Racial/Ethnic and social class differences in preventive care practices among persons with diabetes

Diabetes is the sixth leading cause of death in the United States [1]. It is a major cause of premature mortality and morbidity and poses an enormous economic burden. Direct medical and indirect expenditures attributable to diabetes in 2002 alone were estimated at 132 billion dollars [2, 3]. This disease affects an estimated 18.2 million persons of whom 5.2 million are unaware of having the disease and over 1 million new cases are diagnosed yearly [1, 3]. Heart disease is the leading cause of death among persons with diabetes with mortality rates two to four times higher than those without diabetes [1, 4]. Preventive care practices such as diet, exercise, smoking cessation, eye examinations, foot examinations, and yearly checkups can prevent or delay the incidence and progression of diabetes related complications [4–6]. However, the frequency of preventive care practices is often lower than recommended which results in further progression of disease and mortality [4, 5, 7–9].

The burden of diabetes is greater among minority populations [10, 11]. An estimated 8.4% of Whites, 11.4% of Blacks, and 8.2% of Hispanics 20 years or older have diabetes [1]. The increasing incidence and prevalence of diabetes is seen at disproportionate levels among ethnic minorities. During the period 1980–1994 the age-adjusted prevalence of diabetes increased 33% for Blacks compared to an 11% increase among Whites [10].

Few studies have simultaneously investigated racial/ethnic and social class differences in care practices associated with preventive care behaviors among persons with diabetes. Results from the 1995 and 2001 Behavioral Risk Factor Surveillance System (BRFSS) show variation in preventive-care practices among Whites, Hispanics, and Blacks [4, 5]. In contrast, other studies have found no racial/ethnic differences in preventive care practices for diabetes [8, 12]. Findings in the literature are somewhat inconsistent and limited in their method of exploring factors related to care practices. The present study utilizes recent BRFSS data available to examine racial/ethnic disparities in recommended preventive practices for diabetes, refines past measures of social class with the inclusion of labor force status in addition to educational attainment, and makes use of logistic regression in analyses instead of previously used weighted percentages.

The findings presented in this paper confound past theories and explanations for disparities seen among persons with diabetes. It is well established that ethnic differences in end-stage outcomes for diabetic persons exist. However, our results show Black and Hispanic persons with diabetes have higher levels of preventive care, which is not congruent with observed end-stage diabetes outcomes [18–21]. The question still remains of why disparities in end-stage outcomes persist despite the higher levels of preventive care. One explanation is that Blacks and Hispanics present with more severe disease and a greater number of co-morbidities. Therefore, providers may be more likely to adhere to the recommended care guidelines examined here. Another explanation is that there may be lower adherence to preventive practices not measured here among Black and Hispanic persons. There may also be an economic barrier present in implementing suggested changes after a physicians visit.

Differences in care practices between private and public insurance may also explain results. For example, if it were the case that government sponsored plans facilitated behavioral type preventive care practices (e.g.. eye exam, food exam) easier than private insurers, we may observe the current results. Harris (1999) found that a greater proportion of Whites (91%) and Blacks (89%) had health insurance coverage compared to Hispanics (66%). Of those having health care coverage, a disproportionate number of Blacks and Hispanics had some form of government-sponsored health insurance compared to Whites. Only 56% of Blacks and 45% of Hispanics had private insurance [10]. A recent study conducted in Puerto Rico found that the use of laboratory services like glucose and glycosilated hemoglobin tests was greater in patients who had private insurance [22]. This finding supports the hypothesis that differences in care practices do exist between the private and public insurance sector. This would also support findings from this and other studies that Blacks and Hispanics perform lower levels of glycemic control activities and lab checks in comparison to self-care practices where performance levels are higher compared to Whites [8, 17].

Others have found contrasting results showing that medical/laboratory checks are delivered more frequently than behavioral activities and that Blacks maintain a more favorable laboratory profile [23, 24]. If the hypothesis is true that government sponsored health care promotes self-care activities more than medical/laboratory type practices, and given that a great number of Black persons with diabetes have some form of government health care, then previous findings make sense. Glasgow and Strycker (2000) found self-management activities were recommended less frequently than medical/laboratory checks for persons with diabetes in two private health care systems. This may point towards differences in recommendations and practice between private and government sponsored insurers.

The American College of Emergency Physicians has shown that some 19 and 33 percent of Blacks and Hispanics are uninsured [25]. A study utilizing 1999 BRFSS data also found that more Blacks were uninsured compared to Hispanics and Whites and reported cost as a barrier to visiting a doctor [26]. The cost barrier may further lead to under use of services and essential medication that impact the health of persons with diabetes [27].

The strong patterns observed for social class are not surprising given the demonstrated relationship between social class and chronic illnesses [28–30]. Results in the current study showed a strong association between social class and preventive care practices above the influence of race/ethnicity. This was not too surprising given that race is such a powerful determinant of individuals position within the class structure of society [31]. Lower social class has also been associated with increased diabetes complications [32, 33]. Although we controlled for health plan status in analyses, a residual effect is likely between social class and health insurance since they are so strongly associated and health insurance status differs by level of social class. Persons in higher social class are more likely to be insured compared to persons in lower levels, even though the rising cost of care and employer cutbacks on insurance is creating a growing middle class that is uninsured.

The findings in this paper are subject to a few limitations. First, persons with diabetes may have been missed with the question "Has a doctor ever told you that you have diabetes." Interviewed persons who do not have access to health care services may have answered "no" while they may unknowingly have diabetes. This potential bias may have resulted in obtaining measures of association that are weaker than the true measure since persons who lack health care access are more likely to be diagnosed at more severe stages and not receive preventive care. Second, the social class index used here may not fully capture social class due to the inherent limitations of the two proxy measures used (education and employment status) in capturing such a complex concept. Though there is no agreed upon way to measure social class, education and employment have been used extensively in previous work as proxy measures [34–36]. The combination of the two variables, which was used to create the social class index, has not been validated; however findings using this measure are consistent with previous findings on social class in the literature [34, 36]. Another limitation is that analyses included only non-institutionalized persons, which limit populations to which results can be generalized. Recall bias is another limitation as persons were asked to remember past preventive care practices. The extent to which the bias inherent in this telephone survey affects results is unknown. Percent missing for key variables was as follows: yearly doctor visit (4%), foot examination (10%), eye examination (0.3%), smoking (0.2%), health insurance coverage (0.4%), and employment (0.1%). Subsequently, results may not reflect true measure of association between race/ethnicity and preventive care practices, especially if missing is systematic. A final limitation is the way in which race/ethnicity was determined. Respondents were given a list of responses to choose which may have omitted responses for some race/ethnic groups.

The BRFSS is a telephone survey that does not reach persons without phone service. Persons missed by this survey method are more likely to be poor and minority [37]. Thus, results may be an overestimate of preventive care practices among minorities since high-risk persons may be excluded. Future studies should include these high-risk populations who may have been excluded in the present study.

These results suggestthat recenthealth education and promotion efforts targetedtowardminority populations have been effective.However, serious complications from diabetes still continue to be highest among minorities. Future research should examine theinfluenceof type of insuranceon diabetes patients' access to laboratory preventive care practices. Finally, prevention efforts should be geared toward earlier screening anddiagnosis and better access to preventive and medical services for bothminorities and whites.