Introduction

Diabetes mellitus is a major and growing global health challenge, with prevalence projected to reach 783 million by 2045 and annual healthcare expenditures exceeding US$1 trillion, underscoring the urgency of effective, scalable management strategies1,2. Recognising the need for scalable interventions, mobile health (mHealth), defined as “medical and public health practice supported by mobile devices such as mobile phones, patient monitoring devices and wireless technologies”, emerges as a promising subset of digital health3.

mHealth interventions in diabetes management leverage ubiquitous smartphones to support self‑management encompassing blood glucose tracking, insulin dose support, education, and remote monitoring and have demonstrated improvements in glycaemic control, quality of life, self‑efficacy, and treatment satisfaction across diverse populations and delivery models4,5,6,7.

Parameters such as quality of life, incremental cost-effectiveness ratios (ICERs), and healthcare costs are essential in evaluating these interventions. Studies indicate mHealth apps improve quality of life and treatment satisfaction, reduce HbA1c levels, and better manage hyperglycemic episodes8,9. Average annual cost savings per patient using mHealth interventions can be substantial, with reductions of $93.63 per patient annually10,11.

However, findings across studies vary reported no statistically significant reduction in HbA1c levels after three months of using the BlueStar mobile app. Nonetheless, exploratory analysis indicated a dose-response effect each additional day of app usage was associated with a 0.016-point decrease in HbA1c, suggesting potential long-term clinical benefit12. A cluster-randomized controlled trial showing significant improvements in HbA1c and fasting blood glucose levels with daily self-monitoring of blood glucose (SMBG) uploaded via a mobile app13,14.

Goyal et al. evaluated a mobile app for self-management in adolescents with type 1 diabetes, finding no significant change in primary outcomes but a significant improvement in HbA1c for those with higher SMBG frequency14. Huo et al. assessed telemedicine’s effectiveness in participants with cardiovascular disease and diabetes, reporting significant reductions in HbA1c and fasting blood glucose levels15,16.

This systematic review and meta‑analysis therefore consolidates contemporary evidence on both clinical effectiveness (e.g., HbA1c, glycaemic parameters, patient‑reported outcomes) and cost‑effectiveness (e.g., ICERs, healthcare costs, cost savings), incorporating measures such as DSQOL, DQOL, and DALYs. The goal is to provide decision‑makers with robust, integrated estimates of the value and sustainability of mHealth strategies in diabetes care17,18,19.

Studies demonstrated that a nurse coaching program using mHealth technology significantly improved diabetes self-efficacy and physical activity20. The cluster-randomized trial in Bangladesh showed that community-based groups and mobile messaging were cost-effective in preventing type 2 diabetes and intermediate hyperglycemia21.

Islam et al. (2020) conducted a randomized controlled trial in Bangladesh to assess the cost-effectiveness of a mobile phone text messaging intervention for type 2 diabetes. Glucose monitoring app for type 2 diabetes in Japan, highlighting the app’s potential to improve long-term health outcomes while maintaining cost-effectiveness22,23,24.

Comprehensive cost-effectiveness analyses and robust clinical evidence are essential for scaling mHealth interventions in diabetes care. Evaluations should incorporate measures such as DSQOL, DQOL, DALYs, and economic metrics like ICERs and healthcare costs to establish both clinical and economic viability. This evidence can guide policymakers in resource allocation, ensuring sustainability.

This systematic review and meta-analysis synthesizes recent trials and economic evaluations to assess the impact of mHealth apps on clinical outcomes (e.g., HbA1c, quality of life) and cost-effectiveness indicators (e.g., ICERs, healthcare costs, savings), providing decision-makers with updated evidence on the value of mHealth strategies in diabetes management.

Methodology

This systematic review and meta-analysis aimed to evaluate the clinical and cost-effectiveness of mobile health (mHealth) interventions in diabetes care. The review protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) under the registration number CRD420250656558. The review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The review included both quantitative (clinical effectiveness) and economic evaluations (cost-effectiveness) to provide a comprehensive assessment of mHealth interventions in diabetes care.

Search strategy

A comprehensive and systematic search of electronic databases, including PubMed, Google Scholar, and Cochrane Library, was conducted to identify relevant studies while minimizing bias. The search strategy combined keywords and Medical Subject Headings (MeSH) terms related to “Mobile Health (mHealth) Intervention”, “diabetes”, “clinical effectiveness” and “cost-effectiveness”. Boolean operators (AND, OR) were used to broaden or refine the search 2024 as given in Table 1. For the purpose of this review, mobile health (mHealth) interventions were defined as healthcare interventions primarily delivered through mobile or wireless digital technologies that enable patient self-management, monitoring, education, or bi-directional communication. Eligible mHealth technologies included smartphone applications, SMS or app-based messaging systems, mobile-connected glucose monitoring devices, and app-supported telemonitoring platforms. Interventions were required to involve an active digital component enabling patient interaction, data input, feedback, or remote monitoring. Studies in which the intervention was delivered exclusively through conventional telephone calls without a digital or automated mHealth platform were not considered eligible.

The search period was limited to studies published between 2014 and 2024 to reflect the rapid evolution and widespread adoption of contemporary smartphone-based mHealth technologies. Earlier studies often relied on basic mobile phone interventions with limited functionality, reducing relevance to current digital health ecosystems.

Table 1 Search strategy used for the literature search.
Full size table

To minimize publication bias, supplementary searches of grey literature were performed using Google Scholar, including theses and conference proceedings where full study details were available. Reference lists of included studies and relevant reviews were also hand-searched to identify additional eligible studies.

Study selection was conducted independently by two reviewers. Titles and abstracts were screened first, followed by full-text assessment based on predefined inclusion and exclusion criteria. Reviewers were not blinded to authors or journals. Any disagreements were resolved through discussion, and consensus was reached with the involvement of a third reviewer when necessary.

Inclusion and exclusion criteria

Inclusion criteria focused on studies involving participants diagnosed with Type 1 Diabetes Mellitus (T1DM), Type 2 Diabetes Mellitus (T2DM) or those at risk of developing T1DM and T2DM. Only studies assessing mHealth interventions those utilising mobile devices, internet platforms, or computers for diabetes management were considered. Additionally, included studies were required to be randomized controlled trials (RCTs), partial or full economic evaluations, and other relevant comparative studies that reported on key health outcomes related to diabetes management, such as glycated hemoglobin (HbA1c), systolic blood pressure (SBP), diastolic blood pressure (DBP), self-efficacy, quality of life, and patient satisfaction. Furthermore, only peer-reviewed articles published in English from 2014 to 2024 were included.

Exclusion criteria eliminated studies that did not meet specific standards. This included any study for which the complete text cannot be obtained or that failed to adhere to basic scientific rigor. Studies focusing on populations outside the target demographic, such as gestational diabetes mellitus, cancer patients, or other non-target groups, were excluded. Studies in which the intervention was delivered solely through conventional unstructured telephone calls without the use of mobile applications, automated messaging systems, or digital data capture were excluded, as these do not meet the operational definition of mHealth., studies reported in languages other than English were also excluded. Finally, studies that did not report relevant clinical and economic outcomes, as specified in the inclusion criteria, were excluded from this review. By adhering to these defined criteria, this systematic review aimed to synthesize high-quality evidence regarding the cost-effectiveness of mHealth interventions for diabetes management, ultimately contributing valuable insights into their economic viability and clinical efficacy.

Study selection and data extraction

All the screening and selection process of the papers from the literature were done by all the authors. For the study selection, firstly, the studies were screened based on the relevance using the titles and the abstract. Next, the studies were screened based on the criteria listed in the inclusion and exclusion criteria. The remaining studies will be used in the results and data extraction is performed.

For the data extraction, we first extracted the data related to the clinical parameters of the patient which include HbA1c, fasting blood glucose (FBG), systolic blood pressure (SBP) and diastolic blood pressure (DBP) to compare those clinical outcomes between mHealth intervention with the common treatment interventions. Next, we analyzed the economic outcome of the mHealth intervention to interpret the Cost-Effectiveness Ratio. Partial economic evaluations (e.g., cost analyses without formal ICER or QALY estimation) were included to capture the real-world cost implications of mHealth interventions for which full economic modeling was not feasible. These studies were analyzed narratively and not pooled quantitatively, in line with methodological recommendations for heterogeneous economic outcomes.

Lastly, analysis is performed to obtain the effect of mHealth on the quality of life, health behaviors and the satisfaction of the patient. Two authors extracted the information that meets the inclusion criteria from the studies using the data extraction and statistical analysis tool. If there is any discrepancy in the study included, the third reviewer was assigned to resolve the conflict. Other co-authors reviewed and verified the accuracy of the information extracted.

Outcome measures

Primary clinical outcomes included changes in glycated hemoglobin (HbA1c). Secondary outcomes included fasting plasma glucose, body mass index, and blood pressure.

Economic outcomes included cost savings, healthcare utilization, and incremental cost-effectiveness indicators. Where reported, disability-adjusted life years (DALYs) and quality-adjusted life years (QALYs) were extracted and analyzed narratively due to heterogeneity in measurement approaches.

Risk of bias and quality assessment

The Cochrane Risk of Bias tool was employed to assess the quality of included RCTs. The Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist was used to evaluate the quality of economic evaluations. Studies were not excluded based on subjective judgments of quality. Methodological quality and risk of bias were assessed using the Cochrane Risk of Bias tool for randomized controlled trials and the CHEERS 2022 checklist for economic evaluations. Study quality was considered during interpretation of findings rather than as an exclusion criterion. Publication bias was assessed using funnel plots and Egger’s regression test. Sensitivity analyses were performed to examine the robustness of the findings.

Statistical analysis

A meta-analysis was conducted using Review Manager software to synthesize clinical and economic outcomes. For continuous outcomes such as HbA1c and blood glucose levels, weighted mean differences (WMD) with 95% confidence intervals (CIs) were calculated. Heterogeneity was assessed using the I² statistic, with values > 50% indicating significant heterogeneity. A random-effects model was applied if substantial heterogeneity was detected; otherwise, a fixed-effects model was used.

For cost-effectiveness analysis, incremental cost-effectiveness ratios (ICERs) were calculated to determine the additional cost required per QALY gained. The results were visualized using a forest plot, where individual study estimates were represented as squares, with their size indicating study weight in the meta-analysis. The pooled ICER estimate was displayed as a diamond, with confidence intervals reflecting the overall precision of cost-effectiveness estimates.

Results

A comprehensive literature search was conducted across three major databases: Google Scholar, PubMed, and the Cochrane Library, yielding a total of 7,511 records (Google Scholar: 7,350, PubMed: 140, Cochrane Library: 21). Duplicate records (n = 164) and records marked as ineligible by automation tools i.e., Rayyan (n = 7,210) were removed, along with 33 records excluded for other reasons, resulting in 107 records for screening. During the screening phase, 56 records were excluded based on relevance to the study topic. Of the remaining 51 records sought for retrieval, 12 could not be retrieved. Subsequently, 39 reports were assessed for eligibility. Of these, 29 were excluded for reasons such as the intervention not involving an mHealth application (n = 9) such as telemonitoring or telehealth, a non-target study population (n = 9) e.g., gestational diabetes patient groups, or a lack of relevant clinical and economic outcomes reported (n = 11). Ultimately, 10 studies met the inclusion criteria and were included in the review. This systematic process ensured a rigorous and targeted selection of literature relevant to the study objectives as presented in Fig. 1. In line with the predefined operational definition of mHealth described in the Methods section, telemonitoring and telehealth interventions were included only when they incorporated a mobile or app-based digital platform enabling patient interaction, data transmission, or automated feedback.

Fig. 1
Fig. 1
Full size image

PRISMA flow diagram detailing the identification, and inclusion of studies in the systematic review.

Table 2 Characteristics of included studies on digital health interventions for type 2 diabetes.
Full size table

Table 2 summarizes the general features of all ten included articles about the clinical effectiveness and cost-effectiveness of using mHealth interventions in patients with diabetes. From 2014 to 2024, ten studies were conducted in Australia, Bangladesh, China, the United States of America, Europe, Egypt, India, Spain, Italy, and Germany. Among the ten studies, cost-effective analysis (CEA) was conducted in 5 studies while clinical effectiveness was conducted in 9 studies.

Clinical outcomes

Primary outcomes

HbA1c

The clinical effectiveness of mobile health (mHealth) interventions in improving HbA1c levels among individuals with diabetes, as reported across multiple studies. Among the nine studies reporting HbA1c outcomes, seven (78%) demonstrated statistically significant reductions in HbA1c favoring mHealth interventions, with effect sizes ranging from − 0.29% to − 1.31%. Two studies reported non-significant changes as represented in Table 3.

Table 3 Assessment of HbA1c reductions in intervention compared to control groups across selected studies.
Full size table

The meta-analysis showed a time-dependent effect of mHealth interventions on HbA1c. No baseline differences were observed between groups (0.03%, 95% CI: −0.15 to 0.21; p = 0.74). At 3 months, mHealth interventions significantly reduced HbA1c (− 0.61%, 95% CI: −0.95 to − 0.26; p = 0.0006) with no heterogeneity (I² = 0%).

By 6 months, the effect attenuated and was no longer significant (− 0.31%, 95% CI: −0.66 to 0.04; p = 0.09; I² = 85%). Pooling all end‑of‑study time points demonstrated a small but significant reduction in HbA1c favoring mHealth interventions (− 0.31%, 95% CI: −0.52 to − 0.10; p = 0.004), although heterogeneity remained high (I² = 86%). Overall, across all follow‑up periods, the pooled HbA1c reduction was − 0.20% (95% CI: −0.40 to − 0.01; p < 0.05) (Fig. 2).

Fig. 2
Fig. 2
Full size image

Comparing the effects of mHealth interventions on HbA1c control.

Fig. 3
Fig. 3
Full size image

Change in HbA1c at the end of studies that compared mHealth intervention and usual care.

The forest plot in Fig. 3 illustrates the effectiveness of mHealth interventions in reducing HbA1c levels compared to usual care, based on data from eight studies. The pooled mean difference is −0.31 (95% CI: −0.52 to −0.10, p = 0.004), indicating a statistically significant reduction in HbA1c levels favoring the intervention. This pooled reduction is slightly below the commonly referenced MCID for HbA1c (≈ 0.4%), suggesting that while statistically significant, the clinical relevance may be modest. Most studies demonstrated reductions in HbA1c ranging from − 0.90 to −0.12, with only one study25 showing a slight positive mean difference of 0.15, favoring usual care. However, the heterogeneity across studies is high (I² = 86%), reflecting variability in study populations, designs, or intervention implementations. Despite this variability, the results suggest that mHealth interventions can effectively reduce HbA1c levels, though further investigation is needed to address inconsistencies and optimize implementation (Table 4).

Fasting plasma glucose (FPG)
Table 4 Comparison of fasting plasma glucose (FPG) at the baseline, end of the study and FPG across intervention and control Group.
Full size table

Three studies reported fasting plasma glucose outcomes. Two studies demonstrated greater reductions in FPG in the intervention group compared to control, with mean reductions ranging from 1.0 to 1.2 mmol/L, while one large cluster trial reported no statistically significant difference between groups.

Fasting plasma glucose (FPG)
Table 5 Comparison of total participants with impaired fasting plasma at the baseline and end of the study across intervention and control group.
Full size table

At baseline, approximately 5% of participants in both groups had impaired fasting plasma glucose (intervention: 186/4,063; control: 200/4,070). After two years, the proportion increased to 43.9% in the intervention group and 47.3% in the control group. The difference between groups was not statistically significant (unadjusted OR: 0.99; adjusted OR: 1.02; p > 0.9), indicating that the mHealth messaging intervention did not reduce the risk of impaired fasting glucose Table 5.

Secondary outcomes

Body mass index (BMI)
Table 6 Comparison of body mass index (BMI) at the baseline, end of the study and changes in body mass index across intervention and control group.
Full size table

Body mass index (BMI) was reported as a secondary outcome in three studies Table 6. Across studies, BMI changes were small and inconsistent, with mean changes ranging from a reduction of − 0.1 kg/m² to an increase of + 0.4 kg/m² during follow-up. While some control groups showed modest BMI reductions (up to − 0.5 kg/m²), intervention groups generally demonstrated minimal change or slight increases. Overall, mHealth interventions were not associated with clinically meaningful reductions in BMI.

Systolic blood pressure (SBP)
Table 7 Comparison of systolic blood pressure (SBP) in terms of mean SBP, SD and sample size between intervention and control group.
Full size table

Five studies evaluated blood pressure outcomes reported in Table 7. The pooled mean difference for systolic blood pressure was − 1.38 mmHg (95% CI: −3.43 to 0.67; p = 0.19), while in Table 8, diastolic blood pressure showed a modest but statistically significant reduction (− 0.96 mmHg, 95% CI: −1.44 to − 0.47; p = 0.0001) as reported in Fig. 4.

Diastolic blood pressure (DBP)
Table 8 Comparison of diastolic blood pressure (DBP) in terms of mean SBP, SD and sample size between intervention and control group.
Full size table
Fig. 4
Fig. 4
Full size image

Forest plot showing the effect of mHealth interventions on Systolic Blood Pressure (SBP) and Diastolic Blood Pressure (DBP) compared to usual care.

While HbA1c, FPG, and PPG served as the primary clinical outcomes, this meta-analysis also evaluated the impact of mHealth interventions on secondary outcomes, including systolic and diastolic blood pressure, in comparison to usual care, i.e., control group. A total of 5 studies were included to evaluate the effect of mHealth interventions on systolic blood pressure (SBP). The pooled mean difference for SBP was − 1.38 mmHg (95% CI: −3.43 to 0.67; p = 0.19), favoring the intervention group, although this result was not statistically significant. There was moderate heterogeneity among studies (I2 = 28%), indicating some variability in effect sizes.

Similarly, five studies assessed the effect on diastolic blood pressure (DBP). The pooled mean difference was − 0.96 mmHg (95% CI: −1.44 to −0.47; p = 0.0001), showing a statistically significant reduction favoring the intervention group. Heterogeneity was negligible (I2 = 0%), indicating consistent findings among the included studies. There were no significant subgroup differences between SBP and DBP reductions (p = 0.69), suggesting similar effectiveness of mHealth interventions in addressing both measures of blood pressure.

Economic outcomes

Cost-effectiveness analysis (CEA)

Table 9 Cost-effectiveness analysis of mHealth interventions compared to usual care.
Full size table

Cost-effectiveness outcomes were reported in five studies Table 9. Overall, mHealth interventions consistently demonstrated cost savings compared to usual care, primarily through reduced hospitalizations and outpatient visits. Reported incremental cost-effectiveness ratios (ICERs) ranged from cost-saving values (e.g., −$3.21 per unit improvement in HbA1c control rate) to $978 per percentage-point reduction in HbA1c, well below commonly accepted thresholds. Annual per-patient savings varied from $88 to $881, confirming high economic viability across diverse settings. Figure 5 illustrates the pooled economic evaluation quality assessment using CHEERS compliance scores.

Direct cost

Table 10 Comparison of direct costs related to diabetes care between the intervention and control groups.
Full size table

Direct cost analyses

Table 10 consistently showed that mHealth interventions reduced overall healthcare costs compared to usual care, despite initial setup expenses. Reported annual per-patient savings ranged from approximately $449 to $881, primarily driven by fewer hospitalizations, reduced outpatient visits, and lower medication costs. While some categories (e.g., laboratory fees) increased slightly due to proactive monitoring, the net effect remained strongly cost-saving across all studies.

Table 11 Simplified cost savings (in USD) as reported by Fritzen et al. in 5 countries.
Full size table

Pan-European analysis Table 11 showed that mHealth interventions generated greater cost savings than usual care across five countries. Annual savings in intervention groups ranged from $3.49 million in the United Kingdom to $37.42 million in Germany, consistently exceeding control group savings (e.g., Germany: $37.42 M vs. $36.18 M; Italy: $20.74 M vs. $19.84 M). These differences, though modest in some cases, reinforce the economic benefit of integrating telemedicine-based glucose monitoring into routine diabetes care.

DALY costs and QALY

Table 12 Comparison of DALY costs and QALY between intervention and control group.
Full size table

Economic outcomes related to DALYs and QALYs were reported in three studies (Table 12). PLA interventions (not included in the meta-analysis) demonstrated DALY costs ranging from $124 to $2,551 per DALY averted, highlighting affordability in resource-limited settings but not representing mHealth interventions. For mHealth studies, quality-of-life measures showed minimal change: Warren et al. reported no difference in SF-6D scores (0.65 vs. 0.65), and Bonnie et al. observed a small, non-significant QALY increase (13.9 to 14.1 vs. 13.85 to 14.0). Overall, mHealth interventions did not produce clinically meaningful improvements in health-related quality of life.

CHEERS checklist assessment and risk of bias in economic evaluations

A methodological evaluation of the economic studies included in this review was conducted using the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) 2022 checklist. This 28-item checklist assesses the quality, transparency, and reporting integrity of health economic evaluations across core domains such as analytical perspective, time horizon, discounting, costing, and uncertainty analysis.

Overall, the quality of reporting among the 11 included studies varied considerably. CHEERS compliance ranged from 46% to 78%, with an average adherence of 59%, indicating moderate quality and a notable risk of reporting bias. Notably, higher compliance scores were observed in studies conducted in high-income settings, such as Warren et al. and Fritzen et al., both scoring above 70%. These studies demonstrated more comprehensive reporting of cost data, analytical methods, and sensitivity analyses.

Fig. 5
Fig. 5
Full size image

Funnel plot assessing publication bias in economic evaluations of mHealth interventions using CHEERS checklist scores.

Methodological quality of economic evaluations varied (CHEERS compliance: 46%–78%; Table 13). Common gaps included poor reporting of discount rates (50% of studies), limited sensitivity.

analysis (adequate in 55%), and unclear costing perspective in 36%. Nearly half did not specify.

whether real-world or modeled cost data were used, and QALY/DALY valuation methods were inconsistently applied in three studies. Funding and conflict-of-interest declarations were missing in 27%. Egger’s test (p = 0.812) indicated no significant publication bias (Fig. 5). Studies with higher CHEERS scores provided more rigorous evaluations and clearer policy implications, while lower-scoring studies lacked granularity but remain relevant for LMIC contexts where cost-effectiveness thresholds differ.

Table 13 Reporting quality and risk of bias in economic evaluations of mhealth interventions according to CHEERS 2022.
Full size table

Discussion

This systematic review and meta-analysis indicates that mHealth interventions for type 2 diabetes are associated with modest but meaningful short-term improvements in glycaemic control and consistent economic benefits, with attenuated clinical effects over longer follow-up. Taken together, the findings suggest that mHealth is well-suited to catalyse early behaviour change and care optimisation, but maintaining durable clinical gains requires attention to intervention intensity, patient engagement, and integration with broader care model4,5. Early reductions in HbA1c were observed across diverse settings and delivery models26,27,28. with a pooled short-term effect that supports the clinical usefulness of mHealth in the initial months of therapy. The attenuation by 6 months likely reflects waning engagement, habit formation challenges, and variability in intervention components and adherence29.

Early reductions in HbA1c were observed across diverse settings and delivery models26,27,28, with a pooled short-term effect that supports the clinical usefulness of mHealth in the initial months of therapy. Attenuation by 6 months likely reflects waning engagement, habit formation challenges, and variability in intervention components and adherence12,30,31. This pattern underscores a practical implication: mHealth tools may be most effective when paired with structured reinforcement (e.g., coaching, feedback loops, goal resetting) timed at points when drop-off typically occurs.

Multiple mechanisms plausibly underpin early HbA1c improvements: timely prompts and feedback that enhance self-monitoring and medication adherence; remote data flows that enable more responsive clinical adjustments; and patient-facing education delivered in bite-sized, context-aware formats5,10. Heterogeneity in longer-term effects likely mirrors differences in (i) technology mix (CGMs and connected meters versus SMS-only messaging), (ii) baseline risk and population characteristics, and (iii) implementation fidelity and health-system context6,7,13. Dose–response findings further support the role of sustained engagement: greater app use and data uploads are associated with better outcomes12,30.

Effects on fasting plasma glucose, BMI, and blood pressure were small and inconsistent, which is unsurprising given that many interventions targeted self-monitoring and treatment adherence rather than structured diet, physical activity, or antihypertensive optimisation. In practice, weight and blood pressure typically improve when mHealth is integrated with behavioural and lifestyle programmes, clinician escalation protocols, and social support32,33,34. These results therefore argue for multifactorial strategies combining mHealth with nutrition counselling, exercise regimens, and medication titration pathways rather than standalone apps.

Economic evaluations consistently favoured mHealth, with cost savings driven by reduced hospitalisations, fewer outpatient visits, and more efficient medication use26,27,35,36. Reported ICERs ranged from cost-saving to highly favourable; pan-European modelling suggested large aggregate savings via telemedicine-enabled glucose monitoring. Direct cost analyses showed net savings even when setup costs were included, highlighting near-term budget impact alongside potential longer-term value through complication avoidance. These signals are coherent across high- and middle-income contexts, provided that programmes achieve adequate uptake and are embedded within care pathways.

Quality-of-life findings were generally neutral over short follow-up26,37, suggesting that improvements in daily functioning and well-being may require longer duration or broader behavioural components. PLA interventions in Bangladesh were highly cost-effective in DALY terms but were not mHealth and were not pooled in the meta-analysis; nonetheless, they illustrate how community strategies can be affordable and scalable in resource-limited settings21. Elsewhere, QALY-based assessments reported mixed results38, indicating that health-utility gains may lag glycaemic improvements and depend on intervention scope and duration.

CHEERS compliance was moderate, with gaps in discounting, sensitivity analysis, and costing perspective, which complicates cross-study comparability and formal pooling of economic outcomes. Importantly, Egger’s test did not indicate publication bias, but variability in reporting standards and intervention components remains a key driver of heterogeneity27,36. Future evaluations would benefit from standardised economic frameworks, prespecified perspectives (health-system versus societal), and transparent treatment of uncertainty.

For clinicians and programmes, the findings support deploying mHealth to accelerate early glycaemic improvement and reduce utilisation, with planned reinforcement to sustain benefit. For payers and policymakers, the consistent cost-saving signals justify reimbursement pilots and scaled implementation, especially where hospitalisation reduction is achievable. In LMIC settings, tailoring to local infrastructure and cost thresholds is essential; hybrid models that combine low-cost messaging with targeted face-to-face support may maximise value11,21.

Limitations

The heterogeneity in study designs, populations, and intervention modalities complicates the generalizability of findings. Moreover, most studies focused on short-term outcomes; thus, evidence on long-term sustainability is limited. Future research should prioritize standardised methodologies and explore combinations of mHealth tools with telemedicine to enhance efficacy.

Conclusion

This systematic review and meta-analysis highlight the clinical and economic effectiveness of mobile health (mHealth) interventions in managing type 2 diabetes mellitus (T2DM). The findings demonstrate significant short-term improvements in glycemic control and substantial cost savings through reduced healthcare utilization. However, challenges persist in sustaining long-term benefits and addressing broader health outcomes such as weight management and blood pressure control. Despite these limitations, mHealth interventions offer a promising approach to enhancing diabetes care, particularly in resource-constrained settings.

Future directions

Future research should prioritize several key areas to maximize the potential of mHealth interventions for T2DM management. Firstly, standardized methodologies and longer-term studies are essential to assess the durability of clinical and economic benefits. Secondly, integrating mHealth tools with behavioral support systems, such as structured exercise regimens and nutritional counseling, may enhance comprehensive health outcomes. Thirdly, addressing barriers to scalability, including data privacy concerns, adherence challenges, and ambiguous reimbursement models, is crucial for widespread adoption. Lastly, exploring the cost-effectiveness of mHealth interventions in diverse settings and populations will help tailor solutions to meet specific healthcare needs and resource constraints. By addressing these challenges and limitations, mHealth interventions can be optimized to provide sustainable and equitable diabetes care globally.