weights, clustering, and stratification design of MEPS to yields nationally representative estimates for the non-institutionalized U.S. population.22 OOP expenditures for all years were adjusted to the 2014 dollar using the consumers price index obtained from the Bureau of Labor Statistics (BLS) (http://data.bls.gov/cgi-bin/cpicalc.pl).23
MEASURES
Variables of interest
The dependent variable in this study was the OOP expenses for the calendar year of each observation. The OOP expenses were defined as the amount paid by the family for medical provider visits, non-physician services, hospital inpatient stays, emergency room visits, dental visits, home healthcare, and prescription medications. The primary independent variable was race/ethnicity that was categorized into: NHW, NHB, Hispanic, or others.
The sociodemographic controlled covariates included age, sex, marital status, education, health insurance, Metropolitan Statistical Area (MSA) status, region, and year category. Binary indicators of comorbidities were based on a positive response to a question, “Have you ever been diagnosed with…?” Cardiovascular disease (CVD) indicated a positive response to coronary heart disease, angina, myocardial infarction, or other heart diseases. The income level was defined as a percentage of the poverty level and grouped into four categories: poor (<125%), low income (125% to less than 200%), middle income (200% to less than 400%) and high income (≥ 400%).
Statistical analyses
The demographic characteristics of adults with diabetes are presented by racial/ethnic category. Chi-squared (χ2) tests were used to examine the bivariate relationship between race/ethnic groups across covariates. Unadjusted means were computed to compare OOP expenses over time. Medical expenditures have a large fraction of observations at zero and are positively skewed, where the assumption of the normality of the error term is not satisfied. Thus, a two-part model was used to estimate adjusted incremental OOP expenses.24,25 In the two-part model, a probit model was estimated for the probability of observing a zero versus positive OOP expenditure. Conditional on having positive expenditure, a generalized linear model (GLM) was then estimated in the second part.26 The margins command in Stata was used to calculate marginal effects and their standard errors from the combined first and second parts of the final model.26 For the two-part model, the use of GLM in the second part had an advantage over log Ordinary Least Squares (OLS) since it relaxes the normality and homoscedasticity assumptions and avoids bias associated with retransforming to the raw scale.26
We conducted diagnostic test using the modified Park test to examine the model fit. The results of the modified Park test verified the use of a gamma distribution with a log link as the best fitting GLM to get consistent estimations of coefficients and marginal effects of OOP expenditures. Multicollinearity was checked for predictors of the two-part model accounting for the complex survey design. Variance inflation factor (VIF) for all predictors used in the two-part model indicated no multicollinearity problems.
We used standard pairwise comparison methods of Sidak, Scheffe, and Bonferroni to compare the pooled mean OOP expense over time. We also used these methods to test for statistically significant differences between OOP expenditures and OOP health services among poverty levels (poor vs. low income, poor vs. middle income, poor vs. high income, low vs. middle income, low vs. high income, and middle vs. high income).27,28 We compared mean OOP expenses between three year groups (2002/05 vs. 2006/09, 2002/05 vs. 2010/11, 2006/09 vs. 2010/11). All analyses were performed at person-level using Stata 14.29 Estimates that were statistically significant at the p < 0.05 level are discussed in this paper.
RESULTS
Table 1 shows the sample demographics by race and ethnicity of adults with diabetes in this study. For the overall sample, nearly 48% was NHW, followed by 23% who were Hispanic, 22% who were NHBs, and 7% who identified with other races and/or ethnicities. Thirteen percent of the sample was between the ages of 18–44 years, 47% was 45–64 years, and 40% was 65 years of age and older. Fifty-one percent of the overall sample was comprised of women. Fifty-eight percent was married, 32% not married, and 9% never married. Approximately 26% had less than a high school education, while 35% and 39% had a high school diploma and attended college, respectively. Sixty-one percent had private insurance, 31% had public insurance, and 8% were uninsured. The majority of the sample (80%) lived in an urban, MSA area. Nearly 18% resided in the Northeast, 21% in the Midwest and West, and 40% in the South. Approximately 20%, 16%, 31%, and 33% self-reported their poverty income ratio as poor, low, middle, and high, respectively. In terms of comorbid conditions among the entire sample, 73% reported hypertension, 32% reported cardiovascular disease, 10% reported stroke, 5% reported emphysema, 56% reported joint pain, 49% reported arthritis, and 14% reported asthma.
Table 1.
Sample demographics among adults with diabetes (N = 17,702).
| Non-Hispanic White (n = 8483) |
Non-Hispanic Black (n = 3932) |
Hispanic (n = 4083) |
Other (n = 1204) |
P-Value | |
|---|---|---|---|---|---|
| Age (years) | <0.001* | ||||
| 18–44 | 53.5 | 18.4 | 20.7 | 7.4 | |
| 45–64 | 62.7 | 16.1 | 14.1 | 7.1 | |
| 65+ | 70.5 | 13.2 | 10.5 | 5.8 | |
| Sex | <0.001* | ||||
| Male | 68.4 | 12.3 | 12.7 | 6.6 | |
| Female | 61.0 | 18.0 | 14.3 | 6.7 | |
| Marital status | <0.001* | ||||
| Married | 69.4 | 10.4 | 13.3 | 6.9 | |
| Unmarried | 60.6 | 20.6 | 12.5 | 6.3 | |
| Never married | 48.2 | 27.2 | 18.8 | 5.8 | |
| Education | <0.001* | ||||
| Less than high school | 49.5 | 17.2 | 27.9 | 5.4 | |
| High school | 69.8 | 15.5 | 9.5 | 5.2 | |
| College | 70.9 | 13.5 | 7.1 | 8.5 | |
| Insurance | <0.001* | ||||
| Private | 71.9 | 12.7 | 9.1 | 6.3 | |
| Public | 55.2 | 20.0 | 17.7 | 7.2 | |
| Uninsured | 45.1 | 15.9 | 32.2 | 6.8 | |
| MSA | <0.001* | ||||
| Non-MSA | 79.3 | 10.1 | 5.9 | 4.8 | |
| MSA | 60.9 | 16.6 | 15.5 | 7.1 | |
| Region | <0.001* | ||||
| Northeast | 67.7 | 15.7 | 10.8 | 5.8 | |
| Midwest | 78.4 | 12.5 | 5.0 | 4.1 | |
| South | 61.8 | 21.4 | 12.3 | 4.5 | |
| West | 53.4 | 5.8 | 26.9 | 13.9 | |
| Poverty income ratio | <0.001* | ||||
| Poor | 48.7 | 23.8 | 20.7 | 6.8 | |
| Low income | 56.1 | 18.3 | 18.9 | 6.7 | |
| Middle income | 66.6 | 13.6 | 13.3 | 6.5 | |
| High income | 76.6 | 10.1 | 6.7 | 6.6 | |
| Hypertension | <0.001* | ||||
| No | 66.0 | 10.5 | 16.4 | 7.1 | |
| Yes | 64.1 | 16.9 | 12.5 | 6.5 | |
| Cardiovascular disease | <0.001* | ||||
| No | 61.3 | 16.1 | 15.3 | 7.2 | |
| Yes | 71.8 | 13.3 | 9.6 | 5.3 | |
| Stroke | <0.001* | ||||
| No | 64.2 | 15.0 | 13.9 | 6.9 | |
| Yes | 67.6 | 17.4 | 10.3 | 4.7 | |
| Emphysema | <0.001* | ||||
| No | 63.6 | 15.7 | 13.9 | 6.8 | |
| Yes | 84.1 | 6.8 | 5.1 | 3.9 | |
| Joint pain | <0.001* | ||||
| No | 60.1 | 15.9 | 16.8 | 7.3 | |
| Yes | 68.1 | 14.8 | 10.9 | 6.2 | |
| Arthritis | <0.001* | ||||
| No | 60.3 | 15.3 | 16.6 | 7.8 | |
| Yes | 69.1 | 15.2 | 10.4 | 5.3 | |
| Asthma | <0.001* | ||||
| No | 64.3 | 15.1 | 14.0 | 6.6 | |
| Yes | 66.4 | 16.5 | 10.3 | 6.8 | |
| Categories by year | 0.759 | ||||
| 2002–2005 | 64.8 | 15.6 | 12.9 | 6.6 | |
| 2006–2009 | 65.3 | 14.7 | 13.5 | 6.6 | |
| 2010–2011 | 63.1 | 15.7 | 14.5 | 6.7 |
All numbers represent percentages.
Statistically significant at p < 0.05.
Table 2 shows the unadjusted means for OOP expenditures over time. The average OOP for the sample from 2002 to 2005 was $2117, followed by $1665 for years 2006–2009, and $1391 for 2010–2011. The results of the pairwise comparison tests showed the unadjusted mean OOP expense variations over time and were consistent across the three periods. Compared to 2002/05, the unadjusted mean for 2006/09 was $452 lower (95% CI: −602, −302), and the unadjusted mean for 2010/11 was $726 lower (95% CI; −892, −560). Compared with 2006/09, the unadjusted mean for 2010/11 was $273 lower (95% CI: −410, −137).
Table 2.
Unadjusted mean scores for out-of-pocket (OOP) expenditures over time.
| Years | Average OOP | 95% Confidence interval |
|---|---|---|
| 2002–2005 | $2117.74 | (2014.6, 2220.9)a |
| 2006–2009 | $1665.10 | (1584.6,1745.5)a |
| 2010–2011 | $1391.22 | (1302.1,1480.3)a |
Statistically significant at 95% Confidence Interval.
The unadjusted mean OOP expenditures and OOP health service expenditures of poverty category based on race/ethnic groups are presented in Table 3. For NHWs, the unadjusted mean OOP expenditure ranged between $1820 for poor to $$2082 for high income groups. For NHBs, the unadjusted mean OOP expenditure ranged between $1349 for poor to $1704 for high income group. For Hispanics, the unadjusted mean OOP costs ranged between $987 for poor to $1474 for high income. For other race/ethnic groups, the unadjusted mean OOP ranged between $1085 for poor to $1572 for middle income. We also used the multiple comparison test when the OOP expenditures and OOP health services expenditures were statistically significant different among poverty levels. The results of the pairwise comparison test showed that the mean OOP expenditures for Hispanic groups were $487 higher (95% CI: 70, 904) for high income compared with the poor. For NHBs, OOP office-based expenditure for high income was $125 higher (95% CI: 17; 233) compared with low income. For Hispanics, OOP office-based expenditure for high income was $153 higher (95% CI: 32, 274) compared with the poor. All other OOP expenditure and OOP health service expenditure were not statistically significant at 95% CI across income/poverty categories.
Table 3.
Unadjusted mean of out-of-pocket (OOP) expenditures of poverty category by racial/ethnic group (95% confidence interval).
| NHW | NHB | Hispanic | Other | |
|---|---|---|---|---|
| Total OOP expenditures | ||||
| Poor | $1820 (1656,1984) | $1394 (1105,1684) | $988 (852,1123) | $1086 (847,1325) |
| Low income | $2042 (1888,2197) | $1413 (1223,1603) | $1128 (974,1283) | $1097 (849,1346) |
| Middle income | $1998 (1877,2119) | $1466 (1227,1705) | $1231 (1094,1368) | $1571 (1269,1875) |
| High income | $2082 (1949,2216) | $1705 (1369,2041) | $1475 (1195,1755) | $1252 (1064,1442) |
| Prescription expenditure | ||||
| Poor | $1233 (1119,1349) | $904 (814,995) | $710 (605,815) | $819 (607,1032) |
| Low income | $1412 (1294,1530) | $1069 (896,1241) | $834 (703,965) | $682 (508,857) |
| Middle income | $1197 (1132,1262) | $975 (847,1104) | $788 (683,893) | $995 (736,1255) |
| High income | $1176 (1078,1274) | $833 (736,929) | $889 (680,1098) | $644 (537,751) |
| Office-based expenditure | ||||
| Poor | $221 (143,301) | $157 (34,280) | $83 (63,102) | $99 (61,138) |
| Low income | $204 (178,232) | $125 (103,147) | $115 (84,146) | $142 (89,195) |
| Middle income | $290 (250,330) | $223 (97,350) | $166 (139,193) | $229 (123,336) |
| High income | $326 (294,359) | $251 (173,328) | $236 (148,324) | $217 (144,290) |
| Inpatient expenditure | ||||
| Poor | $82 (43,123) | $91 (34,148) | $69 (8129) | $12 (−5,29) |
| Low income | $61 (40,83) | $40 (19,60) | $29 (−2,61) | $80 (−39,200) |
| Middle income | $50 (38,64) | $70 (−3143) | $75 (14,136) | $38 (3,73) |
| High income | $50 (31,70) | $139 (−12,289) | $59 (−17,136) | $37 (10,63) |
| Outpatient expenditure | ||||
| Poor | $43 (27,58) | $22 (8,36) | $13 (7,18) | $16 (4,29) |
| Low income | $60 (35,87) | $17 (10,25) | $17 (9,26) | $41 (24,58) |
| Middle income | $57 (42,73) | $39 (−0.20,78) | $36 (14,58) | $23 (7,38) |
| High income | $59 (48,72) | $42 (24,59) | $24 (12,37) | $46 (−1,93) |
| ER expenditure | ||||
| Poor | $25 (15,35) | $84 (−21,189) | $25 (2,47) | $13 (−1,27) |
| Low income | $20 (13,28) | $20 (−4,45) | $21 (8,34) | $9 (4,15) |
| Middle income | $19 (11,27) | $9 (5,13) | $15 (9,20) | $4 (1,8) |
| High income | $20 (14,26) | $60 (−24,143) | $19 (7,31) | $10 (4,17) |
Table 4 shows estimates for the adjusted two-part model for OOP expenditures across race and ethnicity. Compared to NHWs, all the other racial/ethnic groups had significantly lower OOP costs per year when adjusted for sociodemographic characteristics, comorbid conditions, and time. Compared to NHWs, NHBs had significantly lower OOP expenditures by more than $480 (95% CI: −621.30, −.51; p < 0.001). Hispanics had significantly lower OOP expenses of more than $590 compared to NHWs (95% CI: −727.38, −455.00; p < 0.001). Finally, the ‘Other’ racial/ethnic category had significantly lower OOP expenses of nearly $645 compared to NHWs (95% CI: −803.07, −484.47; p < 0.001).
Table 4.
Adjusted regression for out-of-pocket (OOP) expenditures across race/ethnicity.
| β-Coefficient | 95% Confidence interval | P-value | |
|---|---|---|---|
| Race | P < 0.001a | ||
| Non-hispanic black | −481 | (−621.30, −341.51) | |
| Hispanic | −591 | (−727.38, −455.00) | |
| Other | −643 | (−803.07, −484.47) | |
| Age | P < 0.001a | ||
| 45-64 | 299 | (146.90, 453.00) | |
| 65+ | 559 | (375.94, 742.18) | |
| Sex | P < 0.001a | ||
| Females | 273 | (155.53,392.15) | |
| Marital status | |||
| Unmarried | −1 | (−123.78, 120.78) | P = 0.981 |
| Never married | −101 | (−290.96, 88.36) | P = 0.295 |
| Education | |||
| High School | 116 | (0.33, 233.63) | P = 0.049a |
| College | 479 | (328.22, 631.53) | P < 0.001a |
| Insurance status | |||
| Public | −166 | (−298.63, −35.28) | P = 0.013a |
| Uninsured | 962 | (642.18, 1282.23) | P < 0.001a |
| Metropolitan statistical area | |||
| MSA | 8 | (−108.96, 125.76) | P = 0.888 |
| Region | |||
| Midwest | 193 | (48.64, 338.59) | P = 0.009a |
| South | 309 | (181.58, 437.33) | P < 0.001a |
| West | 165 | (6.45, 324.68) | P = 0.041a |
| Income | |||
| Low income | 89 | (−45.88, 224.54) | P = 0.195 |
| Middle income | 189 | (44.43, 334.88) | P = 0.010a |
| High income | 343 | (170.06, 517.49) | P < 0.001a |
| Comorbidity burden | |||
| Hypertension | 312 | (203.48, 421.43) | P < 0.001a |
| Cardiovascular disease | 367 | (251.93, 483.39) | P < 0.001a |
| Emphysema | 185 | (−37.93, 409.34) | P = 0.104 |
| Joint pain | 113 | (13.92, 213.26) | P = 0.026a |
| Arthritis | 268 | (160.81, 376.02) | P < 0.001a |
| Asthma | 190 | (53.95, 327.16) | P = 0.006a |
| OOP differences by time (years) | P < 0.001a | ||
| 2006–2009 | −516 | (−640.71, −393.06) | |
| 2010–2011 | −808 | (−939.77, −676.37) | |
Reference groups: NHW, 18–44 years, men, married, less than high school, private insurance, non-metropolitan statistical area, northeast region, poor income, absence of comorbidities (hypertension, cardiovascular disease, emphysema, joint pain, arthritis, and asthma), and OOP differences for years 2002–2005.
NHW, Non-Hispanic White.
Statistically significant at p < 0.05. β = Beta coefficient. Adjusted for covariates: sex, education, income, age, marital status, insurance, region, time, and comorbidity burden.
DISCUSSION
In this nationally representative sample of adults with diabetes, individuals belonging to racial/ethnic minority groups had significantly less OOP expenses compared to NHWs. These observed differences in OOP expenditures among the different racial and ethnic groups can be explained by higher healthcare utilization among NHW compared to racial/ethnic minorities.13,30–33 We are unable to determine whether this utilization variance is due to provider driven bias in treatment and prescription regimen or varying needs of minority patients compared to NHWs. Chen et al. associated differences in receipt of appropriate Diabetes Self-Management Education with varying outcomes for patients, with Hispanic groups being the least likely to receive appropriate education and have the least amount of improvement in care measures.4 Likewise, NHBs were likely to receive proper self-care education, but not have favorable outcomes due to less motivation to follow through with care.4 Also, our findings of OOP expenses decreasing significantly over time are supported by evidence from previous studies in the literature.34
In this nationally representative sample of adults with diabetes, OOP expenses decreased significantly over time. Previous research indicates there has been a 5% decrease in OOP expenses between 2001 and 2011.7 In a comparison of healthcare expenditures in patients with and without diabetes, patients with diabetes experienced a decrease in medical OOP expenses from 2005 to 2011 (11). This overall decrease in average OOP expenses may be due to policy changes in Medicaid and Medicare that limit the number of brand-name drugs prescribed at one time.13,35 Many states created legislative provisions that encouraged physicians to prescribe less costly pharmaceuticals. With prescription drugs accounting for 60% of OOP costs for diabetes patients, prescribing quality medications that cost less could be advantageous by leading to more overall cost reductions for care among patients.11 An additional explanation for the decrease in OOP expenses is that patients are not able to afford the costs of prescriptions, equipment, and supplies such as testing strips for self-monitoring, resulting in these items not being taken or used to self-manage care.
The findings of our study are important because they offer information that can direct and inform future research, policy, and clinical practice. While overall OOP expenditures continue to decrease over time, additional research is needed to identify the factors contributing to the differences in OOP expenses among racial/ethnic groups. In a study to assess medical spending associated with diabetes, Hu et al. found significant disparities in total and OOP expenditures by age, gender, education, race/ethnicity, and insurance status.35 Costs associated with prescription medications for diabetes have been shown to contribute to higher OOP expenses for personalized and individualized diabetes management.5,6,15 Many patients, particularly minority patients with diabetes, reported non-adherence to medication regimens due to high prescription costs despite Medicare’s efforts to provide more affordable options mentioned in prior literature.13,36–38 More specifically Hispanic and NHB patients of lower income categories were most likely to report this cause for non-adherence.15 Additionally, lower utilization of healthcare services by racial and ethnic minority groups may have contributed to a decline in OOP expenses for these specific populations.2,4 It will be valuable to reevaluate diabetes management guidelines and develop strategies that will increase access to care, while reducing financial burden to patients. Continued expansion of diabetes education with the additional component of customized approaches to individual patient lifestyles and innovative technologies would likely improve glycemic control and contribute to better health outcomes across all racial groups. Clinicians should continue to educate patients on the complications associated with uncontrolled diabetes and the patients’ pivotal role in action to prevent such complex comorbities. By doing so, this can also contribute to reductions in OOP and total medical expenditures for care associated with diabetes management.
There are study limitations that are noteworthy to mention. First, the design of this study was cross-sectional. As such, we cannot infer causal relationships. Second, the data was self-reported; therefore, over- and underreporting in responses is possible. Given the validity of the dataset and extensive testing undertaken by AHRQ, however, this is not as significant a concern as with other studies. Third, given that MEPS data does not distinguish between a diagnosis of type 1 and type 2 diabetes, we are unable to distinguish the specific diagnoses of the participants included in our sample. However, as 95% of patients with diabetes are diagnosed with type 2 diabetes, this nationally representative sample likely has similar proportions.
CONCLUSIONS
In this nationally representative sample of adults with diabetes, individuals belonging to the racial/ethnic minority groups had significantly less OOP expenses compared to NHWs, and OOP expenditures significantly decreased over time. Future research should continue to investigate this area to determine concrete reasons for the racial and ethnic differences in costs and offer solutions for addressing these differences. It is important to discover ways to begin narrowing the multifaceted health disparities gaps between racial groups.
Supplementary Material
Footnotes
Conflict of interest: The authors report no potential conflicts of interest relevant to this article.
Publisher's Disclaimer: Disclaimer: This article represents the views of the authors and not those of AHRQ, the Medical University of South Carolina, or the Medical College of Wisconsin.
Guarantors: LEE, JSW, and KB are the guarantors of the study and take full responsibility for the work as a whole, including the study design, access to data, and the decision to submit and publish the manuscript.
Supplementary data related to this article can be found at https://doi.org/10.1016/j.jnma.2018.04.004.
Contributor Information
Makiera Simmons, Center for Health Disparities Research, Department of Medicine, Medical University of South Carolina, Charleston, SC 29425, USA
Kinfe G. Bishu, Center for Health Disparities Research, Department of Medicine, Medical University of South Carolina, Charleston, SC 29425, USA; Department of Medicine, Division of General Internal Medicine and Geriatrics, Medical University of South Carolina, Charleston, SC 29425, USA
Joni S. Williams, Department of Medicine, Division of General Internal Medicine, Medical College of Wisconsin, 9200 W. Wisconsin Ave, Milwaukee, WI 53226, USA; Center for Patient Care and Outcomes Research, Medical College of Wisconsin, 8701 Watertown Plank Road, Room H3165, Milwaukee, WI 53226, USA
Rebekah J. Walker, Department of Medicine, Division of General Internal Medicine, Medical College of Wisconsin, 9200 W. Wisconsin Ave, Milwaukee, WI 53226, USA; Center for Patient Care and Outcomes Research, Medical College of Wisconsin, 8701 Watertown Plank Road, Room H3165, Milwaukee, WI 53226, USA
Aprill Z. Dawson, Department of Medicine, Division of General Internal Medicine, Medical College of Wisconsin, 9200 W. Wisconsin Ave, Milwaukee, WI 53226, USA; Center for Patient Care and Outcomes Research, Medical College of Wisconsin, 8701 Watertown Plank Road, Room H3165, Milwaukee, WI 53226, USA
Leonard E. Egede, Department of Medicine, Division of General Internal Medicine, Medical College of Wisconsin, 9200 W. Wisconsin Ave, Milwaukee, WI 53226, USA; Center for Patient Care and Outcomes Research, Medical College of Wisconsin, 8701 Watertown Plank Road, Room H3165, Milwaukee, WI 53226, USA
REFERENCES
- 1.Centers for Disease Control and Prevention. (2017). National Diabetes Statistics Report. Atlanta, GA: Centers for Disease Control and Prevention, U.S. Dept. of Health and Human Services; 2017. [Google Scholar]
- 2.Campbell JA, Walker RJ, Smalls BL, & Egede LE (2012. December). Glucose control in diabetes: the impact of racial differences on monitoring and outcomes. Endocrine, 42(3), 471–482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Walker RJ, Strom Williams J, & Egede LE (2016. April). Influence of race, ethnicity and social determinants of health on diabetes outcomes. Am J Med Sci, 351 (4), 366–373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chen R, Cheadle A, Johnson D, & Duran B (2014. Nov-Dec). US trends in receipt of appropriate diabetes clinical and self-care from 2001 to 2010 and racial/ethnic disparities in care. Diabetes Educat, 40(6), 756–766. [DOI] [PubMed] [Google Scholar]
- 5.Meng YY, Pickett MC, Babey SH, Davis AC, & Goldstein H (2014. May). Diabetes tied to a third of California hospital stays, driving health care costs higher. PB2014–3 Policy Brief UCLA Cent Health Policy Res, 1–7. [PubMed] [Google Scholar]
- 6.Trogdon JG, & Hylands T (2008. December). Nationally representative medical costs of diabetes by time since diagnosis. Diabetes Care, 21(12), 2307–2311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Li R, Barker LE, Shrestha S, et al. (2014. June). Changes over time in high out-of-pocket health care burden in U.S. adults with diabetes, 2001–2011. Diabetes Care, 37(6), 1629–1635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Diabetes Care, 39(suppl 1), (2016. Jan), S1–S2. [DOI] [PubMed] [Google Scholar]
- 9.U.S. Centers for Disease Control and Prevention. Diabetes Public Health Resource. (2011). National Diabetes Fact Sheet; [Internet], 2011. Available from: www.cdc.gov/diabetes/pubs/estimates11.htm. Accessed June 16, 2016. [Google Scholar]
- 10.Walker RJ, Gebregziabher M, Martin-Harris B, & Egede LE (2014. Nov-Dec). Independent effects of socioeconomic and psychological social determinants of health on self-care and outcomes in Type 2 diabetes. Gen Hosp Psychiatry, 36(6), 662–668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ozieh MN, Bishu KG, Dismuke CE, & Egede LE (2015. October). Trends in health care expenditure in U.S. adults with diabetes: 2002-2011. Diabetes Care, 38(10), 1844–1851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.E H, Sandhu, Newton N, McCulloch KM, Ramsey DK, S D., & Grothaus LC. (2001. January 10). Effect of improved glycemic control on health care costs and utilization. JAMA, 285(2), 182–189. [DOI] [PubMed] [Google Scholar]
- 13.Li R, Bilik D, Brown M, et al. (2013). Medical costs associated with type 2 diabetes complications and comorbities. Am J Manag Care, 19(5), 421–430. [PMC free article] [PubMed] [Google Scholar]
- 14.Zhuo X, Zhang P, Barker L, Albright A, Thompson TJ, & Gregg E (2014. September). The lifetime cost of diabetes and its implications for diabetes prevention. Diabetes Care, 37(9), 2557–2564. 10.2337/dc13-2484. [DOI] [PubMed] [Google Scholar]
- 15.Tseng C, Tierney E, Gerzoff R, et al. (2008. February). Race/ethnicity and economic differences in cost-related medication underuse among insured adults with diabetes. Diabetes Care, 31(2), 261–266. [DOI] [PubMed] [Google Scholar]
- 16.Robbins JM, Vaccarino V, Zhang H, & Kasl SV (2001. January). Socioeconomic status and type 2 diabetes in African American and non-hispanic white women and men: evidence from the third national health and nutrition examination survey. Am J Public Health, 91(1), 76–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lee J-A, Liu C-F, & Sales AE (2006). Racial and ethnic differences in diabetes care and health care use and costs. Prev Chronic Dis Public Health Res Practice Policy, 3(3), 1–12. [PMC free article] [PubMed] [Google Scholar]
- 18.Williams DR, Costa MV, Odunlami AO, & Mohammed SA (2008. November). Moving upstream: how interventions that address the social determinants of health can improve health and reduce disparities. J Public Health Manag Pract, (14 suppl), S8–S17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Menzin J, Korn JR, Cohen J, et al. (2010. May). Relationship between glycemic control and diabetes-related hospital costs in patients with type 1 or type 2 diabetes mellitus. J Manag Care Pharm, 16(4), 264–275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Agency for Healthcare Research and Quality (AHRQa). (2013a). Methodology Report # 27, Sample Design of the 2011 Medical Expenditure Panel Survey Insurance Component. Available from: http://meps.ahrq.gov/data_files/publications/mr27/mr27.pdf. Accessed August 22, 2014.
- 21.Agency for Healthcare Research and Quality (AHRQb). (2013b). Medical Expenditure Panel Survey, 2011 Medical Conditions. Available from: http://meps.ahrq.gov/mepsweb/data_stats/download_data/pufs/h146/h146doc.pdf. Accessed August 20, 2014. [PubMed]
- 22.Agency for Healthcare Research and Quality (AHRQc). (2013c). Medical Expenditure Panel Survey. 2011 Full Year Consolidated Data File. Available from: http://meps.ahrq.gov/mepsweb/data_stats/download_data_files.jsp. Accessed August 18, 2014.
- 23.CPI Inflation Calculator. Washington, DC, U.S. Bureau of Labor Statistics. Available from: http://data.bls.gov/cgi-bin/cpicalc.pl. Accessed 15 July 2015. [Google Scholar]
- 24.Cameroon AC, &Trivedi PK. (2010). Microeconometrics using stata. revised edition. College Station, Texas: StataCorp LP. [Google Scholar]
- 25.Manning WG, & Mullahy J (2001). Estimating log models: to transform or not to transform? J Health Econ, 20, 461–494. [DOI] [PubMed] [Google Scholar]
- 26.Belotti F, Deb P, Manning WG, & Norton EC (2015). twopm: two-part models. Stat J, 15(1), 3–20. [Google Scholar]
- 27.Mihaylova B, Briggs A, O’Hhagan, & Thompson SG. (2011). Review of statistical methods for analyzing healthcare resources and costs. Health Econ, 20, 897–916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Azkaya G, & Ercan I (2012). Examining multiple comparison procedures according to error rate, power type and false discovery rate. J Mod Appl Stat Meth, 11, 348–360. [Google Scholar]
- 29.StataCorp. (2015). Stata: Release 14. Statistical Software. College Station: StataCorp LP. [Google Scholar]
- 30.Nwasuruba C, Osuagwu C, Bae S, Singh KP, & Egede LE (2009). Racial differences in diabetes self-management and quality of care in Texas. J Diabetes Complicat, 23, 112–118. [DOI] [PubMed] [Google Scholar]
- 31.Canedo JR, Miller ST, Schlundt D, Fadden MK, & Sanderson M (2017). Racial/ethnic disparities in diabetes quality of care: the role of healthcare access and socioeconomic status. J Racial Ethnic Health Disparities. 10.1007/s40615-016-0335-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Fiscella K, Franks P, Doescher MP, & Saver BG (2001). Disparities in health care by race, ethnicity, and language among the insured: findings from a national sample. Medical Care, 40(1), 52–59. [DOI] [PubMed] [Google Scholar]
- 33.Sambamoorthi U, Shea D, & Crystal S (2003). Total and out-of-pocket expenditures for prescription drugs among older persons. Gerontol, 43(3), 345–359. [DOI] [PubMed] [Google Scholar]
- 34.Campbell JA, Bishu KG, Walker RJ, & Egede LE (2017). Trends of medical expenditures and quality of life in US adults with diabetes: the medical expenditure panel survey, 2002 – 2011. Health Qual Life Outcome, 15(70), 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hu R, Shi L, Pierre G, Zhu J, & Lee D (2015). Diabetes and medical expenditures among non-institutionalized U.S. adults. Diabetes Res Clin Pract 1–12. 10.1016/j.diabres.2015.02.016. [DOI] [PubMed] [Google Scholar]
- 36.Ashkena R, & Abrahamson MJ (2006). Medicare coverage for patients with diabetes. J Gen Intern Med, 21(4), 386–392. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Iuga AO, & McGuire MJ (2014). Adherence and health care costs. Risk Manag Healthc Pol, 7, 35–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Mojtabai R, & Olfson M (2003). Medication costs, adherence and health outcomes among medicare beneficiaries. Health Aff, 22(4), 220–229. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.