Fortunately, digital health has been pretty active when it comes to innovating for diabetes management. Just this month Fitbit invested $6 Million in a startup working on continuous glucose monitoring, while diabetes management company Intuity Medical brought in $40 million in equity financing. That news closely followed the launch of Ascensia’s Diabetes Challenge, a global innovation competition aimed at supporting startups and entrepreneurs with solutions for managing type two diabetes. The list of digital diabetes management platforms goes on.
But what is the big picture of digital therapies for diabetes? When we step back from individual apps and devices, how is digital health shaping the future of diabetes management?
Earlier this month, a group of researchers in Israel published an article reviewing the state of digital health technology and diabetes management. They explain that digital health is transforming diabetes management with innovations around EMR data, telemedicine, apps and decision support software.
Electronic medical records have obvious benefits for the daily operations of hospitals and clinics. But EMRs are also changing the way medicine is delivered to populations – especially populations living with chronic diseases like diabetes.
The potential of EMR data lies in the ability to analyze data retrospectively. Unlike clinical trials, this type of research can consider the health behavior of hundreds of thousands of individuals spanning decades. For example, researchers have used mathematical models based on EMR data to define biomarkers that predict diabetes and its complications, identify drug targets, characterize differential responses to therapy, correlate genetic and clinical data and create predictive economic models.
Of course there are limitations to research based on EMRs. This data wasn’t designed for research, so it isn’t always accurate enough for analysis. However, new sources of digital data will likely shape diabetes therapies in the coming decades.
“Incorporation of EMR data with data derived from personal sensors and wearables may represent the future of data management in diabetes,” according to the recent review article. “Combining these data with patients’ genomic, proteomic, and metabolomics profile may enable one to derive clinical recommendations that are based upon an integrated picture of medical, social, and biological data.”
APPS AND DIGITAL CLINICS
The nature of diabetes requires individual patients to be active participants in their own care. To successfully manage diabetes and improve outcomes, patients have to constantly manage their eating, exercise and insulin levels.
For several decades now, various forms of telemedicine have been used to improve what experts call “diabetes self-management education” (DSME). These include telephone calls, interactive voice response calls, text message services, video calls, email and mobile technologies like smartphone apps.
Market research estimates that 71 percent of people with diabetes own a smartphone. Apps are thus a natural method for improving diabetes self management and education.
Market research estimates that 71 percent of people with diabetes own a smartphone. Apps are thus a natural method for improving diabetes self management and education. However, apps designed to diagnose, cure, mitigate, treat or prevent diabetes are considered devices, which require regulatory clearance. Few apps have been cleared by the FDA or published clinical outcomes in peer-reviewed journals.
Still, there are thousands of apps on the market designed to improve diabetes management. Research reviews divide these into seven categories:
- Information provision is the simplest app function, and this has evolved from simple displays into interactive platforms tailored to the user’s profile and preferences.
- Calorie and carbohydrate counting apps help patients search food databases and look up nutritional information.
- Weight loss apps range from simple trackers to more complex systems that provide incentives for healthy behaviors.
- Physical activity apps often incorporate wearables and sensors to help users record physical activity, goal setting and motivation.
- Recording glucose levels or insulin doses is another app function. While glucometers store glucose levels and insulin pumps can calculate insulin doses, some apps incorporate both of these functions.
- Medication adherence apps give users reminders to take medications or insulin on time. They can also analyze data regarding drug compliance and share it with healthcare professionals or authorized family members.
- Enhancing motivation and compliance is perhaps the most important and challenging function of diabetes apps. The newest apps are using principles from machine learning to selectively apply coaching and motivation methods that are most likely to work based on an individual’s lifestyle and previous experience.
The important question to ask about these apps is, “are they working?” A recent meta-analysis summarized 87 trials to see how diabetes-related telemedicine interventions affect HbA1c and other measures like quality of life and mortality. HbA1c refers to glycated hemoglobin, which is often measured to get an overall picture of what our average blood sugar levels have been over a period of weeks or months.
The meta-analysis found a big variation in the effectiveness of digital diabetes interventions, but the overall conclusion was somewhat discouraging. Overall telemedicine solutions reduced HbA1c by 0.28 percent, meaning they were not very effective. As the recent review article concludes, “although patient-centered health communication technology appeared to be a most promising approach to curb the increasing diabetes epidemic, the magnitude of the effect and the sustainability of results appear to be somewhat discouraging.”
There are several reasons why research might find that digital diabetes management solutions are less than effective. User fatigue can curb compliance before patients see the full benefits of an intervention. Many apps also ask too much of patients, like constant data input. There are also inherent limitations to stand-alone apps that don’t connect patients to their healthcare provider.
Researchers argue that there is better potential for digital clinics that combine data from mobile apps or wearables and interactions with healthcare professionals.
DECISION SUPPORT SOFTWARE
With the increasing use of glucometers, continuous glucose monitors, insulin pumps, and other wearables, patients with diabetes are generating enormous amounts of data. As with other areas of medicine, the quantity of this data far exceeds the analytic ability of the human mind. New software is making this data more accessible to providers and leveraging machine learning methods to support decision making.
Most decision support software for diabetes management is at an early stage, but it is already being incorporated into care. Algorithms range from simple alerts in the EMR to tools that recommend the best therapy for patients based on analysis of differential responses. “The clinical validity and safety of these algorithms needs further investigation,” according to the latest review article of digital therapies for diabetes management.
A CONCEPTUAL MISMATCH
The review article ends with an intuitive insight that is nevertheless a useful reminder: “The computer industry is conceptually different from clinical care.” While technology develops rapidly, often based on beta testing and trial and error, medicine progresses more cautiously. Clinical care is highly regulated and rightly bases decisions on scientific evidence with the aim of minimizing errors.
This conceptual mismatch – between the computer industry and clinical care – leads to a tricky situation. While healthcare is often slow to accept new technologies, patients accustomed to life in the digital age insist on getting state of the art digital health. As the medical and hi-tech industries continue to collaborate on innovative diabetes solutions, researchers say new technologies need to be scrutinized carefully to ensure they are cost-effective and improve clinical outcomes.