Effects of user characteristics on the usability of 2 home connected medical device : The case of Smart 3 Angel 4

The Smart Angel connected medical device allows ambulatory surgery patients to monitor 10 their health by taking their own blood pressure and oxygen levels and by answering a health 11 questionnaire from home. This preventive device must necessarily be "usable" by patients with 12 different profiles. The objective of this article is, therefore, to better understand the links between 13 certain characteristics of potential patients and usability. We conducted an experimental study 14 involving thirty-six participants, investigating the effects of four patient characteristics (i.e. age, 15 education, technophilia and health literacy) on usability measured in terms of effectiveness, efficiency, 16 and satisfaction. The results show a moderate correlation between age, health literacy and usability. 17 However, there is a weak correlation between technophilia and usability and no relationship 18 between the level of education and usability. This study provides theoretical insights into the effects 19 of user characteristics by means of personas in usability (ISO 9241-11). 20


23
Outpatient surgery has been on the rise in recent years. The operations performed are 24 increasingly complex and dangerous for the patients who have to manage their convalescence at 25 home. The Smart Angel device is a Home Connected Medical Device (HMD), specifically designed to 26 prevent post-surgical complications related to outpatient surgery. The purpose of this device is to 27 facilitate the patient's return home by maintaining a link with the hospital. This requires patients to 28 use the device upon returning home after the operation, sending all their vitals three times a day for 29 one week, before returning the equipment to the Hospital Centre. This post-operative follow-up may 30 also enable patients to manage their convalescence better by avoiding the all-too-frequent returns to 31 the emergency services or outpatient consultations [1]. Currently, this system is at an early design 32 stage. Like any medical device, this tool must follow safety and quality standards [2], as well as In the context of health technologies such as connected medical devices that are increasingly 140 becoming part of the patient's life, studies on the correlation between health literacy and usability are still rare and/or exploratory. Monkman and Kushniruk [21] propose an assessment of usability by more understandable to users with a reliable level of health literacy. Czaja et al. [28] were able to 145 show, using an electronic personal health record system, that populations with low literacy levels had more difficulty using these tools. Kim and Xie [29] conducted a systematic review of articles examining the impact of low health literacy on the use of e-health devices. Based on 74 studies, the 148 authors conclude that the major barrier to accessing and using online health information for 149 individuals with low literacy is strongly related to website usability. Jensen et al. [18] found that 150 participants with low levels of health literacy (as measured by REALM) used health technologies less.

151
Those with low levels of health numeracy (as measured by TOFHLA) would have limited access to 152 these technologies. This latter finding is consistent with those of Kaufman et al. [24] who also 153 concluded that low numeracy can be a barrier to using a telemedicine system. Finally, to our 154 knowledge, no experimental studies have empirically characterized links between health literacy and 155 usability in terms of efficiency, effectiveness, and satisfaction.

157
1.3. Objective of the study 158 We have seen that the complexity of use of medical devices resides essentially in usability 159 problems [29], all the more so as they must be usable by patients with diverse profiles [33,34]. In this 160 sense, consideration of user characteristics including age, education, technophilia (IT and medical 161 device experience) and health literacy are important factors to consider in the design of a connected 162 medical device for the patient's home, such as the Smart Angel device. However, to our knowledge, 163 no study involving all these four characteristics has been conducted. Moreover, the relation between 164 these characteristics and usability is still little exploited in the literature. Thus, the aim of this work is 165 to better understand the relationships between the four user characteristics (i.e. age, educational level, 166 technophilia and health literacy) and the usability (ISO 9241-11: effectiveness, efficiency, and satisfaction) 167 of a connected medical device intended for the patient's home.

168
To do this, we formulated four hypotheses:

171
(H2) Users with a low level of technophilia (IT and medical device experience) will be less 172 effective, efficient, and satisfied with the Smart Angel connected medical device than those with a high 173 level of technophilia. (e.g. [11,23,27]).

174
(H3) The level of education will not affect the effectiveness, efficiency, and satisfaction with the 175 Smart Angel connected medical device. (e.g. [11,19,20]

206
• A procedure for using the connected objects (blood pressure monitor and pulse oximeter) 207 in which the patient finds out the information one step at a time and can initiate the 208 connection and then the measurement using these objects. Once the blood pressure or 209 oxygenation measurement has been taken, the patient's health data is displayed on a 210 coloured gauge (from green to orange) according to the level of severity of the constant 211 collected.

212
• A form in which the user answers a questionnaire with various items related to general 213 health, pain, sleep, and nausea. These items are presented either in SCQ format (e.g.

214
"How are you feeling today? Good, not good, not good at all") or on a Likert scale (e.g. left is the form allowing an overview of the subjective state of health, and on the right is the procedure 221 for using the monitor.            The correlations between user characteristics (see Table 1

349
Previous experience of using medical devices that allow users to take their own blood pressure    Table 2 below:

372
After correlation analysis, participants were clustered according to the HLS-EU-Q16 measures 373 (

385
(measured using NVS) the fewer manipulation errors they make (i.e., they are more effective), the 386 faster they manipulate (i.e., they are more efficient), and the higher their SUS score will be (i.e., they 387 will be more satisfied).

388
After analyzing the correlations, the participants were clustered according to the NVS 389 measurements (

412
We made four hypotheses that age, technophilia, and health literacy would have an impact on 413 usability, while education level would not.

414
Our first hypothesis was that older users would be less effective, efficient, and satisfied with the 415 device compared to younger users. We can partially validate this hypothesis. The results show that 416 the younger the individuals are, the less likely they are to make manipulation errors (i.e. they are 417 more effective) and the faster they manipulate the device (i.e. they are more efficient). On the other 418 hand, we did not observe any difference between the age of the subjects and the SUS score 419 (satisfaction

443
We also observed a correlation between experience with medical devices and usability. However,

444
previous experience in the use of a blood pressure monitor had no impact on usability. Conversely,

445
previous experience in the use of a pulse oximeter had a significant effect on effectiveness.

446
Participants who had previously manipulated a pulse oximeter made significantly fewer errors than 447 those who had never manipulated a pulse oximeter. In contrast, previous experience using a pulse 448 oximeter had no effect on efficiency and satisfaction. All subjects who reported previous use of a

454
Finally, the fourth hypothesis postulated that health literacy influences usability (effectiveness, 455 efficiency, and satisfaction). The HLS-EU-Q16 scores showed no effect on usability (Figure 4). In  (Figure 4). This suggests that the higher 460 the literacy level of the participants, the fewer manipulation errors they make (i.e., the more effective 461 they are), the faster they are (i.e., the more efficient they are) and the higher the SUS score will be (i.e.,

462
the more satisfied they are As a result of this study, four research perspectives can be suggested.

484
First, the relevance of the personas method in the prototype evaluation phase has never been engineers, and even future users), but much more rarely used in an evaluation framework. To Secondly, the training carried out by the researcher could be adapted according to the literacy