Impact of changes in diabetes care models and advanced technologies on two-year glycemic control after COVID-19 lockdown in adults with Type 1 Diabetes Authors: Lutgarda Bozzetto, Roberta Lupoli, Maria Masulli, Annalisa Creanza, Raffaele De Angelis, Clemente Giglio, Manuela Macera, Carmen Rainone, Giovanni Annuzzi, Brunella Capaldo Published: 2025 ABSTRACT AimsThe COVID-19 pandemic led to significant shifts in lifestyle and clinical management for individuals with Type 1 Diabetes (T1D) that, during the COVID-19 lockdown, showed an improved glycemic control, characterized by increased time-in-range (TIR) and reduced glucose variability. This study aimed to assess the long-term trends and factors influencing glycemic control over two years post-lockdown in adults with T1D.MethodsA longitudinal study was conducted on 207 adults with T1D (mean age 38 ± 13 years), with 50.2% using multiple daily insulin injections and 49.8% on continuous subcutaneous insulin infusion. Continuous glucose monitoring (CGM) metrics—including TIR, time-above-range (TAR); time-below-range (TBR), and coefficient of variation (CV)— were assessed before, during, and at 6, 12, and 24 months post-lockdown. Participants also completed an online questionnaire evaluating lifestyle habits before, during, and one year after lockdown.ResultsThe glycemic improvements observed during lockdown persisted for two years. TIR 70–180 mg/dl remained significantly higher, while TAR > 250 mg/dl and TBR < 54 mg/dl were significantly lower at all post-lockdown time points (p < 0.05 for all). These improvements were not attributed to lifestyle changes, which remained relatively stable one year after lockdown. Telemedicine was implemented being the prevalent model of care during this period. Notably, the improvement in glycemic control concerned the participants transitioning to advanced insulin delivery systems during the two-year observational period (p < 0.001).ConclusionsGlycemic control improvements in adults with T1D were sustained for two years post-lockdown and related to transitioning to more advanced insulin delivery systems.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13098-025-02022-x. Introduction The COVID-19 pandemic led to significant shifts in healthcare delivery, profoundly affecting the management of chronic diseases, including diabetes. The necessity for social distancing measures expedited the transition of most patients with type 1 diabetes (T1D) from traditional in-person consultations to a telemedicine-based care model. This shift was facilitated by advancements in remote glycemic monitoring technologies and cloud-based data management systems, allowing healthcare providers to maintain continuous patient oversight even during periods of home confinement. Several studies have investigated the effects of lockdown restrictive measures on glycemic control among individuals with T1D [1, 2] reporting an overall improvement characterized by increased time in blood glucose control, decreased glucose variability, and stable or slightly reduced hypoglycemia rates. This unexpected improvement in glycemic control has been attributed to several factors, including more structured daily routines, increased self-care practices, reduced physical activity levels, and more regular sleep patterns [1, 3, 4]. Notably, the decrease in physical activity was associated with reduced glucose variability [1], underscoring the complex relationship between nutrient intake, insulin therapy, and physical exercise in people with T1D. The use of telemedicine has also played a role, as demonstrated by the observation that patients who attended telemedicine compared to non-attenders had a significant improvement in glucose control [5]. As public health restrictions have eased and society has begun returning to pre-pandemic routines, there is a growing interest in understanding the long-term impact of these changes on glycemic control. Few studies have longitudinally assessed glycemic control post-lockdown [6, 7], with some indicating a sustained improvement in glucose management that coincided with an increased utilization of wearable diabetes devices among young adults with T1D [8]. Therefore, in light of the transformative impact of the pandemic on diabetes care models and the rapid adoption of new technologies, this study aims to explore the trajectory of glucose control in the two years following the lifting of COVID-19 strict restrictions in adults with T1D and to identify factors influencing their glucose metrics during this period. Methods A cohort of adults with T1D, 96 females and 111 males, mean age 38.4 ± 12.7 years (M ± SD) attending the Diabetes Outpatient Clinic of the Federico II University Hospital of Naples, users of continuous glucose monitoring (CGM) systems for at least 6 months before enrolment, participated in the study. All of them had spent a two-week period with >70% CGM use before (January-February 2020) and during the first lockdown (March-April 2020) keeping the same CGM system. This cohort’s data on glucose control during the lockdown have been previously published [1]. One hundred-four patients were on multiple daily insulin injections (MDI) and 103 were on continuous subcutaneous insulin infusion systems, of whom 44 on non-integrated insulin pump (CSII), 39 on Low-Glucose Suspend system (LGS), 19 on Hybrid Closed-Loop system (HCL), and 1 on Advanced Hybrid Closed-Loop system (AHCL) (Table 1). For the participation in this retrospective study, a written informed consent was obtained from all subjects. All methods were carried out in accordance with relevant guidelines and the Federico II University Ethics Committee regulations. Table 1Insulin regimen and continuous glucose monitoring (CGM) systems used by the participants before and 2 years after lockdownBeforelockdown2 years after lockdownMultiple daily injections (MDI) (n, %)104 (50.2)96 (46.4)CGM systems used by the MDI participants (n,%) Freestyle libre99 (95.2)87 (90.6) Eversense4 (3.8)3 (3.1) Guardian 31 (1.0)2 (2.1) Dexcom G602 (2.1) Glucomen day01 (1.0)Continuous subcutaneous insulin infusion (n,%)103 (49.8)111 (53.6) Non-integrated insulin pump4430 Low-glucose suspend398 Hybrid closed-loop197 Advanced hybrid closed-loop166 Table 1Insulin regimen and continuous glucose monitoring (CGM) systems used by the participants before and 2 years after lockdownBeforelockdown2 years after lockdownMultiple daily injections (MDI) (n, %)104 (50.2)96 (46.4)CGM systems used by the MDI participants (n,%) Freestyle libre99 (95.2)87 (90.6) Eversense4 (3.8)3 (3.1) Guardian 31 (1.0)2 (2.1) Dexcom G602 (2.1) Glucomen day01 (1.0)Continuous subcutaneous insulin infusion (n,%)103 (49.8)111 (53.6) Non-integrated insulin pump4430 Low-glucose suspend398 Hybrid closed-loop197 Advanced hybrid closed-loop166 Insulin regimen and continuous glucose monitoring (CGM) systems used by the participants before and 2 years after lockdown Starting from March 2020, our centre moved to a telehealth-based care model and continued to offer remote care through video calls, messages, e-mails, and videos. All information collected during the virtual visits was uploaded to electronic health records and the visit report was shared via e-mail. The CGM systems used by the participants were FreeStyle Libre (n = 130), Guardian 3 (n = 47), Dexcom G6 (n = 18), Eversense (n = 12), and Glucomen Day (n = 1). The download of the data was carried out using dedicated platforms (Libreview, Abbott Diabetes Care; Care Link, Medtronic; Diasend; Dexcom Clarity, Dexcom Inc; Eversense Diabetes Management System). The following two-week CGM metrics were analyzed: time-in-target range (TIR) (70–180 mg/dl [3.9–10.0 mmol/L]), time-above-target range (TAR) (> 180 mg/dl [10.0 mmol/L]) and (> 250 mg/dl [13.9 mmol/L]), time-below-target range (TBR) (< 70 mg/dl [3.9 mmol/L] and < 54 mg/dl [3.0 mmol/L]) expressed as the percentage (%) of all CGM readings, mean glucose, glucose measurement indicator (GMI), and glycemic variability expressed by the coefficient of variation (CV%). CGM metrics were collected before, during, and 6, 12, and 24 months after the end of lockdown. Qualitative information on lifestyle was obtained through an online questionnaire administered during the lockdown and after 12 months. The evaluated items were: (1) physical activity (a score based on type and frequency of activity), (2) changes in dietary habits (food intake, meal timing, number of snacks), (3) changes in sleeping pattern (duration, bedtime, awake time). The one-year questionnaire also investigated modalities and satisfaction/preferences of the type of visits (remote/face-to-face) made during the observational period. Statistical analysis Continuous variables are expressed as mean ± standard deviation unless otherwise stated; categorical variables are expressed as percentages. Changes over time of each parameter were assessed by analysis of variance for repeated measures, and a post-hoc t-test analysis was performed between time points. A general repeated measure general linear model analysis was performed having time as intra-subject variable and sensor device and insulin delivery mode as inter-subject factors. Questionnaire-derived lifestyle data were analysed by chi-squared test. For all analyses, a two-sided p-value < 0.05 was considered significant. Statistical analysis was performed with the Statistical Package for Social Science SPSS 27 software (SPSS Inc). Results The glucose metrics monitored over the two-year study period are shown in Fig. 1. Relative to the pre-lockdown period (55.6 ± 17.6%), TIR 70–180 mg/dl was significantly higher during the lockdown (58.2 ± 18.1%, p = 0.002) and at 6 (58.7 ± 15.8%, p = 0.003), 12 (61.2 ± 15.1%, p < 0.001), and 24 months post-lockdown (59.9 ± 17.8%, p = 0.001). TAR 180–250 mg/dl decreased from the pre-lockdown value (26.8 ± 11.2%) to 24.8 ± 9.9% (p = 0.033) at 2 years, while TAR > 250 mg/dl was significantly lower than the pre-lockdown value (14.8 ± 15.5%) at 6 months (12.0 ± 10.2%, p = 0.007), 1 year (10.7 ± 9.9%, p < 0.001), and 2 years (13.0 ± 11.5%, p = 0.048). TBR 70 − 54 mg/dl remained relatively unchanged throughout the observation period, whereas TBR < 54 mg/dl was significantly lower than the pre-lockdown value (1.42 ± 2.39%) at 6 months (0.55 ± 1.16%), 1 year (0.63 ± 1.47%), and 2 years (0.57 ± 1.32%) (p < 0.001 for all). Glycemic variability, as expressed by the coefficient of variation (CV), was 35.9 ± 7.0% before the lockdown, decreased significantly during the lockdown (34.7 ± 6.3%, p < 0.001), remaining significantly lower at 6 months (34.9 ± 5.1%, p = 0.007), but not at 1 year (35.2 ± 5.8%, p = 0.22) and 2 years (35.5 ± 6.7%, p = 0.48). Compared to the pre-lockdown period (7.7 ± 1.3%), GMI decreased to 7.3 ± 0.7%, 7.2 ± 0.7%, and 7.4 ± 0.8% at 6 months, 1 year, and 2 years, respectively (p < 0.001 for all). Fig. 1Two-week CGM metrics before, during, and 6, 12, and 24 months after lockdown in the study participants. TIR, time-in-range; TAR, time-above-range; TBR, time-below-range; GMI, glucose measurement indicator. Mean ± SEM. *p < 0.05 vs. before lockdown Two-week CGM metrics before, during, and 6, 12, and 24 months after lockdown in the study participants. TIR, time-in-range; TAR, time-above-range; TBR, time-below-range; GMI, glucose measurement indicator. Mean ± SEM. *p < 0.05 vs. before lockdown Similar results were observed when the analysis was restricted to the participants who had the same CGM brand/model in the pre-lockdown period and at two-year final observation (n = 184). Compared to pre-lockdown, after two years, TIR was increased (60.2 ± 17.8 vs. 55.6 ± 17.7, p = 0.002), TAR180 − 250 and TAR> 250 were decreased (24.4 ± 9.8 vs. 26.9 ± 10.9, p = 0.017, and 13.3 ± 11.9 vs. 14.9 ± 15.4, p = 0.205, respectively), TBR70 − 54 was not significantly changed (2.51 ± 2.87 vs. 2.82 ± 3.00, p = 0.233). and TBR< 54 was decreased (0.557 ± 1.31 vs. 1.38 ± 2.34, p < 0.001). The results on lifestyle changes obtained from the online questionnaire administered one year after lockdown are reported in Table 2. Compared to the lockdown period, physical activity, food intake, and body weight were mostly unchanged. There was a trend towards a shorter sleep duration as participants woke up earlier than during the lockdown. Table 2Changes in participants’ lifestyle one year after lockdown compared to lockdown, as derived from the online questionnaire (n = 180)UnchangedIncreasedDecreasedTotal physical activitya95 (55.2)34 (19.8)43 (25.0)Food amount106 (58.9)23 (12.8)51 (28.3)Regularity of mealtimes122 (67.8)25 (13.9)33 (18.3)Number of snacks110 (61.1)26 (14.4)44 (24.4)Sleep duration103 (57.2)22 (12.2)55 (30.6)Body weight80 (44.4)60 (33.3)40 (22.2)UnchangedLaterEarlierBedtime117 (65.0)22 (12.2)41 (22.8)Waking time98 (54.4)11 (6.1)71 (39.4)a n = 172. Data are n (%). χ2 test, p-value < 0.001 for all variables Table 2Changes in participants’ lifestyle one year after lockdown compared to lockdown, as derived from the online questionnaire (n = 180)UnchangedIncreasedDecreasedTotal physical activitya95 (55.2)34 (19.8)43 (25.0)Food amount106 (58.9)23 (12.8)51 (28.3)Regularity of mealtimes122 (67.8)25 (13.9)33 (18.3)Number of snacks110 (61.1)26 (14.4)44 (24.4)Sleep duration103 (57.2)22 (12.2)55 (30.6)Body weight80 (44.4)60 (33.3)40 (22.2)UnchangedLaterEarlierBedtime117 (65.0)22 (12.2)41 (22.8)Waking time98 (54.4)11 (6.1)71 (39.4)a n = 172. Data are n (%). χ2 test, p-value < 0.001 for all variables Changes in participants’ lifestyle one year after lockdown compared to lockdown, as derived from the online questionnaire (n = 180) a n = 172. Data are n (%). χ2 test, p-value < 0.001 for all variables A high percentage (85%) of the participants utilized remote visits, and most of them expressed a high/medium level of satisfaction with remote visits (Fig. 2). Asked about the modality of visit they would prefer to continue with, the majority of the participants (67%) indicated their preference for the combination of the two modalities. Fig. 2Modality of diabetes visits in the first year after the Covid-19 lockdown, satisfaction with remote visits, and modality the participants would prefer to continue with, as assessed by the online questionnaire Modality of diabetes visits in the first year after the Covid-19 lockdown, satisfaction with remote visits, and modality the participants would prefer to continue with, as assessed by the online questionnaire Table 2 outlines the changes in insulin regimen modality from baseline to the end of the observation period. Two years after lockdown, the number of participants on CSII only slightly increased, while there was a striking difference in the number of participants on Advanced Hybrid Closed-Loop systems who were 60% of those on CSII. In Fig. 3, a Sankey diagram shows the individual changes in insulin delivery mode between pre-lockdown and the two-year observation. Fig. 3Individual changes in insulin delivery mode between pre-lockdown and the two-year observation. MDI, multiple daily injections; CSII, non-integrated insulin pump; LGS, low-glucose suspend system; HCL, hybrid closed-loop system; AHCL, advanced hybrid closed-loop system Individual changes in insulin delivery mode between pre-lockdown and the two-year observation. MDI, multiple daily injections; CSII, non-integrated insulin pump; LGS, low-glucose suspend system; HCL, hybrid closed-loop system; AHCL, advanced hybrid closed-loop system To assess the impact of the changes in insulin regimen on glycemic control, we compared glucose metrics between participants who maintained the same insulin regimen and those who switched to a more complex insulin delivery system. TIR 70–180 mg/dl was significantly higher in participants who changed their insulin regimen (n = 70) compared to those who did not (n = 125) (69.0 ± 14.8% vs. 54.7 ± 17.3%; p < 0.001) (Fig. 4). Both TAR > 180 mg/dl (21.1 ± 8.15% vs. 26.8 ± 10.2%; p < 0.001) and TBR < 70 mg/dl (1.65 ± 1.81% vs. 2.83 ± 3.10%; p = 0.004) were significantly lower in participants who transitioned to a more complex insulin regimen (Fig. 4). Fig. 4CGM metrics two years after lockdown in the participants who maintained the same insulin regimen as before lockdown (NoChange, n = 125) and those who switched to a more complex insulin delivery system (Change, n = 70). TIR, time-in-range; TAR, time-above-range; TBR, time-below-range. Data are expressed as Mean ± SEM CGM metrics two years after lockdown in the participants who maintained the same insulin regimen as before lockdown (NoChange, n = 125) and those who switched to a more complex insulin delivery system (Change, n = 70). TIR, time-in-range; TAR, time-above-range; TBR, time-below-range. Data are expressed as Mean ± SEM To disentangle the effects of changing sensor type or insulin delivery mode during the observation period on CGM metrics, a repeated measure general linear model analysis was performed having time as intra-subject variable and sensor device and insulin delivery mode as inter-subject factors. The results show that the changes during the observation period were independent of the changes in CGM type while restricted to the individuals who changed insulin delivery mode by transitioning to a more advanced technology (Fig. 5). Fig. 5Estimated marginal means with 95% IC of glucose time-in-range (TIR) during the observation period, according to changes in type of CGM or insulin delivery mode, by repeated measures general model analysis Estimated marginal means with 95% IC of glucose time-in-range (TIR) during the observation period, according to changes in type of CGM or insulin delivery mode, by repeated measures general model analysis Discussion Results from our study, along with findings from other cohorts, consistently indicate that glycemic control in people with T1D improved during the COVID-19 pandemic lockdown compared to pre-lockdown periods. This improvement was likely driven by a more structured lifestyle, enhanced self-management practices—including more consistent meal and insulin timing—and reduced physical activity [1–4]. However, this positive trend was observed primarily in cohorts from developed countries. In contrast, glycemic control often deteriorated in developing countries, likely due to restricted access to telemedicine and limited technological resources [9–11]. Our study, which assessed CGM metrics and factors influencing glucose control in adults with T1D over the two-year period following the COVID-19 lockdown, shows that the improvement observed during the lockdown was not only sustained but also enhanced in some cases over the subsequent two years. Specifically, we found a notable increase in TIR 70–180 mg/dl, alongside reductions in TAR, particularly for TAR >250 mg/dl, and TBR < 54 mg/dl, when compared to pre-lockdown levels. These findings align with similar observations in a cohort of emerging adults with T1D attending a multidisciplinary clinic in a high-income country, where glycated hemoglobin levels were, on average, 0.4% lower than prior to the pandemic [8]. When considering the factors contributing to this sustained improvement in glucose control, we noted that, one-year post-lockdown, lifestyle changes were relatively minimal. Physical activity levels remained stable overall, although there was a noticeable shift in eating habits, with more frequent snacking reported by many patients, leading to weight gain in about 30% of the cohort. This is consistent with global trends identified through smartphone health apps that track daily step counts. Research has shown that, as of early 2022, worldwide step counts had not returned to pre-pandemic levels and remained significantly lower than baseline levels [12]. Thus, while physical activity remained reduced, it is unlikely that lifestyle changes alone could account for the observed sustained glycemic improvements. Our data strongly suggest that the increased adoption of advanced insulin delivery systems played a key role in the long-term improvement in glycemic control. Notably, the number of patients transitioning to hybrid closed-loop insulin delivery systems increased from 1 patient prior to the lockdown to 66 patients after two years. This transition was associated with significant improvements in glucose metrics, including increased TIR and reductions in both TAR and TBR among patients who upgraded their insulin delivery systems. Notably, these individuals exhibited a mean 13.6% increase in TIR after the switch compared with baseline (95% CI 9.5–17.8, p < 0.001). These findings are consistent with previous research, which has demonstrated that transitioning from multiple daily injections or open-loop insulin delivery to advanced closed-loop insulin systems significantly improves glycemic parameters, quality of life, and psychological well-being in individuals with T1D [13–16]. The lower transition rate to AHCL among individuals on MDI may reflect that they are generally more resistant to adopting advanced technologies than those already using CSII. Existing barriers to change may have become more pronounced during and after the lockdown, when the post-Covid context further hindered transitions despite the use of telemedicine. The efficacy of telemedicine in diabetes care has been well-documented, particularly for individuals with poor compliance or suboptimal metabolic control [17, 18]. Further evidence has reinforced the clinical efficacy of telemedicine throughout the pandemic [19–22], and many researchers now advocate for the expanded use of telehealth, either as a complement to or a replacement for in-person visits. In a smaller cohort of individuals with T1D, it was found that only those who attended telemedicine visits saw a significant improvement in glycemic metrics during the lockdown, while those who did not attend telemedicine visits did not experience any change in their glycemic control [5]. Importantly, our patients received training on these technologies via periodic remote sessions conducted by specialists, which facilitated the implementation and effective use of advanced insulin delivery systems [23]. This highlights the vital role that telemedicine played during the pandemic, enabling continued management of diabetes and improvements in glycemic control during times when in-person consultations were limited. In our study, an impressive 85% of patients utilized telemedicine services during the year following the lockdown. One of the strengths of this study is its longitudinal design, spanning two years, which provides a rare opportunity to assess the long-term effects of the COVID-19 pandemic on glycemic control in individuals with T1D. Additionally, the use of direct CGM measurements allowed for reliable, robust data collection. However, a limitation of the study is the reliance on self-reported questionnaires to assess lifestyle factors, which may introduce biases and inaccuracies in reporting. Conclusion Our study demonstrates a sustained improvement in CGM metrics over the two years following the COVID-19 lockdown, mostly attributable to the increased use of advanced diabetes-related technologies. The integration of telemedicine and advanced automated insulin delivery systems has facilitated improved patient-provider communication and sustained glycemic control, despite the challenges imposed by the pandemic. The findings suggest that a hybrid care model, combining both traditional in-person and telemedicine services, offers significant benefits in managing T1D, optimizing glycemic control, and improving quality of life. Moving forward, expanding the use of these technologies beyond the pandemic may provide continued improvements in care for individuals with T1D. Supplementary Information Supplementary Material 1 Supplementary Material 1