The long-term effectiveness of glycemic control in a diabetes case management
program and the challenges of diabetes care in Taiwan
Chih-Cheng Hsu, MD, DrPH1,2, Tong-Yuan Tai, MD, PhD3,4
1
Division of Geriatrics and Gerontology, Institute of Population Health Sciences,
National Health Research Institutes, Zhunan, Taiwan
2
Department of Health Services Administration, China Medical University, Taichung,
Taiwan
3
Taipei Jen-Chi Relief Institution, Taipei, Taiwan
4
Formosan Diabetes Care Foundation, Taipei, Taiwan
Running title: effectiveness and challenges of diabetes care in Taiwan
Corresponding author
Chih-Cheng Hsu, MD, DrPH
Division of Geriatrics and Gerontology, Institute of Population Health Sciences,
National Health Research Institutes, 35 Keyan Road, Zhunan, Miaoli County, Taiwan
Tel: +886-37-246-166 ext. 36336
Fax: +886-37-586-261
Email: cch@nhri.org.tw
1
Abstract
Aims
This study aims to evaluate the long-term effects on glycemic control of a domestic
diabetes care program focusing on case management and to discuss the challenges in
quality of diabetes care in Taiwan.
Methods
We randomized (clinic-based) 1060 study subjects recruited from the DMIDS
(Diabetes Management through an Integrated Delivery System) project in 2003–2005
into the intervention group (n=789 from 27 clinics) and control group (n=271 from 7
clinics). The quarterly self-care and nutrition education program in the intervention
group was delivered by case managers, and usual care was given to the control group.
The intervention program was provided from 2003 to 2007. A multivariable mixed
model analysis was used to assess 3.5-year intervention effects on glycemic control.
Results
The glycated hemoglobin (HbA1c) in the intervention group (especially for those with
baseline HbA1c in 7–9% and >9%) was significantly lower than that in the control
group since the sixth month after recruitment and the intervention effects could
sustain for 3 years.
Conclusions
The current study shows that the case management provided in the pay-forperformance (P4P) diabetes care program in Taiwan is effective to improve glycemic
control for at least 3 years. However, previous research indicated poverty is associated
with not only higher diabetes incidence but inequality of diabetes care in Taiwan
despite universal health coverage. The sicker were also shown less likely to be
enrolled to the P4P diabetes care program. Additional cost-effectiveness studies and
more health policy reform are needed to optimize diabetes care in Taiwan.
2
Introduction
The global prevalence of diabetes mellitus (DM) has been escalating in the past
decade and is expected to continuously increase by about 55% over the next 20 years
[1]. In Taiwan, like elsewhere in the world, DM is a major public health problem. Its
prevalence increased by 48.0% from 2000 to 2008 in spite of a relatively stable ageadjusted incidence rate (by only 2.2% increase) seen for the same period [2]. The fact
that diabetes could incur high mortality (since 1987 it has been the fourth or fifthleading cause of death in Taiwan) and high healthcare costs (4.3 times higher for
diabetes patients compared to those without diabetes) indicates that diabetes imposes
significant health and financial burdens in Taiwanese society [3].
In response to the growing challenges of this devastating disease, the National
Health Insurance Administration in Taiwan has initiated a multidisciplinary diabetes
care program since 2001 aimed at bettering diabetes care [4]. As a result, the annual
check-up rate of all recommended diabetes tests — including glycated hemoglobin
(HbA1c), lipid profile, urine analysis, and fundus examination — has significantly
increased in the past 10 years [2]. However, health providers have been reported less
likely to enroll sicker patients to attend this pay-for-performance (P4P) diabetes care
program [4,5]. A study also showed poverty was associated with inequality of diabetes
care despite the universal coverage policy implemented in Taiwan since 1995 [6].
Furthermore, to the best of our knowledge, there is only one randomized controlled
trial (RCT) demonstrating improved short-term (1-year) glycemic control after
receiving the multidisciplinary management in the designated P4P program [7]. This
study aimed to evaluate the long-term (≥ 3 years) effectiveness of the P4P program on
glycemic control. We also wanted to discuss the current challenges to optimizing
diabetes control in Taiwan.
Methods
We followed the recruited participants in the Diabetes Management through an
Integrated Delivery System (DMIDS) project (NCT00288678 ClinicalTrial.gov) until
December 2007. The project, which was sponsored by the National Health Research
Institutes (NHRI) has been described elsewhere [7]. Briefly, we recruited 1209 type 2
diabetes patients from 40 community clinics during 2003–2005. For the current study,
we excluded participants recruited from clinics with enrollment size fewer than 10
(n=70 from 13 clinics). Those who had incomplete data were also excluded (n=79).
The remaining 1060 study subjects recruited from 27 clinics were selected for further
analysis. By study design, the participants in each clinic were randomized into 2
groups: the usual care group (n=537) and the NHRI-intervention (NHRI-I) group
3
(n=523). The NHRI-I group underwent a quarterly self-care and nutrition education
program delivered by NHRI-hired case managers. However, the physicians in 7 out of
27 clinics decided to hire a local case manager to provide case management for those
(the Local-I group, n=266) originally allocated into the usual care group in their
clinics. Therefore, we had 3 study groups: the NHRI-I group (n=523 from 27 clinics),
the Local-I group (n=266 from 7 clinics), and the usual care group (n=271 from 20
clinics). The case management, carried out from recruitment through the end of 2007,
in both the NHRI-I and Local-I groups followed the same protocol designated by the
NHI Administration [4,7]. The only distinction between these two was the
sponsorship for the case management: the case managers in the NHRI-I group were
hired by NHRI (5 case managers shared by 27 clinics), while the case managers in the
Local-I group were hired by individual clinics (7 case managers serving a total of 7
clinics). All case managers in the NHRI-I and Local-I groups were qualified by the
NHI Administration. For the purpose of analysis, we further regrouped those in the
NHRI-I and the Local-I groups into the intervention group (n=789). As previously
indicated, the baseline characteristics — including demographics and the biochemical
profiles of the intervention and control groups — were comparable [7]. Finally, we
stratified the participants into 3 classes according to their baseline HbA1c for
subgroup analysis: < 7%, 7–9%, and > 9%. A multivariable mixed model analysis was
used to assess intervention effects on HbA1c control. The covariates adjusted in the
models included age, gender, education, DM duration, baseline HbA1c, and region.
The clinic was treated as a random effect. Analyses were performed with SAS
software, version 9.1 (SAS Institute, Cary, NC). A two-sided P value < 0.05 was
considered statistically significant.
Results
Figure 1 (A) shows the HbA1c in the intervention group was significantly lower than
that in the control group since the sixth month after recruitment and the intervention
effects could last for at least 3 years. Figure 1 (B) indicates the levels of glycemic
control in the NHRI-I and Local-I groups were both better than that in the control
group; the HbA1cs after the sixth month in the Local-I group looked even lower,
although not having reached a significant level, than that in the NHRI-I group
throughout the 3-year intervention period. Figure 2 displays results from subgroup
analysis. In general, the glycemic control in the intervention groups, regardless of
their baseline status, was better than in the control group. The significant intervention
effects could last for 2.5 years and 3 years for those with baseline HbA1c > 9% and
7–9%, respectively. However, for those who had good glycemic control at recruitment
(HbA1c < 7%), the intervention did not seem to be able to provide glycemic
4
improvement, except for a significantly lower HbA1c level seen in the 36th month.
Discussion
This is the first and also the only RCT study to date in Taiwan to demonstrate longterm effects on glycemic control in the NHI Administration-initiated P4P diabetes
care program. This study is in accord with and even strengthens the results shown in
our previous, small-scale research project focusing on the short-term (1-year)
glycemic effect of case management for the study subjects recruited in the city of
Kaohsiung [7]. We previously reported the intervention group had greater
improvements than the control in fasting plasma glucose (-13.4 vs. 16.9 mg/dL,
p=0.007) and HbA1c (-0.7% vs. -0.2%, p=0.034) at 12 months. The improvement of
HbA1c in the intervention group (beta=0.100; SE=0.033; p=0.004) could further be
independently explained by the reduction in carbohydrate intake (-31.24 ± 61.53 and
7.15 ± 54.09 grams/day for the intervention and control group, respectively). This
RCT study demonstrates that the current P4P diabetes care program can effectively
improve long-term glycemic control in type 2 diabetes in primary care clinics in
Taiwan through the betterment of diet control.
In addition to implementing the P4P diabetes care program since 2001, the
Ministry of Health and Welfare in Taiwan has been promoting nationwide public
health campaigns aiming at improving diabetes care, such as diabetes-awareness
initiatives, patient self-help organizations, and accreditation for diabetes care
providers [8]. Through such efforts the health system in Taiwan has become
increasingly supportive of diabetes care. As a result, both diabetes patients’ self-care
behaviors and health providers’ accountability have improved over the last decade
[2,9]. That such endeavors are not fully sufficient by themselves, however, is
indicated by the fact that diabetic nephropathy accounts for more than 45% of cases of
end-stage renal disease requiring dialysis in Taiwan [10]. The big challenges we are
facing today and probably for the next decade include the rising prevalence of
diabetes and the disparity in quality of diabetes care.
The barriers of healthcare accessibility were expected to be minimized after
the NHI program implementation in 1995. But the universal and comprehensive
coverage featured in this greatly appreciated nationwide single-payer health insurance
program did not seem to be able to eliminate all concerns about inequality of
healthcare in Taiwanese society. For example, critics often worry that the
reimbursement design of the P4P diabetes care program may disparage the sicker and
the elderly as unwelcome patients for program enrollment [4,5]. In one of our
previous studies [6], we also found that economically disadvantaged people were
more likely to develop type 2 diabetes; their incidence density was estimated to be
5
20.4 per 1000 person-years, which was about 50% more than the general population
(hazards ratio = 1.5; 95% CI = 1.3–1.7). Furthermore, compared to their middleincome counterparts, the poor with diabetes were more likely to get their first-time
DM diagnosis during a course of hospitalization (adjusted odds ratio [aOR] = 2.2, P <
0.001) and less likely to visit diabetes clinics (aOR = 0.4, P < 0.001) after the
diagnosis was established. The poor with diabetes were also less likely than their
middle-income counterparts to be prescribed for the test of HbA1c (aOR = 0.6; 95%
CI = 0.4–0.9), low-density lipoprotein cholesterol (aOR = 0.4; 95% CI = 0.2–0.7),
triglycerides (aOR = 0.5; 95% CI = 0.4–0.8), and retinopathy (aOR = 0.4; 0.2–0.9).
All of these epidemiological features indicate that socioeconomic background is
related to the vulnerability of diabetes development and the quality of diabetes care in
Taiwanese society.
The current study provides some empirical evidence confirming the NHI
Administration-initiated P4P diabetes care program is an effective model for glycemic
control in Taiwan. However, several other studies also indicate that universal
coverage in the NHI program is not a silver bullet on its own. That less than 25% of
diabetes patients are currently enrolled in the P4P diabetes care program indicates a
need to promote it more. Facing the challenges of the rapid growth of diabetes
prevalence, without further reforming health policy, we may not be able to optimize
and equalize quality of diabetes care for all disadvantaged groups.
Acknowledgements
This project was supported by grants (96A1-HDPP08-017) funded by the National
Health Research Institutes of Taiwan. We thank all co-investigators (Drs. MengChuan Huang, Shyi-Jang Shin, Hung-Jen Yang, Chwen-Tzuei Chang, Chih-Jen
Chang, Ching-Hui Loh, Mao-Tsu Fuh, and Hsing-Yi Chang) for their enthusiastic
support. We also extend our deep appreciation to the 40 general practitioners and
1029 diabetes patients who participated in this study for their contributions.
6
Figure legends
Fig. 1
Comparison of mean HbA1c in different study groups in 0–42 months since
recruitment. (A) mean HbA1c in the intervention vs. control group. (B) mean HbA1c
in the NHRI-I, Local-I, and control groups. A multivariable mixed model analysis was
used to assess intervention effects on HbA1c control. The covariates adjusted in the
models included age, gender, education, DM duration, baseline HbA1c, clinic, and
region. * indicates P < 0.05 comparing the intervention group with the control group.
Fig. 2
Comparison of mean HbA1c in the intervention vs. control group in 0–42 months
since recruitment for different subgroups according to baseline HbA1c. (A) for those
with baseline HbA1c < 7%, (B) for those with baseline HbA1c in 7–9%, and (C) for
those with baseline HbA1c > 9%. A multivariable mixed model analysis was used to
assess intervention effects on HbA1c control. The covariates adjusted in the models
included age, gender, education, DM duration, baseline HbA1c, clinic, and region.*
indicates P < 0.05 comparing the intervention group with the control group.
7
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
Whiting DR, Guariguata L, Weil C, Shaw J. IDF diabetes atlas: global
estimates of the prevalence of diabetes for 2011 and 2030. Diabetes Res Clin
Pract 2011;94(3):311–21.
Jiang YD, Chang CH, Tai TY, Chen JF, Chuang LM. Incidence and
prevalence rates of diabetes mellitus in Taiwan: analysis of the 2000–2009
Nationwide Health Insurance database. J Formos Med Assoc
2012;111(11):599–604.
Lin T, Chou P, Lai MS, Tsai ST, Tai TY. Direct costs-of-illness of patients
with diabetes mellitus in Taiwan. Diabetes Res Clin Pract 2001;54 Suppl
1:S43–6.
Chang RE, Lin SP, Aron DC. A pay-for-performance program in Taiwan
improved care for some diabetes patients, but doctors may have excluded
sicker ones. Health Aff (Millwood) 2012;31(1):93–102.
Chen TT, Chung KP, Lin IC, Lai MS. The unintended consequence of
diabetes mellitus pay-for-performance (P4P) program in Taiwan: are patients
with more comorbidities or more severe conditions likely to be excluded from
the P4P program? Health Serv Res 2011;46(1 Pt 1):47–60.
Hsu CC, Lee CH, Wahlqvist ML, Huang HL, Chang HY, Chen L, Shih SF,
Shin SJ, Tsai WC, Chen T, Huang CT, Cheng JS. Poverty increases type 2
diabetes incidence and inequality of care despite universal health coverage.
Diabetes Care 2012;35(11):2286–92.
Huang MC, Hsu CC, Wang HS, Shin SJ. Prospective randomized controlled
trial to evaluate effectiveness of registered dietitian-led diabetes management
on glycemic and diet control in a primary care setting in Taiwan. Diabetes
Care 2010;33(2):233–9.
Health Promotion Administration, Ministry of Health, ROC (Taiwan).
Prevention and Control of Major Chronic Diseases. [Internet]. Available from
http://www.hpa.gov.tw/English/ClassShow.aspx?No=201401280005.
Accessed 15 March 2014.
Li CL, Lin NY, Wang HH, Hsu CC, Hurng BS, Chang HY. A population
study on changes in diabetes self-care behaviors in Taiwan between 2001 and
2005. Prev Med 2010;50:308–9.
United States Renal Data System. Chapter 12. International comparison in the
2012 USRDS Annual Data Report (volum 2): Atlas of End-Stage Renal
Disease in the United States. Available from
http://www.usrds.org/2012/pdf/v2_ch12_12.pdf. Accessed 15 March 2014.
8
Figure 1
(A)
(B)
9
Figure 2
(A)
(B)
(C)
10