Diabetes and technology in 2030: a utopian or dystopian future?

Artificial intelligence in paediatric endocrinology: conflict or cooperation

Artificial intelligence (AI) in medicine is transforming healthcare by automating system tasks, assisting in diagnostics, predicting patient outcomes and personalising patient care, founded on the ability to analyse vast datasets. In paediatric endocrinology, AI has been developed for diabetes, for insulin dose adjustment, detection of hypoglycaemia and retinopathy screening; bone age assessment and thyroid nodule screening; the identification of growth disorders; the diagnosis of precocious puberty; and the use of facial recognition algorithms in conditions such as Cushing syndrome, acromegaly, congenital adrenal hyperplasia and Turner syndrome. AI can also predict those most at risk from childhood obesity by stratifying future interventions to modify lifestyle. AI will facilitate personalised healthcare by integrating data from 'omics' analysis, lifestyle tracking, medical history, laboratory and imaging, therapy response and treatment adherence from multiple sources. As data acquisition and processing becomes fundamental, data privacy and protecting children's health data is crucial. Minimising algorithmic bias generated by AI analysis for rare conditions seen in paediatric endocrinology is an important determinant of AI validity in clinical practice. AI cannot create the patient-doctor relationship or assess the wider holistic determinants of care. Children have individual needs and vulnerabilities and are considered in the context of family relationships and dynamics. Importantly, whilst AI provides value through augmenting efficiency and accuracy, it must not be used to replace clinical skills.

Multimodal digital phenotyping of diet, physical activity, and glycemia in Hispanic/Latino adults with or at risk of type 2 diabetes

Digital phenotyping refers to characterizing human bio-behavior through wearables, personal devices, and digital health technologies. Digital phenotyping in populations facing a disproportionate burden of type 2 diabetes (T2D) and health disparities continues to lag compared to other populations. Here, we report our study demonstrating the application of multimodal digital phenotyping, i.e., the simultaneous use of CGM, physical activity monitors, and meal tracking in Hispanic/Latino individuals with or at risk of T2D. For 14 days, 36 Hispanic/Latino adults (28 female, 14 with non-insulin treated T2D) wore a continuous glucose monitor (CGM) and a physical activity monitor (Actigraph) while simultaneously logging meals using the MyFitnessPal app. We model meal events and daily digital biomarkers representing diet, physical activity choices, and corresponding glycemic response. We develop a digital biomarker for meal events that differentiates meal events into normal and elevated categories. We examine the contribution of daily digital biomarkers of elevated meal event count and step count on daily time-in-range 54-140 mg/dL (TIR54-140) and average glucose. After adjusting for step count, a change in elevated meal event count from zero to two decreases TIR54-140 by 4.0% (p = 0.003). An increase in 1000 steps in post-meal step count also reduces the meal event glucose response by 641 min mg/dL (p = 0.0006) and reduces the odds of an elevated meal event by 55% (p < 0.0001). The proposed meal event digital biomarkers may provide an opportunity for non-pharmacologic interventions for Hispanic/Latino adults facing a disproportionate burden of T2D.

Virtual Reality Meets Diabetes

BACKGROUND

This article provides a detailed summary of virtual reality (VR) and augmented reality (AR) applications in diabetes. The purpose of this comparative review is to identify application areas, direction and provide foundation for future virtual reality tools in diabetes.

METHOD

Features and benefits of each VR diabetes application are compared and discussed, following a thorough review of literature on virtual reality for diabetes using multiple databases. The weaknesses of existing VR applications are discussed and their strengths identified so that these can be carried forward. A novel virtual reality diabetes tool prototype is also developed and presented.

RESULTS

This research identifies three major categories where VR is being used in diabetes: education, prevention and treatment. Within diabetes education, there are three target groups: clinicians, adults with diabetes and children with diabetes. Both VR and AR have shown benefits in areas of Type 1 and Type 2 diabetes.

CONCLUSIONS

Virtual reality and augmented reality in diabetes have demonstrated potential to enhance training of diabetologists and enhance education, prevention and treatment for adults and children with Type 1 or Type 2 diabetes. Future research can continually build on virtual and augmented reality diabetes applications by integrating wide stakeholder inputs and diverse digital platforms. Several areas of VR diabetes are in early stages, with advantages and opportunities. Further VR diabetes innovations are encouraging to enhance training, management and treatment of diabetes.

The Growing Role of Technology in the Care of Older Adults With Diabetes

The integration of technologies such as continuous glucose monitors, insulin pumps, and smart pens into diabetes management has the potential to support the transformation of health care services that provide a higher quality of diabetes care, lower costs and administrative burdens, and greater empowerment for people with diabetes and their caregivers. Among people with diabetes, older adults are a distinct subpopulation in terms of their clinical heterogeneity, care priorities, and technology integration. The scientific evidence and clinical experience with these technologies among older adults are growing but are still modest. In this review, we describe the current knowledge regarding the impact of technology in older adults with diabetes, identify major barriers to the use of existing and emerging technologies, describe areas of care that could be optimized by technology, and identify areas for future research to fulfill the potential promise of evidence-based technology integrated into care for this important population.

Digital Connectivity: The Sixth Vital Sign

Digital health and telehealth connectivity have become important aspects of clinical care. Connected devices, including continuous glucose monitors and automated insulin delivery systems for diabetes, are being used increasingly to support personalized clinical decisions based on automatically collected data. Furthermore, the development, demand, and coverage for telehealth have all recently expanded, as a result of the COVID-19 pandemic. Medical care, and especially diabetes care, are therefore becoming more digital through the use of both connected digital health devices and telehealth communication. It has therefore become necessary to integrate digital data into the electronic health record and maintain personal data confidentiality, integrity, and availability. Connected digital monitoring combined with telehealth communication is known as virtual health. For this virtual care paradigm to be successful, patients must have proper skills, training, and equipment. We propose that along with the five current vital signs of blood pressure, pulse, respiratory rate, temperature, and pain, at this time, digital connectivity should be considered as the sixth vital sign. In this article, we present a scale to assess digital connectivity.

Level of Digitalization in Germany: Results of the Diabetes Digitalization and Technology (D.U.T) Report 2020

BACKGROUND

New diagnostic and therapeutic technologies are increasingly changing the treatment of people with diabetes (PWD), along with increased usage of digital tools. To date, however, there is little data to which level and how diabetologists and PWD implement digitalization. Also, not much is known about the view of diabetologists on the current status and future developments in this respect.

METHOD

In an online survey, diabetologists working in clinics and practices across Germany provided responses regarding their view on digitalization and the adoption of new technologies in diabetology to 56 questions. These comments reflect the opinion of several experts about the current importance and use of specific digital/technological topics.

RESULTS

Overall, 326 diabetologists took part in the survey. They reported a positive attitude (75.8%) toward new technologies and digitalization, and they see more advantages rather than disadvantages. Younger age of the diabetologists was significantly associated with a more positive attitude (r = -0.176; P < .01), and there was no gender effect (P = .738). On average, in each practice, 5.5% of PWD are using an insulin pump for therapy, 4.8% a real-time continuous glucose monitoring system, 16.9% an intermittent scanning continuous glucose monitoring system, and 0.3% an automated insulin delivery (AID) system. With respect to digitalization, the three most important current topics are software for glucose data analysis (average rank on a scale from one to six, with one being the most important: 2.4), compatibility with other systems (2.9), and AID systems (3.8)).

CONCLUSIONS

This survey, which is going to be repeated annually, showed that the diabetologists who participated predominantly have a positive attitude toward new technologies and digital applications and were aware of the associated advantages. However, perceived disadvantages need to be addressed to enable wider adoption of new technologies and digital solutions.

Leveraging advances in diabetes technologies in primary care: a narrative review

Primary care providers (PCPs) play an important role in providing medical care for patients with type 2 diabetes. Advancements in diabetes technologies can assist PCPs in providing personalised care that addresses each patient's individual needs. Diabetes technologies fall into two major categories: devices for glycaemic self-monitoring and insulin delivery systems. Monitoring technologies encompass self-measured blood glucose (SMBG), where blood glucose is intermittently measured by a finger prick blood sample, and continuous glucose monitoring (CGM) devices, which use an interstitial sensor and are capable of giving real-time information. Studies show people using real-time CGM have better glucose control compared to SMBG. CGM allows for new parameters including time in range (the time spent within the desired target glucose range), which is an increasingly relevant real-time metric of glycaemic control. Insulin pens have increased the ease of administration of insulin and connected pens that can calculate and capture data on dosing are becoming available. There are a number of websites, software programs, and applications that can help PCPs and patients to integrate diabetes technology into their diabetes management schedules. In this article, we summarise these technologies and provide practical information to inform PCPs about utility in their clinical practice. The guiding principle is that use of technology should be individualised based on a patient's needs, desires, and availability of devices. Diabetes technology can help patients improve their clinical outcomes and achieve the quality of life they desire by decreasing disease burden.KEY MESSAGESIt is important to understand the role that diabetes technologies can play in primary care to help deliver high-quality care, taking into account patient and community resources. Diabetes technologies fall into two major categories: devices for glycaemic self-monitoring and insulin delivery systems. Modern self-measured blood glucose devices are simple to use and can help guide decision making for self-management plans to improve clinical outcomes, but cannot provide "live" data and may under- or overestimate blood glucose; patients' monitoring technique and compliance should be reviewed regularly. Importantly, before a patient is provided with monitoring technology, they must receive suitably structured education in its use and interpretation.Continuous glucose monitoring (CGM) is now standard of care for people with type 1 diabetes and people with type 2 diabetes on meal-time (prandial) insulin. Real-time CGM can tell both the patient and the healthcare provider when glucose is in the normal range, and when they are experiencing hyper- or hypoglycaemia. Using CGM data, changes in lifestyle, eating habits, and medications, including insulin, can help the patient to stay in a normal glycaemic range (70-180 mg/dL). Real-time CGM allows for creation of an ambulatory glucose profile and monitoring of time in range (the time spent within target blood glucose of 70-180 mg/dL), which ideally should be at least 70%; avoiding time above range (>180 mg/dL) is associated with reduced diabetes complications and avoiding time below range (<70 mg/dL) will prevent hypoglycaemia. Insulin pens are simpler to use than syringes, and connected pens capture information on insulin dose and injection timing.There are a number of websites, software programs and applications that can help primary care providers and patients to integrate diabetes technology into their diabetes management schedules. The guiding principle is that use of technology should be individualised based on a patient's needs, desires, skill level, and availability of devices.

Clouds and Silver Linings: COVID-19 Pandemic Is an Opportune Moment to Democratize Diabetes Care Through Telehealth

With the recent pivot to telehealth as a direct result of the COVID-19 pandemic, there is an imperative to ensure that access to affordable devices and technologies with remote monitoring capabilities for people with diabetes becomes equitable. In addition, expanding the use of remote Diabetes Self-Management Education and Support (DSMES) and Medical Nutrition Therapy (MNT) services will require new strategies for achieving long-term, effective, continuous, data-driven care. The current COVID-19 pandemic has especially impacted underserved US communities that were already disproportionately impacted by diabetes. Historically, these same communities have faced barriers in accessing timely and effective diabetes care including access to DSMES and MNT services, and diabetes technologies. Our call to action encourages all involved to urge US Federal representatives to widen access to the array of technologies necessary for successful telehealth-delivered care beyond COVID-19.

Conversational Agents in Health Care: Scoping Review and Conceptual Analysis

BACKGROUND

Conversational agents, also known as chatbots, are computer programs designed to simulate human text or verbal conversations. They are increasingly used in a range of fields, including health care. By enabling better accessibility, personalization, and efficiency, conversational agents have the potential to improve patient care.

OBJECTIVE

This study aimed to review the current applications, gaps, and challenges in the literature on conversational agents in health care and provide recommendations for their future research, design, and application.

METHODS

We performed a scoping review. A broad literature search was performed in MEDLINE (Medical Literature Analysis and Retrieval System Online; Ovid), EMBASE (Excerpta Medica database; Ovid), PubMed, Scopus, and Cochrane Central with the search terms "conversational agents," "conversational AI," "chatbots," and associated synonyms. We also searched the gray literature using sources such as the OCLC (Online Computer Library Center) WorldCat database and ResearchGate in April 2019. Reference lists of relevant articles were checked for further articles. Screening and data extraction were performed in parallel by 2 reviewers. The included evidence was analyzed narratively by employing the principles of thematic analysis.

RESULTS

The literature search yielded 47 study reports (45 articles and 2 ongoing clinical trials) that matched the inclusion criteria. The identified conversational agents were largely delivered via smartphone apps (n=23) and used free text only as the main input (n=19) and output (n=30) modality. Case studies describing chatbot development (n=18) were the most prevalent, and only 11 randomized controlled trials were identified. The 3 most commonly reported conversational agent applications in the literature were treatment and monitoring, health care service support, and patient education.

CONCLUSIONS

The literature on conversational agents in health care is largely descriptive and aimed at treatment and monitoring and health service support. It mostly reports on text-based, artificial intelligence-driven, and smartphone app-delivered conversational agents. There is an urgent need for a robust evaluation of diverse health care conversational agents' formats, focusing on their acceptability, safety, and effectiveness.

Web-Based and mHealth Technologies to Support Self-Management in People Living With Type 2 Diabetes: Validation of the Diabetes Self-Management and Technology Questionnaire (DSMT-Q)

BACKGROUND

A growing number of web-based and mobile health (mHealth) technologies have been developed to support type 2 diabetes self-management. Little is known about individuals' experiences with these technologies and how they support self-management. Appropriate tools are needed to understand how web-based and mHealth interventions may impact self-management.

OBJECTIVE

This study aimed to develop an instrument, the Diabetes Self-Management and Technology Questionnaire (DSMT-Q), to assess self-management among people living with type 2 diabetes who use web-based and mHealth technologies.

METHODS

A total of 36 candidate questionnaire items, drafted previously, were refined using cognitive debriefing interviews (n=8), expert consultation, and public patient involvement feedback. Item reduction steps were performed on survey data (n=250), and tests of validity and reliability were subsequently performed.

RESULTS

Following amendments, patients and experts found 21 items relevant and acceptable for inclusion in the instrument. Survey participants included 104 (41.6%) women and 146 (58.4%) men. Two subscales with high construct validity, internal consistency, and test-retest reliability were identified: "Understanding individual health and making informed decisions" and "Confidence to reach and sustain goals."

CONCLUSIONS

Analyses confirmed good psychometric properties in the DSMT-Q scales. This tool will facilitate the measurement of self-management in people living with type 2 diabetes who use web-based or mHealth technologies.

Multisensor Systems and Arrays for Medical Applications Employing Naturally-Occurring Compounds and Materials

The significant improvement of quality of life achieved over the last decades has stimulated the development of new approaches in medicine to take into account the personal needs of each patient. Precision medicine, providing healthcare customization, opens new horizons in the diagnosis, treatment and prevention of numerous diseases. As a consequence, there is a growing demand for novel analytical devices and methods capable of addressing the challenges of precision medicine. For example, various types of sensors or their arrays are highly suitable for simultaneous monitoring of multiple analytes in complex biological media in order to obtain more information about the health status of a patient or to follow the treatment process. Besides, the development of sustainable sensors based on natural chemicals allows reducing their environmental impact. This review is concerned with the application of such analytical platforms in various areas of medicine: analysis of body fluids, wearable sensors, drug manufacturing and screening. The importance and role of naturally-occurring compounds in the development of electrochemical multisensor systems and arrays are discussed.

The burden of type 2 diabetes in Europe: Current and future aspects of insulin treatment from patient and healthcare spending perspectives

Due to the progressive nature of type 2 diabetes (T2DM), initiation of insulin therapy is very likely in the disease continuum. This article aims at highlighting the current situation with regard to insulin therapy in people with T2DM in Europe and at presenting the associated unmet need. Challenges for both people with T2DM and healthcare professionals include clinical inertia also derived from fear of hypoglycaemia, weight gain and injections as well as increased need for a comprehensive diabetes management. We compare national and international guidelines and recommendations for the initiation and intensification of insulin therapy with the real-world situation in six European countries, demonstrating that glycaemic targets are only met in a minority of people with T2DM on insulin therapy. Furthermore, this work evaluates currently recorded numbers of people with T2DM treated with insulin in Europe, the proportion not achieving the stated glycaemic targets and thus in need to enhance insulin therapy e.g. by a change in means of insulin delivery including, but not limited to, insulin pens, wearable mealtime insulin delivery patches, patch pumps, and conventional insulin pumps with continuous subcutaneous insulin infusion.

Diabetes Healthcare Professionals Use Multiple Continuous Glucose Monitoring Data Indicators to Assess Glucose Management

BACKGROUND

Continuous glucose monitoring (CGM) offers multiple data features that can be leveraged to assess glucose management. However, how diabetes healthcare professionals (HCPs) actually assess CGM data and the extent to which they agree in assessing glycemic management are not well understood.

METHODS

We asked HCPs to assess ten de-identified CGM datasets (each spanning seven days) and rank order each day by relative glycemic management (from "best" to "worst"). We also asked HCPs to endorse features of CGM data that were important in making such assessments.

RESULTS

In the study, 57 HCPs (29 endocrinologists; 28 diabetes educators) participated. Hypoglycemia and glycemic variance were endorsed by nearly all HCPs to be important (91% and 88%, respectively). Time in range and daily lows and highs were endorsed more frequently by educators (all Ps < .05). On average, HCPs endorsed 3.7 of eight data features. Overall, HCPs demonstrated agreement in ranking days by relative glycemic control (Kendall's W = .52, P < .001). Rankings were similar between endocrinologists and educators (R² = .90, Cohen's kappa = .95, mean absolute error = .4 [all Ps < .05]; Mann-Whitney U = 41, P = .53).

CONCLUSIONS

Consensus in the endorsement of certain data features and agreement in assessing glycemic management were observed. While some practice-specific differences in feature endorsement were found, no differences between educators and endocrinologists were observed in assessing glycemic management. Overall, HCPs tended to consider CGM data holistically, in alignment with published recommendations, and made converging assessments regardless of practice.

Digital Diabetes Care System Observations from a Pilot Evaluation Study in Vietnam

Digital technologies are gaining an important role in the management of patients with diabetes. We assessed clinical outcomes and user satisfaction of incorporating a digital diabetes care system in diabetes clinics of a developing country. The system integrated a wireless blood glucose monitor that communicates data to any smartphone utilizing a patented acoustic data transfer method, a mobile-app, and cloud-based software that stores, analyzes, and presents data. Five hospital endocrinology clinics in Vietnam sequentially recruited all patients willing to join the study, providing they had a smartphone and access to internet connectivity. Face-to-face visits were conducted at baseline and at 12 weeks, with monthly digital visits scheduled in the interim and additional digital visits performed as needed. HbA1c levels were measured at baseline and at 12 weeks (±20 days). The study included 300 patients of whom 279 completed the evaluation. Average glucose levels declined from 170.4 ± 64.6 mg/dL in the first 2 weeks to 150.8 ± 53.2 mg/dL in the last 2 weeks (n = 221; p < 0.001). HbA1c levels at baseline and 12 weeks declined from 8.3% ± 1.9% to 7.6% ± 1.3% (n = 126; p < 0.001). The digital solution was broadly accepted by both patients and healthcare professionals and improved glycemic outcomes. The durability, scalability, and cost-effectiveness of this approach merits further study.

Digital Health Interventions for Diabetes: Everything to Gain and Nothing to Lose

IN BRIEF The traditional approach to integrating new therapies involves long, expensive roadmaps with evidence generation required for multiple stakeholders, most notably regulators and clinicians. More recently, new technologies such as insulin delivery systems and continuous glucose monitoring devices have become mainstream without complete clinical evidence being available when they were first introduced. There is tremendous enthusiasm from investors, industry, and people with diabetes regarding the potential of digital health to add value to diabetes care, and this enthusiasm exists despite a paucity of high-quality clinical evidence from traditional randomized clinical trials. Moreover, the potential of diabetes digital health technologies has been recognized by the U.S. Food and Drug Administration and other regulators, who are changing their approaches to allow easier, earlier access to diabetes software and devices. This wager that digital health will add value makes sense.

Health Care Professionals' Clinical Perspectives and Acceptance of a Blood Glucose Meter and Mobile App Featuring a Dynamic Color Range Indicator and Blood Sugar Mentor: Online Evaluation in Seven Countries

BACKGROUND

Despite many new therapies and technologies becoming available in the last decade, people with diabetes continue to struggle to achieve good glycemic control. Innovative and affordable solutions are needed to support health care professionals (HCPs) to improve patient outcomes.

OBJECTIVE

To gather current self-management perceptions of HCPs in seven countries and investigate HCP satisfaction with a new glucose meter and mobile app featuring a dynamic color range indicator and a blood sugar mentor.

METHODS

A total of 355 HCPs, including 142 endocrinologists (40.0%), 108 primary care physicians (30.4%), and 105 diabetes nurses (29.6%), were recruited from the United Kingdom (n=50), France (n=50), Germany (n=50), India (n=54), Algeria (50), Canada (n=51), and the United States (n=50). HCPs experienced the OneTouch Verio Reflect glucose meter and the OneTouch Reveal mobile app online from their own office computers using interactive demonstrations via webpages and multiple animations. After providing demographic and clinical practice insights, HCPs responded to statements about the utility of the system.

RESULTS

Concerning current practice, 83.1% (295/355) of HCPs agreed that poor numeracy or health literacy was a barrier for their patients. A total of 85.9% (305/355) and 92.1% (327/355) of HCPs responded that type 2 diabetes (T2D) and type 1 diabetes (T1D) patients were aware of what represented a low, in-range, or high blood glucose result. Only 62.0% (220/355) felt current glucose meters made it easy for patients to understand if results were in range. A total of 50.1% (178/355) and 78.0% (277/355) of HCPs were confident that T1D and T2D patients took action for low or high results. A total of 87.0% (309/355) agreed that the ColorSure Dynamic Range Indicator could help them teach patients how to interpret results and 88.7% (315/355) agreed it made them more aware of hyper- and hypoglycemic results so they could take action. A total of 83.7% (297/355) of HCPs agreed that the Blood Sugar Mentor feature gave personalized guidance, insight, and encouragement so patients could take action. A total of 82.8% (294/355) of HCPs also agreed that the Blood Sugar Mentor provided real-time guidance to reinforce the goals HCPs had set so patients could take steps to manage their diabetes between office visits. After experiencing the full system, 85.9% (305/355) of HCPs agreed it was beneficial for patients with lower numeracy or health literacy; 96.1% (341/355) agreed that it helped patients understand when results were low, in range, or high; and 91.0% (323/355) agreed that the way it displayed diabetes information would make patients more inclined to act upon results. A total of 89.0% (316/355) of HCPs agreed that it would be helpful for agreeing upon appropriate in-range goals for their patients for their next clinic visit.

CONCLUSIONS

This multi-country online study provides evidence that HCPs were highly satisfied with the OneTouch Verio Reflect meter and the OneTouch Reveal mobile app. Each of these use color-coded information and the Blood Sugar Mentor feature to assist patients with interpreting, analyzing, and acting upon their blood glucose results, which is particularly beneficial to keep patients on track between scheduled office visits.

The Big Picture on the "AI Turn" for Digital Health: The Internet of Things and Cyber-Physical Systems

This article offers an analysis of the ways in which digital health innovations are being coproduced by mainstreaming of artificial intelligence (AI), the Internet of Things (IoT), and cyber-physical systems (CPS) in health care. CPS blurs the boundaries between the physical and virtual worlds, and creates a dynamic digital map of all things in existence that can be analyzed in ways that are much more sophisticated than a bar code scanning system. Examples of CPS include self-driving cars, wearables for digital monitoring of heart arrhythmias, industrial AI-powered robots in smart factories and health robots delivering home care services to disabled persons and rural communities. Another interesting prospect of digital health powered by AI, IoT, and CPS is remote phenotypic data capture and characterization of pharmaceutical outcomes in clinical trials in ways that are user centric and meaningful to patients. For rural or remote communities with limited access to medical product information, the IoT could bring about pharmacy and health services innovation. There are unprecedented societal challenges at intersections of digital health with AI, IoT, and CPS as well. For example, the physical and virtual worlds markedly differ in speed, scale, and temporalities, as do our physical self and digital footprints. Our efforts to map and develop effective solutions to societal corollaries of AI, IoT, and CPS need to bear in mind such asymmetries between the physical and virtual worlds. A societal issue such as privacy may emerge in different forms and intensities in the physical and virtual contexts. Digital data are highly fluid and can rapidly move across spaces and places, whereas the physical data and humans are much slower and exist in different scales than our digital footprints. It is therefore timely for the system sciences and integrative biology communities to critically engage with digital health and the related technologies, such as AI, IoT, and CPS.

Application of Machine Learning to Identify Clustering of Cardiometabolic Risk Factors in U.S. Adults

Aims: The aim of this study is to compare some machine learning methods with traditional statistical parametric analyses using logistic regression to investigate the relationship of risk factors for diabetes and cardiovascular (cardiometabolic risk) for U.S. adults using a cross-sectional data from participants in a wellness improvement program. Methods: Logistic regression was used to find the relationship between individual risk factors, predictor and cardiometabolic risk. Supervised machine learning methods were used to predict risk and produce a ranking of variables' importance. A clustering method was used to identify subpopulations of interest. Predictors were divided into those that are nonmodifiable and those that are modifiable. Results: The population comprised 217,254 adults of whom 8.1% had diabetes. Using logistic regression, six variables were identified to be negatively related and eleven were positively related to cardiometabolic risk. Three supervised machine learning classifiers (random forest, gradient boosting, and bagging) were applied with average AUC to be 0.806. Each classifier also produced a ranking of variables' importance. Four subgroups were identified with a k-medoid clustering algorithm, which were mainly distinguished by gender and diabetes status. Conclusions: The study illustrates that machine learning is an important addition to traditional logistic regression in terms of identifying important cardiometabolic risk factors and ranking their importance and the potential for interventions based on lifestyle and medications at an individual level.

Digital Diabetes Data and Artificial Intelligence: A Time for Humility Not Hubris

In the future artificial intelligence (AI) will have the potential to improve outcomes diabetes care. With the creation of new sensors for physiological monitoring sensors and the introduction of smart insulin pens, novel data relationships based on personal phenotypic and genotypic information will lead to selections of tailored, effective therapies that will transform health care. However, decision-making processes based exclusively on quantitative metrics that ignore qualitative factors could create a quantitative fallacy. Difficult to quantify inputs into AI-based therapeutic decision-making processes include empathy, compassion, experience, and unconscious bias. Failure to consider these "softer" variables could lead to important errors. In other words, that which is not quantified about human health and behavior is still part of the calculus for determining therapeutic interventions.