Abstract
The global burden of controlling and managing diabetes mellitus (DM) is a significant challenge. Despite the advancements in conventional DM therapy, there remain hurdles to overcome, such as enhancing medication adherence and improving patient prognosis. Digital therapeutics (DTx), an innovative digital application, has been proposed to augment the traditional disease management workflow, particularly in managing chronic diseases like DM. Several studies have explored DTx, yielding promising results. However, certain concerns about this innovation persist. In this review, we aim to encapsulate the potential of DTx and its applications in DM management, thereby providing a comprehensive overview of this technique for public health policymakers.
Introduction
Chronic diseases, including DM, account for over 70% of China’s total health expenditure, placing a significant financial burden on both the families of patients and society at large (1, 2). Based on the International Diabetes Federation (IDF), the global prevalence of DM in individuals aged 20–79 years old was estimated to be 536.6 million people in 2021. Without effective interventions, this figure is projected to escalate to 783.2 million by 2045. Although type 2 DM accounts for approximately 90% of these cases, the incidence of type 1 DM is also on an upward trend (3, 4). The management of DM and its associated complications necessitates long-term medical services, which invariably incur substantial costs. It was estimated that global diabetes-related health expenditures were at 966 billion USD in 2021 and would increase to 1054 billion USD by 2045 (4). In the United States, the proportion and total amount of out-of-pocket expenses for chronic diseases among the elderly population are escalating annually, particularly among those in economically disadvantaged situations (5). In China, the total costs of diabetes as a percentage of GDP will increase from 1.58% to 1.69% during 2020–2030, indicating a more rapid growth in the economic burden of diabetes than China's economic growth (6). A survey by the China Foundation for the promotion of health shows that the cost of a single hospitalization for DM often exceeds half of the average annual income for urban residents and is 1.5 times the average annual income for rural residents in China. Consequently, families dealing with chronic diseases face a higher risk of catastrophic medical expenditure compared to the general population.
The current strategies for DM treatment and management warrant careful scrutiny. Conventional DM therapies face challenges, including enhancing medication adherence and improving patient prognosis, which necessitate resolution. Poor adherence to medication for primary chronic diseases can lead to increased treatment costs and suboptimal patient outcomes. Consequently, nations worldwide are seeking strategies to alleviate the burden on their healthcare systems and embrace innovative treatment modalities (7). The advent of information and communication technology (ICT), encompassing digital health, digital medicine, and digital therapeutics (DTx), has revolutionized the management of chronic diseases, particularly DM. Despite the availability of a plethora of solutions, including telephonic consultations, short message service (SMS), websites, mobile health applications, remote monitoring devices, and advanced artificial intelligence (AI) systems, these have not been universally adopted as standard treatments. This limited adoption can be attributed to the absence of supportive policies and legislation, unsustainable reimbursement models, inefficient business strategies, and concerns regarding health data security and privacy (7, 8). Mobile health (mHealth) programs have demonstrated safety and efficacy in reducing body weight and HbA1C levels in prediabetic adults, suggesting a potential for reducing DM risk over the long term (9). mHealth technologies have rapidly gained popularity due to their wide appeal among individuals, accessibility, and ability to reach broad populations at low cost (10).
An International Panel on Diabetes Digital Technologies has recommended the development of novel DM care delivery methods to broaden access to care, alleviate the burden on individuals living with diabetes, enhance efficiency, and reduce unsustainable financial liabilities for health systems and payers. DTx, as defined by the Digital Therapeutics Alliance, involves the delivery of evidence-based therapeutic interventions via high-quality software programs to patients (https://dtxalliance.org/). As illustrated in Fig. 1, these digital interventions prescribed by doctors aim to treat, manage, or prevent medical conditions or diseases by efficiently analyzing the different transmitted bundles of patient’s data in the cloud to achieve personalized and AI-enhanced solutions to optimize patient care and health outcomes. The Digital Therapeutics Alliance classifies DTx into four categories: addressing medical conditions, managing or preventing medical disorders, optimizing medication use, and treating medical disorders or diseases (11).
Currently, DTx products, either on the market or in development, are available for various chronic conditions, including DM, cancer management, anxiety, depression, and insomnia (https://dtxalliance.org/; 12). An increasing number of startups are launching DTx products and securing substantial investments (13). From 2011 to 2019, venture capital firms invested nearly $1.3 billion in related startups, with DM being one of the three major disease targets. Over 89 transactions have been made, with a total investment of approximately $1294.8 million and an average investment of $14.5 million per startup. Traditional pharmaceutical and high-tech companies have also joined the race to develop digital health products (14). DTx has been shown to significantly improve glycemic control and provide diabetes care independently or combined with medications, devices, or other therapies, which could optimize patient care and health outcomes and substantially reduce treatment costs (15). Therefore, this review aims to explore the potential of DTx and its applications in DM management, with the goal of providing a comprehensive overview of this technology for public health policymakers.
Methods
To ensure the comprehensiveness and representativeness of the literature, we conducted searches using the primary databases PubMed, Scopus, Web of Science, IEEE Xplore, CNKI, and Google Scholar. The search employed keywords and free terms such as ‘Digital Therapeutics’, ‘DTx’, ‘Diabetes Mellitus’, ‘Diabetes’, “DM’, ‘Diabetes Management’, ‘Digital Health’, ‘Digital Technology’, ‘Type 1 Diabetes Mellitus’, ‘Type 1 ‘Diabetes’, ‘T1DM’, ‘Type 2 Diabetes Mellitus’, ‘Type 2 Diabetes’, and ‘T2DM’. We included studies published between January 1998 and April 2024 that explored the use of DTx in managing diabetes, covering randomized controlled trials, cohort studies, case-control studies, and high-quality review articles. Studies were selected based on the relevance of their titles and abstracts, followed by full-text reviews. Studies published before January 1998, and those not explicitly addressing DTx in diabetes management were excluded. To ensure the quality and reliability of the included studies, two independent reviewers conducted the selection process, resolving discrepancies through discussion and consensus. Themes and headings in the main text were generated through thematic analysis, identifying and categorizing recurring topics from the included studies to provide a structured overview of the findings.
Effectiveness and efficiency of DTx in DM management
Among the various treatment modalities, DTx has more effectiveness and efficiency in two aspects. As shown in Fig. 2, one of the critical elements in improving outcomes for patients with DM and other chronic health conditions is enhancing their ability to manage their lives and chronic conditions (16). Diabetes self-management education (DSME) and behavioral support integrated into DTx can mitigate the risk of diabetes-related complications and improve glycemic control (17). One Drop, a DTx solution, incorporates an evidence-based mobile app with Bluetooth-connected glucose meters and in-app guidance from certified diabetes educators. A 3-month study revealed that participants who used both One Drop and an activity tracker had significantly lower HbA1C levels than those who used only One Drop during the same period (18). Furthermore, One Drop has been shown to significantly reduce HbA1c in T2DM patients and may enhance the likelihood of individuals with DM receiving effective DSME and support (19). In another instance, Pack Health conducted a retrospective cohort analysis based on real-world data. The analysis demonstrated that T2DM participants who completed a multichannel digital health guidance program and provided a complete HbA1c dataset improved not only their physical and mental health but also their blood sugar control and body weight. The cohort analysis revealed that individuals with the highest blood glucose risk (HbA1C > 9% at enrollment) achieved the most significant changes in all clinical and patient-reported outcomes, except for weight loss (20). The Artificial Pancreas (AP) is a medical device that represents a significant advancement in diabetes management. It integrates continuous subcutaneous insulin infusion pumps and continuous glucose monitoring (CGM) to enhance glycemic control. These components work together to improve glycated HbA1c levels and alleviate diabetes-related distress. The AP provides a sophisticated control algorithm to automatically regulate blood glucose levels in ambulatory patients with type 1 DM, thereby effectively reducing the management burden on both physicians and patients (21, 22). Additionally, a recent outpatient study of a wearable AP, utilizing a meal-informed model predictive control system, has demonstrated its ability to reduce postprandial glycemic excursions (23).
Effectiveness and efficiency of DTx in DM management.
Citation: Endocrine Connections 13, 9; 10.1530/EC-24-0219
The global COVID-19 pandemic underscored the importance of enabling telemedicine and digital diabetes technologies during periods of isolation (16). Therefore, it is appropriate and efficient to leverage DTx to address the pressing issue of insufficient and untimely delivery of offline medical services. Similarly, in people-centered primary health care, most general practitioners (GPs) deliver quality healthcare to the majority of the population within their local communities (24). DTx can provide GPs with a means to enhance their diagnostic and treatment capabilities, addressing issues related to difficult and untimely access to medical treatment (25). Meanwhile, DTx also incorporates support from cutting-edge technologies such as machine learning, deep learning (DL) algorithms, and intelligent AI. These technologies can effectively enhance the efficiency of programs and address fundamental clinical problems. According to Lancet, AI models possess diagnostic capabilities comparable to professional doctors, suggesting that AI in DTx could provide effective assistance to medical practitioners (26). For instance, a multi-scale dynamic fusion module combined with graph convolution operations based on deep learning techniques is beneficial for the correct diabetic retinopathy grading and is more accurate and diverse for the extraction of lesion information (27). AI can augment healthcare providers' efficiency, allowing them to dedicate more time to unique human skills, such as relationship building, empathy, and human judgment to guide and advise (28).
Regulations and oversight for DTx applications
Like other drugs and medical devices, DTx requires clinical trials to validate their efficacy and safety and formulate relevant regulations. Currently, many countries are promoting the commercialization of DTx through collaborative efforts involving regulatory authorities, pharmaceutical firms, and medical experts across various sectors. Initially, DTx received approval and entered the market primarily for the treatment of chronic diseases, particularly diabetes (29). During the COVID-19 pandemic, the U.S. Government demonstrated a high level of interest in digital healthcare. Major companies launched their core products to apply for FDA approval. A government agency head expressed a desire to see more evidence-based procedures (30). An increasing number of DTx products have received innovative regulatory approval systems. In response, the FDA launched a digital health software Pre-certification (PreCert) pilot program designed to expedite regulatory scrutiny of companies demonstrating quality and organizational excellence in software development. Pre-certified developers can bring less risky devices to market without additional FDA review or via a more streamlined premarket review (31). Similarly, the corresponding departments in the UK and Germany began to launch a treatment evaluation program and a new digital bill, respectively, to promote the use of telemedicine and ensure better medical data for research purposes (32). In Asia, the Shukang APP received approval from the National Medical Products Administration (NMPA) of China and was directly prescribed by doctors to patients, marking the beginning of DTx in China. Japan and Korea have issued relevant documents since 2017 to guide and standardize the market (33, 34, 35, 36).
The current body of evidence regarding the safety and efficacy of mobile health applications, particularly for diabetes, remains limited. In response, the European Association for the Study of Diabetes (EASD) and the American Diabetes Association (ADA) have collaborated to conduct a joint review of the existing landscape of digital health technologies for diabetes, as well as the practices of regulatory authorities and organizations. The Diabetes Technology Working Group, established by the ADA and the EASD, seeks to complement previously published reviews, position statements, and guidelines related to digital health applications (37). Currently, the USA is leading the way in DTx, and the FDA's certification policy serves as a guide for the rest of the world. Here is an explanation of the US government’s policy: In response to the rapid growth of digital health applications, the FDA has sought to distinguish between applications that require oversight and those that do not. In 2015, it issued a guidance document (updated in 2019) applicable to mobile medical applications as defined in Section 201 of the Federal Food, Drug, and Cosmetics Act (FDCA) (38). This definition includes applications and software designed to be accessories to regulated medical devices. However, the guidelines state that the FDA intends to exercise ‘enforcement discretion’ over those deemed to pose a low risk to users (e.g. apps that offer diabetics incentives to achieve their health goals or provide tools to track their health information). Therefore, using this ‘risk-based’ approach, mobile applications that calculate insulin doses fall under regulation, while applications that simply organize and/or provide health or nutritional information are not overseen. The FDA lists approved applications in its 510(k) and Premarket Approval (PMA) databases, and its registration and listing database (39). Welldoc's diabetes digital health platform, BlueStar, has received its 11th 510(k) clearance from the FDA. Notably, Welldoc was the pioneer recipient of clearance for a CGM-informed bolus calculator specifically tailored for adults who manage their diabetes with multiple daily injections of insulin. This recent clearance empowers BlueStar to deliver bolus insulin dose recommendations based on the most recent glucose readings and the rate of change from a compatible CGM device (40).
It is crucial to note that FDA approval can be sought and obtained for safety or indications for digital treatments or devices, or both. For instance, a mobile medical application (MMA) used to capture, store, and catalog measurements from a Class II medical device (such as heart rate, lung murmurs, blood sugar, or pulse oxygen saturation), or an MMA sensor device that can be used to control its function, is considered a class II medical device and needs to be submitted for FDA approval. However, it is clear that merely capturing, storing, and recording measurements from sensor devices does not necessarily have any therapeutic effect on disease. FDA approval of the device does not imply that the device is therapeutic. Conversely, an MMA can have therapeutic effects and can be used to treat medical or psychological conditions without seeking or obtaining FDA approval. This is particularly true for DTx, which fall within the FDA’s enforcement discretion (38). In conclusion, the current approval policy in the USA may have some nuances, but this can be attributed to the diversity of application functions, making it challenging to find a completely adaptive method to determine its rationality. All these issues require the consideration and exploration of all sectors of society, gradual improvement of relevant laws and regulations, and the introduction of high-quality DTx into the consumer market.
Safety and privacy concerns for DTx in application
While modern information technology has proven effective for patient prognosis, issues such as evidence-based DTx and the relationship between electronic information technology-based treatment methods and practical application by doctors have impeded the realization of personalized precision therapy. Poor integration of the entire health information technology (HIT) workflow can have serious negative effects. Integrating these tools and technologies into medical devices that clinicians use for disease management and health behavior change is within DTx’s capabilities (41). Studies have shown that access to medical records benefits both patients and clinicians. Patients can actively participate in the management of their health care by ensuring that clinicians have an updated and accurate overview of their medical records (42, 43). European legislation protects the rights of patients to access clinical data when needed and controls who can view and alter this information (44). However, there are still concerns about the safety risks of sharing electronic health records (EHR) and registry data (45).
Diabetes apps primarily allow individuals with DM to monitor their data and discuss it with health professionals. Users may believe their health data stored in apps is private, but a 2014 study of Android smartphone diabetes apps suggests otherwise. The study found that most apps do not widely use data transmission encryption and often share information with third-party servers, or even lack specific privacy protocols, leading to health issues such as reduced data security and serious threats to user data privacy and security (46, 47, 48). These issues can only be addressed if mhealth providers improve the way they communicate and store data.
Patients with DM have high safety requirements when using wireless diabetes devices such as glucose monitors, continuous glucose monitors, and insulin pumps to detect blood glucose levels and insulin doses. Medical devices are susceptible to security breaches. There have been reports of users remotely accessing data from insulin pumps and controlling their functions without their knowledge. Johnson & Johnson has issued a warning about a safety flaw in its insulin pumps, a situation that can be life-threatening (49). Even with multiple layers of validation, errors may occur infrequently throughout the program, but some errors are clinically significant. In the absence of regulation, the responsibility for any adverse consequences of using these apps falls on the clinician, whose choice and trust in the app can impact prescription outcomes (50). Will they accept DTx as their primary prescription, or will they be skeptical about it as an adjunct? This also reduces trust among healthcare providers to use them and raises questions about their safety and authenticity. From a patient's perspective, the medical selection process is riddled with DTx of varying quality, making it more difficult to find the right one. There are still risks in terms of application security and data privacy, increasing users' distrust of the application and the possibility of putting it at risk, and seriously impacting the overall rollout and future vision of DTx.
The data stored in health data applications should be fully encrypted to prevent serious malicious attacks and adverse consequences. The use of blockchain technology is one choice for data encryption (51). The application scenarios of blockchain technology are further embodied in helping governments and citizens prevent and control disease and raise health awareness by collecting glucose via a smartphone and sending it to a cloud platform. The blockchain digital encryption system, combined with decentralized storage, private key signatures, tamper-proofing, and other technologies, improves information reliability and carries out DM diagnosis, monitoring, and research, and adopts public health movements (52). Currently, some scholars have studied and evaluated existing app review methods (53). Stoyanov et al. extensively studied relevant literature and have summarized and refined a tool called MARS to classify and evaluate the quality of mobile health applications (54). It can also be used to provide checklists for the design and development of new high-quality health applications, which are carried out from the aspects of engagement, functionality, aesthetics, information quality, as well as the subjective quality of applications. The Diabetes Technology Association’s ‘Safety Standard for Diabetes Equipment’ is also a reference standard (55). Lee et al. studied that when developing diabetes-related apps, a comprehensive usability assessment must be made by using definitions such as the ISO9241-11 usability definition or the mobile application rating scale (56).
Ethical and social issues during the adoption of DTx
The largest group of diabetes-related spending is among those aged 60–69 (57). This age group’s limited familiarity with technology can hinder the widespread use of DTx. Younger patients generally have a better command of smartphones than older individuals. Therefore, applications specifically designed for older individuals suffering from diabetes must consider the technological capabilities of this demographic. Aside from age, there is a negative correlation between individual socioeconomic status and the prevalence, incidence, and mortality of DM. In other words, people with better personal financial situations are more likely to have first access to the latest diagnostic technology to improve their outcomes, and there is a correlation between personal education level and access to the latest technology. Therefore, the emergence of new technology may create a digital divide at the social level, leading to unequal health outcomes (58, 59). In addition, currently, available DM management apps may not be available in languages other than the single language in which they were designed, or may not be accessible to specific populations with certain physical or mental disorders (e.g. color blindness, blindness, and hearing impairment). Furthermore, people from remote areas and areas of extreme socio-economic poverty may not have access to smartphone technology or the cost of acquiring and activating a smartphone.
Cost efficiency of DTx in DM management
DTx has shown significant promise in improving cost efficiency and is appealing from a health economics perspective, particularly in the management of diabetes. Compared to conventional therapeutics, DTx offers substantial advantages, including reduced development time and cost, while significantly improving clinical outcomes. For instance, DTx can cut development costs by up to 50%, accelerating the time-to-market for new diabetes treatments (60). The prescription digital therapeutic BlueStar for type 2 diabetes demonstrated substantial cost savings by reducing hospital admissions by 29% and emergency visits by 19%. Additionally, it improved glycemic control, resulting in a positive return on investment (ROI) (61).
Moreover, telemedicine programs and mHealth applications for diabetes management have proven effective in reducing healthcare costs and improving patient outcomes. A study on a mobile app for diabetes management revealed that it reduced HbA1c levels by an average of 0.8% and saved approximately $1500 per patient annually due to fewer complications and hospital visits (62). Research published by Liebertpub demonstrated that personalized digital interventions reduced hospital readmission rates for diabetes-related complications by 25%, resulting in significant cost savings for healthcare systems (63). Additionally, appropriate reimbursement policies for DTx are crucial for enhancing patient health and ensuring the financial stability of the National Health Insurance (NHI) fund. The research indicated that such policies could lead to a 20% increase in patient uptake of these solutions, further reducing long-term healthcare costs for diabetes management (64).
Clinical approval and quality
It is important to note that most existing apps on the market generally lack quality clinical trial evidence. They often rely on evidence from small, uncontrolled, and short-duration experimental studies aimed at testing the feasibility of their technology or programs rather than their long-term effectiveness or other key issues. For diabetes, an ideal target disease for digital health technology due to its high incidence and chronic nature, its prognosis is primarily improved by data detection and lifestyle changes. The number of relevant apps has surged by 330%, but there are still no effective studies to prove their long-term prognosis improvement effect for diabetic patients (65). Some studies have revealed that even FDA-certified programs like BlueStar might show poor or no effectiveness due to factors such as low utilization rate, insufficient persistence time, and decreased patience of users during clinical trials (66). The main reason for the lack of relevant clinical research evidence is the continuous iterations and updates of digital technology-based health apps by developers. This is quite different from traditional drugs that can be developed within a certain period without the need for frequent changes and verifications. Furthermore, even a small, short-term RCT with a small cohort requires at least 1-2 years from experimental design to final publication, which is challenging to achieve in today's rapidly evolving Internet era (67, 68). From the perspective of the RCT experimental method, the core of its difference from other types of clinical research lies in the use of placebo and the implementation of a blind method by operators, which is difficult to practically perform in the process of program-based digital technology verification. Even if RCTs have been published, these problems are worth scrutinizing.
Recently, health authorities have advocated for the use of real-world evidence (RWE) studies and pragmatic RCTs to better reflect clinical practice (69). Consequently, it is essential to develop and apply innovative statistical methods, including causal inference, big data analysis, and machine learning (ML) algorithms, to analyze DTx data in diabetes management and inform decision-making. Causal inference techniques such as propensity score matching (PSM), inverse probability weighting (IPW), and instrumental variable (IV) analysis are crucial in controlling for confounding factors and simulating the conditions of an RCT. In evaluating a DTx intervention like a mobile app for diabetes management, PSM can match app users with non-users based on factors, such as age, gender, BMI, diabetes duration, and baseline HbA1c levels, thereby minimizing bias in the estimated treatment effect (70). Tools like Hadoop and Spark facilitate the efficient processing of large datasets generated by digital health technologies. In diabetes management, big data analytics can integrate diverse data sources, including electronic health records (EHRs), wearable devices, and mobile apps, to offer a comprehensive view of a patient’s health. This integration allows for the identification of trends in blood sugar levels, medication adherence, and lifestyle factors that may not be evident when analyzing data from a single source (71, 72). Additionally, ML can be applied to analyze data from CGMs and insulin pumps to predict glucose trends and optimize insulin dosing. Algorithms can learn from historical glucose data to provide personalized insulin recommendations, thereby enhancing glycemic control and reducing hypoglycemic episodes (73, 74). It is well known that most apps on the market today are designed by startups, which lack the financial strength of traditional medicine companies to perform high-quality clinical trials. Therefore, it is challenging to find long-term and high-quality clinical evidence. Only years of iteration and accumulation can promote the birth of high-quality DTx products.
If DTx is to be considered a routine prescription, its safety and effectiveness must be measured by the standards of traditional drugs, that is, the improvement of prognosis and all possible adverse results. According to the clinical pharmacology of traditional drugs, a certain dose–exposure–response relationship may exist in DTx, which should also be verified in clinical trials. For instance, to achieve the purpose of improving prognosis, a certain dose (duration) of DTx is used regularly, and the long-term focus on the electronic screen may impair patients’ vision, hearing, and central nervous system (CNS) or influence their musculoskeletal system to a certain extent. Therefore, the optimal exposure dose–response relationship, that is, the optimal duration of efficient prognosis improvement under the condition of satisfying the basic safety of patients, should be found under DTx technology. However, due to the unique nature of digital technology verification experiments, a new experimental endpoint needs to be set to measure and evaluate the experimental results of new technology before studying the exposure dose–response of DTx (75). At this point, by providing a new experimental endpoint or an innovative digital marker for clinical research of new technology, digital technology can serve as an innovation or supplement to traditional clinical trials. Also, the application of digital endpoints in experimental data collection and analysis will be superior to traditional methods. Then, considering the DM DTx experiment, should one or several new digital endpoints be set to supplement or replace traditional experimental endpoints like GHB and blood glucose levels, and highlight the effectiveness or safety of DTx technology? (76, 77) In the absence of relevant clinical trial standards, the rationality and safety of digital endpoints as a new standard have not been verified. Nowadays, few RCT experiments have chosen the digital endpoint as the endpoint standard. There are still big difficulties in the brand-new clinical trials for its uncertain exposure or dose (frequency and duration of program use), response endpoints (new digital endpoint), and dose–exposure–response relationship in DTx. The characteristics of DTx technology require continuous updates and upgrades by the developers. Also, there are practical problems in the application of the blind method in traditional clinical RCT and immoral or impossible situations that may exist in the control group using placebo. Even issues on the data privacy of participants, ethics, and security in the experiment need to be solved urgently. FDA provides a new perspective on the unique consideration of clinical trials of medical devices, which may be of reference significance in the above situations (78). It is well-known that the recruitment of volunteers for clinical trials, especially RCTs with placebo, has always been a core issue. The serious shortage of participants will greatly delay the normal progress of the experiment or reduce the quality of the experimental results, and even terminate the experiment. A MyHeart Counts study proves that digital technology (Apple Research Kit) is a potentially effective way to recruit volunteers, but the quality of recruitment is not clear yet. Under the background of increasing time, cost, and technical requirements of clinical trials, the addition of digital health technology is inevitably a choice (79).
Conclusion
Nowadays, DTx has emerged as a new option for DM management. The potential of DTx in DM management is promising, provided that more clinical trials are conducted and issues related to regulatory policy, safety and privacy concerns, and ethical and social issues are addressed.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the study reported.
Funding
This work was supported by a grant from the Postgraduate Research and Practice Innovation Program of Jiangsu Province (nos KYCX22_3363 and KYCX21_3104) and the Jiangsu Students' Platform for Innovation and Entrepreneurship Training Program (nos 202310304221E and 202310304124Y).
Author contribution statement
SL, HW, and JCT contributed to the conception and design of the study and completed the drafting, critical review, and final approval. JT and YC provided the logistical support. All authors contributed to manuscript revision and agreed to be accountable for all aspects of the work.
Acknowledgements
The project is granted by the Engineering Research Center of Integration and Application of Digital Learning Technology, Ministry of Education (1311016). The figures are created with https://www.biorender.com/.
References
- 1↑
Teng H, Cao Z, Liu J, Liu P, Hua W, Yang Y, & Xiong L. Health status and burden of health care costs among urban elderly in China. Asia-Pacific Journal of Public Health 2015 27(Supplement) 61S–8S. (https://doi.org/10.1177/1010539515571579)
- 2↑
Zhou M, Sun X, & Huang L. Chronic disease and medical spending of Chinese elderly in rural region. International Journal for Quality in Health Care 2021 33. (https://doi.org/10.1093/intqhc/mzaa142)
- 3↑
Patterson CC, Karuranga S, Salpea P, Saeedi P, Dahlquist G, Soltesz G, & Ogle GD. Worldwide estimates of incidence, prevalence and mortality of type 1 diabetes in children and adolescents: results from the International Diabetes Federation Diabetes Atlas. Diabetes Research and Clinical Practice, 2019 157 107842. (https://doi.org/10.1016/j.diabres.2019.107842)
- 4↑
Sun H, Saeedi P, Karuranga S, Pinkepank M, Ogurtsova K, Duncan BB, Stein C, Basit A, Chan JCN, Mbanya JC, et al. IDF Diabetes Atlas: global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045. Diabetes Research and Clinical Practice 2022 183 109119. (https://doi.org/10.1016/j.diabres.2021.109119)
- 5↑
Schoenberg NE, Kim H, Edwards W, & Fleming ST. Burden of common multiple-morbidity constellations on out-of-pocket medical expenditures among older adults. Gerontologist 2007 47 423–437. (https://doi.org/10.1093/geront/47.4.423)
- 6↑
Liu J, Liu M, Chai Z, Li C, Wang Y, Shen M, Zhuang G, & Zhang L. Projected rapid growth in diabetes disease burden and economic burden in China: a spatio-temporal study from 2020 to 2030. Lancet Regional Health. Western Pacific 2023 33 100700. (https://doi.org/10.1016/j.lanwpc.2023.100700)
- 7↑
Lemstra M, Nwankwo C, Bird Y, & Moraros J. Primary nonadherence to chronic disease medications: a meta-analysis. Patient Preference and Adherence 2018 12 721–731. (https://doi.org/10.2147/PPA.S161151)
- 8↑
vbdata.cn. VCBeat Research. White Paper on Digital Therapeutics in China Chongqing, China: vbdata.cn, 2021. (available at: https://cdn.vcbeat.top/upload/report/47/96/22/21/607fea2e49222.pdf)
- 9↑
Everett E, Kane B, Yoo A, Dobs A, & Mathioudakis N. A novel approach for fully automated, personalized health coaching for adults with prediabetes: pilot clinical trial. Journal of Medical Internet Research 2018 20 e72. (https://doi.org/10.2196/jmir.9723)
- 10↑
Miki T, Nohara M, & Nomura K. Effectiveness of mHealth interventions to promote physical activity and reduce sedentary behaviours on work-related productivity and performance: a systematic review protocol. BMJ Open 2024 14 e080240. (https://doi.org/10.1136/bmjopen-2023-080240)
- 11↑
Digital Therapeutics Alliance. Combining Technology and Evidence-Based Medicine to Transform Personalized Patient Care. Arlington, VA USA: Digital Therapeutics Alliance, 2018. (available at: https://www.dtxalliance.org/wp-content/uploads/2018/09/DTA-Report_DTx-Industry-Foundations.pdf)
- 12↑
Hong JS, Wasden C, & Han DH. Introduction of digital therapeutics. Computer Methods and Programs in Biomedicine 2021 209 106319. (https://doi.org/10.1016/j.cmpb.2021.106319)
- 13↑
American Medical Association. Medical innovation and digital snake oil: AMA CEO speaks out. Chicago, IL, USA: American Medical Association, 2016. (available at: https://www.ama-assn.org/practice-management/digital/medical-innovation-and-digital-snake-oil-ama-ceo-speaks-out)
- 14↑
Klonoff DC, Evans B, Zweig M, Day S, & Kerr D. Is digital health for diabetes in an investment bubble? Journal of Diabetes Science and Technology 2020 14 165–169. (https://doi.org/10.1177/1932296819867742)
- 15↑
Joshi S, Verma R, Lathia T, Selvan C, Tanna S, Saraf A, Tiwaskar M, Modi A, Kalra S, K V, et al. Fitterfly diabetes CGM digital therapeutics program for glycemic control and weight management in people with type 2 diabetes mellitus: real-world effectiveness evaluation. JMIR Diabetes 2023 8 e43292. (https://doi.org/10.2196/43292)
- 16↑
Kaufman N. Digital therapeutics: leading the way to improved outcomes for people with diabetes. Diabetes Spectrum 2019 32 301–303. (https://doi.org/10.2337/ds19-0012)
- 17↑
Pal K, Dack C, Ross J, Michie S, May C, Stevenson F, Farmer A, Yardley L, Barnard M, & Murray E. Digital health interventions for adults with type 2 diabetes: qualitative study of patient perspectives on diabetes self-management education and support. Journal of Medical Internet Research 2018 20 e40. (https://doi.org/10.2196/jmir.8439)
- 18↑
Osborn CY, Hirsch A, Sears LE, Heyman M, Raymond J, Huddleston B, & Dachis J. One drop app with an activity tracker for adults with type 1 diabetes: randomized controlled trial. JMIR mHealth and uHealth 2020 8 e16745. (https://doi.org/10.2196/16745)
- 19↑
Kumar S, Moseson H, Uppal J, & Juusola JL. A diabetes mobile app with in-app coaching from a certified diabetes educator reduces A1C for individuals with type 2 diabetes. Diabetes Educator 2018 44 226–236. (https://doi.org/10.1177/0145721718765650)
- 20↑
Martin M, Patterson J, Allison M, O'Connor BB, & Patel D. The influence of baseline hemoglobin A1c on digital health coaching outcomes in adults with type 2 diabetes: real-world retrospective cohort study. JMIR Diabetes 2021 6 e24981. (https://doi.org/10.2196/24981)
- 21↑
Jones RW, Gianni F, Despotou G, & Katzis K. The artificial pancreas: reducing safety risk via intra-peritoneal insulin delivery. Studies in Health Technology and Informatics 2017 238 56–59. (https://doi.org/10.3233/978-1-61499-781-8-56)
- 22↑
Nwokolo M, & Hovorka R. The artificial pancreas and type 1 diabetes. Journal of Clinical Endocrinology and Metabolism 2023 108 1614–1623. (https://doi.org/10.1210/clinem/dgad068)
- 23↑
Del Favero S, Bruttomesso D, Di Palma F, Lanzola G, Visentin R, Filippi A, Scotton R, Toffanin C, Messori M, Scarpellini S, et al. First use of model predictive control in outpatient wearable artificial pancreas. Diabetes Care 2014 37 1212–1215. (https://doi.org/10.2337/dc13-1631)
- 24↑
Khatri RB, Wolka E, Nigatu F, Zewdie A, Erku D, Endalamaw A, & Assefa Y. People-centred primary health care: a scoping review. BMC Primary Care 24 236. (https://doi.org/10.1186/s12875-023-02194-3)
- 25↑
Luo Y, Wu H, Liao X, Zhao T, Cui N, Li A, Sun X, Zhang P, Huang Y, Zhang X, et al. A guideline-based decision tree achieves better glucose control with less hypoglycemia at 3 months in Chinese diabetic patients. Diabetes Therapy 2021 12 1887–1899. (https://doi.org/10.1007/s13300-021-01075-1)
- 26↑
Liu X, Faes L, Kale AU, Wagner SK, Fu DJ, Bruynseels A, Mahendiran T, Moraes G, Shamdas M, Kern C, et al. A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis. Lancet. Digital Health 2019 1 e271–e297. (https://doi.org/10.1016/S2589-7500(1930123-2)
- 27↑
Wang Y, Wang L, Guo Z, Song S, & Li Y. A graph convolutional network with dynamic weight fusion of multi-scale local features for diabetic retinopathy grading. Scientific Reports 2024 14 5791. (https://doi.org/10.1038/s41598-024-56389-4)
- 28↑
Fogel AL, & Kvedar JC. Artificial intelligence powers digital medicine. npj Digital Medicine 2018 1 5. (https://doi.org/10.1038/s41746-017-0012-2)
- 29↑
Wang C, Lee C, & Shin H. Digital therapeutics from bench to bedside. npj Digital Medicine 2023 6 38. (https://doi.org/10.1038/s41746-023-00777-z)
- 30↑
Morning Consult. COVID-19 accelerated use of digital therapeutics, but coverage issues and regulatory questions could slow their momentum. Washington DC, USA: Morning Consult, 2021. (available at: https://pro.morningconsult.com/articles/digital-therapeutics-use-coverage-polling)
- 31↑
U.S. Food and Drug Administration. FDA selects participants for new digital health software precertification pilot program. Silver Spring, MD, USA: FDA. (available at: https://www.fda.gov/news-events/press-announcements/fda-selects-participants-new-digital-health-software-precertification-pilot-program)
- 32↑
AOK-Bundesverband. Gesetz für eine bessere Versorgung durch Digitalisierung und Innovation/Digitale-Versorgung-Gesetz (DVG). Berlin, Germany: AOK-Bundesverband, 2020. (available at: https://www.aok.de/pp/reform/dvg/)
- 33↑
Liu W, Shi X, Lu Z, Wang L, Zhang K, & Zhang X. Review and approval of medical devices in China: changes and reform. Journal of Biomedical Materials Research. Part B, Applied Biomaterials 2018 106 2093–2100. (https://doi.org/10.1002/jbm.b.34031)
- 34↑
Kim Y. Korea publishes guidelines for digital therapeutics. In Korea Biomedical Review. Seoul, Korea, Korean Doctors’ Weekly Co., Ltd,. 2020. (available at: https://www.koreabiomed.com/news/articleView.html?idxno=9072)
- 35↑
Ministry of Health, Labour and Welfare. Release of Clinical Trial Guidance to Facilitate the Speedy and Accurate Approval and Development of Medical Devices. Tokyo, Japan: Pharmaceuticals and Medical Devices Agency, 2017. (available at: https://www.pmda.go.jp/files/000237132.pdf)
- 36↑
Pharmaceuticals and Medical Devices Agency. Notifications and Administrative Notices. Tokyo, Japan: Pharmaceuticals and Medical Devices Agency. (available at: https://www.pmda.go.jp/english/review-services/regulatory-info/0003.html).
- 37↑
Fleming GA, Petrie JR, Bergenstal RM, Holl RW, Peters AL, & Heinemann L. Diabetes digital app technology: benefits, challenges, and recommendations. A consensus report by the European Association for the Study of Diabetes (EASD) and the American Diabetes Association (ADA) diabetes technology working group. Diabetes Care 2020 43 250–260. (https://doi.org/10.2337/dci19-0062)
- 38↑
U.S. Food and Drug Administration. Policy for device software functions and mobile medical applications. Silver Spring, MD, USA: FDA.(available at: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/policy-device-software-functions-and-mobile-medical-applications)
- 39↑
U.S. Food and Drug Administration. Device software functions including mobile medical applications. Silver Spring, MD, USA: FDA. (available at: https://www.fda.gov/medical-devices/digital-health-center-excellence/device-software-functions-including-mobile-medical-applications)
- 40↑
welldoc. Welldoc Receives 11th 510(k) Clearance from FDA for Award-Winning Diabetes Platform BlueStar® - Welldoc. Columbia, MD, USA: welldoc, 2023. (https://www.welldoc.com/news/welldoc-11th-510k-clearance-fda-diabetes-platform/)
- 41↑
Lawrence K, Rodriguez DV, Feldthouse DM, Shelley D, Yu JL, Belli HM, Gonzalez J, Tasneem S, Fontaine J, Groom LL, et al. Effectiveness of an integrated engagement support system to facilitate patient use of digital diabetes prevention programs: protocol for a randomized controlled trial. JMIR Research Protocols 2021 10 e26750. (https://doi.org/10.2196/26750)
- 42↑
Ferreira A, Correia A, Silva A, Corte A, Pinto A, Saavedra A, Pereira AL, Pereira AF, Cruz-Correia R, & Antunes LF. Why facilitate patient access to medical records. Studies in Health Technology and Informatics 2007 127 77–90.
- 43↑
Damen DJ, Schoonman GG, Maat B, Habibović M, Krahmer E, & Pauws S. Patients managing their medical data in personal electronic health records: scoping review. Journal of Medical Internet Research 2022 24 e37783. (https://doi.org/10.2196/37783)
- 44↑
Council of Europe. Recommendation no. R (97) 5 of the Committee of Ministers to member states on the protection of medical data. International Digest of Health Legislation 1998 49 502–508.
- 45↑
Leigh S, Baines R, Stevens S, Garba-Sani Z, Austin D, & Chatterjee A. Walk a mile in my shoes: perspectives towards sharing of health and experience data among individuals living with sickle cell disorder. mHealth 2024 10 4. (https://doi.org/10.21037/mhealth-23-18)
- 46↑
Blenner SR, Köllmer M, Rouse AJ, Daneshvar N, Williams C, & Andrews LB. Privacy policies of android diabetes apps and sharing of health information. JAMA 2016 315 1051–1052. (https://doi.org/10.1001/jama.2015.19426)
- 47↑
Huckvale K, Prieto JT, Tilney M, Benghozi PJ, & Car J. Unaddressed privacy risks in accredited health and wellness apps: a cross-sectional systematic assessment. BMC Medicine 2015 13 214. (https://doi.org/10.1186/s12916-015-0444-y)
- 48↑
He D, Naveed M, Gunter CA, & Nahrstedt K. Security concerns in android mHealth apps. AMIA Annual Symposium Proceedings 2014 2014 645–654.
- 49↑
Klonoff DC. Cybersecurity for connected diabetes devices. Journal of Diabetes Science and Technology 2015 9 1143–1147. (https://doi.org/10.1177/1932296815583334)
- 50↑
Bierbrier R, Lo V, & Wu RC. Evaluation of the accuracy of smartphone medical calculation apps. Journal of Medical Internet Research 2014 16 e32. (https://doi.org/10.2196/jmir.3062)
- 51↑
Saeed H, Malik H, Bashir U, Ahmad A, Riaz S, Ilyas M, Bukhari WA, & Khan MIA. Blockchain technology in healthcare: a systematic review. PLoS One 2022 17 e0266462. (https://doi.org/10.1371/journal.pone.0266462)
- 52↑
Fernández-Caramés TM, Froiz-Míguez I, Blanco-Novoa O, & Fraga-Lamas P. Enabling the Internet of mobile crowdsourcing health things: a mobile fog computing, blockchain and IoT based continuous glucose monitoring system for diabetes mellitus research and care. Sensors (Basel) 2019 19 3319. (https://doi.org/10.3390/s19153319)
- 53↑
BinDhim NF, Hawkey A, & Trevena L. A systematic review of quality assessment methods for smartphone health apps. Telemedicine Journal and e-Health 2015 21 97–104. (https://doi.org/10.1089/tmj.2014.0088)
- 54↑
Stoyanov SR, Hides L, Kavanagh DJ, Zelenko O, Tjondronegoro D, & Mani M. Mobile app rating scale: a new tool for assessing the quality of health mobile apps. JMIR mHealth and uHealth 2015 3 e27. (https://doi.org/10.2196/mhealth.3422)
- 55↑
Diabetes Technology Society. Standard for Wireless Diabetes Device Security (DTSec). Burlingame, CA, USA: Diabetes Technology Society, 2017. (available at: https://www.diabetestechnology.org/dtsec/DTSec%20Standard.pdf)
- 56↑
Lee JL, & Kim Y. Evaluation of mobile applications for patients with diabetes mellitus: a scoping review. Healthcare (Basel) 2024 12 368. (https://doi.org/10.3390/healthcare12030368)
- 57↑
American Diabetes Association. Statistics About Diabetes. Arlington, VA, USA: ADA, 2024. (available at: https://diabetes.org/about-diabetes/statistics/about-diabetes)
- 58↑
Meisters R, Albers J, Sezer B, de Galan BE, Eussen SJPM, Stehouwer CDA, Schram MT, van Greevenbroek MMJ, Wesselius A, Koster A, et al. Socioeconomic inequalities in health-related functioning among people with type 2 diabetes: longitudinal analyses in the Maastricht Study. BMC Public Health 2024 24 73. (https://doi.org/10.1186/s12889-023-17553-z)
- 59↑
Hansen AH, Bradway M, Broz J, Claudi T, Henriksen Ø, Wangberg SC, & Årsand E. Inequalities in the use of eHealth between socioeconomic groups among patients with type 1 and type 2 diabetes: cross-sectional study. Journal of Medical Internet Research 2019 21 e13615. (https://doi.org/10.2196/13615)
- 60↑
Choi H, Kim B, Kim I, Kang JG, Lee Y, Lee H, & Park MH. Analysis of the status and future direction for digital therapeutics in children and adolescent psychiatry. Soa Chongsonyon Chongsin Uihak 2023 34 192–203. (https://doi.org/10.5765/jkacap.230044)
- 61↑
Davison NJ, Guthrie NL, Medland S, Lupinacci P, Nordyke RJ, & Berman MA. Cost-effectiveness analysis of a prescription digital therapeutic in type 2 diabetes. Advances in Therapy 2024 41 806–825. (https://doi.org/10.1007/s12325-023-02752-2)
- 62↑
Gentili A, Failla G, Melnyk A, Puleo V, Tanna GLD, Ricciardi W, & Cascini F. The cost-effectiveness of digital health interventions: a systematic review of the literature. Frontiers in Public Health 2022 10 787135. (https://doi.org/10.3389/fpubh.2022.787135)
- 63↑
Clements M, Kaufman N, & Mel E. Using digital health technology to prevent and treat diabetes. Diabetes Technology and Therapeutics 2023 25 S90–S108. (https://doi.org/10.1089/dia.2023.2506)
- 64↑
Sim B, Ju JH, Kim B, & Lee JY. Stakeholders' perceptions regarding digital therapeutics reimbursement in South Korea: qualitative study. JMIR mHealth and uHealth 2023 11 e47407. (https://doi.org/10.2196/47407)
- 65↑
Kerr D, Gabbay RA, & Klonoff DC. Finding real value from digital diabetes health: is digital health dead or in need of resuscitation? Journal of Diabetes Science and Technology 2018 12 911–913. (https://doi.org/10.1177/1932296818771200)
- 66↑
Agarwal P, Mukerji G, Desveaux L, Ivers NM, Bhattacharyya O, Hensel JM, Shaw J, Bouck Z, Jamieson T, Onabajo N, et al. Mobile app for improved self-management of type 2 diabetes: multicenter pragmatic randomized controlled trial. JMIR mHealth and uHealth 2019 7 e10321. (https://doi.org/10.2196/10321)
- 67↑
Mumtaz H, Riaz MH, Wajid H, Saqib M, Zeeshan MH, Khan SE, Chauhan YR, Sohail H, & Vohra LI. Current challenges and potential solutions to the use of digital health technologies in evidence generation: a narrative review. Frontiers in Digital Health 2023 5 1203945. (https://doi.org/10.3389/fdgth.2023.1203945)
- 68↑
Durán CO, Bonam M, Björk E, Hughes R, Ghiorghiu S, Massacesi C, Campbell A, Hutchison E, Pangalos MN, & Galbraith S. Implementation of digital health technology in clinical trials: the 6R framework. Nature Medicine 2023 29 2693–2697. (https://doi.org/10.1038/s41591-023-02489-z)
- 69↑
Morales DR, & Arlett P. RCTs and real world evidence are complementary, not alternatives. BMJ 2023 381 736. (https://doi.org/10.1136/bmj.p736)
- 70↑
Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behavioral Research 2011 46 399–424. (https://doi.org/10.1080/00273171.2011.568786)
- 71↑
Raghupathi W, & Raghupathi V. Big data analytics in healthcare: promise and potential. Health Information Science and Systems 2014 2 3. (https://doi.org/10.1186/2047-2501-2-3)
- 72↑
Kuo MH, Sahama T, Kushniruk AW, Borycki EM, & Grunwell DK. Health big data analytics: current perspectives, challenges and potential solutions. International Journal of Big Data Intelligence 2014 1 114–126. (https://doi.org/10.1504/IJBDI.2014.063835)
- 73
Obermeyer Z, & Emanuel EJ. Predicting the future - big data, machine learning, and clinical medicine. New England Journal of Medicine 2016 375 1216–1219. (https://doi.org/10.1056/NEJMp1606181)
- 74↑
Deo RC. Machine learning in medicine. Circulation 2015 132 1920–1930. (https://doi.org/10.1161/CIRCULATIONAHA.115.001593)
- 75↑
Chung JY. Digital therapeutics and clinical pharmacology. Translational and Clinical Pharmacology 2019 27 6–11. (https://doi.org/10.12793/tcp.2019.27.1.6)
- 76↑
Redfield MM, Anstrom KJ, Levine JA, Koepp GA, Borlaug BA, Chen HH, LeWinter MM, Joseph SM, Shah SJ, Semigran MJ, et al. Isosorbide mononitrate in heart failure with preserved ejection fraction. New England Journal of Medicine 2015 373 2314–2324. (https://doi.org/10.1056/NEJMoa1510774)
- 77↑
Snipelisky D, Kelly J, Levine JA, Koepp GA, Anstrom KJ, McNulty SE, Zakeri R, Felker GM, Hernandez AF, Braunwald E, et al. Accelerometer-measured daily activity in heart failure with preserved ejection fraction: clinical correlates and association with standard heart failure severity indices. Circulation. Heart Failure 2017 10 e003878. (https://doi.org/10.1161/CIRCHEARTFAILURE.117.003878)
- 78
Faris O, & Shuren J. An FDA viewpoint on unique considerations for medical-device clinical trials. New England Journal of Medicine 2017 376 1350–1357. (https://doi.org/10.1056/NEJMra1512592)
- 79↑
Sharma A, Harrington RA, McClellan MB, Turakhia MP, Eapen ZJ, Steinhubl S, Mault JR, Majmudar MD, Roessig L, Chandross KJ, et al. Using digital health technology to better generate evidence and deliver evidence-based care. Journal of the American College of Cardiology 2018 71 2680–2690. (https://doi.org/10.1016/j.jacc.2018.03.523)
