m-Health 2.0: New perspectives on mobile health, machine learning and big data analytics
Introduction
Mobile health (m-Health) is considered one of the most transformative drivers for healthcare delivery innovations in modern times and has been repeatedly called the biggest technological breakthrough of our modern times [1]. In recent years, big data has become increasingly synonymous with mobile health, however the key challenges of ‘Big Data and mobile health’, remains largely untackled. This is becoming particularly important with the deluge of the structured and unstructured data sets produced from the numerous mobile health systems globally. With both m-Health and big data becoming increasingly popular buzzwords within the global healthcare market communities, there is still lack of proper understanding of the most appropriate computational and analytical framework and tools required for this rapprochement, considering the rapid technological developments of the former with the challenges associated with the applications and utilization of the latter. The global usage of the mobile health applications is increasing rapidly and so is the voluminous data sets generated from these and other smart connected devices leading to complex, voluminous and multi-dimensional mobile health data that are collected and stored globally on explosive levels. This game-changing trend is largely propelled by the unprecedented global usage of the Internet connected devices, and the massive amounts of smart phone data generated by services and applications linked to these devices. As a consequence, there is major push on how to better manage, optimise and analyze this volume of data. Also, and more importantly how to convert these into meaningful information that can benefit patients, clinicians and other stakeholders. The recent developments from major corporations of the likes Apple (HealthKit), Google (DeepMind), Microsoft towards developing smarter mobile healthcare systems are evident of these trends.
To understand this further, it is essential to illustrate the original and basic structure of m-Health. Mobile health (m-Health) was first defined as ‘mobile computing, medical sensors and communication technologies for healthcare’ [2]. It has evolved since then into a major global healthcare delivery innovation and technological area albeit mostly driven by successful business and market sectors. It is still aiming to reach the tipping point from the healthcare delivery efficiency and efficacy aspects and for large scale clinical adoption [2], [4], [5]. This rapid popularity is mainly driven by the fast innovative developments of the three technological pillars on m-Health: telecommunications, computing and medical sensing as shown in Fig. 1 [2]. Billions of smart phones and Internet connected devices connected to tens of thousands of mobile health applications (Apps) are used worldwide by patients, clinicians and healthcare providers. These are continuously generating deluge of structured and unstructured datasets that are disclosing new healthcare insight opportunities but also producing challenges. The rapprochement between m-Health and big data is not new or surprisingly incidental, considering that big data science falls within the original computing pillar represented in one of the technological pillars of m-Health shown in Fig. 1
The transformative technological leap associated with the introduction of the first smart phone more than a decade ago had a major if albeit controversial impact on the evolution of m-Health. This technological breakthrough allowed the inevitable amalgamation of the three building technological pillars into a single and unifying smart communications device. This also enabled powerful computational tools to be embedded within the smart phones and other devices (e.g. smart watches, wearable monitors) to generate massive and ubiquitous mobile health data sets.
This innovation step also propelled the dawn of the ‘smart phone centric m-Health’ era with the patient centric approach becoming the focal point of this model [2]. Today, smart phone m-Health applications (Apps) are increasingly connected to variety of wearable sensors and Internet of Things (IOT) device in variety of healthcare applications [2], [3], [12]. These are also becoming increasingly pervasive in every day health, wellness and clinical areas and other healthcare applications. These include for example smart cardiovascular mobile health monitors that are clinically validated as heart monitors used to monitor, diagnose and alert heart patients. Others used for the monitoring and management of diabetic patients by continuously monitoring their blood glucose levels or insulin intakes [2]. Today, these m-Health apps are exceeding the 200,000 barrier listed on the main smart phone markets with increasing trend that seems unstoppable not at least in the foreseeable future. These applications are increasingly transforming current big data repositories from their care episodes to more effective and smarter repositories for preventive approaches applied for different diseases. However, this progress is also fuelling the challenges and risks associated with the big data sets generated from these applications and devices on daily basis. Yet, this data is often either incomplete or unavailable to the healthcare providers, patients and other beneficiaries and remains largely meaningless to the most.
The capability of future mobile health systems to translate and successfully transform this lack of actionable data to a meaningful one remains one of the key challenges in developing smarter more personalised and efficient m-Health delivery systems. This new ‘big m-Health data’ science include the structured and unstructured data sources generated from these systems not only within the human body but also within the wider spectrum of the human health determinants; namely genetics, body, behaviour, social and society levels [6], [11]. This deluge of big m-Health data is also fuelling serious concerns and unmasking the dark side of big data with its bad and ugly characteristics [7]. These include in addition to privacy issues and security threats, the appropriate clinical data base design and data extraction complexities, the accuracy of big data statistical and analytical approaches in different clinical setting and many others.
Section snippets
m-Health 2.0 and big data
Big data are generally categorised into two types [2]:
- (i)
Structured data: These generally refer to the data that has a defined length and format (numbers, dates, strings etc.) for big data. These are the data generated by sources such computers, mobile phone, sensors, web logs, etc. This type of data represents the minority (generally around 20% of the total data generated). Examples of structured health data include data extracted from EPR, EHR, home monitoring and treatment data, medical
m-Health data analytics and machine learning tools
In simple terms, whilst data mining is typically concerned with sifting through the data to search for previously unrecognized patterns and trends, data analytics is mainly about breaking down such data and assessing the impact of these patterns overtime and predicting future trends. Big data analytics promise to deliver important and transformative healthcare insights. However, from the m-health perspective, there is still further work to be carried out in this important area. This lack of
Conclusions
In this paper we presented new perspectives of mobile health (m-Health) with big data analytics and machine learning. We presented the relevant big data issues from the mobile health perspective. In particular we discussed these issues from the relevant technological areas and the building blocks (communications, sensors and computing) that are associated with mobile health and the new concept of (m-Health 2.0). We also presented and discussed the rapprochement of m-Health data analytics with
References (16)
- et al.
Can mobile health technologies transform health care
J. Am. Med. Assoc
(2013) - et al.
M-health: Fundamentals and Applications
(2017) - Istepanian, R S H, Hu, S, N Philip, Sungoor, A., ‘The potential Internet of Things of m-Health Things (m-IOT) for...
- et al.
Introduction to focused issue on mHealth infrastructure: issues and solutions that challenge optimal deployment of mHealth products and services
mHealth
(2017) - et al.
mHealth: Transforming Healthcare
(2014) National surveys of population health: big data analytics for mobile health monitors
Big data
(2015)- et al.
Research based on big data: the good, the bad, and the ugly
J Thorac Cardiovasc Surg
(2016) M-health 2.0 – the future of mobile health from 5G and IOT perspectives
Int. J. of Sensor Network Data Commun.
(2016)
Cited by (71)
An intelligent disease prediction system for psychological diseases by implementing hybrid hopfield recurrent neural network approach
2023, Intelligent Systems with ApplicationsA review of IoT systems to enable independence for the elderly and disabled individuals
2023, Internet of Things (Netherlands)Exploring benefits and ethical challenges in the rise of mHealth (mobile healthcare) technology for the common good: An analysis of mobile applications for health specialists
2023, TechnovationCitation Excerpt :The goal in this ethical context is to provide further suggestions for ethical data management, apart from legal necessity, and follow societal tensions regarding the ethics of data-driven products and services that existing norms still fail to intercept (Loi et al., 2019). However, there is still a lack of proper understanding of the required computational and analytical framework tools for these technological developments) (Istepanian and Al-Anzi, 2018). Some existing literature has investigated whether ethical considerations are among the selection criteria of a user when downloading a health app.
Big data analytics and artificial intelligence technologies based collaborative platform empowering absorptive capacity in health care supply chain: An empirical study
2023, Journal of Business ResearchCitation Excerpt :Certain specific characteristics of artificial intelligence include natural language processing, expert systems, and speech recognition (Turulja, Cinjarevic, & Veselinovic, 2020; Yoon, Lee, & Schniederjans, 2016). When combined, big data analytics and artificial intelligence can bring a synergized association; in which artificial intelligence will not provide any output when there is no data (Istepanian & Al-Anzi, 2018; Kamble, Gunasekaran, Goswami, & Manda, 2018), and managing huge datasets without the involvement of artificial intelligence can be extremely surmounting (Hole, Pawar, & Khedkar, 2019; Kumar, Mookerjee, & Shubham, 2018). Omnichannel healthcare services based on technological interventions like big data analytics and artificial intelligence are emerging as an exciting phenomenon in various business environments, the healthcare industry is not an exception (Qian & Acs, 2013; Yu, Beam, & Kohane, 2018).
The platform development, adherence and efficacy to a digital Brief therapy for insomnia (dBTI) during the COVID-19 pandemic
2022, MethodsCitation Excerpt :Driven by the Internet, Internet of Things, big data and artificial intelligence, the Fourth Industrial Revolution centers on the great integration of networks, information and intelligence [1]. The application of these technologies to medical industry has brought about the booming of digital medicine, and even the emerging of numerous concepts such as m-health and digital therapeutics [2,3]. The advantages of remote monitoring, portability and personalized treatment characteristic of digital therapeutics have been on full display since the breakout of coronavirus disease 2019 (COVID-19) Pandemic in 2020, and thereupon have drastically accelerated the digital revolution of the world [4].
A Review on the Importance of Machine Learning in the Health-Care Domain
2024, EAI Endorsed Transactions on Pervasive Health and Technology