Precision monitoring in diabetes
Combining data from continuous glucose monitoring with self-care and mental health data could help improve diabetes care. A new review in Diabetologia looks at the emerging area of precision monitoring. Dr Susan Aldridge reports.
Glucose monitoring to help achieve treatment goals is at the heart of diabetes management. It has long been known that glucose levels are affected by food and exercise, but recent research has pointed to the importance of other contextual factors, such as sleep, mood, diabetes distress and psychological comorbidities, which also exert an influence on glycaemia. There is increasing availability of devices, apps and other sensors that measure eating, sleep and activity patterns, as well as psychological variables. This opens up the possibility of precision monitoring in diabetes, by generating large datasets that combine continuous glucose monitoring (CGM) with data relating to the above variables.
The current American Diabetes Association (ADA) and European Association for the Study of Diabetes (EASD) consensus statement calls for precision monitoring and defines it as the multimodal assessment of glucose, behaviours, diet, sleep and psychophysiological stress. A review published in a recent issue of Diabetologia, from Norbert Hermanns and colleagues from the German Centre for Diabetes Research, and elsewhere, looks at two aspects of precision monitoring: prognostics and therapeutics. Precision prognostics is about how precision monitoring could be used to improve prediction of diabetes outcomes, and precision therapeutics is concerned with intervention strategies that could be developed following the insights of precision monitoring.
An early example of precision prognostics, dating back to 1997, was the development of the low blood glucose index (a weighted measure of low glucose values), which could be combined with clinical data on severe hypoglycaemia to identify subgroups of people with type 1 diabetes who have different risks for future hypoglycaemia. With today’s technology, much more refined subgrouping of this kind could be achieved.
Mental health and glycaemic control
The burden of self-managing diabetes, day in day out, can affect mental health, with diabetes distress and depression being common among those with the condition. This can have a negative impact on self-care behaviours and quality of life – both of which may affect outcomes. Questionnaires have long been the mainstay of assessing mental health, but do not lend themselves to the precision-monitoring approach. But ecological momentary assessment (EMA), which allows repeated daily sampling of experiences and behaviours, does enable monitoring of mental health states and self-care in the context of CGM data.
A recent systematic review suggested there is no convincing relationship between mood and glucose variability. However, it did show a significant link between postprandial glucose and negative mood in people with type 2 diabetes and a potential positive effect of lower glucose variability on depressive mood in adults with type 1 diabetes. In summary, there is growing evidence that negative mood states are associated with elevated or low glucose values, while glucose values within the normal range are linked with positive mood states. However, the causality and directionality of these association requires clarification. For instance, does seeing low (or high) values on your CGM output make you depressed – or is it the other way round?
There are some current studies that combine EMA and CGM. Their findings could help identify people whose mental health is more strongly influenced by their glycaemic control and those for whom it is not a factor. From the precision therapeutics point of view, different therapeutic strategies might then be offered – for instance, in the first subgroup, mental health interventions might be more effective when including the diabetes context, and in the second maybe glycaemic control and mental health would be better addressed independently.
Behaviour and glycaemic control
EMA has been used to study the impact of behaviour on glycaemic control. One study showed that higher variability in self-care was associated with a higher percentage of glucose values out of range, particularly in the hyperglycaemic range. Another showed that stronger negative affect was associated with fewer glucose checks, especially in teenagers with elevated HbA1c. Meanwhile, people with type 1 who had higher levels of guilt, frustration or diabetes distress had a higher risk of binge eating which led, in turn, to higher postprandial glucose excursions.
Monitoring of behaviours shows that there are direct links between self-care and glycaemic control. Identifying those who have issues with self-care behaviours might allow precision therapeutics through offering specific support and early intervention, particularly if there is a link with mental health problems.
Sleep and glycaemic control
Sleep, of course, is influential in hormonal regulation and circadian rhythm and sleep disturbances can lead to a wide range of physiological and psychological problems. Sleep problems are more common in people with diabetes and, indeed, reduced sleep may even be a cause of type 1. Notably, diabetes self-management with modern devices can actually be a cause of disrupted sleep, through their alerts and alarms.
Meta-analyses have suggested that shorter sleep duration and lower sleep quality are associated with suboptimal glycaemic control. For instance, in adults with type 1, those who slept for six hours or less or had poor self-reported sleep quality had a higher HbA1c. Wearables that can assess both sleep duration and sleep quality could be used to identify those with disordered sleep. And then, interventions to address the sleep issue might help improve glycaemic control and long-term outcomes.
Precision prognostics and therapeutics
Combining and synthesising glucose levels, self-care behaviour and mental health data could contribute to better prediction of long-term diabetes outcomes. All three, or two of the three, data sources could be combined, as appropriate.
The approach to precision monitoring should be tailored to the individual. The authors give one example, of three people with different associations between diabetes distress, assessed via EMA, and exposure to low glucose. There was a strong link in one individual, where it might be appropriate to reduce distress by making strenuous efforts to avoid low glucose – for instance, by adjusting glycaemic targets, changing medication, providing diabetes education or offering a pump or automated insulin delivery (AID). In the other two individuals, there was no obvious link between diabetes distress and low glucose, so a different approach to avoiding hypoglycaemia might be employed with them.
The authors suggest that, going forward, monitoring results could be integrated into treatment decisions. For instance, although AID can increase time in range, 70% seems to be the limit, suggesting maybe we haven’t reached the full potential of these systems. Maybe the integration of monitoring of data on stress, exercise and carbohydrates could be used to inform AID algorithms about upcoming glucose excursion, allowing earlier response?
Precision monitoring of glucose, behaviour and mental health could also be used to trigger so-called just-in-time adaptive interventions, which could be tailored to the individual. For instance, if high glucose, low adherence and high stress are identified, it could trigger a suggestion for a correction on their smartphone.
A roadmap to precision monitoring
Of course, there are several knowledge gaps that need to be overcome before precision monitoring enters the mainstream of diabetes care. The authors have created a roadmap that addresses these issues. First, the quality of the monitoring data itself – we need to know more about its accuracy and reliability. Then there is the issue of causality and directionality – for instance, does depression cause low glucose or is it the other way round? And what about interventions designed to act on monitoring data? These need to be tested in clinical trials. Finally, the precision-monitoring approach needs to be tested in the real world, looking at how acceptable people with diabetes find it to have so much data collected about them. But larger studies will provide useful information on the subgroups of people who can most benefit from precision monitoring.
In summary, precision monitoring in diabetes is a new and developing field of research and clinical care. Monitoring of self-care behaviour and mental health can enhance glucose data by providing a context to the values. Precision monitoring could identify psycho-behavioural phenotypes, who could then benefit from an individualised approach to their care. But to achieve this, there is a need for automated combination and integration of data sources. So, while precision monitoring is very much in its early days, it is a step further towards giving people with diabetes and healthcare professionals new tools to individualise and optimise therapy and care.
To read this article, go to: Hermanns N, Ehrmann D, Shapira A, et al. Coordination of glucose monitoring, self-care behaviour and mental health: achieving precision monitoring in diabetes. Diabetologia online 5 April. https://doi.org/10.1007/s00125-022-05685-7
For more on blood glucose monitoring, enrol on the following EASD e-Learning courses:
Any opinions expressed in this article are the responsibility of the EASD e-Learning Programme Director, Dr Eleanor D Kennedy.
