Reducing health disparities: medical advice received for minorities with diabetes.
Abstract: Objective: to examine the relationships among reported medical advice, diabetes education, health insurance and health behavior of individuals with diabetes by race/ethnicity and gender.

Method: Secondary analysis of data (N = 654) for adults ages [greater than or equal to] 21 years with diabetes acquired through the National Health and Nutrition Examination Survey (NHANES) for the years 2007-2008 comparing Black, non-Hispanics (BNH) and Mexican-Americans (MA) with White, non-Hispanics (WNH). The NHANES survey design is a stratified, multistage probability sample of the civilian non-institutionalized U.S. population. Sample weights were applied in accordance with NHANES specifications using the complex sample module of IBM SPSS version 18.

Results: The findings revealed statistical significant differences in reported medical advice given. BNH [OR = 1.83 (1.16, 2.88), p = 0.013] were more likely than WNH to report being told to reduce fat or calories. Similarly, BNH [OR = 2.84 (1.45, 5.59), p = 0.005] were more likely than WNH to report that they were told to increase their physical activity. Mexican-Americans were less likely to self-monitor their blood glucose than WNH [OR = 2.70 (1.66, 4.38), p < 0.001]. There were differences by race/ethnicity for reporting receiving recent diabetes education. Black, non-Hispanics were twice as likely to report receiving diabetes education than WNH [OR = 2.29 (1.36, 3.85), p = 0.004]. Having recent diabetes education increased the likelihood of performing several diabetes self-management behaviors independent of race.

Conclusions: There were significant differences in reported medical advice received for diabetes care by race/ethnicity. The results suggest ethnic variations in patient-provider communication and may be a consequence of their health beliefs, patient-provider communication as well as length of visit and access to healthcare. These findings clearly demonstrate the need for government sponsored programs, with a patient-centered approach, augmenting usual medical care for diabetes. Moreover, the results suggest that public policy is needed to require the provision of diabetes education at least every two years by public health insurance programs and recommend this provision for all private insurance companies.

Key words: Medical advice, health behavior, minorities, health insurance
Article Type: Report
Subject: Health care disparities (Research)
Medical advice systems (Management)
Minorities (Care and treatment)
Diabetics (Health aspects)
Authors: Vaccaro, Joan A.
Huffman, Fatma G.
Pub Date: 03/22/2012
Publication: Name: Journal of Health and Human Services Administration Publisher: Southern Public Administration Education Foundation, Inc. Audience: Academic Format: Magazine/Journal Subject: Government; Health Copyright: COPYRIGHT 2012 Southern Public Administration Education Foundation, Inc. ISSN: 1079-3739
Issue: Date: Spring, 2012 Source Volume: 34 Source Issue: 4
Topic: Event Code: 310 Science & research; 200 Management dynamics Computer Subject: Company business management
Product: SIC Code: 7389 Business services, not elsewhere classified
Geographic: Geographic Scope: United States Geographic Code: 1USA United States
Accession Number: 304050536

Diabetes leads to complications such as heart disease and stroke, high blood pressure, blindness, kidney disease and nervous system disease; the risk of death for persons with diabetes is twice that of persons without diabetes (Center for Disease Control and Prevention (CDC), 2007). Type 2 diabetes, the most common form (90-95% of all cases) has increased among the general population (National Institute of Diabetes and Digestive Kidney Diseases (NIDDK), 2008) and disproportionately among minorities (particularly African Americans and Hispanics) (National Diabetes Surveillance System (NDSS), 2005). Mexican Americans have the highest rate of diabetes among Hispanics and are 1.7 times as likely to have diabetes as non-Hispanic Whites (CDC, 2007). African Americans are 2.1 times more likely to be diagnosed with diabetes than non-Hispanic Whites (CDC, 2007).

Minorities tend to have less access to and receive a lower quality of health care, even when controlling for insurance status and income (American College of Physicians - American Society of Internal Medicine (ACP), 2000). Adjusting for socioeconomic status reduces the effects of race and ethnicity, but the effects are still apparent in that the lack of cultural competency and appropriate communication skills by health care providers can have many negative health outcomes including micro-and macro-vascular complications (ACP, 2000). It is essential for persons with diabetes to acquire and practice adequate diabetes self-management skills in order to reduce the risk factors that lead to morbidity and mortality associated with diabetes-related complications.

An operational definition of high quality health care for persons with diabetes would include guidance on risk factor control for all of the following: 1) dietary intake and weight management; 2) glycemic and lipid control; and 3) foot and eye care (ADA, 2010). Persons without healthcare coverage may not be receiving the advice to develop the skills necessary to manage their diabetes. Moreover, there may be a discrepancy in the quality of health care received by race and ethnicity or by health insurance coverage. Several national studies have found differences in quality of care within health insurance types.

Coverage analysis has focused on clinical indicators up until publication of the Health Care Coverage Analyses of the 2006 National Healthcare Quality and Disparities Reports (AHRQ, 2008a). Prior to which, the effectiveness of coverage was based on a secondary indicator: clinical health outcomes rather than the quality of care standards by different types of insurance (private or public). For example, the AHRQ reported that a standard of care for obese adults is medical advice to exercise and that Hispanic obese adults were less likely to be given advice to exercise than White obese adults across insurance groups (AHRQ, 2008a). They found Black obese adults who were publicly insured or without insurance were less likely to be given advice to exercise as compared to White obese adults (AHRQ, 2008a).

The National Healthcare Quality and Disparities Reports (2006) confirmed that individuals without insurance had poorer health outcomes and, that publicly insured persons with diabetes were less likely to have their A1C tested and to have the recommended diabetes services (all three examinations: A1C, eye and foot) than those with private insurance (AHQR, 2008a). Compared to privately, publicly insured adults were more likely to report the following difficulties: access to care (limited hours of operation); referrals to specialists; and, timeliness of care for an illness or injury (AHQR, 2008a).

Patient quality of care and insurance status was analyzed for diabetes patients (N = 2018) from 27 federally-funded health centers using information for diabetes processes of care and health outcomes, available from the National Committee for Quality Assurance (NCQA) Health Plan Employer Data and Information Set (HEDIS) (Zhang et al, 2009). The authors found that the quality of diabetes care was worse for persons with Medicaid than those with private insurance, and that the quality of care for those on Medicaid was similar to those with no insurance (Zhang et al, 2009). The investigators suggested cost variables between insurance providers, albeit small differences (co-payments or deductibles) may mediate health service utilization for lower-income individuals (Zhang et al, 2009).

The National Medical Expenditures Panel Survey (MEPS) was conducted to assess the quality of care domains considered important by the Institute of Medicine (IOM): safe, equitable, effective, patient-centered, timely/accessible, efficient care (Ng & Sholle, 2010). Conversely, Nelson et al (2005) found private, Medicaid and Medicare patients had little differences in quality of services; albeit, these findings differ from the majority of studies.

Medical advice reported received may differ by race and ethnicity as a consequence of the communication process. The relationships among medical advice, diabetes self-management, health outcomes by ethnicity and race have not been adequately reported in the literature. Diabetes is a public health problem requiring a multilevel systems approach for prevention and treatment (Glasgow, Wagner, Kaplan, Vinicor & Norman, 1999). The population-based approach advocated by Glasgow et al (1990) includes personal, family, health care team, and community influences that impact the promotion or inhibition of diabetes self-management and lifestyle changes (Glasgow et al, 1999). A key factor, interwoven through each system, is communication. Although medical professionals have established guidelines for effective communication, the complex dynamics of interpersonal relationships make desired outcomes and assessment of the patient- provider communication challenging. For example, the treatment plan for a patient with type 2 diabetes includes an interview that has a standard protocol. Even when the message was intended to be 'culturally sensitive and collaborative,' the communication may not have been received in the manner it was intended by the provider. Moreover, the effectiveness of patient-provider communication has been assessed through health behavior and outcomes as opposed to direct feedback by the majority of health research.

Patient-provider communication is confounded by the health beliefs and values of the dyad. The manner in which the organization and healthcare provider approach the patient depends on individual and organizational cultural competency. According to Cross, Bazron, Dennis & Isaacs (1989), cultural competency is an evolving process of awareness and skills that incorporates values, principles, behaviors, attitudes and policies of working effectively across cultures. As such, measurement of cultural competency is complex and is a factor of variance among the study population. In order for a system to become more culturally competent, Cross et al (1989) identifies the following five elements: value diversity; cultural self-assessment; consciousness of the dynamics of cultural knowledge; and, development of adaptations to diversity. Cross et al (1989) further stated that attitudes, policies and practices are areas that need to be targeted in the movement toward cultural competency.

Cultural competency has also been referred to as cultural sensitivity. Cultural linguistic competency, a narrower type of cultural competency, specifies only the ability to communicate in the client's language either by being bilingual or having a certified interpreter participate in the communication process. The Office of Minority Health has developed 14 national standards on culturally and linguistically appropriate services (CLAS) that are mandated for government agencies at the federal, state and county levels and suggested for use in all health care organizations (OMH, 2007). Even though these standards are specifications for an operational definition of cultural and linguistically cultural competency, definition of the term varies among health care organizations. As applied to the patient- provider relationship, cultural competency is a subjective indicator of the degree to which the provider can interact successfully irrespective of race and/or ethnicity of the pair. While cultural sensitivity might be considered the intention of the provider, cultural competence is the measureable outcomes of patients' satisfaction and their rating of the effectiveness of the advice or education given.

Specifically, the linguistic competency of the provider may influence whether or not the patient receives the intended message. Goode & Jones (2009) developed and revised a definition for linguistic competency that has been widely used in health care and other human service delivery systems. Communication is considered to be linguistically competent if it is delivered effectively to meet the needs of the populations served and is easily understood by such persons (Goode & Jones, 2009).

Understandings of diabetes as a disease and diabetes self-management are influenced by health beliefs (Anderson & Christison-Lagay, 2008). In turn, health beliefs and practices vary by cultural differences, ethnicity and race (Anderson & Christison-Lagay, 2008). It is therefore imperative to uncover the interrelationships of provider-patient communication; ethnicity and race; and diabetes self-management beliefs and practices. Given the newly available national data for medical advice in NHANES 2007-2008, the objective of this study was to compare health care disparities regarding reported medical advice and services necessary for diabetes self-management received from healthcare providers for two minority groups at high risk for diabetes complications: Black, nonHispanics and Mexican-Americans as compared to nonHispanic, Whites.


A secondary analysis was conducted using from the National Health and Nutrition Examination Survey (NHANES) 2007-2008 database comparing race/ethnicity, reported medical advice, diabetes self-management skills and diabetes-related health outcomes. The sample population included male and female adults that were selected from NHANES 2007-2008 database for whom detailed interviews and examinations were available and met the following conditions: adults [greater than or equal to] 21 years and reporting a diagnosis of diabetes and of the following ethnicities: Black, non-Hispanic (BNH); Mexican American (MA) and White, non-Hispanic (WNH).

Sample Size Estimation

Since the main outcome variables, medical advice received and treatment behaviors have not been previously tested by NHANES, the power analysis was based on several clinically important outcomes: fasted blood glucose (FBG), systolic blood pressure (SBP), diastolic blood pressure (DBP), triglycerides (TG), and low-density lipoprotein cholesterol (LDL) as well as a review of the literature. Since these outcomes are paired by ethnicity and continuous, we calculated sample size based on the t-test. Furthermore, our power analysis is based on a two-tailed alpha of 0.05 (95% confidence) and beta of 0.20 (80% power) for each variable. Meta-analyses of short-term dietary interventions by the American Diabetes Association (2007) reported reductions of 15-25 mg of LDL-C and considers this reduction range to be a clinical target for lifestyle interventions. Applying this target (15-25 mg range) for a power analysis, a modest standardized effect size of 0.45 yielded a sample size of 80 per group (Hulley & Cummings, 1988).

Next, we performed a power analysis using FBG. A desired effect of 1 mmol/L or 9 mg/dL was chosen, based on the outcome evaluation of the CANOE trial (Zinman et al, 2006). A standard deviation of FBG for persons with diabetes was found to be 66 with a mean of 146 from an analysis of data collected in our laboratory from Cuban American subjects. Back calculation of a standard effect size of 0.45 yielded an estimated SD of 20 and a sample size of 80 (Hulley & Cummings, 1988). A clinical change of 5 mg/dL in FPG would correspond to a standardized effect size of 0.40 and would require 98 participants per group. Three ethnic groups were compared, so approximately 300 participants would be required to achieve statistical power considering a design effect of 1.0. The design effect (DEFF) is an estimate the variance of a complex sample with respect to that of a simple random sample (DEFF = variance estimate (cluster)/variance estimate (simple random sampling). According to the National Center for Health Statistics (NCHS, 2010) it is difficult to set a single minimum sample size for analysis since DEFF are generally greater than 1.0 for NHANES and differ for each variable, race/ethnicity and age group. As such, for this study, a number of strategies were applied to determine the adequacy of sample size for each analysis.

The following conditions were required for sample size adequacy: 1) A model classification of at least 60%. 2) Category frequencies of at least 30. 3) Odds ratio of at least 1.5. The later cut-off for the OR was chosen, based on preliminary investigation of the design effect range for race explaining medical advice (DEFF = 0.8 - 1.5).

Data Collection

Raw data were extracted from datasets collected from the National Health and Nutrition Examination Survey, 2007-2008 (NHANES 2007-2008) available for public use. More detailed information is available at the NHANES website: This survey contains data for 10,149 individuals of all ages. Data were collected between January 2007 and December 2008. A limited data set from the survey interview and examination is available to the public with the corresponding codebooks.

All NHANES research is generated under the auspices of The National Center for Health Statistics (NCHS), Division of Health and Nutrition Examination Surveys (DHNES), part of the Centers for Disease Control and Prevention (CDC). Since the early 1960's, NHANES were conducted and starting from 1971 to 1994, the surveys were periodically administered. Beginning in 1999 the survey has been conducted continuously. Questions from the NHANES 2007-2008 were taken from previous versions of NHANES with additional questions added based on public feedback.

The NHANES survey design is a stratified, multistage probability sample of the civilian non-institutionalized U.S. population. The stages of sample selection are as follows: 1) Primary Sampling Units (PSUs), which are counties or small groups of contiguous counties; 2) segments within PSUs (a block or group of blocks containing a cluster of households); 3) households within segments; and 4) one or more participants within households. A total of 15 PSUs were visited during a 12-month time period.

The procedure for the household interviews and health examinations are briefly described in the next several paragraphs. First a letter was sent to all selected households to inform respondents that a trained interviewer would visit their home. When the interviewer arrives at the home, identification was shown and the objectives of the survey were explained.

For the household interview, participants were those who understood, agreed to and signed an Interview Consent for the household interview portion of the survey. After the household interview was completed, all interviewed persons were asked to complete the health examination component. Those who agreed to participate were asked to sign additional consent forms for the NHANES health examination component. The interviewer telephoned the NHANES field office from the participant's home to schedule an appointment for the examination and informed the participants that they would receive remuneration as well as reimbursement for transportation and childcare expenses, if necessary. The health examinations were conducted in mobile examination centers (MECs); the MECs provide a standardized environment for the collection of high quality data.

Data Analysis

To achieve their target population, NHANES 2007-2008 oversampled, Mexican Americans, all Hispanics, Blacks, persons 60 years or older and all persons of lower income. Sample weights were constructed and included in the data sets to account for complex sample design and achieve unbiased national estimates. The principle need for sample weights in complex designs is to compensate for unequal probabilities of selection, account for nonresponse, and make sample weights conform to a known population distribution. The base sample weights for interview and MEC are the probability of selection at each stage. The choice of sample weight needs to be based on data file with the smallest sample size (NCHS, 2006). The choice of sample weight for this study was the MEC sample weight: WTMEC2YR. This is because hypotheses for full models included variables with laboratory or anthropometrics. These measurements were taken for a smaller number of participants. The choice of sample weight was based on the data file with the smallest sample size as recommended by the NHANES guidelines. In addition to the base sample weight, the design information for the complex sampling plan included mask variances incorporated into strata (sdmvstra) and primary sampling units (sdmvupsu). Together, the design accounted for unequal probability of selection and reduced the chance of type 1 error (NHANES, 2006; Stiller & Tompkins, 2005). The statistical program used the Taylor series linearization for estimating population characteristics (Stiller & Tompkin, 2005). The sample plan handled the multistage design as a single stage design with replacement.

Data analysis was conducted with IBM-SPSS version 18 with a complex sampling add-on, where p < 0.05 was considered significant. Continuous variables were analyzed for normality by Q-Q plots and when needed, transformed. Participants' characteristics were presented by frequency and percent. Logistic regression was used to determine likelihood of health disparities. Hierarchical logistic regression models were conducted for medical advice by race predicting health behaviors adding variables associated by the literature as covariates. The best model was tested by subtraction of the log-likelihood [chi square] between the full and reduced models.


The general characteristics of the sample are presented in Table 1. For the study participants diagnosed with diabetes, there was no significant difference in reporting having health care between BNH and WNH and over 90% reported having a health care plan. Conversely, approximately 38% of MA diagnosed with diabetes reported not having health coverage. In addition, over 40% MA responded they did not know how many times per year they saw their doctor; less than 25% of BNH and WNH reported they did not recall the number of doctor's visits over the past year. Of those who reported values, there were no significant differences among participants by race; however, those reporting specific frequencies may not be representative of their group. As such, access to health care may differ between participants, in particular, MA and WNH.

There was an overlap of medical insurance type among the participants; in particular, a percentage of those with Medicare where either enrolled in Medicaid (dual eligible) or in a private plan. Considering the varying and ever-changing policies of private insurance, along with reimbursement changes in government-sponsored programs it is not possible to compare insurance type and quality of care with adequate power for this study. As such, the percent of persons belonging to each group were given in Table 1; however comparison were valid for any type of insurance versus no insurance, only.

Table 2 provides odds ratio for medical advice, goals, and diabetes education by race/ethnicity. Controlling for obesity, MA and BNH were approximately twice more likely to report being told to reduce fats or calories and to increase their physical activity or exercise by their medical doctor than WNH. Controlling for education and health insurance, BNH were more than twice as likely to report receiving diabetes education within the past two years than WNH, controlling for age (Table 2), and when controlling for medical insurance [OR = 2.04 (2.26, 3.30), p = 0.006]. There were no differences by gender for medical advice; however, there was a 3-way interaction: race, gender, and reporting receiving a goal for blood pressure associated with systolic blood pressure.

The responses for reporting being given a goal for blood pressure were among the following: a numerical value, which was coded as 'yes'; no goal was given by a healthcare provider, which was coded as 'no'; and, not sure. Since there were not any significant differences between the categories 'not sure' and no', the two categories were collapsed. Using the general linear model, we assessed the association of the following independent variables: race, gender, age, and reporting receiving a goal for systolic blood pressure with the dependent variable measured systolic blood pressure. Race, gender and receiving a goal were not independent. Therefore, full factorial models were conducted to examine the 3-way interaction of race*gender*goal (age as a covariate) with systolic blood pressure as the independent variable. Both the model and the 3-way interaction were significant [F (12, 5) = 6.57, p = 0.025, model; F (2, 15) = 5.38, p = 0.017, interaction]. Age was a significant predictor of SBP [F (1, 16) = 64.1, p < 0.001]. The 3-way interactions were plotted separately by gender (Figures 1, 2).



Effective diabetes self-management behavior has been associated with ongoing and recent diabetes education throughout the literature and by recommendation of the American Diabetes Association (2010). We found individuals with recent diabetes education (< 2 years) were more likely to perform diabetes self-management behaviors than those with less recent or no diabetes education. Individuals reporting they were advised to reduce fat or calories were more likely to report receiving recent diabetes education [OR = 1.71 (1.20, 2.45), p = 0.006] and, the relationship was significant controlling for race [OR = 1.66 (1.13, 2.24), p = 0.012]. Participants reporting being told to increase physical activity or exercise were more likely to have recent diabetes education [OR = 1.73 (1.01, 2.95), p = 0.045]. Race, gender and health insurance were not significantly associated with reporting any of the following behaviors: reducing fat or calories; controlling or losing weight; and, increasing physical activity or exercise.


Due to the many health consequences of diabetes and the nature of the disease, diabetes care is vital to quality of life and survival. Interestingly, diabetes is a disease that can be managed by the individual with appropriate guidance. Yet fewer than 60% of all adults age 40 and over with diagnosed diabetes have their blood glucose, cholesterol, or blood pressure within the recommended levels for adequate control (Agency for Healthcare Research and Quality (AHRQ), 2008b).

Although diabetes care is largely the responsibility of the individual, medical staff plays a vital role in the patient's skill development. In fact, health care providers are the link between the patient and their disease self-management. The communication process between the provider and patient can determine whether or not the patient is informed, motivated and confident enough to make the behavioral changes necessary for diabetes care. Moreover, reduced length of visit and competing demands during visits (such as acute illness), may be barriers to receiving recommended diabetes services (Parchman, Romero & Pugh, 2006).

We found differences among race and among race by gender regarding reported medical advice given. Receiving a goal for blood pressure varied for race by gender. Black, non-Hispanics were twice as likely to report receiving diabetes education in the past two years as WNH. Mexican-Americans were less likely to self-monitor their blood glucose as compared to White, non-Hispanics and the differences were not significant by gender. Our results corroborate the findings in the literature that certain health practices and quality of care differ by race and gender.

Gender differences were found in certain areas of healthcare by a study contracted by the Agency for Healthcare Research and Quality (AHQR); women were found to be more likely to report problems with access to healthcare; however, the effectiveness of care and patient-centeredness measures did not differ by gender (Ng & Scholle, 2010).

When ethnic or racial groups receive, on average, unequal health care or have an imbalance in access to health care, they are considered to have 'health disparities'. The Office of Minority Health (2005) defines health disparities as significant differences between one population and another. The Department of Health and Human Services launched a series of initiatives to eliminate health disparities through the Minority Health and Health Disparities Research and Education Act of 2000 (Office of Minority Health, 2005). The National Diabetes Education Program (NDEP), a governmental and private public health partnership program, was formed in an effort to eliminate the diabetes epidemic by forming programs specifically for African Americans, Hispanics, Native Americans, Alaskan natives, Asian Americans and Pacific Islanders (NDEP, 2007). BNH and WNH had no significant difference in health care coverage (Table 1). In an effort to comply with governmental programs and cut costs, it is possible that health care providers are selecting more BNH to receive diabetes education than WNH, while maintaining or decreasing the numbers of persons sent for diabetes education per year. It is likely that in an effort to eliminate health disparities, inadvertently, another form of health care inequality was formed for persons having health care coverage since the difference remained controlling for health care insurance type. The differences between MA and WNH for diabetes education cannot be assessed for this study since MA were less likely to be covered by health insurance.

Weight management, an important aspect of diabetes self-management, is achieved by reducing fat or calories and increasing physical activity; performing these skills are central among the recommendations for persons with type 2 diabetes by the American Diabetes Association (ADA, 2010). For our sample, there were differences by race/ethnicity in medical advice reported in these effective means of weight management (calorie or fat reduction and physical activity). Controlling for obesity, MA and BNH with type 2 diabetes were more likely to report being told by their doctor to reduce calories or fat in the past year than WNH; and, they were more likely to report being told to increase their physical activity by a doctor in the past year. These findings are contrary to the study hypothesis which predicted that MA and BNH would be less likely to report being given medical advice than WNH. The original hypothesis was based on two assumptions: 1) There are health disparities with access to quality medical care; and 2) The health beliefs of Latinos and non-Hispanic, Blacks might be a factor associated with filtering the provider's advice regarding lifestyle changes.

Diabetes self-management behavior differed by race/ethnicity. Participants from the two minority groups (MA and BNH) were more likely to report engaging in healthy diabetes self-management behaviors such as reducing fat or calories and BNH were more likely to report checking their feet for sores than WNH. BNH and MA were more likely to report being advised to reduce fat or calories and increase physical activity than WNH. In direct contrast to our findings, Oster et al (2006) reported Blacks (n = 984) and Hispanics (n = 428) with diabetes were less likely to monitor their diet than Whites (n = 4623) from a national managed care organization. In addition, they found Blacks were less likely than Whites to exercise.

Patient-centered quality of life and self-management behaviors are considered benchmarks of effective medical advice (Glasgow, Peeples & Skovlund, 2008). Component of quality care for diabetes can be assessed by self-management actions needed to control diabetes (healthy eating, regular physical activity, not smoking, etc); self-evaluation of quality of life (diabetes-specific and general health status); collaborative, specific goals for diabetes self-care; and, patient engagement in decision making and care (Glasgow, Peeples & Skovlund, 2008). As such, the Chronic Care Model (CCM), the Association of Diabetes Educators (AADE), the American Diabetes Association (ADA), and the Institute of Medicine (IOM) advocate that patient-centered measurements be added to the Provider Recognition Performance Measures collected by the National Committee for Quality Assurance (NCQA) (Glasgow, Peeples & Skovlund, 2008).

There are several limitations of this study. First, communication between patient and provider with respect to race/ethnicity was confounded by multiple relationships including the following: cultural/linguistic competency and continuity/frequency of care on the part of the healthcare provider; health behavior knowledge, self-efficacy, motivation and values of the patient; and, the health beliefs of the dyad. Second, although the sample was drawn from a national survey, there were a limited number of MA in certain health care categories (< 30). Hence, lack of significance may be due to limited power rather than a true lack of association. Finally, associations cannot be considered causal, since the data were a single-time point measure. Since there may be differences in quality of care by health insurance type and this may influence health behaviors and health outcomes, health insurance was considered a covariate for all analyses. For this study, health insurance was not significantly associated with medical advice or health behaviors; however, variations in health insurance may have been a confounder. Despite the limitations, a major strength of this study was the use of a national database (NHANES), which has specialized in collecting health data by race/ethnicity. Since this was the first year that NHANES included data concerning medical advice for diabetes self-management behaviors this study was one of the first to use a national database to assess health disparities of reported medical recommendations.


There were differences by race/ethnicity for reporting medical advice given. Black, non-Hispanics and MA were more likely to report receiving advice from their healthcare provider to reduce fat or calories and to increase physical activity. Black, non-Hispanics were twice as likely to report receiving diabetes education than WNH controlling for health insurance. We found race/ethnicity and race/ethnicity by gender differences in several diabetes care behaviors. On the other hand, other diabetes self-management behaviors depended on medical treatment received. Performing diabetes self management behaviors: reducing fat or calories and increasing physical activity were more likely if diabetes education was given within 2 years, independent of race/ethnicity. These findings clearly demonstrate the need for government sponsored programs, with a patient-centered approach, augmenting usual medical care for diabetes. Moreover, the results indicate that public policy is needed to require the provision of diabetes education at least every two years by public health insurance programs and recommend this provision for all private insurance companies. Future studies assessing the relationships among diabetes care, race/ethnicity and gender with respect to patient satisfaction, frequency and access to healthcare are recommended to determine the direction of policy change for persons with diabetes.


Agency for Healthcare Research and Quality (AHRQ) (2008a). Health care coverage analyses of the 2006 National Health Quality and Disparities Reports. Baltimore, MD: U.S. Department of Health and Human Services, Centers for Medicare & Medicaid Services; December, 2008.

Agency for Healthcare Research and Quality (AHRQ) (2008b). National Healthcare Quality Report, 2008. Chapter 2: Effectiveness, Diabetes. AHRQ Publication No. 09-000. Current as of March 2010.

American College of Physicians - American Society of Internal Medicine (ACP) (2000): No Health Insurance? It's enough to make you sick. Latino community at great risk. White Paper, Philadelphia, PA, ACP-ASIM.

American Diabetes Association (2007). Evidence-based nutrition principles and recommendations for the treatment and prevention of diabetes and related complications. Diabetes Care, 30(Suppl. 1), S48S65.

American Diabetes Association (ADA) (2010). Standards of medical care in diabetes - 2010. Diabetes Care, 33 (Suppl. 1), S11-S61.

Anderson, D. & Christison-Lagay, J. (2008). Diabetes self-management in a community health center: Improving health behaviors and clinical outcomes for underserved patients. Clinical Diabetes, 26(1), 22-27.

Centers for Disease Control and Prevention (CDC, 2007). National diabetes fact sheet: General information and national estimates on diabetes in the United States, 2007. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention.

Cross, T. L., Bazron, B. J., Dennis, K. W. & Isaacs, M. R. (1989). Towards a culturally competent system of care: A monograph on effective services for minority children who are severely emotionally disturbed. CASSP Technical Assistance Center, Georgetown University Child Development Center, 3800 Reservoir Road, N.W., Washington, DC 20007.

Glasgow, R. E, Wagner, E. H., Kaplan, R. M., Vinicor, F. Smith, L. & Norman, J. (1999). If diabetes is a public health problem, why not treat it as one? A population-based approach to chronic illness. Annals of Behavioral Medicine, 21(2), 159-170.

Glasgow, R. D., Peeples, M. & Skovlund, S. E. (2008). Where is the patient in diabetes performance measures? Diabetes Care, 31(5), 1046-1050.

Goode, T. and Jones, W. (2009) Definition of Linguistic Competence. National Center for Cultural Competence. Georgetown University Center for Child and Human Development. Box 571485; Washington, DC, 20057-1485. Website:

Hulley, S.B. & Cummings, S.R. (1988) Designing Clinical Research: An Epidemiological Approach. Baltimore: Williams & Wilkins. pp.140-141,

Appendix 13A, 215

National Center for Health Statistics, Center for Disease Control and Prevention. National Health and Nutrition Examination Survey (NHANES) (2006). Analytical and reporting guidelines. Hyattsville, Maryland. Retrieved October 5, 2010. From:

National Center for Health Statistics, Center for Disease Control and Prevention. Continuous NHANES web tutorial: Variance estimation. Retrieved October 12, 2010. From Info1.htm.

National Diabetes Education Program (NDEP) (2007). National diabetes education strategy plan (20082010). Retrieved September 14, 2010 from

National Diabetes Surveillance System (NDSS) (2005). Age-adjusted Prevalence of Diagnosed Diabetes per 100 Population (2005) Online Table. National

Diabetes Surveillance System. Retrieved October 5, 2010 from:

National Institute of Diabetes and Digestive Kidney Diseases (NIDDK). (2008). National Diabetes Statistics, 2007 fact sheet. Bethesda, MD: US Department of Health and Human Services, National Institutes of Health.

Ng, J. & Scholle, S. H. (2010). Disparities in quality of care for midlife adults (ages 45-64) versus older adults (ages > 65). Prepared under contract from the Agency for Healthcare Research and Quality and the Office of the Assistant Secretary for Planning and Evaluation. National Committee for Quality Assurance. Washington, D.C. May, 2010. Office of Minority Health (OMH) (2005). What are health disparities? Retrieved September 15, 2010 from Office of Minority Health (2007). National Standards on Culturally and Linguistically Appropriate Services (CLAS). Retrieved September 24, 2011 from

Oster, N. V., Welch, V. Schild, L. Gazmararian, J. A. & Rask, K. (2006). Differences in self-management behavior and use of preventive services among diabetes management enrollees by race and ethnicity. Disease Management, 9(3), 167-175.

Parchman, M. L., Romero, R. L. & Pugh, J. A. (2006). Encounters by patients with type 2 diabetes-- complex and demanding: An observational study. Annals of Family Medicine, 4(1), 40-45.

Stiller. A. B. & Tompkins, L. (2005). The big four: Analyzing complex sample survey data using SAS[R], SPSS[R], STATA[R] and SUDAAN[R]. Proceedings from the Northeastern Users Group, Inc. Presenters are from the Centers of Disease Control, Division of the National Center for Health Statistics. Retrieved October 5, 2010. From:

Zang, J.X., Huang, E. S., Drum, M. L., Kirchhoff, A. C., Schlichting, J.A., Schaefer, C. T., Heuer, L.J. & Chin, M.H. (2009). Insurance status and quality of diabetes care in community health centers. American Journal of Public Health, 99(4), 742-747.

Zinman, B., Harris, S. B., Gerstein, H., Young, K., Raboud, J. M., Neuman, J. & Hanley, A. J. G. (2006). Preventing type 2 diabetes using combination therapy: design and methods of the Canadian Normoglycaemia Outcomes Evaluation (CANOE) Trial. Diabetes, Obesity and Metabolism, 8(5), 531537.



Florida International University
Table 1
General characteristics of the study population

Variable (b)      MA           BNH          WNH          P Total

Age (years)       56.2[+       57.6[+       60.7[+
                  or -] 1.95   or -] 0.89   or -] 0.65

P-values (c)      0.019        0.012        -            0.002

Gender                                                   0.127

Male              54(48.1)     96(398)      160(50.2)    -

Female            67(51.9)     117(60.2)    128(49.8)    -

Years with        9.68[+       11.6[+       11.6[+
Diabetes          or -] 0.85   or -] 0.67   or -] 0.68

P-values          0.127        0.989        -            0.242

Education                                                <0.001
[greater than     56(41.5)     25(9.2)      37(10.1)     -
or equal to]
8th grade

>8th 75,000           20(15.3)     28(16.1)     38(22.4)     -

Refused           5(3.4)       2(0.6)       4(2.2)       -

Don't know        6(3.9)       5(2.9)       2(0.6)       -

Insurance (d)     MA           BNH          WNH

Yes               82(62.0)     192(85.2)    262(93.1)    < 0.001

Medicare          43(25.4)     96(35.8)     164(47.7)    -

Medicaid          16(11.0)     36(16.4)     25(6.9)      -

Private (e)       39(33.0)     105(47.8)    175(71.3)    -

Doctor's visits   4.63[+       4.44[+       4.17[+       0.549
Per year (f)      or -] 0.68   or -] 0.33   or -] 0.28

Don't Know        56(41.8)     37(16.0)     68(22.2)     < 0.001

Abbreviations: MA=Mexican American; BNH=Black, non-Hispanic; WNH=
White, non-Hispanic (comparison group).

(a) unweighted cases: MA n=131; BNH n=223; WNH n=300. There were
missing responses for income (n=614), education (n=583), years with
diabetes (n=644). Gender and health insurance was weighted for N=622
cases based on MEC (mobile examination center) participants.
(b) Continuous variables are given as (mean [+ or -]SE) were tested
by an ANOVA and categorical variables are given as N (%)
(unweighted count) and were tested by Pearson's
[chi square].
(c) Parital p-values comparing MA and BNH to WNH
(d) Totals may be > 100% since respondents may have private and
Medicare. The p-values are for the log-likelihood [chi square]
of the unadjusted odds ratios. The adjusted likelihood of no
insurance: O[R.sub.MA/WNH]  = 5.73 (2.17, 15.1),p < 0.001;
O[R.sub.BNH/WNH] = 1.90 (0.77, 4.70), p = 0.151, (controlling for age
and education). (e) Private or through company benefits were combined.
(f) Valid cases N = 498, were those who answered a numerical value.
Those who reported that they did not know were tallied as a binary,
don't know versus reported.

Table 2
Medical advice, goals and diabetes education by race

Dependent       Independent
Variable        Variables (a)     OR (95%CI)

Medical Advice, goals
and diabetes education            MA/WNH         BNH/WNH

Told fat/cal    -Race             2.15           1.83 *
                -Obesity          (1.03, 4.46)   (1.16, 2.88)
                ([greater than    p = 0.042      p = 0.013
                or equal to]

Told PA         -Race             2.45           2.84 *
                -Obesity          (1.08, 5.57)   (1.45, 5.58)
                ([greater than    p = 0.034      p = 0.005
                or equal to]

Told Wt         -Race             1.86           1.29
                -Education        (0.72, 4.85)   (0.88, 1.89)
                -Obesity          p = 0.187      p = 0.169
                ([greater than
                or equal to]

Given goal      None              1.00           0.82
A1C                               (0.68, 1.48)   (0.47, 1.42)
(yes)                             p = 0.981      p = 0.444

Given goal      None              2.14           0.99
LDL                               (1.37,3.35)    (0.64,1.54)
(yes)                             p = 0.011      p = 0.972

Given goal      Education         1.78           0.96
LDL                               (1.21, 2.63)   (0.64, 1.46)
(yes)                             p = 0.002      p = 0.972

Diabetes        Age(yrs)          0.75           2.29 *
Education (c)                     (0.40, 1.44)   (1.36, 3.85)
(yes)                             p = 0.366      p = 0.004

Abbreviations: Told fat/cal = reported yes to being told by
healthcare provider to reduce fat or calories; told PA= reported
yes to being told by healthcare provider to increase physical
activity or exercise; told wt = reported yes to being told by a
healthcare provider to control or lose weight; goal A1C = What
does your doctor or other health professional say your "A one C"
level should be? Goal LDL = What does your doctor or other health
professional say your LDL cholesterol should be?

Notes: All models were significant. The 2-degree of freedom
variables for race were significant after the Bonferroni
correction * (p < 0.017).

(a) Control variables for the reduced model.

(b) Represents frequency reporting being given recent diabetes
education (within two years).

Table 3
Diabetes self-management behaviors by race

Dependent    Controls (a)   OR(95% CI) (b)
Behavior                    MA/WNH            BNH/WNH

SMBG         none           3.82 *            1.63
no                          (2.16, 6.76)      (0.81,3.29)
                            p<0.001           p=0.156

SMBG         -health-       2.70 *            1.89
no           insurance      (1.66,4.38)       (1.02,3.49)
             -diabetes      p=0.001           p=0.044

Reduce       -diabetes      2.33              1.28
fat/cal      education      (1.05,5.14)       (0.073,2.26)
yes          -obesity       p=0.038           p=0.367

Increase     -age           1.18              1.1.0
PA           -diabetes      (0.615,2.25)      (0.65,1.85)
yes          education      p=0.603           p=0.718

Reduce Wt    -education     1.41              0.82
yes          -diabetes      (0.082,2.42)      (0.46,1.46)
             education      p=0.202           p=0.474

Check Feet   -gender        0.94              2.40 *
                            0.60,1.47)        (1.67,3.45)
                            p=0.771           p<0.001

Abbreviations: Told fat/cal = reported yes to being told by
healthcare provider to reduce fat or calories; told PA= reported
yes to being told by healthcare provider to increase physical
activity or exercise; told wt = reported yes to being told by a
healthcare  provider tocontrol or lose weight.

(a) Control variables indicated are for the reduced model.
(b) All models were significant after Bonferroni correction at
p < 0.017. The 2-degree of freedom test for race was significant
where indicated by (*).
Gale Copyright: Copyright 2012 Gale, Cengage Learning. All rights reserved.