Use of technology in type 2 diabetes management

Recently, there has been an acceleration of uptake of technology for managing diabetes: as an adjunct to conventional therapy, to improve self-management and to provide education. This presents both challenges and opportunities for general practitioners and people with type 2 diabetes. 

The technology available to help manage diabetes falls into three main categories: 

• information technology – such as mobile phone apps, SMS messaging, wearable technology (eg fitness trackers, smartwatches), web-based programs and clinic-based chronic disease care programs 
• technological innovations for monitoring of glycaemia – such as CGM and flash glucose monitoring (FlashGM), which provide greater insights into glycaemic variability patterns and alarm systems for defined hypoglycaemia or hyperglycaemic excursions 
• technology for medication delivery – such as evolving medication pen devices and continuous subcutaneous infusion of insulin (insulin pumps). Although insulin pump infusions have traditionally been used mainly by people with type 1 diabetes, they are increasingly being used in type 2 diabetes. 

Technology: Clinical utility 

A recent meta-analysis found that information technology such as mobile phone apps and web-based applications combined with standard diabetes care resulted in clinically significant reduction in glycated haemoglobin (HbA1c) in people with type 2 diabetes.² In addition, there is emerging evidence that information technology interventions are associated with: 

• reduced sedentary behaviour (computer, mobile and wearable technologies)³ 
• increased physical activity (online self-tracking program)⁴ 
• improvements in diet and exercise, including understanding of nutrition (counselling delivered via mobile phone messaging).⁵ 

Continuous glucose monitoring 

What is it? 

CGM involves a small sensor being implanted in the subcutaneous tissue to monitor interstitial glucose. ‘Real-time’ CGM continuously records and reports glucose levels, with some remote devices at times using visual or auditory ‘alarms’ to alert users to hypoglycaemia or hyperglycaemia. CGM measures interstitial glucose, and is not exactly equivalent to capillary blood glucose measurement (eg there may be a delay of a maximum of 15 minutes between interstitial glucose being equilibrated with capillary glucose levels), which remains the standard for confirmation of high and low blood glucose levels and treatment decisions. 

FlashGM, also called ‘intermittently viewed CGM’, uses a disc device, worn on the arm, that can be scanned with a reader or smart phone to obtain interstitial glucose results instantly.⁶ These devices do have similar alerts for either low or high blood glucose levels and trends to changing levels. 

How does it help? 

HbA1c is the standard for assessing long-term glycaemic management; however, it does not reflect within-day and day-to-day glycaemic variability that might lead to hypoglycaemia or postprandial hyperglycaemia.⁷ 

CGM can be a useful clinical tool to detect glycaemic patterns, including hypoglycaemia, and hyperglycaemic events, and assist in the assessment of the quality of glycaemic management, evaluate glycaemic variability and patterns of hypoglycaemia.⁸ Clinical situations may include sick-day management, effects of lifestyle changes on glucose and complex insulin initiation and titration. Evidence is less robust in support of the use of CGM in people with type 2 diabetes on non-insulin glucose-lowering medicines or premix insulin.¹ 

Increasingly, standardised reporting that uses the ambulatory glucose profile (AGP)9 is being adopted. AGP represents the modal distribution of interstitial glucose in a graphic form, which allows the identification of issues such as hypoglycaemic risk, glycaemic variability and excessive glycaemic excursions, which informs clinical intervention such as modifying pharmacotherapy or implementing medical nutrition therapy. The Australian Diabetes Society has published a practical guide to interpret CGM and FlashGM data. 

The minimum duration of CGM to obtain enough data to effectively characterise and interpret glycaemia patterns has been reported as at least 7 days.¹ 

Accuracy of CGM 

The accuracy of CGM is often reported as the ‘mean absolute relative difference’ (MARD) between the CGM system values and matched reference values. A MARD of ≤10% is considered desirable.¹⁰ 

Calibration requirements for each sensor may vary. FlashGM sensors do not require calibration; however, discrepancy with SMBG can occur when glucose levels are changing rapidly or in a lower glucose range. Compression on the sensor (eg when lying on it while asleep) can lead to false reporting of hypoglycaemia due to restriction of flow of interstitial fluid around the sensor. Glucose levels should be confirmed with a fingerprick assessment if:⁶ 

• glucose levels are changing rapidly 
• sensors indicate hypoglycaemia or possible hypoglycaemia 
• a person displays symptoms inconsistent with reported glucose levels. 

Continuous subcutaneous insulin infusion (insulin pumps) 

Continuous subcutaneous insulin infusion (CSII) allows for more controlled delivery of insulin compared with injectable insulin, particularly for basal insulin. Pumps deliver basal plus bolus (prandial and correction) doses that can be programmed to change in response to the user’s changing needs (eg meal times, exercise). These integrated systems are referred to as automated insulin delivery (AID) systems. 

Integrated smart insulin pens 

Smart insulin pens provide additional functionality beyond insulin delivery. These additional properties may include: 

• electronic recording of insulin dosing information (type of insulin, time of injection and number of units injected) 
• electronic data sharing with a healthcare professional via an appropriate app, creating an insulin dosing summary 
• integrating insulin dosing and timing with CGM, providing insights into how insulin may affect blood glucose and, combined with compatible apps, allowing overlay of blood glucose data with insulin dosing information. 

Mobile apps, web-based programs, text messaging 

Many practices already use web- and phone-based messaging for recalls, reminders and appointment scheduling (eg mySugr app for calculation of the insulin to carbohydrate ratio). 

More work needs to be done to determine the most effective interventions and the optimal integration of technology with validated models of care for chronic disease management. 

AID systems 

The decision to implement an AID system incorporating CSII and CGM is a case-by-case assessment based on cost–benefit analysis, individualising the decisions according to needs, wishes and capacity. These technologies can be costly and resource intensive, and might increase stress and distress to the person. They also require careful education delivered within specialised trained healthcare team environments. Thus, the introduction, implementation and ongoing use of any complex technology requires high levels of professional support to instruct users about the appropriate use and interpretation of outcomes.^11,12 

Clinicians who recommend these technologies should be experienced in their use or consult experts such as endocrinologists or credentialled diabetes educators. 

The National Association of Diabetes Centres has developed national standards for diabetes technology to guide primary care user. 

Consider the use of CGM or FlashGM in individuals with type 2 diabetes on basal insulin regimens who have suboptimal glycaemic control, based on the recent Asia-Pacific consensus recommendations.¹ 

Individuals who might benefit most from CGM or FlashGM are those:¹⁰ 

• at high risk of hypoglycaemia 
• with hypoglycaemic unawareness 
• with high glycaemic variability. 

Intermittent use of CGM or FlashGM by the person can be a useful adjunct to SMBG. 

Those likely to benefit from CSII most are those: 

• with the most unstable glycaemic levels 
• with recurrent hypoglycaemia 
• who are engaged with the additional offerings of the technology beyond insulin delivery. 

When paired with CSII, the benefits of CGM are added to those of CSII. 

Potential barriers include: 

• cost (insulin pumps are covered by most private health insurers, but consumables are not; the National Diabetes Services Scheme subsidies consumables only for people living with type 1 diabetes) 
• lack of technical or information technology literacy (users need to navigate pump menus, upload pump and/or CGM data, be able to ‘troubleshoot’) 
• level of clinical and technological support that is required from family, healthcare professionals and purveyors of technology 
• the dexterity required to apply infusion sets, CGM sensors and transmitters. 

Recommended glycaemic targets for users of CGM/FlashGM with type 2 diabetes (not during pregnancy) are as follows:¹³ 

• time in range – a target of 3.9–10 mmol/L should be maintained at least 70% of the time 
• time below range – blood glucose levels <3.9 mmol/L should occur for less than 4% of the day (approximately one hour); very low levels (<3.0 mmol/L) should occur for no more than 1% of the day (15 minutes) 
• time above range – blood glucose levels >10 mmol/L should occur less than 25% of the time; very high levels (>13.9 mmol/L) should occur less than 5% of the time. 

The following targets are recommended for older or high-risk individuals with type 2 diabetes:¹³ 

• time in range – a target of 3.9–10 mmol/L should be maintained more than 50% of the time 
• time below range – avoiding hypoglycaemia is a priority in this population, so blood glucose levels <3.9 mmol/L should occur for less than 1% of the day, or 15 minutes 
• time above range – very high blood glucose levels of >13.9 mmol/L should be allowed for less than 10% of the time. 

Battelino et al¹³ have published detailed information about clinical glucose targets for CGM, including CGM-based targets for different diabetes populations (figure 1). 

The AGP enables retrospective analysis of dense data, trends and patterns for people with diabetes and their healthcare team to help achieve appropriate glucose targets and to minimise hypoglycaemia and hyperglycaemia.⁹ 

Artificial Intelligence in diabetes management is emerging mostly related to fully automated insulin delivery that detects blood glucose and adjusts drug delivery or can suspend insulin if hypoglycaemia is detected; however, this application is mostly focused on type 1 diabetes. The article by Guan et al¹⁴ gives an overview of current opportunities for artificial intelligence with type 2 diabetes.