Pervasive Healthcare Monitoring System (PHMS) : An investigation into the acceptance of PHMS by consumers.

(1)

Pervasive Healthcare Monitoring System

(PHMS):

An investigation

into the acceptance of

PHMS by consumers.

bianca iordachioaiei

bianca_iord@yahoo.com

Student ID: 10270876

Supervisor: Dr J.F.M. Feldberg

Programme: MBA part time, 2011 - 2013 Date of submission: 15-Sept-2013 Confidentiality restrictions: none

(2)

bianca iordachioaiei Page 2

Abstract

Health information systems, and in particular Pervasive Healthcare Monitoring Systems (PHMS), promise to revolutionize healthcare with their design for personal and continuous use. Existing studies show that by exploiting mobile wearable medical sensors, the society can achieve significant economic savings, improvement in quality of care and increased life expectancy.

However, despite all technological progress in the area of mobile computing, wireless sens-ing and real-time monitorsens-ing of patient’s vital signs, PHMS devices are still not a common presence in our daily life.

This study attempts to bring its scientific contribution by proposing and empirically testing the acceptance model for PHMS developed in this study. Factors from several reviewed the-ories are put together: perceived usefulness and perceived ease of use from Technology Ac-ceptance Model; privacy, reliability and personalization issues are derived from the theories on Opportunistic and Participatory Sensing; personality traits like Drive to Learn and Drive to Defend are derived from Task Technology Fit Model. To the best of my knowledge this is the first study to test a PHMS acceptance model.

A vignette-based online survey of 104 respondents is used in order to test the hypotheses. Quantitative analysis is done with Partial Least Squares method, incorporating tests for the validity and reliability of the results. Overall, the model explained 71.9% of the variance in intention to use PHMS.

The findings show that Perceived Usefulness and Drive to Learn are the most significant drivers for accepting PHMS (p<0.001), followed by system reliability, confidentiality and pri-vacy (p<0.01). Personalization, interaction and drive to defend, on the other hand, did not prove to be significant in this study.

Based on the detailed key findings, managerial implications are discussed, keeping in mind the stakeholders who might benefit from the results of this study. For example, pa-tients/user’s gain of information and tighter connection to the doctor; objective body measurements for the doctors and improved data accuracy for medical research; lower cost of governmental healthcare expenses; more focused marketing for PHMS distributors or a larger footprint in the social responsibility commitment.

The limitations of this study and directions for future research are also presented.

Keywords: Pervasive Healthcare Monitoring System (PHMS), participatory sensing, wireless,

(3)

Acknowledgement

I would like to express my great appreciation to Dr J.F.M. Feldberg for his valuable and con-structive suggestions during the planning and development of this research work. His wis-dom, knowledge and commitment to the highest standards inspired and motivated me.

(5)

I. Introduction

Health has played an important role in any society, since the beginning of human kind: thanks to the discoveries in medicine, the life expectancy has grown from 59 to 78 years old since 1930 (National Center for Health Statistics, 2011). In the meantime, national reports are showing a significant increase in regards to the health expenses which is estimated to continue. For example, US is spending currently $2.7 trillion and is estimated to raise to $4.9 trillion by 2021 (National Centers for Medicare and Medicaid Services, 2011). According to the same source, the main cause is increasing average age, increased number of elderly and the shortage of medical personnel.

Another significant recent change in today’s society is the growth of mobile communication. According to a study by the World Bank and infoDev titled "Information and Communica-tions for Development 2012", referenced by Farber (2012), the mobile phone was first in-troduced to public use in 1978 and by 2014, the number of active cell phones will bypass the world’s population of 7.3 billion.

A few years ago a new trend emerged in connecting the health problem with the mobile communication explosion. Initiatives like “Living by Numbers” Health Conference1, started last year in New York by WIRED magazine, and “International Conference on Pervasive Computing Technologies for Healthcare”2, started in Innsbruck in 2006, are bringing togeth-er exptogeth-erts from medicine, science, technology and business, in ordtogeth-er to discuss the future of health for the human kind.

The actual introduction of telecommunications in healthcare was done on two levels, start-ing with year 1999: (1) Electronic Health Records (EHR), which is a set of protocols to man-age and share patient data between medical practitioners involved in personal care, and (2) m-Health, which is the approach of using mobile devices in handling EHRs and collecting health data in real-time by monitoring patient’s vital signs, such as body temperature, heart rate, blood pressure. (Klasnja & Pratt, 2012).

Mobile health monitoring is commonly discussed in regards to three main activities, which were highlighted by Gupta et al (2013): disease management (e.g. glucose meter, insulin pump, and heart rate monitor); health and fitness (e.g. heart rate monitor, weighing scale,

1_{http://www.wired.com/wiredscience/health-conference-2012/ (September,3, 2013)} 2

(6)

cardiovascular fitness monitor, strength fitness equipment, smartwatch, MapMyRun type of apps); and independent living, also known as “aging independently” (e.g. medication moni-tor, disease management device, paramedics alerting, personalized toilet, Portal Monitor cameras system).)

Simple and affordable smartwatches (e.g. Garmin Forerunner 10) are becoming very popu-lar3,but there are also more advanced biomedical parameterization devices for monitoring daily life. On a quick online search, one can find watches connected to GPS who can also gather personal pulse, bracelets for counting steps, mobile apps to record daily calories in-take (e.g. DietSense). From the academic world there are even more innovations on the way in regards to wireless sensors used for human’s future well-being: sensor vest that can send emergency SMS (Röcker & Ziefle, 2010), socks with pressure sensors to adjust the running style, online cameras system that could substitute the presence of a home caregiver called Portal Mobile, Tele-Immersion system to help people coping with their own psychological disorders (Gupta et al., 2013).

All these mobile and medical devices are going to gather large datasets that are challenging to store, search, share, visualise and analyse. In today’s business world, this is Big Data as it is explained in the Big Data Guide conveyed as a white paper by Oracle (Sun & Heller, 2012). Similar definition was provided in the research of Chen (2012) when mentioning Big Data and Big Data analytics: exabytes of datasets, including sensor information or social media, that require complex techniques to store, manage, analyse and visualize. The same paper shows the three (change all) step evolution of Big Data analytics, which started with busi-ness intelligence and analytics 1.0 (BI&A 1.0) as dashboards and scorecards from structured data warehouse repositories. BI&A 2.0 was marked by the spread of web-based unstruc-tured content like online encyclopaedia, blogs and social media. BI&A 3.0 is defined by the appearance of mobile and sensed content (Chen, 2012).

Building on a corporate framework, ICT research company Gartner defined four Vs in Big Data analytics: volume of data; velocity -the speed with which information is put into or taken out of the system; variety of data source - including structured traditional RDBMS

3

(7)

ta or unstructured digital content from social networking sites; value- which brings into pic-ture today’s business challenge: how to create maximum value from Big Data analytics? Extending the business challenge to the area of health, one can wonder how can general healthcare improve to spend less and prevent more, by using numbers and data analytics. An astonishing research in this area is done at the University of Groningen through the Life-Line project, involving a total of 165000 people from the north of The Netherlands: each person allows personal data to be collected every five years for 30 years (Stolk et al., 2008). The data ranges from blood analysis, DNA sequence and electrocardiographs to information about lifestyle, psychological and physiological factors.

Interpreting such medical mobile Big Data will require a lot of machine power and human defined algorithms. Some might be sceptical on whether humankind has already reached this level of intelligence and human-machine collaboration, but scientists are already devel-oping and testing advanced ICT systems to provide computing power required, as well as addressing the human demand for privacy and security (Han et al., 2010).

Figure 1 below summarizes the features and requirements of a personalized healthcare sys-tem, as seen by Rodrigues (2013). It includes all the stakeholders and their future roles: medical professionals will always be informed about the current health status, patients can receive assistance before the disease becomes critical, and remote monitoring systems trig-gering emergency assistance in case of need.

Figure 1 Pervasive Computing in Personalized Healthcare System (Rodrigues, 2013)

Putting together the concept of m-Health (the use of mobile devices for collecting patient health data and providing healthcare information to practitioners), the advanced level of

(8)

wireless sensor accuracy, and the capacity to store and handle daily exabytes using the new Big Data concept, “Pervasive Healthcare” is born: healthcare for anyone, anytime, any-where, as introduced by Varshney (2007).

A more recent article wrote that “the term “Pervasive Healthcare’’ was born from the syn-ergy of two fundamental research fields: pervasive computing and communications and e-Health systems. […] Pervasive e-Healthcare systems are targeted at personal and continuous monitoring of single users, in terms of health status (by exploiting wearable and embedded sensors), life style, and social support for better inclusion and independent living “ (Delmastro, 2012, p. 811)

For a complete understanding of potential technological disruptive innovation, Gupta et al. (2013) present how the entire medical care system will change with the arrival of Pervasive Healthcare Monitoring Systems. From the current reactive mode -where the treatment is done based on symptoms and diagnosis- to a future proactive style- where treatment will be preventive, based on system generated warnings and some of the health issues will even be avoided by real time alerts. For example, the smartphone can send a warning when sensing too many calories intake in order to prevent obesity in the longer term. Similarly, smart clothing can adjust the amount of heating generated according to the outside temperature to prevent hypothermia.

A. Research question

Pervasive Healthcare Monitoring is not just one new technology, but an innovative way of combining several technologies - mobility, wireless communications, sensing and Big Data- with the purpose of improving healthcare. National research programmes for healthcare have already been budgeted with more than $50 billion in the last five years and the bene-fits of technology applied in healthcare have already shown improvements of 8% in tariffs drops and 20% reduction in emergency admission (Al-Shorbaji, 2013).

The literature reviews thus far are focusing on the benefits and challenges of implementing PHMS on a large scale outside the controlled environment of a laboratory or a hospital (e.g. Iwaya et al.,2013; Qiang et al.,2011 ). However, as per my research, there has been little published in the area of what drives PHMS consumer acceptance and little seems to be known about actual burdens and user drives to wear a body monitoring device. My research

(9)

so far yielded no model or hierarchy of factors that could explain the acceptance of PHMS by the public…and this is exactly where this paper will focus, trying to answer the below re-search question:

What are the factors driving the adoption of Pervasive Healthcare

Moni-toring Systems?

The current research will provide a more detailed definition of Pervasive Healthcare and its applications on a global level: from developing to developed countries. In attempting to an-swer the above research question, a list of advantages and disadvantages that PHMS may bring to an individual’s daily life will be presented, based on previous studies and theories. This list will consequently be used to derive the factors influencing acceptance of PHMS.

B. Scientific and Managerials Contribution

Scientific

Mobile Health in general and Pervasive Healthcare Monitoring Systems in particular, appear in more and more academic studies, at an increasing rate over the last two years (Akter et al., 2013; Al-Shorbaji, 2013; Alsos et al.,2012; Atienza & Patrick, 2011; Christin et al.,2011; Cristofaro & Soriente, 2013; Dunnebeil et al.,2012; Garcia-Sanchez et al.,2013; Karippacheril et al.,2013; Pawar et al.,2012 ).

As per the focus of this study, various models of influence have been proposed, looking at factors in isolation, including: personality traits (Junglas et al.,2009), task matching (Goodhue & Thompson, 1995), perceived usefulness (Hernandez et al.,2009) and privacy in participatory and opportunistic sensing (Christin et al.,2011; Kapadia et al.,2010).

However this new industry that is being born under our own eyes needs an acceptance model for itself in order to understand what drives users to accept PHMS as part of their daily life and how the three new technology concepts-mHealth, wireless sensing, Big Data- are seen together in this new setting. Figure 2 shows PHMS as part of mHealth applications at the junction of health and technology, with a high influence from regulatory authorities (Qiang et al.,2011).

(10)

Figure 2 mHealth ecosystem (Qiang et al.,2011)

This technology has a wide reach on society: from the poor to the rich (Iwaya et al.,2013), from government to private businesses (Al-Shorbaji, 2013) and from young to old (Rodrigues, 2013). This technology might change the way the medical system works (Coiera, 2004) from reacting to a disease with treatment to proactively treating before the disease (Gupta et al., 2013). However, as with any technology, the gain it might bring is “often ob-structed by user’s willingness to accept and use available systems” (Davis F. , 1989, p. 319). In the area of mHealth, a couple of models have been proposed in the area of EHRs ac-ceptance by the medical stuff: Dunnebeil et al. (2012) modelled physicians’ acac-ceptance of mHealth in ambulatory care and Junglas et al. (2009) studied nurses’ willingness to use mo-bile information technology in the hospital. However, as per my research up to this moment in time, no model has been defined for the acceptance of PHMS by the end user, who is not necessary an expert in ICT or medicine. Therefore, the current study will contribute to the scientific world by introducing for the first time a model for user’s acceptance of wireless wearable monitoring sensors.

Using as basis the Technology Acceptance Model(TAM) (Davis et al., 1989), Task Technology Fit(TTF) (Goodhue & Thompson, 1995) and Opportunistic-Participatory Sensing theories (Burke, et al., 2006), the new model will be customized to the factors specific to healthcare monitoring in a wireless sensing environment, focusing on relevant issues including confi-dentiality, personalization, drive to defend and others.

All in all, this study answers the call for future research posed by Holden and Karsh (2010) “Aside from improved study quality, standardization, and theoretically motivated additions to the model, an important future direction for TAM is to adapt the model specifically to the health care context, using beliefs elicitation methods”. (Holden & Karsh, 2010, p. 169)

(11)

bianca iordachioaiei Page 11 Managerial/Business

Mobile technologies can not physically carry drugs, doctors and equipment between loca-tions, but they can carry and process information in many forms. Figure 3 shows some of the successful healthcare projects in the developing world, where mobile access made a dif-ference: e.g. 97% improvement in the time needed to deliver lab results in Kenya, two days to send by SMS instead of 45 days by normal post (Qiang et al.,2011).

Figure 3 Benefits of mHealth Application (Qiang et al.,2011)

As underlined by Qiang et al. (2011) the biggest risk of the mHealth industry is its own fragmentation: as it usually happens with technological disruptive innovations, many achievements are being proved on a small scale as prototypes, but they don’t manage to reach outside of the laboratory, to the end user. This problem could be solved by standardi-zation of mHealth services and applications. This exact issue of interoperability standards was raised at the eHealth and Personalized Health Care Summit in 2010 at Nice4, but the impact of it on the end user was not measured. In essence, the business world agrees that standardization is required in healthcare applications and this thesis will show it’s impact on consumer’s intention to use the monitoring devices.

Another issue raised in the literature is related to terms like “trust” (Alsos et al., 2012), “pri-vacy” (Cristofaro & Soriente, 2012), and “safety” (Smith & Eloff, 2002) – terms which fail to be clearly defined on the healthcare industry level (Christin et al.,2011). This study will also focus on the influence of system reliability on the final intent to use. The analysis will shine some light onto the effort that manufacturers of such medical high tech devices must put

4

(12)

into building a good reputation and conveying to the consumer the quality of their products and services.

The business benefit of the proposed PHMS acceptance model is that by understanding what drives the public to use wearable body monitoring, manufacturers could refocus their research and investment plans in order to address the area that matters most for the poten-tial consumer (e.g. make it more easy to use).

Furthermore, by knowing the most significant factors, the marketing communication can be adjusted so that it sends the message that will trigger the purchase. As this is all about un-derstanding the customer, this study can help complete the profile of the future wearer of pervasive body monitoring devices and provide useful information for a sustainable ICT business in healthcare.

The results of this study can also bring value to the medical care industry, improving com-munication between patients and medical stuff in regards to PHMS.

Overall, knowing what impacts the use of PHMS is beneficial for all stakeholders involved in the healthcare industry:

 Patients/users: gain information, tighter connection to the doctor, large data availa-bility, easy to gather health readings

 Doctors: would work with objective data, not only subjective symptoms; would han-dle less hospital admissions and more prevention actions; improved data accuracy for research purposes.

 Government: Lower cost of healthcare. A study by Friederici et al. (2012), for exam-ple, mentions as much as 30% cost reduction

 Business sales: Increased sales of health app: for example, Apple store declared 4000 apps on offer in Feb 2010 and 15000 in Sept 2011 (Friederici et al.,2012).

 Business focus: More focused marketing and manufacturing.

 Social responsibility: offering easy access to primary healthcare in developing coun-tries (Iwaya et al.,2013).

(13)

C. Research approach

Since the objective of this thesis is to shine light on the reasons driving the adoption of wearable health monitoring devices, a set of relevant factors will be derived using media reports and academic studies from the last few years.

The new proposed model for PHMS acceptance will be based on three existing theories, en-riching TAM framework with a theory about personality fit to tasks and with a theory to ful-fill the privacy and security challenges when it comes to people-centric sensing applications like wearable body monitoring devices. Specifically:

1) Technology Acceptance Model (TAM): Model introduced by Davis(1985) explaining individual’s behavioural intention to use technology by two main determinants: per-ceived usefulness and perper-ceived ease of use. After being referenced in 424 scientific papers in its first ten years (Venkatesh & Davis, 2000), TAM was proven to be valid in healthcare as well: “The recent increase in the use of TAM appears justified, with many of the relationships specified by TAM repeatedly validated in health care set-tings“ (Holden & Karsh, 2010, p. 166).

2) Task-Technology-Human fit: Goodhue & Thompson(1995) were first to propose the

Task Technology Fit (TTF) model arguing that a strong connection between task and technology, which implies that the technology “fits” the task to be executed, would impact technology’s utilization and ultimately individual’s work performance.

The paper of Junglas et al. (2009) reports on a study examining nurses' decisions to utilize MICTs (Mobile Information and Communication Technology). Their findings indicate that the utilization of healthcare technology is impacted not only by the na-ture of the task and the available technology in the hospital, but also by archetypical human drives of the nurses: drive to bond, drive to learn, drive to defend and drive to secure.

3) Opportunistic and Participatory sensing: With the rise of wireless networks and mobile access to the world wide web, there is more and more human involvement as part of the sensing infrastructure (Klasnja & Pratt, 2012).

Lane et al. (2007) conceptualize a spectrum of conscious human involvement, rang-ing from opportunistic sensrang-ing on one side (the human owner of the mobile wireless

(14)

device is not fully aware of its active applications, as to when and what is provided to external systems) to participatory sensing on the other side (the participants need to continuously control the release of their sensor readings to third party). This spec-trum is discussed in the context of urban sensing systems, where personal mobile phones could be used as data input (e.g. Google Street View), raising a flag as to what are the new security challenges.

This thesis will propose a new PHMS acceptance model with an extended set of variables based on aforementioned theories. The proposed model will be empirically tested using da-ta collected via a survey based on questions from the reviewed literature. Partial Least Squares (PLS) will be used to analyse the data.

D. Structure

As a start, the conceptual background will be presented, in the form of review of the exist-ing literature, focusexist-ing on the main three theories that are goexist-ing to be the basis for the new extended research model. This chapter will present in more depth two additional theories in order for the new proposed acceptance model for PHMS to have the list of variables as complete as possible.

Chapter three will present the research model and develop hypotheses. The method for col-lecting data will be presented in chapter four, including details about the survey design, sample population and the choice of PLS method for data analysis.

Chapter five will present the results, as well as the findings concerning the validity and relia-bility of the data. Finally, chapter six will conclude the key findings of the study, will address the managerial and scientific contribution, and present the imitations and future research suggestions.

(15)

II. Conceptual background

A. Technology Acceptance Model (TAM)

Technology Acceptance Model (TAM) is considered influential and commonly used as sup-port theory for explaining an individual’s acceptance of information systems (Lee et al.,2003) because it suggests a small number of factors which jointly account for usage. The two factors, perceived usefulness (PU) and perceived ease of use (PEOU), are defined as

“the degree to which a person believes that using a particular system would enhance his or her job performance” and “the degree to which a person believes that using a particular sys-tem would be free from effort” (Davis et al., 1989, p. 320), respectively.

PU and PEOU are specific, simple, easy to understand, and can be adjusted through design and implementation for a broad range of technologies and heterogeneous user populations. (Wangpipatwong & Papasratorn, 2008).

TAM is capturing exactly the moment when consumers have to use a new technology, be it forced by job responsibilities or by society’s technical evolution (Banta et al.,1987). By mod-eling and knowing in advance what factors are impacting the acceptance of technology by individuals, organizations can manipulate the communication and timing of the change in order to promote acceptance and therefore commence utilization.

TAM has also been used to predict user’s intention to use an information system after hav-ing a long period of experience with the system, so this model can be applied to understand behavior for both experienced and inexperienced users (Wangpipatwong & Papasratorn, 2008).

Numerous empirical studies have found that TAM consistently explains about 40% of the variance in usage intentions and behavior (Venkatesh & Davis, 2000). TAM’s popularity can be seen by the fact that there were 424 journal citations in the Social Science Citation Index (SSCI) by the beginning of 2000 (Venkatesh & Davis, 2000) and the number reached 698 by 2003 (Lee, Kozar, & Larsen, 2003).

“TAM has been the only one who has captured the most attention of the Information Sys-tems community” (Chuttur, 2009, p. 1) and a set of changes have been applied to it throughout time, “evolving like an organic being” (Lee et al.,2003, p.754). This evolution has been captured in Figure 4.

(16)

Figure 4 Chronological Progress of TAM Research (Lee et al., 2003, p.755)

TAM was first introduced in the academic world by Fred Davis (1985) with his study called “A technology acceptance model for empirically testing new end-user information systems: theory and results”, based on the 10 year old Theory of Reason Action (TRA) developed by Fishbein & Ajzen (1975). TRA’s social psychology view is that attitudes towards a behaviour are determined by personal feelings towards performing the behaviour and subjective norms (Fishbein & Ajzen, 1975).

A thorough comparison was done between TRA and TAM by Davis et al. (1989) and con-cluded that TAM explained better the intention to use information systems and there is lit-tle correlation of the subjective norms to the behavioural intention variables.

Mathieson (1991) chose to do one more type of comparison, between TAM and another theory of Ajzen (1985), Theory of Planned Behaviour (TPB). TPB is TRA enriched with Per-ceived Behavioural Control factor, as seen in Figure 5. For this purpose, Mathieson (1991) designed an experiment to explain the intention to use a computer using each of the two models. The slightly higher accuracy offered by TPB was not enough to overcome the power of TAM, which has a simpler format and is easier to be applied to any system. Replication studies were also done to show TAM’s versatility. For example, Adams, Nelson & Todd (1992) examined TAM in five different applications and found validity and reliability for measurement of PU and PEOU across the different settings.

(17)

Figure 5 TRA and TPB

Other studies showed limitations of TAM and suggested improvements: Adams et al. (1992) suggested to examine the impact of gender and task; Igbaria & Iivari (1995) suggested to include the training and managerial support as determinants for PU and PEOU; Straub et al. (1997) found that national culture played an important role in explaining intention to use information systems.

As an answer to the aforementioned extension proposals coming from the academia, a new millennium version was introduced by Venkatesh & Davis (2000): TAM2, as shown in Figure 6. The novelty was the removal of the “Attitude” component and the addition of more vari-ables, such as subjective norm and output quality, which meant to capture the social influ-ence (e.g. from colleagues or management) in order to compel end users to positively eval-uate and accept IT systems.

Figure 6 Technology Acceptance Model (Davis,1985; Venkatesh & Davis, 2000)

The intention to buy personal goods is followed by the intention to use, but in the case of acquiring new technology for a certain business, the implementation is done before the

(18)

ac-bianca iordachioaiei Page 18

tual use. This is the case with ICT introduction to healthcare industry: hospitals and clinics had to adjust to the increasing demand for better structured data and faster service (Banta H. , 1990). ICT was the answer – and this is how the concepts of eHealth and mHealth were born ( Fieschi, 2003; Hameed, 2003). However, the users of the new technology were yet to adjust: medical stuff and patients alike faced a major challenge having to use this technolo-gy ( Alsos et al.,2012; Davis et al.,2009).

Holden & Karsh (2010) performed a meta-research of all existing TAM and TAM2 studies on healthcare up to 2009 in order to verify that TAM and TAM2 does apply to healthcare. A list of 16 datasets from previous studies was included, with a high heterogeneity: span of 20 years of scientific research in this field, highly variable demographics, and various types of medical care. The results confirmed the two common variances explained in previous stud-ies: the positive effect of perceived usefulness and perceived ease of use towards intention to use. One more important observation is that the attitude variable as a mediator was not present in all studies and this is what prompted me to use the TAM2 setting as the pillar of the future PHMS acceptance model.

B. Usage Intention Model for Electronic Health Services:

A practical study was done by Dünnebeil, Sunyaev, & Blohm (2012) on the request of Ger-many Healthcare Authorities when noticing the adverse reactions of ambulatory medical stuff to the use the newly built nation-wide telemedicine infrastructure. According to the study, even though the German physicians acknowledge the potential advantages of eHealth services, their resistance delayed the implementation of national projects by five years prior to the time of the study. Their proposed model for technology acceptance for e-health in ambulatory care was based on the TAM and brought into discussion more factors through detailed interviews with the German physicians.

(19)

Figure 7 Usage Intention Model for Electronic Health Services (Dunnebeil et al., 2012)

All hypotheses presented in Dünnebeil et al. (2012) were validated with a quantitative anal-ysis based on medical personnel from the Bavarian area, as shown above, Figure 7. Their study concludes that variables like standardization and the current level of IT utilization were the most significant drivers for accepting electronic health services (EHS) in German ambulatory settings and they will be taken into consideration when building the PHMS ac-ceptance model.

C. Service Quality Model for mHealth

The model provided by Akter, D’Ambra & Ray (2013) is a valid and highly reliable instru-ment for measuring the quality of mobile driven services; therefore it will be one of the support theories when building a model for PHMS acceptance.

Akter et al. (2013) dedicated their attention to the service industry involving mHealth, seen this time from the point of view of the patient. The authors pinpointed the two distinctive characteristics of mHealth compared with medicine as we know it and made sure to include them in their study:

1) MHealth is based on mobile technology and its success is connected to the fact that mobile phones reach further into developing countries than any other technology or

(20)

health infrastructure5 - see Figure 8. Therefore the study for the new service quality model was done in Bangladesh, where 44 million people have access to mHealth services provided by the local mobile operators.

Figure 8 Explosive of Mobile Phones in Developing World (Vital Wave Report, 2011)

It can be argued therefore, that the dramatic penetration of mobile phones in low- and middle-income countries will be playing a critical role in reducing the digital di-vide in health care. It can be expected that mHealth will soon transform the face and context of health care service delivery in the developing world by counteracting for the limited number of medical personnel and the difficulties to reach remote areas. 2) Shifts the care paradigm from crisis intervention and treatment to prevention and

self-management.

Figure 9 Service Quality Model for mHealth (Akter et al.,2013, p192)

In the same study, Akter et al.(2013) extend service quality research by “developing and val-idating a higher-order mHealth service quality model on three primary dimensions -system quality, interaction quality and outcome quality- and eight sub dimensions -system reliabil-ity, system efficiency, system privacy, cooperation, confidence, care, and utilitarian and

5

www.vitalwaveconsulting.com/pdf/2011/Vital_Wave_Consulting_mHealth6March09_pdf.pdf (Septem-ber,3, 2013)

(21)

donic benefits of information services” (p.190). Figure 9 emphasizes with quantitative re-sults, that having a good platform, as the world currently has with the mobile infrastructure, is not enough to provide the expected medical service quality. Managers must keep in mind the importance of the interaction and information quality when designing a new mHealth system.

“[…]For instance, perceptions of system quality could be improved by increasing the reliabil-ity, efficiency and privacy of the service system. Likewise, interaction quality could be im-proved by serving customers with a prompt response, adequate knowledge and proper at-tention. In addition, information quality could be enhanced by updating customers on the utilitarian and hedonic benefits of information services (e.g. convenience, cost effective-ness).” (Akter et al.,2013, p. 190)

D. Opportunistic and Participatory Sensing

With the rise of wireless networks and with mobile phones accessing the World Wide Web present in almost every pocket, there is more and more human involvement as part of the sensing infrastructure6. Depending on the degree to which the owner of the sensor can con-trol what to share, the sensing system can range from opportunistic, when the human own-er is not fully in control of what, when and to whom is provided to participatory, when the human owner is continuously in control over the data released by the sensor (Burke et al., 2006).

The study of Lane et al. (2007) provided an evaluation formula to decide under what condi-tions a sensing system would work better in opportunistic or participatory settings. They found that “with higher levels of human involvement, the participatory approach outper-forms the opportunistic. However, as the probability of human cooperation falls, as is likely to happen as sensing queries begin to annoy the user, the opportunistic approach outper-forms the participatory one.” (Lane et al.,2007, p. 4)

On a similar note, Kapadia et al. (2010) argue that even though the system is presented as participatory (e.g. the user has the option to switch off the camera or deny any request), there are a series of security and privacy challenges which make it impossible at the current

6

(22)

time to have an absolute participatory sensing, where the user would have full control over the sensed data. For example, if the user agrees to share with the system photography of the restaurant menu, he might unwillingly share personal information as well: friends’ pres-ence at the table, alcohol price, time and date, etc. The authors present a set of urban sens-ing projects besens-ing currently researched, includsens-ing CarTel, MetroSense, BikeNet, CenceMe, Urban Atmosphere, Urbanet, Mobiscope, and Sense Web. All these systems are collecting large-scale, online data and process context information related to aspects of everyday life – very similar to the function of wearable body monitoring devices. The challenge is to main-tain an acceptable level of data security, and AnonySense is recommended as the best avail-able in terms of security for people-centric opportunistic sensing system.

The security challenges are detailed in three main categories and they will all be taken into consideration in the proposed model for acceptance of wearable body sensors, as per the classification shown in Table I.

Table I Security Challenges in Opportunistic Sensing (Kapadia et al.,2010)

Confidentiality and Privacy issues Integrity issues Availability issues

Challenge 1: Context privacy Challenge 2: Anonymous tasking Challenge 3: Anonymous data reporting

Challenge 4: Reliable data readings Challenge 5: Data authenticity Challenge 6: System integrity

Challenge 7: Preventing data sup-pression

Challenge 8: Participation Challenge 9: Fairness

When combining the pervasive presence of mobile phones with the capabilities of wireless sensor networks, the result is Participatory Sensing: the collection and distribution of data obtained by self-selected participants, like air temperature, traffic conditions, pollution lev-el, and consumer pricing information or medical data (Burke, et al., 2006). The clear security issues when it comes to participatory sensing were identified by Cristofaro & Soriente (2012) as visible risks of privacy loss on the side of data provider, the owner of the mobile sensor, and data consumer. Their proposed solution is Privacy-Enhanced Participatory Sens-ing Infrastructure (PEPSI), which uses efficient cryptographic tools for query and location privacy.

Christin et al. (2011) provide an overview of current participatory sensing applications, em-phasizing their benefits on one side and user privacy threats on the other side. The authors dedicate an entire chapter to evaluate and classify personal health monitoring usage of mo-bile phones as sensors for physiological parameters, as shown in Figure 10. With embedded or external sensors, mobile phones could be turned into: accelerometers, photo cameras,

(23)

temperature sensors, gyroscopes to measure body position, gps to determine location, mi-crophones for sound measurement and/or environmental pollutants detectors.

Figure 10 Medical Sensing Applications (Christin et al.,2011, p. 1934)

All participatory sensing applications, as the examples shown in Figure 10, have four archi-tectural components: sensing, processing, storage and reporting. PHMS, is a subset of par-ticipatory sensing, and therefore the same architecture applies. Christin et al. (2011) identify the specific privacy threats for each of the components and proposes possible solutions. Since is not in the scope of this thesis to go into technical details about system architecture, only a brief overview of their findings will be presented in order to create a clear picture of the variety and complexity of challenges when it comes to protecting privacy in a mobile sensing environment:

 “Privacy in participatory sensing is the guarantee that participants maintain control over the release of their sensitive information. This includes the protection of infor-mation that can be inferred from both the sensor readings themselves as well as from the interaction of the users with the participatory sensing system.” (Christin et al., 2011, p. 1934)

 Privacy threats can be anything that would disclose personal information: time and location, sound samples, pictures and videos, acceleration, environmental data, bi-ometric data and others.

(24)

Figure 11 Mobile Big Data architecture: threats and solutions (Christin et al.,2011, p. 1933)

Possible solutions or countermeasures exist and they are mentioned in Figure 11, but they are yet to be developed to a satisfying level. These include tailored sensing and user prefer-ences, anonymous task distribution (levels of granularity), anonymous and privacy-preserving data reporting, privacy aware data processing, privacy aware data storage, ac-cess control and audit.

The paper of Christin et al. (2011) also discusses at length possible privacy improvements in the area of participatory sensing system. These ideas are also used to identify the factors driving acceptance of pervasive healthcare by the end user:

1. Include the participants in the privacy equation: tailored privacy interface, ease of use, transparency of privacy protection levels, user feedback

2. Provide adaptable privacy solutions: independent to tailored, centralized to distrib-uted system, old to new senses

3. Trade-off between privacy, performance and data fidelity: anonymity vs data quality, multi-party privacy protection, overriding privacy

4. Measurable privacy: generalized privacy measures, provable guarantees for privacy 5. Standards for privacy research: open data sets, open privacy solutions

E. Task – technology – human fit theory

The need to understand the connection between individual traits and technology has been in the attention of information system studies since Goodhue & Thompson (1995). Their pa-per proposed and tested a comprehensive model to highlight the importance of the tech-nology-task fit in order for the end-user to achieve high individual performance.

(25)

The paper goes as far as to suggest that the fit between task, technology and individual characteristics have a higher impact in increased work performance than the actual tech-nology utilization. The entire model is presented in Figure 12, but Goodhue & Thompson (1995) only tested the part highlighted in blue, which is known in the academic world as Technology-to-Performance Model or Task-Technology-Fit (TTF). TTF is defined as “the de-gree to which a technology assists an individual in performing his or her portfolio of tasks. More specifically, TTF is the correspondence between task requirements, individual abilities and the functionality of the technology” - (Goodhue & Thompson, 1995, p. 216)

Figure 12 Technology to Performance Model (Goodhue & Thompson,1995, p. 217)

Goodhue & Thompson (1995) also suggested that that this model could be used to diagnose the quality of the information system present in an organization, and this is an advice that this study on PHMS is going to follow. For environments like healthcare, the fit between task and technology is very important given the required level of accuracy and the potentially fatal consequences of errors.

Fourteen years later, the model was brought back into the foreground, while examining nurses’ decision to utilize or not mobile information communication technologies. This time, Junglas et al. (2009) decided to add the individual characteristics into the TTF model in order to prove the relationship with technology usage and the subsequent increase in work per-formance. This study is of special interest because it focuses on the nurses as the front-line care givers, using a thorough qualitative and quantitative study to analyze the impact of their personality traits on their daily job. The variables and hypothesis are shown below, Figure 13, and managed to explain 70% of variance in nurses’ performance and 56% of the nurses’ decision to utilize mobile system on the hospital premises.

(26)

Figure 13 Task and human drives Fit Model ( Junglas et al.,2009, p 640)

The key findings were that “three dimensions of fits, namely time criticality, user comfort, and workflow fit display a significant influence on overall performance; and identification and workflow fit display a significant influence on utilization.” (Junglas et al.,2009, p. 642).

F. Pervasive Healthcare

With a vast knowledge and research under his belt in the area of Health Informatics and Bio-informatics, Professor Enrico Coiera stated in 2004: “Over the next 20 years, national health systems will have to treat proportionately more people, with more illness, using relatively fewer tax dollars and workers, yet these systems are already under significant strain. To flourish in the coming setting of diminished resources and increased demand […] may re-quire nothing less than the reinvention of health care. […] Since health systems are soci-otechnical systems, where outcomes emerge from the interaction of people and technolo-gies, we cannot design organisational or technical systems independently of each other.” (Coiera, 2004, p. 1197).

Significant advances in communication and sensing technologies have led to the develop-ment of intelligent hand-held and wearable devices, such as smart phones, smart watches, smart clothes and smart homes. Such devices provide a platform to implement a wide range of solutions for health monitoring purposes. Further, their wearable or hand-held nature allows them to be present around the user at all times, making health monitoring a perva-sive activity. “Such technologies that use pervaperva-sive computing capability for health man-agement are called pervasive healthcare systems.” (Gupta et al., 2013, p. 22)

(27)

The power of mobile communications as a disruptive innovation in healthcare industry is underlined by Atienza & Patrick (2011, p. S153), who state that “as electrical lighting was the killer app in the early 1900s that precipitated the rapid development of a comprehen-sive nationwide grid to deliver electricity, […] mobile health (mHealth) may be the killer app for cyber infrastructure for health in the twenty-first century”. The paper also presents facts about the penetration of mobile technology into healthcare: for example, as of 7-Aug-2010, iTunes App Store offers more than 5200 Healthcare and Fitness and more than 3400 Medi-cal iPhone Apps .

PHMS by their very nature alter the traditional treatment model by making it more proac-tive, providing a more patient-centric form of care, as opposed to the reactive caregiver-centric form of today’s medical system (Gupta et al., 2013). Health problems will be detect-ed and addressdetect-ed as soon as they occur, maybe even before any symptom manifests. With-out the need for a personal visit to the doctor, the individual can receive treatment based on a detailed analysis of the personal medical data.

According to Varshney (2007), beside the benefits of the wireless infrastructure in healthcare applications, there are also challenges regarding the implementation and ac-ceptance of this concept. The factors underlined in this paper are going to be the main vari-ables in the proposed model for PHMS acceptance: infrastructure reliability, context aware-ness, and autonomous and adaptable operations.

PHMSes are medical devices tightly coupled to the human body, meant to be worn at all times. Since they are designed for long term monitoring purposes, they require a minimum of energy and employ wireless communications. There are a whole set of human body measurements that can be monitored by these new-generation gadgets: sleep patterns, cardiac activity, food intake, exercise levels, blood chemistry, meta-physiological state (fa-tigue, stress) as wel as external context (location, time, participants).

PHMS can have different forms, as explained by Rocker & Ziefle (2010) in their book “Smart Healthcare Applications and Services”:

 monolithic platform-based solution, a bulky device to be kept at home, which would be used a few times a day to transfer medical data to the hospital. Examples of such projects: LiveNet, AMON and LifeGuard

(28)

bianca iordachioaiei Page 28  textile based solutions, also known as smart clothing, where the sensors are

distrib-uted over a fabric worn by the person making a truly wearable health monitoring system. Examples of such projects: MyHeart, Wealthy, MagIC, SmartVest

 body area network, a set of tiny wireless units forming a network on the entire body and sending the collected data to a central node and then forwarded to a base sta-tion for long term storage and processing. Examples of such projects: CodeBlue, Alarm-Net, Human++, BASUMA, HeartToGo

 mobile phones or wrist bands, also known as smart watches. With the mobile phone being recently considered a medical device, as per Food and Drug Administration (FDA) regulation7, PHMS are expected to become critical health infrastructure, sup-porting Coiera’s vision about the future healthcare. Examples of such devices: FitBit Flex, Nike+ FuelBand, Jawbone UP, Basis B1, BodyMedia Fit Link 8 or Palisades Park, Health PIA (Patrick et al., 2008).

A comparison of the different implementation concepts of PHMSes was done by Ziefle & Rocker (2010) and as seen in Figure 14, more than 80% of respondents would most proba-bly accept such devices, with a preference towards mobile device (nearly 60%, in contrast to 56% for smart clothes, and 51 % for smart homes). Another interesting finding of this study was that younger users prefer a combined device, whereas older users go towards accesso-ries for the current mobile phone.

Figure 14 General willingness to use medical technology (Ziefle & Rocker, 2010)

Based on this literature review by Ziefle & Rocker (2010) and Gupta et al. (2013) the conclu-sion is justified that a model for acceptance of PHMSes has not yet been published. The next chapter will propose a set of variables and hypotheses in an attempt to identify the most

7 http://gregpiche.typepad.com/blog/2011/09/fda-to-regulate-minor-subset-of-smartphone-medical-device-applications.html (Au-gust,28, 2013) 8 http://blog.fixya.com/pr/aug2013/fitness-band-report.html (September,3, 2013)

(29)

significant factors for intention to use health monitoring devices in the form of a smart bracelet.

III. Proposed Research Model and Hypothesis

A. Conceptual Model

In line with the Technology Acceptance Model, the backbone of the model is Perceived Use-fulness and Perceived Ease of Use positively influencing the Intention to use wearable body monitoring systems. These three constructs (Ha, Hb, Hc in Figure 15) are not marked as hy-potheses in the proposed model for PHMS because they were mentioned and proven through extensive studies of TAM (Davis et al., 1989; Venkatesh & Davis, 2000; Adams et al., 1992). The same three constructs were also discussed in other papers related to adoption of technology in medical care (Dunnebeil et al., 2012; Hernandez et al., 2009; Pawar et al., 2012; Holden & Karsh, 2010).

The first additions to the TAM model are the factors influencing Perceived Usefulness and Perceived Ease of Use: personalization, interaction, aesthetics, standardization, and the drive to learn personality trait. Therefore, based on the reviewed literature on technology’s impact in medical care, hypotheses H5, H6, H7 and H8 are proposed and discussed be-low.Another major addition to the TAM model in the newly proposed PHMS acceptance model is the Perceived Security & Privacy (PSP) factor which is hypothesized to have a direct impact on the intention to use wearable body monitoring devices, as well as on the per-ceived usefulness.

(30)

Figure 15 Proposed theoretical model and research hypothesis

A graphical overview of the factors can be found in Figure 15. A brief description of the ex-ogenous variables used for building the PHMS acceptance model is included as an overview in Table II.

Table II Exogenous Factors of PHMS acceptance model

Factor Description

Personalization and Interaction

Elements of the device that give users various opportunities to interact with the system and other users. Ac-cording to (Lee T. , 2005), interaction includes: user control, responsiveness, personalization, and connected-ness.

Drive to learn The need to collect information, examine the environment, make observations about explanatory ideas and _{theories concerning the human body and assess their condition that aids in decision-making concerning} medi-cal intervention. (Junglas, Abraham, & Ives, 2009)

Standardization Standardization is the process of developing and implementing technical standards, with the goal of helping _{with independence of single suppliers, compatibility, interoperability, safety, repeatability, or quality.} Standardization allows consumers to use their new items along with what they already own.

By using standardization, groups can easily communicate through the set guidelines, in order to maintain focus. (Fieschi, 2002)

Aesthetics Visual design qualities that lend a sense of attractiveness or pleasant appearance. This concept is encapsulated _{by codes like ‘cute’, ‘bright’, ‘unique design’, ‘suits me’. (Cyr, Head, & Ivanov, 2006)}

Confidentiality and Privacy

The downside of large-scale, on-line data collection and processing of context information related to aspects of everyday life. The body monitoring device may gather context information which is not part of the main task (e.g. location and timestamp, outside temperature, other devices in proximity, abbreviations or grammat-ical errors offering clues about user’s identity, etc) (Kapadia, Kotz, & Triandopoulos, 2010)

Drive to defend The need to protect oneself reputation, and those that one cares for ( e.g. have proofs for what exactly hap-_{pened, showing that the story is in line with the facts ). This includes the care for personal safety. (Junglas,} Abraham, & Ives, 2009)

System reliability Trustworthiness of data including accurate data readings, high data security , system integrity and system _{availability. (Kapadia, Kotz, & Triandopoulos, 2010)}

Perceived Usefulness (PU) Perceived Ease of Use (PEOU) Perceived Securi-ty&Privacy (PSP) Intention to use wearable body monitoring (PHMS) Drive to learn Drive to defend System reliability Standardization Aestetics Personalization And interaction Confidentiality and privacy H8 H7 H6 H5 H4 H3 H2 H1 Ha Hb Hc

(31)

B. Hypotheses development

All proposed hypotheses are presented below, including the facts referenced in the re-viewed literature which lead to the above proposed PHMS acceptance model.

Various papers on wireless sensing underlined users attention to privacy and security during the decision to be or to not be part of a pervasive sensing environment ( Akter et al., 2013; Burke et al., 2006; Chorppath & Alpcan, 2013; Qiang et al.,2011). The importance of this fac-tor is also noticed by the large amount of academic research focused on technical solutions for increasing privacy protection ( Campbell et al., 2008; Cristofaro & Soriente, 2012; Kapa-dia et al., 2010; Ramanathan & Swendeman, 2012). So therefore I argue:

H1. Perceived Security and Privacy have a positive effect on Intention to Use PHMS

The aforementioned research regarding privacy and security in people-centric systems like PHMS, indicate that perceived privacy and security are determined by other factors, based on which hypothesis H2, H3 and H4 are proposed for the new model.

Data provided by the US National Committee on Vital and Health Statistics in 1999 quanti-fied the current level of medical system’s reliability: “preventable medical errors accounted for 12- 15% of hospital costs, 80% of nurses calculated dosages incorrectly 10% of the time, and 40% of nurses made mistakes more than 30% of the time. Approximately 180000 un-necessary deaths and 1.3 million injuries occurred from inadequate medical treatment” (Fieschi, 2002, p. 86).

The importance of the system reliability in PHMS was raised by Varshney (2007), admitting the need for highly accurate measurements as well as high quality of data transmission and data analysis. Kapadia et al. (2010) defined system reliability not only as accurate data read-ings, but also high data security, system integrity and availability.

Iwaya et al. (2013) reported the lack of reliability in the Brazilian mHealth initiatives and connected it to the poor success of current implemented projects, with a special emphasis in the area of patient monitoring. In order to counteract the possible failures of the sensors, Han et al. (2010) propose a system where expected accuracy and reliability can be achieved by high frequency of readings and redundancy of data.

Therefore, I argue that system reliability as seen by the consumer will have a positive influ-ence in perceived security for using PHMS:

(32)

The impact of human drives in technology acceptance was first referenced in the model of Goodhue & Thompson (1995) as influencing the individual performance level, and then dis-cussed and quantitatively analysed in the patient care setting by the paper of Junglas et al. (2009).

Drive to defend, as used in TTF (Task Technology Fit) theory, is related to the pure human instinct to survive, the instinct to protect oneself and others from threat. The definition was extended to the notion of defending “valued accomplishments whenever we perceive them to be endangered. At the individual level, the drive to defend is activated by perceived threats to one's person, valued objects, status, or beliefs” ( Junglas et al., 2009, p. 639). One issue related to personal safety was raised by Patrick et al. (2008) reminding of the general perception that the use of mobile phones brings a risk of brain tumours, but the pa-per also underlines that this concern is yet to be validated by scientific research. The same paper mentions the discomfort that a mobile phone can create when annoying ringtones are used or an extremely loud conversation is had in public.

An attempt to define a methodology to assess the risk of a healthcare application was done by Smith & Eloff (2002) entitled “Risk Management in Health Care – using cognitive fuzzy techniques (RiMaHCoF)”. This methodology is meant to incorporate human common sense and intuition, along with quantitative analysis.

In the context of healthcare, the authors (Smith & Eloff, 2002) maintain that the drive to de-fend is meaningful when seen as one’s need to counteract perceived medical risks or to maintain a certain reputation among the social circle. Given the special attention on per-ceived security and privacy in wireless sensing technology (Kapadia et al., 2010; Cristofaro & Soriente, 2012; Burke et al., 2006; Smith & Eloff, 2002), it is hypothesized that a person’s drive to defend will have a positive effect on his/her perceived security and privacy, there-fore suggesting:

H3. Drive to Defend has a positive effect on Perceived Security and Privacy

General concerns about confidentiality and privacy have been raised in the literature. Kapa-dia et al. (2010) attempted a classification of security issues by grouping three items under the tag “confidentiality and privacy”: context privacy, anonymous tasking and anonymous data reporting. Some examples of security and privacy concerns include the user’s interest

(33)

to control what type of information is gathered, when and with whom the information is shared and the user’s choice of anonymity within the network. Another such concern is analogous to the user’s control over second-hand smoke, defined as “accidental compro-mise of privacy[…] For example, an application measuring traffic noise might sample a mo-bile phone’s microphone as the custodian stands at a busy city intersection, but the audio sample might also contain fragments of a passerby’s private conversation”. (Campbell et al., 2008, p. 16).s

On a more practical level, Ramanathan & Swendeman (2012) completed an extensive quali-tative survey on two focus groups with regards to acceptance of self-monitoring devices and found that the group of young mothers didn’t see confidentiality and privacy as an im-portant decision factor.

Since the acceptance model for PHMS is referring to a wider population than just young mothers, it is hypothesized that confidentiality and privacy does have a positive effect on overall perceived security and privacy for wearable body monitoring devices:

H4. Confidentiality and Privacy have a positive effect on Perceived Security and Privacy

Based on the work of Pawar et al. (2012), user friendliness and convenience is very im-portant in adoption of a mobile patient monitoring system. Their study presented the in-creased need for personalization and interaction through time, with the various technolo-gies introduced in healthcare. Figure 16 shows how ICT impact in healthcare moved from batch processing in 1950s to telephone based medicine in 1970s to electronic prescriptions in 2010 and personalized predictive healthcarein the recent years.

(34)

Lorenz & Oppermann (2009) went onto analysing which exact part of the user interface mattered for various elderly groups. The findings were that an advanced interface, with lim-ited details and brief summaries about the health status was chosen as number one by half the subjects. The basic interface with only brief summaries ranked second, whereas the op-tion with the most professional display and interacop-tion capabilities was considered accepta-ble only by the respondents familiar with smart phone experience.

Personalization also means configuration of the system’s interface to match the different levels of consumer’s ICT literacy and health understanding (Qiang, Yamamichi, Hausman, & Altman, 2011). The very low end of ICT literacy is more common in developing countries9 and a special attention for the way healthcare monitoring systems could be used by this population as presented by Iwaya et al. (2013) and Karippacheril et al. (2013).

Iwaya et al. (2013) brought this personalization matter in the context of the current chal-lenges, gaps and opportunities in Brazil’s mHealth current implementations: lack of security mechanisms, minimum involvement of health managers, lack of interoperable global stand-ards and the existence of an ubiquitous mobile infrastructure. Karippacheril et al. (2013) showed there are different stakeholders involved in mobile healthcare providers in develop-ing countries which do not have the same power in developed areas of the world: the ser-vice provider and mobile deser-vice manufacturers dominate the mHealth ecosystem.

The different levels of interaction and adaptability of the medical system to the needs of the patient were discussed as a case study for the implementation of technology in UK healthcare, with emphasis on the benefits of flexibility. Service concepts like “walk-in” and “walk-out”, where medical data is gathered in or outside the medical facility are just two of the examples provided by Hameed (2003) .

A more technical approach was presented by Shin et al. (2011) when describing Ano-nySense, an urban sensing project, emphasizing the multitude of reporting and configura-tion capabilities.

A test for interaction and technology was done by Alsos et al. (2012) and showed that in the hospital premises the old paper chart provided more satisfaction given the better perceived

9

(35)

doctor-patient communication compared to scenarios of doctors using a PDA, even though it contained more detailed and easier available patient data.

Satisfaction and trust were found to come from interaction in an “always-on society” (Lee T. , 2005, p. 165), so in line with this reasoning, interaction will be an important factor in PHMS adoption as well.

I argue that a high level of personalization, in the form of reports presenting individual’s monitored health data and the existence of a configurable communication link to the medi-cal personnel or other users, will have a high effect on perceived usefulness for PHMS by the general public, leading to:

H5. Personalization and Interaction have a positive effect on Perceived Usefulness for PHMS

A few early inquiries about technology penetration in medical care were done based on the pure human need to collect information, to examine the environment, to make observa-tions about explanatory ideas and theories concerning the human body. Such studies were conducted by Banta (1990) who interviewed doctors in an attempt to forecast coming changes in hospitals. He used a method developed by himself just three years earlier for “technology forecasting within the broader framework of technology assessment” (Banta et al., 1987, p. 253), involving a less structured survey of expert opinion, using “what-if” types of scenarios. One of the examples hinted the current trend of PHMS: “If by the end of 1997 a host of new home diagnostic tests are going to be marketed, what are the consequences for health going to be?” (Banta et al., 1987, p. 258)

The importance of personality traits in technology acceptance was detailed in two of the main support theories for the acceptance model of PHMS, as presented earlier: the Technology-Fit (TTF) theory of Goodhue & Thompson (1995) and its continuation of Task-Technology-Human-Fit theory of Junglas et al. (2009) .

It is worth mentioning that human drives used as variables in the aforementioned theories are not the same as personality attitudes, which were shown not to impact the future and science of technology: Blind et al. (2001) showed that optimism, pessimism or scepticism have little to do with the technological development of the society.

Pervasive Healthcare Monitoring System (PHMS) : An investigation into the acceptance of PHMS by consumers.