Preliminary clinical outcomes and adoption of continuous glucose monitoring following reimbursement implementation in patients with type 1 diabetes in Thailand
Article information
Abstract
Purpose
Continuous glucose monitoring (CGM) is recommended by clinical guidelines for children and adults with type 1 diabetes mellitus (T1DM) to improve clinical outcomes. In Thailand, CGM was incorporated into the Universal Healthcare Coverage (UHC) program in mid-2023. This study aimed to evaluate preliminary clinical outcomes and device adoption at a single tertiary care center. Glycemic outcomes were assessed before and after CGM use following the UHC reimbursement program and results were compared across 4 groups: self-monitoring blood glucose, CGM, open-loop insulin pump, and hybrid closed-loop (HCL). CGM adherence and parameters were also analyzed.
Methods
This retrospective-prospective study collected and analyzed demographic data, hemoglobin A1c (HbA1c) levels, and CGM parameters.
Results
A total of 142 T1DM patients (median age, 17.3 years; range, 3.5–69.2 years) were included. Baseline HbA1c was 8.1%±1.5%, with no significant differences among groups (P=0.223). The HCL group showed the largest HbA1c reduction at 12 months (-0.99%, P= 0.001), particularly in patients <18 years (-1.21%, P=0.014). CGM users showed improvements in HbA1c (-0.29%) and a higher proportion achieving time in range (TIR) ≥70% at 12 months (69.2% vs. 47.1%, P=0.08), though this was not statistically significant. Preliminary CGM uptake was 12% (17 of 142). The HCL group exhibited higher TIR and better sensor adherence (P<0.05), while other groups showed no significant changes.
Conclusions
The HCL system significantly improved glycemic outcomes, particularly in younger patients. However, CGM adoption remains low, highlighting the need for expanded access, enhanced reimbursement policies, and improved adherence strategies.
Highlights
· Continuous glucose monitoring implementation improved glycemic outcomes in type 1 diabetes mellitus, with the greatest benefit observed in hybrid closed-loop users. However, despite national coverage, uptake remains limited due to ancillary device costs, highlighting the need for expanded and comprehensive reimbursement policies.
Introduction
Type 1 diabetes mellitus (T1DM) is a lifelong chronic condition that can manifest at any stage of life, including childhood, adolescence, and adulthood. Poor glycemic control in individuals with T1DM can lead to the early onset of diabetes-related complications, significantly impacting quality of life and long-term health outcomes. In Thailand, approximately 72% of individuals diagnosed with diabetes before the age of 20 have T1DM [1]. In 2021, an estimated 8.4 million individuals worldwide were living with T1DM, with 0.5 million new cases diagnosed that year [2]. The global burden of T1DM is vast and projected to increase significantly by 2040, with prevalence expected to rise by 60%–107%. The largest relative increase is anticipated in low- and lower-middle-income countries, emphasizing the urgent need for improved diabetes management strategies and expanded access to care, particularly in resource-limited settings [2].
The current standard of care for T1DM includes intensive insulin therapy, structured educational programs, frequent self-monitoring of blood glucose (SMBG), and psychological support. However, traditional SMBG has several limitations, including the need for frequent fingerstick checks (typically 4–8 times per day), which is particularly challenging for children and adolescents. SMBG provides only single-point-in-time glucose measurements, failing to capture glucose trends or rates of change, and requires a high level of patient compliance. Additionally, SMBG often fails to detect critical events such as nocturnal or asymptomatic hypoglycemia, increasing the risk of complications [3]. To address these challenges, the American Diabetes Association 2025 guidelines recommend that continuous glucose monitoring (CGM) should be offered at diagnosis or as soon as feasible for people with any type of diabetes who use insulin, provided they can safely use the device [4]. Beyond CGM, hybrid closed-loop (HCL) insulin pump systems have emerged as a promising technology that integrates CGM with automated insulin delivery [5]. Recent evidence suggests that real-time CGM usage and HCL insulin pump systems contribute to a reduction in hemoglobin A1c (HbA1c) levels, an increase in time in range (TIR), and a lower risk of hypoglycemia, particularly among adolescents and young adults [6,7]. However, access to both CGM and HCL systems remains limited in many regions, restricting their potential benefits for glycemic management. As a middle-income developing country, Thailand is striving to expand access to advanced diabetes technologies through Universal Healthcare Coverage (UHC), ensuring that patients with T1DM can benefit from CGM without an additional financial burden. CGM was integrated into the UHC program in mid-2023.
This study aims to evaluate preliminary clinical outcomes and CGM adoption following the implementation of the UHC reimbursement program in Thailand. Specifically, we assess glycemic outcomes before and after CGM use, comparing results across 4 groups: SMBG with multiple daily injections (MDI), CGM with MDI, open-loop insulin pump, and HCL system. Additionally, CGM adherence and parameters were evaluated to provide insights into real-world device usage and effectiveness. Despite the potential benefits of CGM, several barriers to its widespread adoption remain, including device availability, healthcare provider familiarity, patient education, affordability of accessories, and adherence challenges.
Materials and methods
This study utilized a retrospective and prospective analysis of electronic medical records from patients with T1DM seen at both the Pediatric and Adult Diabetes Clinics at King Chulalongkorn Memorial Hospital, Bangkok, Thailand. The study was approved by the Institutional Review Board (IRB), Faculty of Medicine, Chulalongkorn University (IRB No. 0735/67).
1. Study population
T1DM patients were identified using International Classification of Diseases, Tenth Revision code E109 (T1DM without complications) between January 1, 2023 and July 1, 2024. Eligible patients met the following criteria: (1) T1DM diagnosis with ongoing follow-up, (2) HbA1c ≤12%, (3) use of SMBG or at least one month of CGM usage.
Patients with known or documented severe or uncontrolled psychiatric disorders at baseline were excluded from analysis.
2. Data collection
Baseline characteristics included: (1) Demographics: age, sex, weight, height, body mass index (BMI), age at diabetes onset, and disease duration. (2) Diabetes management: insulin administration method and glucose monitoring type. (3) Clinical history: number of severe hypoglycemia episodes (requiring third-party assistance) and hospitalizations for diabetic ketoacidosis (DKA) or severe hypoglycemia in the 12 months before enrollment and during follow-up
For SMBG users, baseline data were retrospectively collected 12 months prior to the most recent clinic visit. For CGM or insulin pump users, baseline data were recorded at the time of CGM initiation, insulin pump start, or glucose monitoring transition.
3. Clinical and CGM data collection
HbA1c levels were assessed at baseline, 3, 6, and 12 months.
CGM parameters were analyzed at 1–2 weeks, 1 month, 3 months, 6 months, and 12 months postinitiation, including: (1) percent TIR (70–180 mg/dL), (2) percent time below range (TBR) (<70 mg/dL), (3) percent time above range (TAR) (>180 mg/dL), (4) coefficient of variation (%CV), (5) glucose management indicator (GMI), (6) sensor wear time.
For this cohort, the CGM device used was the Medtronic MMT-7040 Guardian 4 glucose sensor (Medtronic MiniMed, Inc., USA), and the HCL insulin pump system used was the MiniMed 780G (Medtronic MiniMed, Inc.).
4. Statistical analysis
Descriptive statistics were used to summarize baseline characteristics. Continuous variables were expressed as mean±standard deviation or median (interquartile range, IQR) depending on data distribution. Shapiro-Wilk test was used to assess normality prior to applying parametric tests. Comparisons of continuous data were reported as median among 4 groups and were performed using the Kruskal-Wallis test for multiplegroup comparisons and the Wilcoxon rank-sum test for 2-group comparisons. Comparisons of continuous data were reported as mean among 4 groups and were performed using 1-way analysis of variance for multiplegroup comparisons and the 2-sample independent t-test. Categorical variables were reported as frequency (percentage) and analyzed using the chi-square test or Fisher exact test, as appropriate. For HbA1c changes over time, a generalized estimating equation (GEE) population-averaged model with a linear function was used to compare trends among different glucose monitoring groups. We did not apply formal post hoc corrections for multiple comparisons (e.g., Bonferroni) due to the exploratory nature of the subgroup analyses and small sample sizes. This has been acknowledged as a study limitation. Comparisons of TIR, %CV and GMI within groups used paired t-tests and among groups used 2-sample independent t-test. Comparisons of TBR and TAR within groups used Wilcoxon sign-rank test and Wilcoxon rank-sum test among groups. All statistical analyses were performed using Stata 18.5 (StataCorp LLC, USA), with a P-value < 0.05 was considered statistically significant.
5. Missing data handling
For longitudinal analyses (e.g., HbA1c and CGM parameters), we used a GEE model, which is robust to missing data under the missing-at-random assumption and does not require imputation. For cross-sectional variables, we performed available case analysis and reported the specific 'N' used.
Results
A total of 142 patients with T1DM from a single tertiary care center were included and categorized into 4 groups based on glucose monitoring and insulin delivery methods: SMBG (n=109), CGM (n=17), open-loop insulin pump (n=4), and HCL system (n=12) (Table 1). In the SMBG group, 97.2% of patients used MDI, while 2.8% followed a conventional insulin regimen. In the CGM group, all patients used MDI. The open-loop group (N=4) had an equal distribution of glucose monitoring methods, with 2 patients using CGM and 2 patients using SMBG.
Baseline characteristics of study participants
The median age of participants was 17.3 (IQR, 12.7–23.9; range, 3.5–69.2) years, with significant differences among groups (P=0.001). The CGM group had the highest median age (28.3 years), while SMBG users had the lowest (16.3 years). Median age at diagnosis was 9 (IQR, 6–12) years, and median diabetes duration was 6.5 (IQR, 3.0–12.3) years, with no significant differences between groups (P=0.442).
For participants ≥18 years, mean BMI was 23.3±4.2 kg/m², with no significant differences between groups (P=0.670). Among participants ≤18 years, the median BMI z-score was -0.14 (IQR, -0.88 to 1.76), also showing no significant variation between groups (P=0.419). The sex distribution was balanced (47.2% female, 52.8% male).
Regarding diabetes-related complications, 5.6% (n=8) of patients had a prior DKA hospitalization, which decreased to 3.5% (n=5) during the study (P=0.739). No patients had a history of severe hypoglycemia in the 12 months prior to the study; however, 2 CGM users (11.8%) experienced severe hypoglycemia events during follow-up (P=0.032). These cases were associated with newly emerging psychiatric conditions and reduced food intake, which likely contributed to the events.
The mean baseline HbA1c was 8.1%±1.5%, with no significant differences between groups (P=0.223) (Table 1). Changes in HbA1c over time across the different glucose monitoring groups are presented in Table 2 and Supplementary Fig. 1. The SMBG group demonstrated a slight increase in HbA1c over time, with a mean change of 0.20 (95% confidence interval [CI], 0.02–0.38). In contrast, the CGM group showed a reduction in HbA1c (-0.29; 95% CI, -0.70 to 0.11), although this change was not statistically significant (P=0.096). The open-loop group exhibited minimal HbA1c change over time (-0.05; 95% CI, -0.40 to 0.30; P=0.657). The HCL group demonstrated the most significant HbA1c reduction, with a mean change of -0.99 (95% CI, -1.32 to -0.67; P=0.001) compared to the other groups.
Changes in hemoglobin A1c over time across different treatment modalities
The proportion of patients achieving HbA1c <7% at different time points is presented in Supplementary Table 1 and Supplementary Fig. 1. In the HCL group, the proportion of patients with HbA1c <7% increased from 8.3% at baseline to 33.3% at 6–12 months, reflecting a substantial improvement in glycemic control. The CGM group also showed an increase in the proportion of patients with HbA1c <7%, though this change was not statistically significant. In contrast, no meaningful changes were observed in the SMBG or open-loop groups, indicating that traditional blood glucose monitoring alone may not be as effective in improving glycemic outcomes.
Age-stratified analysis in Table 3 further highlighted the impact of automated insulin delivery systems on glycemic control. Among patients younger than 18 years, the HCL group demonstrated the largest HbA1c reduction (-1.21; 95% CI, -1.75 to -0.67; P=0.014), suggesting that younger individuals may benefit the most from automated insulin delivery systems. Patients aged 18 years or older in the HCL group showed a similarly significant reduction (-0.83; 95% CI: -1.28 to -0.38; P=0.007). The CGM group also exhibited a slight decrease in HbA1c in adults (-0.38; 95% CI: -0.86 to 0.11), though this change was not statistically significant (P=0.115).
Changes in hemoglobin A1c over time stratified by age group (<18 years vs. ≥18 years)
The comparison of CGM parameters between CGM alone users and HCL system users showed a trend toward improved glycemic control in the HCL group, although most differences were not statistically significant (Table 4). TIR increased in both groups, reaching 74.4% in HCL users and 70.7% in CGM users at 12 months, while TBR remained consistently lower in the HCL group (1% at 12 months, P=0.032 at 6 months). TAR showed a downward trend, decreasing to 22% in HCL users at 12 months, though differences between groups were not significant. CV and GMI remained comparable between the groups throughout the study. Overall, the HCL system demonstrated potential benefits in stabilizing glucose levels, particularly by reducing hypoglycemia and hyperglycemia exposure, but further larger studies are needed to confirm statistical significance.
Comparison of CGM parameters between CGM and HCL system groups
The proportion of CGM users achieving TIR ≥70% showed a non-statistically significant trend toward improvement at 12 months (69.2% vs. 47.1%, P=0.08), while HCL users showed a steady increase from 41.7% to 75%, though not statistically significant (P=0.50) (Table 5). CGM users initially experienced a dip in TIR at month 3 (31.3%) before improving, whereas HCL users demonstrated a more stable trend. These findings suggest that CGM use alone can lead to significant TIR improvement over time, while HCL systems may provide more consistent glycemic control.
Proportion of participants with TIR ≥ 70% over time
Sensor wear time and auto mode utilization were higher in HCL users compared to CGM users, with significant improvements over time (Supplementary Table 2). At baseline, sensor wear time was similar between groups (CGM: 73.4% vs. HCL: 75.3%, P=0.826), but by 6 months, HCL users showed a significant increase to 89.4% (P=0.034), while CGM users remained stable (69.8%–76.0%, P>0.05). By 12 months, sensor wear time remained higher in HCL users (83.6%), though this difference was not statistically significant (P=0.335). Auto mode utilization in the HCL group started at 57.4% at baseline, increased significantly to 90.8% at 3 months (P=0.008), and remained high at 84.6% at 12 months, suggesting consistent engagement with automated insulin delivery. These findings highlight the potential of HCL technology in enhancing sensor adherence and optimizing glycemic control through sustained auto mode use.
Discussion
This study emphasizes the value of the CGM and HCL systems in T1DM management. HCL systems led to the largest HbA1c reductions, particularly in younger individuals. These findings are consistent with real-world data, demonstrating that HCL systems significantly improve glycemic control, enhance metabolic outcomes, and improve quality of life for children and adolescents with T1DM, even among those with low therapeutic engagement [8].
Regarding glycemic control, our TIR findings were similar to previous reports, which showed an increase in TIR from 62.8% to 75.3% [8], and from 69.1% to 76.9% over the course of a year of HCL system use [9]. Similarly, our observed HbA1c reduction was comparable to reports demonstrating a decrease from 7.4% to 6.9% [9] and from 7.6% to 7.1% over the course of one year [10]. While HCL users had consistently higher TIR and lower TAR at all follow-up points, these differences were not statistically significant, likely due to the small sample size, which was a limitation of this study. We also acknowledge that the small subgroup sizes—particularly the open-loop group (n=4)—limit the reliability of estimated means, confidence intervals, and statistical power. However, as this was a real-world retrospective study, the group sizes reflected actual clinical practice following reimbursement implementation. Despite this, the observed trends are encouraging.
Beyond clinical effectiveness, HCL systems have demonstrated cost-effectiveness in T1DM management. A 2023 cost-utility analysis in Singapore found that switching from MDI with intermittently scanned CGM to the MiniMed 780G advanced HCL system resulted in a gain of +1.45 quality-adjusted life years and cost savings of SGD 25,465 (USD 18,723) due to reduced diabetes complications [11]. Similar cost-effectiveness analyses in European countries support the economic viability of integrating advanced diabetes technologies into national healthcare systems [12-15].
Although the CGM group demonstrated a modest HbA1c reduction, the trend suggested potential long-term glycemic benefits, consistent with prior studies. A 2020 randomized trial in adolescents and young adults found that CGM use decreased HbA1c from 8.9% to 8.5% over 26 weeks, while the SMBG showed no improvement [7]. Similarly, the 2021 SWEET study reported significantly lower HbA1c levels in children and adolescents with T1DM using CGM compared to those without sensor use [6]. Government-subsidized CGM access for young people with T1DM has also been shown to be cost-effective compared to a user-funded model in Australia [16]. In summary, the HCL system consistently demonstrated the most significant improvement in HbA1c levels, particularly in younger patients (<18 years). CGM users showed a trend toward improved glycemic control over time, emphasizing the potential benefits of advanced glucose monitoring technologies and automated insulin delivery systems in optimizing T1DM management. However, ongoing education and support are essential to help patients maintain auto mode use above 90%–95%, which is important for optimizing pump therapy and achieving maximal clinical benefit.
Despite the proven advantages of CGM and HCL systems, adoption of diabetes technology remains suboptimal in Thailand. In our study, only 12% of patients adopted CGM, a stark contrast to Australia, where CGM uptake surged from 5% to 79% following the introduction of universal subsidized funding for individuals under 21 [17]. Several barriers to CGM adoption in Thailand persist, consistent with challenges reported in previous studies [18]. Financial constraints remain a primary concern, as the current reimbursement program covers sensors but excludes transmitters, placing a financial burden on patients and limiting long-term CGM use. Additionally, the complex reimbursement process, involving multiple administrative steps, poses challenges for both patients and healthcare providers.
Healthcare professionals also face challenges in CGM utilization due to limited training and familiarity with diabetes technology. Moreover, children and adolescents expressed concerns related to body image, social stigma, skin irritation (exacerbated by humid climates), information and alarm overload, and activity disturbances, all of which negatively impact CGM uptake and adherence. Addressing these barriers through policy enhancements, improved reimbursement structure, provider training, and patient support programs is crucial to optimizing diabetes technology adoption in Thailand.
1. Challenges in implementing diabetes technology in low- and middle-income countries
The integration of diabetes technology in low- and middle-income countries (LMICs) faces unique challenges, including health policy constraints, financial sustainability, and disparities in access [19,20]. Thailand, as a middle-income country, is making strides in expanding UHC to include advanced diabetes technology, but financial and logistical hurdles remain. Global experiences suggest that robust reimbursement policies, physician training, and patient-centered education programs are crucial to improving adoption and adherence [21].
Moreover, healthcare digitalization in LMICs presents infrastructural and regulatory challenges. World Health Organization reports indicate that successful technology implementation requires comprehensive digital health policies, data integration frameworks, and workforce capacity building. Artificial intelligence-driven diabetes management has the potential to revolutionize care delivery, but equity concerns, data privacy issues, and implementation barriers in resource-limited settings must be addressed.
2. Reimbursement schemes for CGM and HCL systems: a global perspective
Reimbursement policies for CGM and HCL systems vary significantly across countries, influencing adoption rates and accessibility. Countries with comprehensive reimbursement models have seen higher uptake and better diabetes management outcomes, underscoring the importance of policy advancements in Thailand.
1) Global reimbursement models
In Australia, CGM reimbursement expanded significantly after the government introduced full subsidies for individuals under 21 years old, leading to an increase in CGM usage from 5% to 79% [17]. The United Kingdom follows the National Institute for Health and Care Excellence guidelines [22]. The National Health Service provides CGM to all individuals with T1DM. Children and young people are typically offered real-time CGM, while adults may choose between real-time CGM or flash monitoring. HCL systems are being introduced gradually over 5 years, prioritizing those with the greatest need, including all children and young people with T1DM, and adults with HbA1c levels of 58 mmol/mol or higher, those experiencing severe hypoglycemia despite optimal management, and individuals who are pregnant or planning pregnancy [23].
In Japan, reimbursement policies prioritize CGM coverage for individuals using insulin pumps, while standalone CGM users face stricter eligibility criteria. In South Korea, despite nationwide reimbursement for CGM devices and insulin pumps, the adoption rate of CGM among the total T1DM population remains low. A recent study identified several barriers to CGM uptake, including age-related disparities in utilization and a lack of proactive promotion and education, suggesting that broad reimbursement alone has not overcome barriers to widespread adoption [24]. While CGM costs in Singapore were previously not reimbursed for individuals with T1DM or type 2 diabetes mellitus [25], recent changes have introduced partial subsidies for children and adults with T1DM [26]. In Taiwan, reimbursement policies and pricing rules for CGM and automated insulin delivery systems are being revised continuously to improve patient access to advanced medical devices [27]. Regional disparities in CGM adoption highlight the need for tailored reimbursement policies. The Asia- Pacific Diabetes Care Advisory Board emphasized that while CGM is becoming the new standard for diabetes management, adoption in Asia-Pacific countries remains inconsistent due to financial constraints, healthcare infrastructure, and policy differences. This underscores the necessity for country-specific strategies to optimize CGM accessibility and adherence, particularly in resource-limited settings [25].
3. Thailand: the need for policy advancement and innovation support
While Thailand has made progress by incorporating CGM into the UHC program, the current reimbursement scheme only covers sensors, which presents a financial barrier for patients. Limited physician training and awareness of diabetes technology further hinder adoption, along with patient-perceived barriers such as social stigma, skin irritation, and information overload.
Moving forward, Thailand must develop more inclusive reimbursement initiatives. Potential policy advancements include: (1) Expanding CGM and insulin pump coverage to include associated supplies under UHC [28]. (2) Introducing tiered reimbursement for HCL systems based on HbA1c levels or patient risk factors to optimize resource allocation. (3) Providing healthcare provider training programs to enhance diabetes technology literacy. (4) Addressing patient concerns through education and support programs. (5) Encouraging local innovation by supporting Thai-developed diabetes technology solutions, ensuring sustainable and cost-effective options for the healthcare system.
Additionally, with rapid advancements in diabetes technology and new CGM and insulin pump brands entering the market, there is hope that greater competition and more choices will reduce cost in the future, making these life-changing technologies more accessible to a broader population. By balancing policy reforms and innovation support, Thailand can enhance diabetes care accessibility and improve long-term health outcomes for individuals with T1DM.
In conclusion, our findings reinforce the efficacy of HCL systems in optimizing glycemic outcomes and highlight the need for improved access to diabetes technology in Thailand. Addressing reimbursement constraints, healthcare provider training, and patient-perceived barriers is critical to maximizing the impact of CGM and HCL systems. Future research should focus on long-term cost-effectiveness analyses and policy frameworks that facilitate wider adoption and sustainability of these technologies in Thailand and other LMICs. As diabetes technology advances, broader adoption and lower costs could improve long-term outcomes for T1DM patients.
Supplementary materials
Supplementary Tables 1-2 and Supplementary Fig. 1 are available at https://doi.org/10.6065/apem.2550096.048.
Proportion of participants with hemoglobin A1c <7% at different time points
Sensor wear time and auto mode utilization over time
Hemoglobin A1c (HbA1c) change over time across different treatment modalities. SMBG, self-monitoring of blood glucose; CGM, continuous glucose monitoring; lb/ub, lower bound/upper bound.
Notes
Conflicts of interest
There are no potential conflicts of interest relevant to this article to report.
Funding
This study received no specific grants from any funding agency in the public, commercial, or not-for-profit sectors.
Data availability
The data that support the findings of this study can be provided by the corresponding author upon reasonable request.
Acknowledgments
We would like to thank Mrs. Jiratchaya Sophonphan and Ms. Suphab Aroonparkmongkol for their assistance in this study.
Author contribution
Conceptualization: NY, NN, WL, NL, TS; Data curation: NY, NN, WL, JJ, NL, TS; Formal analysis: NY, NN, WL, PP, TS; Methodology: NY, NN, WL, TS; Project administration: NY, PP, JJ, NL, TS; Visualization: TS; Writing - original draft: NY, TS; Writing - review & editing: NN, WL, PP, JJ, NL, TS