Oucsace.cs.ohiou.edu

Can a Computer be a Caregiver?
Peter Whitehouse, MD
Cynthia Marling
Richard Harvey, MD
Abstract
with memory or attention problems and progresses to in-volve language, skilled motor acts, and perception. Ac- This paper critically examines the issues involved in companying these cognitive problems are behavioral distur- computer-assisted and automated caregiving for patients bances, which vary widely but are quite common. These be- with Alzheimer’s disease and related dementias. First, thenature of the problems experienced by those who care for havioral disturbances include wandering, agitation and sleep dementia patients is explicated. Next, a high-level overview disturbances, as well as more serious and traditional psy- of the ways in which computers might assist these caregivers, chiatric conditions such as hallucinations, delusions and de- now and in the future, is presented. Finally, the ethical issues pression. Moreover, patients have varying degrees of insight raised by the advent of automated caregiving are explored.
into the condition that they are suffering from. Even earlyin the disease, they may not recognize their own cognitive Keywords:
der Care, Computing Ethics, Alzheimer’s Disease, AI in The onset of dementia is often insidious, and it may only be in retrospect that the problem is clearly defined to havebegun years before medical attention is sought. Several Introduction
changes in personality, memory and difficulty focusing at- Comforting and caring for people affected by dementia is tention are usually unaccompanied by any significant prob- a human challenge that many more individuals will face as lems with activities of daily living. As the disease pro- the world continues to age. Alzheimer’s disease, the most gresses, however, the cognitive problems grow both in sever- common dementia, is estimated to affect up to 50% of those ity and breadth, and the patient becomes unable to accom- over the age of 85 (National Institute on Aging 2000). The plish complex activities of daily living. In the severe stage of most rapidly growing segment of the populations in west- the disease, the patient loses most communication functions, ern Europe and North America are those over the age of is perhaps only able to repeat simple words, and cannot even 85. The rate of aging is even greater in the less industri- perform basic activities of daily living such as toileting and alized countries, such as China and India. Thus, more and more human beings around the world will face this challenge The act of caring for a patient with dementia requires pa- of providing care for relatives, friends and clients affected tience, creativity and fortitude. The illness often progresses by these devastating conditions. The question we pose in over five to ten years. The needs of the patient change, and this essay is whether computers can assist in the process of often the caregivers change as families are overwhelmed and providing care for victims of progressive cognitive impair- more and more formal caregivers are introduced to help fam- ment. When we first asked the question, “Can a computer ilies with their informal care. At first the patient may be at be a caregiver?” just a few years ago, the question seemed home, requiring minimal supervision, but then will require largely rhetorical. It seems less so now, but it is still not easy increasing attention and perhaps need community services to answer, as technical innovations, social expectations and like home visitors and day care. In the later stages of the ethical considerations interplay. We will divide our com- disease, patients frequently require residential care, which ments into the nature of the caring relationship, the current can take place in a variety of different organizations, in- and potential future use of computers, and the ethical issues cluding assisted living, nursing homes and special care units that arise with this envisioned future use.
specifically designed for patients with progressive cognitiveimpairment. These special care units are being developed The Nature of Care
without a clear sense of what defines them or what the crit-ical features are. Sometimes, there are only minor changes Patients affected by dementia by definition are impaired in in existing environment and programming. However, some more than one cognitive area. The impairment usually starts organizations have built entirely new buildings staffed in dif- Copyright c 2002, American Association for Artificial Intelli- ferent ways to avoid the medical model and create a space gence (www.aaai.org). All rights reserved.
that is friendly for patients with cognitive impairment. For example, devices are used to help patients recognize their own rooms, such as personal mementos in display cases be- More recently, we developed a CD-ROM at the UMAC side their doors. For patients who wander, facilities may for counseling caregivers and other family members of be designed with indoor and outdoor circular tracks that Alzheimer’s disease patients about their own genetic risk avoid dead-ends in which patients may enter other resident’s of getting Alzheimer’s disease. A simulated interview be- tween a professional and a caregiver is employed. Care- Studies show that caregiver burden is a significant com- givers can select frequently asked questions to get answers ponent of the caregiving experience (Ory et al. 1999). Care- about Alzheimer’s disease, genetics, genetic counseling and giver burden can affect family members as well as pro- fessional staff in the community and institutional services.
At the National Hospital for Neurology and Neurosurgery Contributing to the burden is the inability to communicate in London, we have developed CANDID, the Counselling effectively with dementia patients and the frustration that this engenders. Moreover, patients may be unable to rec- information, advice and clinical management service for ognize their caregivers, even their spouses of 50 years, and younger people with dementia that began operation in 1995 may heap abuse upon the individuals who love them most and who are the most intimate with them. In the later stages http://candid.ion.ucl.ac.uk/candid/ that provides information of the disease, patients have problems with personal hygiene for caregivers. There is also a team of trained nurses and and become incontinent of urine and stool. Increasing phys- counsellors who answer questions over the internet or by ical disabilities may prevent patients from moving about in- telephone. For patients who have registered at our hospi- dependently, so that considerable physical energy can be re- tal, we aim to provide sound clinical management advice, as While these early efforts are educational and advisory, The Current and Future Use of Computers
we are moving toward systems that can provide more di-rect care. One computer capacity we can take advantage At first glance, caring for a patient with dementia appears to of is memory, which is one of the difficulties dementia pa- be a uniquely human experience. Past personal relationships tients have. Computer memory can be used to supplement with the patient enables family caregivers to be sensitive to and improve the memories of patients with dementia. For the patient’s moods and needs. However, researchers are example, the patient who is having difficulty remembering already building systems that can understand and emulate people’s names could have images of significant people and human emotions (Picard 1997). It may not be long before their names stored in a memory-enhancing program. Train- practical caregiving systems are enhanced with this capabil- ing programs tailored to each individual patient could be de- ity. Of course, computers already play a significant role in providing health care for patients. They may be most evi- In early stages of the disease, small digital recorders dent in tracking medical records and the financial aspects of and/or personal digital assists (PDAs) may provide memory health care. They are also routinely used for medical imag- aid to individuals (even those with normal age-related mem- ing, analyzing samples of blood and urine, providing access ory impairment like the authors). One such device, primar- to the medical literature, and a host of other useful functions ily suited for those with frontal lobe brain injuries, is already (van Bemmel and Musen 1997). What we wish to explore marketed for this purpose (Attention Control Systems, Inc.
here, between the routine and the dream, are the imminent 2001; Levinson 1997). At the UMAC, we are exploring the near-term possibilities for computers to provide “hands on,” further use of PDAs. Much of the attention in the area of direct care to patients, assisting in day-to-day activities.
dementia is now on preventing age-related cognitive deteri- We do not believe that computers should fully replace oration by identifying individuals at risk for progressive de- human beings, but rather provide assistance in a variety of mentia like Alzheimer’s disease. The label Mild Cognitive ways. We have long been using computers to provide educa- Impairment (MCI) is applied to individuals who have some tional information, advice and moral support to caregivers.
degree of cognitive difficulty but not enough to be labeled At the University Memory and Aging Center (UMAC), there as demented. Individuals with MCI are at increased risk of has been an online virtual caregiver support group for fif- developing progressive dementia. PDAs have technical ca- teen years. This system was developed by Dr. Kathleen pabilities that may provide useful assistance to people with Smyth, of the UMAC, in concert with the Cleveland chap- MCI. Screen resolutions, static and video image capabili- ter of the Alzheimer’s Association (Smyth and Harris 1993; ties and sound allow the creation of psychologically salient Smyth, Feinstein, and Kacerek 1997). It began as part of sources of information for individuals with MCI. These can the Cleveland Free-Net, one of the world’s first city-wide include not only the standard PDA functions to help track free public access computer networks. As such, it was a contacts and schedule, but also visual reminders, like pho- text-based bulletin board, accessed by caregivers from ter- tographs of relatives and other family related information.
minals in their homes. This in itself was significant, as care- Removable media such as Smart Media Cards or Sony Mem- givers are frequently unable to leave patients unsupervised ory Sticks can store integrated programs of information to to attend meetings and participate in support group activi- help an individual deal with particular circumstances. For ties. Early system users labeled themselves the “computer example, a removable memory device could provide infor- family.” This system has since been updated to run over the mation about specific locations, like a relative’s house, the individual’s own house or any other location frequented by and social interaction for the lonely. Our own preliminary the individual. Location specific information could include, research on a Robotic Caregiver’s Assistant (RCA), at Ohio for example, the names and photographs of particular people University and the UMAC, aims to extend this work by focusing on the special needs of the cognitively impaired.
Devices that have been used by the individual throughout Making sure the patient is safe, not only from falls, but also life and that do not have to be learned anew are likely to be from dangers like kitchen hazards, cigarette fires and wan- most successful. In the years to come, many patients with dering, is a primary concern. This would someday permit dementia will in fact have had PDAs and home computers patients to be supervised by human beings with less inten- with which they are familiar. Computer displays that are sity. The RCA is envisioned as an intelligent monitor, as- personalized and display identically wherever the patient is sisting the human caregiver by identifying events that are would also help. Just as computer interaction has been made unusual and alerting them to the possibility of danger.
easier for children, so too could interface mechanisms be The computer might also assist in making life more pleas- modified to permit easier access for patients with dementia, ant for dementia patients. For patients who wish to stay in even those who are relatively unfamiliar with computers.
touch with current events, the computer could monitor cur- Other enabling technologies for caregivers are computer rent events that fit the interest profile of an individual patient.
vision, speech recognition, and intelligent robotics. A com- Moreover, the computer could present the material in a for- puter could monitor the visual and auditory world in the pa- mat suitable to the patient’s cognitive level and reading abil- tient’s space and learn to adjust to specific patterns. It could ity. As the disease progresses and the patient’s abilities to track and learn the patterns of activity that the patient en- remember recent events diminishes, patients frequently ex- gages in and identify changes in those patterns. For exam- ist in a past world. Many patients in long-term care facilities ple, if the patient fell to the floor unconscious, that event think that their parents are still alive and work through con- could be detected by sensors with software tuned to recog- flict situations in their minds with people who are long dead.
nize falling or the ensuing silence and lack of motion.
The computer, with some forethought on the part of family Ongoing research projects are currently addressing this members, could store static images, voice tracks, videos of possibility. One of these is the Aware Home Research Ini- family events, and favorite songs and movies. Thus, as the tiative (Georgia Institute of Technology Future Computing patient becomes more demented, the computer could choose Environments Group 2001; Kidd et al. 1999). Here, a three- to play selections that reflect the patient’s tastes at that par- story house, or “living laboratory,” has been outfitted with ticular moment. As the patient dwells more and more in the video cameras and a variety of sound, motion and touch sen- past, it is likely that music and movies from years ago would sors. While this project focuses on home computing technol- be found more pleasant than current movies, for example.
ogy for all ages, surveillance and assistance of the elderly is Moreover, the computer could track the patient’s response to the multi-media event, for example, noting vocalizations At the Alzheimer’s Society, we are working on the Safe and whether or not the patient actually watches the screen, at Home project, in cooperation with the Northampton Bor- to determine whether the selection of audio visual material ough Council, the Northamptonshire County Council, and the Northampton Healthcare NHS Trust (Northampton Bor- At the UMAC, we are exploring the use of individual- ough Council 2001). We have outfitted a demonstration ized DVD archives for patients. Increasingly, families are house with a clock/calendar for patient orientation, an alarm maintaining archives of personal history. Frequently, these that sounds if a patient gets out of bed at night, a gas de- are not well organized, but computers can assist in their or- tector and shut off device, and a doorside pressure mat that ganization, storage and display. Previous writing projects, alerts caregivers if a patient leaves the house. In one study, school records, birthday cards, home videos, photographs, eighteen dementia patients and their caregivers were invited and favorite music can all be digitized. Inexpensive writable to visit the house and try out these devices. Fourteen care- DVD technology allows the creation of personal archives for givers agreed to install one or more devices in their own use in situations where patients could obtain information, homes for a trial. Over the next six months, the health care pleasure or both from reviewing significant family events.
costs for patients in this experimental group were compared There are many other ways in which computers could po- to those of patients in a control group without access to the tentially help. For example, one problem dementia patients devices. Costs for the experimental group rose by 4%, com- have is that of dressing appropriately. A computer could pared to 46% for the control group, primarily because more maintain an inventory of the patient’s clothes and identify control group patients were institutionalized during this pe- what the patient is wearing as well as help the patient in se- riod. Two thirds of participating caregivers felt the technol- lecting clothes that match. It could engage the patient in ogy they used helped to delay institutionalization.
speech and other forms of activity therapy to maintain com- The Nursebot project is specifically tailored to the needs munication capabilities as long as possible. It could track of the cognitively intact elderly (Pollack et al.
changes in a patient’s speech as the patient becomes more Nursebot Project: Robotic Assistants for the Elderly 2000).
demented. It could test patients’ cognitive ability to monitor Falling is a hazard for many elderly, and detecting falls is effects of a new drug therapy, thus allowing research to be one part of this effort. An intelligent robotic assistant is planned that could provide surveillance, daily reminders, re- More futuristic challenges also await. Some degree of mote doctor visits, manipulation of objects for the arthritic, cognitive change accompanies aging in all of us. Many actu- ally believe that Alzheimer’s disease is a continuum, rather greater ethical problems. The need for patients to receive than a discrete disease, as the concept of MCI suggests. The help and support has to be balanced with our moral require- question then becomes whether and how computers can care ment to maintain their confidentiality.
for people with milder memory and attention problems. As An issue less frequently discussed is that of the inter- the baby boomer generation achieves elderly status, might play between technology and pharmacology (Whitehouse computers be the first to recognize the subtle shifts in the and Marling 2000). Most efforts to improve the quality of cognitive abilities of their users? Will computers diagnose life of an individual with cognitive impairment focus on im- dementia based on a pattern of inefficiencies in their use? proving the care system and the administration of appropri- More importantly, will computers be able to adapt to their ate medications. A computer could be a link in the complex aging users, presenting information in ways that account for relationship between technology and pharmacology. A com- their changing cognitive abilities? More importantly still, puter could track medication use and assess the patient at can computers take advantage of the positive aspects of ag- relevant intervals to assess the effectiveness of the therapy.
ing, like the potential for wisdom to emerge from a lifetime Moreover, drugs may allow patients to attend more to the of experiences? Can computers help people reflect on their computer. A system that could help a patient dress appropri- lives, make associations, and contribute to the personal, in- ately, for example, would only work if the patient attended to tellectual and even spiritual development of us all as we age? it, but dementia patients may have attention deficits. In chil-dren with Attention Deficit Disorder, stimulant drugs like Ethical Issues in the Use of
methylphenidate have been found to improve attention in Computers for Caregiving
classroom settings. Many clinicians believe that the cur-rent generation of drugs used to treat Alzheimer’s disease, From the above discussion, it seems clear that even with cur- namely cholinesterase inhibitors such as donepezil, galan- rent capabilities and particularly with future ones, comput- tamine and Reminyl, act by enhancing attention.
ers will be able to provide a greater role in caring for pa- drugs used to treat psychosis and behavioral disturbances in tients with dementia. After all, one only needs to recognize dementia patients, such as the neuroleptics, may also affect the ubiquity of televisions in nursing facilities, day care pro- attention in various ways. Thus, for the caregiving situation grams and homes. If the passive television set were replaced of the future, we need to recognize that interactions between by an active computer with peripherals allowing it to intel- computer-enhanced caregiving and the use of medications ligently monitor the environment, it is quite possible that the quality of lives of patients and caregivers would improve New and complex ethical issues could develop around the considerably. But what are the ethical issues that might be relationship between the patient and the computer. The com- puter is, after all, an information processing entity just like The first issue is that of invasion of privacy. A computer human beings and other animals. In biomedical ethics, we with a video camera that is constantly on would represent consider the moral position of animals based on their own some kind of intrusion into the private space of an individ- particular interests and possibilities for suffering. Clearly ual. Would this be “Big Brother” watching the demented animals have moral status in our discussions in a variety patient for purposes of control rather than for purposes of of areas of human life. Could we ever imagine a situa- increasing autonomy and quality of life? Finding an appro- tion in which a computer would achieve moral status? Cer- priate balance would clearly be an issue, but one that could tainly it could be viewed as a possession of an individ- be resolved on an individual basis, because the situation mir- ual human being. Thus a demented patient, who became rors in many ways the issues involved in human supervision very dependent on the computer, and all of the memo- of patients with dementia. Individual contracts with patients ries and images it stored, would suffer greatly if this com- and caregivers could be developed to allow a computer to monitor certain things and not others. For example, many imagine a situation in which the computer itself would be actions occur in the bathroom, and one would have to define viewed as an individual entity with rights? It may seem what level of visual supervision occurred in this space.
far-fetched, but on the other hand, differences in capabil- Maintaining the confidentiality of patient records is an- ities between biological and physical devices for process- other major issue. There is the potential to archive vast ing information may be more challenging to define than amounts of information about individual patient behav- we imagine. Clearly biological creatures, from amoebae iors. Like privacy, confidentiality is not a new ethical issue to human beings, are programmed for growth, develop- for those who employ computer technology in health care ment and death. They have interests and intents. Com- (Goodman 1998). However, it is one that has been especially puter agent researchers draw from this biological basis in significant for us with the CANDID service. Over twelve the popular Belief Desire Intention (BDI) model. Here, soft- hundred people call our telephone helpline each year, and ware and/or robotic agents are given human-like motivations at least a thousand people access our Web site every week.
to independently pursue their own goals (Bratman 1987; People with dementia rarely call, and therefore, we are of- ten giving advice and information to third parties. The staff Could a computer measure the quality of life of its care must be constantly vigilant: discussing patient details with recipient and gain reward from this? Could the reward be a spouse is likely to be acceptable, but discussing a patient distribution of its caring approaches to other information with more distant relatives or a separated spouse presents systems in “support” groups? Could systems of computers compete for resources based on health outcomes of popula- Pollack, M. E.; McCarthy, C. E.; Tsamardinos, I.; Ramakr- tions of patients under their care? When these things become ishnan, S.; Brown, L.; Carrion, S.; Colbry, D.; Orosz, C.; possible, where do we draw the line between excluding an and Peintner, B. 2002. Autominder: A planning, monitor- information processing entity from moral dialogue and al- ing, and reminding assistive agent. In Proceedings of the lowing them in our ethical conversations? Perhaps we will Seventh International Conference on Autonomous Intelli- have to ask the computers of the future that question.
gent Systems. To be published in March.
Rao, A. S., and Georgeff, M. P. 1995. BDI agents: From Acknowledgments
theory to practice. In Proceedings of the First InternationalConference on Multi-Agent Systems, 312–319. Cambridge, This work was funded, in part, by the Takeyama Founda- tion and by the Alzheimer’s Society of the UK. The authorsgratefully acknowledge this support. The authors would also Smyth, K. A., and Harris, P. B. 1993. Using telecomputing like to thank Jared Bendis and David Chelberg for technical to provide information and support to caregivers of persons with dementia. The Gerontologist 33:123–127.
Smyth, K. A.; Feinstein, S. J.; and Kacerek, S.
References
The Alzheimer’s disease support center: Information andsupport for family caregivers through computer-mediated Attention Control Systems, Inc. 2001. The planning and communication. In Brennan, P. F., and Schneider, S. J., execution assistant and training system. World Wide Web eds., Information Networks for Community Health. New Bratman, M. E. 1987. Intentions, Plans, and Practical van Bemmel, J. H., and Musen, M. A., eds.
Reason. Cambridge, MA: Harvard University Press.
Handbook of Medical Informatics. Heidelberg, Germany: Georgia Institute of Technology Future Computing En- Whitehouse, P. J., and Marling, C. R.
enhancement uses of biotechnology, ethics, cognitive en- http://www.cc.gatech.edu/fce/house/house.html.
hancement. In Murray, T. H., and Mehlman, M. J., eds., Encyclopedia of Ethical, Legal & Policy Issues in Biotech- Medicine. Cambridge, UK: Cambridge University Press.
Harvey, R. J.; Roques, P. K.; Fox, N. C.; and Rossor, M. N.
1998. CANDID - Counselling and diagnosis in demen-tia: A national telemedicine service supporting the care ofyounger patients with dementia. International Journal ofGeriatric Psychiatry 3(6):381–388.
Kidd, C. D.; Orr, R. J.; Abowd, G. D.; Atkeson, C. G.;Essa, I. A.; MacIntyre, B.; Mynatt, E.; Starner, T. E.; andNewstetter, W. 1999. The aware home: A living laboratoryfor ubiquitous computing research. In Proceedings of theSecond International Workshop on Cooperative Buildings.
Levinson, R. 1997. The planning and execution assistantand trainer (PEAT). Journal of Head Trauma Rehabilita-tion 12(2):85–91.
National Institute on Aging. 2000. Progress Report onAlzheimer’s Disease, 2000. Bethesda, MD: National Insti-tutes of Health. NIH Publication No. 00-4859, availableonline at http://www.alzheimers.org/pubs/prog00.htm.
World Wide Web site, http://www.northampton.gov.uk/news.archive%5Fnews/pr273.htm.
Nursebot Project: Robotic Assistants for the Elderly. 2000.
World Wide Web site, http://www.cs.cmu.edu/ nursebot.
Ory, M. G.; Hoffman, R. R.; Yee, J. L.; Tennstedt, S.; andSchulz, R. 1999. Prevalence and impact of caregiving: Adetailed comparison between dementia and nondementiacaregivers. The Gerontologist 39(2):177–185.
Picard, R. W. 1997. Affective Computing. Boston, MA:MIT Press.

Source: http://oucsace.cs.ohiou.edu/~marling/caregiver.pdf