Impact on diabetes-related health outcomes using MyWay Diabetes platform in Somerset, UK: An interrupted time-series analysis

Background and introduction
Diabetes affects 10% of the global population, accounting for 10% of health spending. MyWay Diabetes (MWD) (https://somerset.mydiabetes.com) is a digitally-enabled diabetes self-management tool which aims to improve diabetes management through providing patient-centred personalised access to electronic health records, structured education, and other self-management features. Since MWD was implemented within NHS Somerset in 2018 as an adjunct to existing healthcare services, no observational research on its impact on clinical outcomes has been conducted.

Aims and objectives
We aim to characterise trends in diabetes-related health outcomes (e.g. HbA1c, lipids, blood pressure etc.) for MWD users within NHS Somerset, by comparing trends through time with respect to date of MWD registration.

Methods
An interrupted time-series analysis in MWD users with type 1 diabetes (T1DM) or type 2 diabetes (T2DM) in Somerset, UK, compared pre- and post-MWD registration trends to estimate differences in health outcomes, using quarterly timepoints within 48-month pre-MWD registration and 24-month post-MWD registration periods. Generalised estimating equations modelling was conducted, which included adjusting for participant baseline characteristics. We also conducted sub-group analyses for age, sex and MWD usage.

Results
A total of 750 and 6457 people with T1DM and T2DM were included in the analysis, respectively. The interrupted time-series analysis (see figure 1) showed that most health outcomes improved significantly for people with T2DM between pre- and post-MWD registration (except total, HDL, Non-HDL, and LDL cholesterol). HbA1c was reduced by 8.7 mmol/mol at 24 months post-MWD registration. Amongst those with T2DM, improvements were greater in women, the highest MWD usage group and the youngest and oldest age sub-groups. All health outcomes for T1DM were left unchanged.

Discussion and conclusion
Most outcomes significantly improved for people with T2DM post-MWD registration compared to trends pre-MWD registration, while left unchanged for people with T1DM. The large HbA1c reduction for T2DM is unusual for a low-cost scalable digitally-augmented self-management intervention and is similar in scale to commencing pharmacological treatment but considerably cheaper and without the risk of side effects. This research demonstrates the positive impact of MWD on diabetes-related outcomes and adds to the broader evidence base on digital health interventions for diabetes.