Nationwide implementation of a diabetes self-management and network system improves outcomes in type 1 diabetes: real-world evidence from Thailand Authors: Thipaporn Tharavanij, Prapai Dejkhamron, Petch Rawdaree, Supawadee Likitmaskul, Jeerunda Santiprabhob, Chaicharn Deerochanawong, Sirimon Reutrakul, Wannee Nitiyanant Published: 2025 ABSTRACT BackgroundCurrently, there is a lack of real-world studies examining the impact of diabetes self-management education for individuals with type 1 diabetes (T1D) in resource-limited countries. This study aimed to evaluate the clinical outcomes of the nationwide implementation of the Diabetes Self-Management Program and Network System (DSMP-NS) in Thailand in conjunction with Thai Type 1 Diabetes (T1D) and Diabetes Diagnosed Age before 30 years Registry, Care and Network (T1DDAR CN).MethodsDSMP-NS provided free-of-charge intensive insulin treatment, Diabetes Self-Management Education, blood glucose monitoring (BGM) supplies, urine ketone strips, and established a network system across 37 hospitals. A prospective observational study was conducted after the implementation of DSMP-NS. The composite outcome was defined as achieving good glycaemic control (HbA1c meeting the goal: <7.5% in children aged < 18 years or < 7.0% in adults) or a ≥ 0.5% reduction if baseline HbA1c level was above the goal without diabetic ketoacidosis (DKA) and severe hypoglycaemia. Data were analyzed using an intention-to-treat approach.ResultsOf 481 participants (median age of 15.8 years; interquartile range: 12.2, 20.5), 305 individuals (63.4%) were children. At baseline, 74 individuals (16.1%) were obese. A total of 412 participants completed the 12-month follow-up. The composite outcome achievement rose significantly compared with baseline (43.0% vs. 17.6%, p < 0.001). Adults had significantly lower mean HbA1c ± SD levels at the 12th month (8.2 ± 2.0% vs. 8.8 ± 2.3%, p = 0.005). The DKA incidence decreased from 22.1 (95%CI 18.6–26.3) to 10.2 (7.6–13.8) per person-year, p < 0.001. Older age at enrolment and BGM ≥ 3 times/day were positively associated with the composite outcome, whereas baseline obesity showed an inverse association.ConclusionsDSMP-NS implementation for T1D resulted in improved glycaemic control and a reduced occurrence of acute diabetes complications. Older age and frequent BGM monitoring (≥ 3 times/day) are positively associated with favourable outcomes, while baseline obesity was inversely associated.Trial registrationNA (a prospective real-world study).Supplementary InformationThe online version contains supplementary material available at 10.1186/s12902-025-02123-8. Background The prevalence of Type 1 diabetes (T1D) varies among countries, ranging from 1% to 15% of people with diabetes [1]. Thailand has been categorized as having a low incidence of T1D [2, 3]. However, recent nationwide population-based longitudinal data in Thailand have revealed that the crude annual incidence rate in individuals aged under 20 years increased over time from 4.8 to 5.4 per 100,000 from 2015 to 2020 [4]. The American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) suggest using an intensive insulin regimen and performing capillary blood glucose monitoring (BGM) pre-meals and additional BGM to prevent and detect hypoglycaemia for most adults with T1D [5]. For children, the ADA recommends monitoring glucose levels up to 6–10 times daily [6]. Although insulin analogs (e.g., glargine, aspart, lispro) are included in Thailand’s national list of essential medicines for T1D, a study conducted from 2005 to 2016 showed that only half of people with T1D received intensive insulin treatment (IIT) [7, 8]. Furthermore, because BGM supplies were not reimbursable, 40% of them performed BGM three or more times daily, while 17% did not monitor at all [7, 8]. Consequently, only 16% of people with T1D achieved their glycaemic goals [8]. Numerous obstacles, including the need to pay out-of-pocket for BGM supplies, a shortage of diabetes educators to provide diabetes self-management education and support (DSMES), and the absence of reimbursable diabetes education services, contributed to this situation. Several studies, such as Dose Adjustment for Normal Eating (DAFNE) [9] and the Diabetes Teaching and Treatment Program (DTTP) [10, 11]have shown the efficacy of DSMES programs [12]. However, real-world studies of such programs in resource-limited countries, including Thailand, are currently lacking. The Diabetes Self-Management Program and Network System (DSMP-NS) for T1D was first launched by the Diabetes Association of Thailand under the Patronage of Her Royal Highness Princess Maha Chakri Sirindhorn (DAT) in 2011. This program integrates DSMES with the support of medical supplies and a networking system. The initial results revealed successfully enhanced diabetes self-management skills for individuals with T1D and their families [13]. Following this success, the Thai-Type 1 Diabetes and Diabetes Diagnosed Age Before 30 Years Registry, Care and Network (T1DDAR CN) was established in 2014 to leverage the standard of care for T1D in 31 hospitals [7]. Preliminary results have demonstrated the program’s cost-effectiveness [14]. Recognizing its impact, the National Health Security Office (NHSO), the provider of universal health coverage to most Thais (74% of the population) [15] has implemented the broader coverage for T1D, with comprehensive treatment benefits, including insulin, BGM supplies, urine ketone strips, and DSMES. The network has since been strengthened to cover the entire country. In 2019, the participating hospitals in T1DDAR CN expanded to 37 hospitals. This manuscript aimed to explore the first-year real-world clinical outcomes following the full implementation and operationalization of DSMP-NS, supported by the NHSO’s universal health coverage. Thus, it was ethical to omit a control group receiving conventional insulin therapy for comparison. We hypothesized that DSMP-NS would result in an increased proportion of participants achieving better glycaemic control and a reduced occurrence of acute diabetes complications, including diabetic ketoacidosis (DKA) and severe hypoglycaemia (SH). Additionally, we aimed to explore characteristics and factors associated with the outcomes. Methods A prospective observational within-subject comparative study was conducted across 37 hospitals in T1DDAR CN, evaluating outcomes in 481 participants with T1D. Baseline data were collected in February 2019, followed by subsequent assessments performed over the course of one year commencing after DSMP-NS. At baseline, 240 participants were receiving care at hospitals classified in the early developing phase of DSMP-NS, which comprised team development, healthcare personnel training, and network establishment. The inclusion criteria were T1D children aged < 18 years and adults, who were willing to participate in DSMP-NS for at least 3 months. T1D was defined by the presence of positive pancreatic autoantibodies, absolute insulin deficiency (c-peptide < 0.6 ng/mL), or a clinical diagnosis of hyperglycaemia with or without ketoacidosis, requiring insulin therapy within the first year. Exclusion criteria included individuals unable to comply with DSMP-NS. Purposive sampling was utilized. Participants were withdrawn if they lost follow-up, were referred to non-participating hospitals, or withdrew consent. Training and DSMP-NS Each hospital had a multidisciplinary team of healthcare professionals, including physicians, nurses, and dietitians. They received annual training in delivering DSMES, including 11 education modules covering general knowledge, diabetes complications, insulin treatment, glycaemic monitoring, carbohydrate counting, sick day care, exercise and social activity, special situations, pregnancy, and psychosocial issues. These modules were developed to suit Thai T1D needs, considering language, diet, and social activities. These educational tools were utilized for both children and adults. DSMP-NS integrated four components: 1) DSMES; (2) promoting the use of IIT; (3) providing free-of-charge BGM devices, glucose test strips, and urine ketone strips; and (4) a referral network for appropriate care. The NHSO supported self-monitoring equipment in 456 participants (94.8%), while DAT or the hospitals supported the others. Participants were encouraged to perform BGM at least three times daily and received at least six essential DSMES modules within the first year and periodically thereafter. For children and adolescents, parents/guardians participated in the training. The additional DSMES modules (e.g. exercise and social activity, special situations, pregnancy, and psychosocial issues) were introduced as appropriate. IIT emphasized insulin analogs with mealtime dosing, targeting HbA1c < 7% in adults and < 7.5% in children. Participants were instructed on smoking cessation, alcohol consumption, and abstaining from recreational drugs. Teleconsultation, glucose data sharing, and diabetes camps were implemented per facility practices. Physicians adjusted insulin according to national and international guidelines and individual needs [16–19]. Networking system The network consisted of Hub, Node, and Level 2 Plus (L2Plus) hospitals, each with a trained multidisciplinary care team and varying endocrinologist and staff availability. For complex T1D participants, consultation or referral to higher-level hospitals would be as needed. Hubs and Nodes also supported regional training. This study covered 19 Hubs, 14 Nodes, and 4 L2Plus hospitals across all 13 health regions of Thailand (Fig. 1). Fig. 1Shows Hubs, Nodes, and L2Plus hospitals covering all 13 health regions of Thailand. The Hub hospitals provided diabetes management, trained multidisciplinary teams, were consultant of Node and L2 Plus, and arranged network training in their respective health regions. The Node hospitals and L2 Plus provided care to T1D and would refer participants with complex conditions to higher-level hospitals when necessary Shows Hubs, Nodes, and L2Plus hospitals covering all 13 health regions of Thailand. The Hub hospitals provided diabetes management, trained multidisciplinary teams, were consultant of Node and L2 Plus, and arranged network training in their respective health regions. The Node hospitals and L2 Plus provided care to T1D and would refer participants with complex conditions to higher-level hospitals when necessary Data collection Baseline data included demographic details, age at diagnosis, anthropometric measurements, pancreatic autoantibodies, insulin usage, BGM frequency, DKA and SH episodes in the previous year, chronic complications, comorbidities, and latest HbA1c levels. Data were obtained from medical records; when these were unavailable, information (e.g. DKA, SH, and BGM frequency) was collected from hospitalization narratives provided by participants and documented by medical personnel. Routine blood chemistries, including lipid profiles and renal function, were evaluated at local laboratories. Participants had scheduled visits every 1 to 4 months, totalling 3 to 12 visits per year based on physician judgment. Data on diabetes complications, self-care, BGM frequency, and routine exams (weight, height, HbA1c) were collected at 3 and 6 months (± 45 days) and every 6 months (± 90 days) thereafter. This report covered the first 12 months, using the most recent data per timeframe. All data were recorded in RedCAP (Vanderbilt University, USA). Microvascular complication screening, including diabetic retinopathy, neuropathy, and nephropathy, was performed annually for participants with T1D diagnosed for ≥ 5 years or after the age of 10 and with ≥ 2 years of diabetes duration. Data accuracy was ensured through audit reviews and site confirmation. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies [20]. The primary outcome was the proportion achieving all three components of the composite outcome: (1) Improved glycaemic control (the latest HbA1c met goal (< 7.5% in children or < 7.0% in adults) or HbA1c decrement ≥ 0.5% in the participants with HbA1c more than goal at baseline), (2) No DKA, and (3) No SH during the visit. Secondary objectives included glycaemic control, DKA and SH incidence, and factors associated with the primary outcome following DSMP-NS. Definition Obesity in children was defined as a body mass index (BMI) >2 standard deviations (SD) from the Z score [21] or ≥ 25 kg/m2 in adults [22]. SH was defined as hypoglycaemia requiring assistance [23]. DKA was diagnosed with plasma glucose >200 mg/dL, serum ketone ≥ 3 mmol/L or moderate-to-marked ketonuria, serum bicarbonate < 18 mEq/L, anion gap >10 mEq/L or arterial/venous pH < 7.3 [24, 25]. Diabetic retinopathy was identified via retinal photography [16]. Diabetic nephropathy was diagnosed when urine albumin-to-creatinine ratio (UACR) was more than 30 mg/gm creatinine on at least 2 occasions within 6 months [16]. Peripheral neuropathy was determined by loss of protective sensation, evaluated by monofilament or pinprick tests. Peripheral arterial disease was defined as a decrease or absence of the dorsalis pedis and/or posterior tibial artery, an ankle-brachial index < 0.9, or foot gangrene. Dyslipidaemia was defined as the calculated or measured serum low-density lipoprotein (LDL) ≥ 100 mg/dL or ongoing lipid-lowering treatment. Autoimmune thyroid diseases were established through medical documentation of Graves’ disease or Hashimoto’s thyroiditis. Glycaemic control was categorized into three groups based on glycaemic target achievement: (1) Good control-The latest HbA1c levels met the goal; (2) Poor control-The latest HbA1c > 9.0%; and (3) Fair control-The latest HbA1c levels were between good and poor control. HbA1c changed post-DSMP-NS was classified as: (1) Improved-The latest HbA1c met the goal or decreased by ≥ 0.5% from baseline in participants whose baseline HbA1c was higher than the goal [26]; (2) Not improved-HbA1c levels remained above the goal with a change of < 0.5%; (3) Worse-HbA1c levels remained above the goal with an increase of >0.5% from baseline. Sample size calculation The sample size was calculated based on the proportion of T1D who had HbA1c < 7% without DKA and SH occurrence. Preliminary data from the early phase of DSMP-NS in Thailand during 2016–2017 revealed that T1D with HbA1c < 7.0% at baseline and after 1 year were 12.8% and 16.8%, respectively [14]. The DKA incidence decreased from 16.1 to 7.1 per 100 person-year, and the SH incidence also decreased from 33.5 to 15.9 per 100 person-year. The probability of the composite endpoint at baseline and 1 year was 0.072 and 0.131, respectively. With the hypothesis of a correlation between paired observations of 1%, a 95% confidence level, and 80% power to determine a difference, this study required a minimum of 397 participants. Statistical analysis Data were analyzed using Stata/IC version 14.0 for Windows (Stata Corp LP, College Station, TX, USA) with an intention-to-treat approach. Normally distributed variables were reported as mean, SD, and 95% confidence interval (95% CI), while non-normally distributed data were expressed as median and interquartile range. Categorical data were presented as numbers or percentages. For within-group comparisons, repeated measures mixed models, ordered logistic regression, and Poisson regressions were used to analyse continuous, ordinal, and binary outcomes, respectively. The Chi-square test compared categorical data between groups. Logistic regression and linear mixed models identified factors associated with outcomes, and Fisher’s protected least-significant difference (LSD) was used for multiple comparisons. Results A total of 481 participants were enrolled in this study, comprising 305 children and 176 adults. At 12 months, the results of 412 participants (85.7%) were statistically analysed, and their findings were presented (Fig. 2). One participant passed away due to congestive heart failure. Fig. 2Consort diagram Consort diagram The majority of T1D participants were children aged under 18 years old (63.4%). The median (interquartile range) age at enrolment was 15.8 (12.2, 20.5) years, and diabetes duration was 5.6 (3.0, 8.8) years. Among adult participants (aged 18 years or older), females were predominant. Adults had higher rates of obesity, longer diabetes duration, were less likely to perform BGM ≥3 times/day, and had better HbA1c levels than children at enrolment (HbA1c 8.8 ± 2.3% vs. 9.5 ± 2.5%, p = 0.001 in adults and children, respectively) (Table 1). IIT was applied to 93% of the participants. Hypertension, dyslipidaemia, autoimmune thyroid diseases, and microvascular complications such as diabetic retinopathy and diabetic nephropathy were significantly more prevalent in adults compared to children (Table 1). All participants completed 6 essential DSMES modules during the 12-month study. Table 1Baseline characteristicsBaseline characteristicsTotal(N = 481)Children(n = 305)Adults(n = 176)Female293 (60.9)174 (57.0)119 (67.6)*Age at diagnosis (years)9.7 (6.0–12.6)8.3 (4.9–10.8)13.5 (9.5–18.1)*Diabetes duration (years)5.6 (3.0–8.8)3.9 (2.4–6.8)9.0 (6.1–14.3)*Age at enrolment (years)15.8 (12.2–20.5)13.1 (10.3–15.4)22.4 (19.9–26.5)*Body weight at enrolment (kg)50.1 (16.5)44.3 (16.2)59.9 (11.8)*Obesity at enrolment (%)74/460 (16.1)24/286 (8.4)50/174 (28.7)*Positive pancreatic autoantibody (%)197/252 (78.6)147/184 (79.9)50/68 (73.5)Insulin Regimen†- Conventional (%)30 (6.3)19 (6.3)11 (6.3)- Basal bolus (%)446 (93.3)284 (93.4)162 (93.1)- Continuous subcutaneous insulin infusion (%)2 (0.4)1 (0.3)1 (0.6)BGM ≥ 3 times/day291/472 (61.7)205/301 (68.1)86/171 (50.3)*HbA1c at baseline (%)9.3 (2.4)9.5 (2.5)8.8 (2.3)*Diabetic retinopathy25/336 (7.4)1/194 (0.5)24/142 (16.9)*Diabetic nephropathy54/333 (16.2)17/189 (8.9)37/144 (25.7)*Diabetic neuropathy7/196 (3.6)1/83 (1.2)6/113 (5.3)Hypertension15/481 (3.1)2/305 (0.7)13/176 (7.4)*Dyslipidemia63/481 (13.0)19/305 (6.2)44/176 (25.0)*Cerebrovascular disease1/481 (0.21)0/305 (0.0)1/176 (0.6)Autoimmune thyroid diseases25/481 (5.2)11/305 (3.6)14/176 (7.9)*Data are presented in mean (SD) or median (interquartile range)BGM: blood glucose monitoring* Significant different between age groups (p < 0.05)† Insulin regimen was categorized as follows: conventional insulin treatment (basal, basal plus, premixed or self-mixed insulin 1–3 injections per day) and intensive insulin treatment (multiple daily injections using long-acting insulin with premeal insulin analogs or continuous subcutaneous insulin infusion) Table 1Baseline characteristicsBaseline characteristicsTotal(N = 481)Children(n = 305)Adults(n = 176)Female293 (60.9)174 (57.0)119 (67.6)*Age at diagnosis (years)9.7 (6.0–12.6)8.3 (4.9–10.8)13.5 (9.5–18.1)*Diabetes duration (years)5.6 (3.0–8.8)3.9 (2.4–6.8)9.0 (6.1–14.3)*Age at enrolment (years)15.8 (12.2–20.5)13.1 (10.3–15.4)22.4 (19.9–26.5)*Body weight at enrolment (kg)50.1 (16.5)44.3 (16.2)59.9 (11.8)*Obesity at enrolment (%)74/460 (16.1)24/286 (8.4)50/174 (28.7)*Positive pancreatic autoantibody (%)197/252 (78.6)147/184 (79.9)50/68 (73.5)Insulin Regimen†- Conventional (%)30 (6.3)19 (6.3)11 (6.3)- Basal bolus (%)446 (93.3)284 (93.4)162 (93.1)- Continuous subcutaneous insulin infusion (%)2 (0.4)1 (0.3)1 (0.6)BGM ≥ 3 times/day291/472 (61.7)205/301 (68.1)86/171 (50.3)*HbA1c at baseline (%)9.3 (2.4)9.5 (2.5)8.8 (2.3)*Diabetic retinopathy25/336 (7.4)1/194 (0.5)24/142 (16.9)*Diabetic nephropathy54/333 (16.2)17/189 (8.9)37/144 (25.7)*Diabetic neuropathy7/196 (3.6)1/83 (1.2)6/113 (5.3)Hypertension15/481 (3.1)2/305 (0.7)13/176 (7.4)*Dyslipidemia63/481 (13.0)19/305 (6.2)44/176 (25.0)*Cerebrovascular disease1/481 (0.21)0/305 (0.0)1/176 (0.6)Autoimmune thyroid diseases25/481 (5.2)11/305 (3.6)14/176 (7.9)*Data are presented in mean (SD) or median (interquartile range)BGM: blood glucose monitoring* Significant different between age groups (p < 0.05)† Insulin regimen was categorized as follows: conventional insulin treatment (basal, basal plus, premixed or self-mixed insulin 1–3 injections per day) and intensive insulin treatment (multiple daily injections using long-acting insulin with premeal insulin analogs or continuous subcutaneous insulin infusion) Baseline characteristics Data are presented in mean (SD) or median (interquartile range) BGM: blood glucose monitoring * Significant different between age groups (p < 0.05) † Insulin regimen was categorized as follows: conventional insulin treatment (basal, basal plus, premixed or self-mixed insulin 1–3 injections per day) and intensive insulin treatment (multiple daily injections using long-acting insulin with premeal insulin analogs or continuous subcutaneous insulin infusion) Glycaemic control and acute diabetic complications during follow-up At the 12th month, 43% of participants met the composite outcome compared to 17.6% at baseline (Table 2). Adults showed a significantly lower mean HbA1c at the 12th month compared to baseline (8.2 ± 2.0% vs. 8.8 ± 2.3%, p = 0.005). The proportion of participants achieving good glycaemic control significantly increased at the 6th and 12th months, compared to baseline, especially in adults (Fig. 3). More children achieved good glycaemic control compared with baseline, but the difference was not statistically significant over 12 months (Fig. 3). HbA1c reduction of at least 0.5% compared with baseline was achieved in approximately one-third of participants (34.7%) at the 12th month. More than half of adults significantly improved their glycaemic control in the 6th month (Fig. 4). In contrast, children had a higher proportion of worse glycaemic control at the 12th month compared to the 3rd month (Fig. 4). There was a significantly increased number of participants who monitored BGM ≥ 3 times/day at the 12th month compared to baseline (68.9% vs. 61.7%, at 12th and 0 months, respectively, p = 0.005) (Table 2). Table 2Clinical characteristics and glycaemic control comparing at baseline and 12-month periodCharacteristicsTotalChildrenAdultBaseline12th monthsBaseline12th monthsBaseline12th months(n = 481)(n = 412)(n = 305)(n = 269)(n = 176)(n = 143) Body weight (kg.) 50.1 (16.5)52.2 (16.2) ***44.3 (16.2)47.5 (16.2) ***59.9 (11.8)60.6 (12.3) ***Available record473385298246175139 Obesity (%) † 74 (16.1)72 (19.0)*24 (8.4)29 (12.0)50 (28.7)43 (31.4)Available record460379286242174137 Composite outcome (%) ‡ 83 (17.6)163 (43.0)***54 (17.9)95 (38.3)***29 (17.0)68 (51.9)***Available record472379301248171131 HbA1c (%) (SD) 9.2 (2.4)9.1 (2.4)9.5 (2.5)9.6 (2.5)8.8 (2.3)8.2 (2.0) **Available record472388301251171137 Glycaemic control § Good (%)86 (18.2)90 (23.2) **56 (18.6)49 (19.5)30 (17.5)41 (29.9) **Fair (%)177 (37.5)134 (34.5)*97 (32.2)75 (29.9)80 (46.8)59 (43.1)Poor (%)209 (44.3)164 (42.3)148 (49.2)127 (50.6)61 (35.7)37 (27.1) ***Available record472388301251171137 HbA1c reduction at least 0.5% from baseline (%) NA132 (34.7)NA81 (32.7)NA51 (38.6)Available recordNA380NA248NA132 BGM frequency (times/day) 2.8 (1.3)2.9 (1.2) *2.9 (1.2)3.0 (1.2)2.4 (1.3)2.7 (1.2) *Available record472411301269171142 Proportion of BGM ≥ 3 time/days (%) 291 (61.7)283 (68.9) **205 (68.1)194 (72.1)86 (50.3)89 (62.7) **Available record472411301269171142 DKA incidence (per person-year)22.110.2 ***23.211.4 **20.37.9**(95%CI)(18.6–26.3)(7.6–13.8)(18.8–28.7)(8.1–16.2)(14.9–27.5)(4.4–14.3)** SH incidence (per person-year)4.03.44.03.74.12.7(95%CI)(2.6–6.3)(2.0-5.6)(2.3-7.0)(2.0-6.8)(2.0-8.4)(1.0-7.2)Data are presented in n (%) or mean (SD)*** p < 0.001, ** p < 0.01, * p < 0.05 compared with baselineNA: not applicable; DKA: diabetic ketoacidosis; SH: severe hypoglycaemia; BGM: blood glucose monitoring† “Obesity” was defined as body mass index Z score > + 2SD in children and ≥25 kg/m2 in adults‡"Composite outcome” defined as improved glycaemic control without occurrence of DKA and SH“Improved glycaemic control” defined as the latest HbA1c met goal (< 7.0% in adults or < 7.5% in children) or HbA1c decrement ≥ 0.5% in the participants with HbA1c more than goal at baseline§Glycaemic control; Good: HbA1c < 7.5% in children and < 7.0% in adults; Fair: HbA1c 7.5-9.0% in children and 7.0–9.0% in adults; Poor: HbA1c > 9.0% Table 2Clinical characteristics and glycaemic control comparing at baseline and 12-month periodCharacteristicsTotalChildrenAdultBaseline12th monthsBaseline12th monthsBaseline12th months(n = 481)(n = 412)(n = 305)(n = 269)(n = 176)(n = 143) Body weight (kg.) 50.1 (16.5)52.2 (16.2) ***44.3 (16.2)47.5 (16.2) ***59.9 (11.8)60.6 (12.3) ***Available record473385298246175139 Obesity (%) † 74 (16.1)72 (19.0)*24 (8.4)29 (12.0)50 (28.7)43 (31.4)Available record460379286242174137 Composite outcome (%) ‡ 83 (17.6)163 (43.0)***54 (17.9)95 (38.3)***29 (17.0)68 (51.9)***Available record472379301248171131 HbA1c (%) (SD) 9.2 (2.4)9.1 (2.4)9.5 (2.5)9.6 (2.5)8.8 (2.3)8.2 (2.0) **Available record472388301251171137 Glycaemic control § Good (%)86 (18.2)90 (23.2) **56 (18.6)49 (19.5)30 (17.5)41 (29.9) **Fair (%)177 (37.5)134 (34.5)*97 (32.2)75 (29.9)80 (46.8)59 (43.1)Poor (%)209 (44.3)164 (42.3)148 (49.2)127 (50.6)61 (35.7)37 (27.1) ***Available record472388301251171137 HbA1c reduction at least 0.5% from baseline (%) NA132 (34.7)NA81 (32.7)NA51 (38.6)Available recordNA380NA248NA132 BGM frequency (times/day) 2.8 (1.3)2.9 (1.2) *2.9 (1.2)3.0 (1.2)2.4 (1.3)2.7 (1.2) *Available record472411301269171142 Proportion of BGM ≥ 3 time/days (%) 291 (61.7)283 (68.9) **205 (68.1)194 (72.1)86 (50.3)89 (62.7) **Available record472411301269171142 DKA incidence (per person-year)22.110.2 ***23.211.4 **20.37.9**(95%CI)(18.6–26.3)(7.6–13.8)(18.8–28.7)(8.1–16.2)(14.9–27.5)(4.4–14.3)** SH incidence (per person-year)4.03.44.03.74.12.7(95%CI)(2.6–6.3)(2.0-5.6)(2.3-7.0)(2.0-6.8)(2.0-8.4)(1.0-7.2)Data are presented in n (%) or mean (SD)*** p < 0.001, ** p < 0.01, * p < 0.05 compared with baselineNA: not applicable; DKA: diabetic ketoacidosis; SH: severe hypoglycaemia; BGM: blood glucose monitoring† “Obesity” was defined as body mass index Z score > + 2SD in children and ≥25 kg/m2 in adults‡"Composite outcome” defined as improved glycaemic control without occurrence of DKA and SH“Improved glycaemic control” defined as the latest HbA1c met goal (< 7.0% in adults or < 7.5% in children) or HbA1c decrement ≥ 0.5% in the participants with HbA1c more than goal at baseline§Glycaemic control; Good: HbA1c < 7.5% in children and < 7.0% in adults; Fair: HbA1c 7.5-9.0% in children and 7.0–9.0% in adults; Poor: HbA1c > 9.0% Clinical characteristics and glycaemic control comparing at baseline and 12-month period DKA incidence (per person-year) SH incidence (per person-year) Data are presented in n (%) or mean (SD) *** p < 0.001, ** p < 0.01, * p < 0.05 compared with baseline NA: not applicable; DKA: diabetic ketoacidosis; SH: severe hypoglycaemia; BGM: blood glucose monitoring † “Obesity” was defined as body mass index Z score > + 2SD in children and ≥25 kg/m2 in adults ‡"Composite outcome” defined as improved glycaemic control without occurrence of DKA and SH “Improved glycaemic control” defined as the latest HbA1c met goal (< 7.0% in adults or < 7.5% in children) or HbA1c decrement ≥ 0.5% in the participants with HbA1c more than goal at baseline §Glycaemic control; Good: HbA1c < 7.5% in children and < 7.0% in adults; Fair: HbA1c 7.5-9.0% in children and 7.0–9.0% in adults; Poor: HbA1c > 9.0% During the DSMP-NS period, there was a statistically significant increase in mean body weight (52.3 ± 16.1 kg vs. 50.1 ± 16.5 kg at 12th month and baseline, respectively, p < 0.001) (Table 2). This weight gain was observed in both pediatric and adult populations. Additionally, the prevalence of obesity showed a corresponding significant increase (Table 2), especially in children (12.0% vs. 8.4% at 12th and baseline, respectively, p < 0.05). Fig. 3Proportion of participants with HbA1c levels meeting the treatment goal. ** p < 0.01, * p < 0.05 compared with baseline. Treatment goals were HbA1c < 7.5% in children or < 7.0% in adults Proportion of participants with HbA1c levels meeting the treatment goal. ** p < 0.01, * p < 0.05 compared with baseline. Treatment goals were HbA1c < 7.5% in children or < 7.0% in adults Fig. 4Comparing glycaemic control categories at 3, 6 and 12 months after DSMP-NS. Data was presented in numbers (%). Improved defined as the latest HbA1c met goal (< 7.5% in children or < 7.0% in adults) or HbA1c decrement ≥ 0.5% in the participants with HbA1c more than the goal at baseline. Not improved was defined as HbA1c was more than the goal at baseline, and HbA1c decrement was less than 0.5% at follow-up. Worse defined as HbA1c was more than the goal and increased by at least 0.5% at follow-up compared with baseline. *p < 0.05 compared with 3 months Comparing glycaemic control categories at 3, 6 and 12 months after DSMP-NS. Data was presented in numbers (%). Improved defined as the latest HbA1c met goal (< 7.5% in children or < 7.0% in adults) or HbA1c decrement ≥ 0.5% in the participants with HbA1c more than the goal at baseline. Not improved was defined as HbA1c was more than the goal at baseline, and HbA1c decrement was less than 0.5% at follow-up. Worse defined as HbA1c was more than the goal and increased by at least 0.5% at follow-up compared with baseline. *p < 0.05 compared with 3 months The incidence rate of DKA significantly decreased after the participants enrolled in DSMP-NS (22.1 (95% CI 18.6–26.3) vs. 10.2 (95% CI 7.6–13.8) episodes per person-year at the baseline and 12th month, respectively, p < 0.001) (Table 2), with both children and adults achieving similar results. There was no significant increase in SH episodes (Table 2). Both children and adults followed the same trend. Factors associated with glycaemic control and acute diabetic complications Factors associated with the composite outcome (adjusted for sex) included older age (adjusted odds ratio (aOR) 1.06, 95% CI 1.03 to 1.09, p = 0.000) and performing BGM at least 3 times/day (aOR 2.02, 95% CI 1.38 to 3.53, p = 0.001). An increase in BGM frequency by one time per day corresponded to a 39% increase in the likelihood of achieving good or better glycaemic control without DKA or SH (OR 1.39, 95% CI 1.13–1.70, p = 0.002). DSMP-NS with free-of-charge strips led to a significantly increased frequency of BGM (Table 2). Conversely, the presence of obesity at enrolment was negatively associated with the composite outcome, which was improved glycaemic control without acute diabetes complications during the visit (aOR 0.51, 95% CI 0.27 to 0.96, p = 0.038). Whereas weight gain, education, income, diabetes duration, and hospital levels (Hub, Node/L2Plus) revealed no significant association. Discussion The DSMP-NS intervention, which encompassed strengthening the knowledge and skills of healthcare personnel, consultation systems, DSMES, the provision of insulin analogs and BGM equipment, and the networking system, led to a significant improvement in T1D care outcomes in a real-world clinical setting. This study clearly emphasized the effectiveness of this program, as evidenced by a significant increase in the proportion of participants who met the favourable composite outcome (achieving glycaemic goals and/or a decreased HbA1c of at least 0.5% without DKA and SH occurrence), as well as a reduction in DKA episodes at the 12th month of the study. Furthermore, despite IIT, the incidence of SH did not increase, likely due to our comprehensive care approach and education on hypoglycaemia recognition and management. However, both adults and children experienced a significant increase in body weight alongside improved glycaemic control, consistent with the results of a previous study [27]. T1D is a chronic disease that requires lifelong insulin therapy. Despite advancements in insulin generation, BGM technologies, and insulin pumps, data from the T1D Exchange registry in the United States during 2016–2018 revealed that only 17% of youth and 21% of adults achieved good glycaemic control [28]. DSMES and BGM are the cornerstones of diabetes management and have been emphasized in guidelines to enhance positive diabetes self-care behaviour and address psychological aspects [29, 30]. DSMES also improves glycaemic control, decreases acute diabetic complications, and leads to a better quality of life for individuals with T1D [31, 32]. In our study, the glycaemic control of children and adults differed. Children had worse baseline glycaemic control compared to adults. Furthermore, after DSMP-NS, the mean HbA1c and the proportion of those with good glycaemic control were more pronounced in adult participants. In contrast, children had no significant changes in these parameters despite having numerically more frequent BGM than at baseline. These results were in agreement with other published studies, which showed that children had higher HbA1c levels compared to adults [8, 28]. Possible explanations include erratic lifestyles and hormonal surges contributing to difficulties in blood glucose control, especially in adolescents. In addition, poor adherence to diabetes-related tasks and psychosocial issues might influence poorer glycaemic control in children, particularly among adolescents [33]. Achievement of glycaemic goals in youth was different depending on age. This underlines that parental involvement is essential in T1D care. In the current study, older age and more frequent BGM monitoring were independently associated with the composite outcome, similar to previous data [8, 34, 35]. An increase in BGM frequency by one time per day corresponded to a 39% increased likelihood of achieving good or better glycaemic control without DKA or SH. On the other hand, obesity at baseline was negatively associated with the composite outcome, which is consistent with the previous study [36]. This effect may be attributable to increased insulin resistance. Although body weight and obesity prevalence increased significantly, weight gain during the one-year follow-up was not associated with the composite outcome. These changes may reflect behavioural influences or insulin-related mechanisms. This study did not find an association between socioeconomic status and the composite outcome, which differed from a previously published study [37]. This discrepancy could be because the DSMP-NS provided insulin analogs and medical devices free of charge. While not currently universally reimbursable in Thailand, our results demonstrated the benefits of increasing access to standard diabetes care, which should be incorporated into the other health reimbursement schemes, including social security and civil service welfare. The strength of our study was being the first real-life observational cohort study in an upper to middle-income/limited resource setting for T1D care in Southeast Asia. It confirmed the effectiveness of DSMP-NS under universal health coverage in the real-world clinical setting. While previous studies demonstrated the effectiveness of DSMES [9, 32, 38], we further enhanced the implementation by incorporating systematic management of 13 health regions, 37 hospitals, along with the cooperation of academic societies as a network, and the National Health Security Office. The T1DDAR CN education strategy emphasized individualized management according to participants’ needs and treatment standards using the self-care behaviours framework of the Association for Diabetes Care & Education (ADCES), adapted to Thai society and culture [39]. The improvement in clinical outcomes likely results from a combination of more frequent BGM, improved self-care behaviours, and the use of IIT. This study has several limitations. First, there was a potential for selection bias, as only participants who were motivated in self-care, willing to perform frequent BGM, and use IIT were included in the program. Additionally, some self-reported data at baseline might have underestimated the prevalence of SH and DKA. Behavioural changes (beyond BGM), which have been demonstrated to contribute to better glycaemic control [40], were not assessed in this study. The variability in the number of follow-up visits may also have influenced the quality of care. Furthermore, it is also important to note that the COVID-19 pandemic disrupted the routine face-to-face management system and might have affected the outcomes after the pandemic. Finally, this study employed a pre- and post-intervention approach to evaluate the effectiveness of intensive insulin therapy, with the pretreatment phase serving as the control. While this study confirmed the short-term efficacy and feasibility of DSMP-NS in real life through T1DDAR CN, a long-term study should be conducted to explore the cost-effectiveness, chronic diabetic complications, quality of life, and mortality after the intervention. The results support that the DSMP-NS should be implemented throughout the country, and that the government should focus on establishing evidence-based, structured diabetes education programs [30], developing certified diabetes educators, and creating national standards for DSMES [41]. Reimbursement of DSMP-NS should be incorporated into all national health schemes, and regular audits of such programs should be performed to ensure desirable outcomes and continuous quality improvement. Additionally, emerging approaches such as continuous glucose monitoring (CGM)-based dynamic markers and machine learning models could further enhance screening and tracking of disease progression. Integrating these tools into nationwide programs like DSMP-NS could help identify individuals at risk earlier and personalize intervention strategies [42, 43]. Conclusions The DSMP-NS intervention in conjunction with T1DDAR CN improved T1D glycaemic control, reducing DKA episodes through free medical and monitoring supplies, improvement of care, and self-management. Increasing age and frequent BGM (≥ 3 times/day) were linked to the composite outcome, while obesity had a negative impact. This real-world study highlights the importance of education, supporting and promoting BGM, and networking in T1D care in resource-limited settings. Sustainable national policies are needed to ensure access to supplies, DSMES, and healthcare. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: List of Thai-Type 1 Diabetes and Diabetes diagnosed Age before 30 years Registry, Care and Network (T1DDAR CN) investigators Supplementary Material 1: List of Thai-Type 1 Diabetes and Diabetes diagnosed Age before 30 years Registry, Care and Network (T1DDAR CN) investigators