https://orcid.org/0000-0003-4875-4912
Figures
Abstract
Aims
We aimed to assess a) whether using the digital intervention, MyWay Diabetes (MWD), during the COVID-19 pandemic in Greater Manchester was associated with improvements in glycaemia and cardiovascular risk factors; and b) the cost-effectiveness of MWD.
Methods
From 01st July 2019–30th April 2022, we assessed risk factor changes in MWD users (n = 507) with T2D, adjusting for changes in a matched control group of non-users (n = 10,135). Economic evaluation was conducted using United Kingdom Prospective Diabetes Study Outcomes Model version 2.1 simulating 1,000 patients over 40-years based on observed risk factor changes.
Results
Over 2-years follow-up, MWD use was associated with a clinically significant reductions in HbA1c of 3.1 (95%CI: 1.5 to 4.6) mmol/mol and decreases in systolic blood pressure (1.4 (0.2 to 2.6) mmHg) and cholesterol (0.1 (0.02 to 0.2) mmol/L) when compared to non-users. Including medication costs, MWD generated more health (0·015 QALYs) and lower costs (−£67 per person) compared to usual care. Excluding medication costs, MWD generated QALYs at a cost of £389 each, below the NICE threshold of £20,000–30,000/QALY.
Citation: Ravindrarajah R, Gittins M, Paterson L, Rogers G, Tilson G, Wake DJ, et al. (2026) Clinical and cost-effectiveness of the digital intervention, MyWay Diabetes, in people with type 2 diabetes living in Greater Manchester during the COVID-19 pandemic. PLoS One 21(6): e0349232. https://doi.org/10.1371/journal.pone.0349232
Editor: Ennio Polilli, Pescara General Hospital, ITALY
Received: August 13, 2025; Accepted: April 27, 2026; Published: June 25, 2026
Copyright: © 2026 Ravindrarajah et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data analysed in this study are subject to the following licenses/restrictions: Electronic health records are, by definition, considered “sensitive” data in the UK by the Data Protection Act and cannot be shared via public deposition because of information governance restriction in place to protect patient confidentiality. Access to data is available only once approval has been obtained through the individual constituent entity controlling access to the data. The Greater Manchester Care Record (GMCR) is a database that is populated with data from primary care (general practitioners), secondary care (acute and community hospitals), mental health trusts and 7 out of 10 adult social care areas across an entire geographical region. A total of nine secondary care organisations (including 12 hospitals) and 2 mental health Trusts contribute data to the GMCR. The primary purpose of the GMCR is for direct patient care as it provides clinicians with information relevant to their patient encounters that would ordinarily be inaccessible. There are ethical and legal restrictions on sharing our study data because they contain potentially identifying and sensitive participant information. Sharing these data would breach the conditions of our ethical approval and could expose us to liability under the UK Data Protection Act 2018. Requests for data access, to replicate or extend our presented work, may be made to Health Innovation Manchester, which co-ordinates the GM Care Record on behalf of NHS Greater Manchester and partner organisations. The primary email contact address for general enquiries and research information is: info@healthinnovationmanchester.com.
Funding: Innovate UK and NHS England (ref: 104658). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Doogie Brodie is development manager at MyWay Digital Health. Deborah J Wake is a co-founder and shareholder of MyWay Digital Health (MWDH). Martin K Rutter reports consulting fees from Eli Lilly and Brigham and Women’s Hospital and modest stock ownership in GSK, all unrelated to this work. Michael Morgan-Curran is CEO at Asclepius MedTech Limited, unrelated to this work. Scott G Cunningham is co-founder, director and CTO at MyWay Digital Health and reports grants from Scottish Government, outside the submitted work; personal fees, non-financial support and other from MyWay Digital Health. Naresh Kanumilli reports fees for lectures and advisory boards for BI, Lily, AZ, Novartis, Sanofi Manarini, Novo, Abbott, unrelated to this work. George Tilston is funded by the National Institute for Health and Care Research (NIHR) Manchester Biomedical Research Centre (BRC) (NIHR203308) All other authors report no conflict of interests. EK is part-funded by the NIHR HeatlhTech Research Centre in Emergency and Acute Care (NIHR205301) and the Manchester British Heart Foundation (BHF) Centre for Research Excellence (RE/24/130017). All authors have completed the conflict-of-interest disclosure forms. In accordance with Innovate UK’s funding agreement, the University of Manchester team conducted data analysis and manuscript writing independently of MyWay Digital Health. Both teams collaborated on study design and the Innovate UK funding application, and the MyWay Digital Health team provided feedback on the final manuscript draft; however, they did not influence the study’s results or conclusions. There are no patents, products in development or marketed products associated with this research to declare. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
Introduction
Diabetes is a serious long-term condition causing premature death and complications such as cardiovascular disease (CVD). Diabetes affects more than 371 million people worldwide, of whom 90% have type 2 diabetes (T2D). Effectively managing T2D requires individuals to make daily lifestyle choices and self-management decisions, guided by the knowledge, support, and advice provided by healthcare services.
Current strategies for T2D management are often inadequate, with data from the UK’s National Audit Office survey showing only 8.2% of patients in 2015 with T2D attended a self-management programme. [1] This highlights the urgent need for innovative interventions. Practical, clinical, and financial considerations make digital solutions an appealing option for supporting T2D self-management. Digital solutions offer clear benefits over usual care when face-to-face care is limited, such as during the COVID-19 pandemic, and in people with limited mobility.
Our study aimed to assess whether using the MyWay Diabetes (MWD) digital intervention during the pandemic was associated with improvements in glycaemic control and CVD risk factors in people with T2D, compared to a matched non-user group with T2D. Additionally, we evaluated the cost-effectiveness of MWD based on risk factor changes associated with its use.
Methods
The development of the MWD intervention, along with the aims and methodology of this study have been published.[2] Briefly, we aimed to assess the clinical and cost-effectiveness of MWD in people with T2D in Greater Manchester (GM) during the COVID-19 pandemic. While we initially planned to evaluate three other digital interventions offered to MWD users (Changing Health, Oviva, and MyCognition) [2] uptake was very low, so findings are presented for MWD only.
MWD supports diabetes self-management by providing multimedia resources including traditional information, interactive educational tools and videos on complications.[2] Unlike other digital interventions, the MWD platform directly integrates with NHS primary-care data, enabling seamless access to patient records and more personalised data-driven support. It also allows users to access their primary care clinical records, offering personalised information and explanations about their diagnosis, clinical test results and medication. Content related to key diabetes indicators, such as HbA1c, blood pressure and body weight, is recorded to empower users and improve self-management. The four interventions were initially offered to consenting participants between 01st July 2019 and 31st March 2021, [2] but only MWD was offered from 31st March 2021–30th April 2021. Since then, MWD has been provided as a clinical service across GM.
Design, setting and study duration
To evaluate effectiveness, we compared outcome changes in MWD users with a matched control group, adjusting for risk factors assessed at the time of intervention. The MWD start date was used as the ‘intervention date’ for matched controls. Primary care data from 1st April 2016 enabled risk factor assessment prior to the study starting in July 2019. Post-intervention data was used up to 31st March 2022.
MyWay Digital Health (MWDH) securely supplied NHS numbers of MWD users to GMCR, allowing their identification. The GM Care Record (GMCR) [3] provided University of Manchester researchers access to pseudonymised data on users and non-users within a secure data environment (SDE).[4] Data was accessed for analysis through the SDE from 01st January 2022 till 30th June 2024.
Inclusion criteria for MWD participants were provision of informed consent; age ≥ 18 years; T2D determined from primary care diagnostic codes (Supplementary Material e1.1 in S2 File); registration with a GP in GM consenting to data linkage with MWD, self-certified understanding of English or access to support from family or friends; and access to a digital device [mobile phone and email].
Inclusion criteria for controls for non-users of MWD; age ≥ 18 years; T2D determined from primary care records, registration with a GP in GM; and have not opted out for use of Primary Care data in research.
Sample size
We estimated that with a ratio of 1:10 for participants to controls and assuming a standard deviation[SD] of 15 mmol/mol for HbA1c, [5] 86 participants and 430 controls would provide 80% power to detect a significant (p < 0.05) 5 mmol/mol between-group difference in HbA1c.
Matching for birth year (within 1 year) and sex was performed, with each MWD user matched to up to 20 controls not using MWD to increase statistical power. Age and sex were defined at the MWD start date for each user, which served as the index date for the corresponding matched control non-users.
The primary endpoint was the mean post-intervention HbA1c difference between MWD users and controls, adjusted for pre-intervention levels.[6] Secondary outcomes included differences in body weight, blood pressure, lipid levels and urinary albumin creatinine ratio [UACR]. We used multiple imputations with chained equations to handle missing data, generating 10 complete datasets, combined using Rubin’s rules.[7] The imputation model was closely aligned to the analytical models and included outcomes and relevant covariates regardless of missingness, as per best practice; [8] [9] We also assessed population uptake of MWD and medication use (Supplementary Material e1.2 in S2 File).
Covariates
Models were adjusted for: age, sex, ethnicity, deprivation, smoking and time between pre- and post-intervention outcome assessments, since T2D is a progressive condition with worsening outcomes over time. MWD users had shorter time intervals between outcomes measured pre- and post-intervention compared to controls (Supplementary Table e2.01 in S2 File). Models were also adjusted for baseline risk factors, including: HbA1c, systolic blood pressure, diastolic blood pressure, cholesterol (total, HDL and LDL), creatinine (UACR), triglycerides and body weight. Ethnicity was self-reported. Area-level deprivation was measured using the Index of Multiple Deprivation 2019 (IMD), based on GP postcode.
Statistical methods assessing clinical effectiveness
Descriptive statistics compared characteristics of MWD users and controls, with data reported at baseline (up to 2 years pre-intervention) and follow-up (up to two years post-intervention). For outcomes with multiple values, we selected the pre-intervention value closest to the intervention date and selected the post-intervention value with the longest follow-up. We present means (SD) for continuous and frequencies (%) for categorical variables. After multiple imputation, we used multivariable linear regression models to estimate adjusted differences in post-intervention outcomes between MWD users and controls. Additional models included interaction terms to examine the influence of ethnicity and deprivation on study results.
Sensitivity analyses assessed the impact of a) excluding participants who also used Changing Health, Oviva and MyCognition interventions; b) limiting the analysis to individuals with complete pre- and post-intervention outcome data.
Health economic analysis
To project lifetime costs and effects associated with short-term observed risk factor changes, we used the United Kingdom Prospective Diabetes Study (UKPDS) Outcomes Model version 2.1 (UKPDS-OM2.1) to conduct an economic evaluation from the English NHS perspective.[10, 11] Supplementary Material e3 in S2 File provides comprehensive methods, with Tables e3.01–4.05 highlighting key model parameters and patient characteristics. Briefly, UKPDS-OM2.1 is a validated, patient-level simulation model that predicts relevant events (e.g., myocardial infarction) and long-term outcomes (life expectancy, quality-adjusted life-years [QALYs], and costs) for groups with different baseline characteristics and longitudinal risk profiles.[10] We followed NICE’s reference case, discounting costs (pounds sterling [£] for 2021/22) and QALYs at 3·5% per year over a 40-year time horizon.[12]
Our simulated cohort was based on that used for national policymaking.[12] We derived 1-year treatment effects from the significant differences in pre-and-post-intervention risk factors when comparing treatment groups (HbA1c, SBP, and LDL-cholesterol), and applied these effects to simulated people in the intervention group, with no treatment effect applied to usual care. We also modelled best- and worst-case scenarios based, respectively, on the upper and lower 95% confidence bounds of the relevant effects. In the base-case analysis, we assumed 2-year cost and 1-year effect durations, with natural disease progression thereafter predicted using UKPDS progression equations.[13] We explored alternative durations of cost and effect in sensitivity analysis. Supplementary Material e3.11, Figures e4.01–4.05 in S2 File illustrate the underlying risk-factor progression trajectories for each scenario.
MWDH advised that the full annual licence fee for GM (182,075 people with T2D in February 2024) was £313,557. This fee is currently subject to a negotiated discount; however, following NICE’s stipulation that analyses should only reflect commercial arrangements when prices are nationally guaranteed for a technology’s lifetime, we used the full fee in our base case (see Supplementary Material e3.2 in S2 File). We apportioned the fee among users (n = 16,194 in December 2023) to derive a cost-per-user of £38 over a 2-year cost period. This cost was only applied to people in the model who remained alive – equivalent to assuming that uptake of new user uptake over time offsets attrition among existing users.
Since UKPDS-OM2.1 does not incorporate medication costs, we undertook an additional analysis (described in Supplementary Material e1.3 in S2 File), using baseline and post-intervention participant data from GMCR, mapped to NHS Prescription Service Drug Tariffs (June 2023).[14] Briefly, we estimated post-intervention mean total medication costs for each group using a generalised linear model (log-link and gamma family) adjusting for age, sex, ethnicity, deprivation, smoking status, and medication costs in the 2 years before intervention.
Sensitivity analyses explored a range of assumptions including duration and magnitude of effects and costs, uptake levels, intervention costs, and parameter inputs. We also examined different starting populations: people newly diagnosed with T2D (NICE’s ‘initial therapy’ cohort) and the GM population from our study.
Ethical approval was granted by the Northwest GM South Research Ethics Committee (reference 261268).
Results
A total of 507 people with T2D registered to use MWD, matched for age and sex (1:20) with 10,135 controls not using MWD. Among these MWD registrants, 15 (3%) also used MyCognition, 7 (1.4%) also used Changing Health, 0 (0%) also used Oviva and 485 (95.6%) accessed the MWD intervention.
Baseline characteristics of MWD users (cases) and controls are presented in Table 1. Mean age and sex distributions were similar in both groups (54 years; 47% female). 69% of MWD users were White, 14% were Asian, and 7% were of Black ethnicity. In controls, there was a higher proportion of Asian ethnicity (24%) but a lower proportion of Black ethnicity (5%) individuals compared to cases. Approximately half of the participants in both groups were from the most deprived IMD. Diabetes duration was ~ 9 years in both groups.
Baseline HbA1c, blood pressure, cholesterol levels were similar between MWD users and controls. However, MWD users had a higher mean body weight (~2 kg) than controls (92.5 kg vs. 90.2 kg) and a lower smoking prevalence (10.5% vs. 18.4%).
The primary analyses included 10,642 individuals (507 MWD users matched with 10,135 controls), following multiple imputation. The extent of missing data is presented in Supplementary Table e2.02 in S2 File. Across outcomes, the proportion of missing data was consistently higher at the post-intervention timepoint than at baseline for both DMW users and controls. For the primary outcome (HbA1c), missingness increased from 3.9% to 21.9% among DMW users and from 13.1% to 31.2% among controls. Similar patterns were observed for secondary outcomes: systolic and diastolic blood pressure missingness rose from 8.7% to 21.5% in users and from 16.1% to 33.4% in controls; total cholesterol missingness increased from 8.1% to 28.2% in users and from 17.5% to 36.8% in controls; and body weight missingness rose from 10.8% to 30.0% in users and from 20.8% to 41.5% in controls. Exploratory outcomes also showed substantial missingness, with post-intervention missing data ranging from 22.1% to 65.3% among users and 31.7% to 74.9% among controls. Smoking status had the lowest missingness overall (1.4% in users and 0.5% in controls). After up to 2 years of follow-up, the multivariable-adjusted post-intervention mean HbA1c level was significantly lower in MWD users than in controls (−3.1 mmol/mol, 95%CI: −4.7 to −1.6) (Table 2; Table e2.01). The corresponding adjusted systolic blood pressure difference was −1.7 mmHg (95%CI: −3.0 to −0.4); total cholesterol was −0.1 mmol/L (95%CI: −0.2 to −0.001); and LDL-cholesterol was −0.1 mmol/L (95%CI: −0.2 to −0.001). No significant difference was observed for the remaining secondary outcome, body weight (Table 2; Table e2.01).
When an ethnicity interaction term was added to the model, compared to White MWD users, there were no statistically significant HbA1c changes from baseline for Asian MWD users (−3.3 mmol/mol, 95% CI (−7.1, 0.4)) or other (Black/Mixed/other) ethnic groups (−4.5 mmol/mol, 95%CI (−9.3, 0.24)). Similarly, deprivation (measured by IMD) showed no significant interaction with study outcomes.
Sensitivity analyses
Excluding participants who also used Changing Health and MyCognition interventions had no effect on main study outcomes (Supplementary Table e2.03 in S2 File). Restricting the analysis to matched individuals with complete pre- and post-intervention outcome data provided similar results (Figs 1–4; Supplementary Table e2.05 in S2 File).
Health economics analysis
The UKPDS-OM2.1 model suggests that short-term benefits observed with MWD, compared with usual care, would translate into small reductions in all simulated long-term diabetes-related complications: e.g., 17 fewer myocardial infarctions per 100,000 people over 10 years, equivalent to a number needed to treat of 520 to prevent one event (Supplementary Material e4 in S2 File).
There was a 17% lower change in medication usage in MWD users compared with controls, driven by a post-intervention increase in prescribing in controls that we did not observe in MWD users. This resulted in a 32% lower change in cost of medicines when compared with controls (Supplementary Material e2.1, Tables e2.08–e2.09 in S2 File).
In the base-case cost–utility analysis (Table 3), MWD generates more health (0·015 QALYs) and lower costs (−£67 per person) than usual care, meaning it is ‘dominant’ – both better and, in the long run, cheaper compared with usual care. Excluding medication costs, MWD is no longer net-cost-saving, but generates QALYs at a cost of £389 each, far below the NICE threshold of £20,000–30,000/QALY. 4.1 Table e4.07.
Fig 5 depicts the impact of intervention cost and uptake on cost-effectiveness. We can see that an intervention with MWD’s observed effects is likely to represent good value for money as long as uptake exceeds 5% and cost-per-patient-per-year is less than £10. Under most circumstances, it will also be net-cost-saving, unless we adopt pessimistic effect estimates and exclude estimated medication costs.
Across a series of sensitivity analyses, all results indicated the intervention was cost-saving or cost-effective by NICE standards (Supplementary Material e4.2, Tables e4.04-4.06 and section 4.5 in S2 File) MWD remains dominant if we assume costs and effects are longer-lasting or permanent but, in these scenarios, has marginally worse cost effectiveness if we exclude medication costs; this is because the diminishing returns of the intervention are very slightly outweighed by its diminishing costs.
Supplementary Material e4.3–e4.5 4.3 and 4.4 in S2 File show scenario analyses with different starting populations. In both cases, health gains are slightly lower but cost reductions slightly greater, with the net result that cost–utility results are similar to the base case.
Discussion
Main findings
Over a period of up to 2-years, including the COVID-19 pandemic, the use of MWD was associated with a clinically significant 3.1 mmol/mol reduction in HbA1c, alongside reductions in systolic blood pressure and LDL-cholesterol compared to non-users. Our analysis also demonstrated that MWD is net cost-saving from an NHS perspective.
Strengths and limitations
Our study has several strengths: a) the cohort from this pragmatic real-world observational study is likely to be more representative of the target population than one recruited to a more costly and time-consuming RCT, and is less prone to be affected by control group contamination through MWD’s Massive Open Online Courses and other online content; b) the inclusion of a carefully selected, prospectively studied matched control group, along with extensive covariate adjustment, allowed for stronger causal inferences regarding risk factor changes attributed to MWD use; c) sensitivity analyses demonstrated the robustness of our main findings; d) our economic analysis adhered to national decision-making standards based on NICE guidance, [12] exploring effects across different patient populations and estimating time-path trajectories using published equations rather than assuming linear relationships.
Our study limitations are: a) the observational design, rather than an RCT, limits causal inference; b) the follow-up of up to 2 years restricts assessment of long-term MWD effects; c) the COVID-19 pandemic may have influenced health behaviours, and limited GP practice engagement (now mitigated by a streamlined NHS login); d) early adopters may be more motivated and health-conscious than non-users; e) the study was limited to people with T2D in Greater Manchester, requiring evaluations of other geographical areas and those living with type 1 diabetes; f) limited statistical power reduced the robustness of interaction and sub-group analyses; g) data on MWD usage/participant engagement was not collected, limiting our ability to assess the impact of motivation and sustained use on effectiveness; h) the health economics analysis was limited by: 1) the absence of a probabilistic sensitivity analysis (this was not computationally feasible), which limits assessment of parameter uncertainty; 2) having a base-case assumption that as people die new people join – an open cohort would be ideal but cannot be achieved with the UKPDS outcomes model; [15] 3) using the UKPDS outcomes model [10, 11] based on an older cohort that did not use newer drug classes; [16] and, 4) reliance on short-term data without knowing long-term treatment effects and/or costs.
Comparison with prior literature
No other diabetes digital intervention uses primary care data to provide population-level personalised automated online support, making direct comparison with other interventions challenging. The combination of online structured education, NHS data provision, remote monitoring support through a data app and open access content is unique.
However, the HbA1c reductions associated with MWD use are consistent with the clinical impacts of a wide range of T2D digital interventions as highlighted in several systematic reviews and meta-analyses.[17–19] These interventions included strategies such as self-monitoring glucose and CVD risk factors, reminders to initiate self-care activities, automated feedback on self-reported data, real-time support through text messages and/or phone consultation, personalised feedback from clinicians on data uploads and AI-driven systems providing personalised self-care training. Whilst many of these systems required considerable clinical input and resources, the HbA1c improvements reported (3.3–4.3 mmol/mol), [17–20] were comparable to those associated with MWD use: 3.1 (95% CI: 1.5 to 4.6) mmol/mol.
In the meta-analysis by Cui et al., four studies assessed the impact of digital interventions on blood pressure.[17] These interventions included providing blood pressure monitors, mobile apps, prompting monitoring and recording, and sending real-time educational messages. Although the results were not statistically significant (e.g., systolic blood pressure reduction 2.6 (5.6 to −0.4) mmHg; p = 0.08, compared to usual care), the magnitude of the reduction was similar to that associated with MWD use (1.6 (0.3 to 2.8) mmHg).
Three studies from the same meta-analysis evaluated the effect of digital interventions on lipids.[17] While no statistically significant results were found, e.g., total cholesterol lowering of 0.15 (0.6 to −0.3) mmol/L, the magnitude of the reduction was similar to that associated with MWD use 0.10 (0.00 to 0.20) mmol/L.
As observed in our observational study, Cui et al showed no significant impact of digital interventions on body weight across the four studies that evaluated this outcome.[17] Likewise, a recent review of systematic reviews found that mobile applications were not associated with significant improvements in body weight, blood pressure, or lipids.[21]
Other studies of the cost-effectiveness of digital interventions for people with T2D have shown reductions in medication use, but no clear evidence of beneficial impacts on health or social care utilisation or costs.[22]
Our finding that MWD is net cost-saving aligns with the overall conclusion of the MWD evaluation in Scotland, although our study employed more conservative assumptions (including assuming a full fee in base case; see Supplementary Material, e3.2 in S2 File), resulting in a lower estimate of cost-effectiveness.[23] This is also consistent with general trends in the literature showing that digital interventions supporting T2D self-management are good value for money for healthcare systems, [24] primarily due to lower labour costs compared with delivering traditional lifestyle interventions. For example, a cost-effective UK in-person diabetes education and self-management intervention reported slightly higher QALYs gained (0·0392) but with higher relative costs (+£82 per person) largely driven by the intervention costs (+£209 per person).[25]) Digital interventions have the added potential to become more cost-effective/saving as uptake increases, spreading the cost across more users.[26]
Mechanistic insights
Although our study was not designed to investigate mechanisms behind MWD’s impact on risk factors, improvements in diet or physical activity, driven by increased awareness of key lifestyle factors, could have played a role. Notably, MWD users did not increase their diabetes medication prescriptions (while controls did), and there were no intervention-related changes in body weight, suggesting that these factors were not mediating factors.
Our finding that MWD use was associated with a smaller change in prescribing costs requires careful consideration. The difference-in-differences design of our analysis makes residual confounding alone an unlikely explanation. While it’s possible that individuals who signed up for MWD had varying propensities to receive diabetes medications (which we may not have fully accounted for), it is harder to argue that participants self-selected as those who would not alter their medication use over time. In fact, the intervention cohort had higher baseline exposure to diabetes medications than controls, yet the groups converged during the intervention period. This suggests that the intervention may have helped prevent the escalation of prescribing in MWD users, which was observed in the control group. Given the uncertainty surrounding these dynamics, we provide full cost-effectiveness results both with and without estimated medication costs. Reassuringly, MWD remains good value in both scenarios. While the smaller change in medication costs observed among MWD users is encouraging, it may reflect differences in prescribing behaviour or medication adherence rather than a direct effect of the intervention, and this potential influence should be considered when interpreting the findings.
Clinical implications
We demonstrated that MWD can be effectively delivered to people with T2D during the COVID-19 pandemic, showing clinically meaningful improvements in glycaemic control and improvements in blood pressure and cholesterol that would be expected to reduce risks for long-term diabetes-related complications. Since MWD does not directly involve clinical staff input, it could lessen clinical primary care workload. Given the ‘real-world’ design (not an RCT), the findings are likely to be highly generalisable. We also showed that MWD was cost-effective and potentially cost saving when the cost of medication was included, making it appealing to financially constrained health services.
Future research opportunities
Future studies in the post-COVID-19 period should target larger diverse ideally nationwide cohorts, including people with type 1 diabetes, allowing greater statistical power for subgroup analyses by ethnicity [27] and deprivation for example. It would be useful to compare effects of MWD and other online interventions including the NHS Healthy Living programme.[28] Large cohorts would be needed to assess if combining these interventions offers added benefits. Further studies should assess medication adherence [29] and effects in subgroups defined by age and digital literacy. A scoping review concluded that T2D digital interventions can be cost-saving or cost-effective but poor reporting quality limits comparisons between interventions and decision making when commissioning services. Therefore, general improvements to study reporting could be valuable.[30]
Conclusions
Using MWD for up to 2-years, including during the COVID-19 pandemic, was associated with a clinically relevant improvement in HbA1c and improvements in blood pressure and cholesterol when compared to non-users. At a population level, these changes would likely reduce risks for long-term diabetes complications. We found MWD to be net cost-saving from an NHS perspective, which could be attractive to financially constrained health services.
References
- 1.
Digital N. National diabetes audit 2016-2017 full report – care processes and treatment targets England and Wales. 2018.
- 2. Goldthorpe J, Allen T, Brooks J, Kontopantelis E, Holland F, Moss C, et al. Digital Interventions Supporting Self-care in People With Type 2 Diabetes Across Greater Manchester (Greater Manchester Diabetes My Way): Protocol for a Mixed Methods Evaluation. JMIR Res Protoc. 2022;11(8):e26237. pmid:35976184
- 3.
Greater Manchester Care Record. https://gmwearebettertogether.com/ 2024.
- 4.
NHS E. Control of patient information (COPI) notice 2020. https://digital.nhs.uk/coronavirus/coronavirus-covid-19-response-information-governance-hub/control-of-patient-information-copi-notice Accessed 2023 October 1.
- 5. Lean ME, Leslie WS, Barnes AC, Brosnahan N, Thom G, McCombie L, et al. Primary care-led weight management for remission of type 2 diabetes (DiRECT): an open-label, cluster-randomised trial. Lancet. 2018;391(10120):541–51. pmid:29221645
- 6. Wright AK, Suarez-Ortegon MF, Read SH, Kontopantelis E, Buchan I, Emsley R, et al. Risk Factor Control and Cardiovascular Event Risk in People With Type 2 Diabetes in Primary and Secondary Prevention Settings. Circulation. 2020;142(20):1925–36. pmid:33196309
- 7. Rubiin DB. Multiple imputation after 18 years. Journal of the American Statistical Association. 1996;91(434):473–89.
- 8. Kontopantelis E, White IR, Sperrin M, Buchan I. Outcome-sensitive multiple imputation: a simulation study. BMC Med Res Methodol. 2017;17(1):2. pmid:28068910
- 9. Hippisley-Cox J, Coupland C, Brindle P. Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study. BMJ. 2017;357:j2099. pmid:28536104
- 10. Hayes AJ, Leal J, Gray AM, Holman RR, Clarke PM. UKPDS outcomes model 2: a new version of a model to simulate lifetime health outcomes of patients with type 2 diabetes mellitus using data from the 30 year United Kingdom Prospective Diabetes Study: UKPDS 82. Diabetologia. 2013;56(9):1925–33. pmid:23793713
- 11.
University of Oxford. UKPDS Outcomes Model. https://www.dtu.ox.ac.uk/outcomesmodel/ 2021.
- 12.
National Institute for Health and Care Excellence. Type 2 diabetes in adults: management (update). Health economic model report (NICE guideline NG28). 2022 [Available from: https://www.nice.org.uk/guidance/ng28/evidence/health-economic-model-report-pdf-10959500845
- 13. Leal J, Alva M, Gregory V, Hayes A, Mihaylova B, Gray AM, et al. Estimating risk factor progression equations for the UKPDS Outcomes Model 2 (UKPDS 90). Diabet Med. 2021;38(10):e14656. pmid:34297424
- 14.
NHS business services authority. Electronic drug tariff. https://www.drugtariff.nhsbsa.nhs.uk/#/00838270-DC/DC00838243/Home 2023.
- 15. O’Mahony JF, Newall AT, van Rosmalen J. Dealing with Time in Health Economic Evaluation: Methodological Issues and Recommendations for Practice. Pharmacoeconomics. 2015;33(12):1255–68. pmid:26105525
- 16. Si L, Willis MS, Asseburg C, Nilsson A, Tew M, Clarke PM. Evaluating the ability of economic models of diabetes to simulate new cardiovascular outcomes trials: A report on the ninth Mount Hood diabetes challenge. Value Health. 2020;23(9):1163–70.
- 17. Cui M, Wu X, Mao J, Wang X, Nie M. T2DM Self-Management via Smartphone Applications: A Systematic Review and Meta-Analysis. PLoS One. 2016;11(11):e0166718. pmid:27861583
- 18. Romadlon DS, Tu Y-K, Chen Y-C, Hasan F, Chiu H-Y. Comparative efficacy of digitally assisted interventions for glycated haemoglobin levels among patients with type 2 diabetes: A systematic review and component network Meta-Analysis. Diabetes Obes Metab. 2023;25(11):3279–89. pmid:37519284
- 19. Tanhapour M, Peimani M, Rostam Niakan Kalhori S, Nasli Esfahani E, Shakibian H, Mohammadzadeh N, et al. The effect of personalized intelligent digital systems for self-care training on type II diabetes: a systematic review and meta-analysis of clinical trials. Acta Diabetol. 2023;60(12):1599–631. pmid:37542200
- 20. de Souza Ferreira E, de Aguiar Franco F, Dos Santos Lara MM, Levcovitz AA, Dias MA, Moreira TR, et al. The effectiveness of mobile application for monitoring diabetes mellitus and hypertension in the adult and elderly population: systematic review and meta-analysis. BMC Health Serv Res. 2023;23(1):855. pmid:37573312
- 21. Represas-Carrera FJ, Martínez-Ques ÁA, Clavería A. Effectiveness of mobile applications in diabetic patients’ healthy lifestyles: A review of systematic reviews. Prim Care Diabetes. 2021;15(5):751–60. pmid:34275771
- 22. Morris T, Aspinal F, Ledger J, Li K, Gomes M. The Impact of Digital Health Interventions for the Management of Type 2 Diabetes on Health and Social Care Utilisation and Costs: A Systematic Review. Pharmacoecon Open. 2023;7(2):163–73. pmid:36495462
- 23. Cunningham SG, Stoddart A, Wild SH, Conway NJ, Gray AM, Wake DJ. Cost-Utility of an Online Education Platform and Diabetes Personal Health Record: Analysis Over Ten Years. J Diabetes Sci Technol. 2023;17(3):715–26. pmid:34986658
- 24. Willems R, Annemans L, Siopis G, Moschonis G, Vedanthan R, Jung J, et al. Cost effectiveness review of text messaging, smartphone application, and website interventions targeting T2DM or hypertension. NPJ Digit Med. 2023;6(1):150. pmid:37596488
- 25. Gillett M, Dallosso HM, Dixon S, Brennan A, Carey ME, Campbell MJ, et al. Delivering the diabetes education and self management for ongoing and newly diagnosed (DESMOND) programme for people with newly diagnosed type 2 diabetes: cost effectiveness analysis. BMJ. 2010;341:c4093. pmid:20729270
- 26. Paterson L, Rennick-Egglestone S, Gavan SP, Slade M, Ng F, Llewellyn-Beardsley J, et al. Development and delivery cost of digital health technologies for mental health: Application to the Narrative Experiences Online Intervention. Front Psychiatry. 2022;13:1028156. pmid:36419974
- 27. Sabahi A, Jalali S, Ameri F, Garavand A, Negahban A. The effect of using mobile health on self-management of type 2 diabetic patients: A systematic review in Iran. J Educ Health Promot. 2023;12:358. pmid:38143998
- 28. Hawkes RE, Benton JS, Cotterill S, Sanders C, French DP. Service Users’ Experiences of a Nationwide Digital Type 2 Diabetes Self-Management Intervention (Healthy Living): Qualitative Interview Study. JMIR Diabetes. 2024;9:e56276. pmid:39024002
- 29. Chong CJ, Bakry MM, Hatah E, Mohd Tahir NA, Mustafa N. Effects of mobile apps intervention on medication adherence and type 2 diabetes mellitus control: A systematic review and meta-analysis. J Telemed Telecare. 2023. pmid:37254516
- 30. Tornvall I, Kenny D, Wubishet BL, Russell A, Menon A, Comans T. Economic Evaluations of mHealth Interventions for the Management of Type 2 Diabetes: A Scoping Review. J Diabetes Sci Technol. 2025;19(1):179–90. pmid:37395212