Requirements of an Application to Monitor Diet, Physical Activity and Glucose Values in Patients with Type 2 Diabetes: The Diameter

Missing Data Statistics Provide Causal Insights into Data Loss in Diabetes Health Monitoring by Wearable Sensors

BACKGROUND

Data loss in wearable sensors is an inevitable problem that leads to misrepresentation during diabetes health monitoring. We systematically investigated missing wearable sensors data to get causal insight into the mechanisms leading to missing data.

METHODS

Two-week-long data from a continuous glucose monitor and a Fitbit activity tracker recording heart rate (HR) and step count in free-living patients with type 2 diabetes mellitus were used. The gap size distribution was fitted with a Planck distribution to test for missing not at random (MNAR) and a difference between distributions was tested with a Chi-squared test. Significant missing data dispersion over time was tested with the Kruskal-Wallis test and Dunn post hoc analysis.

RESULTS

Data from 77 subjects resulted in 73 cleaned glucose, 70 HR and 68 step count recordings. The glucose gap sizes followed a Planck distribution. HR and step count gap frequency differed significantly (p < 0.001), and the missing data were therefore MNAR. In glucose, more missing data were found in the night (23:00-01:00), and in step count, more at measurement days 6 and 7 (p < 0.001). In both cases, missing data were caused by insufficient frequency of data synchronization.

CONCLUSIONS

Our novel approach of investigating missing data statistics revealed the mechanisms for missing data in Fitbit and CGM data.

AI and semantic ontology for personalized activity eCoaching in healthy lifestyle recommendations: a meta-heuristic approach

BACKGROUND

Automated coaches (eCoach) can help people lead a healthy lifestyle (e.g., reduction of sedentary bouts) with continuous health status monitoring and personalized recommendation generation with artificial intelligence (AI). Semantic ontology can play a crucial role in knowledge representation, data integration, and information retrieval.

METHODS

This study proposes a semantic ontology model to annotate the AI predictions, forecasting outcomes, and personal preferences to conceptualize a personalized recommendation generation model with a hybrid approach. This study considers a mixed activity projection method that takes individual activity insights from the univariate time-series prediction and ensemble multi-class classification approaches. We have introduced a way to improve the prediction result with a residual error minimization (REM) technique and make it meaningful in recommendation presentation with a Naïve-based interval prediction approach. We have integrated the activity prediction results in an ontology for semantic interpretation. A SPARQL query protocol and RDF Query Language (SPARQL) have generated personalized recommendations in an understandable format. Moreover, we have evaluated the performance of the time-series prediction and classification models against standard metrics on both imbalanced and balanced public PMData and private MOX2-5 activity datasets. We have used Adaptive Synthetic (ADASYN) to generate synthetic data from the minority classes to avoid bias. The activity datasets were collected from healthy adults (n = 16 for public datasets; n = 15 for private datasets). The standard ensemble algorithms have been used to investigate the possibility of classifying daily physical activity levels into the following activity classes: sedentary (0), low active (1), active (2), highly active (3), and rigorous active (4). The daily step count, low physical activity (LPA), medium physical activity (MPA), and vigorous physical activity (VPA) serve as input for the classification models. Subsequently, we re-verify the classifiers on the private MOX2-5 dataset. The performance of the ontology has been assessed with reasoning and SPARQL query execution time. Additionally, we have verified our ontology for effective recommendation generation.

RESULTS

We have tested several standard AI algorithms and selected the best-performing model with optimized configuration for our use case by empirical testing. We have found that the autoregression model with the REM method outperforms the autoregression model without the REM method for both datasets. Gradient Boost (GB) classifier outperforms other classifiers with a mean accuracy score of 98.00%, and 99.00% for imbalanced PMData and MOX2-5 datasets, respectively, and 98.30%, and 99.80% for balanced PMData and MOX2-5 datasets, respectively. Hermit reasoner performs better than other ontology reasoners under defined settings. Our proposed algorithm shows a direction to combine the AI prediction forecasting results in an ontology to generate personalized activity recommendations in eCoaching.

CONCLUSION

The proposed method combining step-prediction, activity-level classification techniques, and personal preference information with semantic rules is an asset for generating personalized recommendations.

Thermosensitive Microneedles Capable of On Demand Insulin Release for Precise Diabetes Treatment

As a novel painless and minimally invasive transdermal drug delivery method, microneedles have solved the challenges of microbial infection and tissue necrosis associated with multiple subcutaneous injections in patients with diabetes. However, traditional soluble microneedles cannot switch drug release on and off according to the patient's needs during long-term use, which is one of the most critical elements of diabetes treatment. Herein, an insoluble thermosensitive microneedle (ITMN) that can control the release of insulin by adjusting the temperature, enabling the precise treatment of diabetes is designed. Thermosensitive microneedles are produced by in situ photopolymerization of the temperature-sensitive compound N-isopropylacrylamide with the hydrophilic monomer N-vinylpyrrolidone, which is encapsulated with insulin and bound to a mini-heating membrane. ITMN are demonstrated to have good mechanical strength and temperature sensitivity, can release significantly different insulin doses at different temperatures, and effectively regulate blood glucose in type I diabetic mice. Therefore, the ITMN provides a possibility for intelligent and convenient on-demand drug delivery for patients with diabetes, and when combined with blood glucose testing devices, it has the potential to form an integrated and precise closed-loop treatment for diabetes, which is of great importance in diabetes management.

A Digital Coach (E-Supporter 1.0) to Support Physical Activity and a Healthy Diet in People With Type 2 Diabetes: Acceptability and Limited Efficacy Testing

BACKGROUND

A healthy lifestyle, including regular physical activity and a healthy diet, is increasingly part of type 2 diabetes (T2D) management. As many people with T2D have difficulty living and maintaining a healthy lifestyle, there is a need for effective interventions. eHealth interventions that incorporate behavior change theories and tailoring are considered effective tools for supporting a healthy lifestyle. The E-Supporter 1.0 digital coach contains eHealth content for app-based eHealth interventions and offers tailored coaching regarding physical activity and a healthy diet for people with T2D.

OBJECTIVE

This study aimed to assess the acceptability of E-Supporter 1.0 and explore its limited efficacy on physical activity, dietary behavior, the phase of behavior change, and self-efficacy levels.

METHODS

Over a span of 9 weeks, 20 individuals with T2D received daily motivational messages and weekly feedback derived from behavioral change theories and determinants through E-Supporter 1.0. The acceptability of the intervention was assessed using telephone-conducted, semistructured interviews. The interview transcripts were coded using inductive thematic analysis. The limited efficacy of E-Supporter 1.0 was explored using the Fitbit Charge 2 to monitor step count to assess physical activity and questionnaires to assess dietary behavior (using the Dutch Healthy Diet index), phase of behavior change (using the single-question Self-Assessment Scale Stages of Change), and self-efficacy levels (using the Exercise Self-Efficacy Scale).

RESULTS

In total, 5 main themes emerged from the interviews: perceptions regarding remote coaching, perceptions regarding the content, intervention intensity and duration, perceived effectiveness, and overall appreciation. The participants were predominantly positive about E-Supporter 1.0. Overall, they experienced E-Supporter 1.0 as a useful and easy-to-use intervention to support a better lifestyle. Participants expressed a preference for combining E-Supporter with face-to-face guidance from a health care professional. Many participants found the intensity and duration of the intervention to be acceptable, despite the coaching period appearing relatively short to facilitate long-term behavior maintenance. As expected, the degree of tailoring concerning the individual and external factors that influence a healthy lifestyle was perceived as limited. The limited efficacy testing showed a significant improvement in the daily step count (z=-2.040; P=.04) and self-efficacy levels (z=-1.997; P=.046) between baseline and postintervention. Diet was improved through better adherence to Dutch dietary guidelines. No significant improvement was found in the phase of behavior change (P=.17), as most participants were already in the maintenance phase at baseline.

CONCLUSIONS

On the basis of this explorative feasibility study, we expect E-Supporter 1.0 to be an acceptable and potentially useful intervention to promote physical activity and a healthy diet in people with T2D. Additional work needs to be done to further tailor the E-Supporter content and evaluate its effects more extensively on lifestyle behaviors.

An automatic and personalized recommendation modelling in activity eCoaching with deep learning and ontology

Electronic coaching (eCoach) facilitates goal-focused development for individuals to optimize certain human behavior. However, the automatic generation of personalized recommendations in eCoaching remains a challenging task. This research paper introduces a novel approach that combines deep learning and semantic ontologies to generate hybrid and personalized recommendations by considering "Physical Activity" as a case study. To achieve this, we employ three methods: time-series forecasting, time-series physical activity level classification, and statistical metrics for data processing. Additionally, we utilize a naïve-based probabilistic interval prediction technique with the residual standard deviation used to make point predictions meaningful in the recommendation presentation. The processed results are integrated into activity datasets using an ontology called OntoeCoach, which facilitates semantic representation and reasoning. To generate personalized recommendations in an understandable format, we implement the SPARQL Protocol and RDF Query Language (SPARQL). We evaluate the performance of standard time-series forecasting algorithms [such as 1D Convolutional Neural Network Model (CNN1D), autoregression, Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU)] and classifiers [including Multilayer Perceptron (MLP), Rocket, MiniRocket, and MiniRocketVoting] using state-of-the-art metrics. We conduct evaluations on both public datasets (e.g., PMData) and private datasets (e.g., MOX2-5 activity). Our CNN1D model achieves the highest prediction accuracy of 97[Formula: see text], while the MLP model outperforms other classifiers with an accuracy of 74[Formula: see text]. Furthermore, we evaluate the performance of our proposed OntoeCoach ontology model by assessing reasoning and query execution time metrics. The results demonstrate that our approach effectively plans and generates recommendations on both datasets. The rule set of OntoeCoach can also be generalized to enhance interpretability.

Health to Eat: A Smart Plate with Food Recognition, Classification, and Weight Measurement for Type-2 Diabetic Mellitus Patients' Nutrition Control

The management of type 2 diabetes mellitus (T2DM) is generally not only focused on pharmacological therapy. Medical nutrition therapy is often forgotten by patients for several reasons, such as difficulty determining the right nutritional pattern for themselves, regulating their daily nutritional patterns, or even not heeding nutritional diet recommendations given by doctors. Management of nutritional therapy is one of the important efforts that can be made by diabetic patients to prevent an increase in the complexity of the disease. Setting a diet with proper nutrition will help patients manage a healthy diet. The development of Smart Plate Health to Eat is a technological innovation that helps patients and users know the type of food, weight, and nutrients contained in certain foods. This study involved 50 types of food with a total of 30,800 foods using the YOLOv5s algorithm, where the identification, measurement of weight, and nutrition of food were investigated using a Chenbo load cell weight sensor (1 kg), an HX711 weight weighing A/D module pressure sensor, and an IMX219-160 camera module (waveshare). The results of this study showed good identification accuracy in the analysis of four types of food: rice (58%), braised quail eggs in soy sauce (60%), spicy beef soup (62%), and dried radish (31%), with accuracy for weight and nutrition (100%).

A Digital Lifestyle Coach (E-Supporter 1.0) to Support People With Type 2 Diabetes: Participatory Development Study

BACKGROUND

A healthy lifestyle, including regular physical activity and a healthy diet, is becoming increasingly important in the treatment of chronic diseases. eHealth interventions that incorporate behavior change techniques (BCTs) and dynamic tailoring strategies could effectively support a healthy lifestyle. E-Supporter 1.0 is an eCoach designed to support physical activity and a healthy diet in people with type 2 diabetes (T2D).

OBJECTIVE

This paper aimed to describe the systematic development of E-Supporter 1.0.

METHODS

Our systematic design process consisted of 3 phases. The definition phase included the selection of the target group and formulation of intervention objectives, and the identification of behavioral determinants based on which BCTs were selected to apply in the intervention. In the development phase, intervention content was developed by specifying tailoring variables, intervention options, and decision rules. In the last phase, E-Supporter 1.0 integrated in the Diameter app was evaluated using a usability test in 9 people with T2D to assess intervention usage and acceptability.

RESULTS

The main intervention objectives were to stimulate light to moderate-vigorous physical activities or adherence to the Dutch dietary guidelines in people with T2D. The selection of behavioral determinants was informed by the health action process approach and theories explaining behavior maintenance. BCTs were included to address relevant behavioral determinants (eg, action control, self-efficacy, and coping planning). Development of the intervention resulted in 3 types of intervention options, consisting of motivational messages, behavioral feedback, and tailor-made supportive exercises. On the basis of IF-THEN rules, intervention options could be tailored to, among others, type of behavioral goal and (barriers to) goal achievement. Data on these variables could be collected using app data, activity tracker data, and daily ecological momentary assessments. Usability testing revealed that user experiences were predominantly positive, despite some problems in the fixed delivery of content.

CONCLUSIONS

The systematic development approach resulted in a theory-based and dynamically tailored eCoach. Future work should focus on expanding intervention content to other chronic diseases and lifestyle behaviors, enhancing the degree of tailoring and evaluating intervention effects on acceptability, use, and cost-effectiveness.

Nutrition-Related Mobile Application for Daily Dietary Self-Monitoring

Nutrition apps for mobile devices such as smartphones are becoming more widely available. They can help ease the arduous chore of documenting intake for nutritional assessment and self-monitoring. This allows people to control food intake, support their participation in physical activities, and promote a healthy lifestyle. However, there remains a lack of research regarding systematic analysis mapping studies in this area. The objective of this study is to identify dietary self-monitoring implementation strategies on a mobile application. This study analyzed 205 journals from the Scopus database using the descriptive-analytic method. The records used in this exploration study were those released between 2007 and 2021 that were collected based on the keywords “dietary self-monitoring,” or “nutrition application,” or “nutrition apps,” and “calorie application.” Data analysis was conducted using the VOSviewer and NVivo software analytical tools. The results show that research studies on dietary self-monitoring increased in 2017. Results also indicated that the country that contributed the most to this topic was China. The study on mobile applications for dietary self-monitoring revealed seven clusters of dominant themes: attitude to improved dietary behaviors, parameters for disease diagnosis, noncommunicable diseases, methods, nutrition algorithms, mobile health applications, and body mass index. This study also analyzed research trends by year. The current research trends are about dietary self-monitoring using a mobile application that can upgrade people's lifestyles, enable real-time meal recording and the convenience of automatically calculating the calorie content of foods consumed, and potentially improve the delivery of health behavior modification interventions to large groups of people. The researchers summarized the recent advances in dietary self-monitoring research to shed light on their research frontier, trends, and hot topics through bibliometric analysis and network visualization. These findings may provide valuable guidance for future research and perspectives in this rapidly developing field.

Design and Development of a Mobile-Based Self-Care Application for Patients with Type 2 Diabetes

INTRODUCTION

Mobile-based applications play a leading role in changes in life-style, improve medication adherence, and provide a unique opportunity to aid patients with type 2 diabetes mellitus (T2DM) elevate their healthcare level. Therefore, we aim to design and develop a mobile-based self-care application for patients with T2DM.

METHODS

The present study was an applied and developmental study to design and develop a mobile-based self-care application for people living with T2DM conducted in 2020. The design and development of the T2DM self-care application were done in 2 main phases of determining the key features and capabilities, and design and development of the T2DM self-care mobile app.

RESULTS

We identified the main model and a set of capabilities and features for the T2DM self-care application. By content analysis on 32 different applications and a previous study by the author, 18 features were extracted for the T2DM self-care mobile app. JAVA programming languages were used to design T2DM applications. Moreover, because of the cost-effectiveness, the Android operating system (AOS) was selected as a platform, and because of the widespread use of smartphones; these phones were chosen as the format of T2DM self-care application.

CONCLUSIONS

In this study, we design and develop a mobile-based self-care application for patients with type 2 diabetes that shows potential in solving the shortcomings of mobile apps for diabetes care. By utilizing the T2DM self-care mobile app we are able to deploy a self-care application with a wide range of functionality such as text messaging, blood glucose monitoring, insulin dose suggestions, educational messaging, metabolic management, pedometer counts, and reporting. Future studies are needed to develop self-care applications for a different type of diabetes with different functions of diabetes care.

Human Coaching Methodologies for Automatic Electronic Coaching (eCoaching) as Behavioral Interventions With Information and Communication Technology: Systematic Review

BACKGROUND

We systematically reviewed the literature on human coaching to identify different coaching processes as behavioral interventions and methods within those processes. We then reviewed how those identified coaching processes and the used methods can be utilized to improve an electronic coaching (eCoaching) process for the promotion of a healthy lifestyle with the support of information and communication technology (ICT).

OBJECTIVE

This study aimed to identify coaching and eCoaching processes as behavioral interventions and the methods behind these processes. Here, we mainly looked at processes (and corresponding models that describe coaching as certain processes) and the methods that were used within the different processes. Several methods will be part of multiple processes. Certain processes (or the corresponding models) will be applicable for both human coaching and eCoaching.

METHODS

We performed a systematic literature review to search the scientific databases EBSCOhost, Scopus, ACM, Nature, SpringerLink, IEEE Xplore, MDPI, Google Scholar, and PubMed for publications that included personal coaching (from 2000 to 2019) and persuasive eCoaching as behavioral interventions for a healthy lifestyle (from 2014 to 2019). The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework was used for the evidence-based systematic review and meta-analysis.

RESULTS

The systematic search resulted in 79 publications, including 72 papers and seven books. Of these, 53 were related to behavioral interventions by eCoaching and the remaining 26 were related to human coaching. The most utilized persuasive eCoaching methods were personalization (n=19), interaction and cocreation (n=17), technology adoption for behavior change (n= 17), goal setting and evaluation (n=16), persuasion (n=15), automation (n=14), and lifestyle change (n=14). The most relevant methods for human coaching were behavior (n=23), methodology (n=10), psychology (n=9), and mentoring (n=6). Here, "n" signifies the total number of articles where the respective method was identified. In this study, we focused on different coaching methods to understand the psychology, behavioral science, coaching philosophy, and essential coaching processes for effective coaching. We have discussed how we can integrate the obtained knowledge into the eCoaching process for healthy lifestyle management using ICT. We identified that knowledge, coaching skills, observation, interaction, ethics, trust, efficacy study, coaching experience, pragmatism, intervention, goal setting, and evaluation of coaching processes are relevant for eCoaching.

CONCLUSIONS

This systematic literature review selected processes, associated methods, strengths, and limitations for behavioral interventions from established coaching models. The identified methods of coaching point toward integrating human psychology in eCoaching to develop effective intervention plans for healthy lifestyle management and overcome the existing limitations of human coaching.