Clinical and psychological effects of the JADEC Diabetes Card System assessed using the Problem Areas in Diabetes (PAID): A retrospective observational study Authors: Yu Kouchi Published: 2026 ABSTRACT ABSTRACTAims/IntroductionDiabetes self‐management education and support (DSMES) is essential for long‐term diabetes care; however, its psychological impact in routine outpatient practice remains unclear. This study evaluated the clinical and psychological effects of DSMES using the Diabetes Card System, a visually structured educational tool developed in Japan.Materials and MethodsThis retrospective observational study included 141 patients with type 2 diabetes who received DSMES using the Diabetes Card System at a single outpatient clinic between April 2022 and July 2025. Changes in Problem Areas in Diabetes (PAID) scores and glycated hemoglobin (HbA1c) over 6 months were analyzed. Multivariable linear regression analyses were performed to identify factors associated with changes in PAID (ΔPAID) and HbA1c (ΔHbA1c). Additional analyses were conducted according to treatment modification status.ResultsTotal PAID scores and HbA1c levels significantly improved over 6 months. Baseline PAID score was the only factor associated with ΔPAID. ΔHbA1c was associated with baseline HbA1c, duration of diabetes, and age, but not with baseline PAID. Greater ΔPAID was observed in patients with higher baseline PAID scores irrespective of HbA1c levels. Treatment intensification was associated with greater reductions in HbA1c, whereas PAID improvement did not differ according to treatment modification status.ConclusionsDSMES using the Diabetes Card System was associated with reductions in diabetes‐related distress and improvements in glycemic control in routine outpatient care. Psychological improvement appeared independent of patient background factors, whereas glycemic improvement was associated with disease‐related indicators. The Diabetes Card System may represent a practical DSMES tool in routine outpatient practice. INTRODUCTION Diabetes mellitus is a chronic condition that requires long‐term management. Effective glycemic control and the prevention or delay of complications depend on patients' engagement in diabetes self‐management education (DSME) and ongoing diabetes self‐management education and support (DSMES) 1 , 2 . However, many individuals with diabetes experience a substantial psychological burden associated with disease management. In Japan, diabetes education has traditionally relied on physician‐led explanations and brief lifestyle counseling during routine consultations, making it challenging to provide standardized support that adequately reflects individual patients' levels of understanding and personal values. The Japan Association for Diabetes Care and Education (JADEC) has developed and recommends the Diabetes Card System, which enables DSMES tailored to individual patient backgrounds and is widely used in routine clinical practice 3 . The Diabetes Card System is a visual, educational tool–based approach to DSMES, in which key information on pathophysiology, lifestyle modification, and complication prevention is systematically organized into individual cards. By presenting information visually, this system reduces patients' cognitive burden and allows both standardization of educational content and individualized support—features not typically achieved through conventional explanation‐based education alone. Standardization of DSMES is also recognized internationally as a major challenge; therefore, evaluating the real‐world clinical and psychological effects of the Japan‐specific Diabetes Card System in outpatient settings is of considerable importance. In addition, the system allows healthcare providers to combine multiple cards according to each patient's disease status and lifestyle context, enabling targeted support 4 , 5 , and it is expected to contribute to the reduction of diabetes‐related psychological burden 6 , 7 . At Kouchi Clinic, DSMES using the Diabetes Card System has been implemented in routine clinical practice; however, reports on its clinical usefulness remain limited. The Problem Areas in Diabetes (PAID) scale is a 20‐item self‐administered questionnaire that assesses diabetes‐related distress and is widely used internationally 8 . PAID enables visualization of psychological burden and is also useful as an outcome measure for evaluating the effectiveness of DSMES 8 , 9 , 10 , 11 . In this study, we retrospectively analyzed changes in PAID scores and HbA1c levels before and after the use of the Diabetes Card System among patients with diabetes who transferred from other institutions to our clinic in order to examine its clinical and psychological usefulness. MATERIALS AND METHODS Study design and participants This study was conducted as a retrospective observational study at Kouchi Clinic (7‐28, Teppou‐machi, Shinjo, Yamagata, Japan). Eligible participants were patients with type 2 diabetes mellitus who first visited our clinic between April 1, 2022, and July 31, 2025, and received DSMES using the Diabetes Card System developed by JADEC. Exclusion criteria were age <18 years, severe dementia, pregnancy‐related diabetes, and cases deemed inappropriate for inclusion by the principal investigator. Outcomes The primary outcome was the change in the total Problem Areas in Diabetes (PAID) score over the 6‐month period before and after the introduction of the Diabetes Card System. Secondary outcomes included changes in glycated hemoglobin (HbA1c) levels. The PAID questionnaire is a 20‐item self‐administered scale assessing diabetes‐related distress. Each item is scored from 1 to 5, and the total score is converted to a 0–100 scale for analysis. In this study, the Japanese version of the PAID questionnaire was used. The English wording of each item presented in this report was translated by the authors from the original Japanese version for reporting purposes and may differ slightly from the original English PAID questionnaire developed by the American Diabetes Association (ADA). Data collection The following data were extracted from outpatient electronic medical records: patient identification number, age, sex, duration of diabetes, history of pharmacological therapy, presence of comorbidities, and diabetes‐related complications, referral source (diabetes specialist or non‐specialist), PAID scores, and HbA1c values. Clinical data were collected at seven time points: at the initial visit and at 1 to 6 thereafter. Intervention using the diabetes card system Implementation of the Diabetes Card System began 1 month after the initial visit. In principle, the following cards were used sequentially: Month 1: “Blood glucose,” Month 2: “What is diabetes?,” Month 3: “Indicators of glycemic control,” Month 4: “Urine testing,” Month 5: “Diabetes and complications,” Month 6: “Hyperglycemia and dehydration.” However, the number, content, and order of cards were flexibly adjusted according to each patient's clinical condition and lifestyle background. DSMES using the Diabetes Card System was primarily provided by a Certified Diabetes Care and Education Specialist (CDCES) experienced in the use of the system. Educational sessions were incorporated into routine outpatient care and typically required approximately 3 min per visit, with one card topic discussed at each encounter. Treatment modifications during the 6‐month observation period were reviewed using electronic medical records and classified into three categories: intensification, no change, and de‐intensification. Intensification was defined as the addition of glucose‐lowering medications, dose escalation of existing medications, or initiation of injectable therapy. No change was defined as the absence of substantial modification of glucose‐lowering therapy, including medication switches without an increase in treatment intensity. De‐intensification was defined as dose reduction, discontinuation of glucose‐lowering medications, or withdrawal of injectable therapy. Discontinuation criteria Criteria for discontinuation of the study were as follows: (1) request for withdrawal by the patient or a legally authorized representative; (2) determination after study initiation that the case did not meet the eligibility criteria; (3) inability to evaluate outcomes due to major protocol deviations; or (4) other reasons deemed appropriate by the principal investigator. Statistical analysis Statistical analyses were performed using SPSS version 25 (IBM, Armonk, NY, USA). Changes in PAID scores and HbA1c levels before and after the intervention were analyzed using paired t‐tests. Associations between outcome measures and background factors were examined using multivariable linear regression analysis. To evaluate the potential influence of treatment modifications, additional analyses were performed. Patients were categorized into an intensification group and a non‐intensification group (defined as the combination of the no‐change and de‐intensification groups). Differences in changes in HbA1c and PAID scores between groups were assessed using independent‐samples t‐tests. A two‐sided P value <0.05 was considered statistically significant. Ethics statement Information on the study purpose and data protection policy was disclosed to eligible participants through an institutional opt‐out procedure, and data collection and analysis were conducted based on this consent framework. RESULTS Patient characteristics are shown in Table 1. Among the 141 participants, 82 were men (58.2%) and 59 were women (41.8%). The mean age was 62.3 ± 11.2 years, the duration of diabetes was 8.1 ± 8.9 years, and the HbA1c level at the initial visit was 7.49 ± 1.55%. Comorbidities included hypertension in 86 patients (61.0%), dyslipidemia in 64 (45.4%), neuropathy in 11 (7.8%), retinopathy in 11 (7.8%), nephropathy (stage 3 or higher) in 12 (8.5%), cardiovascular disease in 5 (3.5%), and cerebrovascular disease in 4 (2.8%). Referral sources included diabetes specialists for 36 patients (25.5%) and non‐specialists for 105 patients (74.5%). Regarding treatment, 121 patients (85.8%) were receiving oral antidiabetic agents and 15 patients (10.6%) were receiving injectable therapies. The primary outcome, the total Problem Areas in Diabetes (PAID) score, is shown. The PAID is a scale in which higher scores indicate greater diabetes‐related psychological distress. The total PAID score significantly decreased from 50.20 ± 15.76 before use of the Diabetes Card System to 45.05 ± 16.13 at 6 months (95% CI: 2.552 to 7.746, P < 0.001) (Figure 1). Among the 20 PAID items, scores increased for Q1 (lack of clear goals; 95% CI: −0.764 to −0.342, P < 0.001) and Q3 (fear of diabetes; 95% CI: −1.783 to −1.352, P < 0.001). In contrast, scores decreased for 10 items: Q5 (loss of enjoyment related to eating; 95% CI: −0.479 to −0.088, P = 0.005), Q6 (depression about living with diabetes; 95% CI: 0.174 to 0.563, P < 0.001), Q8 (feeling overwhelmed by diabetes; 95% CI: 0.018 to 0.422, P = 0.033), Q9 (worry about hypoglycemia; 95% CI: 0.001 to 0.382, P = 0.049), Q12 (concerns about complications; 95% CI: 0.268 to 0.669, P < 0.001), Q14 (non‐acceptance of diabetes; 95% CI: 0.057 to 0.468, P = 0.013), Q15 (lack of trust in healthcare providers; 95% CI: −0.648 to −0.203, P < 0.001), Q16 (loss of energy; 95% CI: 0.162 to 0.576, P = 0.001), Q19 (difficulty coping with complications; 95% CI: 0.034 to 0.434, P = 0.022), and Q20 (burnout; 95% CI: 0.093 to 0.460, P = 0.003) (Table 2). Table 1Baseline characteristics of the study participantsVariablesOverall (n = 141) † SexMale82 (58.2%)Female59 (41.8%)Age (years)62.3 ± 11.2Duration of diabetes (years)8.1 ± 8.9Baseline HbA1c (%)7.49 ± 1.55Hypertension86 (61.0%)Dyslipidemia64 (45.4%)Neuropathy11 (7.8%)Retinopathy11 (7.8%)Nephropathy (stage ≥3)12 (8.5%)Cardiovascular disease5 (3.5%)Cerebrovascular disease4 (2.8%)Referral source (specialist/non‐specialist)36 (25.5%)/105 (74.5%)Oral medications121 (85.8%)Injectable therapy (insulin/incretin agents)15 (10.6%)HbA1c, glycated hemoglobin. † Data are presented as mean ± standard deviation or number (percentage). Baseline characteristics of the study participants HbA1c, glycated hemoglobin. Data are presented as mean ± standard deviation or number (percentage). Changes in total PAID score over 6 months. Total Problem Areas in Diabetes (PAID) scores significantly decreased over the 6‐month study period. Data are presented as mean ± standard deviation. *P < 0.001 vs baseline (paired t‐test). Table 2Changes in individual PAID item scores from baseline to 6 monthsPAID item (English wording)Baseline mean ± SD † 6‐month mean ± SD † Mean difference (baseline ‐ 6 months) ‡ 95% CI P‐valueQ1. Feeling discouraged with diabetes1.96 ± 0.1862.52 ± 1.23−0.553−0.764 to −0.342<0.001Q2. Feeling scared when thinking about living with diabetes1.97 ± 0.1672.14 ± 1.16−0.170−0.368 to 0.0280.091Q3. Worrying about the future and serious complications1.14 ± 0.3502.71 ± 1.29−1.567−1.783 to −1.352<0.001Q4. Feeling that diabetes controls your life1.89 ± 0.3091.99 ± 1.04−0.092−0.277 to 0.0920.325Q5. Feeling deprived regarding food and meals2.67 ± 1.212.38 ± 1.110.2840.088 to 0.4790.005Q6. Feeling depressed because of diabetes2.89 ± 1.302.52 ± 1.220.3690.174 to 0.563<0.001Q7. Feeling overwhelmed by diabetes2.45 ± 1.142.24 ± 1.100.2130.000 to 0.4260.050Q8. Feeling angry when thinking about diabetes2.35 ± 1.232.13 ± 1.010.2200.018 to 0.4220.033Q9. Worrying about low blood sugar2.47 ± 1.192.28 ± 1.070.1910.001 to 0.3820.049Q10. Feeling constantly concerned about food and eating2.27 ± 1.132.11 ± 1.070.163−0.031 to 0.3580.100Q11. Feeling alone with diabetes3.11 ± 1.192.90 ± 1.120.206−0.023 to 0.4340.077Q12. Feeling worried about complications3.62 ± 1.293.15 ± 1.180.4680.268 to 0.669<0.001Q13. Feeling burned out by the constant effort2.79 ± 1.182.81 ± 1.18−0.021−0.234 to 0.1920.844Q14. Feeling that diabetes limits your freedom2.41 ± 1.222.15 ± 1.180.2620.057 to 0.4680.013Q15. Feeling that diabetes is taking away your energy1.84 ± 1.201.42 ± 0.7090.4260.203 to 0.648<0.001Q16. Feeling exhausted because of diabetes2.38 ± 1.172.01 ± 0.9750.3690.162 to 0.5760.001Q17. Feeling that diabetes takes too much time and energy1.67 ± 0.9761.65 ± 0.9270.021−0.141 to 0.1840.796Q18. Feeling guilty when you get off track1.79 ± 0.9671.67 ± 0.8490.121−0.057 to 0.2980.181Q19. Feeling that dealing with diabetes is overwhelming2.51 ± 1.162.28 ± 1.040.2340.034 to 0.4340.022Q20. Feeling burned out2.07 ± 1.091.79 ± 0.8830.2770.093 to 0.4600.003PAID, problem areas in diabetes. † Data are presented as mean ± standard deviation. ‡ Mean differences were analyzed using paired t‐tests. Changes in individual PAID item scores from baseline to 6 months PAID, problem areas in diabetes. Data are presented as mean ± standard deviation. Mean differences were analyzed using paired t‐tests. Changes in HbA1c, a secondary outcome, are shown in Figure 2. HbA1c levels improved significantly at all time points from the initial visit to 6 months. The mean HbA1c was 7.49 ± 1.55% at the initial visit (95% CI: 0.285 to 0.544, P < 0.001), 7.08 ± 1.21% at 1 month (95% CI: 0.471 to 0.857, P < 0.001), 6.83 ± 0.98% at 2 months (95% CI: 0.495 to 0.946, P < 0.001), 6.77 ± 1.02% at 3 months (95% CI: 0.574 to 1.026, P < 0.001), 6.69 ± 0.90% at 4 months (95% CI: 0.572 to 1.020, P < 0.001), 6.70 ± 0.85% at 5 months (95% CI: 0.572 to 1.020, P < 0.001), and 6.68 ± 0.72% at 6 months (95% CI: 0.581 to 1.047, P < 0.001). Changes in HbA1c over 6 months Glycated hemoglobin (HbA1c) levels significantly improved at all time points during the 6‐month follow‐up. Data are presented as mean ± standard deviation. *P < 0.001 vs baseline (paired t‐test). Table 3 shows the results of the multivariable linear regression analysis with ΔPAID as the dependent variable. The coefficient of determination for the model was R 2 = 0.273 (adjusted R 2 = 0.240), and the overall model was statistically significant [F(6, 134) = 8.367, P < 0.001]. In the analysis of coefficients, only the baseline PAID score was significantly associated with ΔPAID (B = −0.459, 95% CI: −0.608 to −0.309, β = −0.493, P < 0.001), whereas age, sex, duration of diabetes, referral source, and baseline HbA1c were not significantly associated with ΔPAID (all P > 0.10). Table 3Multivariable linear regression analysis for ΔPAID § VariableUnstandardized coefficient, B 95% CIStandardized coefficient, β P‐valueAge (years)0.005−0.012 to 0.0220.0420.561Sex (female = 1) † −0.071−0.403 to 0.261−0.0280.673Duration of diabetes (years)0.009−0.008 to 0.0260.0750.304Referral source (specialist = 1) ‡ 0.084−0.268 to 0.4360.0300.642Baseline HbA1c (%)−0.049−0.189 to 0.092−0.0560.490Baseline PAID score § −0.459−0.608 to −0.309−0.493<0.001 Model statistics: R 2 = 0.273; Adjusted R 2 = 0.240; Model P < 0.001. † Sex was coded as female = 1, male = 0. ‡ Referral source was coded as specialist = 1, non‐specialist = 0. § ΔPAID indicates the change in PAID score from baseline to 6 months (6 months minus baseline). Multivariable linear regression analysis for ΔPAID § Model statistics: R 2 = 0.273; Adjusted R 2 = 0.240; Model P < 0.001. Sex was coded as female = 1, male = 0. Referral source was coded as specialist = 1, non‐specialist = 0. ΔPAID indicates the change in PAID score from baseline to 6 months (6 months minus baseline). Table 4 shows the results of the multivariable linear regression analysis with ΔHbA1c as the dependent variable. The coefficient of determination for the model was R 2 = 0.819 (adjusted R 2 = 0.811), and the overall model was statistically significant [F(6, 134) = 101.084, P < 0.001]. Among the explanatory variables, baseline HbA1c (B = −0.806, 95% CI: −0.879 to −0.734, β = −0.892, P < 0.001), duration of diabetes (B = 0.025, 95% CI: 0.013 to 0.036, β = 0.158, P < 0.001), and age (B = −0.014, 95% CI: −0.024 to −0.005, β = −0.115, P = 0.004) were significantly associated with ΔHbA1c. Sex, referral source, and baseline PAID score were not significantly associated with ΔHbA1c (all P > 0.10). Table 4Multivariable linear regression analysis for ΔHbA1c § VariableUnstandardized coefficient, B 95% CIStandardized coefficient, β P‐valueAge (years)−0.014−0.024 to −0.005−0.1150.004Sex (female = 1) † −0.025−0.176 to 0.126−0.0140.744Duration of diabetes (years)0.0250.013 to 0.0360.158<0.001Referral source (specialist = 1) ‡ −0.069−0.244 to 0.105−0.0370.436Baseline PAID score−0.001−0.008 to 0.005−0.0150.732Baseline HbA1c (%)−0.806−0.879 to −0.734−0.892<0.001Model statistics: R 2 = 0.819; Adjusted R 2 = 0.811; Model P < 0.001. † Sex was coded as female = 1, male = 0. ‡ Referral source was coded as specialist = 1, non‐specialist = 0. § ΔHbA1c indicates the change in HbA1c from baseline to 6 months (6 months minus baseline). Multivariable linear regression analysis for ΔHbA1c § Model statistics: R 2 = 0.819; Adjusted R 2 = 0.811; Model P < 0.001. Sex was coded as female = 1, male = 0. Referral source was coded as specialist = 1, non‐specialist = 0. ΔHbA1c indicates the change in HbA1c from baseline to 6 months (6 months minus baseline). In the subgroup analysis, patients with baseline PAID scores ≥50 showed significantly greater reductions in PAID scores than those with baseline PAID scores <50. This association was observed in both HbA1c subgroups (<7 and ≥7%). Among patients with HbA1c <7%, the mean difference in ΔPAID was 16.69 (95% CI 9.44–23.95, P < 0.001), whereas among those with HbA1c ≥7%, the mean difference was 9.63 (95% CI 2.99–16.26, P = 0.005) (Table 5). Table 5Subgroup analysis of ΔPAID § by baseline PAID and HbA1c categoryHbA1cBaseline PAID <50 ‡ Baseline PAID ≥50 † 95% CI P‐value<7%3.93 ± 11.29 (n = 40)−12.77 ± 18.25 (n = 26)9.44 to 23.95<0.001≥7%−1.19 ± 13.63 (n = 27)−10.81 ± 13.96 (n = 48)9.63 to 2.99–16.260.005Values are presented as mean ± standard deviation unless otherwise indicated. Independent‐samples t‐tests were used to compare ΔPAID between patients with baseline PAID ≥50 and PAID <50 within each HbA1c category. A negative mean difference indicates greater improvement (reduction) in the PAID ≥50 group. † High PAID group: baseline PAID score ≥50. ‡ Low PAID group: baseline PAID score <50. § Mean difference indicates the difference in ΔPAID between groups. Subgroup analysis of ΔPAID § by baseline PAID and HbA1c category Values are presented as mean ± standard deviation unless otherwise indicated. Independent‐samples t‐tests were used to compare ΔPAID between patients with baseline PAID ≥50 and PAID <50 within each HbA1c category. A negative mean difference indicates greater improvement (reduction) in the PAID ≥50 group. High PAID group: baseline PAID score ≥50. Low PAID group: baseline PAID score <50. Mean difference indicates the difference in ΔPAID between groups. Treatment was intensified in 78 patients (55.3%), unchanged in 51 patients (36.2%), and de‐intensified in 12 patients (8.5%). To evaluate the potential influence of treatment modifications on study outcomes, additional analyses were performed. Patients were categorized into an intensification group (n = 78) and a non‐intensification group (n = 63; no change, n = 51; de‐intensification, n = 12). Reductions in HbA1c were significantly greater in the intensification group than in the non‐intensification group (ΔHbA1c: −1.09 vs −0.47, P = 0.008). In contrast, changes in PAID scores did not significantly differ between groups (ΔPAID: −6.41 vs −3.59, P = 0.287). Analyses are summarized in Table 6. Table 6Additional analyses according to treatment modification statusVariableIntensification (n = 78) † Non‐Intensification (n = 63) ‡ P‐valueΔHbA1c § −1.09 ± 1.46−0.47 ± 1.250.008ΔPAID § −6.41 ± 13.99−3.59 ± 17.370.287Values are presented as mean ± standard deviation. † The non‐intensification group included patients with no change in glucose‐lowering therapy (n = 51) and those who underwent de‐intensification (n = 12). ‡ Intensification was defined as the addition of glucose‐lowering medications, dose escalation of existing medications, or initiation of injectable therapy. § ΔHbA1c and ΔPAID indicate changes from baseline to 6 months (6 months minus baseline). Additional analyses according to treatment modification status Values are presented as mean ± standard deviation. The non‐intensification group included patients with no change in glucose‐lowering therapy (n = 51) and those who underwent de‐intensification (n = 12). Intensification was defined as the addition of glucose‐lowering medications, dose escalation of existing medications, or initiation of injectable therapy. ΔHbA1c and ΔPAID indicate changes from baseline to 6 months (6 months minus baseline). DISCUSSION In this study, we retrospectively examined the effects of DSMES delivered in an outpatient setting using the Diabetes Card System on diabetes‐related distress and glycemic control. As a result, the total PAID score significantly improved over a 6‐month period, clearly indicating a reduction in psychological burden. In contrast, scores increased for some items (Q1 and Q3). This finding may reflect that the educational intervention enhanced patients' understanding of diabetes and awareness of their condition, thereby making previously latent anxieties and concerns more explicit. HbA1c levels also decreased significantly during the same period, suggesting that support using the Diabetes Card System may have led to changes in patients' self‐management and treatment‐related behaviors. Furthermore, multivariable analysis showed that the only factor associated with ΔPAID was the baseline PAID score, whereas age, sex, duration of diabetes, referral source, and baseline HbA1c were all non‐significant. In contrast, ΔHbA1c was significantly associated with baseline HbA1c, duration of diabetes, and age, while baseline PAID was not associated with ΔHbA1c. These findings suggest that psychological improvement and glycemic improvement are based on different mechanisms and may progress as independent processes. The improvement in PAID observed in this study may be associated with the structured support design of the Diabetes Card System as a visual educational tool and with its content being tailored to individual culture and daily practices 12 , 13 . Improved disease understanding through structured education may have enhanced patients' confidence in self‐management and reduced uncertainty regarding diabetes care, thereby contributing to reductions in diabetes‐related distress. This interpretation is consistent with the current DSMES consensus framework, which emphasizes that diabetes education improves not only knowledge but also self‐efficacy, self‐management behaviors, and psychosocial outcomes. Structured and patient‐centered educational interventions may help individuals better understand their condition, increase confidence in self‐management, and develop more effective coping strategies, thereby reducing diabetes‐related distress. Furthermore, ongoing support integrated into routine clinical practice is considered an essential component of DSMES, as continuous reinforcement may facilitate sustained behavioral and psychological benefits 11 , 12 , 13 , 14 . Outpatient diabetes care in Japan faces structural challenges due to strict time constraints, making it difficult to sufficiently enhance patient understanding within limited consultation time 15 . The Diabetes Card System was designed to address this challenge and is expected to reduce working memory load through visual information while correcting variability in educational delivery among providers (intervention fidelity) 16 , 17 . As a result, consistent psychological improvement independent of patient background may have been achieved. From the perspective of quality assurance in DSMES, this finding is of substantial significance, and the Diabetes Card System may have international value as a diabetes education model developed in Japan. The association between baseline PAID score and improvement in PAID observed in this study is consistent with previous reports 18 , 19 , 20 . Part of this finding may also reflect regression to the mean, whereby patients with higher baseline scores tend to show greater improvement over time regardless of intervention. However, diabetes‐related distress is inherently influenced by patients' values, life circumstances, and treatment experiences 8 , 21 , 22 , and it has been reported that patients with higher levels of distress have greater potential for improvement, making the effects of DSMES more apparent 23 . In addition, a change of 5 points or more in the PAID score is considered the minimal clinically important difference (MCID) 24 . Therefore, the magnitude of improvement observed in this study represents not merely a statistical difference but also a clinically meaningful reduction in psychological burden in real‐world practice. A key feature of the Diabetes Card System is its use of visually organized educational materials to facilitate patient understanding. Compared with verbal information, visual information imposes a lower burden on working memory, and ease of understanding is regarded as a key factor directly linked to behavioral change in frameworks such as the Health Belief Model 2 4 , 25 , 26 and self‐efficacy theory 22 , 27 , 28 . By presenting information through concrete tools in the form of cards, patients may find it easier to perceive their condition as personally relevant, which could lead to greater acceptance of treatment and increased patient engagement. In this study, PAID scores improved regardless of background factors such as age, sex, and referral source, which is thought to reflect the effects of reduced cognitive load and more standardized understanding across patients. In the present study, educational sessions required only approximately 3 min per visit, suggesting that the system may be feasible even in busy outpatient settings. The Diabetes Card System may therefore provide a structured and visual approach to continuous patient education that can be readily incorporated into routine outpatient practice. The finding that baseline HbA1c level and duration of diabetes were associated with improvement in HbA1c is reasonable and consistent with existing evidence in diabetes research 29 , 30 . Greater improvement among patients with higher baseline HbA1c is likely attributable to pathophysiological characteristics in which the effects of treatment intensification and behavioral change are more readily reflected 31 . Multiple studies have reported that patients with a longer duration of diabetes show smaller reductions in HbA1c after therapeutic interventions, suggesting decreased treatment responsiveness associated with long‐standing disease 32 , 33 . In this study, age was also a significant factor, which may reflect reduced metabolic function and treatment responsiveness in older patients 34 . In contrast, baseline PAID was not associated with ΔHbA1c, supporting one of the main implications of this study that psychological improvement and glycemic improvement occur independently. This independence between improvement processes is clinically important. Glycemic improvement primarily depends on pharmacological treatment, behavioral change, and metabolic characteristics, whereas psychological improvement is shaped by different factors, such as understanding of educational content, communication with healthcare providers, patients' values toward treatment, and a sense of reassurance 35 , 36 . Therefore, these findings indicate that to appropriately evaluate the effects of DSMES, psychological outcomes should be treated as independent endpoints in addition to glycemic outcomes 23 . Additional analyses demonstrated that reductions in HbA1c were significantly greater among patients who underwent treatment intensification than among those without treatment intensification. This finding suggests that changes in pharmacological therapy contributed, at least in part, to the observed improvements in glycemic control. Therefore, improvements in HbA1c cannot be attributed solely to DSMES using the Diabetes Card System. In contrast, changes in PAID scores did not differ significantly according to treatment intensification status. This finding suggests that the observed reduction in diabetes‐related distress may not be explained solely by medication adjustments. Rather, factors related to the educational intervention itself, such as improved disease understanding, enhanced communication with healthcare providers, and increased confidence in self‐management, may have contributed to the psychological benefits observed in this study. In the subgroup analysis of PAID, greater reductions in PAID scores were consistently observed among patients with higher baseline PAID scores. This finding suggests that patients experiencing greater diabetes‐related distress may derive particular benefit from DSMES using the Diabetes Card System. Notably, this association remained significant in both HbA1c subgroups (<7% and ≥7%), suggesting that the potential for psychological improvement was more strongly related to baseline distress than to baseline glycemic status. These findings support the concept that psychological improvement may represent a universal effect that can be achieved independently of metabolic management. This study has several limitations. First, because this was a retrospective, single‐center observational study without a control group, causal inferences should be interpreted with caution. Second, although additional analyses according to treatment modification status were performed, residual confounding due to medication changes cannot be completely excluded. Therefore, the observed improvements in HbA1c should be interpreted with caution. Third, the Japanese version of the PAID questionnaire used in this study has not been formally validated. Therefore, the findings should be interpreted with consideration of potential limitations in measurement validity. Fourth, other psychosocial factors that may influence distress, such as depression, anxiety, and self‐efficacy, were not evaluated. Fifth, although multiple Certified Diabetes Care and Education Specialists used the Diabetes Card System in this study, complete consistency in educational content and instructional skills (intervention fidelity) was not fully ensured. Variability in how the cards were used and in the depth of explanations may have existed, potentially introducing some degree of measurement error in psychological outcomes. However, the observation that meaningful effects were obtained despite such real‐world variability indicates the high implementation feasibility of the Diabetes Card System and supports its suitability for routine clinical practice. Finally, because the Diabetes Card System was developed specifically in Japan, its international generalizability requires further investigation. DSMES using the Diabetes Card System may represent a practical approach associated with improvements in glycemic control and reductions in diabetes‐related distress in outpatient care. Future studies using prospective designs and control groups, together with optimization of the intervention process including standardization of educator training, are warranted to further clarify the effectiveness of support using the Diabetes Card System. CONCLUSION In conclusion, DSMES using the Diabetes Card System may represent a practical approach associated with improvements in glycemic control and reductions in diabetes‐related distress in routine outpatient care. DISCLOSURE The author declares no conflict of interest. Approval of the research protocol: The study protocol and public disclosure of the research were reviewed and approved by the JADEC Clinical Research Ethics Committee (approval No. 250801–2‐1). Informed consent: N/A. Registry and the registration no. of the study/trial: N/A. Animal studies: N/A.