Development and usability evaluation of a bedside robot system for inpatients

Abstract

BACKGROUND:

Many inpatients become anxious or frightened about scheduled treatment processes, and medical staff do not have sufficient time to provide emotional support. The recent advancement of information and communications technology (ICT) and the use of artificial intelligence (AI), including robots, in the health care field is being put to the test.

OBJECTIVE:

This study aimed to develop a bedside robot system to deliver information and provide emotional support to inpatients and to evaluate the usability and perceptions of the developed robot.

METHODS:

The first stage was accomplished by deriving essential functions from the results of user demand surveys on robots and by implementing a prototype by mapping each essential function to the robot’s hardware and software. For the second stage, the robot was assessed for usability and perceptions in a simulation center, a hospital-like environment, by 10 nurses, 10 inpatients, and family caregivers. Usability and perception were evaluated using the think-aloud method, a survey, and individual interviews.

RESULTS:

Based on the usability evaluation, the perceived usefulness, ease of use, and satisfaction were 5.28 $\pm$ 1.27 points, 5.42 $\pm$ 1.55 points, and 5.27 $\pm$ 1.46 points out of 7, respectively. It was found that overall, the robot was positively perceived by participants. As a result of the qualitative data analysis, the participants perceived the robot as an object that had the positive effect of providing emotional support through communication.

CONCLUSIONS:

The bedside robot in this study, which incorporated human-robot interaction (HRI) technology, is an alternative suited to the new normal era that will contribute to ensuring that patients have more self-directed hospital stays as well as emotional support through information delivery and communication.

Keywords

Robotics artificial intelligence inpatients patient-centered care point-of-care systems caregivers

1. Introduction

Inpatients and family caregivers may develop a passive attitude or become more dependent on medical staff due to a lack of medical information pertaining to patients. Many inpatients and family caregivers become anxious or frightened about scheduled treatment processes, and medical staff do not have sufficient time to provide emotional support. From a holistic view of nursing, emotional demands are regarded as inseparable from physical and mental needs [1]. The point-of-care (POC) approach involves medical staff providing timely medical services at the patient’s side [3]. The purpose of this approach is to gain a proper understanding of the health needs of patients and to provide timely and necessary treatment and care. Recently, several domestic and foreign medical institutions have implemented inpatient portal services through bedside terminals in the form of a tablet PC with a POC approach [3, 4]. The detailed functions differ depending on the system, but they mainly deliver medical information and enable data searches through web browsers and television [5].

Recent advancements in information and communications technology (ICT) and the use of artificial intelligence (AI), including robots, in the health care field are being put to the test [3]. Robot systems integrating human-robot interaction technology (HRI) with bedside terminals are expected to provide emotional support to patients and help them lead more proactive hospital lives [6, 7]. Currently, health-related robots are mostly developed for use at home or in senior residential complexes, and it is difficult to utilize these robots because they do not fully meet the needs of inpatients. In addition, there have been attempts by the medical community to develop and utilize robots to meet health care needs, but research on user acceptance is insufficient [8].

Various factors that contribute to such systems being accepted and consistently utilized by users must be considered in preparation for the development of robots to meet health care needs, and it is important to investigate users’ needs in this process [2, 6]. Furthermore, the development of robots to be deployed for health management requires the participation of medical staff with a strong understanding of the patients and the environments the robots will be used in [9]. Unlike laboratory research, insight into users’ experiences with robots and knowledge on how this can be applied to the environments the robots will be used in can be gathered via research conducted in such environments [10]. Since families are important caregivers who provide support for inpatients, it is also necessary to reflect their opinions.

This study was designed to develop a bedside robot system to deliver information and provide emotional support to inpatients and to evaluate the usability and perceptions of the proposed robot system. The specific aims were to (1) develop a bedside robot that supports inpatients through a user requirements analysis process and (2) evaluate the usability and perceptions of the developed bedside robot to improve the acceptance of future robot applications.

2. Methods

2.1 Study design

This study comprised two stages: the development of a bedside robot system and the evaluation of the usability and perceptions of the system. The first stage was accomplished by deriving essential functions from the results of user demand surveys administered previously on robot system development [11] and then implementing a prototype by mapping each essential function to the robot’s hardware and software. For the second stage, the robot was assessed for usability and perceptions in a simulation center, a hospital-like environment.

2.2 Robot development

The bedside robot system was developed following a user requirements analysis process consisting of four stages: information gathering, user needs identification, envisioning and evaluation, and requirements specification [12]. Before proceeding with the main study, a pilot study was conducted to collect information on beside robot systems and to design a robot based on the results of the user requirements definition [11]. In the pilot study, the needs of primary users – inpatients and medical staff (doctors and nurses) – were identified through surveys and interviews during the information gathering stage, and the essential functional elements of the bedside robot system were defined based on the collected information in the user needs identification stage. Essential functions, including a main frame for robot utilization and operation, log-in and log-out, a memo function (text, voice, and video), a timer setting/alarm function, scheduling, content viewing, voice calling, voice commands, telepresence, an IoT (Internet of Things) equipment linkage function, and a user emotion recognition function, were identified [2]. During the envisioning and evaluation stage, a prototype robot that could carry out the identified essential functions was implemented. The first revision of the prototype was completed based on feedback and correction processes from nine interim evaluations by four experts with master’s degrees or higher in nursing information science (Fig. 1). The robot’s hardware consisted of a webcam for telepresence, a 7-inch display, a two-axis motion control board, a PC, and a power supply (Table 1).

Table 1
Hardware specification of the bedside robot

Items	Specification
Dimensions/Weight	W $\times$ D $\times$ H: 250 $\times$ 265 $\times$ 340 mm/4.5 kg
Head	7inch (800 $\times$ 480)
Head pitch	Adjustable angle: $+$ 10 ${}^{\circ}$ $\sim$ $-$ 10 ${}^{\circ}$ (motorized)
Head yaw	Adjustable angle: $+$ 60 ${}^{\circ}$ $\sim$ $-$ 60 ${}^{\circ}$ (motorized)
Camera/Video	8M pixels/720p
Microphone/Speaker	With camera/6W
Wireless network	802.11 a, b, g, or n
Power	DC 19V 5A

Figure 1.

Bedside robot.

2.3 Evaluation of usability and perceptions

2.3.1 Research design

As a single group post-design study using mixed methods to assess the usability and perception of bedside robots, a prototype of a bedside robot was produced to perform defined essential functions. An evaluation was conducted once for the participants, and the robot was deployed according to certain scenarios and tasks. Usability and perception were evaluated through the think-aloud method, a survey, and individual interviews. The think-aloud method is effective in assessing users’ ability to interact with ICT [13, 14]. A survey was administered using a self-reporting type questionnaire, and interviews were conducted to identify problems or improvements that may arise from the use of robots in ward settings.

2.3.2 Participants

Ten nurses currently working at medical institutions and 10 family caregivers or adults 18 years of age or older who had been hospitalized for less than 7 days within the last 3 months and who volunteered to participate in the usability evaluation were selected via convenience sampling. Nielsens’ previous work showed that 12 participants can be a sufficient sample size to discover the majority of usability issues [13].

2.4 Measurements

2.4.1 Evaluation scenario and task development

The researchers developed a case scenario and 14 tasks, which were reviewed by three nursing informatics experts. Cases were based on the hospitalization of patients who needed operations. Examples of tasks included “registration using face recognition” and “call the nurse using voice command.” For each task, the participants rated the difficulty level of performing each function on a 5-point Likert scale.

2.4.2 General characteristics

The general characteristics of the subjects included gender, age, occupation – in the case of nurses – clinical experiences, and inpatient departments, in the case of a patient or family caregiver. Gender and age as sociodemographic traits for the general characteristics of the subjects, occupation type, clinical experiences for nurses, and medical department for patient or family caregivers were surveyed.

2.4.3 Usability of robot

Since no test tools have been developed for this study, professors in medical information science and nursing information science were commissioned to modify and supplement a Korean translation of the perceived ease of use and perceived usefulness measurement tool developed by Davis [16] and the user satisfaction tool developed by Wang and Liao [17]. The usability evaluation tools have a total of 24 questions: nine questions for perceived usefulness, six questions for perceived ease of use, and nine questions for satisfaction. Questions were based on a 7-point Likert scale. The perceived usefulness and perceived ease of use tools were validated, and a Cronbach’s $\alpha$ of 0.85 and 0.93, respectively, were obtained. The study by Wang and Liao reported appropriate content validity and criterion validity for the user satisfaction tool. In this study, the user satisfaction tool had a Cronbach’s $\alpha$ of 0.77.

2.4.4 Godspeed questionnaire series

The Godspeed questionnaire series (GQS) was used to measure the users’ perception of the robot. This is one of the most widely used questionnaires in the field of HRI, and it has been cited more than 600 times as of October 2018 [18]. The tool, developed in Germany, has been translated into diverse languages. It was translated into Korean by a professional translator and was revised based on the advice of a relevant expert. The GQS consists of five scales that are relevant for evaluating the perception of social HRI. The scales include anthropomorphism (five items), animacy (six items), likeability (five items), perceived intelligence (five items), and perceived safety (three items). The scales consist of 5-point semantic differentials, such as “fake-natural.”

2.4.5 Interview questions

Since it was difficult to fully understand the circumstantial factors of robot acceptance based only on quantitative surveys, the participants were individually interviewed for 20–30 minutes after they filled out the questionnaire. The main interview questions used in this study were extracted through a literature review (Table 2).

Table 2
Interview guide

Participants	Interview questions
All	• How do you think the bedside robot would help patients’ hospital stay? • What is your impression of the appearance of the bedside robot? – Do you approve of the robot’s appearance? – What do you think about the size and shape of the robot? • What is your opinion on the robot’s movement? – Is the speed of the robot’s movement appropriate? • Are there any difficulties using the robot? – What difficulties (if any) did you experience? – How can it be improved?
Nurse	• How do you think the utilization of bedside robots would help the nurses with their tasks? • What are the administrative problems with the use of bedside robots? – Are there any problems in terms of managing/maintaining robots? – Are there any suggestions for improvement in terms of managing/maintaining robots? • Are there any clinical problems that arise from the use of the bedside robots, such as infections or interruptions of the treatment process?
Patient and family caregiver	• What are your emotions toward the robots? – Do you feel comfortable around the robot? – Do you feel that the robot is someone you can have conversations with? – Do you feel that the robot is a conscious being?

Participants

Interview questions

All

•

How do you think the bedside robot would help patients’ hospital stay?

•

What is your impression of the appearance of the bedside robot?

–

Do you approve of the robot’s appearance?

–

What do you think about the size and shape of the robot?

•

What is your opinion on the robot’s movement?

–

Is the speed of the robot’s movement appropriate?

•

Are there any difficulties using the robot?

–

What difficulties (if any) did you experience?

–

How can it be improved?

Nurse

•

How do you think the utilization of bedside robots would help the nurses with their tasks?

•

What are the administrative problems with the use of bedside robots?

–

Are there any problems in terms of managing/maintaining robots?

–

Are there any suggestions for improvement in terms of managing/maintaining robots?

•

Are there any clinical problems that arise from the use of the bedside robots, such as infections or interruptions of the treatment process?

Patient and family caregiver

•

What are your emotions toward the robots?

–

Do you feel comfortable around the robot?

–

Do you feel that the robot is someone you can have conversations with?

–

Do you feel that the robot is a conscious being?

2.5 Data collection and procedure

Data for this study were collected from July 22 to July 28, 2019. Subject recruitment began with an announcement about recruitment for the sake of evaluating the usability and perceptions of the bedside robot, and then the researchers explained the purpose of the study to the potential subjects and proceeded with the study after receiving written consent from the subjects to participate in the usability evaluation. The specific procedure went as follows: 1) subjects received individual training on how to use the robot for approximately 20 minutes, and then a user manual was distributed; 2) subjects had a brief conversation with the bedside robot on what day of the week it was and the current weather, and they spent time getting used to the robot; 3) subjects utilized the robot under certain scenarios and discussed aspects that caused inconvenience or needed improvement using the think-aloud technique; 4) a survey was administered on the usability and perceptions of the robot after subjects used the robot; and 5) individual interviews were conducted to identify problems or possible improvements that were brought to the users’ attention while using the robot system.

2.6 Data analysis

SPSS Statistics for Windows version 22.0 (Released 2013; IBM Corp., Armonk, NY, USA) was used for the quantitative data analysis. The general characteristics, usability, and GQS of the subjects were expressed as descriptive statistics (constants, percentages, means, and standard deviations) and were analyzed using the Mann-Whitney U test to find differences in usability and GQS between the medical staff and patients’ family caregivers.

The qualitative data acquired using the think-aloud technique and interviews were analyzed using content analysis. First, the researchers attempted to grasp the overall meaning by immersing themselves in and reading all of the written data without categorizing it. Second, the data were read exactly as the participants conveyed them, and coding words, sentences, and paragraphs containing major thoughts or concepts were selected. Third, related codes were collected and classified into sub-topics, and the main topic was identified. Fourth, the degree of consistency among the researchers informed the process of determining whether to continue the data collection through discussions, extracting concepts, or naming topics. To maintain the consistency of the research and the neutrality of the data analysis, daily research logs were kept, and five regular meetings were held.

2.7 Ethical considerations

The research plan was reviewed and approved by the institutional review board of Seoul National University Hospital (IRB No. 1706-174-864). After explaining the purpose and method of the study to the subjects who wished to participate voluntarily, the researchers recruited those who expressed their intent to participate and signed the consent form. To prevent the leakage of personal information and questionnaire contents, all data were converted into electronic files by the researchers, and the converted data were stored with a password set on a personal computer so that only the researchers could gain access.

3. Results

3.1 Development of a bedside robot for inpatients

The main functions of the robot developed based on the needs of the subjects are listed in Table 3, and the main screens are shown in Fig. 2. The architecture of the bedside robot and the division of authority by the user are shown in Fig. 3.

Table 3
Robot’s main functions

Function	Explanation
Main frame UI	Frame for operating the robots.
Face recognition sign in	The user’s face is registered with the camera for automatic sign in by facial recognition.
Linking IoT equipment	Monitoring and alarming of patient exercise and sleep using IoT equipment.
Telepresence function • Video transmission • Audio transmission	A function to have face-to-face conversations with the user in a remote location (simulation center corresponding to a nursing office).
Memo function 1: Text memo	Storing memos as text files in the robot.
Memo function 2: Audio memo	Storing memos as audio files in the robot.
Memo function 3: Video memo	Storing memos as video files in the robot.
Timer setting alarm function	Setting an alarm in a remote location or from the robot by having the time set up in advance for different circumstances.
Schedule maintenance	Event notification by bedside robot for events registered on the robot or in a remote location.
Content viewing (video)	Playing videos on surgery information and hospital information.
TV and movie watching	Television and movie player.
Voice call	Calling a PC in a remote location (simulation center corresponding to the nursing office) through voice call in a simulated emergency situation.
Voice command	Controlling certain functions through voice recognition, e.g., end of use.
Pain assessment	Storing the patient’s pain assessment via VAS. Graph display available for trend identification.
Emotional content 1: Decision module for robot’s response based on recognized user information	Showing appropriate gesture/sound responses by recognizing emotions through the user’s expression and voice.
Emotional content 2: Story room	Carrying out simple conversations with robots.

UI: User interface, IoT: Internet of Things, VAS: Visual Analog Scale.

Figure 2.

Main screens of bedside robot. (A) Registration using face recognition, (B) Main screen, (C) Monitoring of activities using IoT, (D) Pain assessment.

3.2 Quantitative results

Ten nurses, five inpatients, and five family caregivers participated in the study. All nurses were female, with a mean age of 29.80 $\pm$ 9.62. As for the patients and caregivers, 70.0% were female, with a mean age of 41.90 $\pm$ 15.10; and 50% were from the internal medicine ward, with the rest from the surgical ward (Table 4).

Table 4
Survey participants ( $N=$ 20)

Type		$n$	%
Nurses ( $n=$ 10)
Gender	Female	10	100.0
Age (y)	Mean $\pm$ SD	29.80 $\pm$ 9.62
Clinical experiences (months)	Mean $\pm$ SD	88.20 $\pm$ 120.82
Inpatients or family caregiver ( $n=$ 10)
Gender	Male	3	30.0
	Female	7	70.0
Age (y)		41.90 $\pm$ 15.10
Departments	Medical ward	5	50.0
	Surgical ward	5	50.0

SD: Standard deviation.

Figure 3.

System architecture of the bedside robot.

3.2.1 Difficulty of task execution by function

Results from 14 tasks showed that the functions were relatively easy to use, with a mean difficulty score of 1.69 $\pm$ 0.89 points out of 5, where patients and family caregivers scored 1.60 $\pm$ 0.85 points and nurses scored 1.78 $\pm$ 0.92 points (Table 5). Patients and family caregivers expressed a higher level of ease with the functions, but the difference between the two groups was not statistically significant.

In terms of the difficulty of performing tasks for each function of the robot, “find ‘activity information’ and number of steps from yesterday” and “ask for the date by voice command” had the highest levels of difficulty, with scores of 2.55 $\pm$ 1.24 points and 1.85 $\pm$ 1.06 points out of 5, respectively, while “choose a chat room to chat with the robot” and “when the ‘call nurse’ popup appears, select yes or call the nurse by voice command” had 1.35 $\pm$ 0.57 points and 1.40 $\pm$ 0.66 points, respectively, as the easiest tasks. In the “registration using face recognition” category, patients and caregivers reported it to be easier to perform than nurses did.

Table 5
The degree of difficulty of tasks

Functionalities and tasks		The degree of difficulty (M $\pm$ SD)
		Patients and FCG ( $N=$ 10)	Nurses ( $N=$ 10)	Total ( $N=$ 20)	Z ( $p$ )
1	Registration using face recognition	1.33 $\pm$ 0.71	2.00 $\pm$ 0.82	1.68 $\pm$ 0.80	$-$ 2.128	(0.033)
2	Find ‘activity information’ and number of steps of yesterday	2.50 $\pm$ 1.43	2.60 $\pm$ 1.17	2.55 $\pm$ 1.24	$-$ 0.354	(0.723)
3	Select ‘pain management’ using voice command	1.90 $\pm$ 1.10	1.50 $\pm$ 0.97	1.70 $\pm$ 1.00	$-$ 0.944	(0.345)
4	Record pain scale (location: head, scale: 7)	1.40 $\pm$ 0.70	1.60 $\pm$ 0.97	1.50 $\pm$ 0.81	$-$ 0.448	(0.654)
5	When the ‘call nurse’ popup appears, select yes or call the	1.20 $\pm$ 0.42	1.60 $\pm$ 0.84	1.40 $\pm$ 0.66	$-$ 1.125	(0.260)
	nurse by voice command
6	Select ‘voice note’ using voice command	1.90 $\pm$ 0.88	1.60 $\pm$ 0.97	1.75 $\pm$ 0.89	$-$ 0.945	(0.345)
7	Record a voice note. “When can I take a shower?”	1.50 $\pm$ 0.53	2.00 $\pm$ 0.94	1.75 $\pm$ 0.77	$-$ 1.258	(0.208)
8	Save the title ‘question list’	1.80 $\pm$ 1.03	1.80 $\pm$ 1.03	1.80 $\pm$ 0.98	0.0	( $<$ 1.000)
9	Call the nurse using voice command	1.80 $\pm$ 0.92	1.90 $\pm$ 1.29	1.85 $\pm$ 1.06	$-$ 0.082	(0.935)
10	Ask for the date by voice command	1.30 $\pm$ 0.48	1.70 $\pm$ 0.67	1.50 $\pm$ 0.59	$-$ 1.420	(0.156)
11	Watch the video you want by selecting the video menu	1.40 $\pm$ 0.70	1.60 $\pm$ 0.70	1.50 $\pm$ 0.67	$-$ 0.781	(0.435)
12	Move home screen with voice commands	1.60 $\pm$ 0.84	1.80 $\pm$ 0.92	1.70 $\pm$ 0.84	$-$ 0.580	(0.562)
13	Choose a chat room to chat with the robot	1.20 $\pm$ 0.42	1.50 $\pm$ 0.71	1.35 $\pm$ 0.57	$-$ 1.037	(0.300)
14	Go back home and log out with voice commands	1.50 $\pm$ 0.53	1.70 $\pm$ 0.67	1.60 $\pm$ 0.58	$-$ 0.640	(0.522)
Total		1.60 $\pm$ 0.85	1.78 $\pm$ 0.92	1.69 $\pm$ 0.89	$-$ 0.533	(0.594)

3.2.2 Evaluation results for usability

From the usability evaluation, the perceived usefulness, ease of use, and satisfaction were 5.28 $\pm$ 1.27 points, 5.42 $\pm$ 1.55 points, and 5.27 $\pm$ 1.46 points out of 7, respectively. When comparing the nurse group to the patient and family caregiver group, the latter group expressed significantly higher perceived usefulness than the former ( $p=$ 0.049), and ease of use and satisfaction were slightly higher in the patient and family caregiver group, but there were no statistically significant differences (Table 6).

Table 6
Perceived usefulness, perceived ease of use, and satisfaction of the robot

Categories	Patients & FCG ( $N=$ 10)	Nurses ( $N=$ 10)	Total ( $N=$ 20)	Z ( $p$ )
	M $\pm$ SD
Perceived usefulness	5.76 $\pm$ 0.97	4.80 $\pm$ 1.34	5.28 $\pm$ 1.27	$-$ 1.968 (0.049)
Perceived ease of use	5.58 $\pm$ 1.55	5.25 $\pm$ 1.53	5.42 $\pm$ 1.55	$-$ 0.682 (0.495)
Satisfaction	5.48 $\pm$ 1.45	5.06 $\pm$ 1.43	5.27 $\pm$ 1.46	$-$ 1.022 (0.307)

FCG: Family caregiver

3.2.3 Evaluation results for perceptions of the robot

While the evaluation results for perceptions of the robot showed that the patients and family caregivers evaluated the robots more positively than the nurses did for most of the evaluation items, there were no statistically significant differences. In the anthropomorphism categories, the patients and family caregivers reported that the robot was more “natural,” “humanlike,” “lifelike,” and “moving elegantly” than the nurses did. The patients and family caregivers felt that the robot was “alive” and “lifelike,” and the nurses felt that the robot was “interactive” and “responsive” in the animacy categories. The patient and family caregiver group showed higher scores than the nurse group in all evaluation items in the likeability, perceived intelligence, and perceived safety categories (Table 7).

Table 7
Users’ perception of the robot

Categories		Patients ( $N=$ 10)	Nurses ( $N=$ 10)	Z ( $p$ )
		M $\pm$ SD
	Scale (1–5)
Anthropomorphism	Fake/Natural	3.10 $\pm$ 1.10	2.90 $\pm$ 1.45	$-$ 0.349 (0.727)
	Machinelike/Humanlike	3.10 $\pm$ 1.10	2.60 $\pm$ 1.35	$-$ 0.987 (0.324)
	Unconscious/Conscious	3.40 $\pm$ 0.97	3.20 $\pm$ 1.62	$-$ 0.155 (0.877)
	Artificial/Lifelike	2.70 $\pm$ 0.82	2.40 $\pm$ 1.26	$-$ 0.976 (0.329)
	Moving Rigidly/Moving elegantly	2.90 $\pm$ 1.10	2.50 $\pm$ 1.18	$-$ 0.921 (0.357)
Animacy	Dead/Alive	3.70 $\pm$ 0.48	3.60 $\pm$ 1.35	$-$ 0.204 (0.838)
	Stagnant/Lively	4.30 $\pm$ 0.67	4.10 $\pm$ 0.99	$-$ 0.288 (0.774)
	Mechanical/Organic	3.00 $\pm$ 1.25	3.00 $\pm$ 1.33	$-$ 0.155 (0.877)
	Artificial/Lifelike	3.00 $\pm$ 1.33	2.80 $\pm$ 1.48	$-$ 0.388 (0.698)
	Inert/Interactive	3.50 $\pm$ 0.71	3.70 $\pm$ 1.25	$-$ 0.433 (0.665)
	Apathetic/Responsive	3.90 $\pm$ 0.99	4.10 $\pm$ 0.88	$-$ 0.398 (0.690)
Likeability	Dislike/Like	4.60 $\pm$ 0.52	3.90 $\pm$ 1.10	$-$ 1.479 (0.139)
	Unfriendly/Friendly	4.20 $\pm$ 0.63	3.80 $\pm$ 1.23	$-$ 0.561 (0.575)
	Unkind/Kind	4.40 $\pm$ 1.07	4.40 $\pm$ 0.52	$-$ 0.764 (0.445)
	Unpleasant/Pleasant	4.40 $\pm$ 0.70	3.80 $\pm$ 1.23	$-$ 1.048 (0.295)
	Awful/Nice	4.30 $\pm$ 0.82	4.10 $\pm$ 0.99	$-$ 0.406 (0.684)
Perceived intelligence	Incompetent/Competent	4.30 $\pm$ 0.48	3.80 $\pm$ 1.03	$-$ 1.150 (0.250)
	Ignorant/Knowledgeable	4.10 $\pm$ 0.99	3.50 $\pm$ 0.97	$-$ 1.418 (0.156)
	Irresponsible/Responsible	3.70 $\pm$ 1.06	3.20 $\pm$ 1.14	$-$ 1.066 (0.286)
	Unintelligent/Intelligent	3.60 $\pm$ 0.84	3.30 $\pm$ 1.25	$-$ 0.594 (0.553)
	Foolish/Sensible	3.90 $\pm$ 0.99	3.50 $\pm$ 1.08	$-$ 0.865 (0.387)
Perceived safety	Anxious/Relaxed	4.30 $\pm$ 0.67	3.70 $\pm$ 1.25	$-$ 1.124 (0.261)
	Calm/Agitated	3.10 $\pm$ 1.10	4.10 $\pm$ 0.99	$-$ 1.965 (0.049)
	Quiescent/Surprised	3.20 $\pm$ 0.92	3.30 $\pm$ 1.25	$-$ 0.197 (0.844)

3.3 Qualitative results

The analysis of the qualitative content led to the aggregation of the advantages and disadvantages of using robots as experienced by the users into three topics: usability, expected effects, and recommendations for improvement. Detailed analysis outcomes are shown in Table 8.

Table 8
Users’ experience feedback regarding the bedside robot ( $N=$ 20)

Categories	Participants	Type	Sample response
Usability
Robot’s shape and appearance	N1,2,3,4,5,6,8,10(8) P1,2,4,5,6,8,9,10(8)	Positive	Cute and pretty appearance, favorable color
	N1,2,3,4,5,8(6) P2,4,7,10(4)	Negative	The robot is larger than expected and needs space to be placed Need to be slim and light robot/Restriction of movement due to big size
Robot’s ease of use	N1,2,3,4,5,6,7,10(8) P2,4,9(3)	Positive	Smooth screen touch Menu configuration is appropriate Works well without failure/Sense of stability
	N1,2,5,8(4) P1,2,4,5,6,7,8,10(8)	Negative	Delay of voice recognition/Requires accurate announcement of voice processing to avoid repeating commands
Personalized functions and services	N3,5,6,7,8,10(6) P1,2,4,7,8,10(6)	Positive	Daily schedule sharing functions that help patients adapt to the hospital Voice saving, video note during rounding is useful/Less burdensome than a person
	N1,2,3,5,6,10(6) P1,2,5,6,7,8(6)	Negative	Requesting the ability to select the font size, color, and contents according to the patient (age, education level)/Small letters for elderly patients
Expected effects
Benefits of content	N1,2,3,4,5,7,8(7)	Positive	Helpful for immediate pain assessment and intervention
	P1,2,3,4,5,6(6)	Positive	Offering interesting entertainment Meeting friends via telepresence function
Emotional support with communication	N1,2,6,10(4) P1,2,3,4,6,7,8,10(8)	Positive	Emotional care is possible without guardian Feeling good with a robot that can talk/Realistic conversation Receiving energetic energy due to robot’s presence Not lonely and happy to have someone to talk to/Become attached to robots
	N5(1) P5,9(2)	Negative	Unnatural tone/Unfamiliar system Disturbance of patient stability in physical situations where it is difficult to hear the robot
Potential to reduce human resources	N1,2,3,5,7,10(6) P1,4(2)	Positive	Patient care and work support available with video call function Staying with patients on behalf of busy medical staff Can be used as a substitute for medical staff
	N3,5,7(3) P3,5,6(3)	Negative	Need a moving robot that can help when transfer More likely to load work in the event of a robot failure (Need staff for error management)
Recommendations for improvement
Need for additional contents	N1,2,4,5,7,8(5)	Focused on nursing	Recommended robots focused on educational functions such as pain, fall down, pressure injury, diet management, medications and examinations
	P1,3,4,5,6,7,8(7)	Focused on interest	Diversification of educational and entertainment contents that will free patients from boredom
	N1,2(2) P3,4,5,6,7,8(6)		Internet connection for searching for music or videos

4. Discussion

In this study, a bedside robot for patients that enables information delivery and provides emotional support by adding HRI functions to the bedside terminal for inpatients was constructed, and its usability and perceptions were evaluated. In conjunction with ongoing research on patient portals and bedside information terminals that deliver information for inpatients, the development of a bedside robot and the users’ evaluations of its usability and value described in this paper have significant clinical applications, filling the void in research on the development of robot systems that promote positive hospital experiences through emotional support. In particular, because it adhered to the recommendations for participation by medical staff who fully understand the patients and the environment the robots will be used in, and it considered various factors regarding the actual users – patients in the case of a robot for health care [3, 6, 9] – this study is meaningful in that a bedside robot was developed and then evaluated by nurses and patients in a hospital environment to establish the usability and perceptions of the robot.

For a health care robot to be successfully adopted by users, it is essential that adequate ease of use is implied when using the robot [6, 7]. The main functions of the bedside robot developed in this study consisted of activity information, pain management, emergency calls, scheduling, voice notes, a story room, educational videos, logout, and more. The evaluation of 14 tasks’ difficulty level based on the users’ experiences showed that the robot’s functions were relatively easy to use, with a mean difficulty score of 1.69 $\pm$ 0.89 points out of 5, and the perceived usefulness, ease of use, and satisfaction from the usability evaluation received 5.28 $\pm$ 1.27 points, 5.42 $\pm$ 1.55 points, and 5.27 $\pm$ 1.46 points out of 7, respectively, showing usability of more than a moderate to high level, which implies that the requirements of HRI were met. This demonstrates that for the development of the robot in this study, the robot system implementation based on the user requirements survey [11], which provided an accurate representation of what the inpatients needed the most in the hospital setting, made it possible to reflect on the inpatients’ needs, resulting in a bedside robot with improved usability [23, 24, 25].

The analysis of the interviews with the users led to the aggregation of the advantages and disadvantages of using robots as experienced by the users into three topics: usability, expected effects, and recommendations for improvement. In terms of usability, despite its cute appearance, the robot was quite bulky, imposing space constraints for patients sharing a room. Furthermore, whereas stable touchscreen or menu button use was possible with the robot, it was slow in voice recognition, and the process of recognizing the user’s voice was not clearly visible to the subjects, resulting in the inconvenience of having to repeat voice commands. This was deemed as a request for improved accuracy in the responses to speech recognition via the system that is triggered by voice commands when performing a specific task with the robot. While the schedule management and voice notes were helpful in ensuring an active hospital stay, it was shown that the patients needed a function to configure a font size or screen color to suit them, and it is speculated that older patients will have difficulty using the system due to the lack of choice of fonts or screen compositions suitable for the subject. For the future deployment of robots in hospitals based on users’ opinions, it is critical that a strategy for using a robot that respects individuality through a variety of options in consideration of the patient’s age and gender is in place.

When comparing the perceptions of the robot between nurses and patients, it was found that overall, the robot was perceived positively, and in particular, the patients and family caregivers felt that the robot was more alive, conscious, kind, and intelligent than the nurses did. The interview results showed that the patients expected the robot to help ensure an active hospital stay or positive hospital experience more than the nurses did. There was a concern from one nurse regarding older patients having a difficult time with the robot, but the perceived ease of use and satisfaction of the patients were high, and most tasks could be performed without great difficulty.

Whereas the nurses had the impression that immediate pain assessment and management were helpful in patient care in terms of the robot’s function as an expected effect, the patients and family caregivers perceived it as an object that provided amusement and interest during a boring hospital stay, could be relied on while sharing a hospital room, and had the positive effect of providing emotional support through honest communication that could not be had with the medical staff. This is in keeping with the results of a previous study [8], which reported that patients showed an interest in conversations with the robot itself, including conversations about health care. With the recent fast-paced advancement of robots with intelligent functions, technologies for interactions such as recognizing the user’s current state through an analysis of their voice or visual elements and responding to requests have become possible [21]. While some users were disturbed by the robot’s awkward mechanical voice and unfamiliar mechanical system, most patients became accustomed to the unfamiliar robot and had the positive experience of gradually developing an affinity for the robot as their usage time accumulated. A previous study on the use of intelligent robots in the form of animals [26] showed that the robots’ facial expressions, head movements, recognition of the user’s emotions using audiovisual data, and speech function significantly reduced depression in the elderly. This is in line with the outcome of this study, implying that HRI will have a positive effect on the user’s emotional domain. In addition, the use of a series of bedside robots for communication between the patient and the nurse through video calls, the voice function of the robot during rounding, active pain management after surgery, and goal setting during early ambulation promote an active health care approach for the patients and provide evidence that robots can fill the gaps in the interaction between medical staff and patients in the overwhelmed Korean health care environment. However, despite the positive effects of these robots, without additional personnel dedicated to their use and management, they may increase the workload of the medical staff in the event of technical problems, so securing robot management personnel should be a priority.

This work was motivated by bedside robots fixed next to the patient’s bed with a restricted range of activity. Currently, various forms of robots, including humanoid robots [27], animal-like robots, walkers that assist users in moving, and robots with attached tablets to enable video chats with medical staff or families in other locations, are being developed to serve specific groups of users with different needs [8]. Therefore, it is speculated that the limitations of the robots can be further reduced so that they will be able to relieve patients from anxiety during transport to examination rooms; and, if the robots’ range of activity is enhanced, they will be able to assist nurses with various tasks. In addition, in times like the present, with the COVID-19 outbreak, the strategic use of the various functions offered by bedside robots, such as telepresence conversations with caregivers, providing sources of music and videos, and providing someone to converse with, might go a long way in relieving emotional difficulties such as anxiety and loneliness that patients feel due to the strict management of visitors such as family members. Furthermore, it is expected that if a crisis comparable to COVID-19 breaks out in the future, bedside robots could be used as another resource capable of taking on given roles and performing certain health care tasks on behalf of the medical staff in response to staff shortages due to strict infection control.

In addition to usability and the expected effects, the nurses mainly asked for upgrades to the contents provided by the robot, such as specific patient care education on pain, falls, bedsores, diet management, and drugs, while the patients suggested expanding the use of robots to alleviate the weariness of hospital stays through various pastimes, including music and videos from the Internet, in addition to beneficial educational materials necessary for the treatment process. In the future, we can expect robots optimized for clinical use that strike a balance between disease treatment and entertainment. This will be achieved by putting together contents for the robots based on the improvement plans derived from the nurses’ and patients’ experiences of using bedside robots.

There are some limitations to this study. The study was conducted within a simulation center, rather than an actual hospital environment, with fewer subjects evaluating the robot. Additionally, no evaluation of the programs for nurses was conducted. Furthermore, there is a high likelihood that patients and family caregivers who are interested in and familiar with robots participated due to the convenience sampling. Therefore, it is necessary to recruit more subjects for future studies to examine the effectiveness and acceptability of robots as a function of a range of characteristics (treatment department, age, gender, disposition, patient’s main diagnosis, hospitalization department, operation status, etc.). Lastly, it is recommended that the effectiveness of a more diverse range of robots be evaluated through additional programs for nurses.

5. Conclusions

Robots in the field of health care have the essential purpose of improving the health and quality of life of users, and they are expected to show significant growth to meet the demands arising from shortages of medical professionals, higher quality of life, and demographic changes, such as the rise in the elderly population. In particular, the bedside robot in this study, which incorporated HRI technology, is an alternative suited to the new normal era that will contribute to ensuring that patients have more self-directed hospital stays as well as emotional support through information delivery and communication.

Footnotes

Conflict of interest

The authors have no conflicts of interest to disclose.

Funding

This work was based upon work supported by the Ministry of Trade, Industry & Energy (MOTIE, Korea) under Industrial Technology Innovation Program (no. 10063098, “Telepresence Robot System Development for the Support of POC (Point of Care) Service Associated With ICT Technology”). This work was supported by the Gachon University research fund of 2020 (GCU-202002130001).

References

Raghubir

. Emotional intelligence in professional nursing practice: a concept review using Rodgers’s evolutionary analysis approach. Int J Nurs Sci. 2018; 5(2): 126-130.

Lee

Kim

. Scenario-based assessment of user needs for point-of-care robots. Healthc Inform Res. 2018; 24(1): 12-21. doi: 10.4258/hir.2018.24.1.12.

Yen

Walker

Smith

JMG

Zhou

Menser

McAlearney

. Usability evaluation of a commercial inpatient portal. Int J Med Inform. 2018; 110: 10-18.

Yoo

Lee

Baek

Ryu

Chung

Kim

, et al. Development and user research of a smart bedside station system toward patient-centered healthcare system. J Med Syst. 2015; 39(9): 86.

Baird

. The new Royal Adelaide Hospital: the age of the digital hospital dawns in South Australia. Paper presented at the Proceedings of the Sixth Australasian Workshop on Health Informatics and Knowledge Management. 2013; 142.

Broadbent

Stafford

MacDonald

. Acceptance of healthcare robots for the older population: review and future directions. Int J Soc Robot. 2009; 1(4): 319-330.

Hancock

Billings

Schaefer

Chen

De Visser

Parasuraman

. A meta-analysis of factors affecting trust in human-robot interaction. Hum Factors. 2011; 53(5): 517-527.

Goher

Mansouri

Fadlallah

. Assessment of personal care and medical robots from older adults’ perspective. Robotics Biomim. 2017; 4(1): 5.

Deneke

Lehane

Kandler

Menchini

Laaksoharju

Obaid

. Using rapid prototyping to explored design implications for a pill-dispensing social agent. Paper presented at the Proceedings of the 5th International Conference on Human Agent Interaction. 2017.

10.

Frennert

Eftring

Östlund

. Case report: implications of doing research on socially assistive robots in real homes. Int J Soc Robot. 2017; 9(3): 401-415.

11.

Lee

Piao

Lee

Byun

Kim

. The purpose of bedside robots: exploring the needs of inpatients and healthcare professionals. Computers, Informatics, Nursing: CIN. 2020; 38(1): 8-17.

12.

Maguire

Bevan

. User requirements analysis. Paper presented at the IFIP 17th World Computer Congress. 2002.

13.

Nielsen

. Estimating the number of subjects needed for a thinking aloud test. Int J Hum Comput Stud. 1994; 41(3): 385-397.

14.

Van Someren

Barnard

Sandberg

JAC

. The think aloud method: a practical approach to modelling cognitive. London Academic Press. 1994.

15.

Nielsen

. Usability inspection methods. Paper presented at Conference companion on Human factors in computing systems. Boston. 1994.

16.

Davis

. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 1989; 13(3): 319-340.

17.

Wang

Liao

. The conceptualization and measurement of m-commerce user satisfaction. Comput Human Behav. 2007; 23(1): 381-398.

18.

Bartneck

Kulić

Croft

Zoghbi

. Measurement instruments for the anthropomorphism, animacy, likeability, perceived intelligence, and perceived safety of robots. Int J Soc Robot. 2009; 1(1): 71-81.

19.

O’Leary

Sharma

Killarney

O’Hara

Lohman

Culver

, et al. Patients’ and healthcare providers’ perceptions of a mobile portal application for hospitalized patients. BMC Med Inform Decis Mak. 2016; 16(1): 123.

20.

Kelly

Hoonakker

Dean

. Using an inpatient portal to engage families in pediatric hospital care. J Am Med Inform Assoc. 2016; 24(1): 153-161.

21.

Petrie

Darzentas

Carmien

. Intelligent support technologies for older people: an analysis of characteristics and roles. In: Langdon, Pat, Lazar, Jonathan, Heylighen, Ann and Dong, Hua, (eds.) Breaking down barriers: usability, accessibility and inclusive design. Springer, Cham, Switzerland. 2018; pp. 89-97.

22.

Dautenhahn

. Socially intelligent robots: dimensions of human–robot interaction. Philos Trans R Soc Lond B Biol Sci. 2007; 362(1480): 679-704.

23.

De Graaf

Allouch

Klamer

. Sharing a life with Harvey: exploring the acceptance of and relationship-building with a social robot. Comput Human Behav. 2015; 43: 1-14.

24.

Heerink

Kröse

Evers

Wielinga

. Assessing acceptance of assistive social agent technology by older adults: the Almere model. Int J Soc Robot. 2010; 2(4): 361-375.

25.

Whelan

Murphy

Barrett

Krusche

Santorelli

Casey

. Factors affecting the acceptability of social robots by older adults including people with dementia or cognitive impairment: a literature review. Int J Soc Robot. 2018; 10(1): 643-668.

26.

Kargar

BAH

Mahoor

. A pilot study on the eBear socially assistive robot: implication for interacting with elderly people with moderate depression. Paper presented at the Humanoid Robotics (Humanoids), 2017 IEEE-RAS 17th International Conference. 2017.

27.

Piezzo

Suzuki

. Feasibility study of a socially assistive humanoid robot for guiding elderly individuals during walking. Future Internet. 2017; 9(3): 30.