No. 2, April 2005
Adherence to Oral Hypoglycemic Agents in Hawaii
Rachel Lee, Deborah A. Taira, ScD
Suggested citation for this article: Lee R, Taira DA. Adherence to oral hypoglycemic agents in Hawaii. Prev Chronic Dis [serial
online] 2005 Apr [date cited].
Available from: URL: http://www.cdc.gov/pcd/issues/2005/
Adherence to oral hypoglycemic agents is essential to reducing
the poor health outcomes of populations at high risk for
developing diabetes and its chronic complications. The goal of
this study was to identify characteristics of patients in Hawaii least
likely to adhere to oral hypoglycemic agents.
This retrospective administrative data analysis included
prescription refill claims for oral hypoglycemic agents from
January 1, 1999, through June 30, 2003 (n = 20,685). Multivariate
logistic regression analysis was used to examine the relationship
between adherence and patient characteristics.
Adherence was found to be strongly associated with age and
ethnicity. Relative to the age subset 55 to 64 years, adherence
increased as age increased, reaching a peak at age 74 (odds ratio
[OR] 1.1; 95% confidence interval [CI], 1.0–1.20). Past the age of
85, adherence declined (OR 0.90; 95% CI, 0.82–0.98). Relative to
white patients, the odds ratio of adherence was highest for
Japanese patients (OR 1.20; 95% CI, 1.0–1.30) and lowest for
Filipino patients (OR 0.78; 95% CI, 0.68–0.90). Gender was not
associated with adherence.
Differences in adherence to oral hypoglycemic agents were
found to be related to ethnicity and age. Adherence was found to
be lowest in younger patients and Filipino patients. This is a
significant finding considering that younger diabetic patients
have been shown to have the poorest glycemic control and worst
health outcomes. Although the literature on adherence to oral
hypoglycemic agents and health outcomes in Filipino patients is
limited, studies support an increased risk for developing
diabetes in this group. This information can be used to target
younger patients and Filipino patients to improve their adherence
to oral hypoglycemic agents.
Back to top
Patient adherence to a prescribed regimen of oral hypoglycemic
agents to prevent diabetes is generally low and difficult to
maintain, even in populations with adequate access to health care
and drug coverage (1,2). This problem poses serious consequences
for Asian-Pacific Islanders, who have a higher genetic
predisposition than whites for developing diabetes and its chronic
complications (3-6). The Asian-Pacific Islander
population in Hawaii is large and heterogeneous, composed of
individuals of Japanese, Chinese, Filipino, Korean, and Hawaiian
ancestry. Comparing the different ethnic groups in Hawaii that
compose the Asian-Pacific Islander category reveals significant
health disparities among them. Diabetes has been found to be
three to seven times more prevalent in Hawaiians and three to
four times more prevalent in Filipinos and Japanese than whites
(7). In addition, Hawaiians have the highest prevalence of
diabetes reported for any Polynesian or part-Polynesian group,
and mixed ancestry has not been shown to diminish the risk for
type 2 diabetes (8,9). Native Hawaiians on average have less
education and lower income and are more likely to live in rural
areas with less access to medical services than other ethnic
groups living in Hawaii. These aspects of the Native Hawaiian
population further elevate its risk for developing diabetes and
its chronic complications.
Evidence suggests that dietary and lifestyle changes place
ethnic minorities at higher risk for developing diabetes. The
Honolulu Heart Program, a study of 8006 Japanese men born from
1900 to 1919, correlated a westernized and sedentary lifestyle
with an increased risk for developing diabetes. These studies
found that Japanese men who retained a Japanese lifestyle and
diet were less likely to develop diabetes than those Japanese men
who followed a westernized lifestyle (10,11).
Different cultural beliefs and dietary practices among the
various ethnic groups in Hawaii affect the outcome of diabetes
and adherence to medications. The literature suggests that
traditional cultural beliefs about family and the social support
it provides when family members are ill play an important role in
either encouraging or preventing individuals from seeking medical
care for diabetes (12,13). Because minority populations and
patients facing socioeconomic barriers to health care access have
been shown to have the worst adherence to medications and poor glycemic control (14), determining the association between ethnicity and
adherence to oral hypoglycemic agents among Japanese, Chinese, Filipino,
Hawaiian, and white patients in Hawaii will help reveal the ethnic
disparities that exist in Hawaii and identify those groups who most need to
be targeted for intervention.
Back to top
The patients in this study were enrollees in a large health
care plan in Hawaii from January 1, 1999, to June 30, 2003, who met
their algorithm for diabetes, had drug coverage, and filled at
least one prescription for one of the following oral hypoglycemic
agents: sulfonylurea, metformin, thiazolidinedione, and
a-glucosidase inhibitors. To determine if a patient had diabetes,
we used a diabetes algorithm that followed two decision paths.
The first path looked for the presence of a diabetes diagnosis
with specific treatment services; the second identified the
presence of diabetes based on specific drugs and medical
supplies, provided the member did not have gestational diabetes.
The characteristics that we controlled for were age, gender,
ethnicity, island of residence, morbidity level, year of treatment,
and type of coverage (i.e., HMO, PPO, Medicare cost contract).
Morbidity level was assessed using the Johns Hopkins
University’s Ambulatory Care Groups (ACGs), derived from the mix
of a member’s diagnoses. In the original study, 51 combinations
or ACGs resulted from applying multivariate techniques to
maximize variance explained in use of services and ambulatory
care charges (15). This method can be applied to large
populations with numerous types of diagnoses to predict
ambulatory care use and cost of care and to determine the burden
of morbidity. High morbidity was defined as a morbidity level of
four or five on a five-point scale.
Sources of data
Patient age, gender, island of residence, morbidity level, comorbid conditions, and type of coverage were obtained from
administrative data. This data did not include ethnicity of the
patient. Self-reported ethnicity information was available for a
portion of members in the study population from existing
health-plan member satisfaction surveys. The satisfaction
survey was a mailed questionnaire filled out by members to
describe their experiences with health care services; on that
survey, members were asked to check all that apply among 17
ethnic categories. These categories were chosen to be consistent
with the Hawaii Department of Health’s Hawaii Health
Surveillance Program. In most cases, members who marked more than
one race or ethnicity were categorized as “mixed.”
Any member who marked Hawaiian, however, was classified as
Hawaiian regardless of what other categories he or she may have
marked, because so few persons are of Hawaiian-only ancestry. We
examined ethnic differences in oral adherence for the six main
ethnic/racial groups in Hawaii: Japanese, Chinese, whites,
Hawaiians, Filipinos, and Koreans. Members who marked other
ethnicities, no ethnicity, or were mixed but non-Hawaiian were
excluded from the analyses.
Calculation of adherence
Treatment adherence for this study was calculated allowing
small gaps between prescriptions. The maximum allowable gaps were
based upon possession ratios, calculated as follows:
Possession ratio = days supplied for first
prescription/(fill date of second prescription − fill date of first
A claim separated from a previous claim with a possession ratio of 0.8 or
greater was considered adherent.
Days of adherence were calculated from the date of the first
prescription until the end-of-supply of the last claim. For
isolated claims, claims not within a possession ratio of 0.8 of
other claims, the days of adherence were taken as the days of
supply on the claim. The days of adherence for other claims were
calculated as follows:
Days of adherence = date of last adherent prescription − date of first
prescription + days of supply on last compliant prescription
The average days of adherence per year were calculated as the
days of adherence divided by the days of enrollment in a drug
plan since the first prescription date. Patients needed to have
some drug enrollment and medical enrollment to be in the study.
They did not need to be continuously enrolled for the entire study period.
Days of adherence were looked at and adjusted by days of
enrollment on an annual basis. The average number of days of
enrollment per year was 354 days.
Multivariate logistic regression analysis was used to examine
the relationship between adherence and patient characteristics.
All analyses were performed using Stata V7.0 statistical software
(StataCorp, College Town, Tex).
Back to top
Among the 39,536 patients on an oral hypoglycemic agent,
20,685 unique numbers met the inclusion criteria for the study
(Table 1). The mean age (± SD) of the study sample was 63.7
years (± 12.01). Of the study sample, 54.4% were male and 45.6% were
female. Among the study sample, 20.7% had high morbidity; 77.3%
lived on Oahu, 12.1% on Hawaii, 5.8% on Kauai, 6.3% on Maui, 0.4%
on Lanai, and 0.7% on Molokai. The distributions of oral
hypoglycemic agents were as follows: sulfonylurea 44.0%, metformin 31.2%,
a-glucosidase inhibitors 1.6%, and
thiazolidinedione 23.2%. During the study period, 19.9% of
patients also used insulin. Of these patients, 16.0% used
sulfonylurea, 19.8% used metformin, 31.0% used a-glucosidase
inhibitors, and 33.0% used thiazolidinedione. Patients on thiazolidinedione and a-glucosidase inhibitors have a higher
percentage of insulin use.
Adherence differed according to drug class, age, ethnicity,
and island of residence (Table 2). Overall adherence to oral
hypoglycemic agents was low at 61.4%. Relative to sulfonylurea,
the odds ratio of adherence was highest for metformin, followed
by thiazolidinedione, and lowest for
Relative to the age subset 55 to 64 years, adherence increased
with older patients, reaching a peak at age 74, then decreased
for patients aged 85 and older. Age was controlled for because metformin and thiazolidinedione are less likely to be used among
the elderly due to contraindications and adverse effects.
Japanese patients were the most likely to be adherent, followed
by Chinese, whites (referent group), Hawaiians, and Filipinos.
Relative to Oahu, Lanai had the lowest odds ratio.
Compared with the fee-for-service members, HMO members were less
likely to adhere to medication therapy. Adherence of members with
Medicare coverage was similar to that of members in the
fee-for-service plan. Members with the lowest morbidity level
(i.e., fewest comorbid conditions) tended to be the least adherent to
medication. Adherence tended to improve with increased morbidity but declined
for those with the highest morbidity level. Gender and the year of treatment
were not significantly associated with adherence.
Back to top
Patient adherence to oral hypoglycemic agents is integral to
reducing the health care costs and chronic complications of
diabetes. Identifying which patients are at greatest risk for nonadherence to oral hypoglycemic agents is an important first
step toward developing interventions that improve adherence.
Older patients had better adherence than younger patients to
oral hypoglycemic agents, with adherence declining after age 85. A possible explanation for the better adherence among the
older patients is that they are more knowledgeable and
experienced with using the medications. However, with increasing
age and burden of disease, adherence becomes more difficult to
maintain over time. The finding of an association between age and
adherence to oral hypoglycemic agents has helped to identify
younger diabetic patients as a susceptible group requiring
intervention to improve adherence. The diagnosis of diabetes made
at an earlier age means these individuals will have an increased
duration of exposure to hyperglycemia and as a consequence
increased severity of microvascular complications (16). In
addition, recent trends show an increased incidence of type 2
diabetes in younger individuals belonging to minority groups
(17). This is reflected in current screening guidelines that
recommend diabetes testing earlier (before the age of 45) for
anyone belonging to a high-risk group, such as Asian-Pacific
Ethnicity was also found to be a significant factor in adherence to oral
hypoglycemic agents. Japanese patients had the highest adherence, and Filipino
patients had the lowest adherence. Worrisome about this finding is that
Filipinos in Hawaii have the highest prevalence of diabetes among the four
largest Asian groups, which include Chinese, Japanese, Filipino and Korean (18).
Among Filipino males living in Hawaii, diabetes is the third leading cause of
death, and in Filipino females it is the second leading cause of death (19).
More research is needed to understand the risk factors that contribute to
Filipino susceptibility to developing diabetes and the reasons behind the low
rates of adherence to oral hypoglycemic therapy in this group.
More studies also need to be done on patients living on the island of Lanai
to find out the reasons behind the low rates of adherence. One possibility to
consider is the rural setting of the island, which poses barriers to health care
access for persons with diabetes.
This study did not examine the effects on adherence of
combination drug therapy either with or without concomitant
insulin use. Combination therapy has been shown to be associated
with poorer rates of adherence than monotherapy (20-21). In
clinical practice, sulfonylurea and metformin are usually
prescribed by themselves as a first-line therapy because these
two drugs are the only ones to demonstrate decreased vascular
risk. Furthermore, although thiazolidinedione and
a-glucosidase inhibitors are indicated for use as monotherapy,
they are used more in combination therapy, thereby contributing
to the decreased rates of adherence in these two medications.
Insulin has been associated with lower adherence rates; however,
it would be premature to make any conclusion at this point
without further data on the use of combination therapy in this
Other limitations to our results are that the database for
ethnicity is not complete, patients may have received free
samples from their physicians, and the use of medical refill
claims is an indirect method of measuring adherence. It is not
known if patients are actually ingesting their medications;
however, filling a prescription is a necessary first step for
doing so. Another limitation was the lack of HbA1c levels, which
would have provided a means for linking nonadherence with poor
health outcomes in the population. To be able to link poor health
outcomes with nonadherence in susceptible groups, such as younger
patients and Filipino patients, would strengthen the conclusion
that these groups are at high risk for the chronic complications
of diabetes and need to be targeted for intervention. In
addition, this study did not examine the effects of combination
drug use on adherence.
Despite these limitations, prescription refill claims can be
used as a tool to predict patient characteristics at highest risk
for low adherence. Previous studies that have attempted to
correlate demographic characteristics such as race, age, and
gender with adherence have had inconsistent results (22-24).
However, these studies used a small sample population. This study
used a large sample population, and the trends demonstrated for
race and age were consistent with other studies that used a large
sample population (25-27).
The findings in this study highlight the need to target
younger patients, Filipino patients, and patients living in rural
areas, such as the island of Lanai, for better glycemic control.
Even with adequate access to health care services and drug
coverage, there was low adherence to oral diabetic medications.
This is a disturbing finding considering that the benefit of
intensive glycemic control has been demonstrated by the U.K. Prospective
Diabetes Study (28).
Back to top
Funding for this study was received from Hawaii Medical
Back to top
Corresponding Author: Rachel Lee, John A. Burns School of Medicine, 2015
Mott-Smith Dr, Honolulu, HI 96822. Telephone: 808-533-3125.
Author Affiliations: Deborah A. Taira, ScD, John A. Burns School of Medicine,
Honolulu, Hawaii, Hawaii Medical Service Association.
Back to top
- Brown JB, Nichols GA, Glauber HS, Bakst A.
follow-up of antidiabetic drug use, nonadherence, and mortality
in a defined population with type 2 diabetes mellitus. Clin Ther 1999;21(6):1045-57.
- Venturini F, Nichol M, Sung JCY, Bailey KL, Cody M, McCombs
sulfonylureas in a health maintenance
organization: a pharmacy record-based study. Ann Pharmacother
- Abate N, Chandalia M.
The impact of ethnicity on type 2
diabetes. J Diabetes Complications 2003;17:39-58.
- Fujimoto WY. Diabetes in Asian and Pacific Islander
Americans. In: National diabetes data goup, diabetes in
America (NIH Publication No. 95-1468, 2nd ed.).
Bethesda (MD): National Institute of Diabetes and Digestive and
Kidney Diseases, National Institutes of Health; 1995. pp. 661-81.
- McNeely MJ, Boyko EJ.
Type 2 diabetes prevalence in Asian
Americans: results of a national health survey. Diabetes Care
- Sloan NR.
Ethnic distribution of diabetes mellitus in
Hawaii. JAMA 1963;183(6):419-24.
- Maskarinec G.
Diabetes in Hawaii:
from insurance claims data. Am J Public Health
- Busch J, Easa D, Grandinetti A, Mor J, Harrigan R.
People in Hawaii? An overview of ethnic healthy disparities in
Hawaii for the Healthy People 2010 initiative targeted health
concerns. Hawaii Med J 2003;62(1):10-4.
- Grandinetti A, Chang HK, Mau MK, Curb JD, Kinney EK, Sagum R, et al. Prevalence of
glucose intolerance among native Hawaiians in two rural
communities. Native Hawaiian Health Research (NHHR) Project. Diabetes Care 1998;21(4):549-54.
- Huang B, Rodriguez BL, Burchfiel CM, Chyou P, Curb JD,
prevalence of diabetes among
Japanese-American men in Hawaii. Am J Epidemiol
- Burchfiel CM, Curb JD, Rodriguez BL, Yano K, Hwang LJ, Fong KO, et al. Incidence and
diabetes in Japanese-American men. The Honolulu Heart Program. Ann
- Mau KM, Glanz K, Severino R, Grove JS, Johnson B, Curb
lifestyle behavior change in native Hawaiians.
Initial findings from the Native Hawaiian Diabetes Intervention
Program. Diabetes Care 2001;24(10):1770-5.
- Wang C, Abbott L, Goodbody A, Hui W, Rausch C. Development of a community-based diabetes management program for
Pacific Islanders. Diabetes Educ 1999;25(5):738-46.
- Schectman JM, Mohan MN, Voss JD.
between diabetes metabolic control and drug adherence in an indigent population. Diabetes Care 2002;25(6):1015-21.
- Starfield B, Weiner J, Mumford L, Steinwachs D. Ambulatory care groups: a categorization of diagnoses for
research and management. Health Serv Res 1991;26(1)53-74.
- Harris MI, Eastman RC, Cowie CC, Flegal KM, Eberhardt MS.
Racial and ethnic differences in glycemic control of adults with
type 2 diabetes. Diabetes Care 1999;22:403-8.
- Rosenbloom AL, Joe JE, Young RS, Winter WE.
epidemic of type 2 diabetes in youth. Diabetes Care
- Carter JS, Pugh JA, Monterrosa A.
diabetes mellitus in minorities in the United States. Ann Intern
- Braun KL, Yang H, Onaka AT, Horiuchi BY.
Hawaii: ethnic variations in life expectancy and mortality, 1980
and 1990. Hawaii Med J 1996;55(12):278-83.
- Dailey G, Kim MS, Lian JF.
Patient compliance and
persistence with antihyperglycemic drug regimens: evaluation of a
Medicaid patient population with type 2 diabetes mellitus. Clin
- Melikian C, White TJ, Vanderplas A, Dezii CM, Chang E.
Adherence to oral antidiabetic therapy in a managed care
organization: a comparison of monotherapy, combination therapy,
and fixed-dose combination therapy. Clin Ther
- McDonald HP, Garg AX, Haynes BR.
patient adherence to medication prescriptions. JAMA
- Steiner JF, Prochazka AV.
assessment of refill
compliance using pharmacy records: methods, validity, and
applications. J Clin Epidemiol 1997;50:105-16.
- Vermiere E, Hearnshaw H, Royen PV, Denekens J.
adherence to treatment: three decades of research. A
comprehensive review. J Clin Pharm Ther 2001;26:331-42.
- Schectman JM, Bovbjerg VE, Voss JD.
medication-refill adherence in an indigent rural
population. Med Care 2002;40:1294-300.
- Monane M, Bohn RL, Gurwitz JH, Glynn RJ, Levin R, Avorn J.
antihypertensive therapy among elderly Medicaid
enrollees: the roles of age, gender, and race. Am J Public Health
- Charles H, Good CB, Hanusa BH, Chang CC, Whittle J.
differences in adherence to cardiac medications. J Natl Med Assoc
- UK Prospective Diabetes Study (UKPDS) Group.
blood-glucose control with sulphonylureas or insulin compared
with conventional treatment and risk of complications in patients
with type 2 diabetes (UKPDS 33). Lancet 1998;352(9131):837-53.
Back to top