Socially Dependable Design
Socially Dependable Design: The Challenge of Ageing Populations for HCI. Mark A. Blythe Andrew F. Monk Kevin Doughty
1
Socially Dependable Design
Socially Dependable Design: The Challenge of Ageing Populations for HCI. Mark Blythe, Andrew F Monk and Kevin Doughty Centre for Usable Home Technology, University of York
[email protected] [email protected] [email protected]
Corresponding author: Mark Blythe Department of Computer Science University of York York, YO10 5DD, England
Direct Line 01904 434764 Fax: 01904 432767 Email:
[email protected]
2
Socially Dependable Design
3
Socially Dependable Design: The Challenge of Ageing Populations for HCI.
Mark Blythe, Andrew F Monk and Kevin Doughty Centre for Usable Home Technology, University of York
Abstract This paper considers the needs of an ageing population and the implications for Human Computer Interaction (HCI) research. The discussion is structured around findings from interviews with medical and care professionals and older people. Various technologies are being successfully used to monitor for falls and other emergencies, and also to assess and manage risk. The design of this technology is currently driven by a medical model of client needs and takes little account of the social context of the home. The design challenges for HCI are to make this technology attractive, provide privacy, allow informed choice and reduce rather than increases the isolation currently felt by many older people. It is argued that the ageing population presents a fundamental challenge to HCI in the need for socially dependable systems. Socially dependable systems take account of social context, the need for sociability and are accessible to all who need them. Keywords Ageing population, independent living, telecare, assistive technology, smart homes, dependability, hazard, risk management, community, responsibility, privacy.
Socially Dependable Design
1
4
Introduction
1.1 Technology and the Ageing Society You live, wherever you are, in an ageing society. By 2050 there will be more people in the world over sixty than there will be people under age fifteen. This is a historic reversal of proportions and it is unprecedented in human history (United Nations, 1999). It is a global phenomenon, occurring in both the “more” and “less developed” world. Declining fertility rates and increased longevity will have far-reaching and profound consequences, year on year, for the rest of our lives. In 1999 15% of the European population were eighty years of age or older; in 2050, barring holocaust or pandemic, that will have doubled. It is statistically likely that we will live longer than any other generation could have possibly hoped or feared. Lifestyle magazines may flatter us that, because we are all living longer, thirty is the new twenty, fifty the new forty and so on, and to an extent this is true. Age is a social as well as a biological construct. Indeed the term “elderly” is problematic: many people do not like it when it is attached to them, it may not reflect their abilities and it can be stigmatising. This paper uses the term current in UK government literature “older people” meaning people over sixty, although this is far from value neutral. While it is important not to stereotype people and assume that older people are necessarily frail, it is also important not to romanticise or patronise older people. Of course we should not assume that all older people will become less able but we cannot ignore the facts of our inevitable physical decline or underestimate the challenges that the ageing society presents in terms of provision, care and support. There are approximately 6 million informal (that is unpaid) carers in the UK, ninety percent of these are family members and most are women, many of these are themselves older people (Age Concern 2002). Formal and informal care provision is already strained. How then will the ageing
Socially Dependable Design
5
population be cared for in the coming decades? What new forms of support will arise when there are twice as many people over sixty as there are now? The answers to these questions are increasingly formulated in terms of technologies for independent living. Older people will live in their own homes for longer with technological support. Institutional care is expensive and governments are well aware of the potential savings to be made if people continue to live in their own homes. The UK government, for example, has set a deadline of 2010 for telecare to be available in every home that needs it, although there is considerable doubt about whether this deadline will be met (Curry et al 2002). Perhaps the most prevalent current technology used in support of independent living has a purely monitoring function. Quite frail older people are given the confidence to live independently through telecare devices that automatically call for help should they fall or experience some other domestic emergency (see section 3.2). Technology can assist with activities of daily living by compensating for problems of mobility, manual dexterity, also for sensory and cognitive deficits. If you have mobility problems, so called "smart home technology" can open a door for you or allow you to control lighting and so on (see for example van Berlo 2002). Deafness can be ameliorated through specialised communication devices such as text phones. There is also much current interest in reminder and advice systems for people with cognitive problems. The market for this kind of technology is relatively small and so there are not, in general, the cost savings that come with mass markets. This problem has been addressed by a movement known as Inclusive Design, promoting the idea that all products should be suitable for the widest population (Coleman 2001). The principle here is that, with ingenuity, it should be possible to design technologies that
Socially Dependable Design
6
both older and younger people can use. Power tools designed for people with arthritis can be equally attractive to people with no such problems. 1.2
HCI and the needs of older people
HCI researchers have reacted in two main ways to the needs of the increasing population of older users. One response has been to characterise the reduced cognitive, sensory and motor capabilities of older users, as a population, in comparison to a younger population of users, and then to adapt user interfaces accordingly. This parallels the development of technology to assist people with their activities of daily living noted above, but where the activity concerned is limited to "using a computer". Accessibility legislation in the UK now requires websites to be usable by people with visual disabilities, i.e., they must be accessible to someone using a screen reader. "Accessibility Interfaces” was a theme at the CHI 2003 conference New Horizons featuring a number of important innovations. Such work has the laudable objective of increasing the accessibility of general purpose computer systems. However, to concentrate solely on personal computers and the internet would be to overlook the tremendous opportunities offered by new developments in ubiquitous computing and digital communication devices. It is important that parts of the population are not excluded from the benefits of new technologies because of poor ergonomic design. However, there is another approach, which is to ask whether we really understand the needs of older people at a higher level than simple ergonomics, that is, to set out higher level requirements for a class of systems that use technology to support older people living independently. These requirements will drive the development of new products exploiting ubiquitous computing and communication devices in the service of improving the quality of life of older people. This is the topic of this paper, structured around a number of
Socially Dependable Design
7
interviews conducted with care professionals and older people in a small town in the north of England. It begins with a discussion of safety by identifying some of the hazards and risks of independent living for older people, drawing on Accident and Emergency (A&E) admission statistics and the experiences of ambulance crews. It goes on to consider the problems of managing risk with technology through a case study of a Warden Call service and interviews with Occupational Therapists and older people. Finally, it considers the issues of isolation and costs. What emerges is a need for socially dependable systems. Socially dependable systems take account of social context, the need for sociability and the principle of open access for all.
2
The interviews
The home presents a range of diverse challenges for ethnographers seeking to inform design (e.g. Blythe et al 2002, Crabtree 2003). We adopted an approach that began with a review of relevant statistical data from Social Trends, the annual UK government publication on demographic changes. This was supplemented with data from the UK Home Accident Surveillance System (HASS 2001) on A&E hospital admissions for 1999. (The UK government stopped funding for the collection and analysis of these data in 2003). These sources gave us some understanding of the key quantitative issues in terms of ageing and the hazards of living at home. These data were further explored in a series of interviews with medical and care professionals. Data collection was, then, staged, moving from government statistics, to professionals and then to users themselves. There were three sets of qualitative data collected for this study and three methods used: statistical prompts; ethnographic observation and interviews; and the technology biography. All three draw on Flanagan’s critical incident method
Socially Dependable Design
8
(Flanagan 1954). Statistical Prompts were used as the basis of a series of interviews carried out with ambulance crew over a number of daytime and evening visits to the station. The statistical prompts were designed to elicit stories that would illustrate the HASS statistics. The interviews began with questions about the most common call outs they attended for older people. These were followed by prompts based on the most frequent types and locations for accidents resulting in A&E admissions. Further interviews were carried out with Occupational Therapists (OTs) employed by the Local Authority. These OTs were responsible for assessing the needs of people with disabilities living independently and supplying them with equipment designed to enable them to live in the community longer. A more in-depth ethnographic study was conducted with a Warden Call service. Here customers’ homes are fitted with an alarm system operated by telephone and pendants so that calls can be made to a primary carer or a warden at any time. These care professionals were asked about the kinds of problems their clients faced and how they were able to respond to them. A researcher accompanied a mobile warden on her home visits for the day, this offered greater insight into the warden’s concerns and role and also afforded introductions to older people living independently. Each of the clients visited agreed to take part in a brief informal interview that focussed on their daily routine and any current problems they were having. The technology biography (Blythe, et al 2002) was used to interview older people in their own homes. Table 1 provides a summary of professionals and older people interviewed. Of course, we can make no statistical argument concerning how representative this sample is and findings are not generalised to other contexts.
Socially Dependable Design
9
PLEASE INSERT TABLE 1 ABOUT HERE
Client Confidentiality is an important element of care and health work and for this reason the names of all participants have been changed. The next section will outline some of the hazards and risks faced by older people living independently. Unless otherwise indicated italicised quotes are taken from recorded interviews. Two quantitative studies are described in Section 4 where the method for these studies is described with the results obtained. 3
Safety: beware of everything
3.1 Risks and Hazards In 1999 there were 1.2 million falls resulting in admission to an A&E ward. There were 500,000 collisions,100,000 burns, 35,000 poisonings and 16,000 chokings (HASS 2001). Of course a range of factors complicate the likelihood of an accident befalling a particular individual, and one of the most significant of these is age. Over 75 year olds are far more likely to have a fall resulting in an A&E admission than other adults. The consequences of a fall for older people living alone can be, literally, grave. Many of the ambulance crew interviewed pointed out that relatively minor accidents can result in very serious injuries: “their [older people's] skin will just tear, it’s like tissue paper. It loses its elasticity as you get older, the slightest knock can tear it” . There are, then, serious risks for any older person living independently. In safety engineering a distinction is drawn between "risks" and "hazards" where a risk is some likelihood and severity of harm and a hazard is something that might cause harm or damage (see Monk et al., submitted for a discussion of this distinction and how it relates to other definitions of risk). What then are the hazards in
Socially Dependable Design 10
the home, the things likely to cause the accidents that can have potentially disastrous results? The American and British societies of gerontology list the most common risk factors for falls found in 16 studies as: muscle weakness, history of falls, gait deficit, balance deficit, user assistive device, visual deficit, arthritis, impaired activities of daily living, depression, cognitive impairment and age greater than 80 years (Panel on Falls Prevention 2001). One of the hazards identified by all three groups of interviewees in this study was clutter: “too much furniture in the home. OT’s hate those, rugs, they always trip over them.” This kind of clutter is such a well-recognised hazard that the guidelines on falls suggest that older people discharged from hospital should have “a facilitated environmental home assessment” (Panel on falls prevention, 2001). Floors that are too smooth were seen by some of the paramedics as nearly as much of a hazard as clutter, especially if the older person is in their socks. Cables and loose wiring were, perhaps unsurprisingly, also identified as hazards: “They leave yards and yards of cable and they tend to pull it wherever they go. They get up to go to the loo, answer the phone and they fall over the cable.” Clearly there are a number of wireless technologies that might mitigate these hazards but it is important to note that in one sense safety can never be guaranteed. When asked if he thought the accidents he was called out to were mainly preventable one paramedic responded with an emphatic no. “It doesn’t matter what you do, there’s always a margin whereby you’re going to have an accident, just by turning, because of the nature of an elderly person, because they’re frail. Yes you can make things safe to a degree, but it’s not going to stop them knocking a pan off a stove or a kettle over.” In short, for older people in the uncontrolled environment of the home, almost anything could be construed as a hazard and almost any activity could be thought of as taking a risk.
Socially Dependable Design 11
3.2
Technology for risk management
Defining risk in terms of the likelihood and severity of harm makes clear how risk is to be managed, that is, reduced to a "tolerable" level. If the consequences of harm are severe but can be made very unlikely, then the risk may be seen as tolerable. Likewise, if the likelihood of harm is moderate but the consequences of that harm can be made negligible then this might also be thought of as tolerable. Risk management then is a process of identifying intolerable risks and then taking measures, either to reduce the likelihood of harm, or, to reduce its severity. The various technological solutions proposed to reduce risk can be seen in this light. The monitoring technologies described in the introduction, for example, work mainly by reducing the severity of harm. They do not prevent accidents; they aim instead to get help quickly when accidents occur (Williams et al. 2000). A detector communicates some potential emergency to a specialised telephone sometimes termed a "carephone". This automatically contacts a call centre where an operator can summon help. These systems have grown out of community alarm services where the only "detector" is a button or pull cord to summon help from an onsite warden. There are a wide variety of sensors that wirelessly communicate with the carephone. The carephone is able to identify which sensor has been activated and to pass this information on to the call centre. Personal triggers can be worn as pendants, brooches or on the wrist. Sensors can be bought at a reasonable price to detect: flood, extreme temperature, carbon monoxide, natural gas and smoke. An important criticism of the simple telecare systems described above is that they are purely reactive, getting help when something goes wrong but doing nothing to prevent things going wrong in the first place. Telecare suppliers are now beginning to market more sophisticated sensors that move some way towards answering that
Socially Dependable Design 12
criticism. The bed occupancy sensor, for example, takes information from a clock and a pressure pad and can infer that a client has got out of bed at a time when they are normally sleeping. It can be used to switch on lights guiding them to the bathroom (preventing falls) and raising an alarm via the carephone if they are not back in bed within a previously set duration (reacting to a potential fall). The "wandering client" sensor is designed for people in the early stages of dementia. It combines data from movement detectors and a clock to infer that the client is about to leave the house at an inappropriate time. It can then play a warning message, perhaps recorded by a relative, in an attempt to make them realise what they are doing. Monitoring systems are designed to provide confidence: confidence for the client, that if they fall they can get help; confidence for the client's relatives that the client is safe, and confidence to authorities in health and social services responsible for the client's health. If they are successful in this aim a purely reactive system could prevent the downward spiral to dependency often experienced by older people after a fall or stroke. However, even the simplest monitoring systems raise difficult, and yet to be resolved, problems for HCI research in terms of how the technology can be made to fit the social context of the home. Much of this technology comes from manufacturers used to designing equipment for medical settings and is often ugly or stigmatising when placed in a normal home. The fall detector, for example, is designed to be worn on a belt. It senses a jolt and then a change of attitude from vertical to horizontal. Many of the clients at risk of falls are older women and may not want to wear a belt. More generally, many older people strongly resist the installation of telecare monitoring technology. The equipment that OTs offer may be rejected despite strong recommendations by health professionals. This OT quotes a common complaint ‘We don’t want our homes looking like a clinic or a hospital’.
Socially Dependable Design 13
Various solutions suggest themselves. Firstly, more attention should be paid to aesthetic design to make the technology attractive through the design of its form. A personal trigger could be made as jewellery, for example. Secondly, providing additional functions could lead to earlier adoption so making it more mainstream. This would have the side benefit of allowing people to become familiar with it before the onset of cognitive decline. This then is one element of socially dependable design. Monitoring systems will not achieve their potential to provide confidence and safety for all those that could benefit from their use until their design takes more account of the social context in which they are used, that is, the social context of the home. A design that has the potential to stigmatise its user is reinforcing a particular view of the place of older people in society. This issue is further developed in the next section which considers the design problems that arise from the sometimes contrasting needs of the different stakeholders involved: professional carers, relatives and so on.
4 4.1
Whose life is it anyway?
Freedom of Choice
Each of the professionals interviewed related stories in which older people were taking risks that they, as carers, did not necessarily approve of, but that they would not wish to control. The manager of the warden service for instance reported that one of the reasons that falls are so common is that “there’s a considerable number of customers who have drink problems.” One of the mobile wardens was sympathetic "if I was in my eighties and I had nothing to do I might reach for the bottle”. Although clutter is a well recognised hazard and one that OTs frequently point out, it is not
Socially Dependable Design 14
necessarily something that the older person will want to do anything about as this OT explains: “We do ask them to take rugs out of the way, and move furniture. Especially if you’re using a hoist, you need space. But again, quite often the customer doesn’t want that and you are in their own home so you can make suggestions but you can’t say ‘this is what you will do!" One of the least dependable aspects of telecare systems is the need to wear some form of personal trigger. Table 2 presents some data collected by the third author using telephone interviews in Scotland in 2003. Clients were telephoned at home during daytime hours and asked where their “pendant” was. All had been assessed as at risk and in need of a community alarm system. At the time, a pendant was the only form of personal trigger available and so the "correct" answer to the question “where is your pendant” was "round my neck". Only 21 percent of those telephoned gave this answer. All the other locations are unlikely to be accessible were the respondent to fall.
PLEASE INSERT TABLE 2 ABOUT HERE
The problem of people not wearing pendants also emerged very clearly from the interviews with personnel in the warden call scheme. Customers of the warden call service were advised to keep their pendants on at all times, especially in the shower, but they seldom did “What we advise when a new customer comes onto the system we say ‘hang your pendant by your bed at night or on your bedside table, but when you want to go to the loo during the night put it on’. But they don’t. They feel more independent not having their pendant on but it actually gives them more independence because they can get help if they need it” (Senior Warden). Although the pendant was
Socially Dependable Design 15
designed to give them more independence, it was sometimes choosing not to wear it that best illustrated this very quality. The freedom to choose is perhaps best viewed as a question of balancing risks. This was exemplified when Val, a mobile warden, visited a new customer to install the carephone. Nora was 79 and had osteoporosis, her flat was on the third floor of a local council estate. She had recently injured her leg and the only way up to the flat was via three flights of concrete steps. Val had tried to persuade her to move home but Nora had refused, she had lived there for thirty seven years: ‘it’s her home, it’s got a nice view, if anybody wanted to break in they’d need a big ladder’ so basically I can’t shift her.” On more than one occasion during the visit Val told Nora “if you do have a fall on those stairs it’ll be your last”. This was a risk that Nora was prepared to take because the risks of moving were perceived, by her, as being much larger. Independence is an emotional as well as a functional state. Whether the freedom to choose to accept risks judged by authorities to be important is acknowledged as a right or not, the fact remains that people will take risks with or without permission “people will battle on. They will continue to clean windows, climb ladders, risk taking - people make their choices” (Age Concern Rep. Field Notes). Risk management involves a complex set of decisions. How likely is the risk? How severe would the harm be if the risk were realised? And crucially, how great is the cost of risk management? For the older person, a loss of privacy, independence or dignity may be too high a price to pay for the mitigation of even a high likelihood of severe physical harm. This view of the balance of risks is at odds with the view of statutory or medical authorities who have different priorities to the older person when it comes to managing risk. The families may have yet another view. Here, the manager of the warden service reflects on the different priorities of clients and their family members
Socially Dependable Design 16
“often the families are keener and often the health professionals are keener [on installing equipment] just because they think people need our equipment, they have to understand we can’t make anybody have our equipment, that’s infringing that person’s rights.” While these professional carers will always try to give the client the last say in decisions about the technology that gets installed in their houses, as we shall see in the next section, they also have a "duty of care" that forces them to focus on issues of safety that may conflict with the priorities of the client. Another problem is informed consent. Section 4.3 discusses some new developments in monitoring technology where it may be hard for the client to understand what the system is doing. If they don't understand the form of the data that the system is collecting and transmitting to other people how can they be said to have approved its installation? 4.2
Responsibility and liability
The principle of allowing the older person freedom of choice can sometimes conflict with the formal responsibilities of care professionals who have a "duty of care" that lays them open to litigation if certain procedures are not followed. The occupational therapists interviewed noted that there were a number of accountability procedures that they adhered to in the assessment process (largely signed documents for paper tracking of advice and decisions). The warden call service was liable to prosecution in the event of a failure of the service. The call operator at the warden service, then, was engaged on a daily basis, in a form of risk assessment. The call operator noted that all of the calls that came into the centre were recorded. “It’s for our safety as much as anything. If something goes wrong, if someone says ‘oh she pressed her button and nobody answered’ we can get the call page up and say ‘well I’m sorry no calls came through from her on that day’ or ‘there was a call that came through on that day but
Socially Dependable Design 17
the lady said it was a false alarm.’ One of the key problems for the operator is assessing what the caller says and selecting an appropriate response. Here, a senior warden and co-ordinator of the service relates a typical complaint. “General complaints are ‘why didn’t you go to my mum when she put a call on?’ Luckily all the calls are taped on voice recordings of all the calls that come through. So when you get that call from the mum and you say ‘is everything alright, do you need any help’ and when she says ‘no, I’m fine’ [laughs] and when the daughter phones up and say’s ‘my mum really needed a doctor’ and then you say ‘Well unless she tells us that, we can’t do anything about it’.” This is a crucial problem not only for the warden service but for any computer supported system which relies on self report and reluctance to report is an especially difficult problem with regard to the assessment of need. This may be because of denial, "I am not old", "I may not really be ill" or a general reluctance to "bother anyone" (Doughty and Williams 2002). Often simply asking an older person if they are having difficulty with an activity of daily living is not enough and keen observation is necessary in the assessment of need as this OT relates “I’ve got a lady who hasn’t got very good grip and she always, when she has a cup of tea with you, she only ever has the cup half full because she has a little bit of a tremor so you can say ‘why do you have your cup half full, is it quite difficult? […] Again it is observation.” It is difficult to imagine an algorithm, however complex, that would replace Mary observing a client making a cup of tea during an assessment or the call operator judging a caller’s tone of voice as they describe their problem. Sometimes there is no substitute for human contact. Only human beings or organisations take responsibility, whether it is for deciding what technology is needed as part of a larger care package, as in the case of this OT, or for day-to-day decisions made on the basis of data provided by technology
Socially Dependable Design 18
as with the call centre operator. Taking responsibility for something lays you open to liability if something goes wrong. As we shall see in the next section, adding technology to the equation adds a new party to liability, the technology provider. 4.3
Privacy, informed consent and responsibility as design issues
One response to the problems of "denial" (refusing to admit that you need the technology) and "non-compliance" (e.g., not wearing a fall detector or the alarm pendant) has been to use subtler monitoring technologies. Inactivity monitors take input from pressure pads and passive infrared movement detectors and assess the general level of activity. If this is not as expected, given the time of day etc., then an alarm can be raised. So called "lifestyle monitoring" systems have been proposed to take this a step further (Williams et al. 2000; Perry et al. 2004). These systems monitor activities such as running water, opening food cupboards and the fridge, moving from room to room, and so on. Computer algorithms can then be used to make inferences about whether the client is eating or drinking regularly, their general level of health etc. To illustrate how this works consider Figure 1 which presents some low-level data displayed on a hour-by-hour basis for one 24 hour period. It was collected by the third author in an intermediate care facility in the north of England in 2002. A number of simple sensors were set up in a flat occupied by an older person. The sensors were connected to a computer that recorded the data thus generated for a number of weeks.
PLEASE INSERT FIGURE 1 ABOUT HERE
Socially Dependable Design 19
Figure 1 then records the activity of the older person throughout a 24 hour period as reflected in the output of these sensors. Such plots can be used to infer room occupancy and the participation in certain events (such as eating, bathing and going to bed) at appropriate times. This may demonstrate that an individual gets up during the night, as in this case, or the fact that they leave home sometimes at unusual times. Many problems manifest themselves as behavioural problems over longer periods of time. For example, the length of time spent in bed every night, together with the time spent not on one's feet, can be used to show subtle changes of capability in the development of a disease. In Figure 2 Client 1 suffers from arthritis which limits the time spent standing or walking. The result is that the time spent in bed or sitting on a chair slowly increases over a period of months. Client 2 has a mild form of dementia which is gradually getting worse. This can be seen as a slow decline in the time spent in bed and in a chair as she becomes more anxious and agitated. These data were also collected, using similar equipment to that used for Figure 1, by the third author in two different locations in North Wales in 2000.
PLEASE INSERT FIGURE 2 ABOUT HERE
This kind of sensing is much less visible to the client than a fall detector or a flood detector as what constitutes an alarm is hidden from the user. While this may get round the problems of denial and non-compliance, it does so by effectively taking away the client's right to make an informed choice. "Denial" and "non-compliance" are medical terms reflecting a medical perspective on the priorities for risk management. A more user-centred approach would focus on design issues concerned with privacy, acceptability and informed choice. What, if any, raw activity data goes
Socially Dependable Design 20
out of the house? Who is able to inspect it and is the client aware of this? When an inference is drawn from activity data, e.g., "you are on the way to the toilet" or "you are in need of more care", are you aware this inference has been drawn and who it has been communicated to? These are hard design problems that HCI researchers could make a real contribution to. Lifestyle monitoring also brings into sharp relief the design issues facing all monitoring technology with regard to responsibility and liability. From a user-centred design perspective this is a question of how automated inferences from sensors are presented and acted upon outside of the home. Given that such inferences are always going to be fallible what are the consequences of false alarms? How does the design of the procedure at the call centre allow for false inferences? Imagine a scenario where a daughter can text a request for information about the status of her ageing mother who is slightly demented and who has a history of leaving the house at inappropriate times. Without this service the daughter is forced to visit and check that her mother is in bed and asleep every night at 11.00 p.m. Now let us say that the message sent back to the daughter's mobile phone is "Mrs Jones has been in bed since 10.30, you will be sent another text if she leaves the bed for more than 20 minutes before 7.00 am". If the data this inference was drawn from was faulty and the mother was actually in a confused state on a busy road the manufacturer of the equipment or the provider of the service could be in trouble. To cover themselves the manufacturer could guard the inference with some sort of probability "Our data indicate that there is a 95% probability that Mrs Jones has been in bed since....". but this may make the system less attractive. Medical instruments do not normally give this kind of information. A digital thermometer could theoretically say "you are well" instead it says "36". This puts responsibility back in the court of the user to obtain the
Socially Dependable Design 21
knowledge needed to correctly interpret the data. Similarly, a lifestyle monitoring system that aggregates activity data over a periods of days and makes inferences about the general health and well being of an older person could present the data summary as a figure, e.g., "the weighted activity index for Mrs Jones is 55.6". Returning to the issue of informed choice, a strong argument can be made that, as well as the client knowing what information has been sent, the client should understand how it will be interpreted. Medical and engineering models of dependability are not sufficient to address the problems of privacy and independence that are inherent to these technologies. Socially dependable design in this context means taking a user-centred perspective that takes account of the full social context in which a lifestyle monitoring system will be used. HCI researchers are used to considering the whole socio-technical system and the differing needs of multiple stakeholders. HCI research could again make a contribution in these difficult areas of designing for privacy, informed consent and responsibility. 5 5.1
Technology and community
Between Independence and Isolation
Age Concern identifies isolation as one of the major problems for older people, some of whom feel a crippling sense of loneliness (Age Concern 2002). The demographics indicate that this problem will in future get worse rather than better. Single occupants are the fastest growing demographic group in the West. Ulrick Beck argues that new forms of “individualised” living will have dramatic impacts on the care needs of future generations of older people, these new forms of living include “living alone, single parenthood, non-marital cohabitation, childless marriage, serial marriage or
Socially Dependable Design 22
“living apart together, with a partner in separate dwellings” (Beck: 2002: 130). Although the medical and care professionals interviewed for this study noted that incidents of total isolation amongst older people were much rarer than in previous years, it is still possible for people to “drop off the map”. Neighbours who have not seen someone around for a while, or postal workers who notice a pile up of mail, might alert the ambulance service to such isolated individuals: “we might be called from a third party saying ‘oh, I haven’t seen them around, their curtains haven’t been drawn for a while’. Because they’re so detached from society sometimes they go into a hermit like state” (Paramedic). The living conditions of such isolated individuals can be shocking as this emergency medical technician relates “you go in and you can tell that humans from outside have not been in there for possibly a couple of years, they store papers which is a big fire hazard. Also there’s excrement on the floor, cat food, it’s squelching under your feet, you’re sticking to the floor. One of my first jobs, this guy was in his chair unable to move for some reason, we weren’t too sure if it was medical or not, but he hadn’t had any visitors so he’d just go to the toilet for maybe the last 3 weeks in the chair, he got like drenched because the urine had gone down to his shoes, but nobody had been round to him.” This last incident is a stark illustration of independent living as social isolation. One of the most important aspects of the mobile warden service is personal contact. The manager of the service noted that alarms were often used not for emergency calls but for company. The highest category of call logged on their system was “false alarm”. But the manager of the service, said she would not want to reduce the number of false alarms. Often false alarms were not false alarms, they were reassurance calls. “We have what we call our regulars. There’s one gentlemen who calls 3 to 4 times a night, in fact he’s made 300 calls on his last quarterly phone bill
Socially Dependable Design 23
[…] He’s lonely, he’s got a catheter and his nights are long.” Customers could say “Oh I just hit the pendant accidentally” and being able to ring for reassurance without having to say that this is what they were doing was psychologically important. For some, living in their own homes with the help of technology preserves social networks and is less isolating than moving to an institutional setting. For others, independent living may increase isolation. While ambulance crew and wardens noted that neighbours would alert them if they thought something was wrong, this does not necessarily indicate strong communities of people in daily ongoing contact, perhaps just the reverse. How, then, are older people to live in such communities without becoming isolated? The technological developments described in the introduction and section 3.2 are generally justified in terms of reduced costs. Some of the costly care provided by human carers can be taken on by machines. How can we ensure that the availability of new technological solutions does not increase isolation? 5.2
Costs, access and provision
Despite the existence of social provision in the form of a welfare state, care in 21st century Britain is ultimately something to be bought and sold. In Europe there is a “pluralist” system of access to assistive technologies, it is funded through a mixture of individual subscription, insurance, state benefits, and contributions form charities. In the UK there is no uniform policy, the manager of the warden call service noted that often customers could not pay their bills. This was dealt with flexibly “We don’t go down the road of whipping out the equipment, that would result in a letter to the councillor, Evening Press job [laughs]”. Other forms of provision would be investigated, charities, previous employers and so on. When asked about unmet needs, a senior warden said that these were usually financial “We just don’t have any control over that. They’re either in line for that [welfare] benefit or they’re not”.
Socially Dependable Design 24
Social class is a crucial indicator of whether an older person’s needs are likely to be met or not as the ambulance service area manager pointed out “if your mum’s ill and you’ve got a bit of money, then your mum is going to get the treatment because you’re either going to pay for it or shout out. In the poorer areas it’s more accepted.” He went on to note that in less affluent areas both sets of relatives may be working and so unable to make daily phone calls or offer other support. There are then issues of access to consider in the development of technology that addresses the care gap in the ageing society. Expensive solutions like smart homes may only be available to the relatively well off and out of reach of the majority of older people. It is for this reason that a principle of access for all must be a defining element of socially dependable systems. 5.3 Designing for social contact and social accessibility Advances in communication technology provide many opportunities for reconnecting people to the community and preventing isolation. The way that some users of the warden call service subverted its original purpose to serve their need for reassurance and human contact is illustrative of what can be done. That the manager of the service was quite happy with this is indicative of a general agreement about the need it satisfies. This section describes two uses of the telephone to reduce social isolation in a cost effective way: recreational telephone conferencing and a volunteer-based online shopping service. The telephone has an important role in home life (Lacohee and Anderson 2001; Crabtree, et al 2003). Telephone conferencing uses specialised network equipment (a conferencing switch) to connect more than two telephone receivers simultaneously. Telephone conferences are being used successfully to link up isolated individuals in friendship groups. Because the intelligence in this system is
Socially Dependable Design 25
in the network, rather than the telephone receiver in the home it is available to most older people; 95% of UK homes have landline telephones (ONS 2004). Hackney Borough Council use telephone conferences to link up groups of isolated older people as a part of their Friendship Link scheme. This befriending scheme supports elderly individuals through recreational telephone conferences and weekly one-to-one telephone contact. The content of the talk is not foregrounded in the accounts of the participants, the value seems to be in terms of human contact (Reed and Monk 2004). The second example of how the telephone can be used innovatively to reduce isolation cost effectively emerged from the field studies reported here. As previously noted, a mobile warden, was accompanied on visits to three of her clients. Each of the individuals visited had great difficulties with shopping and relied on carers, relatives or neighbours for help. Also, a number of participants in the field study pointed out how important the social aspects of shopping can be to older people. One of the drivers of the dial-a-bus service based at the ambulance station noted “it is a social thing, it means they’re not stuck in their four walls and they can actually go out and see somebody”. Jack, an older person living in warden controlled sheltered accommodation described the pleasure of shopping as “meeting people, talking to the shopkeeper, talking to people in the queue […] Even now when I go for me bloody paper sometimes can be quarter of an hour’s chat, you know, any shop I go in that doings”. After conducting the fieldwork discussed in earlier sections the first author devised and co-developed “Net Neighbours” to meet some of the needs identified. Net Neighbours allows volunteers to shop online for older people, providing them with a vital service and social support by phone. Informally, PC users sometimes place online orders for friends or relatives who do not have access to the Internet. The Net
Socially Dependable Design 26
Neighbours scheme aims to extend this kind of appropriation, making it widely available and dependable. Volunteers are registered and police checked with the local branch of the national charity Age Concern, they undertake training on the principles of the charity and the Net Neighbours procedure. Age Concern pairs them with an older person; they are introduced either over the phone or in person depending on preference. The volunteer takes orders over the phone and, crucially, has a chat as well. In the last five years there has been a significant decline in the number of hours that have been volunteered to UK charities. Net Neighbours allows people to volunteer from their office or home in “bite sized” chunks of time. The scheme is currently being piloted at the University of York. Eleven university volunteers have been recruited and after further trials the scheme will be expanded. Blythe and Monk (2005) describe the iterative, participatory design methods used in developing this service. As with the other technologies described in this paper there is a danger that such a scheme might increase an older person’s sense of isolation. For this reason volunteers are encouraged to chat to the older person as well as get their shopping; it is described as a part shopping and part-befriending scheme. If it is found that the scheme is effective in terms of getting requested groceries to an older person’s door but that it increases a sense of social isolation then Net Neighbours will have failed and could not be thought of as socially dependable design. 6
Discussion
It is an almost unquestionable orthodoxy in the UK that older people should live independently in their own homes for as long as possible. Undoubtedly it is something that most British older people want themselves, they do not want to go into care homes, they will fight for their independence. This is culturally specific and would be
Socially Dependable Design 27
looked on with horror in other areas of the world. Genevieve Bell’s ethnographic work in Asia reminds us of very different attitudes to age and ageing in non-western countries. In India and in China the notion of having older people live independently (i.e. alone) would be regarded as quite strange (Bell 2002). Why would they not want to live with their families? It could be argued that we in the West have transferred responsibility for our parents to paid care workers and are now preparing to transfer it to machines. A move towards socially dependable design might go some way to refuting this notion. This paper has looked at the various ways that technology can support independent living, illustrating individual needs through interviews with older people and the people who deal with those needs. As we have seen, the physical risks, their likelihood and consequences are well understood as are the hazards that lead to accidents in the home. Systematic procedures can be borrowed from other disciplines to manage these risks. Techniques developed for the design of "critical" systems can be adapted to include social and psychological issues. Recent work in HCI has focussed on the dependability of technologies for older people living independently (e.g. Dewsbury et al. 2003; Baxter et al., 2004) by broadening the concerns to include social and psychological issues such as isolation and privacy. With the increasing sophistication of sensors and the algorithms used to detect accidents, providing dependable accident detection systems has become a question of development rather than research. HCI knowledge can help in this process of technology development by providing user-centred design methods to ensure usability and acceptability. Although safety may be the main concern of some relatives and authorities with a statutory requirement to provide medical and social care, the older people themselves worry equally about other things such as loneliness, privacy and the
Socially Dependable Design 28
freedom to live as they wish. These are much more poorly understood issues where the user-centred approach of HCI researchers could bring great bonuses. For example, a better understanding of our need for human contact and how different communication technologies might best support that need could considerably improve the quality of life of many isolated older people. The need for a move from a medically oriented focus on safety towards a usercentred focus on quality of life is, then, one theme to emerge strongly from the fieldwork. Issues of access, provision and cost came up throughout the interviews. These are hard to solve, moral and political problems but that does not mean that technologists should not be involved in solving them. Two examples of socially dependable design were presented. Friendship groups using telephone conferencing and the Net Neighbours shopping scheme both fulfil the requirements of being available to all who need them and catering for social needs. These are not technologies that transfer the burden of care from people to machines; rather these are technologies that allow people to support people, they are in this sense, socially dependable. To a large extent access to technology comes down to cost and this in turn depends to a large extent on high volume leading to low unit cost. Inclusive design aims to ensure that “products and services address the needs of the widest possible audience, irrespective of age or ability”. Advocates of the movement argue that “getting things right for older users leads to real improvements in usability and customer satisfaction, which translate into better products for people of all ages” (Coleman: 2001: 37). Others have pointed out (Newell & Gregor 1999) that many of the most successful products were inspired by designs for "extreme" users with sensory or physical disabilities, the audio cassette, predictive text and central locking
Socially Dependable Design 29
for cars are often quoted in this connection. The inclusive design movement is an important one that has resulted in many innovative approaches to new and old technology. But a broader conceptualisation of the term “inclusive” might take into account not just different physical ability but different socio-economic power. Older people are excluded from technology not only by physical disability. Over 75 year olds are far more likely to suffer financial hardship than other age groups and may be excluded from technology simply because they cannot afford it. George Orwell (1965) once pointed out that lampposts are only of benefit to everyone because the wealthy cannot turn them on when they are approaching or off when they have passed. Bill Gates’ experience of old age will probably be very different to ours; but perhaps we can persuade him to make a few lamp posts rather than flash lights before we all head off into the dark. Some technologies are inherently beneficial to a whole society rather than certain individuals within it. When considering what a technology is for we also, implicitly, ask whom it is for: a knob may be for turning but is it for turning by someone with arthritis? Changing notions of community must be recognised as design challenges for systems that aim to support the ageing population. Questions like – is this a technology that will be of benefit to a whole community or just a part of it, should be paramount in the development of new computer systems. This paper has argued then for the necessity of socially dependable systems that take account of the social context of use, the need for sociability and the principle of access for all.
References Age Concern (2002) Concerned About Ageing? The needs of older people: key issues and evidence. London, Age Concern England.
Socially Dependable Design 30
Baxter, G.D. and Monk, A.F., Doughty, K., Blythe, M. and Dewsbury, G. (2004) Standards and the Dependability of Electronic Assistive Technology, in Keates, S., Clarkson, J., Langdon, P. and Robinson, P. (Eds.) "Designing a more inclusive world", Proceedings of the second Cambridge Workshop on Universal Access and Assistive Technology, London: Springer, 247-256. Beck, U. (2002) Individualisation: institutionalised individualism and its social and political consequences. Sage, London. Bell, G. (2002) ICTs in Asia: A cultural account. Proceedings of Asia Pacific Economics and Business Conference 2002, 479-489. Blythe, M. and Monk, A.F. (2005) Net Neighbours: adapting HCI methods to cross the digital divide. Interacting with Computers, 17, 35-56. Blythe, M., Monk, A. F. & Park, J. (2002). Technology biographies: field study techniques for home use product development. In L. Terveen & D. Wixon (Eds.), CHI 2002, Minneapolis. Extended Abstracts. City: ACM Press. 658-9. Coleman R., (2001) Living Longer: The New Context For Design. London, The Design Council. Crabtree, A., Rodden, T., Hemming, T. and Benford, S. (2003) Finding a place for UbiComp in the home. in Proceedings of UBICOMP 2003, Heidelberg, Springer-Verlag, 208-226. Curry R.G, Tinoco M.T, Wardle D. (2002) The Use of Information and Communication Technology (ICT) to Support Independent Living for Older and Disabled People. London, Department of Health. Dewsbury G, Sommerville I, Clarke K, Rouncefield M (2003) A Dependability Model of Domestic Systems, in Anderson, Felici & Littlewood (Eds), Computer
Socially Dependable Design 31
Safety, Reliability And Security: 22nd International Conference, Safecomp 2003, Proceedings, Lecture Notes In Computer Science, 2788, Heidelberg, Springer-Verlag, 103-115. Doughty K & Williams G.(2002) Towards a Complete Home Monitoring Solution, Official publication of the RoSPA Home Safety Conference, Stratford Upon Avon, November 2001, Birmingham, RoSPA. Flanagan, J.C. (1954) The Critical Incident Technique, Psychological Bulletin, 51, 327-358. HASS (2001) Home Accident Surveillance System Working For A Safer World: 23rd annual report of the home and leisure accident surveillance System 1999 data. London, HMSO, Department of Trade and Industry. Lacohee, H. and Anderson, B (2001) 'Interacting With The Telephone' International Journal of Human-Computer Studies, 54, 666-699. Monk, A.F., Hone, K., Lines, L., Dowdall, A., Baxter, G., Blythe, M.B. and Wright, P. (Accepted subject to corrections) Towards a practical framework for managing the risks of selecting technology to support independent living. Applied Ergonomics. Newell, A., & Gregor, P. (1999). Extra-ordinary human-machine interaction - what can be learned from people with disabilities? Cognition Technology and Work, 1(2), 78-85. ONS (2004). Social Trends 34 : Office for National Statistics. London. The Stationary Office. Orwell, G. (1965) The Decline of the English Murder and Other Essays. London, Penguin.
Socially Dependable Design 32
Panel on falls prevention (2001) Guidelines for the Prevention of Falls in Older People. Journal of the American Geriatrics Society, 49, 664-672. Perry, M., Dowdall, A., Lines, L. and Hone, K. (2004) Multimodal and ubiquitous computing systems: Supporting independent-living older users. IEEE Transactions on Information Technology in Biomedicine, 8(3), 258-270. Reed, D.J. and Monk, A.F. (2004) Using familiar technologies in unfamiliar ways and learning from the old about the new. Universal Access in the Information Society, 3, 114-121. United Nations (1999) Population Ageing 1999 United Nations publication, (ST/ESA/SER.A/179), Sales No. E.99.XIII.11. Van Berlo A. (2002) Smart Home Technology: have older people paved the way? Gerontechnology, 2, 77-87. Williams, G., Doughty, K., and Bradley, D.A. (2000) Safety and risk issues in using telecare. Journal of telemedicine and telecare, 6, 249-262.
Tables
Socially Dependable Design 33
Participants Key Concerns Identified Ambulance Crew
9
Falls, Isolation
Age Concern staff
3
Isolation, Fear of Crime
Occupational
2
Trip hazards, clutter
1
Independence
Mobile Wardens
4
Freedom to take risks
Older People
4
Isolation, Usability
Therapists Shop Mobility Manager
Table 1. Summary of participants interviewed and key concerns.
Answer
%
By my bed
31
In a drawer
22
Round my neck
21
Round a door
9
handle/peg Not sure
6
Somewhere else
11
Table 2. Data from a telephone survey of 200 community alarm users, "where is your community alarm pendant at this moment?"
Socially Dependable Design 34
Figure Legends
Figure 1. Low level lifestyle monitoring data. The x axis shows the hour of the day from 1 a.m to midnight, the y axis indicates the output from various activity sensors.
Figure 2. Lifestyle monitoring, hours per day spent in bed or bed and chair plotted to illustrate long term trends in two clients.