Key points are not available for this paper at this time.
To the Editor: Type 2 diabetes (T2DM) is a significant health issue in China, with the age-standardized prevalence rising 2.5-fold from 1990 to 9.96% in 2021.1 Despite this rise, only 32.9% of patients received treatment, and just 50.1% achieved controlled blood sugar.1 The Chinese Diabetes Society (CDS) advocates for self-monitoring of blood glucose (SMBG) to help patients track levels and manage their condition.2 While SMBG has been shown to lower HbA1c in type 1 diabetes and untreated type 2 diabetes (T2D), large-scale evidence for its effect in insulin-treated T2D is still limited.3 Digital tools like mobile apps and WeChat mini-programs are increasingly used in treating conditions such as cardiovascular disease, insomnia, and diabetes.4 Therefore, we conducted TRIO, a mobile health-based diabetes management program for T2D patients starting basal insulin in China. TRIO does not stand for a longer phrase. Instead, it represents the three key pillars of diabetes management: physician, nurse, and patient, working in unison. Glucose readings and insulin dosages were automatically uploaded to a WeChat mini program for monitoring and adjustment. The program also featured glucose data visualization, diabetes education, and patient-physician communication. This study, part of the TRIO program, assessed SMBG frequency and adherence to guidelines in T2D patients starting basal insulin using digital tools. It also explored the link between SMBG adherence and glycemic control, highlighting the potential of mobile health technologies for improving self-management. This 12-week prospective cohort study within the TRIO program involved 113,074 patients from 578 hospitals between December 2019 and December 2020. Eligible participants were adults with T2D, inadequate glycemic control (HbA1c ≥7%), and initiating basal insulin. Patients were followed up at 1, 2, 4, 8, and 12 weeks, with HbA1c measured at week 12. SMBG and fasting blood glucose (FBG) data were collected via smart glucose monitors. Adherence to SMBG guidance was categorized by testing frequency. Outcomes included glycemic control (HbA1c, FBG) and hypoglycemia rates. Baseline characteristics were summarized using percentages (categorical) and mean ± standard deviation (continuous). Differences by adherence were analyzed with chi-squared and One-way ANOVA (Analysis of variance) tests. SMBG and FBG data were described as means. Glycemic control outcomes were assessed with ANCOVA (Analysis of covariance) and logistic regression, adjusting for confounders. Hypoglycemia rates were analyzed using logistic and Poisson regression, including a composite endpoint of glycemic control without hypoglycemia. All analyses were two-tailed and conducted using SAS 9.4 (SAS Institute Inc., Carey, NC, USA) with the significance level set at 0.05. As part of the TRIO program, the study was conducted among a subset of patients in the TRIO program who were dispensed with smart glucose monitoring equipment. A total of 14,084 patients from 242 hospitals were enrolled Supplementary Figure 1, https://links.lww.com/CM9/C592. The mean age was 55.6 years, with 57.9% (8158/14,084) male. Comorbidities were present in 64.3% (8227/12,804), with peripheral neuropathy, hypertension, and hyperlipidemia being the most common. Approximately 30.8% (4319/14,014) had a family history of diabetes. Most patients (73.3%, 10,326/14,084) tested SMBG <4 times per week and had longer diabetes duration, higher baseline SBP and DBP, and lower HbA1c compared to those with better adherence Supplementary Table 1, https://links.lww.com/CM9/C592. Among 14,084 patients, 539,672 SMBG records were uploaded by 12,235 patients, while 1849 patients did not upload any data. The average SMBG testing frequency was 3.19 times per week. The percentage of patients uploading SMBG data declined from 76.04% at week 1 to 39.24% at week 12 Supplementary Table 2, https://links.lww.com/CM9/C592. Among patients uploading SMBG data, testing frequency dropped from 5.94 ± 7.64 at week 1 to 1.67 ± 3.52 at week 12 Supplementary Figure 2A, https://links.lww.com/CM9/C592. Patients testing <4 and 4–7 times per week showed consistent declines, while those testing ≥7 times experienced an initial increase, followed by a significant drop. For FBG data, the percentage of patients uploading results decreased over time, with a slight initial increase before a continuous decline in testing frequency Supplementary Table 3 and Supplementary Figure 2B, https://links.lww.com/CM9/C592. A total of 3775 patients had HbA1c data at both baseline and follow-up, with mean HbA1c decreasing from 9.36% to 6.98%, representing an unadjusted change of −2.38%. All adherence groups showed reductions, with changes of −2.09%, −2.74%, and −3.49% for <4, 4–7, and ≥7 times of SMBG testing per week, respectively. After adjusting for baseline HbA1c and confounders, least square means for HbA1c change were −2.34%, −2.44%, and −2.51% for the respective groups. Compared to the <4 times group, greater reductions were observed in the more adherent groups: LS mean differences of −0.10 (95% confidence interval CI: −0.18, −0.02) for 4–7 times and −0.17 (95% CI: −0.26, −0.08) for ≥7 times (P for trend <0.001). Notably, patients testing ≥7 times had higher baseline HbA1c, which was accounted for in the analysis Supplementary Table 4, https://links.lww.com/CM9/C592. At week 12, 52.9% of patients achieved HbA1c <7%. Significantly higher proportions were seen in patients testing 4–7 times (Odds ratio OR: 1.28, 95% Confidence Interval CI: 1.06, 1.53) and ≥7 times (OR: 1.46, 95% CI: 1.18, 1.81) compared to those testing <4 times Supplementary Table 4, https://links.lww.com/CM9/C592. Using HbA1c cut-offs of <7.0% for patients <60 years and <7.5% for those ≥60 years, 64.4% (3045/4725) of patients reached target levels at week 12. Patients testing 4–7 times and ≥7 times weekly had numerically higher proportions reaching target HbA1c compared to those testing <4 times. However, the difference between <4 and 4–7 times was not significant (63.5% vs. 66.4%; OR 95% CI: 1.19 0.99, 1.44). The difference between <4 and ≥7 times remained significant (63.5% vs. 67.3%; OR 95% CI: 1.41 1.13, 1.75) Supplementary Table 4, https://links.lww.com/CM9/C592. Regarding FBG data, among 5471 patients with both baseline, levels decreased from 9.43 mmol/L to 7.09 mmol/L, with an unadjusted mean change of –2.43 mmol/L. All adherence groups showed reductions, but better adherence correlated with significantly greater changes. Using FBG cut-offs of <7.0 mmol/L or <6.1 mmol/L, significantly more patients in the 4–7 and ≥7 times SMBG testing groups achieved target levels compared to those testing <4 times Supplementary Table 4, https://links.lww.com/CM9/C592. Subgroup analyses in Supplementary Table 5, https://links.lww.com/CM9/C592 show that in patients with baseline HbA1c ≥9.0%, the HbA1c reduction for those testing 4–7 and ≥7 times weekly compared to <4 times was significant, while it was not significant in the <9.0% subgroup. Significant interaction terms (P <0.001) indicated a stronger positive effect of adherence on HbA1c change in the ≥9.0% group. For reaching target HbA1c levels, the ≥9.0% subgroup showed significantly higher proportions, except between <4 times and 4–7 times using age-based cut-offs. This trend was not significant in the <9.0% subgroup due to smaller sizes. The benefit of testing ≥7 times vs. <4 times was more notable in the ≥9.0% group (P = 0.039, OR: 1.56 vs. 1.09) Supplementary Table 5, https://links.lww.com/CM9/C592. The favorable trends for FBG change and target achievement remained consistent across both subgroups Supplementary Table 5, https://links.lww.com/CM9/C592. A total of 9667 hypoglycemic events were identified from 539,672 SMBG records across 12,235 patients, averaging 3.16 events per person-year. Event rates correlated with adherence: 1.66 events for <4 SMBG tests weekly, increasing to 3.80 for 4–7 tests and 10.0 for ≥7 tests. Using a cut-off of SMBG <3.0 mmol/L, the overall event rate was 1.26 events per person-year, also rising with adherence. Event rates per test were similar regardless of cut-off Supplementary Table 6, https://links.lww.com/CM9/C592. In the analysis of the composite endpoint, the proportion of patients achieving glycemic control without hypoglycemic events tended to decrease with adherence to SMBG guidance. However, this decrease was not significant when comparing patients with <4 vs. 4–7 SMBG tests per week for FBG <6.1 mmol/L without SMBG ≤3.9 mmol/L (P = 0.746) Supplementary Table 7, https://links.lww.com/CM9/C592. Sensitivity analyses using a trajectory model Supplementary Tables 8 and 9, Supplementary Figure 3, https://links.lww.com/CM9/C592 or FBG adherence groups Supplementary Tables 10–14, https://links.lww.com/CM9/C592 or treating numbers of SMBG tests or days per week as continuous variable in the model Supplementary Table 15, https://links.lww.com/CM9/C592 showed that better patient adherence was associated with better glycemic control, which was consistent with the base case findings. In this study, most patients adhered to self-monitoring guidance during the first week, with a majority uploading SMBG data and testing more frequently than recommended. This initial adherence rate surpassed that of previous cross-sectional studies in China, where only 27.5% of patients were deemed adherent.5 The higher adherence in our study may be attributed to the engagement fostered by providing smart glucose monitoring devices at enrollment. While adherence decreased after the first week, high-adherence patients continued to test more frequently than others. This study provides valuable longitudinal data on SMBG adherence in China, addressing a significant evidence gap from previous research.5 The decreasing trend in adherence indicates a need for ongoing motivation. While previous studies show that counseling improves adherence, they were limited in sample size.6 Given the large T2D population in China, mHealth innovations, such as regular reminders and interactive interfaces, could enhance patient engagement and motivation. Moreover, SMBG is recommended by the American Diabetes Association, International Diabetes Federation, and CDS for diabetes management. Its benefits in glycemic control are well-documented in type 1 diabetes and non-insulin-treated T2D patients. This study addresses the unclear relationship between SMBG adherence and glycemic control in insulin-treated T2D patients in real-world settings. In conclusion, this is a large-scale cohort study in China assessing adherence to blood glucose self-monitoring guidance in T2D patients. It highlights the importance of sustained patient-driven behavior changes for effective management. The study found that adherence to self-monitoring guidance was positively associated with glycemic control in patients initiating basal insulin. Conflicts of interest None.
Guo et al. (Mon,) studied this question.