Broadly speaking, one of such uses has been to support innovation and decision making, and one area where this has gained traction recently is the health care industry, particularly elderly care. Elderly people represent a growing proportion of the world’s population. As noted by the World Population Ageing Report [2], the population of people in the world aged 60 and older has risen from 205 million in 1950 to almost 810 million in 2012, and is projected to double by 2050, reaching 2 billion (approximately 20% of the world’s population and an increase of about 181% between 2010 and 2050). In addition, we are living longer. It is anticipated that we can expect to have an increase in life expectancy at birth from 86 years for males and 92 years for females by 2050 [3]. Building technologies to empower the elderly to stay healthy and active and, as importantly, participate in decisions pertaining to their health care is at the center of studies on personalized elderly care. Achieving this remains challenging primarily because of huge differences in individual care needs and the progressive nature of ageing [4].

Personalization entails processes that encourage and respect the contributions elderly persons can make to their own health—from the perspective of their values, goals, past experience, and knowledge of their own health needs—with the underlying goal of providing for them care that is “predictive, personalized, preventive and participatory” [5]. However, existing approaches often treat the elderly as compliant individuals, and deliver services through a series of discrete care episodes that hardly take account of the varied needs and lifestyles of the elderly [6]. With exabytes of data generated on a daily basis about our activities from digital footprints and online activities, recent research efforts are focusing on exploiting these massive data sets to address the challenge of personalized elderly care. But, despite the potential ability of Big Data to add to our knowledge of ourselves in novel ways and to enrich our understanding of people’s latent needs and preferences [7-9], deriving valuable insight from Big Data poses numerous challenges.

In this chapter, we describe the opportunities and challenges that leveraging Big Data analytics for elderly care presents. It is argued that though potentially viable as an approach to raise elderly care to a level where the uniqueness of each elderly person is adequately recognized [10], it does pose inherent and enormous social, context-aware, and technical challenges that research cannot afford to ignore. We present a review of research in personalized services for elderly care and explore the challenges in the area of Big Data analytics for elderly care. Based on the discussion, ACTVAGE—a context-aware lifestyle-oriented framework for supporting personalized elderly care and independent living is proposed. The framework combines systematic capture of past lifestyles and knowledge of current activities and user context. It then applies rigorous analytics to build a complete picture of the elderly person's lifestyle and contextual needs in order to build a formal representation of the lifestyle concept for system design. Based on the representation, required services, including social networking, self-diagnosis and monitoring, advisory, entertainment, exercise, dietary, reminder, and local events services are developed to offer individually tailored and lifestyle-oriented support for active ageing and independent living.

Existing research approaches to developing personalized elderly care technologies can be broken down into two broad categories: (1) monitoring and surveillance, whereby electronic technologies are deployed to keep track of an elderly person’s medical condition and daily activities, and automatically alert health care staff when intervention is required. Research in this category focuses primarily on smart homes and the monitoring activities of daily living. (2) Assistive technologies, which can range from a simple device to help a physically impaired old adult, for example, with Parkinson disease to turn on a tap without much risk, to a personal robot that is able to spoon-feed an elderly person or administer medication without the need for a health care staff. In what follows, we briefly discuss a number of research efforts in these areas. These efforts overlap, and it is not unlikely to find, for example, a project that integrates the concept of a smart home and the technique of activities of daily living monitoring [33], or a smart home that employs a robot care assistant [29]. Our goal here is to critically review research in these areas in relation to the key challenges of personalized elderly care as discussed above.

Big Data refers to data that are so diverse, complex, and large-scale combining both structured and unstructured information contents from multiple sources [59] so much that it is technically challenging, if not impossible, to be processed using existing traditional software and analytic methods, such as relational databases and management information systems. Boyd and Crawford [60] outline a definition that highlights so interestingly both the opportunities and challenges of Big Data in elderly care. According to the authors, Big Data is “a technological, cultural, and scholarly phenomenon” that rests on the interplay of three things, namely, (1) technology, taking advantage of increased computing power and algorithmic accuracy to gather, analyze, link, and compare large data sets; (2) analysis, drawing on large sets to identify patterns in order to make economic, social, technical, and legal claims; and (3) mythology, the computational turn in thought and research that large data sets offer a higher form of intelligence and knowledge that can generate insights that were previously impossible, with the aura of truth, objectivity, and accuracy. Data are not technically “neutral” [61], and if considered otherwise, become, in the words of Lisa Gitelman [62], an “oxymoron.” Deriving actionable value from Big Data, therefore, requires that research takes cognizance of the contexts that have shaped the data collection, the forms of arguments and analysis that can, and are, made with the data, as well as a clear indication of what the data do not, and cannot, account for [63]. Big Data⁴ for elderly care comes from three major sources. The first, popularly referred to as Big Data, involves analytics of multiple types of data across a population, typically over a period of time, potentially from multiple sources, of structured and unstructured nature, and of heterogeneous kinds of objects. This has been a by-product of ubiquitous smart technologies that have pervaded every aspect of our life. The second comes from what Cornell NYC Tech professor Deborah Estrin has called small data, the output of a whole host of pervasive tracking processes about any one individual, e.g. data from quantified self [61]. The third source of Big Data refers to what ethnographers, such as Tricia Wang [64], have called thick data, and includes more qualitative kinds of data, obtained, for example, from interviews and direct observational studies, which are needed to bridge and/or reveal knowledge gaps in quantitative and algorithmic Big Data. Besides the large volumes of data available to the health care industry, driven by digitization of EHRs, laboratory systems, physician notes, medical correspondence, regulatory requirements, and care plans [65-67], which many experts say double every five years, as well as data from social media and public health records, current approaches in the use of Big Data for elderly care tend to focus on the Big Data and small data categories. These approaches leverage ubiquitous sensor-enabled devices and such techniques as sociometeric badges [9] and the concept of quantified self⁵ [68]. Though potentially capable of providing statistical regularities in a “measured life” [69] and a mathematical explanation for understanding human dynamics [9] and personal health [70], these techniques have been critiqued for their tendency to hide rather than reveal the very essence of what makes us who we are [63], as they may automatically affect how “the self is understood, experienced, and practiced more broadly” [71].

The biggest opportunity that Big Data offers for the personalization of elderly care lies not only in the availability of large volumes of data but also in the diversity of both the types and sources of data. This opportunity, it could be argued, will potentially be harnessed at the confluence of health care and personal location data, which according to a recent report from McKinsey Institute, constitute two of the five domains that would be greatly transformed by Big Data [72]. As noted earlier, one of the major challenges in personalized elderly is the difficulty in specifying the concept of personalization and operationalizing it for computational design. This is all the more challenging in view of the fact that elderly people usually suffer multiple age-related ailments, such as dementia and Parkinson disease, that impair considerably their cognitive and physical capabilities. In such cases, health care practitioners often resort to information from historical records and family members in order to gain an understanding of the patient. By digitizing, combining, and effectively using Big Data, the health care industry stands an unprecedented chance to gain a significant understanding of the elderly patient. This will lead to more targeted treatment outcomes and increase the chances of detecting diseases at earlier stages when they can be treated more easily and effectively. In addition, certain care outcomes, such as allergies, risks of infection or hospital-acquired illness; or the possibility of identifying patients who may choose elective surgery or benefit from proactive care or lifestyle changes; or the possibility of comorbid conditions due to environmental, family, or lifestyle factors, etc. can be more effectively estimated on the basis of historical data [13]. Big Data is deeply related to the phenomenon of e-Science [73], computationally intensive science that is usually carried out in highly distributed computing environments, using immense data sets and requiring grid computing. From this perspective, Big Data will potentially play a vital role as a part of e-Social Science in collecting, processing, and analyzing social and behavioral data [74] for personalized elderly care.

The framework consists of four layers, namely, the data layer, the analytics layer, the context-awareness layer, and the services layer (see Figure 5.2). The approach will combine systematic data capture and rigorous analytics to uncover the actual and complete picture of the elderly person's lifestyles and build a formal representation of the lifestyle concept for system design. Lifestyle information is integrated with context-aware capabilities that are dynamically configurable using the user’s location, identity, time, and an understanding of user activities of daily living in order to provide personalized recommendations. Based on the recommendations, a set of active ageing services are developed, using the capabilities of mobile devices and cloud-based services, to offer individually tailored and lifestyle-oriented services for active ageing and independent living, including social networking, self-diagnosis and monitoring, advisory, entertainment, exercise and dietary, reminder and local events services. The e-services will leverage on advances in ICT, including pervasive computing and usability engineering, for the provisioning of rich and platform independent e-services.

In developing the conceptual design of ACTVAGE, we draw on the seven principles and best practices to guide the development of Big Data analytics systems [74,90,85]. Exploiting Big Data analytics, as noted by Chen and Zhang [74], requires not only new technologies but also new ways of thinking. In the remainder of this section, we describe the steps and activities of the proposed framework and illustrate its application.

It is usually the norm that people with dementia, especially in rich economies where attaining great age is increasingly the case, are cared for in residential homes by professional (though not highly paid) carers whose job is typically viewed as having low status [10]. They are usually under pressure to balance work and administrative duties and are left with little time to gain sufficient understanding of their residents. In addition, these carers regularly encounter challenging behavior from the residents, including refusal to eat and take medication, inconsistent behaviors, and physical and verbal aggression that is difficult to diagnose and resolve [10]. No two residents are alike, and what has worked for one might not work for another. Existing solutions have leveraged techniques in person-centered care. However, there remains the need to gain an in-depth understanding of a dementia resident, particularly in cases where families do not have much valuable information about the resident. Here, we discuss how the various steps in ACTVAGE could enable a carer gain an in-depth understanding of a dementia resident’s life history, values, and interests for personalized care (see Figure 5.6).

The system acquires data using both Big Data and thin data driven approaches (Facebook and Twitter data) as well as thin based approaches (semistructured interviews) via the Data Aggregator (see Figure 5.4). The data sets acquired are preprocessed, transformed, and stored in the digital life history repository. As shown in Figure 5.5, the repository contains data about several aspects of a resident’s past life events, interests, values, and preferences—modeled as a lifestyle graph (in step two). The Aggregator is designed to allow the carer and family members to remotely upload information about a resident using separate bespoke mobile apps and web applications, such as Digital Life Story.⁷ Context information about user activity is captured via wall-mounted cameras. The carer is able to enter queries and perform data analysis aimed to construct user lifestyle profile using the analytics and query engine. Services are composed in the recommendation engine. The system interface allows the carer to perform these actions as well as to view and analyze system output.

This chapter contributes to research efforts that aim to highlight the opportunities and challenges of Big Data analytics and lay out foundations for the design and development of applications that leverage the Big Data to address the problem of the world’s ageing population. We present an overview of research in personalized elderly care, and explore the opportunities and challenges in leveraging Big Data analytics for personalized elderly care and independent living. In particular, we have focused our discussion of the opportunities and challenges to be addressed in relation to the five-dimensional V’s of Big Data, as well as the multistep process of Big Data analytics, while acknowledging that other challenges remain that are not covered in this chapter. Based on this discussion, we propose ACTVAGE—a context-aware lifestyle-oriented framework for supporting personalized elderly care and independent living. The framework will potentially improve elderly care through the development of a model of ICT-enabled support for independent living that is grounded in an understanding of a history of user lifestyles and personal preferences, as well as knowledge of user activity context. By developing a set of lifestyle-oriented active ageing e-services, the proposed framework makes a novel contribution to elderly health care through an approach that exploits the lifestyle concept combined with context-aware user support, which itself has received considerable attention in ICT and computer science research, as a foundation for informing the design of ICT-enabled active ageing services and products. In addition, the approach will lead to the development of usable technologies for encouraging independent living, reducing reliance on institutionalized care, and promoting active social and healthy life into old age.

In a compelling paper, Choudhury et al. [95] present an approach that leverages Big Data, in particular Facebook data, to identify women at risk of postpartum depression at an early stage and provide them access to appropriate services and support. They employed statistical analysis and survey techniques to demonstrate the feasibility of Big Data as a tool to detect, characterize, and predict postpartum depression in new mothers, and as a result, recommend treatment plans to avoid recurrence. Like Choudhury and her colleagues, this chapter makes a case that Big Data has the potential to revolutionize care by providing us access to the vast amounts of data about individuals via self-reports, transaction data, and online activities as ground truth for developing ICT-based services for health care. We propose this innovative approach in a way that integrates digital life history with user activity context.

The chapter has focused on a high-level description of the ACTVAGE framework and lacks specific details about a number of the techniques included in the proposed framework, for example, life event modeling, ontology development, graph analytics, context information processing, and service composition. Future work will focus on developing specific details of the framework as well as prototype and evaluate systems based on it within a real-world application context. Our goal here is primarily to present the conceptual architecture of a new framework that combines knowledge of user lifestyle and knowledge of user activity context toward tackling the problem of personalized elderly care, and to show how the conceptualization was informed by meta-analysis of studies in personalized elderly care. As such, we have pointed to research studies, tools, and techniques that could be leveraged in developing and implementing solutions based on the proposed framework. In addition, we have illustrated the applicability of the framework to support elderly people living with dementia using a real-world example scenario. Integrating context and history of lifestyles into a single framework presents a lot of challenges. The value to be derived from Big Data is both analytically and technically bound to the context of production and use of the data, which as Dourish [87] notes is “a slippery notion” [p. 29]. Context is a central issue for Big Data analytics system. The power of Big Data to deliver on the promise of differentiated and useful insights lies precisely in the ability of analytics algorithms and systems to take into account both the context that has shaped data production and the context within which analytics-driven recommendations would be utilized. The overarching goal of the work described in this chapter is to lay a foundation for combining Big Data analytics with context-aware computing as a novel way to address the problem of personalized elderly care. The proposed framework will equally apply to many problem areas that leverage Big Data analytics.