Diabetes management and vaccination rates in the southeast United States, 2000 through 2007

Abstract

Objective

To examine trends in diabetes management, by race and gender, in the southeast U.S.

Design

Population-based survey

Setting

Southeast U.S. (i.e., Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, and Tennessee).

Participants

Whites and blacks surveyed through the CDC Behavioral Risk Factor Surveillance System between 2000 and 2007 in the southeast U.S.

Main Outcome Measures

Total diabetes management included personal factors (checking blood glucose and feet), healthcare provider factors (visiting a doctor, checking hemoglobin A1c, checking feet) and vaccinations (influenza, pneumonia).

Results

Vaccination levels were low throughout the study period, and a racial disparity in vaccination rates developed due to the increasing trends in prevalence in whites relative to blacks. Personal diabetes management increased significantly in all race/gender groups with greater increases in black than white men, resulting in white men having lower point estimates of prevalence than black men. Healthcare provider diabetes management increased statistically significantly in all race/gender groups with the exception of black men. Considering vaccinations together with personal and provider diabetes management criteria, diabetes management improved significantly in all race/gender groups, but remained low in 2007 (i.e., 8.8%, black women; 14.0%, white women; 11.5%, black men; 12.8% white men).

Conclusion

Emphasis should be placed on improving vaccination levels, diabetes patient self-management and provider-management in the southeast U.S. Although diabetes management improved over the study time period, in 2007 the percent of individuals meeting all of the diabetes management criteria examined remained very low ranging from 8.8% in black women to 14.0% in white women.

INTRODUCTION

In 2006 diabetes was the sixth leading cause of death in the United States¹. Studies have shown that the prevalence of diabetes has dramatically increased over the past 20 to 30 years, and has disproportionately affected minority populations with higher disease prevalence and worse disease outcomes ^{2, 3}.

The southeast U.S. is the region with the highest prevalence of heart disease, stroke, congestive heart failure and renal failure in the U.S. ^4-6 and has a higher percentage of African Americans relative to the remainder of the U.S. Diabetes management has been shown to differ between racial groups, often influenced by socioeconomic factors, access to care, and disease awareness and knowledge, and may influence the racial disparities in diabetes outcomes ^7-13. Effective diabetes management, including self-care, healthcare provider-care, and getting vaccinated against influenza and pneumonia, is vital in reducing diabetes morbidity and mortality.

As diabetes prevalence continues to rise, it is important to understand secular trends and racial disparities in disease management in individuals with diabetes. Therefore, this study aimed to examine the trends in diabetes prevalence and diabetes management, by race and gender, in the southeast U.S. between 2000 and 2007.

METHODS

STUDY POPULATION AND DATA

This study utilized the Center for Disease Control and Prevention’s (CDC) Behavioral Risk Factor Surveillance System (BRFSS). The survey uses random digit dialing to collect health data from individuals aged 18 and older from all 50 states, the District of Columbia, the U.S. Virgin Islands, Guam and Puerto Rico. The median state survey response rate for the BRFSS ranged from 48.9% to 53.2% between 2000 and 2007¹⁴.

Respondents located in the southeast U.S. who were surveyed between 2000 and 2007 were included in this study. The Department of Health and Human Services’ Region four, used to define the southeast U.S, includes Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, and Tennessee. Each survey respondent was weighted to account for differences in probability of selection, the number of residential telephones in the household, the number of adults in the household, non-coverage, and non-response.

DEFINITIONS

Diabetes was defined based on a positive response to the following question, “Have you ever been told by a doctor that you have diabetes?” Women who reported diabetes only when pregnant were classified as non-diabetic individuals. Non-Hispanic white and non-Hispanic black racial groups were determined from self-reported race and ethnicity. All diabetes management variables, including vaccination rates, were determined from self-reported data. Management variables were determined based on recommendations by the American Diabetes Association, the CDC, Healthy People 2010 and the questions included in the BRFSS^14-18. An individual was considered to personally manage their diabetes if they reported checking their blood glucose levels at least once a day and if they checked their feet for sores or irritations at least once per day. An individual was considered to manage their diabetes at the healthcare provider level if they visited a doctor at least twice a year, had their hemoglobin A1c checked at least twice a year, and had their feet checked for sores or irritations by a healthcare professional at least once a year. A respondent was adequately vaccinated if they received an influenza vaccination within the previous year and if they had ever received a pneumonia vaccination. A respondent was said to have total diabetes management if they managed their diabetes on a personal and healthcare provider level and they were adequately vaccinated.

STATISTICS

BRFSS weights that account for differences in probability of selection, the number of residential telephones in the household, the number of adults in the household, non-coverage, and non-response were not used for the initial descriptive characterization of the study sample (ie. Table 1), but were used for all subsequent analyses¹⁴.

Table 1.

Unadjusted^* characteristics (percent or mean) of respondents with diabetes by year.

	2000 (n=2198) %(se)	2001 (n=2868) %(se)	2002 (n=3505) %(se)	2003 (n=4685) %(se)	2004 (n=5458) %(se)	2005 (n=6569) %(se)	2006 (n=7748) %(se)	2007 (n=13092) %(se)
White Males	30.8 (1.0)	31.0 (0.9)	31.1 (0.8)	30.2 (0.7)	29.8 (0.6)	29.7 (0.6)	31.7 (0.6)	32.8 (0.4)
White Females	43.0 (1.1)	44.1 (1.0)	43.6 (0.9)	42.4 (0.8)	45.0 (0.7)	44.4 (0.6)	44.3 (0.6)	45.0 (0.5)
Black Males	7.6 (0.6)	7.7 (0.5)	7.5 (0.5)	8.2 (0.4)	7.0 (0.4)	7.9 (0.4)	6.9 (0.3)	6.3 (0.2)
Black Females	18.5 (0.9)	17.1 (0.7)	17.8 (0.7)	19.2 (0.6)	18.1 (0.5)	18.1 (0.5)	17.1 (0.5)	15.8 (0.3)
Age (years ± se)	59.4 (14.1)	58.7 (14.9)	60.2 (14.3)	59.8 (14.2)	60.9 (13.9)	61.2 (13.5)	61.6 (13.7)	62.6 (13.2)
High School	68.3 (1.0)	70.7 (0.9)	71.4 (0.8)	71.9 (0.7)	73.2 (0.6)	74.3 (0.5)	76.2 (0.5)	77.4 (0.4)
Access to Care	82.7 (0.8)	85.8 (1.9)	90.2 (0.5)	80.7 (0.6)	81.3 (0.5)	82.8 (0.5)	83.4 (0.4)	84.3 (0.3)

^*In this table participant characteristics have not been weighted to account for the BRFSS study design and the n reflects the actual number of individuals interviewed.

Data were analyzed using the SAS System (version 9.1; SAS institute; Cary, NC) and its complex survey-specific procedures. A multivariate logistic regression model was used to assess the age-adjusted prevalence of each outcome variable in each race/gender group for each year. The model included age, and 31 race/gender/year combination variables. The average age of the diabetic population, 61.5 years was included in the model when solving for each race/gender/year group. The model was weighted to account for the survey design of the data¹⁴.

Logistic regression models including year, three race/gender groups and three interaction terms that were the race/gender groups multiplied by year were used to determine if there was a linear trend in the outcome over the study time period in each race/gender group, and the percent change in outcome prevalence over the study time period. The model also included either age, or age, socioeconomic status, and access to care to control for possible confounding effects. Self-reported education level was used as a surrogate for socioeconomic status; access to care was also self-reported. Wald chi-square probabilities were utilized to determine significant values for year variables, indicating significant linear trends over the study time period in the outcome variable in each race/gender group. Models were solved for year to determine the percent change in outcome prevalence over the study time for each race/gender group. The derivation for the confidence intervals for percent change in outcome prevalence over the study time period is found in the Appendix.

Lastly, a logistic regression model including year, race, gender, year/race, year/gender, race/gender, and year/race/gender variables was used to determine if there were significantly different linear trends in an outcome over the study time period between racial groups. A statistically significant year/race variable indicated that there was a statistically significant difference in the outcome’s linear trend over the study time period between either white and black men or white and black women.

P-values less than 0.05 were taken as statistically significant. No correction for multiple testing has been applied to reported p-values.

RESULTS

Over the study time period, a total of 425,020 individuals responded to the BRFSS in the southeast U.S. Of those, 46,121 (10.9%) had self-reported diabetes. Of the individuals with diabetes, the highest proportion of respondents was from Florida (20.2%), and the lowest proportion of respondents was from Tennessee (6.9%). Over the study time period, 44.4% of the respondents with diabetes were white females, 31.2% were white males, 17.3% were black females, and 7.2% were black males. The average age of respondents with diabetes was 61.5 ± 13.3 years and ranged from 58.7 ± 14.9 to 62.6 ± 13.2 between 2000 and 2007. An average of 74.4% of respondents with diabetes completed high school, obtained their GED, or had a higher education level, and an average of 83.6% had access to care (Table 1). Descriptive characteristics presented above are unweighted and therefore are reflective of the actual study population utilized rather than the population distribution in the southeast U.S.

In 2000, the diabetes prevalence in the southeast U.S. ranged from 7.7% in white women to 20.1% in black women. Over the study time period, there was a significantly increasing age-adjusted trend in diabetes prevalence in white men, white women, and black men, but no significant change in black women in the southeast U.S. The age-adjusted diabetes prevalence increased by 4.22% (95% CI: 3.11-5.34%) in white men, 3.03% (95% CI: 2.26-3.80%) in white women, and 3.53% (95% CI: -0.33-7.38%) in black men. The age-adjusted trend in diabetes prevalence in white women was significantly different than that in black women, with the racial disparity decreasing by 1.25% (95% CI: -2.07-4.57%) over the study time period. These trends remained significant after adjusting for SES and access to care (Table 2).

Table 2.

The percent change of each outcome variable trend over the study time period, the confidence interval, and the p-value for trend.

	Age-Adjusted Percent Change 2000-2007			Fully-Adjusted Percent Change† 2000-2007
	% Change	95% CI	p-value for trend	% Change	95% CI	p-value for trend
Diabetes
White Males	4.22	(3.11-5.34)	<.0001	4.35	(3.23-5.47)	<.0001
White Females	3.03	(2.26-3.80)	<.0001*	2.75	(1.97-3.52)	<.0001*
Black Males	3.53	(−0.33-7.38)	0.0269	3.90	(0.09-7.72)	0.0134
Black Females	1.78	(−.0.78-4.33)	0.092*	1.36	(−1.18-3.90)	0.1897*
Personal Diabetes Management
White Males	4.50	(−0.61-9.61)	.0391*	5.77	(0.30-11.24)	0.0132*
White Females	6.05	(1.64-10.46)	0.0014	6.99	(2.13-11.86)	0.0006
Black Males	15.5	(4.57-26.43)	0.001*	16.75	(5.10-28.39)	0.0008*
Black Females	8.91	(1.67-16.14)	0.0041	10.68	(2.95-18.41)	0.0011
Healthcare Provider Diabetes Management
White Males	7.17	(2.07-12.28)	0.001	2.98	(−2.63-8.60)	0.2135
White Females	7.69	(3.28-12.11)	<.0001	4.38	(−0.67-9.42)	0.0396
Black Males	7.34	(−4.61-19.29)	0.1498	0.02	(−13.22-13.26)	0.9971
Black Females	9.69	(2.41-16.98)	0.0017	3.31	(−4.80-11.42)	0.332
Vaccinations
White Males	7.62	(2.42-12.82)	0.0007	6.91	(0.99-12.84)	0.0071
White Females	6.65	(1.98-11.32)	0.001	7.78	(2.34-13.21)	0.0009
Black Males	−2.07	(−14.35-10.22)	0.6778	3.28	(−10.37-16.93)	0.5582
Black Females	3.63	(−3.28-10.54)	0.2081	4.51	(−3.58-12.60)	0.1826
Total Diabetes Management
White Males	7.02	(4.33-9.71)	<.0001	5.95	(2.67-9.22)	<.0001
White Females	7.35	(4.81-9.89)	<.0001	7.60	(4.54-10.66)	<.0001
Black Males	6.23	(0.68-11.78)	0.003	8.47	(2.18-14.76)	0.0004
Black Females	5.75	(2.38-9.12)	<.0001	6.21	(2.01-10.40)	0.0002

All estimates take the sample design/weighting into effect.

^†Fully Adjusted model is adjusted for age, socioeconomic status, and access to care;

^*indicates significantly different racial trends

Personal diabetes management ranged from 39.9% in black men to 49.4% in white women with diabetes in 2000. All four race/gender groups experienced significantly increasing age-adjusted trends in personal diabetes management prevalence over the study time period, increasing 4.50% (95% CI: -0.61-9.61%) in white men, 6.05% (95% CI: 1.64-10.46%) in white women, 15.50% (95% CI: 4.57-26.43%) in black men, and 8.91% (95% CI: 1.67-16.14%) in black women. The increasing trend in black men was significantly larger than the trend in white men, resulting in the prevalence in black men crossing over that of white men. No racial disparity was seen in personal diabetes management throughout the study time period. All trends in personal diabetes management remained significant after adjusting for SES and access to care (Figure 1).

Figure 1.

Trends in diabetes management in the southeast United States 2000-2007 stratified by sex and race. (A) Personal Diabetes Management; (B) Health Provider Diabetes Management; (C) Vaccination Prevalence; (D) Total Diabetes Management. White Women Black Women White Men Black Men

Healthcare provider diabetes management ranged from 38.5% in white men to 44.6% in black men with diabetes in 2000. White men, white women, and black women experienced significantly increasing age-adjusted trends in healthcare provider diabetes management prevalence over the study time period. Healthcare provider diabetes management increased 7.17% (95% CI: 2.07-12.28%) in white men, 7.69% (95% CI: 3.28-12.11%) in white women, and 9.69% (95% CI: 2.41-16.98%) in black women (Figure 1). However, after adjusting for SES and access to care, the trends in healthcare provider diabetes management in white men and black women failed to reach statistical significance. No significant trend in healthcare provider diabetes management was seen in black men with diabetes.

The vaccination prevalence ranged from 23.7% in black men to 37.3% in white women with diabetes in 2000. There were significantly increasing age-adjusted trends in vaccination in white men and women with diabetes over the study time period, with prevalence increasing 7.62% (95% CI: 2.42-12.82%) in white men and 6.65% (95% CI: 1.98-11.32%) in white women. The trends remained significant after adjusting for SES and access to care. No significant trends in vaccination prevalence were seen in black men or women and the increasing trends in whites led to development of a racial disparity in vaccination prevalence by 2007 (Figure 1).

Total diabetes management prevalence ranged from 1.3% in black men to 6.9% in white males with diabetes in 2000. Over the study time period, there were significantly increasing age-adjusted trends, and age-, SES-, and access to care-adjusted trends in total diabetes management prevalence in all four race/gender groups. Total diabetes management increased 7.02% (95% CI: 4.33-9.71%) in white men, 7.35% (95% CI: 4.81-9.89%) in white women, 6.23% (95% CI: 0.68-11.78%) in black men, and 5.75% (95% CI: 2.38-9.12%) in black women (Figure 1). By 2007 among women a racial disparity emerged with white women [14.0% (95% CI: 12.6, 15.5)] having a higher prevalence of total diabetes management than black women [8.5% (95% CI: 7.1, 10.9)] (Figure 1).

All diabetes management outcomes were stratified by age to determine if there were differences in trends in individuals in each race/gender group < 65 years old and ≥ 65 years old. No significant differences were found, with the following exceptions: while there was no significant trend in personal diabetes management prevalence in white women aged < 65, there was a significantly increasing age-, SES-, and access to care-adjusted trend in white women 65-years-old and older , with the prevalence increasing 12.7% (95% CI: 9.36-21.1%). Similar results were seen in black women where the prevalence increased 24.1% (95% CI: 10.9-37.2%) in black women 65-years-old and older.

DISCUSSION

In summary, we found racial disparities in both genders in vaccination rates in the southeast U.S. Vaccination levels were low throughout the study time period, and a racial disparity in vaccination rates developed during the study time period due to the increasing trends in prevalence in white relative to black individuals. Personal diabetes management increased significantly in all race/gender groups with greater increases in black than white men; moreover, although no racial disparity was seen throughout the study, the prevalence of personal diabetes management in black individuals overcame that in white individuals by the end of the study. In contrast, for healthcare provider diabetes management we found statistically significant improvement in all race/gender groups with the exception of black men. Finally, we found that although total diabetes management was increasing in all race/gender groups in the southeast U.S., prevalence according to our criteria remained very low in 2007 ranging from 8.8% in black women to 14.0% in white women.

Individuals with diabetes are more susceptible to influenza and pneumonia, and they often have higher death rates from influenza than people without diabetes^19-21. Results from this study concur with a few other studies which have shown higher vaccination rates in white than black individuals with diabetes, even after adjusting for socioeconomic status and access to care ^{10, 11}. Furthermore, this study showed that from 2000 to 2007 white individuals with diabetes not only had significantly higher rates of vaccinations than black individuals with diabetes, but their rates were significantly increasing over the study time period while no significant trends in vaccination rates were seen in black men or women with diabetes. As a result, a racial disparity in vaccination rates developed during the study time period. The target vaccination rate for individuals with diabetes is 65% in individuals less than 65 years old and 90% for individuals 65 years old and older ¹⁸. This study found no significant difference in vaccination rates for individuals between the two age groups, but rates for the entire diabetic population are far from desired with white women having the highest vaccination rate in 2007 at 48%, and black women having the lowest vaccination rate at 29%.

This study showed that the majority of individuals with diabetes in the southeast U.S. do not meet management guidelines for diabetes; however, significant increasing trends in total diabetes management in all four race/gender groups studied were observed, which may reflect increasing trends in personal diabetes management. Increases in personal diabetes management in this study could be a result of increased diabetes knowledge or awareness over the study time period. In addition to increased media messages about diabetes, there are 59 Centers for Disease Control and Prevention funded diabetes prevention and control programs (DPCP), which are actively building capacity and implementing programs related to diabetes prevention and control and there is a DPCP in each of the states surveyed for this study. Recently, a focus of the programs has been on improving diabetes care within systems as well as policy changes that improve diabetes care and outcomes. Several states have implemented programs that focus on improving diabetes care and outcomes. For example, the South Carolina Legislature provides funding to improve diabetes outreach, diabetes care by health professionals and surveillance of diabetes by the Diabetes Initiative of South Carolina²². This funding has been used to improve self-care and care by health professionals, implement statewide policies for improving care, as well as providing infrastructure and data to obtain other funding for improving diabetes outcomes²². Other diabetes programs such as REACH 2010: Charleston and Georgetown Diabetes Coalition and REACH programs in North Carolina, Tennessee, and Georgia have focused on decreasing disparities for diabetes and related complications²³, and have organized free influenza and pneumonia vaccinations with community partners.

There were several limitations to this study. First, different race/gender groups may be more susceptible to undiagnosed diabetes, misrepresenting the diabetic population in this study. National estimates show black women have the highest prevalence of undiagnosed diabetes (4.1%), followed by white men (3.5%), black men (3.2%), and white women (1.9%) ³. If this study included all individuals with diabetes instead of those with diagnosed diabetes only, diabetes management levels would have decreased, especially in black women, widening racial disparities in women. Second, this study relies on self-reported data. Self-reported data may be more susceptible to recall bias, an overestimation of positive health behaviors and an underestimation of negative health behaviors. However, self-report may be one of the most accurate measures of personal diabetes management and vaccine rates. Typically, individuals have the most accurate knowledge of how often they check their blood glucose levels and their feet for sores. Individuals could also potentially have a better record of vaccinations, due to the wide variety of ways to receive an influenza vaccination each season. Another limitation of this study was that black males represented a small proportion (7.1%) of the study respondents; even with weighted data, the relatively small number of black men reduces the precision of outcome estimates in black males. This study does not examine behavioral health factors, such as smoking status and physical activity levels or include outcome data such as hemoglobin A1c levels, complication rates, hospitalization rates or emergency department visit rates which may impact racial disparities in diabetes outcomes. Lastly, our definitions of personal and healthcare provider diabetes management were limited to information collected.

A strength of the study is that for each state the BRFSS sample is designed to be representative of the non-institutionalized population and studies have shown the BRFSS is valid against the general population ^{24, 25}.

In summary, the results from this study suggest that a more aggressive diabetes management approach needs to be made across all race/gender groups in the southeast U.S. Though diabetes management experienced increases over the study time period, prevalence according to our criteria remained very low in 2007 ranging from 8.8% in black women to 14.0% in white women. Hence, despite marked improvement at the present rate it will take decades for the majority of individuals with diabetes in the southeast U.S. to meet our minimal criteria for diabetes management.

Specifically, improved self-management and health provider care for diabetes are needed. Effective diabetes self-management includes not only self-care related to monitoring glucose and checking feet, but also eating healthy, increasing physical activity, effectively managing stress, healthy coping with daily living, reducing risks of obesity, managing dental problems, eye problems, high blood pressure and cardiovascular disease as well as taking medications if needed to control glucose, lipids, blood pressure, and other cardiovascular problems. Quality care by health providers includes not only semi-annual visits for care and A1C testing and foot checks, but also care by dentists, eye care professionals, and foot care if indicated. Guidelines for diabetes care by health providers are graded according to current evidence and are updated annually by the American Diabetes Association and other health professional organizations^{17, 26}. These guidelines include recommendations for vaccination for influenza and pneumonia. Additionally, Healthy People 2010 have established goals for diabetes care and outcomes, as well as a goal that 90% of high risk person receive a pneumococcal vaccination and an annual influenza vaccine¹⁸. Finally, interventions are needed that target barriers to influenza and pneumonia vaccines among black individuals with diabetes.