Self-reported Diabetes in Hawaii: 1988-1993

Field Work Summary
by
Matthew J. Shim, MA, MPH

Presented to
The Department of Public Health Sciences - Biostatistics
at
The University of Hawaii at Manoa, School of Public Health

Committee Members
C. S. Chung, PhD
A. R. Katz, MD, MPH

In partial fulfillment of the degree requirements for
Master of Public Health

December 17, 1995


Contents


Background

Diabetes mellitus is a physiological disorder of carbohydrate metabolism. It is characterized by hyperglycemia and glycosuria which results from the body's inadequate production or utilization of the hormone insulin. (Diabetes, 1985).

In the Unites States, diabetes affects fourteen million people and is the fourth leading cause of death. There are two major type of diabetes. Non-insulin-dependent (Type II) diabetes (NIDDM) accounts for 90-95 percent of diagnosed cases and almost all undiagnosed cases. Insulin-dependent (Type I) diabetes (IDDM) accounts or 5-10 percent of diagnosed diabetes. Gestational diabetes occurs in 2-5 percent of pregnancies (National Institute of Health, 1994).

The Centers for Disease Control and Prevention (CDC) (1993a), reported that in 1990, an estimated 78,000 individuals had diabetes in Hawaii, only half of whom had been diagnosed. The 1990 figures list 39,905 persons with diagnosed diabetes. The direct (medical care) and indirect (lost productivity) cost of diabetes in Hawaii was estimated at $156,000,000.

The prevalence of diabetes is unevenly distributed across the ethnic groups in Hawaii. The Chinese, Filipino, Japanese and Pacific Island populations tend to have higher prevalence rates of diabetes (Wood, Uyeda & Yerkes, 1993). Native Hawaiians have been shown to have the highest age-adjusted rate of diabetes of any ethnic group (Johnson, 1989).

Long term complications of diabetes include cardiovascular disease, stroke, hypertension, blindness, end-stage renal disease, neuropathy, amputations and birth defects in babies born to women with diabetes (NIH, 1994).

The 1993 Behavioral Risk Factor Surveillance System (BRFSS) survey of Hawaii residents documented the rates of diagnosed diabetes to be the highest among the Japanese (93/1000) and the Korean (77/1000) populations (Wood, et. al., 1993).

The purpose of this project was to identify trends in diabetes rates over the past eight years of BRFSS data collection (1986-1993). Trends in diabetes rates among Hawaiian/Pacific Islanders and Japanese ethnic groups and trends in diabetes rates among individuals at risk for obesity were of particular interest.


Behavioral Risk Factor Surveillance System (BRFSS)

Initiated in 1986 with the assistance of the CDC, the BRFSS was designed to collect heath risk behavior information from adult residents and monitor the prevalence of the behaviors over time. The Hawaii State Health Department has published yearly finding from 1986 through 1993 (Hawaii State Health Department Health Promotion and Education Branch, 1986-1993).

The sample population (1986-1993) included all state residents over age 18, except for institutionalized persons and on-base military personnel. The BRFSS followed the CDC recommended random digit dialing sampling procedure. Originally, data collected since 1986 were considered for this project.

After reviewing the data, only the 1988-1993 subset was used. Data from 1986 and 1987 could not be used because the surveys did not include any questions about diabetes. For the 1988-1993 subset, the following question was asked regarding diabetes:

"Next, I'd like to ask you about diabetes, sometimes called sugar diabetes. Have you ever been told by a doctor that you have diabetes?"
    Possible responses were:
  1. Yes
  2. No
  3. Don't know/Not sure
  4. Refused

Care must be taken when interpreting this self-reported diabetes data because there is no way to determine actual clinical cases of diabetes mellitus. Also, there is no differentiation made between the three types. A 'yes' answer could include Type I, Type II and gestational diabetes. The assumption, based on current National Institute of Health statistics (NIH, 1994), is that the majority of 'yes' responses equal Type II diabetes.


Data Set

The following variables were extracted from the original 1988-1993 data sets for this analysis:

VariableVariable Description
AgeAge of respondent
DiabetesEver told by doctor - have diabetes
ObesityComputed risk for obesity (120% or more of ideal body weight)
Pregnant Currently pregnant
RaceRace/ethnicity
SexGender
StrataIsland of residence
MMMonth surveyed
YYYear surveyed

The 1988-1993 data set include a total of 11,622 Hawaii residents, 18 to 99 years of age. A new coding scheme was used to reduce the original ten BRFSS ethnicity categories to six. Ethnicity categories were combined into: 1) Caucasian, 2) Hawaiian (including Part-Hawaiian)/Pacific Islander, 3) Chinese/Korean, 4) Japanese, 5) Filipino and 6) African American/Other. Table 1 and Figure 1 provide the ethnic and gender distribution of the respondents.

Figure 1.
Ethnicity Distribution of BRFSS Respondents (1988-1993)
Table 1.
Ethnicity and Gender Distribution of BRFSS Respondents (1988-1993)
EthnicityMaleFemaleTotal
Caucasian212723064433
Hawaiian/Pacific Islander7539781731
Chinese/Korean316372688
Japanese119015292719
Filipino5907551345
African Am/Other349357706
Total5325629711622


Data Analysis

The original intent of this project was to identify trends in diabetes rates over the past eight years of BRFSS data collection (1986-1993). Using the BRFSS data from 19886 through 1990, Chung, Villafuerte, Wood and Lew (1992) found significant increasing trends in lack of exercise and obesity in Hawaii's population. However, due to unexplainable inconsistencies in the data, possibly associated with changes in the survey organizations, time trend analysis on the diabetes data was not possible. Exploration of the data was still possible by calculating prevalence rates of self-reported diabetes and obesity by gender, age group, and ethnicity. It was assumed that the inconsistencies in the data were not related to these factors.

Analysis of the data involved a two-step process, descriptive and analytic. The descriptive portion of the data analysis provided prevalence rates of self-reported diabetes among the total sample. Prevalence rate was defined as: Number of EXISTING cases of diabetes divided by Total Population (1988-1993). The analytic portion involved performing logistic regression analysis to identify potential variables that are associated with self-reported diabetes.


Self-reported Diabetes Rates per 1000 Population

Total population

Self-reported diabetes is defined as any BRFSS respondent who answered 'yes' to the question about diabetes. Of the sample population (11,622), there were 684 cases of self-reported diabetes. Prevalence rates for self-reported diabetes were calculated for the pooled survey population of 1988 through 1993. Male respondents had a prevalence rate of 49.95, females 66.57 and the total population 58.85 (all rates are per 1000 population). Female respondents had a higher prevalence of self-reported diabetes than the male respondents. This is consistent with national statistics showing higher prevalence among women in the U.S. (NIH, 1994).

Ethnicity by Gender

The Japanese had the highest overall prevalence of self-reported diabetes, followed by Filipinos, then Hawaiian/Pacific Islanders, Chinese/Koreans, African Am./Others and lastly Caucasians. Table 2 and Figure 2 represent the self-reported diabetes rates for each ethnicity by gender.

These rates reflect the 1989 Hawaii Health Surveillance Program which found the Japanese with the highest rate (46.7 per 1000), followed by Filipinos (31.3 per 1000), then Hawaiian/Part Hawaiian (23.8 per 1000), Other (19.9 per 1000), Chinese (17.1 per 1000) and Caucasians (10.1 per 1000) (Wood, et. al., 1994).

Wood, et. al. (1994) reported that among Hawaii residents, individuals of Japanese ethnicity had the highest diagnosed diabetes rates (93 per 1000, age and sex adjusted) followed by individuals of Korean ethnicity (77 per 1,000, age and sex adjusted). This was based on BRFSS data collected in 1993.

The Office of Hawaiian Affairs (OHA) (1994) reported that in 1990, approximately 77 per 1000 Hawaiians/Part Hawaiians ages 36-65 had diabetes. OHA also lists diabetes as the most common chronic condition in the Native Hawaiian population and that Native Hawaiians comprised 44% of all the cases reported in the State in 1990. These figures were based on the Hawaii State Health Surveillance Program, Special Tabulation, 1990.

Figure 2.
Self-reported Diabetes Rate:
Ethnicity by Gender
Table 2.
Self-reported Diabetes Rates
(per 1000 population):
Ethnicity by Gender
EthnicityMaleFemaleTotal
Caucasian23.5136.4330.23
Hawaiian/Pacific Islander58.4384.8773.37
Chinese/Korean79.1167.2072.67
Japanese83.1993.5389.00
Filipino59.3287.4275.09
African Am/Other37.2547.6242.49

Age groups by Ethnicity

The sample population was divided into three age categories based on previous NIH defined age-categories (NIH 1985): 18-44 years old, 45-64 years old, and 65 years or older. Table 3 and Figure 3 show the rates for each ethnicity by age group. (Table 3 includes the age group totals). These data are also consistent with national findings that prevalence of diabetes increases with age (NIH 1994, CDC 1993).

Figure 3.
Self-reported Diabetes Rate:
Ethnicity by Age Group
Table 3.
Self-reported Diabetes Rates
(per 1000 population):
Ethnicity by Age Group
Ethnicity18-44 years45-64 years65 years +
Caucasian13.2344.6587.19
Hawaiian/Pacific Islander39.00149.61158.62
Chinese/Korean34.9594.97145.99
Japanese43.58101.04159.12
Filipino36.30130.81170.07
African Am/Other28.6290.91108.11
Total28.2289.07134.19


Calculated Obesity Risk Rates per 1000 Population

Total population

Obesity is a known risk factor for the development of diabetes mellitus (NIH, 1994). For the BRFSS respondents, obesity risk was defined as any person twenty percent or more above 'ideal weight,' based on 1959 Metropolitan Life Insurance Height/Weight tables. This variable was calculated based on the respondents height and weight. Prevalence rates for obesity risk were calculated for the pooled survey population of 1988 through 1993. Male respondents had an prevalence rate of 256.86, females 212.94 and the total population 233.32 (all rates are per 1000 population).

Ethnicity by Gender

Among ethnic groups in Hawaii, Hawaiians, part-Hawaiians and Pacific Islanders have a known risk for obesity (OHA, 1994 and Hawaii State Health Department, 1986-1993). Table 4 and Figure 4 show the obesity risk rates for each ethnicity by gender. The data shows that almost half of all the Hawaiian/Pacific Islanders who responded to the survey did have a risk for obesity.

Figure 4.
Calculated Obesity Rates:
Ethnicity by Gender
Table 4.
Calculated Obesity Rates
(per 1000 population):
Ethnicity by Gender
EthnicityMaleFemaleTotal
Caucasian217.08188.36202.29
Hawaiian/Pacific Islander491.98414.09448.66
Chinese/Korean154.34129.83141.16
Japanese214.04154.73181.07
Filipino217.62172.13192.22
African Am/Other296.51250.73273.65

Age groups by Ethnicity

Table 5 and Figure 5 show the breakdown of obesity rates for each ethnicity by age group. (Table 5 includes the age group totals).

Figure 5.
Calculated Obesity Risk Rate:
Ethnicity by Age Group
Table 5.
Calculated Obesity Risk Rate
(per 1000 population):
Ethnicity by Age Group
Ethnicity18-44 years45-64 years65 years +
Caucasian163.46286.25234.23
Hawaiian/Pacific Islander423.57547.17398.60
Chinese/Korean136.99184.9796.30
Japanese194.17193.17142.86
Filipino200.71193.35137.68
African Am/Other244.49383.18388.89
Total222.54282.00196.17


Self-reported Diabetes Rates per 1000 Population for individuals 'at risk' for Obesity

Ethnicity by Gender

Table 6 and Figure 6 show the rates of self-reported diabetes (for those individuals 'at risk' for obesity) for each ethnicity and gender. Females show both a higher prevalence of obesity and self-reported diabetes across all ethnic groups.

Figure 6.
Self-reported Diabetes Rates (per 1000 population) if 'at risk for Obesity:
Ethnicity by Gender
Table 6.
Self-reported Diabetes Rates
(per 1000 population)
if 'at risk for Obesity
Ethnicity by Gender
EthnicityMaleFemaleTotal
Caucasian51.9582.5566.59
Hawaiian/Pacific Islander78.80143.96112.29
Chinese/Korean83.33127.66105.26
Japanese110.67205.24155.60
Filipino95.24142.86119.05
African Am/Other58.82103.4579.37


Summary of Prevalence Rates

For the 1988-1993 BRFSS respondents:


Logistic Regression Analysis

Logistic regression analysis, using Statistical Package for the Social Sciences for Microsoft Windows (SPSS for Windows®), was performed to identify potential variables that can be used to identify risk factors for self-reported diabetes. Based on the above analysis of prevalence rates of self-reported diabetes, the following hypotheses were tested:

  1. As age increases, risk of self-reported diabetes increases.
  2. Females have a higher risk of self-reported diabetes than males.
  3. As risk of obesity increases so does the risk of self-reported diabetes.
  4. Risk of self-reported diabetes increases if respondents belong to the Hawaiian/Pacific Islander ethnicity group and Hawaiian/Pacific Islander with a risk for obesity.
  5. Risk of self-reported diabetes increases if respondents belong to the Japanese ethnicity group.

Three analyses were performed to test the above hypotheses. Analysis Number 1 was performed to identify risk factors for self-reported diabetes. The only ethnicity factor used in this analysis was if respondents did or did not belong to the Hawaiian/Pacific Islander ethnicity group. Analysis Number 2 was performed to identify risk factors for self-reported diabetes without any ethnicity factors present. Because the Japanese are known to be at higher risk for diabetes, it was decided that factor for Japanese ethnicity should be tested without any other ethnicity factors. Analysis Number 3 included all independent variables used in Analysis Number 2 plus Japanese ethnicity.

Analysis Number 1

All independent variables and interaction terms were entered into the analysis in one step. Table 7 lists the results of Analysis Number 1.

Table 7.
Logistic Regression Summary - Analysis Number 1
Variable B S.E. Wald df Sig. R Exp(B)
Relative Risk
Age Variable 1 (45-64 years vs 18-44 years) 1.5217 0.1658 84.2561 1 0.0000 0.1275 4.5801
Age Variable 2 (65 years + vs 18-44 years) 2.2881 0.1714 178.1118 1 0.0000 0.1865 9.8560
Gender 0.7911 0.1570 25.3892 1 0.0000 0.0680 2.2058
Obesity Risk 0.8864 0.0970 83.4840 1 0.0000 0.1269 2.4265
Hawaiian/Pacific Islander Ethnicity 0.0977 0.1916 0.2605 1 0.6100 0.0000 1.1027
Hawaiian/Pacific Islander x Obesity 0.1659 0.2241 0.5483 1 0.4590 0.0000 1.1805
Age Variable 1 x Gender -0.6821 0.2150 10.0655 1 0.0015 -0.0399 0.5055
Age Variable 2 x Gender -0.8979 0.2183 16.9164 1 0.0000 -0.0543 0.4074
Age Variable 1 x Gender x Hawaiian/Pacific Islander 0.4483 0.2504 3.2043 1 0.0734 0.0154 1.5656
Age Variable 2 x Gender x Hawaiian/Pacific Islander -0.4220 0.3580 1.3894 1 0.2385 0.0000 0.6558
Constant -4.3244 0.1381 980.6941 1

Summary of Analysis Number 1

AGE

The categorical variables of Age (Age Variable 1 and Age Variable 2) were significantly associated with self-reported diabetes (p < .001, R=0.1275 and p < .001, R=0.1865, respectively). Individuals ages 45-64 had a relative risk of self-reported diabetes four and one half times greater than individuals ages 18-44. Individuals ages 65 and older had a relative risk of self-reported diabetes almost ten times greater than individuals ages 18-44.

GENDER

Gender was significantly associated with self-reported diabetes (p < .001, R= 0.0680). Females had a relative risk of self-reported diabetes over two times greater than males.

OBESITY

Obesity was significantly associated with self-reported diabetes. (P < .001, R=0.1269). Individuals who were 'at risk' for obesity, i.e. 120% of ideal body weight, had a relative risk of self-reported diabetes almost two and one half times greater than those with no risk for obesity.

HAWAIIAN/PACIFIC ISLANDER ETHNICITY

The Hawaiian/Pacific Islander ethnicity group was not found to be significantly different from the rest of the population for self-reported diabetes (p > .05, R= 0.0).

Interaction Term: HAWAIIAN/PACIFIC ISLANDER ETHNICITY x OBESITY

The two-way interaction term of belonging to the Hawaiian/Pacific Islander ethnicity group and being 'at risk' for obesity was not associated with self-reported diabetes (p > .05, R=0.0), suggesting that the effect of obesity is not significantly different between the Polynesian and non-Polynesian groups.

Interaction Terms:

AGE VARIABLE 1 x GENDER

The two-way interaction term of Age Variable 1 (45-64 years vs. 18-44 years) and Gender was significantly associated with self-reported diabetes (p < .05, R=-0.0399). The negative direction of this association shows that self-reported diabetes rates for females ages 45-64 decreased. Females ages 45-64 had a relative risk only one half times greater than males of the same age group.

AGE VARIABLE 2 x GENDER

The two-way interaction term of Age Variable 2 (65 years + vs. 18-44 years) and Gender was significantly associated with self-reported diabetes (p < .001, R=-0.0543). The negative direction of this association shows that self-reported diabetes rates for females 65 years or older decreased. Females ages 65 years or older had a relative risk less than one half times greater than males of the same age group.

Table 8 provides the self-reported diabetes rates (per 1000 population) for Gender by Age Group with Figure 7 providing a graphical representation of this interaction.

Figure 7.
Self-reported Diabetes Rates: Age group by Gender Interaction
Table 8.
Self-reported Diabetes Rates
(per 1000 population):
Gender by Age Group
Gender18-44 years45-64 years65 years +
Male18.9081.95135.99
Female36.5894.73132.86






Interaction Terms:

AGE VARIABLE 1 x GENDER x HAWAIIAN/PACIFIC ISLANDER
and
AGE VARIABLE 2 x GENDER x HAWAIIAN/PACIFIC ISLANDER

The three-way interaction terms of age, gender and Hawaiian/Pacific Islander ethnicity was not associated with self-reported diabetes (p > .05), suggesting that the effects of age and gender is not significantly different between the Polynesian and non-Polynesian groups.


Analysis Number 2

This analysis was performed to identify risk factors for self-reported diabetes regardless of ethnicity. All independent variables and interaction terms were entered into the analysis in one step. Table 9 lists the results of Analysis Number 2.

Table 9.
Logistic Regression Summary - Analysis Number 2
Variable B S.E. Wald df Sig. R Exp(B)
Relative Risk
Age Variable 1 (45-64 years vs 18-44 years) 1.5098 0.1656 84.2561 1 0.0000 0.1266 4.5260
Age Variable 2 (65 years + vs 18-44 years) 2.2710 0.1700 7.0000 1 0.0000 0.1859 9.6894
Gender 0.7977 0.1570 0.1569 1 0.0000 0.0686 2.2205
Obesity Risk 0.9595 0.0970 0.0849 1 0.0000 0.1575 2.6104
Age Variable 1 x Gender -0.5989 0.2150 0.2075 1 0.0045 -0.0347 0.5545
Age Variable 2 x Gender -0.9477 0.2183 0.2157 1 0.0000 -0.0584 0.3876
Constant -4.3129 0.1381 0.1352 1

Summary of Analysis Number 2

AGE

The categorical variables of Age (Age Variable 1 and Age Variable 2) remained significantly associated with self-reported diabetes (p < .001, R=0.1266 and p < .001, R=0.1859, respectively). Individuals ages 45-64 still showed a relative risk of self-reported diabetes four and one half times greater than individuals ages 18-44. Individuals ages 65 and older showed a relative risk of self-reported diabetes approximately nine and one half times greater than individuals ages 18-44.

GENDER

Gender remained significantly associated with self-reported diabetes (p < .001, R= 0.0686). Females had a relative risk of self-reported diabetes over two times greater than males.

OBESITY

Obesity remained significantly associated with self-reported diabetes. (P < .001, R=0.1575). Individuals who were 'at risk' for obesity, i.e. 120% of ideal body weight, had a relative risk of self-reported diabetes over two and one half times greater than those with no risk for obesity.

Interaction Terms:

AGE VARIABLE 1 x GENDER and AGE VARIABLE 2 x GENDER

The two two-way interactions of age by gender remained significantly associated with self-reported diabetes. (Note: Refer to Table 8 for self-reported diabetes rates (per 1000 population) for Gender by Age Group and Figure 7 for a graphical representation of this interaction.)


Analysis Number 3

All independent variables and interaction terms were entered into the analysis in one step. Table 10 lists the results of Analysis Number 3.

Table 10.
Logistic Regression Summary - Analysis Number 3
Variable B S.E. Wald df Sig. R Exp(B)
Relative Risk
Age Variable 1 (45-64 years vs 18-44 years) 1.4767 0.1659 79.2045 1 0.0000 0.1235 4.3785
Age Variable 2 (65 years + vs 18-44 years) 2.1962 0.1716 163.7395 1 0.0000 0.1787 8.9909
Gender 0.0870 0.1571 26.3833 1 0.0000 0.0694 2.2411
Obesity Risk 1.0161 0.0859 139.8504 1 0.0000 0.1650 2.7623
Japanese Ethnicity 0.4849 0.0887 29.8654 1 0.0000 0.0742 1.6241
Age Variable 1 x Gender -0.6032 0.2077 8.4313 1 0.0037 -0.0356 0.5470
Age Variable 2 x Gender -0.9880 0.2163 20.8713 1 0.0000 -0.0611 0.3723
Constant -4.4417 0.1385 1029.0240 1

Summary of Analysis Number 3

Age, Gender, Obesity Risk, and the interaction terms of Age by Gender were all significantly associated with self-reported diabetes as in Analysis Number 1 and Analysis Number 2.

JAPANESE ETHNICITY

The Japanese ethnicity group was found to be significantly different from the rest of the population for self-reported diabetes (p < .001, R=0.0742). The Japanese had a relative risk of self-reported diabetes over one and one half times greater than non-Japanese respondents.


Summary of the Logistic Regression Analyses

For the 1988-1993 BRFSS respondents:

Summary

Behavioral Risk Factor Surveillance System in Hawaii

Based on pooled BRFSS survey responses (1988-1993), self-reported diabetes appeared in almost 6% of the surveyed population. Self-reported diabetes was defined as any BRFSS respondent who answered 'yes' to the question: "Have you ever been told by a doctor that you have diabetes?" No differentiation was made between the three types of diabetes (Type I, Type II, and gestational diabetes). This poses a particular problem when interpreting the results of this project.

Because there was only one question asked about diabetes, the assumption of this project was that a 'yes' responses equaled Type II diabetes. Additional questions regarding diabetes on future BRFSS surveys would provide more information on the impact of diabetes in Hawaii. The other questions that need to be asked include: respondents age at time of diagnosis, type of diagnosis (childhood, adult onset or gestational), and type of treatment receiving (none, insulin, non-insulin, etc.).

Risk of Self-reported Diabetes in Hawaii

In Hawaii, the risk of self-reported diabetes increases with age. Hawaii residents have one of the highest life expectancies in the U.S. (Wood, et. al., 1994). As Hawaii's population grows older, diabetes will continue to affect the health of Island residents. Particularly, the Japanese in Hawaii are at greater risk of self-reported diabetes than any other ethnic group.

Obesity is a known risk factor for diabetes (NIH, 1994). This project found that Hawaii residents at risk for obesity, i.e. 120% or more above ideal body weight, were at greater risk of self-reported diabetes than those with no risk for obesity. Chung, et. al. (1992) reported a significant positive trend of obesity among Hawaii's population from 1986 through 1990. If this trend continues, more of Hawaii's population will be at risk for developing diabetes.

This project found that females in Hawaii had an overall greater risk of self-reported diabetes than males. This reflects national findings that show a higher prevalence of diabetes among women (NIH, 1994). An interesting finding was the interaction of age and gender on self-reported diabetes in Hawaii. As age increased, both male and female respondents showed increasing rates of self-reported diabetes. However, as age increased, the rate of self-reported diabetes for females decreased. In the 65 years or older age group, the rate of self-reported diabetes for females fell below the rate for males in the same age group (see Table 8 and Figure 7). One possible explanation is that, in Hawaii, females with diabetes have a higher mortality rate than males. Another possibility is that the survey under-reported diabetes in the older female population. Hawaii does not have a centralized diabetes surveillance system and the Hawaii State Health Department relies on individual organizations for diabetes data (Wood, et. al., 1994). Further research should be conducted to determine the impact of diabetes mortality on women in Hawaii.

Hawaiians, part-Hawaiians and Pacific Islanders have a known risk for obesity (OHA, 1994 and Hawaii State Health Department, 1986-1993). This project found that almost half of all Hawaiian/Pacific Islanders who responded to the BRFSS survey were at risk for obesity. It was hypothesized that Hawaiian/Pacific Islanders and obesity among the Hawaiian/Pacific Islander would show a more serious effect on the risk of self-reported diabetes. This hypothesis was found to be not true.

The Hawaiian/Pacific Islander ethnicity group was not found to be significantly different from the rest of the population for risk of self-reported diabetes. Also, the effect of obesity on risk of self-reported diabetes is not significantly different between the Polynesian and non-Polynesian ethnic groups.

We know that the prevalence of diabetes is unevenly distributed across ethnic groups in Hawaii, particularly the Japanese (Wood, et. al., 1994). For this data, these differences may not be due to lack of access to health care. BRFSS data on doctor's visits (1988-1993) shows that over 80% of respondents, in each ethnic group, reported visiting the doctor within the past 2 years.

Long term complications of diabetes include cardiovascular disease, stroke, hypertension, blindness, end-stage renal disease, neuropathy, amputations and birth defects in babies born to women with diabetes (NIH, 1994). The Hawaii State Health Department estimated the 1992 direct and indirect costs of diabetes in Hawaii to be $558,000,000 (Hawaii State Health Department, 1995). This is an estimated cost increase of three and one half times (up from $156,000,000 in 1990) in just two years. These costs will continue to increase without active surveillance systems and prevention programs for diabetes.


References

Burchfield, C. M., Sharp, D. S., Curb, J. D., Rodriguez, B. L., Hwang, L., & Marcus, E. B. (1995). Physical Activity and the Incidence of Diabetes: The Honolulu Heart Program. American Journal of Epidemiology, 141, 360-368.

Centers for Disease Control and Prevention. (1991). Diabetes Surveillance, 1991. (The Division of Diabetes Translation). Washington, DC: U.S. Government Printing Office.

Centers for Disease Control and Prevention. (1993a). Special Focus: Behavioral Risk Factor Surveillance - United States, 1991 (MMWR CDC Surveillance Summaries, Vol. 42, No. SS-4). Washington, DC: U.S. Government Printing Office.

Centers for Disease Control and Prevention. (1993b). Diabetes in the United States: A Strategy for Prevention. (A Progress Report to the Technical Advisory Committee for Diabetes Translation and Community Control Programs). Washington, DC: U.S. Government Printing Office.

Chung, C. S., Villafuerte, A., Wood, D. W., & Lew, R. (1992). Trends in Prevalences of Behavioral Risk Factors: Recent Hawaiian Experiences. American Journal of Public Health, 82, 1544-1546.

Diabetes. (1985). In Taber's Cyclopedic Medical Dictionary (15th ed., pp. 452-454). Philadelphia: F. A. Davis Company.

Hawaii State Health Department Health Promotion and Education Branch. (1995). Diabetes Statistics for Hawaii. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1986). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1987). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1988). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1989). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1990). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1991). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1992). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Hawaii State Health Department Health Promotion and Education Branch. (1993). Hawaii's Health Risk Behaviors. Honolulu, Hawaii.

Johnson, D. B. (1989). "Diabetes: Epidemiology and Disability" in E. L. Wegner (ed), Social Process in Hawaii, 32. Department of Sociology, University of Hawaii at Manoa.

National Diabetes Data Group. (1985). Diabetes in America: Diabetes Data Compiled 1984 (NIH Publication No. 85-1468). Washington, DC: U.S. Government Printing Office.

Office of Hawaiian Affairs. (1994). Native Hawaiian Data Book. Honolulu, Hawaii.

U.S. Department of Health and Human Services. (1987). Prevalence of Diagnosed Diabetes, Undiagnosed Diabetes, and Impaired Glucose Tolerance in Adults 20-74 Years of Age - United States, 1976-1980. (Vital and Health Statistics, DHHS Publication No. (PHS) 87-1687). Washington, DC: U.S. Government Printing Office.

Diabetes Overview (1994). U.S. Department of Health and Human Services. (NIH Publication No. 94-3235). Washington, DC: U.S. Government Printing Office.

Diabetes Statistics. (1994). U.S. Department of Health and Human Services. (NIH Publication No. 94-3822). Washington, DC: U.S. Government Printing Office.

Wood, D. W., Uyeda, C., & Yerkes, J. (1993). Community Epidemiology Work Group for Diabetes Mellitus (1993 Monograph). Honolulu, HI: University of Hawaii at Manoa, International Center for Health Promotion and Disease Prevention Research.



Copyright © 1995, 1996 by Matthew J. Shim
Home