Introduction
Type 2 diabetes mellitus (T2DM) is a chronic metabolic disease caused by insufficient insulin secretion or reduced insulin efficiency, accounting for approximately 90-95% of all diabetes cases worldwide [
1]. According to the International Diabetes Federation, over 537 million people currently suffer from diabetes globally, and this number is projected to reach 643 million by 2030 [
2]. In China, diabetes has become a significant public health issue, with recent studies showing an estimated prevalence of approximately 11% among adults, equating to over 140 million individuals affected [
3].
For patients with T2DM, maintaining optimal blood glucose levels, specifically an HbA1c level of ≤ 7%, is a primary goal in clinical care. Studies have demonstrated that effective blood glucose control can substantially reduce the risk of diabetes-related complications (by 53-63%) and mortality (by 46%) [
4]. However, successful blood glucose management largely depends on patients’ adherence to healthy lifestyle practices, such as regular exercise, balanced diet, and proper medication usage [
5]. Many patients, however, find it challenging to sustain these behaviors in daily life, making ongoing supervision and follow-up essential in diabetes management.
With advancements in telemedicine, remote nursing interventions have emerged as a convenient, efficient, and cost-effective approach to diabetes care, and have been widely adopted in many countries [
6]. Remote nursing includes a variety of methods such as phone calls, mobile applications, WeChat, SMS, internet platforms, and smart nursing systems [
7], Among these, telephone intervention has become one of the most commonly used tools for nurses to provide remote care to diabetes patients due to its accessibility and interactive benefits [
8].
Although several studies have explored the effects of telephone interventions by various healthcare professionals (such as dietitians, physicians, multidisciplinary teams, and psychologists) on diabetes management [
9‐
11], there remains a lack of evidence-based data regarding the impact of nurse-led telephone interventions on HbA1c levels in T2DM patients. Additionally, there is no consensus on key elements of telephone follow-up, including optimal frequency, duration per call, content of follow-ups, and interval between follow-ups, which are crucial for effective blood glucose control. Clarifying the influence of these factors on glycemic outcomes would help to inform more targeted follow-up protocols in clinical practice.
This study aims to conduct a meta-analysis to systematically evaluate the effects of nurse-led telephone interventions on HbA1c levels in T2DM patients. Through subgroup analyses, we aim to identify the optimal duration, content, frequency, and interval of telephone follow-ups. This will provide clinical nurses with effective follow-up protocols, enhance the quality of remote nursing care, and contribute evidence-based strategies for blood glucose management in T2DM patients.
Methods
This review strictly follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and is registered on the PROSPERO platform (Registration Number: CRD42024578866).
Inclusion and exclusion criteria
To ensure the comprehensiveness of this meta-analysis, we made every effort to include all eligible studies that met the inclusion criteria. Comprehensive database searches were conducted across PubMed, Web of Science, Cochrane Library, Embase, China Biology Medicine (CBM), China National Knowledge Infrastructure (CNKI), Wanfang, and China Science and Technology Journal Database (VIP). Additionally, we manually screened the reference lists of relevant articles to identify additional studies and contacted corresponding authors when necessary to obtain further data or clarifications. These efforts were undertaken to minimize the risk of omitting relevant studies and to ensure that the analysis provides a complete and representative evaluation of the available evidence.
Inclusion criteria: (1) Participants: Adult patients with a confirmed diagnosis of T2DM. (2) Intervention: The experimental group received nurse-led telephone follow-ups in addition to standard care, with a detailed follow-up protocol; the control group received standard care. (3) Outcome measure: The primary outcome was HbA1c levels before and after the intervention, a reliable indicator of average blood glucose levels over the preceding 2–3 months in diabetic patients [
12]. (4) Study design: Randomized controlled trials (RCTs). Exclusion criteria: (1) Interventions not led by nurses (e.g., interventions by physician teams or psychologists); (2) Inaccessible full texts or studies where data could not be extracted; (3) Conference abstracts, study protocols, or reviews; (4) Non-English and non-Chinese literature: To ensure the comprehensiveness and regional applicability of this study, we included Chinese literature. The inclusion of these studies was based on China’s extensive research background and wide implementation experience in nurse-led follow-up interventions. All Chinese studies were translated into English by the research team and assessed according to standardized quality evaluation criteria to ensure the reliability and transparency of the data. (5) Duplicate publications.
Literature search strategy
A comprehensive search was conducted in PubMed, Web of Science, Cochrane Library, Embase, CNKI, VIP, CBM, Wan fang Database for RCTs investigating the effects of nurse-led telephone interventions on HbA1c control in patients with type 2 diabetes. The search period extended from the inception of each database to February 2024. This timeframe was chosen to ensure that all relevant studies, from the earliest available literature on nurse-led telephone interventions to the most recent findings, were included. Nurse-led telephone follow-ups have gained prominence in diabetes care over the past several decades, and this search period comprehensively captures the evolution of this intervention. We used a combination of subject terms and free terms. Boolean operators were employed to refine searches. The search terms included “telehealth,” “telemedicine,” “tele-referral,” “virtual medicine,” “virtual nursing,” “telecare,” “tele intensive care,” “mobile health,” “digital health,” “remote consultation,” “mobile phone,” “telephone,” “smartphone,” “e-mail,” and terms for diabetes such as “Type 2 Diabetes,” “Diabetes Mellitus, Type II,” “Diabetes Mellitus, Noninsulin-Dependent,” “Type 2 Diabetes,” “NIDDM,” and “T2DM.” The detailed search strategy for each database is provided in Appendix Table
1.
Table 1
Basic characteristics of the studies
| 2018 | China | 49 | 49 | Standard Care + Telephone | Standard Care | 2 | ④ | 3 |
| 2021 | China | 60 | 60 | Standard Care + Telephone | Standard Care | 6 | ② | 12 |
| 2013 | China | 100 | 100 | Standard Care + Telephone | Standard Care | 3 | ① | 3 |
| 2018 | China | 62 | 61 | Standard Care + Telephone | Standard Care | 12 | ③ | 12 |
| 2016 | Belgium | 240 | 246 | Standard Care + Telephone | Standard Care | 18 | ② | 5 |
| 2013 | Australia | 221 | 219 | Standard Care + Telephone | Standard Care | 15 | ③ | 8 |
De Vasconcelos et al. [ 16] | 2018 | Brazil | 16 | 15 | Standard Care + Telephone | Standard Care | 6 | ③ | 12 |
Esmaeilpour-Bandbonil et al. [ 17] | 2021 | Iran | 28 | 32 | Standard Care + Telephone | Standard Care | 3 | ④ | 8 |
| 2003 | Korea | 20 | 16 | Standard Care + Telephone | Standard Care | 3 | ④ | 16 |
| 2005 | Korea | 15 | 10 | Standard Care + Telephone | Standard Care | 3 | ④ | 16 |
| 2010 | Iran | 30 | 30 | Standard Care + Telephone | Standard Care | 3 | ④ | 16 |
Odnoletkova and Goderis et al. [ 21] | 2014 | Belgium | 287 | 287 | Standard Care + Telephone | Standard Care | 18 | ② | 16 |
| 2017 | USA | 20 | 21 | Standard Care + Telephone | Standard Care | 3 | ① | 3 |
Table 2
Subgroup analysis of the effects of telephone intervention on overall HbA1c levels
Follow-Up Frequency | 3–5 | 4 | 409/416 | -0.28[-0.64,0.07] | 14.04 | 0.003 | 79 | 1.57 | 0.12 |
| 8 | 2 | 249/251 | -0.36[-0.99,0.27] | 7.15 | 0.007 | 86 | 1.13 | 0.26 |
| 12 | 3 | 138/136 | -0.87[-1.28,-0.46] | 4.94 | 0.08 | 60 | 4.13 | < 0.001 |
| 16 | 4 | 352/343 | -0.92[-1.71,-0.12] | 27.25 | < 0.001 | 89 | 2.26 | 0.02 |
Interval Duration | 1 week | 4 | 165/156 | -0.93[-1.68,-0.17] | 24.99 | < 0.01 | 88 | 2.40 | 0.02 |
| 2 weeks | 5 | 411/415 | -0.46[-1.00,0.09] | 28.96 | < 0.01 | 86 | 1.65 | 0.10 |
| Half a month | 2 | 111/110 | -1.29[-2.43,-0.15] | 5.27 | 0.02 | 81 | 2.22 | 0.03 |
| ≥ 1month | 2 | 461/465 | -0.09[-0.24,0.06] | 0.06 | 0.80 | 0 | 1.12 | 0.26 |
Follow-Up Duration | 10–15 min | 3 | 190/193 | -0.54[-1.02,-0.06] | 26.00 | < 0.01 | 92 | 2.21 | 0.03 |
| 20–25 min | 3 | 65/56 | -1.23[-1.63,-0.83] | 1.08 | 0.58 | 0 | 5.98 | < 0.001 |
| 30 min | 2 | 307/308 | -0.09[-0.25,0.06] | 0.34 | 0.56 | 0 | 1.17 | 0.24 |
Follow-Up Content | ① | 2 | 120/121 | -0.15[-0.30,-0.01] | 0.50 | 0.48 | 0 | 2.07 | 0.04 |
| ② | 3 | 587/593 | -0.42[-0.87,0.04] | 24.53 | < 0.01 | 92 | 1.80 | 0.07 |
| ③ | 3 | 299/295 | -0.37[-0.99,0.25] | 19.94 | < 0.01 | 90 | 1.18 | 0.24 |
| ④ | 5 | 142/137 | -1.17[-1.60,-0.75] | 8.23 | 0.08 | 51 | 5.42 | < 0.01 |
Literature screening and data extraction
Two researchers (Y.C. and T.Z.) with backgrounds in evidence-based medicine independently screened the literature and extracted data. In cases of disagreement, a third researcher (X.K.) was consulted to resolve conflicts. All identified studies were imported into EndNote, and duplicates were removed. Titles and abstracts were screened for initial selection, followed by a full-text review for final inclusion. Extracted data included first author, publication year, country, sample size, intervention details, duration of intervention (in months), and outcome measures.
Quality assessment of included studies
Two researchers (Y.C. and T.Z.) trained in evidence-based medicine independently assessed the quality of each study using the Cochrane Handbook’s criteria for RCTs [
13]. Quality indicators included random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, data integrity, selective outcome reporting, and other sources of bias. Each criterion was assessed as follows, in accordance with the guidelines provided by the Cochrane Handbook for Systematic Reviews of Interventions: Low risk: The study provided adequate methods or information to meet the criterion, with minimal risk of bias. Unclear risk: The study provided insufficient details to fully assess the criterion, or the reported methods lacked transparency. High risk: The study reported methods that clearly introduced bias or failed to meet the criterion.
Studies were categorized into three grades based on the combined assessment of all seven criteria, following the Cochrane Handbook: Grade A (High Quality): A study was rated as Grade A only if all seven criteria were assessed as “low risk.”Grade B (Moderate Quality): A study was rated as Grade B if it had a majority (≥ 4) of “low risk” ratings but included one or more “unclear risk” ratings and no more than one “high risk” rating. Grade C (Low Quality): A study was rated as Grade C if fewer than four criteria were assessed as “low risk” or if two or more criteria were assessed as “high risk. After quality appraisal, studies rated as Grade A or Grade B were included in the meta-analysis. Grade C studies were excluded due to their high risk of bias and potential to compromise the validity of the results.
Statistical analysis
Meta-analysis was conducted using RevMan 5.4 software. Heterogeneity was first assessed; if P > 0.1 and I² < 50%, a fixed-effect model was applied, indicating acceptable heterogeneity. If P ≤ 0.1 and I² > 50%, a random-effect model was applied, and potential sources of heterogeneity were further explored through subgroup analysis. Sensitivity analysis was used to test the robustness of the results, and publication bias was assessed through funnel plot analysis. A significance level of α = 0.05 was applied, with P < 0.05 considered statistically significant.
Discussion
This meta-analysis demonstrates that nurse-led telephone interventions significantly improve HbA1c levels in patients with T2DM. These interventions are particularly valuable for underserved or rural populations, as they address geographical and temporal barriers to accessing professional healthcare services. Our findings highlight that follow-up protocols with specific characteristics—such as at least 12 sessions, conducted weekly or every half month, each lasting 10–25 min, and focusing on medication, diet, exercise, and glucose monitoring—are effective in improving HbA1c outcomes. Among these, protocols involving 16 follow-ups with half-month intervals and session durations of 20–25 min achieved the most significant HbA1c reductions, indicating the importance of optimizing follow-up frequency and intensity to maximize therapeutic benefits.
Our findings align with prior studies demonstrating the benefits of remote interventions in diabetes management. For instance, de Vasconcelos et al. [
16] found that increasing the number of remote interventions significantly improved blood glucose levels, consistent with our finding that higher follow-up frequencies (≥ 12 sessions) lead to better glycemic control. Specifically, our results showed that 12 follow-ups reduced HbA1c by -0.87%, while 16 follow-ups yielded an even greater reduction of -0.92%. These improvements emphasize the importance of reinforcing self-management behaviors through consistent, structured interventions. In contrast, follow-up frequencies of 3–5 or 8 sessions did not yield significant HbA1c improvement compared to the control group, suggesting that fewer follow-ups may be insufficient to sustain behavioral changes. Consequently, we recommend that nurses conduct at least 12 follow-ups for effective HbA1c management in clinical practice.
While 16 follow-ups demonstrated the most significant improvement in HbA1c levels in this study, diabetes management is a lifelong process that requires continuous support beyond the intensive follow-up phase. After completing 16 follow-ups, transitioning to less intensive schedules, such as monthly or quarterly follow-ups, may help maintain the glycemic improvements achieved during the initial phase while reducing the burden on healthcare systems and patients. Additionally, integrating alternative approaches, such as community-based interventions, mobile health applications, or remote glucose monitoring technologies, may provide sustained support for long-term diabetes management. Future research should explore the efficacy and feasibility of these strategies, as well as their impact on glycemic control and other clinical outcomes.
Similarly, intervention duration significantly impacted outcomes; sessions lasting 20–25 min achieved the greatest reduction in HbA1c levels. In contrast, shorter sessions (10–15 min) and longer sessions (> 30 min) were less effective, likely due to insufficient engagement or patient fatigue. This finding aligns with Li et al. [
27], who reported that intervention duration can impact patient outcomes, such as blood pressure management. Thus, we recommend maintaining telephone session durations between 10 and 25 min to ensure sessions are neither too brief to be effective nor too lengthy to cause disengagement. The results also indicate that follow-up intervals of 1 week or half a month are most effective in lowering HbA1c levels, while intervals longer than one month do not provide significant improvements compared to control groups. Shorter intervals may facilitate consistent monitoring and adjustment of patient behaviors, while excessively long intervals may reduce patient adherence due to limited oversight, and overly frequent contacts may lead to patient fatigue. Therefore, we suggest an optimal follow-up interval of 1 week or half a month to balance engagement and adherence. Moreover, comprehensive diabetes management encompasses diet, medication, exercise, glucose monitoring, and diabetes education [
28], which aligns with the most effective follow-up content identified in our study. This finding suggests that addressing these core management aspects in telephone interventions can facilitate improved HbA1c outcomes.
While earlier meta-analyses included diverse healthcare professionals, our study focuses exclusively on nurse-led, telephone-based interventions, enhancing its relevance to real-world clinical settings where nurses often play a primary role in patient follow-up [
29]. Research indicates that education and support provided by nurses enhance patients’ self-management abilities [
30]. The International Diabetes Federation recommends regular health education for all diabetic patients, which is crucial for effective glycemic control [
28]. Asante et al. [
31] found that patients receiving nurse-led telephone interventions showed better adherence to dietary, exercise, glucose monitoring, and foot care practices compared to those receiving standard care. Furthermore, previous meta-analyses suggested that remote interventions by various healthcare professionals effectively reduced blood glucose and systolic blood pressure in T2DM patients [
10], consistent with our findings.
It is important to note that this study exhibits a certain degree of heterogeneity, which likely stems from multiple factors, including patient baseline characteristics, comorbidities, baseline HbA1c levels, follow-up protocols, and regional variability. For instance, baseline HbA1c levels varied significantly across studies, ranging from 5.24% in Xiaomei’s study to 11.06% in Brown-Deacon et al., potentially contributing to variability in intervention effects. Patients with higher baseline HbA1c levels may experience greater improvements due to regression to the mean, while those with near-normal levels may have less room for improvement. Additionally, differences in comorbidities, such as hypertension or cardiovascular disease, may have influenced patient responses to the interventions, adding to the heterogeneity. Variations in follow-up protocols, including frequency, duration, and content, were also notable. For example, studies that implemented more frequent or comprehensive follow-ups showed greater HbA1c improvements, indicating that standardizing follow-up protocols could reduce variability. Furthermore, regional and cultural differences, such as disparities in healthcare systems and access to resources, likely contributed to the variability in the results.
While subgroup analyses helped to identify some potential sources of heterogeneity, residual variability remains, which highlights the limitations of this meta-analysis. Differences in study quality and sample size may have further contributed to the observed heterogeneity, as smaller studies tend to introduce greater variability. Additionally, the variability in healthcare access and cultural norms across regions may limit the generalizability of the findings. Future research should focus on standardizing follow-up protocols, conducting larger, high-quality randomized controlled trials, and exploring contextual factors such as patient adherence and socioeconomic status to address these sources of heterogeneity. Despite these limitations, the consistent trends in HbA1c improvement observed across the included studies provide robust evidence supporting the effectiveness of nurse-led telephone interventions.
Implications
This study provides significant implications at the clinical, educational, organizational, and research levels. Clinically, the findings offer evidence-based guidance for optimizing nurse-led telephone follow-ups to improve glycemic control in T2DM patients, particularly for those in rural or underserved areas. Educationally, the study highlights the need to incorporate structured diabetes education and follow-up training into nursing curricula and professional development programs, ensuring nurses are equipped with the skills to deliver effective remote interventions. At the organizational level, the study underscores the importance of allocating sufficient resources and establishing standardized protocols for nurse-led follow-ups, balancing nursing workloads with patient outcomes. From a research perspective, the findings emphasize the need for further high-quality randomized controlled trials to refine follow-up protocols and investigate their applicability to diverse patient populations with varying comorbidities and baseline conditions.
Strengths and limitations
This study has several strengths. First, it is the first meta-analysis to focus exclusively on nurse-led, telephone-based interventions, which provides clinically relevant evidence for real-world settings where nurses often play a primary role in patient management. Second, the inclusion of subgroup analyses allowed for a more detailed investigation of key factors influencing intervention effectiveness, such as follow-up frequency, duration, and content, offering specific, actionable recommendations for optimizing follow-up protocols. Third, the study included a broad range of populations from diverse geographic regions, improving the generalizability of findings to various healthcare settings. Lastly, rigorous quality assessment and sensitivity analyses ensured the stability and reliability of the results, minimizing the impact of potential biases.
However, this study also has several limitations. First, the included studies varied in quality, underscoring the need for further high-quality randomized controlled trials to strengthen the robustness of these findings. Second, the study exhibited some heterogeneity, likely stemming from differences in patient baseline characteristics, comorbidities, and initial HbA1c levels across studies. Additionally, variations in follow-up protocols, including duration, content, and frequency, may have contributed to outcome differences. Future research should focus on developing more tailored follow-up protocols that consider patients’ comorbidities, disease duration, and baseline conditions to enhance intervention effectiveness. Third, while the inclusion of Chinese literature improves the comprehensiveness of this meta-analysis, it may introduce potential language bias, as studies in other non-English languages were not included. To address this, future meta-analyses should aim to include studies from other non-English languages to provide a more balanced and inclusive analysis. Lastly, although subgroup analyses and sensitivity tests were conducted to explore heterogeneity, some residual variability remains unexplained. Further research is needed to standardize intervention protocols and explore contextual factors such as patient adherence and regional healthcare infrastructure to reduce this variability.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.