UNIVERSITEIT VAN AMSTERDAM
The influence of obtrusiveness on the intention
to advise the use of a wearable sensor of
general practitioners.
Bachelorthesis Tom Hölscher (10288937)
Begeleiders: Tom van Engers & Peter Weijland /2014
In this study obtrusiveness of wearable sensors is researched. The question “Does obtrusiveness have influence on the intention to advise the use of a wearable sensor of general practitioners? is the main question asked. Obtrusiveness is found to have influence on the decision making process of general practitioners. Functional qualities still prove to be of greater importance to general practitioners. Part of the reason why general
practitioners find obtrusiveness important is uncovered. A low level of obtrusiveness results in better use of a wearable sensor by a patient. Correct use has a positive effect on the quality of the measurements made by a device. The diagnose made by general
Content
1 Introduction
2
2 Scope
3
2.1 Wearable sensors 3
2.2 Market segments of wearable sensors 3
3 Literature review
7
3.1 TAM 7
3.2 UTAUT 8
3.3 Extended Technology Acceptance Model for wearable sensors 9
1.1 Framing obtrusiveness in relation to wearable sensors 9
4 Method
13
4.1 Subjects 13
4.2 Questionnaire 13
4.3 Validation questionnaire 14
5 Results
14
5.1 Results open-ended questions 14
5.2 Results closed-ended questions 17
6 Discussion
25
7 Conclusion
26
8 Reflection, Limitations/future research
26
10 Appendix
32
1 Introduction
In the modern day world the influence of technology on our lives continues to grow. Some technologies are more obtrusive than others but in the end the people or users choose if a particular technology is worth using or not. Several technologies are developed that are able to continuously monitor and make biometric measurement of daily activities. “These techniques have the potential to provide remote health monitoring, at-home screening, and rapid notification of critical events such as heart attacks, falls or respiratory problems” (Chan et al, 2013). Biosensing wearables, or wearable sensors are becoming more and more sophisticated, making it possible to measure a growing amount of clinically relevant information. However, a growing number of researches (Bergmann, Chandaria & McGregor, 2012) suggest that user preferences need to be taken into account in order to be able to design devises that will gain acceptance in both a clinical and a home setting. This acceptance is important because wearable sensors become redundant if the clinicians or patients do not accept a device, and therefore do not use it. Even though there has been this rising interest in wearable sensors, many technologies have not been integrated into clinical care because of the limited understanding of the user-centered design issues. Understanding and then integrating user-preferences is therefore an important step in the evolution of wearable sensors. For the wearable sensors to be accepted in the clinical care it is especially important that current trends in patient and clinician preferences are incorporated at an early stage into the design process of prospective wearable sensors. Bergmann et al. (2012) and Chan et al. (2012) found that important factors of user preferences of wearable sensors are divers. Bergmann et al. (2012) found that patients want wearable sensors to be small, discreet, preferably incorporated in everyday objects and unobtrusive. Chan et al. (2012) found that the system efficiency and reliability as well as unobtrusiveness are essential user preferences for wearable sensors. Obtrusiveness is given as an extension of the Technology Acceptance Model when looking at acceptance of wearable sensors by clinicians. Because clinicians is a broad term and one of the main groups of clinicians that advise patients, in the Netherlands, on using wearable sensors are general practitioners, the choice was made to focus on this group of clinicians. Because Bergmann et al., (2012) showed that obtrusiveness was one of the most important factors in the acceptance of wearable sensors by patients the choice was made to also focus on obtrusiveness in this research. The goal of this study was to analyse the user preferences of general practitioners when looking at wearable sensors, but also to find out if obtrusiveness would be as important a factor as it was for patients in the Bergmann et al., (2012) research. This resulted in the following main question:
Does obtrusiveness have influence on the intention to advise the use of a wearable sensor of general practitioners?
2 Scope
10.1 Wearable sensors
Wearable (bio)sensors are devices that convert data gathered through biological recognition elements into a signal output (Chan, Selvaraj, Ferdosi & Narashimhan, 2013). A lot of different types of wearable sensors exist and a separation is made between ‘in vivo’ and ‘ex vivo’ technologies. In vivo (“within the living”) studies or technologies are those in which the effects of various biological entities are tested on the whole, living organism usually animals, including humans. The opposite, ex vivo, (“out of the living”) means that the study or technology takes place outside an organism. In this research both ‘ex vivo’ and ‘in vivo’ technologies make up the group of ‘wearable sensors’. So the term wearable sensor in this research can mean both on- or in-body accessories. The term wearable sensor refers to sensors that are placed on, or implanted in the body. A wearable sensor in this research is defined as ‘a sensor that is worn on, or implanted in, the body with the goal to obtain clinically relevant information’. Some of the standard types of wearable sensors include measurement of movement (via an accelometer), temperature, moisture, location (via GPS), sound, light, heart rate and variability of the heart rate. Other sensors include GSR (galvanic skin response), ECG/EKG (electrocardiography to record the electrical activity of the heart. EMG (electromyography to measure the electrical activity of the muscles), EEG (to read electrical activity along the scalp) and PPG (to measure blood flow volume).
The market of wearable sensors is split in different segments. It is important to define which segment is targeted and analyzed in this research. According to Hoogendoorn M., (personal interview, 02 July 2014) two large segments in the wearable marketspace are the consumer segment which consists of devices focused on the individual consumer, and the clinical segment which consists of devices that provide a supporting role for healthcare professionals. Healthcare professionals are defined as qualified clinicians such as physicians, nurses, general practitioners or any other kind allied health professionals. The clinical segment is a complicated segment. This is because the use cases are specialized and demand a very high validity and accuracy unique per use case (Hoogendoorn, personal interview 02 July 2014). In this research the focus will be on wearable sensors used in the clinical segment
10.2 Market segments of wearable sensors
Individual Wellness
Health monitoring wearable sensors or activity trackers are biosensors meant for consumer individuals. These general purpose devices require accurate measurements but the demands are less strict compared to clinical wearable sensors for healthcare purposes. The
Example
The Jawbone UP24 is a wristband that connects wirelessly to a smartphone by using BLE (Bluetooth Low Energy radio). It tracks a user’s everyday activity, measuring movement and food intake. It lets the user set goals and then positively reinforced the user to reach those goals. Measurements are automatically synced to the user’s smartphone after which they are processed in the jawbone app and insights are given.
Clinical segment
The clinical segment can be split in two. This is because wearable sensor can be used in two different ways in healthcare. Due to the very strict demands of healthcare professionals that work in the second line of healthcare the actual use of wearable sensor in the clinical sector is still in its infancy. Extreme accuracy and validity is demanded as well as an approval of a regulatory instance like de ‘FDA’. In the first line of healthcare (i.e. General practitioners, dentist, ER, etc.) the healthcare professionals will be able to use wearable sensors to monitor their patients. These healthcare professionals will decide for themselves if a patient needs to visit a hospital or a wearable sensor.
Example
The Vital Connect patch sensor consists of two main components: a disposable adhesive patch that houses the ECG and battery, and a reusable electronics module that houses the embedded processor, tri-accelerometer, Bluetooth Low Energy transceiver (Chan et al, 2013). The patch has a typical wear cycle of three days, and continuous remote monitoring of heart rate, respiration, activity, skin temperature, stress, sleep staging, steps and body posture including fall detection/severity is
possible even if the patch is placed in a slightly different location. “These techniques have the potential to provide remote health monitoring, at-home screening, and rapid notification of critical events such as heart attacks, falls or respiratory problems” (Chan et al, 2013).
A smart contact lens that measures glucose levels in tears is being tested by Google (Armstrong-Smith, 2014). Built into the lens, embedded between two layers of soft contact lens material, is a miniature wireless chip and glucose sensor that can, in prototype
4
Figure 1. Jawbone UP24
Figure 2. Disposable Vital Connect patch sensor and reusable electronics
form, generate a reading every second. The aim of the technology is to help diabetic patients managing their disease and a built-in early warning system that will be able to inform the wearer when a glucose threshold is reached is being investigated.
Corporate Wellness
The Corporate wellness segment can be split in two different segments. Both segments exist because of a company’s care towards their employees. The first segment is an extension of the individual wellness segment. A company can motivate their employees to live healthier and, as an incentive, give its employees activity trackers. This segment is named ‘employee wellness’. The second segment is focused more on the safety of the employees and can be called ‘external threat wearables’. These wearable sensors are used for detection of external threats. A threshold is defined. If the threshold is reached a warning follows after which the wearer can bring him- or herself to safety.
Hard hat detect Carbon Monoxide
The exhaust form gasoline-powered hand tools in enclosed spaces places construction workers at risk from carbon monoxide poisoning. A wearable has been developed by the group of Forsyth (2014) that incorporates a sensor into a hard hat. The work takes a pulse oximetry sensor that is used for monitoring O2 and simply adapts it by differing in the number of wavelengths of light that are employed. This way the sensor is able to measure carbon monoxide and will warn the wearer of imminent carbon monoxide poisoning.
Professional sports
A wearable sensor made for sport applications are highly sophisticated devices. The standards for accuracy and validity in the world of professional sports are as demanding as in the clinical sector. The difference is that there is no regulatory bureau approval necessary for sensors meant for professional sports. This is why a lot of experimental techniques and prototypes are tested to enhance training and athletes alike Sometimes a device that proves successful in the professional sports segment is later (after approval) used in the healthcare sector as well.
Subcutaneous (beneath the skin) implantation sensor
At the London Imperial College, UK, A team is involved in an Elite Sport Performance Program ESPRIT (Lo & Yang, 2005). ESPRIT’s objectives
include success at Olympic and other international sporting events and the application of achievements in sport to technological transformations in healthcare. This team has developed oxygen, glucose and lactate sensors for subcutaneous implantation and short term monitoring during
5
Figure 4. Hard Hat Detect Monoxide System
exercise. According to Vadgama, on the members of the ESPIRIT team “The sensors are based on stable membrane components that allow devices to operate inside the body with safety and stability; we have performed human trials on these devices. We have also devised solid-state sensors from metal ions that can be used on the skin for sweat analysis during training” (Armstrong-Smith, 2014). A similar kind of sensor is developed by Joe Wang’s group (Withworth, 2013). Their device uses electrochemical reactions directly on the wearer’s skin. The biosensor is applied to the skin as a temporary tattoo and measures lactate levels in sweat.
Progressive wearable sensor
The amount of wearable biosensors keeps increasing and the complexity of the devices with it. According to some of the leading researchers in the biosensor field, like Joe Wang the next step forward will come in the form of self-powered sensors. Wang is researching this step and is developing a biofuel power source that makes biosensors independent of external power supplies like batteries. The self-powering units borrow the idea of the temporary tattoo (the sport sensor) but utilize the lactate in sweat for energy production. Wang is working with Evgeny Katz to apply the biofuel cell-powered biosensor concept to drug delivery for pain relief (Zhou M., 2013).To achieve this a “nanopharmacy” is activated by injuring biomarkers (excess lactic acid) triggering the release of a drug from one of the electrode. The work is funded by the Office of Naval Research, with the goal of developing a complete nanopharmacy that will
11 Literature Review
11.1 TAM
Davis (1989) and Davis et al. (1989) developed the TAM model which is an adaptation of the Theory of Reasoned Action (TRA) (Ajzen and Fishbein 1980; Fishbein and Ajzen, 1975) as cited in Davis, Bagozzi, & Warshaw (1989). TAM was developed because Davis et al (1989) believed that more fundamental variables, that could be used as determinants when researching computer acceptance, had to be found. Additionally, it was important that TAM could be used for both predicting and explaining. This way both researchers and practitioners could identify failures in a system and take appropriate corrective steps (Davis et al. 1989). So the goal of TAM was to develop better constructs that could predict and explain the acceptance of technology. The technological acceptance model explains user acceptance of a technology based on perceptions of the user (Davis, 1989; Davis et al, 1989). TAM assumes that two particular beliefs that are of primary relevance for technology acceptance exist. The two beliefs are Perceived Usefulness (‘U’) and Perceived Ease of Use (‘E’). ‘U’ is defined as the “extent to which a person believes that using a system will enhance his or her job performance”, ‘E’ is defined as the extent “to which a person believes that using a system will be free of effort” (Davis 1986; Davis 1989). Information systems (IS) have been researched to great extent. Several efforts in developing and testing models that could help predict system use have been taken (Chau, 1996; Cheney, Mann, & Amoroso, 1986). Among these models TAM is referred as one of the more robust models. Lee et al. (2003) thinks of TAM as one of the most influential models that explains the acceptance of technology. In the TAM several relations can be found. ‘U’ and ‘E’ are both influenced by external variables. ‘U’ en ‘E’ both have a direct effect on the attitude towards using (‘A’). The ‘A’ has a direct effect on the behavioral intention to use (‘BI’). ‘E’ has an extra effect on ‘U’ and ‘U’ has a direct effect on ‘BI’. ‘BI’ both influenced by the relation with ‘U’ and ‘A’ will determine the actual system use. All these relation can be seen in figure 1.
11.2 UTAUT
Explaining user acceptance of new technology is often described as one of the most mature research areas (Venkatesh, Morris, Davis & Davis, 2003). The TAM model (Davis, 1989) has been extended in multiple different researches. The Unified Theory of Acceptance and Use of Technology (UTAUT) developed by Venkatesh et al. (2003) sought to combine the most prominent models in user acceptance of technology. UTAUT, developed after reviewing eight prominent user acceptance models used the theory of reasoned action (TRA)
(Ajzen and Fishbein 1980; Fishbein and Ajzen 1975), technology acceptance model (TAM),
the motivational model (MM) (Davis, Bagozzi, & Warshaw, 1992), the theory of planned behavior (TPB)( Ajzen, 1991), a combined model incorporating TAM and TPB (C-TAM-TPB), model of PC utilization (MPCU) (Thompson, Higgins & Howell, 1991), innovation diffusion theory (IDT) (Rogers, 1995), social cognitive theory (SCT). UTAUT integrates elements of all eight models. The four key constructs in UTAUT are performance expectancy, effort expectancy, social influence and facilitating conditions. These four constructs influence the behavioral intention to use a technology and/or the actual use of the technology. Performance expectancy measures the degree to which a person believes that using the system could help improve his or her performance and is comparable to the usefulness determinant in the TAM. Effort expectancy measures to which extent a person believes the technology will be easy to use and is comparable with the ease of use determinant of the TAM. Social Influence measures to which extent a person is influenced by his or her social environment. Thus, to what extent he or she is influenced by opinions of others that he or she cares about and think he or she should use a technology. Facilitating conditions measure to which extent a person believes an organization has the facilities to assist him or her in the usage of the technology. In the UTAUT four moderating conditions are considered as well. These are gender, age, experience and voluntariness of use (Venkatesh et al., 2003). Gender is considered because it could for instance influence the role of social influence on men and women. Research shows that women are more susceptible to opinions of others (Venkantesh, Morris & Ackerman, 2000; Venkatesh & Morris, 2000). Venkatesh & Morris (2000) also provided evidence that demonstrated the role of experience could moderate the influence of social influence. More experience results in social influence having a smaller effect. Age could be of influence the role of facilitating conditions. Research shows that older people require or like to get more assistance in their tasks (Hall & Mansfield, 1975).
11.3 Extended Technology Acceptance Model for wearable sensors
TAM research has often focused on corporate information technology as well as professional users, often employees, who rely on the analyzed technology to complete their jobs. Because of this, TAM pays exclusive attention to the two cognitive constructs, perceived usefulness and perceived ease of use and is focused on the more utilitarian determinants of users’ intention to accept and to actually use a technology. The acceptance of a wearable technology is certainly influenced by these two constructs. Wearable sensors must
relevant information (Bergmann et al., 2012). However, users’ decision to accept and use a wearable sensor is possible equally influenced by obtrusiveness. Research showed that ‘obtrusiveness’ could be an important factor to consider when developing a wearable sensor (Bergmann & McGregor, 2011). According to Chan, Estève, Fourniols, Escriba, & Campo, (2012) system efficiency, reliability, and unobtrusiveness are essential and wearable sensors must therefore have software and hardware that employ an excellent functionality, efficiency and reliability as well as a high level of unobtrusiveness (Daniel, Cason & Ferrell, 2009).
11.4 Framing obtrusiveness in relation to wearable sensors
While obtrusiveness, by many researches, is seen as an important criterion in the acceptance and success of a wearable sensor. The term obtrusive is often not well defined or is used inconsistently. Obtrusiveness is often interchanged with intrusive and in most cases implies that a wearable sensor generally has to be user-friendly and accessible (Hensel, Demiris & Courtney, 2006). It is important for this research to clearly define what obtrusiveness is in relation to wearable sensors. Based on a model developed by Hensel et al. (2006) a frame work of perceived obtrusiveness by the user is presented. Hensel et al. (2006) define eight dimensions that influence the user perception of obtrusiveness (figure 2).
Figure 8. Dimension of obtrusiveness
dimension includes ‘perception of lack of usefulness’, which is as described earlier an important cognitive construct of the TAM on its own. The usability dimension is also already used as a construct in the TAM. Because obtrusiveness is in seen as a possible extension of the TAM in this research and to avoid using constructs double in the TAM, some dimensions and properties will be excluded from the definition of obtrusiveness of wearable sensors in this research. The most important dimensions and/or properties of obtrusiveness in relation to wearable sensors are chosen. This choice is based on studies mentioning this form of obtrusiveness. If a dimension is described frequently in different articles, making it possible to define a dimension more clearly, the dimension will be added to the definition of obtrusiveness in this research.
Physical dimension
The physical aspects of a technology that are associated with obtrusiveness. Its effect on the user and his or her (home)environment (Bergmann & McGregor, 2011; Haahr, Duun, Thomsen, Hoppe, & Branebjerg, 2008)
- Excessive noise
- Incompatible aesthetics
- Obstruction or impediment of space
Privacy dimension
Invasion of the privacy of a user has been identified as a potential barrier to acceptance of a technology. The importance of confidentiality of patient data will be growing as the information measured and stored will become increasingly personal (Abascal & Nicolle, 2005; Leino-Kilpi,Välimäki, Dassen, Gasull, Lemonidou, Scott & Arndt, 2001)
- Invasion of personal information
- Violation of personal space
Self-concept dimension
Wearable sensors can enhance a users’ life and increase independence. If this is the case research shows that the acceptance of technology, and/or the negatives of a technology increases drastically (Steele, Lo, Secombe, & Wong, 2009). At the same time a user can feel that the fact that he or she needs to use a wearable sensor is a symbol of loss of independence (Tamm, 1999). If a wearable sensor is not discrete or conspicuous a user could feel embarrassed and even stigmatized (Mynatt, Melenhorst, Fisk, & Rogers, 2004)
- Symbol of lost independence
- Cause of embarrassment
Wearable sensors may affect the daily routines of a user or result in the need to develop extra routines. Research shows that a low impact on daily routine has a positive effect on acceptance of wearable sensors as well as a positive effect on the quality of measurements (Frail, Bajo, Corchado & Abraham, 2010). This increase in the quality in measurements is due to the fact that the less the daily routine is changed the more accurate the measurements will represent a user.
- Interference with daily routines
- The need to develop new, extra routines.
These four chosen dimension result in a model that shows the factors that influence obtrusiveness in relation to wearable sensors. Figure 3 shows the representation of important variables that influence the perceived obtrusiveness in relation to wearable sensors.
12 Method
Recent studies showed that obtrusiveness is an important factor in the acceptance of wearable sensors by patients. The studies promoted that developers should consider their target user group at an early stage in the design process. In the Netherlands, general practitioners advise their patients on using a wearable sensor. To obtain relevant information from general practitioners and to investigate their user preferences on wearable sensors a questionnaire was developed and validated. The goal of this questionnaire was to obtain relevant information on the view of general practitioners on obtrusiveness of wearable sensors.
12.1 Subjects
The target group of this study consisted of general practitioners of all ages. A criteria, however, was that the general practitioners that participated had a practice of their own. No further criteria were selected. Paper copies of the questionnaire were dropped at the front desk of a ‘huisartsenpost. A ‘huisartsenpost’ is a place where general practitioners gather during the night. This ensures that there are doctors available during the night, in the region of the huisartsenpost. The assistant working behind the front desk offered the questionnaire to general practitioners that entered the building. The assistant did this for a period of two weeks. A personal interview with a general practitioner was done before the questionnaires were given to the participants. A recapitulation of the results, also with an general practitioner (dr. H.L. Hölscher) was done after the results produced by the questionnaires were analyzed.
12.2 Questionnaire
The questionnaire was designed as a self-complete questionnaire. The questionnaire was split into four sections. The first section explained the goal of the study and the reason why the cooperation of the general practitioners was needed. In the second section the participant was thanked for his or her cooperation and the term ‘wearable sensor’ was defined and explained using examples of existing wearable sensors. The third section was used to explain the different dimensions of ‘obtrusiveness’ in relation to wearable sensors. Knowledge about these dimensions was needed to answer some of the questions in section four. The fourth section consisted of several open ended questions and several questions where the participants were able to answer on a 1 to 5 Likert-scale, with 5 begin the highest (See Appendix A of this thesis for the complete questionnaire). Although according to Williams (2003) and Crawford, Coupe & Lamias (2001) general questions should be asked before specific and open-ended questions in a questionnaire. The choice still was made to ask the general practitioners specific open-ended questions in the beginning of the questionnaire. This was done to ensure that the general practitioners had formed a good idea of what they thought was a wearable sensor before answering the more general questions in questions 4
12.3 Validation of the questionnaire
The questionnaire of this thesis incorporated several identified user issues found by Bergmann et al., (2011) as well as questions that were used in the questionnaire used by Bergmann et al., (2012). These questions were validated and found unambiguous, appropriate and acceptable and were reviewed and complemented by experts on wearable sensors, a medical doctor, a musculoskeletal therapist and three occupational therapists (Bergmann et al., 2012). To ensure that the questionnaire used in this thesis was focused on the general practitioners instead of patients, the questionnaire was subsequently tested. A general practitioner reviewed and complemented the questionnaire; ensuring that the questions asked and information given were clear and understandable for the target group.
13 Results
A total of 14 general practitioners participated in this research. It has to be noted that out of every 10 general practitioners that were offered the questionnaire, an average of 4 returned the questionnaire fully completed. This means that 40% percent of the total number questionnaires given to participants were returned. This could be due to several different reasons. The exact reason for this is unclear. The precise reason is unclear. The fourteen returned questionnaires where all fully completed and could therefore all be used.
13.1 Response to open-ended questions
Question 1. “Have you ever advised the use of a wearable sensor to a patient? If so, which wearable sensor did you advise to use?”
Figure 10. Word cloud generated for question 1
There are three types of wearable sensors that were mentioned often. The ‘24-uurs-RR-meting’ is a device that measures the blood pressure of a patient in a time span of 24 hours. The Holter-meting refers to the Holter-monitor which is a device that measures electrical ECG signals from the heart. An Event-recorder is a portable device that tape-records a
for a period of 24- or 48 hours. The Event-recorder is often used when the Holter-monitor does not produce clear and/or useful results. Event-recording means that the portable device only measures when a particular event (i.e. heart rhythm disorder) occurs. The 24-hours-RR-measurement was mentioned by a 83%of the participants, the Holter-monitor by 60% and the event recorder was mention by a little more than 40% of the participants.
Question 2. “Could you give an example of a situation, syndrome or reason in which you would feel it would be useful to advise a patient to use a wearable sensor?”
Figure 11. Word cloud generated for question 2
Most of the more frequent mentioned terms relate to the wearable sensors the general practitioners already used themselves. In question 2 another sensor frequently mentioned is ‘stappenteller’ which is a lifestyle device that measures the amount of steps a person takes during the day. This is seen as important because the general practitioners saw importance in ‘awareness’. Patients’ becoming more aware of their own lifestyle was seen as a positive development. The most prominent and most mentioned term is ‘Witte-Jassen-Syndroom’ or the ‘White coat syndrome’. The White Coat Syndrome is a widely recognized term. It is a phenomenon in which patients’ exhibit elevated blood pressure in a clinical setting, but not in another setting, for instance a home setting. This is thought to be caused by anxiety or stress from being in a clinical environment. The white jacket syndrome was mentioned by 64% of the participants.
Question 3. “We assume a syndrome, situation or reason like the ones in question 2 are present. What would stop you from advising the use of a wearable sensor even though the right reasons, syndromes or situations are present?”
Figure 12. Word cloud generated for question 3
A substantial amount of participants thought that there were no reasons to not advise the use of a wearable sensor to a patient. This was however, in a situation when the use of a wearable sensor was needed to treat the patient as best as possible. In this case, objections from the patient are negated by the general practitioner and experience of the participants shows that the patient in the end will agree with the use of the wearable sensor. It is noted that the participants cannot see if the patient uses the device correctly ones he or she leaves the practice. This is why the participants also often mentioned ‘usability’, ‘ease of use’ and ‘discomfort’. When the device is obtrusive to a patient, changes are that the patient will not use the device at all, or use the device in a faulty or incomplete manner. The patient is mentioned frequently as well. The patient is naturally taken into account when deciding on whether he or she needs a wearable sensor. When the situation absolutely demands the use of a wearable sensor and a general practitioner strongly advises its use, in almost every case, a patient will accept. In other less urgent situations the wish of the patient can be used not to advise the use of a wearable sensor.
13.2 Response to closed-ended questions.
Results of question 4
“To what extent each participant would way each dimension of obtrusiveness in his or her decision to advise the use of a wearable sensor to a patient”. Please note that this question does not further specify in was situation a wearable sensor is advised to the patient. This means that the answers given depend on the own interpretation of the circumstances of the participants.
Statistics
Physical dimension Privacy dimension Self-concept dimension Routine dimension N Valid 14 14 14 14 Missing 0 0 0 0 Median 3.50 2.00 2.00 2.00 Mode 4 2 1 2 Percentiles 25 3.00 2.00 1.00 2.00 50 3.50 2.00 2.00 2.00 75 4.00 4.00 4.00 3.00
Table 1. Results of participants providing a response on a 5-point Likert scale to the questions that relate to dimensions of obtrusiveness.
Table 1 provides information on which dimension of obtrusiveness was considered most important by the participants. Most prominent are the results of the physical dimension. With a median of 3.5 which is 1.5 points higher that all the other dimensions and a mode of 4. The self-concept dimension has the lowest mode of 1 and a score of 1 at the 1st quartile. The
Table 2. Kendall's Tau-b
In table 2 the results of a Kendall Tau-b correlation test are shown. Two significant positive correlations are found.
- Privacy dimension and self-concept dimension with a positive correlation coefficient of 0.582 which is significant at the level of 0.05 (2-tailed).
- Routine and self-concept dimension with a positive correlation coefficient of 0.687 which is significant at the level of 0.01 (2-tailed).
Table 3. Spearman's rho
In table 2 the results of a Spearman’s rho correlation test are shown. Two significant positive correlations are found
- Privacy dimension and self-concept dimension with a positive correlation coefficient of 0.659 which is significant at the level of 0.05 (2-tailed).
- Routine and self-concept dimension with a positive correlation coefficient of 0.781 which is significant at the level of 0.01 (2-tailed).
In the graphs below the distribution of the responses per response category are displayed in a bar chart. Each bar represents a response category, so to what extent a participant weighed a particular dimension in his or her decision making process (weight lightly, weight average, weight heavily).
In the graphs of the privacy- and the self-concept dimension there are no clear ‘winners’ or ‘losers’ between any of the response categories. The graphs show that the participant found the physical dimension the most important as 50 % of the participants chose the ‘weighs heavily’ category for this dimension. The graphs also show that the participants found the
routine dimension the least important as the ‘weighs lightly category’ was chosen by 50% of the participants for this dimension
Results question 5.
The participants were asked to rank the importance of several properties of a wearable sensor on a 5 point Likert Scale, with 5 being the highest.
Table 4. responders providing a response on a 5-point Likert scale to the questions that relate to the properties of a wearable sensor. Participants were asked on a Likert-scale of 1 to 5 how important the following properties were, with 5 being the highest
In table 4 the median, mode and quartiles of the result of question 5 are shown. The participants had to rank whether they thought a wearable sensor should have either of these properties. Were 1 meant that a property was not important at all to the participants and 5 meaning it was very important to the participants. Several properties that could be related to obtrusiveness like ‘be comfortable’, ‘be compact’, and ‘be discreet’, ‘easy to apply’ scored high. With medians and modes of 4, where ‘be comfortable’ even had a mode of 5. The best scoring property can be related to functional qualities. ‘Provide clear and useful results’ scores best with a median and a mode of 5.
In the graphs below the distribution of the responses per response category are displayed in a bar chart. Each bar represents a response category (disagree, agree, strongly agree, etc).
14 Discussion
A range of questions were asked to uncover if obtrusiveness influences the decision making process of a general practitioner. The open-ended questions were asked to see what kind of experience the participants had with wearable sensors. The open-ended questions were also asked to find out what reasons to advise the use of a wearable sensor the participants thought of as important. The closed-ended questions were asked to find out what dimension of obtrusiveness the participants weighed heaviest in the decision making process of the participants. The closed-ended questions were also asked to find out if the participants considered obtrusiveness, in relation to wearable sensors, important as well as to further analyze the user preferences of the general practitioners. Several more prominent results occurred. General practitioners are most familiar with the Holter-monitor and 24-hours blood pressure measurement. This meant that wearable sensors used by general practitioners at the moment, are commonly used to either measure ECG of the heart, to check up on palpitation or to find heart fluttering. As well as, blood pressure measurement that has to be done over a longer period (like 24 hours). The participants found ease of use, usability and discomfort the most important reasons not to advise the use of a wearable sensor. The preference of the patient is an important factor in the decision making process. The patients do tend to accept the wearable sensors when their general practitioner strongly advises them to because the use of the device will result in a better treatment. Another important finding is that the participants mentioned the ‘white jacket syndrome’. The white jacket syndrome is a phenomenon in which patients exhibit elevated blood pressure in a clinical setting, but not in another setting. This is thought to be caused by anxiety or stress from being in the clinical environment. The use of wearable sensors, according to the participants, is one of the solutions to circumvent the white jacket syndrome. Properties that could be related to obtrusiveness scored high. Giving reasons to assume that obtrusiveness is considered by general practitioners in their decision making process. The results of question 5 also show that although the properties that could be related to obtrusiveness score very high, the highest scoring properties are the ones that could be related to functional demands like ease of use, usability and accurate and valid measurements.
15 Conclusion
This study produced three findings in regard of the question “Does obtrusiveness have
influence on the intention to advise the use of a wearable sensor of general practitioners?”
Firstly, this study promotes that general practitioners do think obtrusiveness is important. Obtrusiveness therefore does have influence on the intention to advise the use of a wearable sensor of general practitioners this study showed that the patients preferences are definitely considered by general practitioners.
Secondly, this study shows that this not the most prominent reason why general practitioners find obtrusiveness important. If a general practitioner strongly advises the use of a wearable sensor a patient will almost always accept. The problem of obtrusiveness for general practitioners comes to play after the patient has left the practice. The more obtrusive a device, the higher the chance that a patient will use a wearable sensor incorrectly or in an incomplete manner. This behavior negatively influences the accuracy and validity of the measurements made by the sensor. The most prominent reason why general practitioners find obtrusiveness important is because a low level of obtrusiveness leads to better use of the sensor by the patient. And better use means better measurements, with these better measurements a general practitioner can produce a better diagnose and thus provide better care.
Thirdly, this study showed that both functional qualities (ease of use, usability) and obtrusiveness are found important by general practitioners. But when a tradeoff has to be made between functional qualities and obtrusiveness, when considering wearable sensor to advise a patient, a general practitioner would prefer the device with the best functional qualities. So, a device of which the measurements are most accurate and most valid.
16 Reflection/ limitations and future research
This study is one of the first studies that researched the user preferences in relation to wearable sensors of clinicians. The generalizability of this study is limited due to the small amount of participants. In future research a bigger group of participants should be gathered. The participants of this study were only familiar with wearable sensors that have proven to be reliable and they used themselves. But these wearable sensors are also much sophisticated than the generation of devices that are being developed at this point in time. It would benefit further research to define more clearly, for the participants, what the near future of wearable sensors holds. This way the participants will have a broader view of wearable sensors than the two devices mentioned most in the questionnaires (Holter-monitor and 24-hours RR measurement). This study shows that obtrusiveness is considered important, what dimension of obtrusiveness is considered important and which properties of wearable sensors are
considered more important than others. Future research must tell to what extent each dimension of obtrusiveness and each of the properties of wearable sensors actually influences the decision making process of clinicians. If this is done, assumptions whether the TAM can be extended with obtrusiveness can be made. In this study the results were not comprehensive enough to support the extension of TAM with obtrusiveness. One of the reasons for this was that obtrusiveness turned out to be a broader construct than ease of use or usability. Looking back it might have been a better choice to use the UTAUT model instead of the TAM because the UTAUT already employs the constructs ´social influence´ and ´facilitating conditions´. In future research a clearer connection of to what extent the dimensions of obtrusiveness influences the decision making process of general practitioners has to be made.
17 References
Abascal, J., & Nicolle, C. (2005). Moving towards inclusive design guidelines for socially and ethically aware HCI. Interacting with Computers, 17(5), 484-505.
Ajzen, I., (1991), The Theory of Planned Behavior," Organizational Behavior and Human Decision Processes (50:2), pp. 179-211.
Anliker, U., Ward, J. A., Lukowicz, P., Troster, G., Dolveck, F., Baer, M., ... & Vuskovic, M. (2004). AMON: a wearable multiparameter medical monitoring and alert
system. Information Technology in Biomedicine, IEEE Transactions on,8(4), 415-427. Armstrong-Smith I., (2014), ‘You wear it so well’, The Analytical Analyst.
Bergmann, J. H., Chandaria, V., & McGregor, A. (2012). Wearable and implantable sensors: The patient’s perspective. Sensors, 12(12), 16695-16709.
Bergmann, J. H. M., & McGregor, A. H. (2011). Body-worn sensor design: What do patients and clinicians want?. Annals of biomedical engineering, 39(9), 2299-2312.
Chau, P. Y. K. (1996). An empirical assessment of a modified technology acceptance model. Journal of Management Information Systems, 13(2), 185–204.
Chan A.M., Selvaraj N., Ferdosi N. and Narashimhan R. (2013) ‘Wireless Patch Sensor for Remote Monitoring of Heart Rate, Respiration, Activity, and Falls’, 35th Annual International
Conference of the IEEE EMBS, Japan.
Chan, M., Estève, D., Fourniols, J. Y., Escriba, C., & Campo, E. (2012). Smart wearable systems: Current status and future challenges. Artificial intelligence in medicine, 56(3), 137-156.
Cheney, P. H., Mann, R. I., & Amoroso, D. L. (1986). Organizational factors affecting the success of end-user computing. Journal of Management Information Systems,
3(1), 65–80.
Compeau, D. R., & Higgins, C. A. (1995b). ComUter self-efficacy: Development of a measure and initial test. MIS quarterly, 189-211.
Crawford, S. D., Couper, M. P., & Lamias, M. J. (2001). Web surveys perceptions of burden. Social science computer review, 19(2), 146-162.
Daniel, K. M., Cason, C. L., & Ferrell, S. (2009). Emerging technologies to enhance the safety of older people in their homes. Geriatric Nursing, 30(6), 384-389.
Davis, F. D. (1986). A technology acceptance model for empirically testing new end user information systems: Theory and results. Doctoral dissertation, Sloan School of Management, Massachusetts Institute of Technology, Cambridge, MA.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS quarterly, 319-340.
Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of comUter
technology: a comparison of two theoretical models. Management science, 35(8), 982-1003. Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1992). Extrinsic and intrinsic motivation to use computers in the workplace. Journal of applied social psychology, 22(14), 1111-1132. Fraile, J. A., Bajo, J., Corchado, J. M., & Abraham, A. (2010). Applying wearable solutions in dependent environments. Information Technology in Biomedicine, IEEE Transactions on, 14(6), 1459-1467.
Forsyth J.B., et al, (2012), ‘Feasibility of Intelligent Monitoring of Construction Workers for Carbon Monoxide Poisoning Automation Science and Engineering’, IEEE Transactions, 57, 80
Gandhi M. and Wang T., (2014), ‘The Future of Biosensing Wearables’, Rock Health Report. Haahr, R. G., Duun, S., Thomsen, E. V., Hoppe, K., & Branebjerg, J. (2008, June). A
wearable “electronic patch” for wireless continuous monitoring of chronically diseased patients. In Medical Devices and Biosensors, 2008. ISSS-MDBS 2008. 5th International Summer School and Symposium on (pp. 66-70). IEEE.
Hall, D. T., & Mansfield, R. (1975). Relationships of age and seniority with career variables of engineers and scientists. Journal of Applied Psychology,60(2), 201.
Hensel, B. K., Demiris, G., & Courtney, K. L. (2006). Defining obtrusiveness in home telehealth technologies a conceptual framework. Journal of the American Medical Informatics Association, 13(4), 428-431.
Hoogendoorn M., personal interview, 02 July 2014. Hölscher H.L., review questionnaire , 20 July 2014.
Lee, Y., Kozar, K.A. and Larsen, K.R.T. (2003), “The technology acceptance model: past, present, and future”, Communications of the AIS, Vol. 12 No. 1, pp. 752-80.
Leino-Kilpi, H., Välimäki, M., Dassen, T., Gasull, M., Lemonidou, C., Scott, A., & Arndt, M. (2001). Privacy: a review of the literature. International journal of nursing studies, 38(6), 663-671.
Lo, B., & Yang, G. Z. (2005, April). Key technical challenges and current implementations of body sensor networks. In Proc. 2nd International Workshop on Body Sensor Networks (BSN 2005).
Mynatt, E. D., Melenhorst, A. S., Fisk, A. D., & Rogers, W. A. (2004). Aware technologies for aging in place: understanding user needs and attitudes. Pervasive Computing, IEEE, 3(2), 36-41.
Rogers, E. M. (2010). Diffusion of innovations. Simon and Schuster.
Steele, R., Lo, A., Secombe, C., & Wong, Y. K. (2009). Elderly persons’ perception and acceptance of using wireless sensor networks to assist healthcare. International journal of medical informatics, 78(12), 788-801.
Tamm, M. (1999). What does a home mean and when does it cease to be a home? Home as a setting for rehabilitation and care. Disability & Rehabilitation, 21(2), 49-55.
Thompson, R. L., Higgins, C. A., & Howell, J. M. (1991). Personal comUting: toward a conceptual model of utilization. MIS quarterly, 125-143.
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS quarterly,27(3).
Venkatesh, V., Morris, M. G., & Ackerman, P. L. (2000). A longitudinal field investigation of gender differences in individual technology adoption decision-making
processes. Organizational behavior and human decision processes, 83(1), 33-60. Venkatesh, V., & Morris, M. G. (2000). Why don't men ever stop to ask for directions? Gender, social influence, and their role in technology acceptance and usage behavior. MIS quarterly, 115-139.
Venkatesh, V., Thong, J., & Xu, X. (2012). Consumer acceptance and use of information technology: extending the unified theory of acceptance and use of technology. MIS quarterly, 36(1), 157-178.
Williams, A. (2003). How to Write and analyse a questionnaire. Journal of
Orthodontics, 30(3), 245-252.
Zhou M., et al., ‘A self-powered ‘Sense-Act-Treat’ System that is Based on a Biofuel Cell and Controlled by Boolean Logic’, Angewandte Chemie International Edition, 51(11), 2686-2689 (2012).
18 Appendix
A. Vragenlijst Section 1
Geachte Meneer/Mevrouw,
Ik zou u graag willen vragen om te participeren in dit onderzoek. Mijn naam is Tom Hölscher en ik studeer momenteel 'Informatiekunde' aan de Universiteit van Amsterdam. Deze
vragenlijst maakt deel uit van een onderzoek dat ik doe als onderdeel van mijn bachelorthesis.
De samenwerking tussen mens en techniek is een belangrijk onderwerp binnen
informatiekunde. In dit onderzoek staat eventuele hinder die patiënten ondervinden van het gebruik van draagbare sensoren centraal. Hinder kan een rol spelen bij de acceptatie van het gebruik van draagbare sensoren. Er wordt steeds meer onderzoek gedaan op dit gebied. Dit gebeurd echter vaak vanuit het oogpunt van de patiënt. Dit onderzoek richt zich juist op degene die het advies geeft om een draagbare sensor te gebruiken, namelijk de arts. De verzamelde informatie zal worden gebruikt om toekomstige ontwikkeling van draagbare sensoren te ondersteunen.
Met vriendelijke groet, Tom Hölscher
B. Filled out questionnaires 1.
