Keywords

1 Introduction

Demographic data [1] show that the ratio of people over 65 years old in Japan has already exceeded 28.1% of the total population (i.e., 1 in every 4 persons). On the other hand, the child population (0–14 years old) decreased to 12.2% of the total population. This means that the productive-age population (15–64 years old) accounted for just 59.7% of the entire population. Japan is facing a severe shortage in the productive-age population, which could lead to a reduction in the quality of nursing and care services for older adults; therefore, robot technologies are expected to be utilized increasingly in such services.

To deal with the issue, the Ministry of Economy, Trade and Industry (METI) and the Ministry of Health, Labour and Welfare (MHLW) worked together to formulate six priority areas [2, 3] in an aging society in which robots are expected to be utilized. They are: (1) transferring support (wearable/non-wearable), (2) mobility support (outdoor/indoor/wearable), (3) toileting support, (4) bathing support, (5) monitoring (care facility/private home) and communication, and (6) care service support. For these priority areas, there are many research and development projects [4, 5] looking to utilize such robot technologies in the elderly care field.

In these research and development activities, attempts have been made to ascertain the effects of the developed devices, and we can find the results in several reports [6, 7] on these projects; however, there is little academic research output. Even in the project reports, there are no detailed explanations as to how the devices were used. It is, therefore, difficult for the target users to understand the validity of the confirmed effects.

Most of these developed devices are new category products, so there is no widely accepted usage of the devices. This can be compared to medicine with unclear dosage and administration. Under such conditions, we consider that it is too early to discuss the effect of the devices; and instead appropriate usage should be discussed at this stage.

To deal with the appropriate usage of the device, the following two points need to be discussed. Firstly, current usage of the device; and secondly, how the users made their decisions, before and after they started to use the device. No research has been conducted to investigate both the usage and related decision-making, so it is worthwhile to investigate these points.

When the current usage of a device is investigated by human observers, it places a heavy mental burden on the device users. In addition to this, it might be possible for the users to use the device to a greater extent or differently than usual, which will affect the result of the research. This relates to the well-known phenomenon of the Hawthorne Effect. We, therefore, decided to use commercial products, which have a logging function to record users’ operations or alert outputs, so as to avoid direct observation in this work. The primary purpose of this work is to discuss the possibility of estimating device usage from their recorded logger data sets.

The target robot care device is a robotic monitoring device (Silhouette Monitoring Sensor, King Tsushin Kogyo Co., Ltd.), which was developed during a five-year national project [4, 5] launched in 2013 for the development of assistive robots in priority areas. This product monitors a person on a bed through silhouette images, and operation logs and alert outputs are recorded in an installed memory card [8].

The usage of the device is dependent on decision-making on the part of the user. An understanding of the device function, target users and safety of the device ought to play a central role in making such decisions; however, there were no previous studies conducted to investigate the level of understanding of the users when they made decisions. In this study, we also explore the levels of user understanding with respect to device function, target users and safety (residual risk); in addition to investigating how users made their decisions.

We conducted interviews in monthly meetings at the target care facility. The first interview was conducted prior to the start of use of the monitoring device. In this meeting, the possibility of using the new monitoring device was discussed. After use of the monitoring device started, we conducted interviews three times during the monthly meetings over a three-month period.

Even if we were to just use the logger data, we could detect any changes in device usage pattern, but it would be difficult to know the reason for the change. But if we combine the logger data with the results from the interviews conducted in the monthly meetings, we can determine the reason for the change. For example, we found a clear reduction in the number of operations during the three months. This type of reduction in operation might be regarded as users simply refusing to use the device; however, from the interview results, we determined the cause of the reduction was due to a better understanding of the characteristics of appropriate target usage of the system, and users thus deciding to limit the usage of the monitoring device.

In the rest of the paper, as the background to this work, we summarize previous studies on the monitoring of the elderly, and show the importance of the purpose of the monitoring. Appropriate selection of technology candidates is also an important factor, but we do not have many commercial products. Next, the method of our investigations and the results of the work are presented, before concluding this paper.

2 Background

This paper focuses on a monitoring system for elderly people; therefore, with respect to the background of this paper, we discuss the monitoring of elderly people in terms of necessity and resent research activities.

Monitoring literally means “to observe someone carefully for a period of time under a clear purpose,” which determines target behavior or the situation of the observation. Observation frequency depends on the purpose, as well. Response actions to monitored behavior or the situation ought to be decided beforehand.

The main purpose of monitoring elderly people is not to restrict their freedom, but to keep them safe while they go about their daily activities. If they are in an unsafe situation, we cannot give them permission to act freely as they want. This means that maintaining a safe situation is important in maintaining the target person’s independence and dignity in living, therefore, we stress the safety of elderly people.

The most fundamental way to monitor target elderly people is direct visual observation, which means that someone stays in the same space with the target people and observes directly on the spot. If it is possible, it is surely the most reliable monitoring method; however, this requires almost the same number of observers as the number of target elderly people.

The alternative is to use an indirect or a remote monitoring system, which enables one to observe someone from a distant place or a different room. A basic remote monitoring system consists of: (1) data acquisition system, (2) data processing system, (3) end terminal, and (4) communication network [9]. Some recent system have a function to (5) store data sets of monitored behavior or situation (See Fig. 1).

Fig. 1.
figure 1

Schematic of basic monitoring system, which consists of: (1) data acquisition system, (2) data processing system, (3) end terminal, (4) communication network, and (5) data storage (database).

Full size image
  1. (1)

    Recently, there are a large number of acquisition system candidates. Some systems include a function to sense essential vital signs such as an electrocardiogram or heart rates, as well as video images of the target persons. The sensor technology varies from sensors attached to the body (with-contact) to ambient sensors attached to the environment (contactless). The with-contact system has advantages in that it enables stable sensing; however, it has disadvantages in that it can be uncomfortable, or that it places restrictions on target users daily activities. For its part, the contactless system does not require direct contact with the body, which might cause discomfort; but the downside is that the system might be affected by external circumstances, such as sunlight for the video system. The technology used in such data acquisition systems has been summarized in several review articles. For example, in [10], inertial sensors, footswitches, pressure sensors, physiological sensors, and GPS sensors; and in [11], with respect to contactless sensors, passive infrared (PIR) sensors, video sensors, sound sensors, floor sensors, and radar sensors are listed. Combinations of contactless sensors and with-contact (wearable) sensors are also explained in [11]. Each technology has its merits and demerits, so the users of the monitoring system are required to select an appropriate technology or combination of technologies for the purpose of monitoring.

  2. (2)

    The data processing system receives output data from the data acquisition system, and processes the acquired data to recognize the target behaviors or situations by the use of algorithms appropriate for the purpose. To process the acquired data, the user is required to set some parameters to fit the monitoring purpose. Such parameters might be required for the data acquisition system as well. Examples of these parameters include thresholds to the output of the sensing device and the frequency of monitoring. After being processed, outputs of this system are transmitted to the end terminal.

  3. (3)

    The end terminal is a personal computer or a smartphone at the place of the observer, who receives the processed data and makes decisions with respect to the next actions to keep the target elderly person safe.

  4. (4)

    The communication network connects these components. The wireless communication network is relatively convenient; but it has the possibility of being affected by connection problems, which reduces the reliability of the system.

  5. (5)

    The data storage stores the monitoring, such as monitored behaviors or situations, and the times of the behaviors or when the situations happened.

As mentioned before, the user needs to select an appropriate combination of these elements; however, we have limited candidates of commercial products. We, therefore, have to try those products that are on the market to find the appropriate product and the appropriate usage of the product to fit the monitoring purpose.

3 Methods

The target robot care device is a robotic monitoring device (Silhouette Monitoring Sensor, King Tsushin Kogyo Co., Ltd.) (See Fig. 2, Fig. 3).

Fig. 2.
figure 2

(Photos courtesy of the company)

Photograph of the monitoring sensor product (Silhouette Monitoring Sensor, King Tsushin Kogyo Co., Ltd.). Left is the sensor unit, and the right images show the applications on a smartphone and a tablet.

Full size image
Fig. 3.
figure 3

(Screen images courtesy of the company)

Captured screen image of the application of the monitoring sensor product (Left), which shows a silhouette image of the monitored person on a bed, along with listed monitored results of four different rooms. The right panel explains that the leftmost of the three icons shows a recognized behavior of the monitored person.

Full size image

This study was approved by the Institutional Review Board on Ergonomic Research of the National Institute of Advanced Industrial Science and Technology. The investigation was carried out in a long-term care facility in Tokyo with the approval of the facility manager. Necessary information regarding the investigation was given and informed consent was obtained from each participant before starting the investigations.

The monitoring sensor device has a logger function to record user operations or alert outputs. The device itself has a function to record video images of the target elderly person on a bed, when the alert occurred and predetermined timing. However, we could not use video images in this work, because the Institutional Review Board did not approve video usage.

3.1 Investigations on Target Device Usage

The monitoring device records each user staffer operation, such as the start and termination of the alert function, and power switch operation. Alert outputs and responses to the alerts also are recorded.

The facility started using five sets of monitoring systems on August 4, 2019. We collected recorded data from the memory cards on November 8, 2019.

3.2 Investigations on Device Users’ Decision-Making

We conducted interviews in monthly meetings at the target long-term care facility, as mentioned in the previous section. The first interview was conducted prior to the start of use of the monitoring device on July 9, 2019. In this meeting, seven lead staffers and the facility manager discussed the possibility of using the new monitoring device. At the end of the meeting, we asked them about the following:

  1. (1)

    Levels of understanding (1: worst – 5: best)

    • Monitoring device functions

    • Expected characteristics of the target elderly people

    • Residual risk of the monitoring device

  2. (2)

    Presence of appropriate target elderly people in the facility (present/absent)

  3. (3)

    The criteria for selecting the target elderly people

  4. (4)

    Expectations to use the monitoring device on the target elderly people

  5. (5)

    Decision-making (Use the device/do not use the device/others).

Use of the device started on August 4, 2019. After the facility started using the device, we conducted interviews three times during the monthly meetings over a three-month period (September 4, October 8, November 8, 2019). In these meeting, the users mostly talked about their findings regarding usage of the monitoring device. At the end of the meeting, we asked about the following:

  1. (1)

    Levels of understanding (1: worst – 5: best)

    • Monitoring device functions

    • Expected characteristics of the target elderly people

    • Residual risk of the monitoring device

  2. (2)

    Changes in their work procedures, and the reason for the change.

  3. (3)

    Change in their ideas with respect to the properties of appropriate target elderly people, and the reason for the changes.

  4. (4)

    If a user or the facility decided to completely stop or temporarily suspend the usage of the device, the reason for stopping or suspending usage of the device.

4 Results

4.1 Investigations on Target Device Usage

The left image in Fig. 4 shows the distribution of the times (hours) when the facility staffers started and ended operations of the alert function over a three-month period. This graph shows that a large percentage of start-end operations were conducted late at night or early morning. Figure 5 shows the distribution of duration times (minutes) for each period in which the alert function was activated. We can see that short durations of monitoring occurred a lot of the time. These results indicate that short monitoring periods frequently occurred, and it is possible that these short monitoring troubled the staffers. We cannot determine the reason for these short monitoring periods solely from the logger data sets.

Fig. 4.
figure 4

The graph on the left shows the time (hours) distribution of start and end count of the alert function, and the graph on the right shows the count of duration period (minutes) for each alert-activated monitoring period.

Full size image
Fig. 5.
figure 5

Distribution in number of alert-activated monitorings per day.

Full size image

Figure 5 shows the distribution of the number of alert start events in one day. At the beginning of the investigation in August, the number of alert activations was high, but the number decreased in the following months. Note that they did not turn off the power of the system, which means that they can watch the video image from the end terminal device.

4.2 Investigations into Device Users’ Decision-Making

Table 1 shows the averaged levels of understanding to each question. The number of participants in the four meetings was 7, 7, 5 and 7, respectively. At the first meeting, which was before using the monitoring device, the number of answers was relatively low; however, they decided to use the monitoring sensor.

Table 1. Averaged levels of understanding of the staffers obtained in the meetings.
Full size table

In the subsequent meetings, which were held after the start of use of the monitoring device, their levels of understandings improved. This indicates that the facility staff had learned about the monitoring device through actually using the device.

At the second meeting, which was the first meeting after the introduction of the device, the facility staff said they had the impression that the device issued alerts too frequently. At this point, by asking about the details of the alerts, we found that they had not changed the parameters to fit the target elderly persons. We, therefore, advised them to consult the company about providing lessons on appropriate parameter settings. At the next meeting, they mentioned that the frequency of alerts had been reduced.

Figure 5 shows that the number of monitoring device usages decreased; therefore we asked the reason for the change in usage. The answer we received was that the facility staff mostly used the monitoring device only for remote cameras, which enabled them to monitor the target elderly persons from a different room. The staff said that they had found the best advantage of using the device in their case was that they could assign a priority to a situation by watching a remote monitored video image, and so they did not use the alert function of the system.

In the final meeting; however, the facility staff said that their usage of the device, which was to use the device for a remote camera, created some confusions with respect to the privacy of the monitored elderly persons. There have already been some discussions on the ethics of monitoring in elderly care facilities, and in the United States, some states have regulation governing the use of camera devices in care facilities [12]. We need more discussions on the ethics of such monitoring products.

5 Conclusions

Both the monitoring of elderly people and support by caregivers are important in the coming days in our country. Remote monitoring robots, which have the ability to recognize a target person’s behavior are one of the key expected solutions to this issue. This work is limited to an investigation conducted in one elderly care facility, but we could see that the usage of such devices is not limited to the way the maker company expected. Our findings suggest that determining the appropriate usage of newly developed products is a key factor in the success of the effective usage of the device. As a matter of fact, some previous works have stressed the importance of “re-invention,” which has the potential to accelerate the acceptance speed of newly developed products [13].