Memory Aids as Collaboration Technology
by
Michael Chi Hung Wu
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Computer Science University of Toronto
© Copyright by Michael Chi Hung Wu 2010
Memory Aids as Collaboration Technology
Michael Chi Hung Wu
Doctor of Philosophy
Graduate Department of Computer Science University of Toronto
2010 Abstract
The loss of memory can have a profound and disabling effect on individuals. People who acquire memory impairments are often unable to live independent lives because they cannot remember what they need to do. In many cases, they rely on family members who live with them. When I carried out ethnographic fieldwork to explore this domain, I observed that individuals with amnesia were surrounded by family members who provided extensive memory support (e.g. reminders). I found that such families very worked closely together to accomplish everyday activities, such as coordinating a family outing or planning a doctor’s appointment. However, these activities were often undermined by family members forgetting. This led me to view memory aids as collaboration technology, rather than as tools that only support an individual’s memory. My dissertation explores this idea and how it can lead to more appropriate designs of assistive technology.
To design collaborative assistive technology, I involved persons with amnesia and their family members in a process of participatory design. The design team included six individuals with amnesia, two neuropsychologists, and myself. Five family members were also involved in later stages. This team envisioned the design of a shared calendar application, called Family-Link, that
I implemented for Palm mobile devices.
ii
I evaluated Family-Link by comparing it to the commercially available Palm Calendar in a six- month study with four families. I found that participants had significantly more shared events when using Family-Link than when using Palm Calendar. Qualitative evidence suggests that
Family-Link increased all participants’ awareness of other family members’ schedules, provided caregivers with a greater a sense of security by enabling them to track their family member with amnesia, and reduced the amount of effort that caregivers needed to coordinate. Family-Link also fulfilled the individual needs of persons with amnesia by providing an information storage and retrieval mechanism. However, persons with amnesia and caregivers differed in their opinions about which features were useful. Family-Link can be a particularly important tool for families where members are not co-located throughout the day.
iii
Acknowledgments
The road to my PhD could not have been possible without the help of a great number people. It would be impossible to mention everyone who has supported me along the way, but I will try my best to highlight those who have made noteworthy contributions.
Firstly, I would like to thank my family: mom, dad, and my two sisters, Jenny and Sylvia (as well as my brother-in-law, Darren). I would also like to thank my closest friend, Floria Lee. They have all been very supportive of my academic pursuits and have been there to listen to me whenever my thoughts weighed heavily.
I owe a debt of gratitude to my supervisor, Dr. Ron Baecker, who has been a wonderful mentor throughout my time at the University of Toronto. He has nurtured my academic growth and given me the tools I need to succeed in research. I really appreciate that he is always willing to find extra time in his demanding schedule to meet with me and provide guidance.
My collaborator from Baycrest, Dr. Brian Richards, is such a patient and knowledgeable person and I have learned a substantial amount of psychology while working with him at Memory-Link. I regret not having learned a single cooking technique from this former sous-chef but will ask him about his recipes soon. Thanks also to my other hospital collaborators Dr. Eva Svoboda, Dr. Kelly Murphy, Dr. Guy Proulx, Dr. Larry Leach, and staff member Ruby Nishioka.
I am grateful for the opportunity to collaborate with talented researchers from academia and industry. I would like to thank Dr. Elliot Cole for being my external examiner and my multidisciplinary committee members Dr. Gillian Einstein (Psychology and Public Health Sciences), Dr. Steve Easterbrook (Computer Science), and Dr. Barry Wellman (Sociology) for helping to shape my thesis with their insights and ideas. During my time at the University of Toronto, I have also been able to work with and learn from Dr. Ravin Balakrishnan and Dr. Khai Truong. I am grateful to Dr. Jeremy Birnholtz who played an instrumental role in assisting me with the data analysis of my ethnographic study. During my internships, Dr. Chia Shen (MERL) and Dr. Michael Muller (IBM Research) have been extraordinary mentors to me. Finally, I would not have even considered pursuing my doctoral degree without the support of Dr. Maria Klawe and I thank her for encouraging me at the very beginning.
iv
I am thankful to my colleagues from DGP and TAGlab. In particular, Karyn Moffatt took the time to review my thesis and share her statistics expertise. I am also grateful to Abhishek Ranjan, Mike Massimi, Alyssa Rosenzweig, Nigel Morris, Cosmin Munteanu, Nick Shim, and Miller Peterson. As well, thanks to my colleagues from other universities, particularly Tony Tang, Kim Tee, and Mark Hancock. On the administrative side of things, Kelly Rankin (KMDI), Joan Allen (DCS), and Linda Chow (Graduate office) were particularly helpful whenever I stopped by their office.
Thanks to my Toronto friends who were there to listen to me and share laughs, especially Vince Cheung, Yue Liu, Tim Tran, Mia Clemente-Short, and Yvonne Chang. I would like to thank my CCF Grad Group friends Hannah and Byron Chan, Amelia and Dan Yu, Henry and Diana Sui, Vicky and Steve Chin, Jacqueline Kwong, Ivanna Yau, Richard Lee, and Bernie Ma. I would also like to thank my Vancouver friends: James Dai, Edward Ho, and Benita Fong.
I appreciate all the financial assistance I have received from my funding organizations: the Natural Sciences and Engineering Research Council of Canada, Health Care, Technology, and Place, the Department of Computer Science at the University of Toronto, and the Ontario Government.
As this Acknowledgments section concludes, I want to end off with an important note. My experience has been deeply enriched by clients of Memory-Link and their family members. I especially would like to thank my study participants and design partners for their time and energy. Working with such extraordinary people has been humbling; they have reinforced in me many intangible lessons such as how to persevere and laugh in the face of unexpected and difficult challenges. I will cherish and remember all the time I spent with them.
v
Table of Contents
Acknowledgments ...... iv
Table of Contents ...... vi
List of Tables ...... xiii
List of Figures ...... xvii
1 Introduction ...... 1
1.1 Research Motivation ...... 1
1.2 Problem Statement ...... 3
1.3 Terminology and Style ...... 3
1.4 Research Objectives ...... 3
1.5 Organization ...... 3
2 Related Work ...... 7
2.1 Anterograde Amnesia ...... 7
2.2 Cognitive Rehabilitation ...... 8
2.2.1 Memory-Link ...... 9
2.3 Memory Aids ...... 10
2.3.1 Electronic Calendar Systems ...... 18
2.4 Theoretical Framework: Distributed Cognition ...... 21
2.4.1 History and Key Concepts ...... 22
2.4.2 Cognitive Systems ...... 23
2.4.3 General Properties of Cognitive Systems ...... 23
2.4.4 Propagation of Representational State Across Media ...... 24
2.4.5 Integrated Framework for Research ...... 25
2.4.6 Theory in Practice ...... 25
2.5 Participatory Design with People with Cognitive Impairments ...... 26
vi
2.5.1 History of Participatory Design ...... 26
2.5.2 Motivation ...... 27
2.5.3 Domain of Applicability ...... 28
3 Exploratory Study: Understanding how Families Cope with Amnesia ...... 31
3.1 Distributed Cognition as a Theoretical Framework ...... 31
3.2 Method ...... 33
3.2.1 Participants ...... 33
3.2.2 Data Gathering ...... 34
3.2.3 Analysis ...... 34
3.3 Results ...... 35
3.3.1 Families Functioning as Cognitive Systems ...... 35
3.3.2 Redundancy in Information and Communication ...... 37
3.3.3 Coordination Processes ...... 39
3.4 Impacts of Memory Impairment on the Cognitive System ...... 41
3.4.1 Amount of Effort Required ...... 41
3.4.2 Differences in Information or Task Prioritization ...... 42
3.4.3 Stress, Tension, and Frustration ...... 43
3.5 Implications ...... 45
3.5.1 Implications for Design ...... 45
3.5.2 Implications for Theory ...... 47
3.6 Limitations ...... 48
3.7 Next Stages...... 49
4 Design and Implementation ...... 50
4.1 Participatory Design with Families with a PwA ...... 50
4.1.1 Participants ...... 50
4.1.2 Method ...... 51 vii
4.1.3 Summary of Design Activities ...... 51
4.2 Family-Link...... 55
4.2.1 Definition of Terms ...... 55
4.2.2 Conceptual Design ...... 56
4.2.3 Usage Scenario ...... 57
4.2.4 User Interface ...... 58
4.2.5 Hardware ...... 65
4.2.6 System Architecture ...... 65
5 Evaluation ...... 70
5.1 Study Design ...... 70
5.2 Participants ...... 72
5.2.1 Neuropsychological Profiles of PwAs ...... 74
5.2.2 Family 1 ...... 77
5.2.3 Family 2 ...... 78
5.2.4 Family 3 ...... 79
5.2.5 Family 4 ...... 80
5.2.6 Family 5 ...... 81
5.2.7 Family 6 ...... 82
5.3 Study Procedure ...... 82
5.3.1 First Training Phase (T1) ...... 82
5.3.2 First Baseline Phase (B1) and Second Baseline Phase (B2) ...... 83
5.3.3 Second Training Phase (T2) ...... 84
5.3.4 Customization ...... 84
5.3.5 First Intervention Phase (I1) and Second Intervention Phase (I2) ...... 84
5.4 Data Collection ...... 84
5.4.1 Interviews and Phone Calls ...... 85 viii
5.4.2 Standardized Paper-Based Questionnaires ...... 85
5.4.3 Electronic Questionnaires ...... 85
5.4.4 Electronic Data Logging ...... 86
5.4.5 Observation Sessions ...... 86
5.4.6 Follow-up Interviews ...... 86
5.5 Pilot Study ...... 87
5.5.1 Observations Resulting from the Change in Hardware ...... 89
5.6 Hypotheses ...... 91
6 Analysis and Results ...... 93
6.1 PwAs’ Anxiety ...... 95
6.2 PwAs’ Completeness of Events ...... 98
6.3 PwAs’ Timeliness of Events Completed...... 100
6.4 Primary Caregiver’s Confidence in Events Being Completed by PwAs ...... 102
6.5 Primary Caregiver’s Effort ...... 105
6.6 Primary Caregiver’s Frustration ...... 108
6.7 Caregiver Burden ...... 112
6.8 Sharing ...... 116
6.8.1 Counting Shared Events ...... 117
6.8.2 What Events Were Shared ...... 124
6.8.3 How Events Were Shared ...... 125
6.9 Awareness ...... 128
6.9.1 PwAs’ Awareness of Caregivers’ Schedules ...... 128
6.9.2 Caregivers’ Awareness of PwAs’ Schedules ...... 129
6.9.3 Caregivers’ Awareness of Other Caregivers’ Schedules ...... 136
6.10 Usefulness of Family-Link ...... 138
6.10.1 Comparing PwAs’ and Caregivers’ Perspective ...... 139 ix
6.10.2 Distributed Family Members ...... 142
6.11 Preference and Feature Comparison ...... 144
6.11.1 Overall Preferences ...... 144
6.11.2 Comparison Using Listed Features ...... 145
6.11.3 Comparing PwAs’ and Caregivers’ Perspectives ...... 147
6.11.4 PwAs’ Previous Experiences with Digital Calendars ...... 150
6.11.5 Continuing to Use Family-Link ...... 153
6.12 Limitations ...... 154
7 Summary and Discussion of Results ...... 156
7.1 Summary of Results ...... 156
7.2 Comparison of Participant Families ...... 159
7.2.1 Family 1 ...... 159
7.2.2 Family 2 ...... 160
7.2.3 Family 3 ...... 160
7.2.4 Family 4 ...... 160
7.2.5 Family 5 ...... 161
7.2.6 Family 6 ...... 161
7.3 Reflections on the Method of Evaluation...... 161
7.4 Study Complications and Considerations ...... 162
7.4.1 Procedural Issues ...... 162
7.4.2 Deployment Issues ...... 165
7.4.3 Training Issues ...... 166
7.5 Individual Differences Affecting the Usability of Collaborative Memory Aids ...... 167
8 Conclusions and Future Work ...... 171
8.1 Summary ...... 171
8.2 Contributions ...... 172 x
8.3 Implications for the Design of Collaborative Memory Aids ...... 172
8.3.1 Allow PwAs opportunity to interact with critical information ...... 172
8.3.2 Allow for personalization of shared resources ...... 173
8.3.3 Support resource sharing without limiting PwAs’ independence ...... 173
8.3.4 Avoid deviating significantly from existing user interface designs in each iteration of the design cycle ...... 174
8.3.5 Maintain orientation of users by providing access to history ...... 174
8.3.6 Provide different user interfaces and functionality for different stakeholders in DC systems even when the task is the same ...... 174
8.4 Future Directions ...... 175
8.4.1 Further Use of Distributed Cognition Theory as a Framework ...... 175
8.4.2 Participatory Design Method ...... 176
8.4.3 Collaborative Memory Aids ...... 178
8.5 Conclusion...... 179
9 Bibliography ...... 181
10 Appendices ...... 192
10.1 Consent Form for Exploratory Study ...... 192
10.2 Information Sheet for Exploratory Study ...... 193
10.3 Interview Questions for Exploratory Study ...... 195
10.4 Diary Booklet for Exploratory Study ...... 196
10.5 Codes from Analysis of Exploratory Study ...... 197
10.6 Family-Link Screenshots ...... 198
10.7 Consent Form for Evaluation ...... 212
10.8 Electronic Questionnaire for Evaluation – Screenshots ...... 219
10.9 Electronic Questionnaire for Evaluation – List of Questions ...... 225
10.10 Training Form – Add an Event to Palm Calendar...... 226
10.11 Training Form – Add an Event to Family-Link ...... 227 xi
10.12 Training Form – Mark an Event Completed in Family-Link ...... 228
10.13 Training Form – Add a Note to an Event ...... 229
10.14 Training Form – Add a Chat to an Event ...... 230
10.15 Palm Device Setup for Evaluation ...... 231
10.16 Beck Anxiety Inventory ...... 232
10.17 The Zarit Burden Interview ...... 233
10.18 Interview Questions for Evaluation – Introduction ...... 235
10.19 Interview Questions for Evaluation – End of Phase ...... 236
10.20 Interview Questions for Evaluation – End of Study ...... 238
10.21 Interview Questions for Evaluation – Follow-up for Primary Caregivers ...... 240
10.22 Questionnaire for Evaluation – Comparison of Systems for Family Member ...... 242
10.23 Questionnaire for Evaluation – Comparison of Systems for Participant ...... 243
10.24 Questionnaire for Evaluation - Usefulness of Family-Link ...... 244
xii
List of Tables
Table 2.1: Summary table of systems that support caregivers and cognitively-impaired individuals...... 16
Table 3.1: Overview of study participants. The first person listed in each case has amnesia. The second person is the primary caregiver. Names have been changes to protect anonymity...... 33
Table 4.1: Overview of activities within each participatory design session...... 51
Table 4.2: Rankings of issues from various design partners. The first six columns are rankings by amnesic participants. The last two columns are neuropsychologists who attended the meeting and participated in the rankings. Ranking scores were assigned to each issue. Scores ranged from 1 to 9, where smaller numbers indicate greater importance...... 53
Table 4.3: Program components, the programming language in which it was written, and the approximate number of lines of code (including documentation)...... 69
Table 5.1: The participants of the study. Each code represents a person. The code consists of a letter concatenated with a number. Codes beginning with A represent PwAs. Codes beginning with C represent primary caregivers. Codes beginning with D or E represent secondary caregivers. The number in the code indicates to which family the participant belongs...... 73
Table 5.2: Memory-Link clients participating in the study and background information...... 74
Table 5.3: Memory Link neuropsychological profiles of study participants prior to injury. A1 and A2 are pilot participants, while A3, A4, A5, and A6 participated in the main evaluation. All data are reported as percentile rankings. See Table 5.4 for interpretations of the percentile rankings...... 75
Table 5.4: Interpretation of percentile rankings for Table 5.3...... 75
Table 5.5 : The specifications of the Palm Zire 72, Palm Centro, and Palm Treo 700p...... 90
xiii
Table 6.1 : The number of events for which the electronic questionnaire was answered by primary caregiver, organized by phases...... 94
Table 6.2 : The interpretations for the various BAI score ranges...... 95
Table 6.3 : Descriptive statistics for anxiety score by phase...... 96
Table 6.4 : Responses to the completeness question and the scores that were assigned. Responses of “Don’t know” were excluded from the data...... 98
Table 6.5 : Descriptive statistics for completeness response by phase. These values are used for the statistical test...... 99
Table 6.6 : Responses to the timeliness question and the scores that I assigned. Responses of “Don’t know” were excluded from the data...... 101
Table 6.7 : Descriptive statistics for timeliness response by phase. These values are used for the statistical test...... 102
Table 6.8 : Responses to the confidence question and the scores that I assigned. Responses of “Don’t know” were excluded from the data...... 103
Table 6.9 : Descriptive statistics for confidence response by phase. These values are used for the statistical test...... 104
Table 6.10 : Responses to the effort question and the scores that I assigned. Responses of “Don’t know” were excluded from the data. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data...... 105
Table 6.11 : Descriptive statistics for effort response by phase. These values are used for the statistical test...... 106
Table 6.12 : Responses to the frustration question and the scores that I assigned. Responses of “Don’t know” were excluded from the data...... 108
xiv
Table 6.13 : Descriptive statistics for frustration response by phase. These values are used for the statistical test...... 109
Table 6.14 : The interpretations for the various Zarit Burden score ranges...... 112
Table 6.15 : Descriptive statistics for caregiver burden score by phase...... 113
Table 6.16 : Examples of shared events occurring during the study...... 118
Table 6.17 : Proportion of events shared for every participant in each phase of the study. “N/A” indicates that there was insufficient data...... 121
Table 6.18 : Descriptive statistics for shared event counts by phase...... 122
Table 6.19 : Descriptive statistics of percent awareness of completion. These values are used for the statistical test...... 133
Table 6.20 : Descriptive statistics of percent awareness of timeliness. These values are used for the statistical test...... 135
Table 6.21 : Responses to the usefulness question and the scores that I assigned. Responses of “Uncertain” was excluded from the data...... 138
Table 6.22 : Descriptive statistics for scores for questions in the Usefulness Questionnaire. .... 138
Table 6.23 : Mean usefulness scores for Family-Link features, grouped by subject type...... 140
Table 6.24 : Responses to the comparison questions and the scores that I assigned. I assigned a score of -1 for Palm Calendar, 0 for Roughly Equivalent, 1 for Family-Link...... 145
Table 6.25 : Mean comparison scores of various features. A larger absolute mean value indicates greater strength in how participants felt about the calendar system. Negative mean values are associated with the Palm Calendar while positive values are associated with Family-Link...... 145
Table 6.26 : Mean comparison scores of various features, grouped by subject type. A larger absolute mean value indicates greater strength in how participants felt about the
xv
calendar system. Negative mean values are associated with the Palm Calendar while positive values are associated with Family-Link...... 148
Table 7.1 : Summary of results from the field evaluation...... 157
Table 7.2 : Summary of individual differences affecting the usability of Family-Link in the study participants...... 170
xvi
List of Figures
Figure 2.1: Baycrest planner with digital alarm and daily schedules...... 10
Figure 2.2: The NeuroPage system. Used with permission (Neuropage, 2008)...... 12
Figure 2.3: MAPS software operating on a PC (left) and handheld device (right). Used with permission (Carmien and Fischer, 2004; Carmien et al, 2008)...... 13
Figure 2.4 : MEMOS software operating on a handheld device. Used with permission (Schulze, 2004)...... 14
Figure 2.5: Integrated research activity map (from Hollan, Hutchins, and Kirsh, 2000)...... 25
Figure 3.1: The activities on Case 5’s wall calendar are written by various family members. They are colour-coded: red for general activities and blue for Peter’s activities. Names have been blurred to preserve anonymity...... 36
Figure 4.1: Storyboard for Family-Link...... 54
Figure 4.2: Storyboard for Family-Link...... 54
Figure 4.3: Screen mock-ups of an early version of Family-Link...... 55
Figure 4.4 : The list of Palm applications. The user taps the Family-Link icon to launch the program...... 58
Figure 4.5 : The main screen of Family-Link, with options to view the Family News, Shared Calendars, and Reminders. A button is also included to manually synchronize the data with the server and an exit button allows the user to leave the application...... 59
Figure 4.6 : The user can select whose calendar to view. The Main button returns the user to the main screen while Search My Calendar enables users to find an event based on search terms...... 59
Figure 4.7 : The main calendar view where events are listed for the day. “[Notes]” indicates that there is a note attached to the event. The Todo is a toggle button that shows events
xvii
that are incomplete. The Event Done button allows users to mark events as completed...... 60
Figure 4.8 : Event details of an event. The two icons appear if there is a notes and a chat attached to the event...... 60
Figure 4.9 : This screen allows users to edit the details of an event, such as who is involved. .... 61
Figure 4.10 : A note can be added to an event...... 61
Figure 4.11 : A chat can be added to an event. Different family members can add messages. .... 62
Figure 4.12 : Family-Link provides additional information about selected dates in relation to the current day...... 62
Figure 4.13 : The Digital Keyboard button brings up the digital keyboard for the user. The Back button enables users to review prior screens...... 63
Figure 4.14 : Four alarms can be set in this family of four. Alarms are specified in relation to the start of the event. For example, “1 min” means one minute before the event start. .. 63
Figure 5.1: The various phases of the study and approximate durations...... 71
Figure 5.2: The timeline of Family 1. See Section 5.5 for details about why this schedule is different from Figure 5.1...... 77
Figure 5.3: The timeline of Family 2. See Section 5.5 for details about why this schedule is different from Figure 5.1...... 78
Figure 5.4: The timeline of Family 3...... 79
Figure 5.5: The timeline of Family 4. C4 was unable to participate in the Follow-up interview. 80
Figure 5.6: The timeline of Family 5...... 81
Figure 5.7: The timeline of Family 6...... 82
Figure 5.8 : Shown from left to right: the Palm Zire 72, Palm Centro, and the Palm Treo ...... 90
xviii
Figure 6.1 : Mean anxiety scores for PwAs across conditions...... 95
Figure 6.2 : Completeness of events as perceived by the primary caregivers by phase. A higher score here indicates that a greater proportion of the events were completed. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data...... 99
Figure 6.3 : Mean timeliness of events by calendar condition. A score below 3 indicates that events were completed earlier. A score above three means that events were completed late. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data...... 101
Figure 6.4 : Mean confidence ratings by calendar condition. A higher score here indicates higher confidence of the primary caregiver. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data...... 103
Figure 6.5 : A lower score here indicates that less effort was required by the caregiver...... 105
Figure 6.6 : A lower score here indicates lower frustration. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data...... 109
Figure 6.7 : This bar graph shows caregiver burden scores across calendar condition...... 112
Figure 6.8: Burden scores by calendar condition for primary and secondary caregivers...... 114
Figure 6.9 : Caregiver burden scores for primary and secondary caregivers. Horizontal reference lines show different burden levels...... 115
Figure 6.10 : Bar graph of caregiver burden scores by family...... 116
Figure 6.11 : Mean shared events counts for each phase of the study...... 122
Figure 6.12 : This figure estimates the awareness of primary caregivers about event completeness in the schedules of PwAs. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.133 xix
Figure 6.13 : This figure estimates how aware primary caregivers were about the timeliness of events in the schedules of PwAs. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.135
Figure 6.14 : Response pie chart for the question, “Overall, which software do you prefer?” .. 144
Figure 6.15 : Feature comparison pie chart based on listed features...... 146
Figure 6.16 : Means comparing the calendars for the 11 features, by subject type...... 147
Figure 6.17 : Counts comparing the calendars for a list of 11 features, listed for all participants...... 149
Figure 6.18 : Pie chart of responses when asked whether they would continue using Family-Link if known shortcomings were addressed...... 153
xx 1
1 Introduction 1.1 Research Motivation
Cognitive disabilities can occur in people of all ages, and involve impairments related to memory, attention, language, executive function, and problem solving. Recently, there has been increasing interest among researchers in the field of human-computer interaction in the development of assistive technologies for people with autism (Hayes and Abowd, 2006), aphasia (Boyd-Graber et al, 2006; Moffatt et al, 2004), dementia (Gowans et al, 2004; Mihailidis, Barbenel and Fernie, 2004; Cohene, Baecker, and Marziali, 2005; Baecker et al, 2007), amnesia (Hersh and Treadgold, 1994; Hodges et al, 2006; Svoboda and Richards, 2009) and other cognitive impairments (Carmien, 2004; Morris, 2003). Such research plays the dual role of seeking solutions for those who need them while simultaneously giving researchers the opportunity to study and learn from individuals who have unique cognitive needs (Newell and Gregor, 1997).
Among the various cognitive disabilities that exist, impairment of memory can have a profound and disabling effect on individuals (Wilson, 1999). A large proportion of memory-impaired individuals are unable to live independent lives. As a result, they often rely on family members who live with them. Unfortunately, in addition to social and caregiver burden, this can lead to a
2 significant economic burden on their families. There is an opportunity cost of caregiving to family members who may be otherwise kept from full-time employment. Median incomes for caregiving families are, on average, 28% below that of non-caregiving families (Wang, 2005). Not only does this represent an economic cost to family income, this also represents a significant economic cost to society. In Canada, the unpaid cost of informal caregiving is estimated to be $50.9 billion Canadian dollars each year (Zukewich, 2003).
It is difficult to estimate the incidence of amnesia. However, traumatic head injury (TBI), one of a number of causes of amnesia, is estimated to occur in 100-600 per 100,000 people every year (Park, Bell, and Baker, 2008). Other common causes of amnesia include oxygen deprivation (e.g. following a heart attack), certain forms of encephalitis, brain tumors, chronic alcoholism, Parkinson’s disease, epilepsy, tuberculosis, and multiple sclerosis. Amnesia is also a common symptom in all stages of Alzheimer’s disease, which was estimated to be 480,000 people in Canada in 2008 and increasing at a rate of over 100,000 new cases per year (Rising Tide: The Impact of Dementia on Canadian Society, 2009).
Prior research suggests that in addition to families bearing much of the responsibility for caregiving (Sachs, 1991; Degeneffe, 2001), the adoption, use, and maintenance of assistive technologies become a family responsibility (Dawe, 2006). Indeed over the past number of years I have observed such families working together to accomplish everyday activities such as planning and coordinating doctors' appointments, relatives' visits, and family outings. However, even in cases where rehabilitative technology is used, these activities can be undermined when one of the members has severe memory impairment.
There are a myriad of books and papers (e.g., Parks and Novielli, 2000) educating families on cognitive impairments and providing practical advice on how to cope with caregiver burden, but the literature is largely missing knowledge about how families actually work together to provide care for individuals with cognitive disabilities. Hand in hand with this is a lack of understanding in the literature on how knowledge of familial behaviour can inform the design of assistive technologies and user interface design. This thesis aims to fill these voids.
3
1.2 Problem Statement
To understand the collaborative needs of families with a PwA, and to develop and test supporting technology.
1.3 Terminology and Style For readability and consistency, I use PwA to refer to the person with amnesia throughout this dissertation. For the plural, PwAs is used. For a family that includes a person with amnesia, the phrase “family with a PwA” is used.
I is used wherever the work was done by myself under the supervision of Dr. Ronald Baecker and with the collaboration of Dr. Brian Richards. Throughout this thesis, I adopted a participatory design approach and We is used to refer to my participatory design team and myself.
The term participants is used to refer to PwAs and family members participating in a study.
1.4 Research Objectives
My objectives have been (1) to gain an understanding of how families with a PwA combat the impact of the memory impairment; (2) to design a system together with end users based on this understanding; and (3) to evaluate the new system in real-world settings.
This dissertation describes research that fulfills each of these three objectives. My research design integrates both ethnographic methods and participatory design to increase the likelihood that the resulting design is grounded in the social context and expressed needs of families with a PwA. I conducted an exploratory study and used distributed cognition theory to analyze the results. The results influenced my participatory design team as we conceptualized the design of a collaborative memory aid. This system was prototyped and pilot tested. Based on results from the pilot study, the system was refined and evaluated with more families.
1.5 Organization
The exploratory study (Chapter 3) addressed my first research objective. The goal of the study was to gain an understanding of how families with a PwA worked together. The research questions included:
4
1. Do individuals and their families work together to reduce the impact of memory issues on the family unit?
2. What are my participants’ strategies for overcoming difficulties with memory?
3. How did memory impairment influence the participant families?
This field study included ten PwAs and their families. I employed an ethnographic method that involved nonparticipant observations of PwAs as they performed their daily activities with their families. As well, semi-structured and freeform interviews were used.
In my data analysis, I drew upon Hutchins’ theory of distributed cognition (1995a). One of the key findings from the exploratory study was that the families worked closely together as cognitive systems that must compensate for memory impairment in one of the members. However, despite best efforts the family system was still susceptible to memory lapses in the PwA, which often put the family system in chaos.
The participatory design approach that I took (Chapter 4) addressed my second research objective. In this phase, memory impaired individuals participated as design team members. This involved 18 design meetings with six PwAs, two neuropsychologists, and one computer scientist. I held 2 follow-up design sessions with five PwAs and five family members, one neuropsychologist and one computer scientist. In these follow-up sessions, I proposed the ideas that were developed by the design team and received additional feedback and direction. Indeed, family members and PwA alike expressed their desire for technology to support their collaborations. Together we designed a memory aid called Family-Link aimed at facilitating collaboration in families with a PwA.
The analysis of the exploratory study gave me a better understanding of how families collaborated and the participatory design work grounded my system design. However, an evaluation was needed to assess the design in real-world settings.
My Evaluation (Chapter 5) addressed the third research objective. I implemented the design and created a prototype called Family-Link for deployment in families with a PwA. To evaluate the system, I designed a six-month study that alternated between use of Family-Link (during Intervention phases), and the Palm Calendar (during Baseline phases) with which the majority of
5 participants were familiar. To guide the study, I derived research goals from the results of the exploratory study and participatory design work. The list of hypotheses I wanted to explore include:
1. PwAs will experience less anxiety during Intervention phases as compared to Baseline phases.
2. PwAs will perform events more completely during Intervention phases as compared to Baseline phases.
3. PwAs will perform events more on time during Intervention phases as compared to Baseline phases.
4. Caregivers will have greater confidence in the ability of PwAs to complete events during Intervention phases as compared to Baseline phases.
5. Caregivers will require less effort to coordinate family activities during Intervention phases as compared to Baseline phases.
6. Caregivers will feel less frustration during Intervention phases as compared to Baseline phases.
7. Caregivers will have less caregiver burden during Intervention phases as compared to Baseline phases.
8. All participants will share more of their events during Intervention phases as compared to Baseline phases.
9. All participants will feel greater awareness of other family members’ schedules during Intervention phases as compared to Baseline phases.
10. Participants will find features of Family-Link useful.
11. Participants will prefer Family-Link.
Two families participated in a pilot study and four families participated in the actual study.
6
The Results (Chapter 6) from the Evaluation revealed that for the measures of collaboration Family-Link stood out better than the Palm Calendar. In particular, the results indicated that Family-Link significantly increased sharing of calendar events between family members and increased awareness, particularly for caregivers.
I summarize and discuss the results (Chapter 7) from my evaluation. There are nuances to designing and deploying such technology. For example, memory deficits in PwAs result in challenges and inefficiencies in acquiring the skill set needed to successfully use such technology. I identified factors that may help identify families that can successfully adopt these kinds of technologies.
There are eight chapters in this dissertation. The present chapter provides an overview. Following this is a literature review that appears in Chapter 2. Chapter 3 describes the exploratory study, which has already been published (Wu et al, 2008). Chapter 4 describes the participatory design process and implementation details. Chapter 5 details the protocol for the evaluation and Chapter 6 reports on the analysis and results. Parts of these two chapters have been accepted for publication (Wu et al, in press). Chapter 7 summarizes and discusses the results, details study complications and considerations, as individual differences affecting the usability of collaborative memory aids. Finally, Chapter 8 concludes the dissertation and highlights areas of future work.
7
2 Related Work
This chapter begins with a description of amnesia and how clinicians have tried to assist PwAs through cognitive rehabilitation. Memory aids are often a core part of cognitive rehabilitation, so an enumeration of research and commercial memory aids is presented. This enumeration includes electronic calendars and shared calendars for the home. Finally, the chapter concludes with a description of distributed cognition (DC). I apply DC in the data analysis of the exploratory study described in the next chapter.
2.1 Anterograde Amnesia
Amnesia can affect anyone – male or female, young or old. It results from neuronal injury to specific brain structures responsible for memory processing. Common causes of amnesia include oxygen deprivation (e.g. following a heart attack), strokes, some forms of encephalitis, tumors, chronic alcoholism, or traumatic head injury (Wilson, 2009).
Anterograde amnesia (Curran and Schacter, 2000) refers to difficulty in consciously recalling activities and events that occur following damage to the declarative episodic memory system. The extent and severity of these impairments to conscious recollection differs between individuals, depending on the location and extent of the injury. Typically the knowledge base and skill sets acquired prior to injury are largely preserved. Amnesia is also characterized by
8 preserved intellectual, problem solving and procedural memory abilities. However, anterograde amnesia undermines one’s ability to perform everyday tasks due to the difficulty in remembering the relevant information necessary for task completion.
Procedural memory (Schacter, 1996) refers to the ability to learn new skills and associations based on prior experiences without the conscious recollection of the experiences. Performance improves through the successive activation of the processing networks involved in accomplishing a task. Procedural memory forms the basis of our ability to acquire skills and habits that require repeated practice (e.g., swimming, touch typing). This memory system can form the basis of rehabilitation techniques that focus on learning skills needed for using memory aids such as a daily planner.
While currently there is no restorative intervention capable of repairing underlying neuronal damage, functional recovery (improvement in day-to-day functioning) can be achieved through compensatory strategies that capitalize on preserved cognitive abilities. In PwAs, procedural memory is an important preserved cognitive ability enabling the acquisition of skills for functional recovery.
2.2 Cognitive Rehabilitation
Common compensatory strategies for amnesia are focused on supporting the injured individual’s memory and include: using a memory notebook to schedule daily events, writing on pieces of paper and placing them around the home as reminders of important tasks, and repeating routines and procedures. While most of these are individual strategies for the person with the memory impairment, family members are often heavily involved in training, testing, and refining the techniques. As well, PwAs often ask family members to remind them of important information.
Technology has also been used to combat a range of memory-related conditions. Kapur et al. (2004) reviewed mechanical memory aids (e.g., pillboxes) and computer-based memory aids for patients with non-progressive brain injury and those with mild to moderate memory deficits. LoPresti et al. (2004) surveyed low-tech and computer-based technological interventions for cognitively impaired individuals. While digital voice recorders, mobile phones, and PDAs with patient-friendly software have helped individuals with mild to moderate memory impairments, there is some evidence that suggests that individuals with more severe memory impairments
9 have difficulty benefiting from such electronic aids (Stapleton, Adams, and Atterton, 2007). A number of researchers and clinicians have had success designing and using technology to assist people with more severe memory deficits (Cole and Dehdashti, 1998; Hersh and Treadgold, 1994; Hodges et al., 2006). Electronic memory aids such as personal digital assistants (PDAs) and pagers were used to meet the scheduling needs of PwAs (Hersh and Treadgold, 1994; Kim et al, 2000; Svoboda and Richards, 2009; Wu, Baecker, and Richards, 2005).
These projects focused on tools for individual rehabilitation, but individuals with cognitive impairments do not live in isolation. Their families are heavily involved in their rehabilitation. For example, many of the user studies described in the papers above required the help of family members. These family members integrated the cognitive aids into the lives of those with impairments, trained them to use the devices in real settings, provided support for the devices when they failed, and evaluated their effectiveness during longitudinal studies.
2.2.1 Memory-Link
This thesis research has been preceded by over two decades of work by researchers from the Memory-Link program at Baycrest (Richards, Leach and Proulx, 1990; Richards, 2007), a major research and clinical setting working with the elderly. Memory-Link is an outpatient service that supports adults who have severe memory problems, focusing on developing and training use of compensatory strategies by tapping into preserved memory systems (i.e., procedural memory).
In order to teach PwA how to use memory aids, Memory-Link researchers developed a training technique (Svoboda and Richards, 2009) based on principles of vanishing cues (Glisky, Schacter, and Tulving, 1986) and errorless learning (Wilson et al, 1994; Wilson and Evans, 1996). Using this technique, they taught clients how to use a paper-based memory aid that consisted of a binder that held a paper day planner and an electronic alarm device (see Figure 2.1). At the beginning of each day, clients were to set alarms for each activity they wanted to do in a day. The alarm device allowed for appointments to be set at half-hour intervals. These memory aids successfully allowed clients to set and perform future tasks.
10
Figure 2.1: Baycrest planner with digital alarm and daily schedules.
Memory-Link researchers then began using Palm devices as they had an integrated software calendar and provided alarm capabilities. The PDAs offered additional benefits in terms of storage capacity and user acceptance that paper-based systems lacked (Svoboda and Richards, 2009). The PalmOS was relatively simple compared with other PDA operating systems at the time and so training was simplified. The training techniques were adjusted and applied to clients successfully (Svoboda and Richards, 2009).
2.3 Memory Aids
Wilson et al (1997) describe a case study of a PwA who developed a sophisticated system of external memory aids (paper, alarms, etc.) over ten years to help him compensate for his memory. While the family members of this person did not participate in the daily operation of the system, they helped refine it.
Oddy and Cogan (2004) followed this work and observed in their own case study that the family played an important role in providing “emotional support” and “active and long-term help in devising, tailoring and monitoring compensatory strategies.”
11
Writing Post-it notes, using a diary or calendar, making a list on paper, writing on the back of your hand, and using alarms on watches are all examples of activities that make use of external aids to help one remember. Such everyday external memory aids can alleviate the difficulties of remembering information, especially in the normal population with normal failures of memory. However, effective use of these memory aids place memory demands on the user that are typically difficult to meet for people who have severe memory impairments. For example, an alarm on a watch requires that the user remember the reason why the alarm was set in the first place. Technological tools may be designed to overcome some of these limitations and thus they hold great promise for assisting individuals with memory impairments.
Over the past decade, a number of researchers and clinicians have had success designing technology to assist memory (Cole, 1999; Cole, 2006; Cole, Dehdashti, Petti and Angert, 1994a, 1998; Lamming, 1994; Gowans et al, 2004; Hayes et al, 2005; Hersh and Treadgold, 1994; Inglis et al, 2003; Rhodes, 1997, Tan et al, 2001). For an enumeration of various memory aids, the reader is referred to the following surveys.
Caprani, Greaney, and Porter (2006) reviewed memory aids for ageing populations, such as systems that use personal digital assistants (PDA) or sensory cueing devices. Kapur, Glisky, and Wilson (2004) reviewed external mechanical (e.g. pillboxes) and computer-based memory aids for patients with non-progressive brain injury and those with mild to moderate memory deficits. LoPresti, Mihailidis, and Kirsch (2004) surveyed low-tech and computer-based technological interventions for cognitively-impaired individuals that have been developed over the past 20 years. In each of these surveys, the authors mentioned that the evaluations show that the memory aids can significantly improve a person's performance when compared to internal (i.e., cognitive) strategies, suggesting that technology has great potential to support older adults and those with cognitive impairments.
Yet the majority of systems reported in these surveys are individual tools for a cognitively- impaired person. The role of the family in these systems is often limited to training and support of these aids (see for example, Hodges et al, 2006). However, there are a few computer tools available for supporting both caregivers and people with memory deficits. Four such systems – NeuroPage, MAPS, MEMOS, and Memojob – are described next.
12
Figure 2.2: The NeuroPage system. Used with permission (Neuropage, 2008).
NeuroPage (Hersh and Treadgold, 1994) is a pager system for assisting memory-impaired individuals in remembering appointments and tasks, such as taking medication. A caregiver uses a desktop computer to input prompting times and textual messages, which are stored on a central server. At the scheduled times, the system automatically transmits those messages to the wearer of the pager. The wearer of the pager is notified by an alarm or vibration and can read the message by the press of a button (see Figure 2.2). This system simple and effective, but is limited in functionality. For example, there is no way for a caregiver to know whether or not a reminder was dealt with appropriately.
In 1995, a pilot study of NeuroPage (Wilson, Evans, Emslie, and Malinek, 1997) involved 15 memory-impaired people. The majority of these clients had mild to moderate impairment. An ABA strategy was used (baseline, treatment, post-treatment). One to two weeks were dedicated to identifying target behaviours (e.g., turning off appliances when done using them, attending appointments), followed by a two- to six-week baseline in which those target behaviours were observed and their success measured. Clients were then given the pager device for 12 weeks, during which the same targets were assessed. At the end of this phase, clients returned their pagers and targets were monitored for a 4 weeks. The results of the study showed that all 15 participants of the study achieved a greater number of target behaviours while using the pager than during the baseline phase. Post-treatment performance varied by individual. From 1997 to 1999, Wilson et al (2001) followed up this pilot study with a 16-week randomized control trial involving 143 people. Target behaviours were measured over 2 weeks as a baseline. Clients were then randomly given a pager or put onto a waiting list for 7 weeks. All clients were assessed for the last two weeks of this phase, then at the end of the 7 week phase, clients returned
13 the pager devices if they had one, or received pagers for 7 weeks if they did not have one. Again, clients were reassessed during the last 2 weeks of this phase. Over 80% of the participants were significantly more successful in daily activities such as taking medication and keeping appointments while using the pager compared to baseline. A majority of these clients maintained improvement 7 weeks after returning the device.
More recently, Teasdale et al (2009) investigated whether NeuroPage could decrease carer strain in caregivers of people with acquired brain injury. 99 caregivers completed a modified caregiver strain questionnaire at the outset of the study, after 7 weeks of use, and (for one group) 7 weeks after withdrawal. Significant reductions in strain were found following NeuroPage use and these reductions continued 7 weeks after withdrawal of NeuroPage. Since October of 2000, NeuroPage has been commercially deployed in the United Kingdom.
Figure 2.3: MAPS software operating on a PC (left) and handheld device (right). Used with permission (Carmien and Fischer, 2004; Carmien et al, 2008).
MAPS (Carmien, 2006) is a guided prompting system that supports diminished executive and memory functions by providing verbal and pictorial prompts to a cognitively-impaired user. The system is a combination of a context-aware mobile prompting aid for cognitively-impaired users, and a PC-based script editor interface for caregivers (see Figure 2.3). A caregiver uses a web browser to create various support scripts that are then shown on a client’s handheld device. The
14 system is fully described in a number of publications (Carmien, 2004; Carmien et al., 2004; Carmien et al., 2005; Carmien, 2005; Carmien, 2008) as well as Carmien’s PhD thesis (2006) — the first doctoral thesis of which I am aware that involves a tool designed specifically for use by both cognitively-disabled individuals and their caregivers.
MAPS was evaluated through traditional usability testing and ethnographic participant observation of the system in real use. Four dyads participated in field trials with MAPS (Carmien, 2008). The trials were carried out in semi-controlled settings and each dyad was observed for 15-22 hours. Results showed that caregivers were able to successfully use the script design environment, but faced operating system issues. Persons with cognitive disabilities were able to follow system prompts, although caregiver interventions were occasionally required.
Figure 2.4 : MEMOS software operating on a handheld device. Used with permission (Schulze, 2004).
MEMOS (see Figure 2.4), a Mobile Extensible Memory Aid System (Schulze, 2004; Schulze at al, 2003; Walther, Schulze and Thone-Otto, 2004; Voinikonis, Irmscher, and Schulze, 2005) is designed to assist the prompting and execution of tasks for memory-impaired patients with head injury. It makes use of a web-based server connected to a pocket computer device through a cellular network. A therapist or caregiver uses the central server to supervise actions of the patient when he/she is outside. The pocket computer is both a portable calendar and mobile phone in an integrated device. In order to make a new appointment or postpone an existing one, patients use the device to call the service centre and record a message onto the answering
15 machine. This message is reviewed by a service centre attendant (i.e., therapist or caregiver) who then enters the task into the system. The task is wirelessly added to the patient’s pocket computer. The added benefit of involving a service attendant is that each task can be checked to ensure that it is not done more than once. As well, the attendant can break a task into multiple steps to guide the patient. As the patient performs the task, each step is confirmed by the caregiver. Finally, important tasks can be marked as “critical”, and if a patient does not mark the task as completed, the responsible person is notified immediately.
MEMOS was evaluated at the Day-care Clinic for Cognitive Neurology at Leipzig University with six persons with mild to moderate memory problems (Schulze et al, 2003). MEMOS was compared to a Palm organization and a mobile phone. Patients fulfilled 94% of all six experimental tasks using MEMOS, compared to fulfilling 80% of tasks using a conventional memory aid and 72% of tasks patients without an electronic memory aid (Walther and Schulze, 2004). The MEMOS research grant ran for three years between March 2002 to February 2005.
Memojog (Szymkowiak et al, 2004; Morrison, Szymkowiak and Gregor, 2004; Szymkowiak et al, 2005) is similar to MEMOS, and is designed specifically for memory-impaired older adults to support memory for prospective tasks. Memojog is composed of a PDA and a web-accessible database. Both are connected to a central server. Any of the user, caregiver, or care professional can make changes to the users’ schedule through the PDA or web-based database. The user can also contact an administrator to make any necessary changes to their schedule. Reminders are prompted by the PDA and the user can accept, postpone, or dismiss these prompts. Users were also able to store personal information on the device such as contact information. In the evaluation, the researchers noted that caregivers were pleased that they could input reminders without needing to go through an administrator. More importantly, the caregivers kept the clients motivated throughout the study. Memojog was a research project that ran for three years between October 2000 and October 2003.
Table 2.1 summarizes the systems mentioned here and compares them to the collaborative memory aid that I developed.
16
NeuroPage MAPS MEMOS Memojog Family-Link Key Reference Hersh and Carmien, 2006 Schulze et al, Szymkowiak N/A Treadgold, 2003 et al, 2004 1994 Users Brain injured Brain injured Brain injured Memory- Persons with patients and patients and patients and impaired anterograde caregivers caregivers caregivers elderly and amnesia and caregivers family members Hardware Pager using PDA and PC- Handheld PDA, server, PDA and Platform radio based editor computer and and web- server technology server accessible system Function Alerts wearer Provides Alerts user to Alerts user to Manages of pager to verbal and perform tasks perform tasks, shared perform tasks visual prompts also able to information for user, hold personal (calendars, scripted by information messages, caregiver on PDA planning) through a PC between family members Communication Tasks input by Scripts are User Anyone can Anyone can & Coordination a caregiver entered by a appointments make changes make Issues through a PC, caregiver, but must be to the primary changes to but wearer user cannot approved by user’s add/modify cannot modify create/edit caregiver. schedule. User shared events times or tasks existing Caregiver has control but not scripts must process over private every request, postponing or events check dismissing conflicts, and prompts supervise tasks Current Status Commercially Research Research Research Reported in deployed in ended in ended in ended in this thesis UK since October 2006 February 2005 October 2003 October 2000
Table 2.1: Summary table of systems that support caregivers and cognitively-impaired individuals.
17
All four of these prior systems are collaborative in the sense that they are used both by individuals suffering from memory problems and their caregivers in scheduling tasks. However, while they have been designed to explicitly include family members in memory rehabilitation, each system offers a slightly different model of how collaboration should occur. One might look at these systems from the perspective of who has control of the information .
Both NeuroPage and MAPS gives control of information to the caregivers and leaves execution of the tasks up to the individuals with cognitive impairment. In NeuroPage, caregivers must schedule all the alarms, leaving memory-impaired individuals with no ability to make modifications or postpone appointments in the system. This limits a memory-impaired person’s autonomy and does not allow them to independently reschedule their activities in the face of changing circumstances that may be encountered during the day. Thus, they have little control of their lives and must ask their caregivers to make changes for them. In the MAPS system, caregivers create, edit, and maintain scripts that are then provided to the cognitively-disabled user. The cognitively-disabled user has no way to create new scripts or customize existing ones through the system and thus no control of information relating to their daily tasks.
MEMOS differs from NeuroPage and MAPS in that it enables people with memory-impairments to request the creation and postponing of their own appointments. However, in both cases, the request must be fulfilled by a caregiver who may decide to cancel the request or expand upon it by adding other related appointments. The control of information in this case is passed from the impaired user to the caregiver, who then passes it back to the user in the form of appointments entered in the system. While caregivers felt less worried about their relatives knowing that they were using the memory aid, it could be argued that MEMOS increases caregiver burden by forcing them to attend to all requests. This may not be an issue if the caregiver is a health care professional, but if the caregiver is another family member, this could add stress to an already stressful situation at home.
Memojog enables anyone using the system to make modifications to the memory-impaired user’s schedule, thus giving everyone control of the appointments. However, appointments made by different people do not appear differently in the system. This can be important for someone with memory-impairments as he or she may not remember who created the appointment (whether or not it was their family member who wants them to perform a task or if it was self-initiated). So
18 while the system gives equal control of information to everyone, the user with memory deficits does not have an understanding of who made what changes to his/her schedule (he/she may have simply forgotten), which can negatively impact their perceptions on their independence.
Independence and control of information appear to be coupled. This perceived independence is important as it can enable someone to feel that they are making progress in their rehabilitation. Many of the PwAs with whom I have spoken have remarked that one of their primary goals in life is to regain the lifestyle and independence that they had before they suffered their brain injuries. While such tools as NeuroPage, MAPS or MEMOS can enable someone with cognitive problems to accomplish tasks prospectively, it comes with the cost of becoming more dependent on a caregiver. The goal of my research is more in line with Memojog – to enablee someone with memory problems to draw support from caregivers while retaining their ability to make decisions and change their own information. However, unlike Memojog a goal of my collaborative memory aid is to enable memory-impaired users to have greater control and be aware of changes in their information.
The above systems focus on prompting and execution of prospective tasks. Neuropage, MAPS, and Memojog includes a digital calendar component to this end. The next subsection surveys existing commercial and research digital calendars for the general population.
2.3.1 Electronic Calendar Systems 2.3.1.1 Personal Calendars
There are a variety of commercially available or free online electronic calendars for personal use such as Windows Calendar, Yahoo Calendar, PreMinder, Mozilla Sunbird, Palm Calendar, and iPhone Calendar. Many are designed for desktop computers but some operate on other digital devices such as cell phones, PDAs, and TabletPCs. These systems all offer basic calendar functionality, which includes creating, editing, and viewing appointments and setting reminders for them.
There is a wide range of research on personal electronic calendars that explore how people use them, how they are visualized, and how they can help manage schedules intelligently.
19
Some research has focused on how people use calendars at work. Kincaid and DuPont (1985) examined office workers’ use of paper calendars and commercially available electronic calendars as part of integrated office systems. They developed recommendations for features for electronic calendars and automatic schedulers.
Some research has focused on visualization, such as with DateLens (Bederson et al., 2004) in which a calendar interface is scalable with fish-eye lens. PerspectiveWall (Robertson et al., 1991) is a calendar that enables users to zoom their views.
There are research systems that aim to have some intelligence in scheduling. Visual Scheduler (Beard and Palanlappan, 1990) is a graphical calendar that has a priority-based automated scheduling algorithm. Users found priority-based time-slots and access to scheduling decision reasoning advantageous. SensiCal (Mueller, 2000) is a calendar that extracts relevant information from calendar items and builds up common sense to help identify conflicts (such as being in a different city than the appointment in your calendar). Augur (Tullio et al., 2002) is a calendar that anticipates and compensates for inaccurate calendar entries and peculiar event names.
2.3.1.2 Shared Calendars
A subset of electronic calendars are networked applications that enable other people to view and sometimes edit personal schedules. A small selection of these applications are presented below.
MS Outlook with Exchange Server is a personal information manager that includes a calendar application. The application is for desktop computers and allows users to create and share calendar events that are synchronized through the Microsoft Exchange Server with other users in an organization. This data can also be synchronized to mobile devices that have Windows Mobile installed.
Yahoo Calendar ( http://calendar.yahoo.com ) and Google Calendar (http://www.google.com/calendar ) are both web-based calendars. They offer the user the option of inviting others to share a personal calendar.
20
VueMinder ( http://www.vueminder.com ) is a shared calendar application for desktop computers. It allows one to publish calendars and reminders online or over a local area network. Reminders can be sent as either emails or text messages to yourself or others.
Cozi ( http://www.cozi.com ) is a family-oriented web-based organizer that includes an online calendar, shopping list, and journal as major components. Reminders to family members can be sent by email or text messages. Cozi also allows family members to view and edit the calendar through a web browser on a mobile device.
CalendarWiz (http://www.calendarwiz.com ) is an online calendar publishing system for large organizations. The system enables users to publish calendars for view by others. However, editing control lies with the original user and not the broader group.
DualDate (“Palm™ DualDate™ 1.1 Getting Started Guide”, n.d.) is an application for Palm devices that allows one to share calendars with other DualDate users. Two calendars can be viewed side-by-side in the Palm. However, receivers are unable to edit the shared calendar. While development on the program appears to have halted since 2002, a free version is available for download.
ClearSync (http://www.clearsync.com ) is a web-based shared calendar system that operates on Palm devices and also work or home personal computers. This system allows families to view and edit shared family calendars and contact lists that are synchronized by the system. The family calendars can be viewed overlaid with work calendars to easily see schedule conflicts, yet the work and personal data does not mix. ClearSync is the closest technology to the system that I design later in this thesis and we will return to discussion of it in Section 4.1.3.
2.3.1.3 Shared Calendars in the Home
There are a few notable research calendar prototypes for the home.
Plaisant et al (2006) develop a shared calendar system for multi-generational family members. The system promotes symmetrical sharing of information by providing an interface that is accessible to older users.
21
Neustaedter and Brush (2006) developed LINC, a calendar system operating on TabletPCs that helps families coordinate everyday activities. Neustaedter and Brush found that coordination in the family was not typically done through LINC, but instead, awareness of activities and changes enabled coordination. Neustaedter, Brush, and Greenberg (2007) conducted a month-long field study of this system with four families. It took approximately two weeks for families to get into the habit of using LINC, but the results suggest that LINC supported existing calendaring routines without disrupting existing social practices. As well, primary schedulers felt that ubiquitous access to the calendars increased family involvement in the calendar routine. LINC was also assessed using a set of guidelines that were created based on information gathered from 44 families’ paper calendar routines (Neustaedter, Brush, and Greenberg, 2009). These evaluations indicated that supporting mobility was important both inside and outside of the home. To this end, newer versions of LINC allows the software program to be installed on PCs at work and on mobile devices. However, users are currently only able to view calendar data on mobile devices and are unable to make changes.
Consolvo et al (2004a; 2004b) designed CareNet, an ambient interactive photo frame to help members of the local care network coordinate care for elders. This system includes a calendar for the elder’s appointments transportation needs. Although the caregivers were able to edit this calendar, the elder was only able to decide who should be allowed to see the events or updates.
Many of the lessons learned from the above systems influenced my design work but while these previous efforts were informative, none of them target the specific needs of my population. The calendar prototype that I will present later in this thesis enables equal access to shared resources for every family member and notifies family members about important changes. Schedules are viewed and edited through a mobile device and information is synchronized through a server. While each of these features offered by my prototype is not novel when considered independently, none of the reviewed systems include this particular combination of features to address the needs of families with a PwA.
2.4 Theoretical Framework: Distributed Cognition
Specific examples of systems as presented in the previous section are useful for informing design and situating my prototype in the literature, yet these systems do not provide theoretical insight
22 on collaboration and memory rehabilitation in the family. A theory that had a strong influence on my ethnographic work (the next chapter) is distributed cognition (DC).
DC is a branch of cognitive science that argues that human cognition and knowledge representations are distributed across individuals, internal/external artifacts, and across time, rather than being confined to the boundaries of an individual. DC provides a framework and analytic methodology for examining the complex interactions between people-and-people and people-and-artifacts in their collaborative activities. DC involves explaining how people work together to solve a problem (distributed problem solving), how verbal and non-verbal actions play a role (what is said, implied, and not said), how coordination mechanisms are used (rules, procedures, etc), how communication occurs as the activity progresses, and how people share and access knowledge (Rogers, 2005).
2.4.1 History and Key Concepts
Distributed cognition was a theory developed by Hutchins and his colleagues in the mid- to late- 80’s and was a radical new paradigm for rethinking cognitive phenomena (Hutchins, 1995a). It has its roots in the cognitive sciences, cognitive anthropology, and the social sciences (Rogers, 2005). The traditional theories of cognition viewed mental processes as occurring within an individual’s head. However, distributed cognition blurs that boundary between what occurs inside an individual and what occurs outside, and extends the cognitive system to include interactions between people, technological devices and other artifacts, and internal/external representations (Scaife and Rogers, 2006). Many of these interdependencies are overlooked by the traditional cognitive perspective (Rogers, 2005).
Key concepts of DC include (Decortis, Noirfalise, and Saudelli, n.d.; Hollans, Hutchins, and Kirsh, 2000):
• Cognitive systems
• The general properties of cognitive systems
• Propagation of representational state across media
• Integrated framework for research
23
Hutchins (1995a) suggests that the traditional cognitive science approach had one shortcoming and that was that it only revolved around the individual. He believed that the existing conceptual framework may still be valid, so he suggested that the units of analysis be larger than one person (Decortis, Noirfalise, and Saudelli, n.d.). Hutchins argued that instead of throwing away traditional cognitive science completely, one could continue to apply their concepts to explain systems involving numerous people and artifacts interacting with each other. This extension would enable one to more reliably determine processes and properties of such systems since observations can be directly made – something that is impossible to do when those details reside in a person’s head. There is the possibility that such processes are actually different than those inside an individual (Rogers, 2005). DC thus includes traditional theories of the mind, such as representations and processes and at the same time, it also mixes in social theories to account for socially-distributed cognition (Scaife and Rogers, 2006).
2.4.2 Cognitive Systems
The distinguishing features of DC are:
1. The central unit of analysis is the functional system, that is, the cognitive process that involves people and artifacts and relationships between them (also known as the cognitive system) (Hollan, Hutchins and Kirsh, 2000; Rogers and Ellis, 1994).
2. DC looks for a broader class of cognitive events that go beyond the individual (Hollan, Hutchins and Kirsh, 2000) while traditional theories look at how individual agents manipulate symbols within their heads.
These two features, when applied to observation of human activity suggest that (list paraphrased from Rogers and Ellis, 1994):
• Cognitive processes may be socially distributed (over other people)
• Cognitive processes may be technologically distributed (over people and artifacts)
• Cognitive processes may be temporally distributed (over time such that products of earlier events affect later ones)
2.4.3 General Properties of Cognitive Systems
The general properties of cognitive systems include:
24
1. In cognitive systems involving groups of people, the properties and processes of the cognitive system as a whole differ from those of the individual participating members (Hutchins, 1995a).
2. Knowledge possessed by members is highly variable and redundant. Knowledge that is shared gives participants awareness of each others’ activities, allowing someone to take over a task if needed (Decortis, Noirfalise, and Saudelli, n.d.). Individuals have different knowledge and expertise and will engage in interactions that help them pool their resources when collaborating. As well, individuals may have overlapping knowledge (i.e., shared task knowledge) enabling them to coordinate actions and develop communicative practices (e.g., nodding at someone means “it’s your turn”). (Rogers and Scaife, n.d.; Rogers, 2005)
3. Access to information is distributed in the cognitive system (Rogers, 1997), allowing for shared knowledge to develop and thereafter coordination.
2.4.4 Propagation of Representational State Across Media
The theory examines how such information is transformed during an activity by focusing on how information is represented across media (Hollan, Hutchins, and Kirsh, 1995). Media is defined to include internal representations, such as a person’s own memory, and external representations, such as a computer or paper-based display. For example, the information on the computer display of a kiosk is transformed into a different state when a person touches one of the on- screen buttons. To analyze how representational states change (or are propagated) across media, attention is given to how information is transformed (Rogers, 2005). A more detailed example from Hutchins (1995b) is: Flying a plane to a higher altitude is a coordinated activity that involves an air traffic controller telling a co-pilot through radio, who then alerts the pilot, who performs the action of moving a knob on an instrument panel. In this case, there are three instances of propagations of representational state, two of which involve verbal communication and one that involves a physical manipulation/change in the position of an instrument.
25
2.4.5 Integrated Framework for Research
Distributed Ethnography Cognition
Workplaces
Experiment Work Materials
Figure 2.5: Integrated research activity map (from Hollan, Hutchins, and Kirsh, 2000).
Recently, there have been attempts to build an integrated framework (see Figure 2.5) for the benefit of human-computer interaction researchers (Hollan, Hutchins, and Kirsh, 2000). Hollan, Hutchins, and Kirsh (2000) identified some core principles of DC that widely apply:
• People use different types of structure in their surroundings
• Maintaining coordination requires effort
• People will off-load cognitive effort to their environments
• Social organization can help improve how cognition is load-balanced
These principles identify phenomena that deserve attention and documentation during ethnographic work. Experimentation is important for understanding the impact of various factors because observational methods can be limited due to the richness of real environments (Hollan, Hutchins, and Kirsh, 2000). One form of ethnographic experiment is the introduction of new work materials since they are important aspects of workplaces (Hollan, Hutchins, and Kirsh, 2000.
2.4.6 Theory in Practice
DC itself is not a method (Perry, 2003) and practitioners are not limited in their choice of data collection techniques, though ethnographic methods are more appropriate than others. Extensive field work is necessary for researchers to become familiar with work practice (Rogers, 1997).
26
DC places a heavy emphasis on conducting ethnographic studies and collecting data on people’s work practices, their communications, and their interactions with different media.
A well-cited example of a DC analysis is Hutchins’ (1995a) analysis on the navigation of a ship. In particular, Hutchins examines the steering of a ship into harbour. In his analysis, there was a focus on the communication that occurred, and a detailed coordination of representational states across media for the activity of plotting a fix. Several members of the team under study were involved in taking and plotting bearings of the ship as it approached the harbour. These navigation activities were highly routinized and required the coordination of many people and artifacts.
A number of researchers have applied DC in various domains of work practice. Hutchins and Klausen (1996) studied cognition within airplane cockpits. Ackerman and Halverson (1998) used DC to examine telephone hotline groups. Garbis and Waern (1999) studied emergency and rescue management as well as underground-line control. Rogers and Ellis (1994) studied shared CAD systems. Flor and Hutchins (1992) studied teams of software programmers.
2.5 Participatory Design with People with Cognitive Impairments
2.5.1 History of Participatory Design
Participatory design (PD) is a design approach that directly involves people in the cooperative design of systems for their use. PD is a branch of a more general design methodology called user-centred design (UCD). In UCD, users are often a source of insight and feedback for designers, but only provide external influence and do not participate in making design decisions. Thus, one fundamental distinguishing feature between the two approaches is that practitioners of UCD design systems for users whereas practitioners of PD design systems with users.
PD originated in Scandinavia and England during the workplace democracy movement of the late 1980’s (Muller and Kuhn, 1993). It spread throughout Europe and by the early 1990’s, began to be used in North America. Part of this popularity was due to the growth of strong labour unions that advocated workers. PD was rooted in a concern about the social effects of new technologies in the workplace. Some major contributions to the evolution of PD were made by Bjerknes, Ehn, and Kyng (1987), Nygaard (1990), Bødker (1991), and Greenbaum and Kyng (1992).
27
PD consists of a diverse collection of practices, many of which share common tenets (Greenbaum and Kyng, 1992). Clement (Participatory Design, n.d.) elaborates on this from a workplace perspective, though these ideas can be more generally applied to other domains:
• Respect users regardless of their workplace status, socioeconomic status, or technical ability. Each PD participant is an expert in what they do.
• Workers can be a source of innovation and design ideas can arise from discussion between persons from diverse histories.
• View systems are networks of people, practices and technology embedded in some organizational structure, rather than as only as software.
• Encourage practitioners to spend time with users in their workplaces to understand the organization and relevant work practices.
• Identify and address issues that arise in the workplace as raised by affected parties. PD provides a forum for people to discuss their issues.
• Discover ways to improve working lives of co-participants.
• Be reflective of one’s own role in the PD process.
2.5.2 Motivation
To situate my research, I explore how PD has been used with populations having cognitive impairments. The main tenets of PD are highly relevant in this context.
• PD attempts to create a close fit between individuals and their preferences through engagement and collaborative work. This fit is important and perhaps more appropriate in populations with cognitive impairments than usual formal design methods which often assume regularity in needs and behaviours across different users.
• Persons with cognitive impairments have often been marginalized and thereby disadvantaged, but the principles of PD advocate respect for all collaborators and thus encourages all participants to contribute.
• While it is true that all people have varying needs at different moments, some participants have needs beyond normal expectation that may make the realization of PD goals difficult. For example, communication is a vital cornerstone of PD, but many practices assume that participants can speak the same language - an assumption that may not be true within
28
international teams. Another example could be participants who cannot effectively speak due to aphasia.
• Domain expertise is extremely important when special needs are considered. It can be extremely difficult for designers to imagine what having a cognitive impairment is like because there is a tremendous gulf in expertise between an impaired and non-impaired individual.
• Active participation and experience by an individual having impairments may lead to improvements in their confidence and the way they learn to deal with their deficits.
A population that is brain-damaged can be complicated in their needs and unique in their abilities. In such cases, there is no “standard profile” of a user and understanding the user is by no means trivial. Given the historical roots of PD and its strengths as a method of simultaneous design and inquiry as presented in this subsection, I chose PD as my primary method of design for my research with PwAs.
2.5.3 Domain of Applicability
Though many researchers have developed assistive technologies for people having various cognitive disabilities (LoPresti, Mihailidis, and Kirsch, 2004), most systems have not been shaped using a participatory design approach. There are a few notable exceptions.
Over two decades, Cole et al. (Cole, 2006, Cole, Dehdashti, Petti and Angert, 1994a, 1994b; Cole and Dehdashti, 1998) have explored interface design with traumatic brain injured patients by using a single subject case study approach commonly applied in cognitive rehabilitation (Sohlberg and Mateer, 2001). In their research, patients guided designers in decisions about interface parameters, such as text and instructions. Clinicians were involved in design sessions focused upon correcting interface characteristics. Many of the interface changes and system functionality of the final products were originally suggested by clinicians and patients.
Fischer and Sullivan (2002) reported on a participatory approach to design transportation systems for persons having cognitive disabilities. Their research methodology involved conducting field studies that examine socio-technical solutions in light of real world constraints and cognitive issues. Though their design team was composed of individuals from a very diverse set of stakeholder communities (including assistive care specialists, family support organizations,
29 urban transportation experts, technology designers, and university researchers), the group did not include any persons who had cognitive impairments.
McGrenere and colleagues (McGrenere et al. 2003; Moffatt, McGrenere, Purves, Klawe, 2004; Davies, Marcella, McGrenere, and Purves, 2004; Tee et al, 2005; Allen, McGrenere, and Purves, 2007; Allen, McGrenere, and Purves, 2008) designed assistive technologies while working with persons who have aphasia, a cognitive disorder that impairs speech and language. They made two observations relating aphasics to the design process. First, the fidelity of their prototypes had a very large impact on aphasic individuals. Second, by using non-aphasic participants to help solve general usability problems, the time with aphasics could be spent focusing on language- specific issues. Boyd-Graber et al. (2006) took this a step further by recruiting speech-language pathologists in place of individuals with aphasia for their design process.
More recently, Madsen et al (2009) used participatory user interface design with autistic adolescents.
All the above investigations have concluded that no single perspective or technique can yield a complete solution. These projects have been deemed ‘participatory’ by having the disabled user or related stakeholders play a role in the initial design and on-going redesign of the system. Yet in many cases, design teams involving users with disabilities were kept to single-subject sessions (of researcher and user) because the variability of the disorders were extraordinarily wide- ranging, thereby making collaboration between impaired users extremely difficult to manage and operate.
In my Masters research (Wu, 2004), I involved cognitively-impaired individuals in substantial ways by giving them the ability to make key decisions by consensus throughout the design lifecycle rather than just influencing external designers through suggestions/feedback at various stages of design. This was achieved by creating a design team that included six cognitively- impaired participants who actively engage in collaborative design discussions. Because such severe memory disorders present unique challenges to group work, it was not immediately clear what techniques could be used. I overcame these challenges by using a combination of design techniques, such as incorporating environmental support and using storyboards as external referents, that were carefully adapted to accommodate these special cognitive needs and reduce
30 the need for spontaneous recall (Wu et al, 2004; Wu, Baecker, and Richards, 2005; Wu, Baecker, and Richards, 2007).
The approach in this thesis extends my Masters research to involve caregivers in the design process as well. This exposed some new challenges, such as the sensitivity of participants and the limited time of caregivers.
31
3 Exploratory Study: Understanding how Families Cope with Amnesia
This chapter describes an exploratory study I conducted to learn more about how families provide memory support for a PwA. I used distributed cognition theory (Huchins, 1995a) in the analysis of the study data. This chapter was previously published in similar form (Wu et al, 2008).
3.1 Distributed Cognition as a Theoretical Framework
In his theory of distributed cognition, Hutchins (1995a) uses the examples of navigators on a ship’s bridge and pilots in a cockpit (Hutchins, 1995b) to make the fundamental argument that cognition in these environments does not just occur within the minds of individuals. Rather, it occurs in “cognitive systems” that involve multiple memory storage and processing units that are both humans and artifacts. These are then used by the system to accomplish specific goals, such as flying a plane (Hutchins, 1995b) or steering a ship (Hutchins, 1995a).
Distributed cognition (DC) helps explain how people use artifacts and work together to solve complex problems, paying particular attention to:
32
• How knowledge is accessed and shared (Rogers, 2006). Information in a cognitive system is encoded into representations and stored into artifacts that can facilitate sharing. For example, in navigating a ship Hutchins (1995a) talks about how information pertaining to where the ship is located, relative to known landmarks, is written into a logbook by the bearing taker for sharing with the plotter.
• How communication occurs as the activity progresses. Communication is a key aspect of a DC system. Representations that need to be communicated between members are propagated to other members or artifacts in the system. For example, the logbook described above is read by a plotter, who then carries out another part of the navigation process by calibrating a tool according to the data. (Perry, 2003)
• How distributed units coordinate . Another key aspect of DC is how members of the system are coordinated with one another. Hutchins (1995a) describes how a ship steers into a harbor by the coordinated activities of multiple people and artifacts working together to plot a fix. No one individual can be said to be navigating the ship. Rather, it is a complex coordinated activity that involves team members carrying out simple individual tasks that, when combined, help to locate the ship and where it is headed. (Perry, 2003)
I shall argue and provide evidence that the families with PwAs that I observed are also representative of distributed cognitive systems. In my case, the goal of the system is to enable the PwA to lead a reasonably normal life, by ensuring that they remember appointments, have meaningful personal relationships, take necessary medications, act on personally relevant future intentions, and so forth.
DC presents a model of how one might expect such a family to behave when providing day-to- day support. For example, one would expect task-related information to be stored in various locations, both human and artificial. One would also expect frequent communication and coordination among family members, in which information is periodically shared and updated. As I will illustrate below, I saw substantial evidence of families behaving as distributed cognitive systems. This systemic approach allows me to suggest novel design recommendations that are likely to substantially improve assistive technology design.
Moreover, the families I studied provide a unique environment for studying DC in action. Unlike Hutchins’ navigators and pilots or Halverson’s air traffic controllers (Hutchin, 1995a, 1995b;
33
Halverson, 1995), the cognitive systems I studied all had a human memory component that was known to be volatile and could not be replaced by a more reliable component, as might be the case in a workplace or artificial memory system. Thus, studying this extreme situation (Newell and Gregor, 1997) enabled me to preliminarily explore the potential impact of memory impairment on cognitive systems more broadly.
3.2 Method
3.2.1 Participants
Subjects and families were recruited from Memory-Link and were trained or were currently training to use Palm Zire 72s. The study began with five such families who had been living with amnesia and were involved with Memory-Link for several years. To diversify my sample, I recruited additional families who were relatively new to the program (see Table 3.1). By the time I observed the tenth family, my primary analysis showed that new observations were not adding significantly to the themes, and so I stopped collecting data. The PwAs were all male. I tried to recruit female PwAs but was unable to do so. Pseudonyms are used here to preserve anonymity. Case Participants Observed Year of Years in Injury Memory- Link 1 Mark, Jane (spouse), Bell (daughter) 1987 5 2 Charles, Linda (spouse) 2002 3 3 Keith, Anna (spouse) 2001 3 4 Stuart, Lily (ex-spouse) Heather (health care worker) 2002 4
5 Peter, Sarah (spouse), Emily (daughter), mother 2004 1
6 Alan, Claire (spouse), two daughters 2004 0 7 Jacob, Eva (sister) 1992 0 8 John, Tessa (spouse), health care worker, mother-in-law, 2004 0 nephew 9 Donald, Judy (spouse), son, daughter 2005 0 10 Eric, Nancy (spouse) 2005 1
Table 3.1: Overview of study participants. The first person listed in each case has amnesia. The second person is the primary caregiver. Names have been changes to protect anonymity.
34
3.2.2 Data Gathering
Detailed interviews or questionnaires alone are unsuitable for PwAs because they have difficulty reporting on their experiences. Thus in addition to semi-structured and freeform interviews, I carried out non-participant observations (Emerson, Fretz, and Shaw, 1995). This method was chosen in order to avoid reliance on self-reported data, which could be compromised due to memory difficulties or caregiving stress (Hawkey et al, 2005).
Nonparticipant observation also allowed me to observe spontaneous events as they unfolded and document processes of which PwAs and their families might not be aware. Observations were carried out in an overt manner with full agreement of my participants, who were encouraged to perform their daily activities in a normal fashion. While participants may have been made slightly uncomfortable by the presence of an observer at first, there was no evidence to suggest that they modified their behavior in the long term.
Hour-long interviews were conducted prior to and following each group’s participation in the study in which the PwA and their primary caregiver were interviewed together. Pre-interviews involved questions derived from my research questions. Post-interviews were completely freeform and used to clarify observations. I originally planned a diary study to occur parallel with the ethnographic study, but the diary study was not carried out because of the realization that I would need to train PwAs to use the booklet and keep it with them at all times. Please see Appendices 10.1 to 10.4 for consent forms, interview questions, and the diary booklet.
Families were observed between November 2005 and September 2006. Typically, I spent 2-3 days and 4-7 hours per day observing each family. In total, I conducted roughly 121 hours of observation that involved 31 people interacting in significant ways. This generated 138 pages of field notes and 331 photos of artifacts and environments.
3.2.3 Analysis
DC provides a lens through which one can understand how a group of individuals accomplishes complex tasks. In analyzing this study, I took the family to be the cognitive system and activities of daily living to be the complex task. In some ways, activities of daily living can be viewed to be as complicated as, say, piloting an airplane. Every day, a PwA must deal with errands, chores, tasks, and appointments that must be remembered and tracked. These are often unstructured,
35 spontaneous, and rely on untrained skills. There is emotion involved that can increase levels of stress and anxiety. Activities of daily living must be remembered to ensure they occur on time, and tracked for reflection or sharing. For example, it is important to track when medications were taken for reporting to health authorities. However, human memory impairment of the PwA makes this difficult. The outcomes can be critical. For example, repeatedly forgetting to take medication can lead to hospitalization or be fatal.
As information and memory are distributed over people, artifacts, and time, I focused my analysis on the complex interactions and information exchange between my study participants and external artifacts. I applied an open coding method (see Appendix 10.5 for codes) but paid particular attention to the important elements of distributed cognition outlined in the theoretical framework section: how knowledge is accessed and shared, how communication occurs, and how coordination is achieved. I also looked at evidence of memory impairment impacting the cognitive system. By being immersed in the lives of my participants, I developed domain expertise that helped to guide my analysis. I began with a coarse-grained level of analysis and found that this was sufficient in providing evidence that families with a PwA act as cognitive systems.
In the next section, I present my results, arguing first that the families worked together as cognitive systems, and then exploring the impact of memory impairment on these systems.
3.3 Results
3.3.1 Families Functioning as Cognitive Systems
The first issue I sought to address in analyzing my data was the extent to which the families I observed constitute cognitive systems in the way Hutchins describes in DC. Despite the impaired persons’ use of assistive technologies designed largely for individual usage, I saw substantial evidence in all of the families I studied that they were actually working together quite closely.
3.3.1.1 Information Storage and Access
As DC would predict, information was indeed stored and exchanged between human minds and external artifacts. Some of these artifacts were computer-based and some were not, but all of them were utilized for storing important information that was needed for everyday living. The
36 families appeared to rely on less volatile memory sources, which tended to be artifacts rather than human memories. I briefly discuss two artifacts that the families incorporated into their lives: wall calendars and handheld PDA cameras.
Wall calendars were used by 6 of 10 families to plan and view prospective appointments (cases 1, 2, 3, 5, 7 and 8). The calendars were placed in high-traffic locations in the home, typically the kitchen. The one- to two- month view enabled coordination and awareness of upcoming activities within the family. The effort that families put into collaborative calendaring was surprising because I believed that the calendaring needs of the PwA would be met by their PDAs calendar application, which they had been trained to use on daily basis. What is interesting here is not how or where information is stored in the artifact itself, but that these artifacts were being used by the group to manage information shared among people. In contrast, prior research on family calendars has shown that typical families have one person who assumes the role of primary scheduler (Neustaedter and Brush, 2006).
Figure 3.1: The activities on Case 5’s wall calendar are written by various family members. They are colour-coded: red for general activities and blue for Peter’s activities. Names have been blurred to preserve anonymity.
37
Case 5: As an example of how these artifacts were used, Peter (the PwA), Sarah, and Emily updated their wall calendar by individually adding their events to the schedule (see Figure 3.1). The notes are colour-coded: red for general events (e.g., work, taking the pet to the veterinarian, hockey game), and blue strictly for Peter’s personal appointments. Planning was often done together, which was important because Emily’s work afforded her some flexibility in choosing which days to keep free to help her father.
In 6 of the 10 cases, the PwA used the built-in camera of their Zire 72. These participants took photos of people and objects and showed them to others to supplement their recollections and also to help trigger their memories, as the following pair of examples show.
Case 3: Keith used his PDA to take photos of his new dog and his daughter skating. He learned by himself to move images taken with his digital camera onto his handheld device. Keith shares these photos with his brother and sister-in-law as well as his friends to tell stories.
Case 10: Using his PDA, Eric took a photo of a couch that he saw at the mall that he wanted to show his wife. He also annotated the photo and sketched the price so that he would not forget. Another time, he took a photo of his bathroom sink drain that he wanted to replace. He brought it to the hardware store and presented it to a sales associate who was then able to identify what he needed.
3.3.2 Redundancy in Information and Communication
Many individual coping strategies for amnesia stress the importance of repetition because of intact procedural memory in PwAs. However, rather than individual repetition, the entire family system was involved in repeating various processes. These redundant processes were used to increase the reliability of information exchange in the family as well as increase the availability of information.
A significant amount of the verbal information exchanged between PwAs and their families was redundant. In large part, this was because the PwAs would forget information and ask for it from their caregivers. As a result, there was a constant updating, checking, changing, and negotiating of upcoming appointments throughout the day. In fact, families often quickly switched from casual conversation to sharing redundant information and then back to casual conversation.
38
Redundancy provided additional avenues for the PwAs to acquire information, thus improving chances that critical information will be remembered or retrieved when necessary. For example, I saw once in Case 5 that a simple piece of paper was used in combination with a phone message. Despite the fact that Peter, the PwA, was asleep, Sarah, his wife, wanted to tell him something. She left a paper note on the dining room table, which was their designated message passing location, and also followed up later in the day by calling home from work to leave a message on the answering machine. Thus, there were two ways that Peter could get the message and the redundancy therefore improved the probability that he would get the information.
Similarly, I saw evidence of redundancy being used to improve ready access to information. Important information was often replicated in multiple external aids so that information could be accessed more easily. For example, while the wall calendar kept shared events relevant to the entire family, these events were often replicated in paper-based planners for family members and in the PDAs for PwAs. This allowed them to access the calendar information from outside the home through their external aids. In fact, Claire (case 6) went so far as to keep old calendars of the past two years (i.e., old appointments) and doctor’s cards (i.e., phone numbers) in her purse.
In one of the families (case 1), Jane coordinated her PDA with a paper-based wall calendar at home to ensure that her shared family data is current. She also maintained a copy of her husband’s appointments in her device, which was similar to what the primary caregivers from cases 6, 8 and 9 did. In these cases, caregivers copied the PwAs appointment schedule into a paper-based day planner that they carried around. Such information was said to be important because it provided a level of awareness of where the PwA was at various times of the day.
It is interesting to note that in cases 2 and 3, the primary caregiver owned a PDA but did not use the device, mentioning that it took too much time to enter data.
All families, however, saw the need to synchronize information, and either communicated immediately as new information arose, or through weekly meetings (cases 1 and 2). As well, some caregivers accessed artifacts owned by the PwAs to check for new appointments and information that might have been added during the course of their days. As is evident from these descriptions, redundancy to ensure ready access to information required substantial effort.
39
3.3.3 Coordination Processes
The groups I observed succeeded because they worked very closely together as a unit. For example, members of each group were aware of the daily events of the PwA and would provide reminders to him about those events along the day. They planned their schedules together and also shared information via artifacts such as photos, photocopies, and notes. I discuss two examples: collaborative planning and running errands.
In all families where the primary caregivers lived in the same household as the PwA (all cases but 4 and 7), collaborative planning of appointments was observed. The discussion surrounding these appointments revolved around ideal timeslots and the availability of family members to help with various tasks or transportation. An example of this interaction follows.
Case 1: After Mark checked messages on the answering machine, Jane reminded him to call George in order to reschedule their lunch meeting. Below is the conversation between Mark and Jane with notes in parentheses.
Jane: “You can have lunch Thursday with George.”
Mark: “No I can’t… I have something.” (takes out PDA)
Jane: “Yes you can. You have a meeting but it’ll be done at 12.”
Mark: (after checking PDA) “Ok” (calls his friend, speaks with him, puts phone down and updates PDA)
Mark: (quickly, to Jane) “Wait a minute, wait a minute.”
(Jane starts speaking but is interrupted)
Mark: “Wait a minute.” (Jane patiently waits 5-10 seconds while Mark finishes)
From a distributed cognition standpoint, there are two particularly interesting things happening in this conversation that make it clear that this planning activity involves a system of individuals, and that this occurs regularly. First is the tightly coupled interaction between Mark and Jane that occurs immediately before and after Mark’s interactions with the phone and his handheld device. Second is the quick switching between use of human and artificial memory sources.
40
Many of these collaborative planning activities were structured to provide early awareness to all members of appointments (e.g., meetings with a doctor) and then provide more detailed information when relevant (e.g., scheduling to determine who can provide transportation).
Another example of tightly coupled action between PwAs and their caregivers was errand- running. This occurred in cases 1 and 5 and at a frequency of once every 1-2 days. Typically, this involved the caregiver driving the PwA to various locations so that important tasks could be completed. The caregivers drove in both cases and the conversations in the car often involved discussion of the tasks at hand, as well as updating each other about events in their lives. Here is an edited excerpt from the field notes of this kind of interaction, again from Mark and Jane.
11:40am: Driving in car: Jane updates Mark on things going on in daughter’s life.
12:00pm: At the bank: As Jane parks car, Mark writes in his PDA that Jane is waiting in car outside. Mark goes by himself to do his banking.
12:15pm: At the drug store: Jane tells Mark about radio broadcast and politics. Jane tells Mark she’s going to shop for something. Mark asks what they are buying.
12:25pm: Driving in car: Mark asks, “So now what are I doing?” Jane explains that they’re going to post office to weigh an item.
Two things are particularly interesting in this interaction. First is the level of detail that must be remembered, as when Mark writes in his PDA that Jane is waiting for him outside the bank. Second is the frequency and ease with which they switch between task-based information (e.g., reminders, updates, etc.) and casual conversation, suggesting that the practices have become ingrained in their daily activities.
I have seen substantial evidence suggesting that the families I observed were behaving as cognitive systems. I now turn to the question of how human impairment impacts these systems.
41
3.4 Impacts of Memory Impairment on the Cognitive System
The families I observed experienced stress and tension not ordinarily seen in cognitive systems due to the impairment of memory of the PwA in each case. While the families were generally able to accomplish their goals successfully using the tactics described above, they often had difficulty in doing so, due to three specific factors: 1) additional effort was required, 2) priorities were misaligned and coordination was difficult, and 3) stress was magnified.
3.4.1 Amount of Effort Required
While the work of remembering and reminding was shared by all members of the system, primary caregivers often bore a significant burden. Family members constantly provided reminders of events and people, kept track of the location of important objects, and helped organize the lives of the PwAs. These activities occurred throughout the day and dominated many conversations. The fact that redundant processes were used meant that extra work was required (e.g., repeating things that were said to the PwAs, triple checking that an important note was received).
Also, redundant information was often stored in different artifacts, and these sometimes needed to be synchronized. Synchronization was often a time-consuming process, as is shown in the following example.
Case 8: While John wrote into his journal, Tessa sat on the sofa and spent 20 minutes updating one of the calendars in the home before a phone call interrupted her. Below is the short conversation between John and Tessa during this time with notes in parentheses.
John: “What are you working on baby?”
Tessa: “Calendar. The calendar for the bedroom, then I have to transfer it over.” (After 10 minutes Tessa goes to the fridge with her calendar to see if she missed any appointments on the fridge calendar. She sees a note on the fridge)
Tessa: “Ok, this” (reading note aloud) “'[health care worker] 4pm’, is this Wednesday or the Wednesday past? Did she make another appointment
42
with you? Do you remember? She had a Wednesday appointment and it just passed. I’m wondering if she booked another appointment.”
John: “Oh I think it’s for this coming Wednesday. There was no room on the calendar [for me to write it].”
Tessa: “Oh, ok. Well she marked you for Tuesday already.” (referring to another upcoming appointment)
John: “Write down her phone number on there just in case, that way if things don’t work out you can call her.”
In this example, updating the calendar required some work in checking multiple memories (John’s memory, fridge calendar, fridge notes). Furthermore, the information on the note was unclear and John did not seem sure himself, leading to additional work that Tessa would need to do to confirm the appointment.
Caregivers of all families reported that they were overwhelmed with the amount of information that they needed to manage, such as taking care of their own schedules, the schedules of the PwA, medication dosages and reminders, or medical information that the health professionals needed to be aware of. The primary caregiver needed to give up their full-time employment in several cases (1, 2, 6 and 8) to free up their time so that they could perform these activities.
3.4.2 Differences in Information or Task Prioritization
While it is common for individuals in groups to have different opinions and priorities, I found this to be particularly problematic in some of the families I studied. In particular, problems stemmed from differences in what was perceived to be important information that should be recorded or remembered, or in who should do the recording or remembering.
Case 2: Linda felt that Charles recorded useless information in his Palm device.
“(He writes) lots of useless things in his Palm like how many times the dog pooped during walks.” (Linda)
43
Charles kept these notes mostly for himself as a running diary. Linda often suggested that he look back at his notes to remember what he has done. However, he did not do this on a regular basis, perhaps because his notes were fairly detailed. Linda does not read Charles’ notes, but said that she would like to hear more about how his day went. He does not share as much as he did before his aneurysm due to his memory lapses. As Charles becomes more independent Linda has fewer cues and knows less about how he is doing. This is important information that she would like to know.
Such situations often led to information being left unrecorded, which could have had significant consequences for the family. The differences also created conflict between members of the group when deciding how to handle the situations.
Case 10: Nancy wanted Eric to report details about his seizures to his doctor but did not trust Eric to accurately do so because he had forgotten important details in the past. Eric did not record information when he was not feeling well and Nancy was not always around to record the details herself, so important information was missed and not relayed to the doctor.
3.4.3 Stress, Tension, and Frustration
The amount of effort required and the misalignment of priorities appeared to evoke high levels of stress in families. In fact, caregivers reported that managing information such as appointments, medications, and reminders on daily basis was a very stressful part of their lives.
All primary caregivers reported that dealing with memory issues was extremely taxing. They devoted a great deal of time and energy to providing reminders and cues when necessary. The amount of work added stress to the caregiver, which may have negatively impacted a couple of the families in the long-term. Two of the ten spouses in my study divorced their husbands with amnesia several years post-injury (one divorce happened many years before my study and the other happened after). I was sensitive to the personal nature of these decisions and did not ask for specific reasons, but it is certainly possible that it was caused by stresses arising from the spouses being cast into a caregiving role unexpectedly. This situation often necessitates that spouses give up their employment and devote much of their energies to taking care of someone full-time (see (Kreutzer, 2007) for a study on divorce rates after brain injuries).
44
Sometimes stress led to tense events during the day when priorities were misaligned. For example, stress can be rooted in spontaneous planning by the PwA. The PwA has difficulty remembering their family members’ appointments and may decide to perform tasks that require their help (e.g., for transportation). When these plans conflict with the family members’ schedules, family members will rearrange their day to accommodate, but this can be a frustrating and laborious job. Stress was evident in other situations as well.
Case 1: After a meeting between Mark, Bell, and a lawyer, Mark left the office before Bell. However, after saying goodbye to the lawyer, Bell did not see Mark outside the office and began to worry. She searched the hallways to no avail. Finally, she found Mark in an adjacent office retrieving a book. Bell was visibly upset and scolded Mark,
“You have to tell me where you go! ...I’ll put you on a leash!” (Bell)
He looked surprised as he himself had not yet realized that he had wandered off. Mark replied in a serious tone,
“I forgot you were with me.” (Mark)
This episode showed that a great deal of tension could have been avoided by a small information update, and that caregivers need to be constantly aware of the activities of the PwA and his whereabouts when outside the home.
Case 9: Rising stress led to a very emotional experience between Donald and his wife and children. Donald recounts that his son lectured him about his memory and mentions that ever since that incident, any mention of his health really stirs up things in his family. Judy explained that it has been hard on everyone and provides a different interpretation,
“[He] thinks [my son] is telling him off… he’s not. [Our son] is trying to explain things. They all looked up to father for help. Now, it’s almost like reversed.” (Judy)
This frustrating experience may be more related to Donald’s lack of awareness of his deficit. Donald feels that there is nothing wrong with him and that his memory has always been bad, so
45 he cannot tell the difference between pre- and post-injury. However, based on his experiences with his family he does see that something is not quite right and comments,
“Sometimes I wonder if I wasn’t happier if I was a lot sicker. It’d be a lot easier to rationalize.” (Donald)
3.5 Implications
Despite these difficulties imposed by amnesia, the family caregiving systems I observed were still able to accomplish their goals by using the tactics described above. In this section I describe design and theoretical implications of my DC approach to the problem of cognitive rehabilitation.
3.5.1 Implications for Design
I saw extensive evidence of families working very hard together to adapt technologies that were not designed with impaired cognitive systems in mind. The findings suggest several strategies for improving the design of assistive technologies, and possibly for improving information sharing in families more broadly.
3.5.1.1 Make Reliable Storage Easy and Available
I observed many cases where information needed to be recorded quickly and reliably for easy access later by multiple individuals. In some cases, as when Mark had to write down where Jane was waiting for him, a PDA or camera was the best way to address this. In other cases, however, as when Eric failed to record information about his seizures because it was too difficult, it was not. Given the frequency with which I saw my participants access this information and its importance in their lives, recording and access must be easy and instantaneous.
When designing for PwAs, memory aids should facilitate the storage of details at the time of the event or immediately afterwards, since recalling details after a delayed interval can be problematic for PwAs. One way to ease the storage of information at the time of the event is to support the capture of rich media types like photographs or video. There are numerous projects demonstrating the benefits of easy capture such as SenseCam (Hodges, 2006), a retrospective memory aid that captures a digital record of the wearer’s day that includes images and sensor
46 logs. The PDA cameras I described accomplish some of this, but sensors and video could augment this substantially.
3.5.1.2 Automate Redundancy, Synchronization, and Tight Coupling
While many existing assistive technologies appear to focus on rehabilitating an individual’s memory or restoring their independence, my observations suggest that technologies designed to help families remember together may be more useful and effective. Design for multiple users in multiple locations raises a number of challenges.
It is important to make it easy to check or browse information from multiple places, such as multiple wall calendars or personal PDAs. Designs should allow users to make frequent updates to stored information, but these updates should involve some sort of “checks and balances” system to prevent significant confusion on the part of PwAs. Designs could, for example, support automated cross-checking of information so that they automatically synchronize information from various artificial sources to maintain accuracy. In automating support for these processes, designers must be careful, however, not to eliminate the important coordination processes that redundancy facilitates – only the extra effort. Automation would save human effort.
In addition, tight coupling of action such as the scheduling of conversation I observed between Jane and Mark could be accomplished even when they are not physically together. This could be supported by, for example, allowing one family member to “suggest” activities that could be approved by another member.
3.5.1.3 Increase Awareness of Information Access and Updates
Many families wanted to know when and how information is processed by others in the system (e.g., whether they received the information, whether they took the correct actions with regards to the information). Designs should therefore incorporate awareness of who makes “read” and “write” actions on information stored in the system and when they do.
For families with a PwA, this might be accomplished by having technology track how information flows from artifact to person and vice versa for various activities that need to be done by the family system. The information could then be presented to family members on a handheld device or PC visualized in some way, perhaps using something similar to the “edit
47 wear and read wear” visualization technique (Hill, 1992) that displays history of author and reader interactions with a document in an abstracted graphical form. This would provide awareness of access and update activities to members of the family.
3.5.2 Implications for Theory
While I used DC primarily as a lens with which to view the collaborative activities of families with PwAs, I also had the opportunity to study a unique type of cognitive system. This perspective gives me some insight that allows me to derive some preliminary implications for DC, though more research is clearly needed to fully explore these ideas.
DC theorists have not specifically examined the existence of cognitive impairment within a DC team. They have not explicitly conceptualized impairment of cognitive processes that are distributed over various people and artifacts. What I observed in my ethnographic study is that the cognitive impairment in one person impacts more than themselves and this has implications for the entire DC system, including external artifacts and other people. In a sense, when one component of the DC unit is impaired, it is like the entire DC unit is impaired. How well DC theory can accommodate impairment of cognitive processes is a crucial test of the theory itself and an interesting area of future work.
Prior DC research has explored memory failures in pilots and air traffic controllers (Halverson, 1994; Halverson, 1995; Hutchin, 1995b), but this research poses the issue as a system reliability problem and addresses it by using redundancy to help reduce the number of errors. There is an implicit assumption in DC that no one component of the cognitive unit is more likely to fail than any another. In the case that the component is a memory source, DC does not explicitly account for the degree of impairment in the memory source, only that it exists. My study illustrates a very profound impairment beyond those mentioned in previous research and this is the key issue in my domain that separates family systems coping with amnesia from other systems. If one can identify unreliable sources in a given system and account for them theoretically, the reliability of the entire system can be improved. While my results showed that the families I observed were robust cognitive systems that coped with the impairment of memory through a variety of strategies, I also saw that this was difficult and stressful for the families, and despite a great deal of effort in attempting to prevent memory failures from occurring, memory failures occasionally occurred and threw the entire system into chaos.
48
The findings suggest that the reliability of human components in a cognitive system can impact stress and affect relationships between all members of that system. While this may not be surprising, one must consider the importance of personal relationships in effective communication, coordination and information sharing (Kraut et al, 1999), which are the very foundations of DC. This is particularly true when it is not possible to remove or replace the component known to be volatile, as is the case with the families I studied. This suggests that the role of relationships in DC (e.g., how the cognitive system is affected by stress, power, feelings of resentment) deserves more careful study. Past research in work environments also advocates incorporating social, cultural and organizational factors into the cognitive framework (Halverson, 1995). However, work domains involve formal relationships between co-workers where role expectations and norms may be clearer than informal relationships in the home where roles are often poorly defined. The focus on the role of relationships is perhaps more important in home settings because individuals may be particularly sensitive where disabilities (or the perception there of) are concerned.
My insights have arisen from studying scenarios with known impairment, which simplified what I was observing. However, all memory is imperfect and there are various degrees of imperfection. Therefore, my observations may be pertinent to many other settings beyond amnesia. For example, this may impact prior research on how persons with Alzheimer's Disease manage their informational needs with family caregivers (Hawkey et al, 2005).
3.6 Limitations
All of my participants with memory impairments were male due to difficulties in recruitment. It would be valuable to compare my results with families having female PwAs.
The results were primarily interpreted by myself and are thus subjective. Yet the primary clinician for the PwAs independently verified that the facts reported were consistent with what he observed himself from his interactions and years of history with his clients. I also recruited a volunteer to observe four support group meetings at Baycrest and thereafter assist in my analysis of field data. She confirmed that my interpretations were consistent with her casual observations of PwAs. This member checking helped to improve reliability of the study results (Easterbrook, Singer, Storey, and Damian, 2008). As well, I approached this ethnography with openness about my expectations and I included documentation of observations that were unusual to me. As a
49 computer science researcher, I do constantly look for technological opportunity. As an attendee of over a dozen support group meetings at Memory-Link, I do have a sense of the concerns and issues that PwAs report. Reflecting on my observations in light of my bias allowed me to have a more balanced perspective.
There may have been a Hawthorne effect (Adair, 1984) and my study participants could have changed their behaviour due to my presence during observations. On the other hand, I strove to maintain non-participant status. The majority of interactions did not actively involve me and the dynamics of the interaction did not appear to be significantly altered. The interview data tended to confirm this.
3.7 Next Stages
The results of this exploratory study informed the work of the next few chapters. Participating family members worked very closely together, but this required extra effort to coordinate, led to differences in prioritization, and increased stress. My observations suggested that PwAs and their families could use support in scheduling their calendars. It appeared that supporting the collaborative efforts of families with a PwA could be very beneficial to the distributed cognitive system. With the realization that technology might be able to provide such support, I set out to design a collaborative memory aid.
50
4 Design and Implementation
The exploratory study suggested that collaborative memory aids may be a useful tool and I set out to design and implement one such aid. In this chapter, I introduce Family-Link, a system designed for families with a PwA.
4.1 Participatory Design with Families with a PwA
To design a collaborative memory aid to support individuals who have amnesia and their memory support networks, I enlisted the help of PwAs and their families.
4.1.1 Participants
My design team was made up of a multidisciplinary group -- including six PwAs, two neuropsychologists, one graphic design student, and myself. The PwAs were selected from the Memory-Link program based on their availability. I was also able to involve five family members in two design sessions.
Involving memory-impaired individuals and their family members throughout the design lifecycle can increase the likelihood that the resulting system will address their expressed needs. By including them in the design team, they are able to make key decisions by consensus rather than merely influencing research at various stages through more traditional user-centred
51 approaches. Also, each participant in the design team can contribute their unique perspectives and skills as all members benefit from mutual learning and respect (Wu, 2004).
4.1.2 Method
Our design team met at Baycrest for approximately an hour and a half every Monday at 11:00am. The design work began in October 2006. The weekly meetings ended on April 2007, but two additional sessions were organized for November 2007 and July 2008. These last two sessions were organized for evening timeslots.
Baycrest was chosen as the location for the design sessions because each PwA was familiar with the location from past experiences with Memory-Link. This removed the need for new route planning and reduced the risk of PwAs getting lost while on their way to or from Baycrest.
We completed a total of 20 design sessions. These sessions covered concept design, requirements analysis, high-level and low-level design, and low-fidelity prototyping. Feedback at each stage was incorporated into the system design.
4.1.3 Summary of Design Activities
Table 4.1 summarizes the meeting activities for the 20 participatory design sessions.
Session Activities 1 Discussion of the results from the ethnographic study 2-4 Viewing demonstrations of existing research and commercial systems 5-6 Brainstorming sessions 7 Ranking issues 8-9 Conceptual design, illustration of ideas, scenarios 10 Reviewing design of the ClearSync system for purposes of comparison 11-12 Brainstorming sessions, explored idea of intelligent planner 13-14 Isolate features to include, refine requirements 15-18 Storyboards, interaction and interface design 19 Low-fidelity prototypes, involvement of some family members 20 Medium-fidelity prototypes, involvement of some family members
Table 4.1: Overview of activities within each participatory design session.
52
In the first design meeting, I presented the results of my ethnographic study findings (see Chapter 3). All PwAs agreed that the findings and conclusions were consistent with their own experiences (four of the six members had participated in the ethnographic study). Over the next three meetings, I presented commercial and research technologies designed for families. I demonstrated LINC (Neustaedter and Brush, 2006), Digital Family Portrait (Mynatt et al, 2001), and SenseCam (Hodges et al, 2006). This provided my design partners with a sense of what collaborative technology could do.
The design members decided that the above mentioned systems did not address their need for family coordination, and were not customized for their use. Two brainstorming sessions followed and in the seventh design meeting, I enumerated all the issues that we had discussed thus far. There were 9 issues in total. I then had design participants individually rank these issues by writing each issue on a card and having participants arrange the cards in order of importance, where 1 was “most important” to 9 being “least important” (see Table 4.2).
There was consistency in the rankings between PwAs, particularly for the top four issues. We noted that these four issues could potentially impact other family members. In the eighth meeting, the group decided to conceptualize a system that addressed the top four issues as they were related to family coordination. In Meeting 9, we discussed a set of pre-made scenarios, which led us to decide to build a shared calendar system. We discussed the requirements of such a system. I learned about a commercial shared calendar, ClearSync, and presented the system to the group in the next meeting. While it provided some sharing capability, the group decided that it was not designed for PwAs and would not be flexible for their use.
53
PwA1 PwA2 PwA3 PwA4 PwA5 PwA6 N1 N2 scoreAverage of PwAs scoreAverage of Neuropsychologists scoreAverage of all design partners
Knowing what is scheduled 1 1 1 1 2 1 1 1 1.1 1.0 1.1 in the near future (tomorrow, next week, etc)
Knowing the schedules of 2 3 2 2 3 3 2 3 2.5 2.5 2.5 family members
Relaying important messages 4 4 5 4 1 2 3 2 3.3 2.5 3.1 to family, friends, doctors, etc
Updating family when 3 2 3 5 4 4 6 4 3.5 5.0 3.9 appointments change
Telling stories of important 6 5 7 9 7 5 7 6 6.5 6.5 6.5 events with photos, etc
Keeping a record of the most 8 7 6 6 5 9 4 7 6.8 5.5 6.5 important events of the past year
Keeping audio recordings of 5 8 4 8 6 8 8 8 6.5 8.0 6.9 shared information
Knowing an appointment 7 9 9 3 9 6 9 5 7.1 7.0 7.1 type from its “ring tone”
Keeping track of emotions 9 6 8 7 8 7 5 9 9.8 7.0 7.4 associated with events
Table 4.2: Rankings of issues from various design partners. The first six columns are rankings by amnesic participants. The last two columns are neuropsychologists who attended the meeting and participated in the rankings. Ranking scores were assigned to each issue. Scores ranged from 1 to 9, where smaller numbers indicate greater importance.
I thus conducted two more brainstorming sessions to discuss how our calendar system would be different from ClearSync. In these meetings, PwAs suggested adding advanced features for the calendar system, including adding some artificial intelligence to the device to optimize activity
54 planning. Ultimately, while these advanced features would be interesting additions to our conceived system, I decided to focus our design team on the sharing aspects of such a system first, and left the intelligent planning to future work. Our team discussed which features were most important for a shared calendar and documented system requirements. The graphic designer on our team helped to develop storyboards (see Figure 4.1 and Figure 4.2), interaction diagrams, and user interface sketches.
Figure 4.1: Storyboard for Family-Link.
Figure 4.2: Storyboard for Family-Link.
We discussed and refined these over three meetings. I developed prototypes for the last two design meetings, which were held at later dates when I could gather family members. A low- fidelity prototype (i.e. digitally drawn user interface diagrams) was presented to our design team as well as some family members (see Figure 4.3). I incorporated feedback from the group into the designs and developed a medium-fidelity prototype (i.e., the shared calendar screens that users could navigate on Treo devices but without network connectivity and data storage). This was demonstrated to the group and feedback was again incorporated into the design.
55
Today’s date is See personal or visible, along shared calendar with next 2-3 jobs See notifications and/or requests
Jobs for Tues, Oct 23, 2007 Check bills See activities What’s Calen Activities and jobs New? (5) dar and Jobs
Indicates number of new items
4
Figure 4.3: Screen mock-ups of an early version of Family-Link.
4.2 Family-Link
We designed a system called Family-Link to support families in achieving their everyday activities by helping them to communicate and coordinate their efforts .
4.2.1 Definition of Terms
Events , alarms , notes , chats , and notifications are objects within Family-Link.
An event is a calendar entry that has a textual title, date, start time and end time. In some cases, events can have no start or end times, denoted by NoTime . Each event can have:
• n alarms associated with it, where n is the number of people in the family.
• A note optionally attached to it.
• A chat optionally attached to it.
An alarm represents an audio alert and optional vibration to get the user’s attention. Alarms are specified in relation to the start time of an event (e.g., 3 minutes before the event, 1 hour before the event).
56
A note is textual information that can be attached to an event.
A chat consists of one or more textual messages that are arranged in chronological order. The time and date are also included for each message. Since each chat is mapped directly to an event, each chat represents a set of messages relating to the event.
A notification is a message to the user that something has changed in Family-Link (e.g. change in title of an event, change in start time of an event, change in content of a note). It is usually accompanied by an audio alarm.
4.2.2 Conceptual Design
Family-Link allows families to:
• Work together to organize and keep track of ongoing events and manage important information necessary for successful completion of them.
• Access the system from many places, including the home, hospital, and on public/private transportation. This will allow users to be more easily reached for consultation, discussion, and decision-making.
• Be notified and presented with information about what other family members have recently done. For example, a caregiver can be informed that their relative with amnesia has relayed important medical information to their doctor.
The system comprises of a number of handheld devices, each one operated by a different family member. These devices wirelessly connect to a server PC that stores all the data. Events created on one device are automatically synchronized on devices operated by other family members.
The system provides a personal calendar that can be edited by the user. Users can create, modify, and delete events on this calendar. The system treats these events as things that can be completed and enables users to mark them in order to keep track of their tasks. The system also provides access to the personal calendars of other family members. Users of Family-Link will be able to set alarms for their own events or the events of others. The system allows users to write chat messages about upcoming events. The discussion happens asynchronously. The system will also enable users are add notes to an event in case they wish to specify additional information for
57 an event. Finally, the system will provide notifications about changes in the system initiated by other family members.
Family-Link is similar to many other calendar systems, despite its minor unique features (the ability to set alarms for other family members, and chat messages being grouped together by the event to which they are attached). What makes Family-Link unique is that its combination of four key features is not offered together in another system.
1. Family-Link enables PwA to create and edit shared resources independent of caregivers or an administrator.
2. Family-Link synchronizes shared information.
3. Family-Link provides notifications of changes.
4. Family-Link allows users to indicate that an event has been completed.
These features when considered together are uncommon in assistive technologies that support prospective tasks.
4.2.3 Usage Scenario
The following scenario describes how our computer system might be used. The three actors in the scenario are John (person with amnesia), his wife Sarah, and their new neighbour Victoria. While John is at the volunteer centre, Sarah is baking at a Victoria’s house and Victoria spontaneously invites both Sarah and John to dinner next Tuesday. Sarah knows that she needs to plan around his various hospital appointments that are inflexible. She grabs her handheld device with Family-Link on it and uses the tool to check John’s personal schedule. Sarah picks a free evening, and then creates the event and indicates that John will participate in the social event. She sets a reminder alarm for John to ring half an hour before the event so that he will be ready when the time comes to leave the house. She also leaves a note in the event to remind John who Victoria is because he has a hard time remembering new people and Victoria just moved into the house across the street a couple weeks before.
A short time passes. John is on the bus going to a volunteer centre. There is quite a bit of ambient noise, but his device alerts him at this moment. The device sounds an alarm and also vibrates in
58 an attempt to get his attention. He takes the device out of his shirt pocket to see why it rang. He sees that Sarah created a dinner at their new neighbour’s house. John creates a note in Family- Link for Sarah to buy some wine to bring to the dinner since he himself is occupied by a doctor’s appointment earlier that day. Regardless of whether he noticed the notification from Sarah while John was on the bus, the appointment would have shown up if John viewed his calendar at a later time.
Sarah’s device now beeps and she checks it. She notices that John created a new note attached to the event and so she views it. She quickly sets a reminder alarm for herself. Sarah tells Victoria that they are both available for dinner on Tuesday.
4.2.4 User Interface
4.2.4.1 Screenshots
Several screenshots of Family-Link are included here as Figures 4.4 to 4.14 to show the program in action. See Appendix 10.6 for additional screenshots.
Figure 4.4 : The list of Palm applications. The user taps the Family-Link icon to launch the program.
59
Figure 4.5 : The main screen of Family-Link, with options to view the Family News, Shared Calendars, and Reminders. A button is also included to manually synchronize the data with the server and an exit button allows the user to leave the application.
Figure 4.6 : The user can select whose calendar to view. The Main button returns the user to the main screen while Search My Calendar enables users to find an event based on search terms.
60
Figure 4.7 : The main calendar view where events are listed for the day. “[Notes]” indicates that there is a note attached to the event. The Todo is a toggle button that shows events that are incomplete. The Event Done button allows users to mark events as completed.
Figure 4.8 : Event details of an event. The two icons appear if there is a notes and a chat attached to the event.
61
Figure 4.9 : This screen allows users to edit the details of an event, such as who is involved.
Figure 4.10 : A note can be added to an event.
62
Figure 4.11 : A chat can be added to an event. Different family members can add messages.
Figure 4.12 : Family-Link provides additional information about selected dates in relation to the current day.
63
Figure 4.13 : The Digital Keyboard button brings up the digital keyboard for the user. The Back button enables users to review prior screens.
Figure 4.14 : Four alarms can be set in this family of four. Alarms are specified in relation to the start of the event. For example, “1 min” means one minute before the event start.
64
4.2.4.2 Menu
There is a hidden menu that can be accessed by tapping onto the top left of the screen. This behaviour is consistent with the majority of Palm programs. The menu has the following tabs and options (arranged in a hierarchy with the three main tabs: Study, Settings, and Other):
• Study: This tab holds a number of administrative functions that are password-protected.
• Set my ID: Set the identification number of the device. It is important that each device has a different number and that the server knows which ID’s are associated with which family members.
• Manage data: Allows clearing of all databases, loading of test data, and clearing of the update database which holds new information that the client needs to send to the server.
• Convert events: Provides a facility for importing events listed in Palm Calendar into Family-Link and vice versa.
• Server settings: Adjust the server IP address or port number. Also can initiate a test to communicate with the server.
• Study phase: Adjust which phase the user currently is in. This automatically sets the synching duration differently (A phase: sync every 1 hour. B phase: sync every 3 hours) and also affects the Questionnaire program in deciding which event set to use (events from the Palm Calendar or from Family-Link).
• Settings: This tab is for users to adjust settings in the program.
• Sync settings: Adjust the duration between automatic syncs with the server. Can also disable/enable the synching.
• Last sync details: Show when the last sync occurred.
• Show/hide sync button: Display or hide the Family Sync button on the main screen. When hidden, the synchronization feature cannot be manually activated.
65
• Set vibrations: Sets the number of vibrations for the device when an alarm goes off. This is ignored if the device does not support vibration.
• Full/minimal confirmations: Display all confirmation messages or only a minimal subset of them.
• Show/hide error messages: Display system error information to assist in troubleshooting. Default setting is disabled.
• Other: This tab hosts miscellaneous options.
• About: Shows versioning information about the program.
4.2.5 Hardware
Family-Link is comprised of a number of Palm applications connected to a server application. The server application can be run on any computer that supports Java. The Palm application can be run on any device running the Palm OS 3.5 and up. However, many of Palm handheld devices do not support network data communication and so those devices would be unable to communicate with the server component. The ideal hardware platforms are thus any device running Palm 3.5+ that supports network data connectivity. The Palm Centro and Palm Treo lines of product are good examples. These products offer CDMA2000 1xRTT and EV-DO connectivity. I originally chose the Palm Centro as the platform to evaluate Family-Link since all but one of our design partners preferred Palm devices and were accustomed to the Palm OS. However, after pilot testing I decided that Treo devices were more suitable for the evaluation portion of my thesis work (see Section 5.5.1 for more detail). The latest Palm device, the Palm Pre ( http://www.palm.com/us/products/phones/pre ), does not operate on Palm OS and was not an option for Family-Link.
4.2.6 System Architecture
Family-Link is a distributed application having a client-server architecture. The server and client operate on separate hardware; the server operates on a PC while the client operates on a Palm handheld device. Each user of Family-Link makes use of a client application.
66
4.2.6.1 Server
The server component synchronizes application-specific content from client components. The server awaits client requests for communication and performs the synchronization process as necessary. The server maintains a global view of all Family-Link data from all clients and saves it regularly. It also contains logic for handling concurrency issues between multiple clients.
The primary goals of the server are to:
• Maintain global data and save it to non-volatile storage (i.e., hard drive).
• Accept connections from clients and synchronize their data with the global data.
• Create data reports of information stored on server
• Log data about client connections
The server offers a command-line tool for administration-purposes only. It is not intended for users of Family-Link.
4.2.6.2 Client
The client component is the main application interface for users of the system. Each client application represents one user of the Family-Link system.
The primary goals of the client are to:
• Allow users to create and view calendar events.
• Alarm the user at set times for specified calendar events.
• Synchronize information with the server at regular intervals.
Although the server stores all the Family-Link data, clients maintain a copy of this information so that the user can access this data without needing to connect with the server. This is a typical thick client that processes data and stores it but relies on occasional access to the server. This significantly improves the speed in which users can access data. As well, users can still access the data when network access is limited or unavailable (such as in the subway, an elevator, or in areas where wireless coverage is finicky).
67
4.2.6.3 Communication
The client and server applications communicate over the Internet.
The server utilizes pull technology rather than push technology. Push technology, where the communication originates from the server, is more ideal in terms of sending data to various clients in a timely manner. This is seen in such systems as instant messaging and email (except the last step from mail server to desktop computer). However, push technology is known for being more resource- and power-intensive, both of which are typically limited on a mobile platform. In my case, the choice of hardware platform limited the choice to pull technology. In pull technology, the initial request for communication originates from the client and the server receives that request and handles it. In this sense, the client (receiver) pulls data from the server (sender). An example of a pull technology is RSS.
To implement the pull technology, the server opens a socket and awaits a client connection on a specified port. The client requests communication with the server via a socket. Once the connection is established, the client identifies itself, sends any new information to the server, and downloads any updates that it should receive. Once the server knows that the client has finished reading the updates, it closes the socket and reclaims any internal resources taken by the communication session.
The order of operations is:
1. Server opens socket
2. Client requests communication with server
3. Server accepts client request and asks client to identify itself
4. Client sends identification number
5. Client sends any new information that it hasn’t told the server yet
6. Server receives information
7. Server processes information to check for concurrency conflicts with other clients
68
8. Server stores any new information in update databases so that other clients get the new information when they poll the server
9. Server checks update database for the current client and sends any new information that the current client hasn’t seen yet
10. Client receives data and updates its own databases
11. Server waits until client has read all the data
12. Client acknowledges that data has been read
13. Server receives acknowledgement and closes socket
If an exception is encountered, the server does not save any of the new changes and reverts back to its state before the communication with the client began. Clients that experience an error in any of these steps will not clear the databases and try to resent the new information on the next connection.
4.2.6.4 Implementation
The server application is written in Java. The client application is primarily written in BASIC with a small portion in C++. The client component is assisted by a number of proxy programs that run on the Palm device. Each proxy program is responsible for a different function.
• FLSharedLib: a shared library that offers access to Palm Calendar’s events. This library uses some source files for accessing calendar information in the Palm.
• FLSync: a proxy program that initiates the client communication with the server.
• FLAttn: a proxy program that gets the attention of the user via sound and vibrations.
A Questionnaire program was also developed to collect information during the evaluation phase in the form of multiple-choice questions asked to the primary caregiver. This application has an interface for the user to respond to those questions. Data stored in this application is uploaded to the server via the main client application.
69
Component Programming Lines of Code Language
Main server application Java 3870
Main client application BASIC 16554
FLSharedLib C++ 1218
FLSync BASIC 64
FLAttn BASIC 162
Questionnaire BASIC 2743
Total Lines of Code 24611
Table 4.3: Program components, the programming language in which it was written, and the approximate number of lines of code (including documentation).
4.2.6.5 Limitations
4.2.6.5.1 Only one client connected at a time
The server handles clients in a first come first served basis. Due to security restrictions of the network on which the server PC resides, the server handles these clients in a single thread of execution. This means that while the server is handling a client, no other client can connect. This resulted in problems if clients connected to the server for long periods of time, but a coding fix mitigated the effects of this issue (see Section 7.4.2.2).
4.2.6.5.2 Data is only as current as latest pull
Since I have pull technology between the client and server, any new information that the server would like to send to a client cannot be sent immediately. The server can only retrieve information when that client initiates communication. One way to simulate push technology is to have the client poll the server frequently, asking constantly if it has any relevant information. Doing this would require increased processing resources from the server and clients. As well, there would be a drain on battery life since communications require use of the antenna. As the server is single-threaded, I limited the frequency, using software settings, of clients polling the server.
70
5 Evaluation
Parts of the following chapter will be described in similar form in a forthcoming paper (Wu et al, in press).
5.1 Study Design
I wanted to evaluate the effectiveness of Family-Link and compare it to the Palm Calendar in everyday settings. Laboratory experiments are typically performed under strictly controlled conditions in locations where the behaviour under study may appear out of context. Instead, I designed the study to evaluate the calendar programs under more ecologically-valid settings. Families with a PwA used either the Palm Calendar or Family-Link in alternating phases throughout the study.
The easiest way to set this up would be for me to compare families using Family-Link with whatever memory aids they currently use. However, the new hardware on which Family-Link operates could be a potentially confounding factor. Also, PwAs may find it difficult to switch hardware platforms between phases.
For my study, I chose an ABAB design (Richards et al, 1999; Barlow et al, 1984) involving baseline and intervention phases. There are four phases in this design: first baseline, first
71 intervention, second baseline and second intervention. Baselines are phases during which measures are taken and used as a basis for comparison with data from other phases. Participants used standard commercial technology during baseline, similar to what they have used in the past. Interventions are phases during which study participants use new technology. The ABAB design enabled me to examine and compare the effects of the presence and absence of Family-Link. Two training phases were introduced to the design.
The order of the study phases are as follows: a training phase (T1), a baseline phase (B1), another training phase (T2), followed the intervention phase in which the collaborative aid is introduced (I1), a return to baseline in which the collaborative aid is withdrawn (B2), and finally an intervention phase where the aid is reintroduced (I2). Figure 5.1 shows the ordering of these phases and their approximate durations. Assessment measures were administered according to procedures outlined in Section 5.4 Data Collection.
Customization (2-3 days) Follow -up T1 B1 T2 I1 B2 I2
Training Baseline Training Intervention Baseline Intervention 3 weeks 4 weeks 3 weeks 4 weeks 3 weeks 3 weeks
Interviews and Questionnaires
Figure 5.1: The various phases of the study and approximate durations.
A Palm smartphone device was given to every participant at the start of B1 for the duration of the study. For their participation and time, participants were allowed to keep the smartphone devices, even if they withdrew in the middle of the study.
Family-Link was installed on each device at the beginning of B1 to allow for data collection. At no point throughout the study were participants without the support of a Palm device. I provided different levels of support at different points in the study. PwAs always had access to either the default Palm Calendar or Family-Link.
72
Each family progressed through the various study phases at their own pace. While I set aside approximate durations for each phase, I adjusted the phase lengths to account for events outside my control, such as family vacations. As well, I advanced participants through the T1 and T2 phases if they completed the training in fewer sessions and included additional sessions if more training was required.
5.2 Participants
I recruited six families from Memory-Link for the study (see Table 5.1 for the person codes used in the remainder of this chapter and the next; see Table 5.2, Table 5.3, and Table 5.4 for information about the PwA of each family). The first two families (Families 1 and 2) participated in a pilot study, while the last four families (Families 3, 4, 5, and 6) participated in the actual study. All families live within the Greater Toronto Area. Each family consisted of two to four members. Only adults over the age of 18 participated in the study. Individual participants fell under one of three categories:
a) PwAs,
b) Primary caregivers (typically a spouse), or
c) Secondary caregivers (other family members).
73
Family Person with Primary Secondary Secondary Amnesia Caregiver Caregiver Caregiver
1 A1 C1 - -
2 A2 C2 - -
3 A3 C3 D3 E3
4 A4 C4 D4 E4
5 A5 C5 - -
6 A6 C6 - -
Table 5.1: The participants of the study. Each code represents a person. The code consists of a letter concatenated with a number. Codes beginning with A represent PwAs. Codes beginning with C represent primary caregivers. Codes beginning with D or E represent secondary caregivers. The number in the code indicates to which family the participant belongs.
For participants with memory impairments, inclusion criteria included a diagnosis of amnesia with otherwise generally intact cognitive functioning and fluency in English. Inclusion criteria for family members included being fluent in English and having no history of cognitive impairments. There were several children of the families who did not participate in the study. In general, these children did not help with the caregiving for the PwA. The study information and consent form can be found in the Appendices (see Appendix 10.7).
74
Client A1 A2 A3 A4 A5 A6 Duration in Memory- 2.5 2.5 3 4 1 6 Link (years) Participate d in Family- Yes Yes Yes Yes No No Link Design Sessions? Age at Start of 34 46 53 45 47 55 Study Diagnosis Encephalitis Hydrocephal Limbic Aneurysm Ruptured Surgical us Encephalitis Aneurysm removal of right temporal neocortex and hippocampus
Occupatio Sales coach Account Truck Computer Radio Cabinet n Prior to at financial executive in driver programm frequency maker; Injury institute telecommuni er engineer graduated as -cations sales mechanical engineer Experience Intermediate Expert PC Minimal Expert PC Expert PC Intermediate with PC skills at user and use of PC user user, built PC skills Technolog work, laptop good for web own PC, y Prior to use at home knowledge surfing surfed web Injury of Microsoft and used Excel programs
Table 5.2: Memory-Link clients participating in the study and background information.
5.2.1 Neuropsychological Profiles of PwAs
I obtained neuropsychological profiles (see Table 5.3) comprised of measures of general intellectual ability, verbal and visual memory, and executive function from the PwAs’ clinical record on file in the Department of Psychology at Baycrest. A brief description of each test and the measures selected for presentation is detailed below.
75
Client A1 A2 A3 A4 A5 A6
Age of Assessment 29 45 51 41 45 46
FSIQ 50 95 47 84 45 50
Verbal Memory 27 10 10 <1 2 33 Learning
Verbal Memory 2 <1 2 <1 3 40 Delay
Visual Memory 39 50 58 10 1 50 Learning
Visual Memory 27 42 69 10 1 45 Delay
Executive 9 12 37 30 1 Unavailable Function
Table 5.3: Memory Link neuropsychological profiles of study participants prior to injury. A1 and A2 are pilot participants, while A3, A4, A5, and A6 participated in the main evaluation. All data are reported as percentile rankings. See Table 5.4 for interpretations of the percentile rankings.
Percentile Ranking Interpretation
≤2nd percentile Impaired
3rd -8th percentile Borderline
9th -24 th percentile Low Average
25 th -74 th percentile Average
≥75 th percentile High Average–Superior
Table 5.4: Interpretation of percentile rankings for Table 5.3.
76
Notice that in terms of overall percentile rankings, A4 and A5 were more similar than A3 and A6 in that the former two PwAs were significantly impaired or borderline for memory learning and delay percentile rankings, with the exception of A3’s verbal memory learning which was low average. A3 and A6 both had average memory learning and delay percentile ranking, with the exception that A3 was low average in verbal memory learning and impaired in verbal memory delay.
Intellectual ability was determined through administration of the Wechsler Abbreviated Scale of Intelligence (Wechsler, 1999). The Full Scale IQ (FSIQ) appearing Table 5.3 is a global composite score based on the measurement of both verbal and non verbal abilities.
Verbal memory was assessed by the California Verbal learning Test (Delis, Kramer, Kaplan and Ober, 1987). This test requires the examinee to recall a 16 word list presented over five trials. Following a 20-minute delay, during which non-interfering tasks take place, the examinee is again asked to recall the list. Two scores appear in Table 5.3: the number of items correctly recalled across the five learning trials, and a measure of retention following the 20-minute delay.
Visual memory was assessed by the Brief Visual Memory Test – Revised (Benidict, 1997). This is multi-trial test of figural learning. Two scores appear in Table 5.3: the number of designs correctly recalled across three learning trials, and a measure of retention following a 20-minute delay.
Executive function was measured through the Wisconsin Card Sorting Test (Heaton, 1981). This is an abstract problem-solving task requiring the examinee to discover rules governing correct performance. It measures the ability to form abstract concepts, shift and maintain set, inhibit prepotent responses, and utilize feedback. The score reported in Table 5.3 is the number of categories completed which provides a measure of concept formation.
These scores are standardized and derived in the context of the normal population for any given age or gender (see Table 5.4 for interpretations of percentile rankings). The percentile score indicates severity of someone's impairment (i.e., how that individual scored in comparison to the normative table). For example, someone who scores at the second percentile performs below 98% of all people of the same age and gender that took the test when establishing the norms.
77
5.2.2 Family 1
Family 1 is a two-person family of A1 and C1. The timeline of Family 1 is shown in Figure 5.2.
Customization Follow -up T1 B1 T2 I1a I1b B2 I2
Training Baseline Training Inter- Inter- Baseline Inter- 2 weeks 3 weeks 8.5 vention vention 2 weeks vention weeks 6 weeks 7 weeks 5 weeks
Interviews and Questionnaire s
Figure 5.2: The timeline of Family 1. See Section 5.5 for details about why this schedule is different from Figure 5.1.
A1 is 34 years old and acquired amnesia in 2005. At the end of that year, he attended Memory- Link and was trained to use the Palm Calendar in early 2006. Before the study, A1 used the Palm Zire 72 for two and a half years. Although he did not use Palm devices before his memory impairment, A1 is not afraid of technology. A1’s participation in the Memory-Link program depended greatly on another Memory-Link client, A2 (from Case 2), because A2 often drove A1 to the psychosocial group meetings at Baycrest. As a result of the time spent together, A1 and A2 have developed a good friendship through the years. In December 2006, A1 joined our participatory design team and contributed to the design of the collaborative software.
A1’s wife, C1, is 37 years old and works part time as a flight attendant. Her work schedule necessitates that she fly out of town for stretches at a time. As a result, the couple goes through a repeated pattern of being apart for a series of days and then being together for a series of days. C1’s work schedule is set every month. At the end of each month, she needs to bid for flights for the upcoming month. The bids are not guaranteed, but once the schedule is set for the month, it does not change. Since C1’s schedule changes from one month to the next, there is a period of uncertainty when A1 does not know when C1 will be in town. This impacts their planning of events as C1 tries to schedule A1’s meetings with doctors when she is in town, thus allowing her to attend the meetings and be aware of the information that the doctors provide. She does not rely
78 on A1 remembering that information nor does she rely on him adding that information in his Palm device. A1 did not participate in our design sessions.
5.2.3 Family 2
Family 2 has two parents who participated in the study, A2 and C2, and two children who did not. The timeline of Family 2 is shown in Figure 5.3.
Customization T1 B1 T2 I1 B2
Training Baseline Training Intervention Baseline 4.5 4 weeks 4 weeks 5 weeks 0.5 weeks weeks
Interviews and Questionnaires
Figure 5.3: The timeline of Family 2. See Section 5.5 for details about why this schedule is different from Figure 5.1.
A2, who has amnesia, is 46 years old. A2 came to Baycrest and was assessed for amnesia in early 2006. He already owned an IPAQ (PocketPC) device and chose to continue using it in lieu of Palm training. Through his own practice, A2 was able to successfully add appointments to the calendar application on his PocketPC. Incidentally, The IPAQ screen was accidentally damaged near the beginning of the study. I trained A2 in the use of the Palm device and had him use it exclusively. In December 2006, A2 joined our participatory design team.
C2 is A2’s wife and is 40 years old. C2 works full-time as manager at a wireless company. She owns a Blackberry for work and uses it to store work-related appointments. C2 did not participate in our design sessions.
A2 and C2 have two daughters, aged 13 and 11, who lived in the same household. The children did not participate in the study. They did not help with the caregiving.
79
5.2.4 Family 3
Family 3 consists of two parents (A3, C3) and two daughters (D3 and E3). The timeline of Family 3 is shown in Figure 5.4.
Customization Follow -up T1 B1 T2 I1 B2 I2
Training Baseline Training Intervention Baseline Intervention 3 weeks 4.5 3.5 3.5 weeks 4.5 weeks 3 weeks weeks weeks
Interviews and Questionnaires
Figure 5.4: The timeline of Family 3.
A3 is a 53-year old man who acquired amnesia in September 2004. He began attending Memory- Link in April 2006 and was trained to use the Palm Zire 72. Before the study, A1 used the Zire for three years. In December 2006, A3 joined our participatory design team.
C3 is A3’s wife and is 49 years old. Prior to the study, C3 did not use any handheld technology. She works full-time at a day care centre. However, during I2 there was a strike by the Canadian Union of Public Employees Local 79 and the Toronto Civic Employees' Union Local 416 which resulted in C3 not working during the entirety of I2. C3 participated in two sessions of our design team.
A3 and C3 have two daughters, D3 and E3, who live in the same household. D3 is 24 years old and is an elementary school teacher. D3 was affected by the strike and did not work during I2. E3 is 19 years old and is a university student. Neither daughter participated in our design sessions.
80
5.2.5 Family 4
Family 4 consists of two parents (A4, C4), one daughter (D4) and a son-in-law (E4). D4 and E4 have a baby that is four-months old. The timeline of Family 4 is shown in Figure 5.5.
Customization T1 B1 T2 I1 B2 I2
Training Baseline Training Intervention Baseline Intervention 3 weeks 4.5 3.5 5 weeks 3 weeks 6 weeks weeks weeks
Interviews and Questionn aires
Figure 5.5: The timeline of Family 4. C4 was unable to participate in the Follow-up interview.
A4 is a 45-year-old man and has been living with amnesia since November 2004. Before his amnesia, A4 worked as a computer programmer. He began attending Memory-Link in 2005 and was trained to use the Palm Zire 72. He used this device for four years before the study began. In December 2006, A3 joined our participatory design team.
C4 is A4’s wife and is 43 years old. She works full-time as a production coordinator at a publishing company. C4 participated in two design sessions.
D4 is the only daughter of the family and is 26 years old. She works in the administrative department of an insurance company. She does not use any handheld technology.
E4 is D4’s husband. He is 35 years old. He works at the same insurance company as D4. E4 has had some past experience with older Palm devices but did not use one for personal organization prior to the study.
D4 and E4 participated in two design sessions. D4 conceived a baby a few months before the study began.
81
5.2.6 Family 5
Family 5 has two participating members: A5 and C5. They have a son but he did not participate in the study. The timeline of Family 5 is shown in Figure 5.6.
Customization Follow -up T1 B1 T2 I1 B2 I2
Training Baseline Training Intervention Baseline Intervention 3 weeks 3 weeks 4 weeks 3.5 weeks 4 weeks 6 weeks
Interviews and Questionnaires
Figure 5.6: The timeline of Family 5.
A5 is a 47-year-old man who was assessed with amnesia in September 2003. He previously worked in the electronics and radio communications industry before his brain injury. A5 began attending Memory-Link at the beginning of 2008. As such, he had a little over a year of experience operating the Palm Zire 72 prior to the study. A5 participated in two sessions of our design team.
C5 is A5’s wife and is 26 years old. She is not currently employed and spends most of her days with A5 and assisting him with transportation as A5 is in a wheelchair. C5 participated in two design sessions.
A5 and C5 have a 23-year-old son who lived in the same household. He had recently finished university and was seeking employment. C5 did not feel that her son participated in providing reminders and daily care for A5. He did not participate in the study.
82
5.2.7 Family 6
Family 6 is a two-person family (A6 and C6). The timeline of Family 6 is shown in Figure 5.7.
Customization Follow -up T1 B1 T2 I1 B2 I2
Training Baseline Training Intervention Baseline Intervention 3 weeks 5 weeks 4 weeks 4 weeks 3 weeks 3 weeks
Interviews and Questionnaires
Figure 5.7: The timeline of Family 6.
A6 is a 55-year-old man who had a brain surgery and acquired amnesia in 1997 and attended Memory-Link in 2003, where he was trained to use the Palm m100. He switched over to the Palm Zire 72 when his m100 device broke down. In total, A6 has approximately six years of Palm experience prior to the study. Before his brain injury, A6 worked as a cabinet maker and enjoyed manual drafting. A6 was not a part of our design team. However, he was a member of the design team in my Masters research in 2003 (Wu, 2004).
C6 is A6’s wife and is 58 years old. Prior to the study C6 worked as a language teacher under contract. Just before the study began, her work contract expired. She began putting together her own business and spent much of her time at home on the computer. C6 did not participate in our design sessions.
5.3 Study Procedure
5.3.1 First Training Phase (T1)
The first phase was a training phase (T1) in which participants were taught how to use the new device. This was particularly important for PwAs as the new hardware carried slight nuances (e.g. location of the power button, location of the stylus, button to disable screen guard) that made it difficult for someone with memory problems to pick up and easily use on first attempt.
83
The hardware platform was new to all study participants. This training also allowed them to become accustomed to the newest version of the PalmOS which ran on the Palm smartphones (this PalmOS had different programs on the Zire 72). The duration of this T1 phase varied depending on individual differences in trainablity but typically lasted 3 weeks.
For PwAs, I applied principles of errorless learning (Wilson et al, 1994) and vanishing cues (Glisky et al, 1986) to the training procedures, aimed at helping PwAs acquire procedural skills. The training techniques were adapted from procedures used by researchers in Memory-Link to train PwAs to use the Palm Calendar (Svoboda and Richards, 2009). This involved having a researcher sit beside and guide a trainee through the skills that needed to be acquired. The training forms can be found in the Appendices (see Sections 10.10, 10.11, 10.12, 10.13, 10.14).
Family members participated in separate training sessions during this time as well. This was particularly important for those who did not know how to use Palm devices. Since family members could rely on intact memory systems when learning, the rigourous techniques described above were not necessary.
For study participants, the following interactions were taught:
• Turning the Phone On and Off (important for study data to be sent to server)
• Resetting the Device (in case of a device freeze)
• Adding Event in Palm Calendar
• Responding to an Alarm from Palm Calendar During T1, PwAs continued to use whatever memory aids (e.g. Palm Zire 72) they had used before the study began.
5.3.2 First Baseline Phase (B1) and Second Baseline Phase (B2)
During the baseline phases (B1 and B2), families used the default Palm Calendar that came pre- packaged with the Palm devices. They did not have access to Family-Link. The first baseline was typically longer than the second because it took extra time for PwAs to become used to the hardware. Before B1 began, the devices were set up using the steps outlined in Appendix (see Appendix 10.15).
84
5.3.3 Second Training Phase (T2)
Prior to the I1 phase, study participants went through a training phase (T2). Participants were trained on how to use the Family-Link software. The duration of this phase varied but typically lasted 3 weeks. PwAs were trained using the same techniques applied during T1. Family members participated in separate training sessions.
For all participants, training was given for the following interactions:
• Adding Event in Family-Link
• Adding a Note in Family-Link
• Adding a Chat Message in Family-Link
• Marking an Event Complete
Training forms can be found in the Appendices.
During T2, PwAs used the default software on the Palm smartphones.
5.3.4 Customization
At the end of T2 and prior to full in vivo implementation (I1), I worked with each family member to configure their device according to their needs (e.g. adjusting default settings, installing Palm PC software, setting up device backup through the PC, copying existing contacts to the new device).
5.3.5 First Intervention Phase (I1) and Second Intervention Phase (I2)
During the intervention phases (I1 and I2), families had full access to Family-Link but did not have access to the default Palm Calendar. The icon for the Palm Calendar was hidden.
5.4 Data Collection
Data collection consisted of electronic logs, face-to-face interviews, phone calls and electronic and paper-based questionnaires.
85
5.4.1 Interviews and Phone Calls
At the end of each baseline and intervention phase, I conducted an individual interview with each participant. This was done in person, if a one-on-one meeting was possible, or done over the phone. Individually speaking to each participant allowed participants to speak freely without needing to consider if their comments could be viewed as judgmental by other family members. As well, PwAs would not take things that their family members remembered for truth. This could be misleading because of the chance of memory failure in family members without impairment. The interviews lasted approximately 45 minutes. Interviews were audio recorded using a Palm Centro if done in person or using Skype Call Recorder ( http://atdot.ch/scr/ ) if done over the phone. The audio tracks were used for transcription and analysis (participants could decline recording with no negative consequences, but none of the participants declined).
I also called the primary caregiver once a week to check on how things were going in the study. These calls only took a few minutes per call and allowed caregivers to report any interesting events that occurred. Notes were recorded for these conversations but audio was not.
5.4.2 Standardized Paper-Based Questionnaires
At the end of each baseline and intervention phase, I met with each participant and administered a paper-based questionnaire package. It took approximately 10 minutes to complete the forms. Paper-based questionnaires examined anxiety in the person with amnesia and caregiver burden in family members. To measure anxiety, the Beck Anxiety Inventory (Beck et al, 1988) was administered (See Appendix 10.16). To measure caregiver burden, the Zarit Burden Interview (Zarit et al, 1980,1985) was administered (See Appendix 10.17).
A comparison questionnaire was created specifically to assess user preferences between the Palm Calendar and Family-Link. As well, Usefulness Questionnaire was designed for assessing the usefulness of the various features of Family-Link. Both these instruments were given during the last interview.
5.4.3 Electronic Questionnaires
I could not rely on all information reported by PwAs due to their memory impairment, so I turned to caregivers for primary reporting. Although caregivers were not around to witness all
86 things that happen to the PwA, they are often aware of — at least at a high level — the progress and status of events. I originally wanted to make telephone calls to caregivers every day to collect information about how well events were being completed, but I soon realized that doing so would be highly disruptive. Calling caregivers, even if done at agreed-upon times, and having them report on daily events interrupted their days and taxed their time.
I created questionnaire software and installed it on all primary caregivers’ smartphones. This allowed the primary caregivers to respond to the questions when it suited them best. Each primary caregiver was asked to run this software and answer its questions every three days during baseline or intervention phases and everyday during the last week of each phase. The software asked five multiple-choice questions and one optional freeform question for every event listed in the calendar of the family member with amnesia (see Appendix 10.8 for screenshots and Appendix 10.9 for the list of questions). The freeform question was an empty field that could be filled with additional comments that the caregiver might wish to add.
Primary caregivers were asked to answer these questions at the end of every two or three days. It took approximately 10-15 minutes to complete this questionnaire.
5.4.4 Electronic Data Logging
Data about events created in either the Palm Calendar or in Family-Link was logged by the handheld device and electronically transmitted to the server PC every hour. Questionnaire data was also logged by the handheld device and transmitted to the server PC automatically.
5.4.5 Observation Sessions
I scheduled one 8-hour observation session with each family during I2. Notes were taken about use of the device and software.
5.4.6 Follow-up Interviews
I interviewed the primary caregivers of each family approximately one month after the end of the study to gather additional information. The interviews were conducted over the phone and were recorded using Skype Call Recorder ( http://atdot.ch/scr/ ). The questions can be found in Appendix 10.21.
87
5.5 Pilot Study
I ran a pilot study with two families (Families 1 and 2) to test the study procedures and assess the practicality of my study design. Based on what I learned from working with these families, I modified the procedures. The procedures mentioned above are the final outcome after pilot work. Changes to the procedures include the following:
• The phase duration was not set to a fixed 3-week duration. Keeping this flexible allowed me to account for any issues such as the family going on vacation and not using the device during that time. Some phases went longer than 3 weeks because it took longer for PwAs to get used to the device.
• Although I put family members of Families 1 and 2 through all the training procedures, I decided this was not necessary after seeing that they learned the procedures more rapidly than the PwAs.
• I removed training for the phone features as they were not used.
• I introduced a customization step to adjust the default settings on the device for each participant. This was particularly important for my work with PwAs and is emphasized in much prior research of technology for persons with cognitive impairment (Cole, 2006; Cole and Matthews, 1999).
• It quickly became evident that the time of caregivers was limited and I was worried about participants withdrawing from the study because too much time was demanded of them, so I removed extra scales that were not essential. I removed the Multidimensional Caregiver Strain Index (MCSI) (Stull, 1996), the Family Adaptability and Cohesion Evaluation Scale (FACES) (Olson et al, 1979; Olson et al, 1985), the McMaster Family Assessment Device (FAD) (Epstein, Baldwin, and Bishop, 1983), a customized Family Involvement Questionnaire, and a customized self-efficacy scale (Bandura, 2006). MCSI was redundant since I was already using the Zarit Burden Interview (Zarit et al, 1980,1985). FAD in particular was time consuming as it was a 60-item scale.
• Since most scales were removed, I shifted from using the 12-item Short Zarit Burden Interview (Bedard et al, 2001), which is more of a screening tool, to using the complete 22- item Zarit Burden Interview (Zarit et al, 1980,1985).
88
• Originally, there was an optional phase after B2 in which participants were able to choose whether to continue using Family-Link or to switch back to the Palm Calendar. I removed this phase to simplify the design.
• Originally, I created a form for assessing how jobs were completed, and used it over the phone to ask PwAs for more information about tasks that were completed during the day. However, PwAs were unable to recall such information with clarity. I then created the electronic questionnaire for caregivers.
• I refined the electronic questionnaire software and interface based on feedback from C1.
• I added a procedure of making brief phone calls to each family to check on participants and help answer any technology questions every week. Some participants did not call me because they assumed that when the technology failed, nothing could be done to resolve it. However, many of these issues were brought up by participants in the brief phone conversations and turned out to be slight misconceptions about how Palm devices or Family-Link worked. The phone conversations helped to resolve those issues. Most of the conversations were with the PwAs as they were available to talk, although I tried to speak with the caregivers whenever possible.
• Palm Treos were used rather than Palm Centros. The PwAs in the pilot study experienced difficulty using the Centro’s physical keyboard and its somewhat small and non-responsive touch screen. This is further detailed in Section 5.5.1.
• For other changes, please see Section 7.4.
Family 1 completed the pilot study, but the intervention phase, I1, was modified. C1 and A1 both reported not having used the tool during the first intervention phase because of various emergencies that came up that involved themselves as well as their extended family or others. For instance, their neighbours’ house had flooded so C1 and A1 provided accommodation to their neighbours while they sorted things out. Another time, both C1 and A1 had to travel to Montreal as C1’s mother was ill. There was a gap between I1a and I1b in which C1 developed bronchitis and was hospitalized. A1 also fell ill during this time. As a result, the study was put on hold. I decided to carry out another intervention phase, I1b, that occurred after the original I1a phase.
89
Family 2 withdrew from the study at the end of I1. The decision was made by C2 because of an incident in which A2 secretly sought out C2’s questionnaire responses without her knowledge. When A2 found the responses, he confronted C2 and expressed displeasure. C2 felt that she could not continue the study because she could not answer the questions honestly if A2 had the slightest chance of getting access to her answers. The Memory-Link neuropsychologist provided individual counseling sessions to the PwA and spouse to address and resolve these issues. C2 also cited a second reason for termination of the study. She reported that the change from IPAQ to Palm was difficult for A2. C2 felt that he did not operate with the Palm as well as he did with his IPAQ.
"Between you and me, A2 is really struggling with not using his IPAQ and all the software options it provides, ease of usability, familiarity etc. I have asked him to get it fixed and use it as the primary tool while still trialing the Palm. I need more help from him and he is really off as of late, there is a difference since he started using the Palm alone. The overall impacts to our family from an organizational perspective and keeping him on task are just too great to continue with Palm alone. I hope you understand... the Palm offers a fraction of the "rest" of the functionality that we need to allow him to work efficiently. The irony is that the family software offers a great option to communicate real time on items that crop up.” (C2, I1)
Since Families 1 and 2 followed different procedures and Family-Link evolved during the pilot study based on feedback, I removed the quantitative data collected from these families in the next chapter.
5.5.1 Observations Resulting from the Change in Hardware
While the choice of a Palm device for the study was appropriate given the prior experiences of our design partners and study participants, slight differences in physical form factor resulted in PwAs having usability issues. The Zire 72 was used pre-study because Memory-Link clinicians chose this platform for their training. Centro devices were used in the pilot study because they supported wireless connectivity while the Zire 72 did not. However, when Treo devices were made available by one of the wireless carriers in Canada, I decided to use them in the actual study because of usability issues with the Centro that were raised in the pilot.
90
The most immediate and obvious difference between the Palm Centro and the Zire 72 is the size of the device (See Table 5.5). The Zire 72 also has a dedicated writing area (separate from the display screen) that uses a propriety single-stroke shorthand handwriting recognition system called Graffiti. In place of the writing area, the Centro and Treo product lines have a plastic physical keyboard.
Palm Dimensions Weight (g) LCD Resolution Text Input Release Device (WxHxD) (cm) Screen (pixel) Method Year Size (cm) Zire 72 7.5 x 11.6 x 1.7 136 5.4 x 5.4 320 x 320 Handwriting 2004 or digital keyboard Centro 5.4 x 10.7 x 1.8 124 3.9 x 3.9 320 x 320 Digital or 2007 physical keyboard Treo 5.8 x 12.9 x 2.3 180 4.4 x 4.4 320 x 320 Digital or 2006 700p physical keyboard
Table 5.5 : The specifications of the Palm Zire 72, Palm Centro, and Palm Treo 700p.
Although the devices have the same resolution display, the screen sizes of the Centro and Treo devices are smaller than that of the Zire 72. My pilot study revealed that A1 and A2 had issues in hitting the smaller targets. This often led to incorrect spelling and corrections that noticeably increased the amount of time taken when creating appointments.
Figure 5.8 : Shown from left to right: the Palm Zire 72, Palm Centro, and the Palm Treo
91
It would seem wise to have a larger screen for seeing and tapping onto targets, particularly when considering designs for an older population in which visual and motor deficiencies are more common. However, recent Palm devices like the Palm Pre (http://www.palm.com/us/products/phones/pre ) have been designed to be smaller and lighter. This may appeal to a younger market, but many older adults would have difficulty accessing such a system. The smaller devices require finger dexterity and sharp vision, both qualities that can become diminished as one ages.
5.6 Hypotheses
I intended to find evidence from the study supporting or refuting the following hypotheses:
1. PwAs will experience less anxiety during Intervention phases as compared to Baseline phases.
2. PwAs will perform events more completely during Intervention phases as compared to Baseline phases.
3. PwAs will perform events more on time during Intervention phases as compared to Baseline phases.
4. Caregivers will have greater confidence in the ability of PwAs to complete events during Intervention phases as compared to Baseline phases.
5. Caregivers will require less effort to coordinate family activities during Intervention phases as compared to Baseline phases.
6. Caregivers will feel less frustration during Intervention phases as compared to Baseline phases.
7. Caregivers will have less caregiver burden during Intervention phases as compared to Baseline phases.
8. All participants will share more of their events during Intervention phases as compared to Baseline phases.
92
9. All participants will feel greater awareness of other family members’ schedules during Intervention phases as compared to Baseline phases.
10. Participants will find features of Family-Link useful.
11. Participants will prefer Family-Link over Palm Calendar.
93
6 Analysis and Results
This chapter presents the analysis of the qualitative and quantitative data collected from the longitudinal study. The quantitative data includes paper and electronic questionnaire responses, and Palm Calendar and Family-Link data collected from the Treo devices. Qualitative data includes responses from phone and in-person interviews. Parts of this chapter will be described in similar form in a forthcoming paper (Wu et al, in press).
In the first seven subsections (6.1 to 6.7), I describe analysis of measures to evaluate the system in general. Note that for 6.2 and 6.3, I did not want to rely on the memory of PwAs for data collection, and I was unable to observe all families simultaneously, so I derived an indirect measurement for hypotheses 2 and 3 (see Section 5.6) based on the primary caregiver. The results in the first seven subsections suggest that Family-Link and Palm Calendar are not significantly different when considering these measurements. This is an achievement given that this is a prototype and the Palm Calendar is a fully developed product.
The next two subsections (6.8 to 6.9) relate to sharing and awareness, which are direct measures of collaboration. I found quantitative and qualitative evidence supporting the hypothesis that Family-Link increased sharing of calendar events between family members and qualitative results supporting the hypothesis that Family-Link increased awareness.
94
Finally, the final two subsections (6.10 to 6.11) relate to the perceived usefulness of Family-Link and to user preferences for one or the other of the two systems.
Quantitative data is analyzed using descriptive statistics and statistical tests (Friedman’s test, Wilcoxon Signed-Rank test, Sign test, Repeated Measures Analysis of Variance 1, and post-hoc comparisons with t-tests and Bonferroni correction where appropriate). The following hypotheses were used for each statistical test that compared means across phases:
• H0: There is no difference between the means across phases for the measurement.
• HA: There exists a difference between the means across phases for the measurement.
The following table (see Table 6.1) shows the number of data points captured by the electronic questionnaire. This data set was used for statistical testing in Sections 6.2 to 6.6, 6.9.2.1 and 6.9.2.2. The quantitative data from the pilot Families 1 and 2 were not included because they participated in the pilot study. However, some qualitative data from these two families were added for discussion. For Families 3 to 6 combined, there are a total of 269 Baseline responses and 252 Intervention responses.
Phase Family 3 Family 4 Family 5 Family 6 B1 60 90 48 16 I1 55 25 50 7 B2 36 0 16 3 I2 56 11 39 9 Total 207 126 153 35
Table 6.1 : The number of events for which the electronic questionnaire was answered by primary caregiver, organized by phases.
C4 did not fill out the electronic questionnaire during B2. This led to missing data. In some statistical tests, this missing data resulted in the Family 4 not being included in the overall analysis. The descriptive statistics tables reflect this missing data and these are the values used in the statistical tests that follow (see Table 6.5, Table 6.7, Table 6.9, Table 6.11, Table 6.13, Table 6.19, Table 6.20). However, all graphs include Family 4’s data to visually illustrate general trends.
1 Ideally, hierarchical linear modeling (HLM) should be used since I have a nested design (e.g. PwAs/caregivers nested in families), but given that N=4 families, I used a slightly less accurate statistical model.
95
It is important to note that statistical testing removes outliers, yet persons with cognitive impairments, with their unique abilities, represent the extremes. I therefore embed qualitative data throughout the quantitative results to present a more complete understanding in each section. For the qualitative data, I applied a coding scheme derived from my hypotheses (see Section 10.5) to the transcripts of the interviews. This helped to classify and organize the quotes.
6.1 PwAs’ Anxiety
Hypothesis: I anticipated that PwAs would experience less anxiety during Intervention phases as compared to Baseline phases.
Analysis: To measure anxiety in PwAs, I administered the Beck Anxiety Inventory (BAI, see Appendix 10.16) to each participant at the end of each phase. According to the Revised BAI Manual (Beck and Steer, 1993), the scoring outcomes are listed below (see Table 6.2).
Total BAI Score Interpretation 0 to 7 Minimal anxiety 8 to 15 Mild anxiety 16 to 25 Moderate anxiety 26 to 63 Severe anxiety
Table 6.2 : The interpretations for the various BAI score ranges.
Figure 6.1 : Mean anxiety scores for PwAs across conditions.
96
The mean anxiety scores across all phases (see Figure 6.1 and Table 6.3) would be interpreted as minimal anxiety according to BAI. This was confirmed by interviews, where all PwAs reported having no anxiety in any of the phases. As A4 put it:
“No, but I’d really have to say that it’s in my nature not to be afraid of a little computer. Because I’m a computer person, right? You know, I’m kinda anxious to get my hands on it and see what it will do. (laughs) So I can see maybe some people be nervous about it definitely. But not me personally because I’ve got a bit of computer experience.” (A4, I1)
Mean Standard N Deviation B1 4.7500 4.27200 4 I1 4.7500 7.08872 4 B2 2.0000 2.44949 4 I2 1.5000 1.29099 4
Table 6.3 : Descriptive statistics for anxiety score by phase.
A repeated measures analysis of variance (RM-ANOVA) was carried out to determine if there was an effect between anxiety scores across phases. The dependent variable was the anxiety scores and the within-subjects variable was the phase. The assumption of sphericity for RM- ANOVA was met. Mauchly's Test of Sphericity was not significant ( χ2(5)=9.601, p=0.133). The results showed that there was no significant effect between the anxiety scores and the phase, F(3,
9)=1.043, p=0.420. Therefore, H0 was accepted and HA was rejected.
The BAI gives proportionately more emphasis to the somatic symptoms of anxiety (which are usually only present at higher anxiety levels) and as such it may have missed the more nuanced feelings of being unsettled that were picked up in interviews. Interviews with the primary caregivers revealed that PwAs were initially uncomfortable with the change in hardware (from Zire 72 to Treo). Hardware novelty may have created some increased anxiety in PwAs not captured in the BAI at the beginning of the study.
C4 commented on this issue.
“I think it’s only been one day... where things just were not met by time…I think it’s apprehension on their part... No matter how many times you’re going to call home and
97
check to make sure it got done, it’s not getting done. I at one point said ‘Just put the thing (Treo) away and leave it alone. I can’t tell you much longer because I’m at work’. So it did cause a major frustration that day… it was probably the first week into it (the study). With the apprehension of something that he’s gotta use.” (C4, B1)
As the following pair of quotes illustrate, C5 spoke about needing to hide the Zire 72 at the beginning of the study because A5 would try to use it instead of the Treo.
“I don’t know if it was having both of them (during the B1 phase) and maybe it was a little confusing. Because once I put old one (Zire 72) away he seemed to be fine. And I’ll snitch on him again. When he got home, he put the calendar that you, you know (hid). He put it (the Palm Calendar) back. But my son took it (Palm Calendar) away again. It’s that separation thing. You know, I guess you’re probably finding it with everybody it’s hard when you get used to the one thing and you depend on it and it becomes a piece of you and then to change is difficult.” (C5, I1)
C3 reported A3 becoming comfortable with using the correct Treo near the end of B1.
“Now there was times where he was grabbing mine (device) and putting entries in and it was his. So maybe that’s partly why I put it away because we decided at some point we needed to get different cases or something because how does he know whose (device) is whose (device)… I think he’s comfortable, I think it’s been a week or more since he’s asked me where it is.” (C3, B1)
C4 explained that the study still presented a transition for A4, particularly in the Intervention phases.
“As you use it, the more you use it, the better you become at it, the more accustomed you are to using it... People find it hard to do something new all the time, because they’re constantly juggling their own. So to have only a 3 week limit as I mentioned to you before – a 3 week limit of doing the new program versus the old one, it makes it difficult for them (PwAs). And then they have more of a, you have more of a challenge on your hand getting them to go into it because they are accustomed to this one. So the length of time to use the new program should technically be extended as I mentioned, for the new
98
program. It’s not the old program that needs to be extended because they’re accustomed to that. They’ve been flying with it for the last year or two years or more. It’s the new program because everybody has a problem with change and then when you’re so accustomed one thing and then you get toppled over to something else.” (C4, I1)
Summary: According to the BAI scores, participants experienced minimal anxiety in all phases. There was no significant difference between anxiety scores across phases. However, interviews revealed that PwAs felt unsettled at the beginning of the study. This was because of PwAs’ unfamiliarity with the new hardware introduced at B1, rather than the effect of the Palm Calendar.
6.2 PwAs’ Completeness of Events
Hypothesis: I anticipated that PwAs would perform events more completely during Intervention phases as compared to Baseline phases.
Analysis: As mentioned previously, event completeness was judged by the primary caregiver of each family. The electronic questionnaire (see Appendix 10.9) asked primary caregivers about the completeness of events listed in the care recipients’ schedules. (“Do you believe this was done?”) A numerical score was assigned to each response (see Table 6.4).
Response Assigned Score All of it 5 Most of it 4 Half 3 Some of it 2 Not at all 1 Don’t know Excluded
Table 6.4 : Responses to the completeness question and the scores that were assigned. Responses of “Don’t know” were excluded from the data.
99
Figure 6.2 : Completeness of events as perceived by the primary caregivers by phase. A higher score here indicates that a greater proportion of the events were completed. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.
Events were mostly completed both conditions (see Figure 6.2 and Table 6.5).
Mean Standard N2 Deviation B1 4.5567 .47078 3 I1 4.2833 .56448 3 B2 4.5033 .86025 3 I2 4.0333 .65546 3
Table 6.5 : Descriptive statistics for completeness response by phase. These values are used for the statistical test.
2 C4 did not fill out the electronic questionnaire during B2. The missing data resulted in the Family 4 not being included in the overall statistical analysis, so N=3 instead of N=4.
100
A Friedman test was conducted to evaluate differences in means for the completeness responses across phases. The test was not significant χ2(3, N=3)=3.400, p= 0.334, so H0 was accepted and
HA was rejected.
Although there was no significant difference between completeness response means across phases, averaging across participants can be misleading (Vicente and Torenvliet, 2000) as the intervention may have had an effect on some families and not others. To assess this, each family was analyzed separately using independent samples t-tests. This does not allow me to generalize beyond my study participants but it may provide insight to individual performance.
A Bonferroni correction was used to address the issues resulting in making multiple t-test comparisons. A statistical level of 0.013 (i.e. 0.05/4) was used to determine significance (i.e., p<0.013 instead of p<0.05). There was one significant difference in the tests, where the primary caregiver of Family 3 believed that events were completed to a greater degree in I2 (5.00) than in B1 (4.57), one-sample t(89)=3.539, p=0.001.
Follow-up interviews with primary caregivers confirmed that there was no perceived difference between phases in completeness of events.
Summary: Results suggest that there were no significant differences between the phases. Events were perceived by caregivers to be mostly completed in both conditions.
6.3 PwAs’ Timeliness of Events Completed
Hypothesis: I anticipated that PwAs would perform events more on time during Intervention phases as compared to Baseline phases.
Analysis: As mentioned previously, event timeliness was judged by the primary caregiver of each family. The electronic questionnaire (see Appendix 10.9) asked primary caregivers about the timeliness of events listed in the care recipients’ schedules. (“Was this done on time?”) A numerical score was assigned to each response (see Table 6.6).
101
Response Assigned Score Very early 1 A little early 2 On time 3 A little late 4 Very late 5 Don’t know Excluded
Table 6.6 : Responses to the timeliness question and the scores that I assigned. Responses of “Don’t know” were excluded from the data.
Figure 6.3 : Mean timeliness of events by calendar condition. A score below 3 indicates that events were completed earlier. A score above three means that events were completed late. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.
The means for each calendar condition were slightly above 3 (see Figure 6.3 and Table 6.7), suggesting that events were, on average, perceived to have been done a little late.
102
N3 Mean Standard Minimum Maximum Deviation B1 3 3.137 .1601 3.00 3.30 I1 3 3.2833 .34675 3.00 3.67 B2 3 3.1300 .22517 3.00 3.39 I2 3 3.0733 .12702 3.00 3.22
Table 6.7 : Descriptive statistics for timeliness response by phase. These values are used for the statistical test.
A Friedman test was conducted but was not significant χ2(3, N=3)=2.520, p= 0.472, so H0 was accepted and HA was rejected.
Each family was further analyzed separately using independent samples t-tests and a Bonferroni correction (i.e., p<0.013 instead of p<0.05). There was one significant difference in the tests, where the primary caregiver of Family 4 believed that events were completed more on time in I2 (3.00) than in B1 (3.32), one-sample t(86.000)=-4.458, p<0.001.
Follow-up interviews with primary caregivers confirmed that there was no perceived difference between conditions in lateness of events.
Summary: As perceived by primary caregivers, events were completed a little late on average. There were no significant differences in timeliness response means across phases, as late events occurred in similar proportions between phases.
6.4 Primary Caregiver’s Confidence in Events Being Completed by PwAs
Hypothesis: I anticipated that caregivers would have greater confidence in the ability of PwAs to complete events during Intervention phases as compared to Baseline phases.
Analysis: The electronic questionnaire (see Appendix 10.9) asked primary caregivers to rate a statement about their confidence in PwAs being able to do events listed in their schedules. (“I
3 C4 did not fill out the electronic questionnaire during B2. The missing data resulted in the Family 4 not being included in the overall statistical analysis, so N=3 instead of N=4.
103 have confidence my family member was able to do this.”) A numerical score was assigned to each response (see Table 6.8). Response Assigned Score Strong agree 5 Agree 4 Neither 3 Disagree 2 Strong disagree 1 Don’t know Excluded
Table 6.8 : Responses to the confidence question and the scores that I assigned. Responses of “Don’t know” were excluded from the data.
Figure 6.4 : Mean confidence ratings by calendar condition. A higher score here indicates higher confidence of the primary caregiver. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.
In all phases, caregivers tended to agree to the statement that they had confidence that their PwA was able to complete the event (see Figure 6.4 and Table 6.9).
104
Mean Standard N4 Deviation B1 4.1633 .29670 3 I1 4.1833 .33620 3 B2 4.22 .387 3 I2 4.2100 .39887 3
Table 6.9 : Descriptive statistics for confidence response by phase. These values are used for the statistical test.
A Friedman test was conducted but was not significant χ2(3, N=3)=1.000, p= 0.801, so H0 was accepted and HA was rejected.
Each family was further analyzed separately using independent samples t-tests and a Bonferroni correction (i.e., p<0.013 instead of p<0.05). There were no significant differences.
The results above were confirmed by interviews with primary caregivers.
“I think there’s a slight edge towards the one you put together, the Family-Link one. A slight edge… Well, the Palm Calendar doesn’t give you a spot to tick off things on the calendar when they’re done. So you’re assuming they’re done just because they went off. Some people mark them, I think using personal and different categories to show whether they’re done. [A3] doesn’t, so whereas the (Family-Link) one, you’d tick it off if you did it. But, I mean, like I said, slight edge because perhaps you’re ticking it off just so that it doesn’t keep reminding you, ‘I know, okay I’ll do it.’ So slight edge.” (C3, Follow-up)
Summary: Across the phases, primary caregivers agreed to the statement that they were confident that PwAs were able to complete events. There was no significant difference between confidence response means across phases. Qualitative evidence did not favour one system over the other.
4 C4 did not fill out the electronic questionnaire during B2. The missing data resulted in the Family 4 not being included in the overall statistical analysis, so N=3 instead of N=4.
105
6.5 Primary Caregiver’s Effort
Hypothesis: I anticipated that caregivers would require less effort to coordinate family activities during Intervention phases as compared to Baseline phases.
Analysis: The electronic questionnaire (see Appendix 10.9) asked primary caregivers to rate a statement about whether effort was required of them in events listed in the PwAs schedules. (“This required effort from me.”) A numerical score was assigned to each response (see Table 6.10). Response Assigned Score Strong agree 5 Agree 4 Neither 3 Disagree 2 Strong disagree 1 Don’t know Excluded
Table 6.10 : Responses to the effort question and the scores that I assigned. Responses of “Don’t know” were excluded from the data. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.
Figure 6.5 : A lower score here indicates that less effort was required by the caregiver.
106
In both conditions, the mean scores suggested that caregivers were neutral in their response, with a slight tendency to disagree that events required effort from them (see Figure 6.5 and Table 6.11). N5 Mean Standard Minimum Maximum Deviation B1 3 2.8033 .55474 2.19 3.27 I2 3 2.6933 .85407 1.71 3.25 B2 3 2.1867 .74568 1.33 2.69 I2 3 2.4800 .61798 1.78 2.95
Table 6.11 : Descriptive statistics for effort response by phase. These values are used for the statistical test.
A Friedman test was conducted but was not significant χ2(3, N=3)=7.000, p= 0.072, so H0 was accepted and HA was rejected.
Each family was further analyzed separately using independent samples t-tests and a Bonferroni correction (i.e., p<0.013 instead of p<0.05). This analysis yielded the following results:
• The primary caregiver of Family 3 had more effort in I1 (3.25) than B2 (2.54), one-sample t(59.634)=4.817, p<0.001.
• The primary caregiver of Family 4 had less effort in I1 (2.00) than B2 (2.53), one-sample t(88.000)=-5.579, p<0.001.
Caregivers in Family 4 reported that Family-Link reduced the effort required to coordinate on a broader scale. When caregivers were away from home, they wanted to know how their PwA were doing during the day. To do this, they would make phone calls to their PwA get this information. C4 reported making fewer phone calls to check on A4.
“It did prevent me from having to call home a lot, yeah it did, and that to me was a huge relief. Because having to work all day and then having to call him to make sure I knew what stage he was at. I think it also bothers him… that I’ve had to call home as much as I
5 C4 did not fill out the electronic questionnaire during B2. The missing data resulted in the Family 4 not being included in the overall statistical analysis, so N=3 instead of N=4.
107
have, whether it’s more because I’m checking to make sure he’s okay or whether it’s a comfort for myself I’m not really sure. I think maybe it’s fifty-fifty (percent).” (C4, I1)
C4 also reported that coordinating secondary caregivers was easier with Family-Link, which translated into fewer phone calls to secondary caregivers.
“To a certain degree, the Family-Link has provided us with the ability to see what were all doing at the same time… So it’s actually given us more time because it’s prevented us from having to sit down and make a phone call and then make another phone call and have to get back to me and now I’m in the middle of cooking dinner… You know, it’s prevented the proverbial ‘You know what? I can get back to her’ because right now I have to cook dinner because that’s more important. Or I could get back to them because right now I have to go walk the dog. I’ll wait till I get back and then I’ll call them but then they’re in bed, so now I’ve got to wait till tomorrow. My time constraints were narrowed down to the point of something taking half an hour as opposed to ‘Oh, I’ll have to get back to them tomorrow. I don’t have the time to go over this with you because that’s not as important right now as getting dinner ready and bathing the baby or having dinner and walking the dog right now. I’ll call mum tomorrow.’ It’s enabled us to visually determine where things are, answer the questions or ask the questions without having to worry whether or not they’re available for that or I’m available for this. Definitely provided more time, which has alleviated a lot of stress because if I had to wait until I call her tomorrow but I really needed answer today. Well if I needed the answer today, the first thing I do is get on the phone and call her and say ‘Listen, I know you guys are really busy but I need to know the answer now.’ Whereas if I’m looking at the Palm, Family-Link program, I could say oh look she’s not (busy)… so next Tuesday is ok for you guys? Yup, ok great thanks. So it’s definitely helped out in that manner.” (C4, I2)
The above quote hints at the issue of awareness of other family members’ schedules. I will return to this in Section 6.9.
Interview data from D4 reinforced C4’s claim about reduced effort in coordination.
“Overall, being able to sync when you’re done is definitely an asset. You want something or need something done for the next day you sync and it’s sent…
108
Instead of having that long conversation of what needs to be done, you could... just sync.” (D4, I1)
Summary: Overall, primary caregivers were mostly neutral in their assessment of whether events required effort from them. Anecdotally, Family 4 experienced less effort in the Intervention phases as compared to Baseline phases. This latter result was attributed to reduced need to make repeated phone calls to monitor the PwA. As well, a reduced need to coordinate activities was reported in Family 4. I will return to this in Section 6.9 when discussing awareness.
6.6 Primary Caregiver’s Frustration
Hypothesis: I anticipated that caregivers would feel less frustration during Intervention phases as compared to Baseline phases.
Analysis: The electronic questionnaire (see Appendix 10.9) asked primary caregivers to rate a statement about whether forgetting of events listed in the care recipients’ schedules frustrated them. (“This led to frustration for me.”) This was an attempt to measure whether problems or forgetfulness of an event might lead to frustration in the caregiver. A numerical score was assigned to each response (see Table 6.12).
Response Assigned Score Strong agree 5 Agree 4 Neither 3 Disagree 2 Strong disagree 1 Don’t know Excluded
Table 6.12 : Responses to the frustration question and the scores that I assigned. Responses of “Don’t know” were excluded from the data.
109
Figure 6.6 : A lower score here indicates lower frustration. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.
In all phases, caregivers tended to disagree with the statement that events caused them frustration (see Figure 6.6 and Table 6.13).
N6 Mean Standard Minimum Maximum Deviation B1 3 2.2367 .38214 1.88 2.64 I1 3 2.1900 .83265 1.43 3.08 B2 3 1.7633 .67678 1.00 2.29 I2 3 2.0000 .40596 1.56 2.36
Table 6.13 : Descriptive statistics for frustration response by phase. These values are used for the statistical test.
6 C4 did not fill out the electronic questionnaire during B2. The missing data resulted in the Family 4 not being included in the overall statistical analysis, so N=3 instead of N=4.
110
A Friedman test was conducted but was not significant χ2(3, N=3)=6.600, p= 0.086, so H0 was accepted and HA was rejected.
Each family was further analyzed separately using independent samples t-tests and a Bonferroni correction (i.e., p<0.013 instead of p<0.05). This analysis yielded the following results:
• The primary caregiver of Family 3 had more frustration in I1 (3.08) than B1 (2.64), one- sample t(106.646)=4.308, p<0.001.
• The primary caregiver of Family 3 had more frustration in I1 (3.08) than B2 (2.29), one- sample t(75.722)=7.709, p<0.001.
• The primary caregiver of Family 3 had less frustration in I2 (2.36) than B1 (2.64), one- sample t(109.512)=-2.802, p=0.006.
C3 experienced frustration at times during the study. Family 3 experienced system crashes because the repeat functionality for the Family-Link was not originally designed for creating daily appointments for an entire year. This increased the load on the synchronization process beyond what the system could handle on a slow network. It is possible to design a system that handles this issue gracefully, but I had not done so up front as I did not anticipate that families would add so many repeated events. When asked about what C3 found stressful, she said,
“At the beginning when we were crashing and locking and trying to sync and all of that kind of thing. And there were other stresses that was pretty stressful... I’m very conscious; I do not want to lose things because that timeline (a memo on his device) is important to him and... if we crash and lose things like that for me to put all of these things together. Sometimes they ask you when you go to appointments, you know, when you had this done and that, and I can’t remember his stuff and my stuff and their stuff you know, so... Just to be conscious of where and when and that kind of thing and so it’s important for me. I have paperwork scattered all over and I try to keep it up. And it’s hard to keep it up. But don’t to lose that (Palm) because it helps keep things up. So I found that part stressful. (C3, B1)
C3 also reported frustration with the repeat functionality of the Family-Link, as the following quote explain in detail.
111
“Until those repeat functions change in the way it gets used I found that it was hard to use it in the way that A3 has been using it. Generally that was frustrating because you can’t undo what he’s doing and so then he just stops doing certain things and I end up keeping track of things again. That really put a cork in how he was (doing) things… If there was a job he wanted to do, if he noticed something needed doing around the house, before he would take that entry and put in ‘put the fungicide on the garden’ for example. And he’d put it on everyday for the week. And when he’d done it, he’d delete it. He’d delete from the day he did it forward. While he couldn’t do it because he couldn’t do it that way. He talked about it and I’d be the one to say ‘Oh, you wanted to put that fungicide on’. So what he started to develop in some sort system and independence of his own, he stopped doing again. Because he couldn’t use it the way he had been. He couldn’t find a way around it.” (C3, I2)
While these issues remained, frustration during the I2 phase was less than it was during the B1 phase, possibly a result of the increased stability of the program after the crashes were fixed.
While there was no significant result in the analysis of Family 4, C4 explained that system speed in processing data, delays in data synchronization and notification overload were issues with the system. Some of this might have contributed to frustration. However, C4 felt that Family-Link relieved stress, which was a term that she often interchanged with frustration throughout interviews.
“It has relieved a lot of stress, personally, right, because I can track him at work. Where other people not working fulltime possibly would find it more ... would probably have more difficulties with it than I did. I personally found it beneficial as I said, I can track him from work. I don’t have to call him or I don’t have to be with him.” (C4, I1)
This is opposite of what was reported during the Baseline phases.
“I can’t track [A4]. I don’t know where he is or what’s done or, you know, unless I pick up the phone and call him.” (C4, B2)
Summary: Primary caregivers tended to disagree with the statement that events caused frustration for them. There was no significant difference between frustration response means across phases.
112
Qualitative data indicated that the Family-Link caused frustration due to the PwA 3 having difficulty with the way the repeat function worked. Caregivers of Family 4 reported issues with synchronization performance and having too many notifications. In contrast, C4 reported that Family-Link relieved stress by enabling her to track A4 from work.
6.7 Caregiver Burden
Hypothesis: I anticipated that caregivers would feel less caregiver burden during Intervention than baseline phases.
Analysis: I used the Zarit Burden Interview (see Appendix 10.17) to assess caregiver burden. It was administered to all caregivers at the end of each phase and was scored according to standard published procedure (see Table 6.14).
Total Score Interpretation 0 to 20 Little or no burden 21 to 40 Mild to moderate burden 41 to 60 Moderate to severe burden 61 to 88 Severe burden Table 6.14 : The interpretations for the various Zarit Burden score ranges.
Figure 6.7 : This bar graph shows caregiver burden scores across calendar condition.
113
Mean caregiver burden was in the mild to moderate range for all phases (see Figure 6.7 and Table 6.15). Mean Standard N Deviation B1 25.13 15.496 8 I1 24.88 15.986 8 B2 23.50 14.794 8 I2 25.00 16.186 8
Table 6.15 : Descriptive statistics for caregiver burden score by phase.
A repeated measures analysis of variance (RM-ANOVA) was carried out to determine if there was an effect between caregiver burden scores and phase. The dependent variable was the caregiver burden scores and the within-subjects variable was the phase. The assumption of sphericity for RM-ANOVA was met. Mauchly's Test of Sphericity was not significant (χ2(5)=4.887, p=0.436). The results showed that there was no significant effect between the number of shared events and the phase, F(3, 21)=0.337, p=0.779. Therefore, H0 was accepted and HA was rejected.
Qualitative data on caregiver burden was mixed. C4 felt that Family-Link had a positive effect on caregiving burden.
“With me working full time I have my own pressures and stress and what they expect of me. For me to constantly be on telephone to know where he was or what was being done or what is being done next… I found that the benefits (of Family-Link) provided me the free time to know that [A4]’s with mom – she’s taking him home and then he walks the dog and takes dinner ready. If I didn’t know by looking at my Palm, I would have to call him, ’Have you eaten?’ ‘I don’t know.’ He could complete (the event in Family-Link). (It) alleviates stress on (my) own personal mind.” (C4, B2)
The concept of awareness is explored in more detail in Section 6.9.
However, C3 felt that Family-Link increased caregiver burden. She noted that A3 had struggled with secondary functionality in Family-Link, in particular, the repeat functionality which behaved differently from the repeat functionality of the Palm Calendar. This resulted in C3
114 needing to play a greater role in helping A3 enter repeated events into the system. This issue could be addressed in the next version of Family-Link by changing its repeat feature to be more similar in functionality to the Palm Calendar.
“There was an increase in stress in that things that were going wrong at the beginning (of I1) and then you know the function that I had to keep explaining to him that you can’t repeat that so many times. ‘Oh I finished this I want to take it off’ ‘Well, it’s a repeated thing [A3], now you have to take it off everyday for…’ ‘Oh ok’ So I felt like I was still babysitting the issues that were left unresolved everyday. And that was stressful. Because I don’t what him to feel like an **** because then he stops asking if he can see that, because he feels like an idiot. So that puts stress on everybody. Do you tell him? You know what you mean? [A3] doesn’t say ‘I don’t know’ very well. Not in his nature. Never was. So he tends to just try and muddle along if he doesn’t know. Or fake it.” (C3, I2)
I will revisit this issue in Section 7.5.
Figure 6.8: Burden scores by calendar condition for primary and secondary caregivers.
A two-factor repeated measures analysis of variance (RM-ANOVA) found a between-subjects effect in which the variable caregiver type has an effect on caregiver burden scores, F(1, 6)=19.334, p=0.005. Figure 6.8 shows the difference in Zarit Burden scores between primary and secondary caregivers.
115
Figure 6.9 : Caregiver burden scores for primary and secondary caregivers. Horizontal reference lines show different burden levels.
Figure 6.9 shows that E3, C3, C5, and D3 had scores that tended to fall in the mild to moderate burden range. D4, C4, and E4 had little or no burden during the study. C6 experienced the most caregiver burden (moderate to severe burden). When asked, C6 commented,
“I would not call it a burden. The word might be a little strong… Yes, a little bit of stress, a little bit annoyed, but it wasn’t burden. It was just frustration. Just some frustration in learning the new thing because it was different.” (C6, Follow- up)
116
Figure 6.10 : Bar graph of caregiver burden scores by family.
A two-factor repeated measures analysis of variance (RM-ANOVA) found a between-subjects effect in which the variable family has an effect on caregiver burden scores, F(1, 4)=99.633, p=0.001. Figure 6.10 shows differences in caregiver burden scores between families. Family 3 had a mean score of 29.3 (mild to moderate burden), Family 4 had a mean score of 11.2 (little or no burden), Family 5 had a mean score of 22.3 (mild to moderate burden), and Family 6 had a mean score of 53.3 (moderate to severe burden). The difference in caregiver burden between families is an interesting factor that we explore further in Section 7.5.
Summary: There was no significant difference in caregiver burden scores across phases. Qualitative analysis revealed that the intervention had a slight positive effect on caregiving burden in that Family-Link allowed for greater awareness. Primary caregivers had greater mean burden scores than secondary caregivers.
6.8 Sharing
Hypothesis: I anticipated that participants would share more events in their calendar during Intervention as compared to Baseline.
117
Analysis: I first examined the quantitative data to measure how many events were shared between participants, and then explored the qualitative data to learn more about how events were added.
6.8.1 Counting Shared Events
In this analysis, I was interested in counting the number of shared events in the family of the PwA. To quantify the number of shared events recorded in the calendars during each phase, I review the concept of an event and then define the concept of a shared event .
As I defined before (see Section 4.2.1), an event is a calendar entry that has a textual title, date, start time and end time. In some cases, events can have no start or end times, denoted by NoTime in both the Palm Calendar and Family-Link. Events in a calendar represent either real-life events in the calendar owner’s life, or things around which the owner can coordinate.
A shared event is an event that appears in more than one calendar of the same type (either Palm Calendar, or Family-Link). This represents a real-life event that involves the PwAs family. The entry must appear on the same day. For example, Bob and Jill could have the shared event, “Family dinner”, at 6pm on a particular day. This would suggest that both persons were planning to participate in the same event. For the analysis below, I am concerned with shared events within each family, so events that might be shared between Families 3 to 6 were not counted (such as a Baycrest meeting).
Identifying shared events can be challenging in practice. For example, two people might refer to the real-life event in different ways. This is possible given that the data entry occurs across multiple devices, possibly at different times and/or places. In the previous example, one person might type “Dinner with Bob” while another might type “Dinner with Jill”, while both intend to refer to the same real-life event. As well, the event might show up on Bob’s calendar at 5:30pm while it shows up on Jill’s calendar at 6pm. To account for these variances, I created a set of heuristics to help determine whether two events, appearing on different calendars, are shared. Two events are shared if the following conditions are met:
• Both events occur on the same calendar day.
• The titles of both events are similar in that the same words, themes, and/or pronouns are used
118
o OR names are used in reciprocal fashion as in the example above “Dinner with Bob” in Jill’s calendar and “Dinner with Jill” in Bob’s calendar).
• There is an overlap in time, o OR one or more of the events is marked with NoTime (an indication that the event is not a timed event), o OR the time differences between the events is within two hours. This allows for some flexibility in case participants write the wrong time. o OR the times do not overlap but the titles are similar and one of the titles mentions that the event is for the better part of the day
There are a couple of heuristics to add. Firstly, spelling errors are ignored in the event titles. This allows for some flexibility but also potentially increase the number of false positives. Secondly, events can be shared three-ways such that the calendar entry appears on three calendars while satisfying the above conditions.
To illustrate the heuristics, Table 6.16 shows examples of shared events occurring in the study. C4’s calendar: 10-12p [D4] and [A4] going to Stars and Strollers Movie E4’s calendar: 5:30am-1:30pm [D4] with [A4]
A6’s calendar: NoTime [Name] M.R.I. C6’s calendar: 10-10:30am MRI [Name]
A6’s calendar: 10-12pm Danforth Ave: Pape to Broadview shoe stores C6’s calendar: 7-12pm Look for [A6] sandal
A6’s calendar: 7:30-9:30am Pick-up sandles BioPed C6’s calendar: 5-3pm Bioped insoles [A6]
A4’s calendar: 5:00-5:00pm going to [A3] and [C3] bring chips… Told [C6] 3 times. If I forget chips its now her fault. I love you Bunny. C4’s calendar: 5-8pm [C3] and [A3] to play cards
A4’s calendar: 5-6am Going over to [D4]s for the day D4’s calendar: 8-9am [A4] coming over E4’s calendar: 5-6am [A4] coming over
Table 6.16 : Examples of shared events occurring during the study.
119
There was an exception to these heuristics, stemming from the fact that there are two families in Family 4. D4 and E4 (wife and husband) are a separate family unit and have events in their calendars that are exclusively shared between themselves but not with A4 or C4. For example, “Ready myself for Walmart with E4”, “Tuxedos at Sid Silver”, and “[Friend] over”. Such events were not explicitly shared with A4 or C4, and so were not included in the final counts. As A4 mentioned in the following quotes at the end of I2:
“I’d say there’s not a need to know, unless it’s something I’m involved with, with them. I know [D4] and [E4] used the scheduler quite a bit. I have no reason to see any of that. I mean a lot of [D4’s] (schedule) is feed (the baby) or take (the baby) to doctor. Their doctor’s appointments. I really… don’t need to know that.” (A4, I2)
“Even though [D4] and [E4] are immediate family, they lead their own lives. It’s not important for me to know what they’re doing all the time.” (A4, I2)
To carry out the analysis, there are a number of events that I excluded. I counted calendar events for each Baseline and Intervention phase, but not events falling on the boundaries of phases (i.e., days during which I met with participants at the end of each phase). This is because on those days, participants could have used both calendars, which may be a confounding factor.
I also only counted events occurring a week after each phase started. This had two benefits. Firstly, this slightly eased problems associated with event conversions between I2 and B2 (see Section 7.4.1.2), where shared events created by Family-Link were automatically added to multiple devices for B2. This would inflate the count of shared events in B2 without participant effort. However, I am trying to measure whether family members go through the effort of making shared events in the first place. Secondly, this could help mitigate learning effects. For example, when using Family-Link for the first time, I found that family members tended to create shared events for the sole reason to try out the procedures and new software. I did not include these events into the counts as they did not represent real-life events.
Since the duration of phases was not uniform across and within families and greater time spent with an aid could affect sharing behaviour, I counted events until the last day of the third week. This was chosen because the shortest phase in all families was three weeks.
120
I applied the heuristics to manually compare all events occurring on each participant’s calendar with events of their family members. The following observations were made:
• Events would sometimes show up on one calendar but not others, though it seemed that the event was an actual shared event. For example, in D4’s calendar was the event “A4 taking me to (Nursery) to apply for a Sunday position” but this did not appear in A4’s calendar. This was not counted as a shared event because it did not appear in more than one calendar.
• A few entries were reminders to send reminders to the PwA, however these were not shared events as they only occurred on one person’s calendar.
• In Baseline phases, shared events on multiple calendars almost always had mismatching start and end times. This can lead to serious coordination issues in the family.
• There were obvious errors in some calendars during Baseline phases. For example, in E3’s calendar, “Dad to Baycrest” was listed every two weeks from 7-10a. However, E3 put the repeating event on the wrong alternating schedule. This was a failed attempt to replicate an event of A3’s in her calendar. Another example was that A3 listed that C3 had work (“C3 work C” from 10-12:30p) in his calendar every Wednesday, but this was not removed for Canada Day, a statutory holiday falling on a Wednesday during the study.
121
B1 I1 B2 I2 % % % % # Events # Events # Events # Events # Shared # Shared # Shared # Shared A3 2 199 1.0 14 165 8.5 2 186 1.1 21 189 11.1 C3 0 16 0.0 12 23 52.2 1 9 11.1 13 13 100.0 D3 2 16 12.5 0 5 0.0 1 8 12.5 0 0 N/A E3 0 6 0.0 0 0 N/A 0 3 0.0 0 0 N/A A4 20 129 15.5 16 123 13.0 2 95 2.1 17 75 22.7 C4 16 55 29.1 14 39 35.9 4 39 10.3 13 25 52.0 D4 4 88 4.5 12 29 41.4 2 97 2.1 11 115 9.6 E4 3 40 7.5 14 71 19.7 2 2 100.0 8 9 88.9 A5 1 51 2.0 27 40 67.5 4 58 6.9 35 56 62.5 C5 0 16 0.0 28 37 75.7 4 5 80.0 35 40 87.5 A6 4 32 12.5 13 24 54.2 2 19 10.5 5 14 35.7 C6 1 13 7.7 13 53 24.5 2 38 5.3 5 29 17.2 Mean 4.4 55.1 7.7 13.6 50.8 35.7 2.2 46.6 20.1 13.6 47.1 48.7
Table 6.17 : Proportion of events shared for every participant in each phase of the study. “N/A” indicates that there was insufficient data.
The counts and proportions of shared events are shown in Table 6.17, while mean counts are graphed in Figure 6.11, and descriptive statistics are shown in Table 6.18.
122
Figure 6.11 : Mean shared events counts for each phase of the study.
Mean Standard N Deviation B1 4.42 6.557 12 I1 13.58 8.361 12 B2 2.17 1.267 12 I2 13.58 11.828 12
Table 6.18 : Descriptive statistics for shared event counts by phase.
A repeated measures analysis of variance (RM-ANOVA) was carried out to determine if there was an effect between shared event counts and phase. The dependent variable was the number of shared events observed and the within-subjects variable was the phase. The assumption of sphericity for RM-ANOVA was not met. Mauchly's Test of Sphericity was significant (χ2(5)=18.76, p<0.05); therefore, the degrees of freedom was corrected using the Greenhouse- Geisser estimates of sphericity ( ε=0.478). The results showed that there was a significant effect between the number of shared events and the phase, F(1.43, 15.77)=9.78, p=0.003. Therefore,
HA was accepted and H0 was rejected.
123
While a significant effect existed, this statistical test does not indicate which phases differed from each other. To explore this further, four paired t-tests were applied (comparing the means from I1 with B1, I1 with B2, I2 with B1, and I2 with B2). The following hypotheses were used:
• H1: There is a significant difference between the mean count of shared events in I1 and the means count of shared events in B1.
• H2: There is a significant difference between the mean count of shared events in I1 and the means count of shared events in B2.
• H3: There is a significant difference between the mean count of shared events in I2 and the means count of shared events in B1.
• H4: There is a significant difference between the mean count of shared events in I2 and the means count of shared events in B2.
A Bonferroni correction was used to address the issues resulting in making multiple comparisons. A statistical level of 0.013 (i.e. 0.05/4) was used to determine significance (i.e., p<0.013 instead of p<0.05). The results were:
• the mean of I1 (13.58) was significantly different from the mean of B1 (4.42), paired t(11)=3.074, p=0.011
• the mean of I1 (13.58) was significantly different from the mean of B2 (2.17) paired t(11)=5.405, p<0.001
• the mean of I2 (13.58) was not significantly different from the mean of B1 (4.42) but the statistic approaches significance, paired t(11)=2.346, p=0.039
• the mean of I2 (13.58) was significantly different from the mean of B2 (2.17), paired t(11)=3.631, p=0.004
Therefore, H1, H2 , and H4 were accepted, while H3 was rejected.
The qualitative data also suggests that more was shared during Intervention than Baseline. C5 explains,
“I find that with the Family-Link thing I use it more. I guess just …now I find myself just using the regular (wall) calendar rather than going and get it (Treo) to put something in it. I just tell [A5] to put it into his. But somehow it seems easier to put it in and send it
124
over to him… With just the regular calendar, I just put in, you know, I put in for myself. There’s less sharing of everything with the (Palm) calendar … like with this aphasia (appointment). If I have Family-Link the aphasia (appointment) is in mine (Treo). But if I’m just using the regular calendar, I wouldn’t put his aphasia (appointment) in there. I probably wouldn’t put anything that he was doing… I guess it comes down to laziness.” (C5, B2)
When asked whether items were shared more with Family-Link or Palm Calendar and who did the sharing, C6 replied:
“We both did. It was mainly with Family-Link. I did not know (his schedule) unless he would tell me. I have no way of knowing. Well it is very (useful) because now I’m back to teaching now I know I (could) look at it and know where he is now and the text chatting I could tell him I could send him a quick note and I would know an hour later he would get it so that (would be) good.” (A6, Follow-up)
6.8.2 What Events Were Shared
Shared events included (actual examples from study are included):
• Events that needed to occur at specified times: o meetings (e.g. A5, aphasia meeting “Aphasia”), o appointments with doctors or health specialists (e.g., C6. “Dr. [Name]”; A3, “Dentist- C3”), o social outings with the family (e.g., A6, “[Name]’s parents house barbeque dinner”; D4, “[name] & [name]’s place for lunch”), o transportation arrangements (e.g., E4, baby’s doctor’s appointment, “([baby]) have a dr appt-grandpa taking me”; A4, “Pick up [D4]”),
• Events that represented general activities or tasks that could be postponed if needed: o family activities (e.g., C4, “Bath Puppy – with [A4]”), o shopping tasks (e.g., A5, “Go to rona to look at door handle”),
• Events that provide awareness of birthdays or holidays: o birthdays (e.g., A4, “[Dog] birthday”) and holidays (e.g., C5, “Canada day”), and
125
o events that indicate the whereabouts of someone or their status (e.g., D4, “[E4] leaves for Calgary”; A4, “[C4] on Holidays”; A3, “[C3] work [community centre]”; C3, “Spouse Support Group”).
6.8.3 How Events Were Shared
6.8.3.1 Beaming Events via Palm Calendar
During the B1 phase, C4 reported using the infrared Beam feature of the Treo to send and receive several appointments from A4’s device. This makes a local copy of the appointment in the receiver’s calendar but the devices do not synchronize the copy. For beaming to function, Both Treos need to have line-of-sight and be typically no more than a few metres apart. This sharing behaviour was not reported by other participants in any other phase, and was not a behaviour that was trained by clinicians at Memory-Link.
6.8.3.2 Creating and Editing Events in PwAs’ Calendars via Family-Link
Family-Link enabled caregivers to create events in the PwAs calendar. C6 made use of this feature, as the following quotes illustrate.
“I liked it. I liked that we could communicate. That I think that was probably the best thing. And that was the things like if I wanted to tell [A5] to remember something, rather than have to depend on him to put it in. I could just put it in and send it to him, which I liked that.” (C5, I1)
”I find even that I miss it now, and I didn’t think I would, honestly (laughs)… Just being able to… put things in and put them over to [A5]. I like that rather than telling him you know ‘put this in’ or ‘put that in, for me’.” (C5, I2)
However, members of Family 3, 4, and 5 rarely did this. Whenever C3 wanted to create a shared event that was the primary responsibility of the PwA, she left a paper note for him and had him enter the event himself.
“I would leave a note for him. If something comes up, leave it on a note. It’s quicker for me to do a lot of things but I can’t live that way. I can’t do that. Because eventually, what is he doing? Then he has no independence. I think it would take 5 steps backwards if I
126
started doing that because he may not remember putting it in. But while he’s doing it he knows he’s doing it. So he has a sense of independence. Sense of self-respect.” (C3, I1)
Family-Link enabled users to modify existing events in another family member’s calendar. C3 mentioned that A3 did not realize that there were changes in his calendar even though the ‘Family News’ feature highlighted the changes. “He doesn’t remember who puts the changes.” (C3, I2) To address this, C3 encouraged A3 to take charge of modifying events in which he was involved.
“If I were to ever change anything, which I have once in a while, I tend to make him in charge of it.” (C3, I2)
“I wouldn’t edit it without chatting to him, not that I couldn’t and he may think and not remember someone else put it in there. I just don’t think it’s my place. I don’t think I should be doing his calendar. It’s his responsibility. If I ever put anything in his calendar, it’s in his presence. Because he should be a part of it because it’s his life. So I wouldn’t put things in. It’s just me. Does that make sense? I like to be able to have access to his calendar so I can see… If I ever put anything in his calendar, it’s in his presence.” (C3, I1)
C4 also spoke about the issue of editing events in A4’s calendar.
“I’ve always had a problem with touching [A4]’s Palm because it’s their calendar. For me to alter his calendar, I found it was difficult .. now I’m messing with his stuff. I find that difficult to do for him. It’s adding one more confusing thing on top. Let him edit the info himself. To do it to him is a confusing situation. Point in case: [D4] put in, ‘put in doctor’s appointment’. He didn’t remember putting the information in there, whereas if he hears it once he puts it in, he goes back and can see it. Validation gets stronger each time as opposed to someone else doing it.” (C4, I2)
To overcome this issue, Family 4 developed a coordination routine where one person would take care of adding the entry into Family-Link after discussion.
“The good thing about us as a family is that we will discuss the situation and then one of us would say I’ll add it, or I’ll do it. And then we were all informed and we took care of
127
things. We did Father’s day and Mother’s day through our Palm Family-Link program which really worked out well. The boys doing their planning and then us doing our planning, which was nice. It was nice for [E4] and [A4] to plan, leaving us out for it. And then it was nice for us to plan, leaving them out of it.” (C4, I1)
It should be noted that A3, A4, A5, and A6 reported having no issues with their family members adding appointments into their calendars. Perhaps this is in part because Family-Link highlighted new appointments that had been put in by somebody else.
“I don’t think it would bother me. It must be important if they’re going to put it in there. I think it’s okay.” (A3, I2)
“She (C4) will add things to my calendar. Not so much changing things but add stuff. If there’s something I have to do, or forgotten to put in to my scheduler, she’ll add it in. And that’s where it’s nice for her and I to be in touch with the Family-Link program. I will forget that I’m supposed to do something, so she’ll put it into my calendar for me. (A4, I2)
6.8.3.3 Creating Events in Other Caregivers’ Calendar via Family-Link
Although caregivers wanted PwAs to input and edit events themselves, caregivers liked the flexibility in being able to add appointments to other caregivers’ calendars. This was mostly utilized in Family 4. C4 explains that the caregivers support one another’s memory this way:
“She (D4) will add things to my calendar. Not so much changing things. If something I have to do, or forgotten to put in, she’ll add it. And that’s where it’s nice Family-Link program. I’ll forget and she’ll put it in my calendar for me.” (C4, I2)
6.8.3.4 Titles of Shared Events
Members of Families 1 and 4 noted in the interviews that more thought was needed in coming up with appropriate titles in shared events during Intervention. For example, A4’s “Doctor’s appointment” might be shared with D4 because she was providing transportation for A4. However, D4 might see it on her device and be confused that the appointment is hers.
128
Summary: Sharing functionality was a key component of Family-Link that participants liked and used. I found quantitative and qualitative evidence that participants shared more events during Intervention phases as compared to Baseline phases. There was a significant difference between mean shared event counts between phases. The mean of I1 was significantly greater than the mean of B1, and the mean of I2 was significantly greater than the mean of B2. While using Family-Link, participants shared in different ways – some caregivers wanted the PwAs to enter events themselves while other caregivers would add the appointment themselves. In both cases, caregivers often shared information with PwAs in order for the event to be created or modified. Titles for shared events sometimes led to confusion.
6.9 Awareness
Hypothesis: I anticipated that participants would feel greater awareness of other family members’ schedules during Intervention as compared to Baseline.
Analysis: I examined three aspects of awareness between Intervention and Baseline: (1) whether PwAs were aware of events in caregivers’ schedules, (2) whether primary caregivers were aware of PwAs schedules, and (3) whether caregivers were aware of other caregiver’s schedules (within the same family).
6.9.1 PwAs’ Awareness of Caregivers’ Schedules
All participants spoke about experiencing increased awareness of other family members’ schedules during Intervention phases.
A3 explained how increased awareness during Intervention provides a way for him to reach his family members if needed.
“(Family-Link) let’s me know what other people are doing: if someone goes to work, or where they are. So if I need them I can get a hold of them.” (A3, I2)
However, this was limited to whether caregivers added information into the program.
“It depends on if they (D3 and E3) put something in it I can see it. But if they don’t, I can’t see… When [D3] goes to work, or she’s going out, or [E3]’s going to a friend’s house, she puts it in, I can see. If they don’t, then I don’t know.” (A3, I1)
129
A4’s knowledge of C4’s schedule reduced panic situations for himself.
“(For) [E4] and [D4], it’s nice to know what they’re doing but not as much as [C4]. I like to know what [C4] is doing, especially when the routine is interrupted. Like she doesn’t come home at work, like if she’s got an appointment of some sort. If I’m sitting at home at 4:30, 5:00 rolls around, 5:30 rolls around and she’s not home I start to go into a panic mode. What’s wrong? Where is she? Is she been into an accident? (A4, B2)
C5 spoke about how A5 was aware of her change in transportation plans from the usual bus ride and how Family-Link provided an alternative method for informing A5.
“Last Thursday, instead of having a bus home from (his) aphasia (meeting), I came to pick him up there... He knew about it (bus) beforehand, but then you know I tried to call him on the way to tell him I was coming to get him and of course his phone was off so I sent (it) on his Palm (using Family-Link)” (C5, I1)
When asked to reflect upon how aware A6 was about C6’s schedule before Family-Link, A6 explained that C6 had different paper calendars at home to keep track of her work. He elaborates,
“Different projects had different calendars. That’s why I didn’t keep track of it. It wasn’t in one place. When I look at the hardcopy of the calendar I wouldn’t know what she doing…” (A6, I1)
6.9.2 Caregivers’ Awareness of PwAs’ Schedules
For caregivers, awareness of PwAs calendars during Intervention enabled them to know what PwAs were doing, know what events were completed, avoid scheduling conflicts and reduce the necessity of making constant phone calls to check up on the PwAs.
C1 mentioned that she used the devices when she was traveling (but not when at home).
“I would probably check it maybe twice a day. I would check it in between flights and again when I got back to the hotel. The first thing I did when I got to hotel was call [A1]. Sometimes I would look at it before [calling him]. Sometimes I would forget about it and call him right away. If the flight was delayed, I would just call and make him not be so concerned.” (C1, I1b)
130
From a medical standpoint, C1 also benefited from greater awareness. C1 also remarked how the lack of events being marked as completed could alert her to whether A1 may be feeling ill or if he experienced a seizure.
“It kept me informed on what he was able to do. It was able to inform me of if he’s had a good day. If he hasn’t done everything I know he had an off day. He’s usually productive. If he hasn’t done anything I know something’s up.” (C1, I1b)
C3 explained that she looked through A3’s calendar to find conflicts.
“I liked to be able to have access to his calendar to see if there was any conflicts. Not to see if he did things because we had those verbal discussions usually when I’m in and out anyways. (C3, I1)
C4 reported a number of benefits of increased awareness provided by the Family-Link program.
“The Family-Link program was extremely beneficial for us ‘cause we found it let us know where each of us were, or kept in contact and full communication, where sometimes we don’t have the option with the actual (Palm Calendar) program.” (C4, I1)
C4 mentioned how being more aware led to less effort required.
“I really enjoyed being able to be more aware of what was going on on a daily basis without having to pick up a telephone every 20 minutes because (I’m) able to communicate with (the) Palm.” (I2)
C4 provided some examples of how Family-Link was used to increase her awareness of A4’s whereabouts and schedule, and how it was as if A4 was in the presence of a secondary caregiver.
“’Cause I’m at work all day so I couldn’t constantly be looking at it. Weekends a lot more, to see where [D4] was or see where [A4] was. Whereas if I were at home and he (A4) had to do something, I found that came in very handy as well. Cause I didn’t have to get up and write a piece of paper and say, by the way I’m heading out. And leave him asleep. Because I try leave him asleep on Saturdays and I would head out with my girlfriend. So then he’d be able to see (me) going ‘Garage sale-ing back at 12’ and then he could wait until 12 or call me on cell phone whichever he preferred at least he was
131
aware of that. And I was also able to get him to do things, such as can you make sure the cat and dog gets fed while I’m gone. That was really quite beneficial. And then if he would disappeared and went out somewhere I knew that he was out on his walk. It was almost like having [D4] or mom being there on full time basis without [D4] or mom being there.” (C4, I1)
C6 spoke about how sharing appointments led to greater awareness during Intervention and an increased sense of security.
“Maybe twice a week (that A6 is outside the home). So that’s why I like the share, the Family-Link. I can always see a chat or whatever, like ‘Where are you? What’s happening, you left at 1 o’clock, It’s 5 o’clock and you’re not back, so where are you?” because I’m always worried. He could cross the street and be hit by a car and I wouldn’t know, so… Yeah, (it gives) me a sense of security to know.” (C6, I2)
“He basically gave me… doctor’s appointments and dentist, stuff like this. And he would share that with me. Even though that’s personal at least I know where he’s going, because he might forget to tell me. But if it’s there and shared then I say ‘Oh, ok’. That’s why I like that. But I think that… we’re dealing with people that have memory problem. For us caregivers, we are always worried, so. And most of us cannot afford a cell phone, so that (Family-Link) is very good. This is my take, anyway, on that.” (C6, I2)
Primary caregivers were not as aware of PwAs’ events during Baseline phases. This was a result of calendar information not being accessible unless caregivers physically accessed the PwAs’ Palm devices. This was seldom done because caregivers viewed the Palm device as a personal and private tool for the PwA. C6 spoke about how she would not be aware of changes during Baseline unless by chance she heard the alarm on his device and was reminded that way.
“Well I don’t know unless I’m here and I hear it. But If I’m not here, I mean if he forgets to tell me, then I don’t know about it.” When asked if she would check his device to see his schedule, she replied, “No, nope. The Zire, I don’t really know how it works so no I don’t check his device. It’s a kinda personal tool at one point so I don’t touch his stuff but but he would be okay if I looked. But I do now (laughs). Well, if I hear a beep because he sets up his alarm every morning to do stretching. Well if he’s in the shower and doesn’t
132
hear the alarm, I’ll go and turn it off. But I’m not going to look to search for his appointment unless he shares it with me. Because to me, this is still personal.” (C6, B1)
An aspect of awareness to consider is whether or not PwAs would mind caregivers seeing their events. Qualitative evidence suggests that PwAs recognized the benefits to keeping their primary caregiver apprised of their events and were keen to allow family members access to their schedules. They were also aware that their primary caregivers were keeping an eye on them.
A3 spoke of how C3 used Family-Link to keep track of himself.
“Just to see what I’m up to, to keep an eye on me, to make sure I take pills, to walk the dog.” (A3, I1)
A4 recognized that C4 uses Family-Link to check on him and his events.
“It’s just the Family-Link program is more versatile, it does more. But just as a day scheduler the (Palm) Calendar worked for just me, (but) I can’t see what the rest of my family is doing. Well for me, more so [C4], I feel like I like to keep an eye. More so I know she checks on me. Because I can mark an event as done that’s what she looks for. She just has to look at my calendar and she can see where I am in my day. Where she looks at me and it’s 3 o’clock at the afternoon and the only thing I’ve done all morning is showered at 10 o’clock, I’m not checking things off or I’ve gone back to bed, right? She keeps an eye on me that way.” (A4, I2)
“I think she (C4) looks at mine a fair bit. To see if I’m on top of things. She knows pretty much what my schedule is for the day, so I know she’ll check mine to make sure I’ve..I’ve walked the dog and had lunch and so forth. And I think it’s more important that way. She keeps an eye on me with it… A lot of her stuff is work stuff too she uses it for. And I have no reason to see it. Not that I don’t care what she’s doing during the day, but most of the stuff doesn’t have to do with me, or us for that matter.” (A4, I2)
6.9.2.1 Primary Caregivers’ Awareness of Completeness of Events
This continues an analysis that was done in Section 6.2. I examine the same data set, but will look at it in a slightly different way. Using data collected from the electronic questionnaire (see
133
Appendix 10.9), I counted the proportion of events in which the primary caregiver gave a response other than “Don’t know” in the Completeness question (see Appendix 10.9). I used this as an estimate of their awareness of the completeness of events.
Figure 6.12 : This figure estimates the awareness of primary caregivers about event completeness in the schedules of PwAs. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.
Mean percent awareness of event completion was near 100% for all conditions (see Figure 6.12 and Table 6.19).
Mean Standard N Deviation B1 99.0000 1.73205 3 I1 98.6667 2.30940 3 B2 99.0000 1.73205 3 I2 100.0000 .00000 3
Table 6.19 : Descriptive statistics of percent awareness of completion. These values are used for the statistical test.
134
A repeated measures analysis of variance (RM-ANOVA) was carried out to determine if there was an effect between percent awareness of completion and phase. The dependent variable was the percent awareness of completion and the within-subjects variable was the phase. The assumption of sphericity was met. The RM-ANOVA results showed that there was no significant effect between the number of shared events and the phase, F(3, 6)=1.000, p=0.455. Therefore, H0 was accepted and HA was rejected.
While there was no statistically significant result, there was qualitative evidence suggesting that Family-Link increased awareness during Intervention phases. C1 explained how Family-Link increased her awareness of completed events.
“It did help in that sense, because I knew when things were actually being accomplished and done because he would confirm them. So that was the good part: to know if things were being done…Well… because I’m not here every day, I could tell what he has planned every day, right. I could tell what he has planned for the day, and then you know as the day goes on and what he’s accomplished and stuff like that. I could have kept track of it a little bit more. Whereas calendar it’s basically his tool and only when I’m home I can verify the tool or see if he’s actually doing it.” (C1, Follow-up)
6.9.2.2 Primary Caregivers’ Awareness of Timeliness of Events Completed
This continues an analysis that was done in Section 6.3. I examine the same data set, but will look at it in a slightly different way. To assess primary caregiver awareness of whether the event was done by the PwA in a timely fashion, I counted the proportion of events in which the primary caregiver gave a response other than “Don’t know” in the Timeliness question (see Appendix 10.9). I used this as an estimate measuring their awareness of the timeliness of events.
135
Figure 6.13 : This figure estimates how aware primary caregivers were about the timeliness of events in the schedules of PwAs. This graph includes data from Family 4 to illustrate general trends but the statistical test below removes Family 4 due to missing data.
Caregivers were mostly aware of the timeliness of events in the schedules of PwAs (see Figure 6.13 and Table 6.20).
Mean Standard N Deviation B1 93.7500 9.08104 3 I1 94.5455 9.44755 3 B2 87.9630 20.84876 3 I2 79.5737 14.48393 3
Table 6.20 : Descriptive statistics of percent awareness of timeliness. These values are used for the statistical test.
A repeated measures analysis of variance (RM-ANOVA) was carried out to determine if there was an effect between percent awareness of timeliness and phase. The dependent variable was the percent awareness of timeliness and the within-subjects variable was the phase. The assumption of sphericity was met. The RM-ANOVA results showed that there was no significant
136
effect between the number of shared events and the phase, F(3, 6)=2.835, p=0.128. Therefore, H0 was accepted and HA was rejected.
The high percentages suggest that primary caregivers knew about event completion and timeliness throughout the study. Interviews revealed that this was because they were in close contact with the PwA every day, speaking about event details over the telephone during the day and at the end of it, and observing cues about whether the event was completed or not (e.g. finding the letter on the kitchen table that was supposed to be mailed earlier in the day). C4 explains this:
“I have to be kept in the loop so there’s never a reason to be concerned about what’s going on and where he is at any given time.” (C4, B1)
“Although Palm program was good for (keeping) him on track, it didn’t keep me informed. So should I let go? Yah maybe (laughs) Do I want to? No, ‘cause I need to know that things are being done at a certain level. He took the dog to the vet. He was there for 9 o’clock he was finished and back home at 10. I had no idea he was at home… but now 10:30 he was supposed to be doing the laundry but that didn’t get done because he’s now overwhelmed by the doctors, you know, by with being there with the dog. It’s just one of those long weeks where you could have used the program as opposed to having to call home to find out what stage he was at.” (C4, B2)
Family-Link improved awareness in Family 4 because of the capability to mark events as completed in A4’s device and the ability to retrieve that information from C4’s Palm. A4 explains:
“Always. Always use it. You mean like, Event Done, right? And that’s also when[C4] looks at my calendar she can see if I mark something as done. That’s how she keeps tabs on me. If something’s not completed she’ll wanna know why.” (C4, I1)
6.9.3 Caregivers’ Awareness of Other Caregivers’ Schedules
C4 also mentioned how awareness was helpful for finding other family members and providing flexible prospective planning.
137
“Keeping in contact with all of us. Finding where each of us was at any given time. Or if I wanted [D4] to do anything with [E4] then we double checked her schedule first and booked her in without picking up the phone and having to contact with her. So it was nice to know we can be flexible with our time when scheduling our weeks or days. It wasn’t just a question of ok well we can’t do that because she’s busy that day. She’s busy up to 3 so she at 5 can technically do this. Or [E4]’s off at 7 so we can do this altogether at 9. So it did have its benefits with respect to that.” (C4, I1)
D4 spoke about how awareness of her parents’ location made her feel safe,
“From my perspective, it definitely helps to keep track of where the parents were which at some point, if the phone’s not answered, really makes me feel safe.” (D4, I1)
She later added that awareness also led to better coordination.
“It was quite handy in that way because they couldn’t run from me at that point, right. (laughs) I knew where they were at all times. There was a couple times where my parents had their little night out or whatnot and it leaves me, not in a horrible panic, but ‘Oh I need to talk to [A4] or mom (C4), I need some sort of conversion’. No one would answer the phone, so I would get onto their Palm and I would jump into [A4]’s calendar. Right they’re out at the puppy park or they were out for dinner. So at least I would know okay, they’re scheduled to be back at 9 so I can call at 9:30. Especially with my mom at work, cause we’re usually going back and forth with emails as well that if she doesn’t respond by email, from nine to eleven what’s going on? I will jump onto her calendar… It was very useful.” (D4, I1)
Summary: I found qualitative evidence that awareness of family schedules was greater during Intervention phases than Baseline phases. Although primary caregivers were aware of event completeness and timeliness to similar degrees between conditions, interviews revealed that this was because they remained in constant contact with PwAs to monitor that information. Qualitative data further suggested that awareness of events during Intervention phases led to a reduction in the amount of effort needed to coordinate; caregivers needed to make fewer phone calls to one another. As well, caregivers reported having a greater sense of security from being able to track their PwAs progress and whereabouts during the day.
138
6.10 Usefulness of Family-Link
Hypothesis: I anticipated that users would find features of Family-Link useful.
Analysis: I administered a Usefulness Questionnaire (see Appendix 10.24) at the end of the study to all participants. The questionnaire had a 4-point Likert scale (see Table 6.21).
Response Assigned Score Not Useful 1 Somewhat Useful 2 Useful 3 Very Useful 4 Uncertain Excluded
Table 6.21 : Responses to the usefulness question and the scores that I assigned. Responses of “Uncertain” was excluded from the data.
N7 Mean Std. Dev. Min. Max. Viewing other family members’ calendars 11 3.27 .786 2 4 Adding events to another family member’s 12 3.17 .835 2 4 calendar Seeing who created events on your calendar 10 2.80 1.135 1 4 Setting alarms for yourself 12 3.08 .793 1 4 Getting reminded until the alarm is 9 3.00 .707 2 4 cleared/snoozed Setting alarms for other people 8 2.38 .744 1 3 Having the changes in calendars 8 3.25 .463 3 4 automatically sync with other calendars Being notified of changes in events 10 3.10 .876 1 4 Attaching notes to an event 10 3.00 .943 1 4 Attaching chat messages to an event 9 2.00 1.000 1 4 Searching for events in your calendar 10 3.30 .675 2 4 Marking an event as completed 12 3.00 1.128 1 4 Seeing if an event is completed or not 12 3.00 .953 1 4 Filtering to see only uncompleted events 8 2.87 .641 2 4
Table 6.22 : Descriptive statistics for scores for questions in the Usefulness Questionnaire.
7 The N value is not 12 for all features because of excluded data (responses of “Uncertain”).
139
Table 6.22 shows the descriptive statistics for the Usefulness responses.
There are two interesting observations from this data. Firstly, “Having the changes in calendars automatically sync with other calendars” was the only feature with the greatest minimum rating (which was a 3). Along with a maximum rating of 4, this suggests that this feature was the most useful. Secondly, “Setting alarms for other people” had a minimum rating of 1 and the lowest maximum rating of all features (which was a 3). This suggests that this feature was the least useful.
To determine whether there was a statistically significant difference between the mean usefulness rating and the average rating of 2.5 for each question, a Wilcoxon Signed-Rank test was carried out for each question. The Wilcoxon Signed-Rank test indicated following Family- Link features were rated significantly higher than the average rating of 2.5:
• Viewing other family members’ calendars (z=-2.377, p=0.017)
• Adding events to another family member’s calendar (z=-2.183, p=0.029)
• Setting alarms for yourself (z=-2.321,p=0.020)
• Having the changes in calendars automatically sync with other calendars (z=-2.640, p=0.008)
• Being notified of changes in events (z=-1.996, p=0.046)
• Searching for events in your calendar (z=-2.521, p=0.012) The following questions resulted in a statistic that approached significance:
• Getting reminded until the alarm is cleared/snoozed (z=-1.811, p=0.070)
6.10.1 Comparing PwAs’ and Caregivers’ Perspective
I was interested to know whether there were differences between PwAs and caregivers in how they responded to the usefulness questions. By grouping the results (see Table 6.23), it seemed that PwAs and caregivers held differing opinions about which features they found useful.
140
nses for for PwAs for PwAsfor Caregivers Caregivers # Respo # Responses Viewing other family members’ calendars 3 2.3 8 3.7 Adding events to another family member’s calendar 4 2.8 8 3.4 Seeing who created events on your calendar 3 2.7 7 2.9 Setting alarms for yourself 4 3.5 8 2.9 Getting reminded until the alarm is cleared/snoozed 4 3.3 5 2.8 Setting alarms for other people 3 2.0 5 2.6 Having the changes in calendars automatically sync 1 3.0 7 3.3 with other calendars Being notified of changes in events 2 3.0 8 3.1 Attaching notes to an event 4 3.3 6 2.8 Attaching chat messages to an event 3 1.7 6 2.2 Searching for events in your calendar 4 3.5 6 3.2 Marking an event as completed 4 3.0 8 3.0 Seeing if an event is completed or not 4 3.0 8 3.0 Filtering to see only uncompleted events 2 3.0 6 2.8 Overall Mean 3.2 2.9 6.9 3.0
Table 6.23 : Mean usefulness scores for Family-Link features, grouped by subject type.
In considering only PwAs’ responses, a Wilcoxon Signed-Rank test was carried out for each question. The Wilcoxon Signed-Rank test found no significant difference between the ratings and the rating of 2.5, but the following features approached significance:
• Setting alarms for yourself (z=-1.857, p=0.063) with a mean rank of 3.5
• Searching for events in your calendar (z=-1.857, p=0.063) with a mean rank of 3.5
• Attaching notes to an event (z=-1.890, p=0.059) with a mean rank of 3.3
In considering only caregivers’ responses (primary and secondary caregivers combined), a Wilcoxon Signed-Rank test was carried out for each question. The Wilcoxon Signed-Rank test found that the following features were rated significantly higher than the neutral rating:
141
• Viewing other family members’ calendars 8 (z=-2.598, p=0.009) with a mean rank of 3.7
• Adding events to another family member’s calendar (z=-2.021, p=0.043) with a mean rank of 3.4
• Having the changes in calendars automatically sync with other calendars ( z=-2.460, p=0.014) with a mean rank of 3.3
It is interesting to note the features that had the greatest difference in means between PwAs and caregivers (mean of PwAs given first, followed by mean of all caregivers).
The events that PwAs found more useful than did caregivers were:
• Setting alarms for yourself (3.5, 2.9)
• Attaching notes to an event (3.3, 2.8)
• Getting reminded until the alarm is cleared/snoozed (3.3, 2.8)
The events that caregivers found more useful than did PwAs were:
• Viewing other family members’ calendars (2.3, 3.7)
• Adding events to another family member’s calendar (2.8, 3.4)
• Setting alarms for other people (2.0, 2.6)
PwAs were most concerned about features that would support their memory. Primary caregivers indicated that the awareness and sharing features were most important to them.
“I liked the fact that we can share I really like that we can share because if something is going on with me, and then he knows what I’m doing and then also If I get something on my side and he’s not aware of it and he gets it. Now were having a BBQ on Sunday, okay. I would put it in mine, I send it to him and now he knows about Sunday we’re having something. The shared calendar for me is a very good idea.” (C5, I1)
8 This includes both PwAs and other caregivers
142
“The best thing is … being able to send stuff to each other like that. You know, to share stuff.” (C6, B2)
6.10.2 Distributed Family Members
Participants from Families 3, 5 and 6 mentioned that the distributed nature of family members was a factor that could influence potential usefulness of Family-Link in such as way that greater distribution would result in greater usefulness.
C3, D3, and E3 were off work for the duration of I2 due to a union strike. C3 felt that Family- Link could be of greater use if the family members were working and thus not home all the time.
“Unfortunately, the timing of some of these phases meant that, for all being around, meant that we couldn’t take full advantage of what it could and would do.“ (C3, I2)
E3, also felt this way and independently raised the point during interviews.
“I found that because of the strike and I wasn’t working and neither was my mom and my sister. And we were home a lot and near each other a lot. And it was much easier to communicate verbally instead of putting it in. it would have been more useful if I had been going to work and not at home as much.” (E3, I2)
Although C5 liked the Family-Link, she spoke about not needing all of its features because she and A5 did not need to communicate over the distance often.
“Yeah, maybe we’re a little bit different because we don’t spend that much time apart. So as far as maybe other families, some are working, some are home. The need to communicate over the distance is important to them… that would just go back to (me) not needing so many of the features. Like having the calendar that communicates, the shared calendar. I like that, and that is something we do use. But we don’t have to communicate over a distance often.” (C5, I1)
A6 believed that Family-Link would have greater potential if C6 was working away from home.
143
“Before when she was working daily. It was harder to keep track. She had things planned in the evening which I didn’t keep track of. Phone calls. Even now she’s busy entering information. But now that she’s home I keep track of it anyway.” (A6, I1)
“I think it can be really beneficial with the right people… Now C6 and I are more together, and we really know what our calendars are. We don’t really change that much, that I’m not aware of, you know.” (A6, I2)
“You see her events are now a lil’ different. She’s working at home, and I’m pretty much aware of what is going on. So before, of course, when two people are separated by a job, you know, that she’s not here and checking on where she is what times she’s back. This program would help in a lot in that way. That is the whole issue. It is Family-Link and it does link family when they are not together or when they’re not quite sure what they’re up to, or they need that information. Like, if we had done this program last year. C6 was busy and had a job, and had I known more what she was doing, had her calendar available at that time, it would have been a lot more help. I could see that for a couple in that situation, that it would be beneficial.” (A6, I2)
Although C6 was not teaching and at home most of the time, she echoed A6’s comments.
“I can see what he’s up to and he can see what I’m up to. Very good especially when I go back to teaching.” (C6, I1).
“Well, it had good things. And I think it was really good the fact that we could share things. I think if I was outside, it would have been very, very helpful to me. It was still helpful here but not as much as outside. So I would find it good to share, good to get the alarm when we have the family news.” (C6, I2)
In referring to “outside”, C6 was speaking about working away from home “And if I was still teaching like I was doing, it would be very useful and helpful.” (C6, I2)
“It would be very helpful for me. Let’s say for example, while teaching I’m not allowed to use my cell phone but at least I would be able to glance at it... And if he had sent me something new, I would be able to at least have a look at it. Because talking, I’m not allowed to have my cell phone. So for me... seeing at least what he is doing it gives me a
144
sense of security. And that I don’t have to talk, I just have to open up my unit. I could have it on my desk and glance at it, and no body knows what I’m doing.” (C6, I2)
Summary: The mean usefulness ratings in 6 questions from the Usefulness Questionnaire were significantly greater than the neutral rating of 3. There were no significant differences in the negative direction for the other 8 questions. Qualitative evidence suggests that PwAs and caregivers found different features of Family-Link useful. Caregivers found the sharing and awareness features useful while PwAs found the individual calendar features useful. Several users felt that Family-Link would be more useful if primary caregivers were not co-located with PwAs during the day.
6.11 Preference and Feature Comparison
Hypothesis: I anticipated that participants would prefer Family-Link over Palm Calendar.
Analysis: I asked participants about overall preference in the Comparison Questionnaire (see Appendix 10.22 and 10.23).
6.11.1 Overall Preferences
Figure 6.14 : Response pie chart for the question, “Overall, which software do you prefer?”
3 preferred Family-Link (A4, E4, C5), 4 preferred Palm Calendar (C3, D3, D4, A5) and 5 had no preference (A3, E3, C4, A6, D6). The overall preferences are shown as a pie chart in Figure 6.14.
145
6.11.2 Comparison Using Listed Features
An overall preference does not communicate which features of a system were important and why. I compiled a list of features for shared calendar systems (see Appendix 10.22 and 10.23) and participants decided whether certain features were performed better using the Palm Calendar, Family-Link, or whether the two systems were roughly equivalent. See Table 6.24 for the scores that were assigned to each response in the comparison questionnaire.
Response Assigned Score Palm Calendar -1 Roughly Equivalent 0 Family-Link 1
Table 6.24 : Responses to the comparison questions and the scores that I assigned. I assigned a score of -1 for Palm Calendar, 0 for Roughly Equivalent, 1 for Family-Link.
N Mean Std. Dev. Min. Max. Plan family events 12 .25 .754 -1 1 Complete responsibilities in a timely fashion 12 -.42 .669 -1 1 Remind to do something 12 -.25 .622 -1 1 Know what your family member is doing 12 .67 .651 -1 1 Attend to and carry out responsibilities 12 -.33 .492 -1 0 Know completion status of responsibilities 12 .33 .778 -1 1 Allow greater participation in making important 12 .50 .798 -1 1 decisions Help stay on track during the day 12 -.17 .577 -1 1 Know what needs to be done in a given day 12 -.33 .492 -1 0 Know what events change during the day 12 .50 .798 -1 1 Handle situations when someone forgets 12 .00 .853 -1 1
Table 6.25 : Mean comparison scores of various features. A larger absolute mean value indicates greater strength in how participants felt about the calendar system. Negative mean values are associated with the Palm Calendar while positive values are associated with Family-Link.
Table 6.25 shows the comparison scores of the various features, grouped by subject type.
146
To determine whether there was a statistically significant difference between the mean comparison rating and the neutral rating of 0 for each question, a Sign test was carried out for each question. The Sign test indicated a significant difference for:
• Know what your family member is doing (p=0.021), in favour of Family-Link.
To examine overall preference for listed features, I counted the number of responses for each of the three categories (Family-Link, Palm Calendar, and Roughly Equivalent). The counts were: 43 for Family-Link, 34 for Palm Calendar, and 55 for Roughly Equivalent. Based on these totals, participants had a slight preference for Family-Link (43.0%) than in the Palm Calendar (34.0%) for the listed features (see Figure 6.15).
Figure 6.15 : Feature comparison pie chart based on listed features.
147
6.11.3 Comparing PwAs’ and Caregivers’ Perspectives
I was interested to know if there were any differences between PwAs and caregivers. I graphed the distributions based on subject type (see Figure 6.16). By visual inspection, primary and secondary caregivers were more similar in their distribution of responses than PwAs.
Figure 6.16 : Means comparing the calendars for the 11 features, by subject type.
148
PwAs Overall Overall Primary Secondary Secondary Secondary Caregivers Caregivers Caregivers Participants Mean of All of Mean Primary and
Plan family events 0.3 0.3 0.3 0.2 0.3 Complete responsibilities in a timely fashion -0.8 -0.3 -0.3 -0.2 -0.4 Remind to do something -0.5 0.3 -0.5 -0.1 -0.3 Know what your family member is doing 0.3 0.8 1.0 0.8 0.7 Attend to and carry out responsibilities -0.5 -0.3 -0.3 -0.2 -0.3 Know completion status of responsibilities -0.5 0.8 0.8 0.7 0.3 Allow greater participation in making important -0.3 0.8 1.0 0.8 0.5 decisions Help stay on track during the day -0.5 0.3 -0.3 0.0 -0.2 Know what needs to be done in a given day -0.3 -0.3 -0.5 -0.3 -0.3 Know what events change during the day -0.3 0.8 1.0 0.8 0.5 Handle situations when someone forgets -0.3 0.5 -0.3 0.1 0.0 Overall Mean -0.3 0.3 0.2 0.2
Table 6.26 : Mean comparison scores of various features, grouped by subject type. A larger absolute mean value indicates greater strength in how participants felt about the calendar system. Negative mean values are associated with the Palm Calendar while positive values are associated with Family-Link.
Table 6.26 shows the breakdown for mean comparison scores grouped by subject type. Everyone indicated a preference for the Palm Calendar for at least some features. In particular, the Palm Calendar was generally more preferred than Family-Link for PwAs. This might be a result of their training at Memory-Link and the amount of experience they had using the Palm Calendar prior to my study. It is not clear how these results would look had PwAs been introduced to Family-Link first.
In considering only PwAs’ responses, a Sign test was carried out to determine if there was a difference in mean comparison rating and the neutral rating of zero. The test found no significant differences.
149
In considering only caregivers’ responses (primary and secondary caregivers combined), a Sign test was carried out to determine if there was a difference in mean comparison rating and the neutral rating of zero. The test found significant differences for:
• Know what your family member is doing (p=0.016), in favour of Family-Link.
• Know completion status of responsibilities (p=0.031), in favour of Family-Link.
• Allow greater participation in making important decisions (p=0.016), in favour of Family-Link.
• Know what events change during the day (p=0.016), in favour of Family-Link.
Figure 6.17 : Counts comparing the calendars for a list of 11 features, listed for all participants.
Figure 6.17 shows the comparison counts for each participant. A5 felt very strongly that Palm Calendar was better than Family-Link for the listed features, while E4 and C6 felt the opposite.
150
One negative aspect of Family-Link that was revealed in the interview was the amount of time required by PwAs for events to be entered. Linear wizards increased the number of stylus taps needed to create an event (due to the need to navigate through multiple screens). Also, confirmations led to more stylus taps. This entire process slowed entry of appointments for PwAs who were used to the Palm Calendar. Also, the system took a second or two to process newly created events, a result of multiple database accesses via NSBasic code.
“I dunno because it’s pretty much all useful. I can’t say there’s anything in it that I dislike or wouldn’t want to use. Other than when you create a new event… Those three screens, I wish they wouldn’t come up. I would rather, and I don’t know how you do it, but I would rather you select that somewhere else… If most my stuff was done with other people or involving other people, then (it should) be part of the process of making a new event. But it’s not - most of my calendar, and I think the rest of the family, is most of the stuff is just to do with me. Going extra screens is just a redundant task that takes time.” (A4, I2)
“It was getting used to the program and trying different things. It took a while to get used to it. Again because I was using a Palm before but I was sort of faster with it, and the only thing I found in general, was it was a lot slower. Because I like to enter information easily, quickly, and that was sort of the only drawback with it.” (A6, I2)
6.11.4 PwAs’ Previous Experiences with Digital Calendars
PwAs had a preference for Palm Calendar. I explored the reasons for this in my qualitative data. The participants’ prior experiences with the digital calendars impacted what they found useful in Family-Link.
A1, A5 and A6 discussed their familiarity with the Palm Calendar and some difficulties they faced as they moved to using Family-Link.
“Palm is very (straight)forward. (I’ve been) using it for 4-5 years. (I’m) really well versed of ins and out of what it allows you to do and what not.” (A1, B2)
He explains how the transition can be difficult,
151
“Transition from something that you’re used to daily routine, trying to do something new, and something happens it forces you back. I can’t take the chance of losing information. Not being able to input critical information I’m able to refer back to when necessary.” (A1, I1b)
A1 explains that the change to Family-Link was drastic.
“Ever since the change to the new software I have noticed that I haven’t been entering as much information into the Palm Pilot as I did previously. The reason being … I’m going from something I’ve been using for 4 years because of changing at the beginning… I noticed it’s a bit more difficult to accept the change to the other program. However I did need to enter information to remind me. I had to do it. That’s what I were using. That was the primary calendar, but there was other types of items I would normally jot down.” (A1, I2)
A1 speaks more about Family-Link,
“I really didn’t find the program to be productive in the way that I thought it would be. Maybe it’s because I’ve become so used to using the Palm Calendar. I was able to colour code events, space things out, different alarm capabilities, amount of space that I was able to input information on, the fonts… The fact of beaming information over. Different alarms I wanted to go off and ring a certain way… The space. The room that I have on the actual screen. I find the Palm Calendar more comfortable to use. I’m not trying to say that the (Family-Link) program is not useful. It’s still a fairly new program. [I’m] not comfortable with it yet.” (A1, I1b)
Prior training appeared to play a role in PwAs’ preferences as the following quote illustrates.
“Well, I like the Palm Calendar because that’s what I was originally trained on.” (A5, I2)
As both C6 and A6 noted, A6 may have rated usefulness based on his extensive practice and experience with another Palm scheduling software, called Datebook5. Datebook5 has a very similar look and feel to the Palm Calendar and is feature-rich, but it does not offer sharing capabilities.
152
“Well the thing is I really don’t know if it’s just a matter of getting used to this one or it’s just that this one doesn’t have all the features that the Zire has. Because I think he mentioned to you Datebook5 so to me I don’t know enough to see what it does. Because Datebook5, I doesn’t know it. Like I say, in anything when there is a change he has to get used to it. But he knew the Datebook5 very well… He played -not played with it- but he was using it everyday. So he did have frustration with it too at the beginning. But I guess he was pretty good. He has been using the Palm for a long time. No, he started with Baycrest. He started using the Palm there. yeah, so I mean, he was going to the weekly session, and he would learn how to use it. But he has been using it for many years. He became very good with the Zire.” (C6, I1).
“Prior to it, again, I was using a different format. I was just trying to get used to using a calendar (Palm Calendar) on Treo. I guess I was just more able to use the program (Datebook5) I had entered on there, being familiar with it. It has a lot more options to it, and of course, then it’s more versatile and just has different things in it… I tried it and liked it and stayed with that one. I don’t know if there are other programs that are better than that, but when something suits you you tend to stay with it.” (A6, B1)
“I used the Zire for longer. I was more familiar with it and faster. And the setup I had, having things displayed in different format, it’s easier that way… In a month, I could attach icons to appointments. Looking at the whole month, I could see the icon attached to it. A little symbol… In Datebook5, I think there were some icons available. Different in that I could enter information like the colour coding, and just the display itself.” (A6, B1)
A6 also explained how becoming aware of new events was helpful but that he had developed a routine around setting up shared events that made sense to him before the study:
“I never did edit her information and when she entered something I more or less knew about it anyway. If she was doing something I didn’t know about it and she entered it and it came through to mine it would be helpful. In fact I’m so used to using the old calendar I would put her into a category. And I could check and go into a quick week view and go by the colour and get an idea of who was involved. And I hadn’t noticed a big improvement with Family-Link. I guess the system I had I sort of made myself. I knew better how to use it. It worked better.” (A6, I1)
153
Two drawbacks of Family-Link were the processing speed and the extra screens that were presented for setting the sharing options when creating new events.
“I know everyone is more familiar with what they’re used to. The old calendar, I used it a number of years. It’s just a matter of learning something new. At the same time, the format was easier with the calendar. Rather than going through all the processes and saying okay for this or that, enter it one time and it’s done. The quick and convenience of that was missing in Family-Link.” (A6, I2)
6.11.5 Continuing to Use Family-Link
Figure 6.18 : Pie chart of responses when asked whether they would continue using Family- Link if known shortcomings were addressed.
When asked whether participants would like to continue to use Family-Link if it was available after the study, 9 participants replied in the affirmative (D3, E3, A4, C4, D4, E4, C5, A6, C6), 1 in the negative (A5), and 2 remained undecided (A3, C3). Figure 6.18 graphically shows this.
Summary: Overall, participants had a slight preference for the Palm Calendar. By examining the preferences based on features, Family-Link had a slight edge. However, in comparing responses based on features, I found one significant finding (Know what your family member is doing) in favour of the Family-Link. While there were no significant preferences of PwAs based on features, caregivers significantly preferred Family-Link over the Palm Calendar for 4 features:
154 knowing what their family members are doing, knowing completion status of responsibilities, allowing greater participation in making important decisions, and knowing what events change during the day. Qualitative data suggested that PwAs felt that Family-Link was not an improvement over the Palm Calendar, despite the latter’s lack of explicit sharing functionality. Further exploration revealed that PwAs prior experience with digital calendars led them to articulate preferences based on user interface elements and features to which they were accustomed. A majority of participants reported wanting to continue using Family-Link if known shortcomings were addressed in future iterations.
6.12 Limitations
Some limitations of the study are described in this section.
• A convenience sample was used to recruit participants and consequently, the results may not be representative of entire population of families with a PwA.
• All PwAs are members of Memory-Link who graduated from the same training program (with the exception of A2) and have been using handheld technology for years. These experiences can be a factor influencing, among other things, individual performance, technology adoption, and preferences.
• All PwAs are adult males, and all primary caregivers are adult females. The results may differ for families not of this configuration (e.g., different genders, ages).
• The participating PwAs were not employed. Other aspects of social class (Kennedy and Wellman, in press), such as prior occupation, social economic status, and wealth/income, may also impact the results of this study.
• The questionnaires and interviews were carried out at the end of each phase and PwAs by definition have difficulty answering questions that require recollection of information. I was aware of this from the outset and collected information from caregivers to confirm responses. However, there were limits to what caregivers knew about PwAs’ daily activities.
155
• There is a limitation in my paper-based instruments in that they were administered once at the end of each phase. Events occurring on the day of sampling (e.g., seizure) could inadvertently impact questionnaire responses, leading to unrepresentative data. I could have administered questionnaires more frequently throughout the study, but I was concerned about participant withdrawal and decided against placing too many demands on my participants given the scope of the study.
• The analysis of counting shared events relied on grouping events based on the day for which they were scheduled. If the shared event occurred during a Baseline phase, it was included in the count for B1 or B2; if the event occurred during an Intervention phase, the shared event was included in the count for I1 or I2. However, events can be created weeks or months in advance. So a PwA using Family-Link during Baseline could potentially create an event that would occur during an Intervention phase. I tried to minimize this issue by not counting shared events during the week after a new phase started. Several phases lasted more than 3 weeks, but the data points after the 3 weeks were also excluded. This helped to temporally separate the phases.
• Questions from the electronic questionnaire were open to interpretation and were asked on a per-event basis. The questions were kept brief to avoid the need for scrolling on the small screen size of the handheld devices, but misinterpretation was possible. For example, there might not have been frustration at the event level (e.g., the caregiver had no problem with PwA being reminded by his device to take pills) but the caregiver could have been frustrated at events that were not listed on the calendar (e.g., PwA keeps looking for the old Palm device instead of using the new one). I included broader questions in the phone interviews to gather this information.
• Primary caregivers may have avoided answering the electronic questionnaire on days when they did not know about the PwAs’ schedules. I encouraged the primary caregivers to fill out the questionnaire even if they did not know the answer to some of the questions (i.e., they could select “Don’t know” in their responses). One reason that primary caregivers rarely did this may have been because they did not want to spend time answering the electronic questionnaire when they did not have information to add.
156
7 Summary and Discussion of Results 7.1 Summary of Results Measures Quantitative Evidence Qualitative Evidence Anxiety No significant difference across PwAs were unsettled at the beginning of the phases. study but it was based on new hardware. Completeness No significant difference across Interviews with primary caregivers confirmed phases. no difference between phases in perceived completeness of events. Timeliness No significant difference across Follow-up interviews with primary caregivers phases. confirmed no difference between phases in perceived lateness of events. Confidence with No significant difference across Evidence did not favour one system over the completeness phases. other. Caregiver effort No significant difference across C3 reported increased effort due to system phases. For the individual family failures during the I1. Caregivers in Family 4 analysis, C3 experienced more reported that Family-Link reduced the effort effort in I1 than B2, but C4 had required to coordinate (e.g., reduced need to less effort in I1 than B2. make repeated phone calls). Caregiver No significant difference across Family-Link caused frustration for Family 3 frustration phases. For the individual family due to A3 having difficulty with the repeat analysis, C3 had more functionality, which behaved differently from frustration in I1 than B1, more Palm Calendar. C4 reported that Family-Link frustration in I1 than B2, but less relieved stress by enabling her to track A4 frustration in I2 than B1. from work and said that it was like having a secondary caregiver around A4.
157
Caregiver No significant difference across Family-Link had a slight positive effect on Burden phases. Primary caregivers had caregiving burden because the system enabled significantly greater mean greater awareness in caregivers of PwAs’ burden scores than secondary schedules. caregivers. Sharing Significant difference between Participants shared more events during shared event counts across Intervention phases as compared to Baseline phases. There were significantly phases. While using Family-Link, caregivers more shared events in I1 than either entered events in the PwA’s calendar B1, and significantly more themselves, or had the PwAs do it. In both shared events in I2 than B2. cases, caregivers often shared information with PwAs in order for the event to be created or modified. Titles for shared events sometimes led to confusion. Awareness No significant difference in Lack of quantitative result may be because awareness of event completeness primary caregivers constantly monitored or timeliness across phases. PwAs. However, I found substantial qualitative evidence indicating increased awareness during Intervention. This led to perceived reduction in the amount of effort needed to coordinate and a greater sense of security in caregivers. Usefulness of The mean usefulness ratings in PwAs and caregivers found different features Family-Link six questions were significantly of Family-Link useful. Caregivers found the greater than the average rating of sharing and awareness features useful while 2.5. PwAs were concerned about PwAs found the individual calendar features features that would support useful. Several users felt that Family-Link memory. Caregivers indicated would be more useful if primary caregivers that the awareness and sharing were not co-located with PwAs during the day. features were most importants. Preference Overall, participants had a slight There were differences in preference based on preference for the Palm whether participants were caregivers or PwAs. Calendar. In comparing Caregivers felt Family-Link was better at responses based on features, I providing sharing, decision-making, and found one significant finding in awareness while Palm Calendar was better at favour of Family-Link. handling calendar functions. PwAs felt that Caregivers significantly Family-Link was not an improvement over the preferred Family-Link over the Palm Calendar, despite its lack of explicit Palm Calendar for four features. sharing functionality. PwAs prior experience A majority of participants with digital calendars led them to articulate reported wanting to continue preferences based on user interface elements using Family-Link if known and features to which they were accustomed. shortcomings were addressed.
Table 7.1 : Summary of results from the field evaluation.
Table 7.1 summarizes the significant results from the previous chapter.
158
Results from the Evaluation suggest that the Palm Calendar and Family-Link are comparable as a scheduling tools. Family-Link was able to perform as well as the Palm Calendar with the exception of a few performance issues and the lack of an effective calendar repeat function.
Family-link excelled in enabling users to share events in their calendars. Participants documented more events that involved family members during Intervention as compared to Baseline. In this sense, shared events are like shared landmarks (Muller et al, 2005), around which participants can coordinate their activities.
Family-Link also allowed study participants to feel more aware of other family members’ schedules. As a result, caregivers of Family 4 appeared to require less effort to coordinate during Intervention because of increased awareness offered by Family-Link. This was a benefit for them in that it saved them time. In a study of caregivers of patients with dementia, Langa et al. (2001) showed that caregiving time increased with the level of severity of the impairment. This might also be true for caregivers of PwAs – if so, small time savings may be important in cases of severe amnesia.
Family-Link did not replace coordination activities in participant families. Rather, users organized coordination activities around and through the device. The chat messaging feature of Family-Link was not utilized. Instead, family members discussed events in person or over the phone and one member would add it to the device. Family-Link became part of the coordination process and eased synchronization.
Caregivers found various sharing and awareness features most useful while PwAs found the individual calendar features most useful. One possible factor that limited the usefulness of Family-Link may be that some participating family members were not distributed and therefore did not benefit as much from Family-Link’s core features.
Ultimately, users were split on their preferences for the Palm Calendar or Family-Link. PwAs tended to prefer Palm Calendar for its calendar functionality that catered to individuals. PwAs learned many nuances of the Palm Calendar through extensive training and experience (e.g. how often the device needs to be charged, how files are backed up to the PC, etc), and demonstrated difficulties switching to new hardware and software. Caregivers tended to prefer Family-Link for
159 its sharing and awareness features. However, three quarters of participants would continue to use Family-Link if current shortcomings are addressed.
It is interesting to note the differences between quantitative and qualitative results for some of our measurements. Quantitative results for caregiver burden and awareness were not significantly different between Intervention and Baseline phases, but the qualitative results presented another view. In some cases, the qualitative data illustrated some limitations of quantitative measures. For example, the quantitative measurements of awareness showed no significant differences but there was substantial qualitative evidence suggesting otherwise. As mentioned earlier, the quantitative data might not have shown significance because of caregiver monitoring. A different measurement of awareness, perhaps a more direct questionnaire item, may have revealed a quantitative trend. This underscores the importance of collecting qualitative data to supplement the quantitative measures. The goal in my analysis was to show rough trends using quantitative data and then explore the trends or lack thereof in more detail. There are also alternative approaches to merging data in mixed methods research for trauma research (Creswell and Zhang, 2009).
7.2 Comparison of Participant Families
7.2.1 Family 1
A1 and C1 found the sharing and awareness aspects of Family-Link to be very useful. C1 felt that she could monitor A1 whenever she traveled and seeing whether he marked events as completed in Family-Link gave her a sense of how he was doing in his day. This information provided a hint about whether he may have experienced a seizure that day and it was important for her to know this so that she could report it to his doctor.
As A1 progressed through the study, A1 felt that he wanted an interface that looked exactly like the Palm Calendar but with the Family-Link features built in. He relied on specific features of the Palm Calendar that he had learned about over the years after his brain injury, such as the categorization and colouring of events which allowed him to easily scan his schedule for the day to identify important tasks. He missed such features as they were unimplemented in Family- Link. A1 felt that Family-Link would be very useful for PwAs who had not been trained to use the Palm Calendar.
160
7.2.2 Family 2
A2 and C2 appeared to really need an application for the PocketPC that included the features of Family-Link, but C2 had a very regimented and full work and family schedule and it appeared that she did not have much time to deal with any bugs in the prototype as the stress levels seemed to increase greatly when there were system errors in early versions of Family-Link.
A2 struggled with the Palm device because he had a great deal of experience using the PocketPC device after his brain injury and relied on several of its features, such as the synchronization between Outlook on his PC at his PocketPC’s calendar and the way the calendar allowed events to be rescheduled. While A2 expressed interest in integration of Family-Link features in the PocketPC, I did not have sufficient time to engineer a cross-platform application that would operate on both the Palm and the PocketPC.
7.2.3 Family 3
Due to a public strike, the caregivers of Family 3 were at home most of the time and did not need to coordinate through Family-Link as much as they might have had they been at work. D3 and E3 rarely used the Palm devices as they already had other paper-based aids to manage their schedules and were reluctant to switching over to using technology. This limited the usefulness of Family-Link because they did not enter events in their calendars for A3 to see.
A3 relied upon the repeat functionality of the Palm Calendar and he struggled with repeated events in Family-Link because they were implemented differently due to different assumptions of the data. This was an issue with C3 because she needed to work with A3 to modify or change repeated events, which A3 did often. This issue can be rectified in future versions of Family- Link with a redesign of the repeat functionality.
7.2.4 Family 4
Members of Family 4 used Family-Link effectively - planning together, sharing events, and checking on one anothers’ schedules frequently. They developed coordination policies around its use, such as discussing events that involved multiple family members and then deciding who would enter them into Family-Link so that the events were not duplicated accidentally. Members of Family 4 also used the beaming feature of the Palm device to share events using the Palm
161
Calendar but this only done during the first Baseline phase. Beaming events required effort and the Palm devices needed to be pointed at one another for the infrared line-of-sight beaming to work.
A4 had developed a strategy where he would categorize events in Palm Calendar as completed if they were done. While Family-Link did not have categories for events, it provided a feature that achieved the same outcome and A4 transitioned to using it without difficulty.
7.2.5 Family 5
Although A5 and C5 spent much of their time together at home or doing errands throughout the day, they valued Family-Link. C5 even felt that she missed the system during Baseline phases and after the study.
A5 did not rely on any particular feature of Palm Calendar that Family-Link did not offer. Coincidentally, out of all the PwAs in the study, he had the least experience with the Palm Calendar (see Table 5.2).
7.2.6 Family 6
While C6 really valued the awareness features of Family-Link, A6 felt that it did not perform as well as another Palm calendar system for individuals (Datebook5) that he had used over the years since his brain injury. C6 felt that Datebook5 outperformed the Palm Calendar as well, although it is much more similar in look and feel and functionality to the Palm Calendar.
Since C6 was working at home throughout the study, A6 felt that the Family-Link system was not as useful as it could have been.
7.3 Reflections on the Method of Evaluation
Choosing the study design for the evaluation of Family-Link proved challenging. At the outset, there was no obvious way to do the evaluation. A controlled experiment in laboratory settings was not possible given the limited ecological validity and my goal of examining system use in everyday settings. Prior research in this area has typically utilized an ABA study design (Wilson, JC, and Hughes, 1997; Yasuda et al, 2002) or a randomized control crossover study (Wilson, Emslie, Quirk, and Evans, 2001; Wilson et al, 2009). ABA studies start with a baseline phase,
162 followed by an intervention phase and another baseline phase at the end. In randomized control crossover studies, participants are randomly assigned one of two groups: intervention or control, and after some period of time, participants switch groups. As with ABA studies, participants experience both intervention and control phases.
I chose an ABAB single-case experimental design for my evaluation. The second intervention phase is typically included in ABAB studies for ethical considerations and for continuing treatment in clinical settings. In our case, use of Family-Link was not continued after the second intervention phase but the additional intervention phase allowed me to make an extra comparison between the second baseline and second intervention phase.
In the analysis of my evaluation data, a number of quantitative measurements did not show significant differences between Family-Link and Palm Calendar (in terms of PwAs’ anxiety, perceived completeness and timeliness of PwAs’ events, caregiver confidence, caregiver effort, caregiver frustration and caregiver burden). Instead, I found evidence that the benefits of Family- Link were in the measures of collaboration, namely sharing of events and awareness. That the other measures did not show significant differences may be important for researchers setting up evaluations comparing multiple interventions for persons with cognitive impairment. In a sense, I compared two interventions: the Family-Link program we created and the Palm Calendar program operating separately on multiple devices. Had I set up the evaluation as a comparison between Family-Link and no digital calendar, the results may have appeared quite differently. An important point to mention is the role of interviews with participants in the absence of quantitative significant findings. The interviews provided rich qualitative data that offered insight to the problem space beyond what the numbers indicated.
7.4 Study Complications and Considerations
7.4.1 Procedural Issues 7.4.1.1 Need for Error-Free Prototypes that Support Critical Features
Some PwAs were frustrated that prototypes lacked sophistication in certain features, such as different calendar views, and repeat functionality. These are typical features of commercial calendar programs, yet Family-Link offered only rudimentary versions of such features. Different PwAs had come to rely upon different features, and had difficulty whenever they were
163 not able to do something that they could do previously (using their old electronic calendar). Another issue was that it was not clear from the start of the study which secondary features were needed. I incorporated a number of settings into Family-Link to increase the chance that it could be customized to each PwA’s needs and this flexibility helped in some cases but not all. In future studies that seek to expand upon the questions investigated in this thesis, researchers should identify which features upon which PwAs rely and ensure that their prototypes support those features. An alternative view is to screen PwAs early to determine if they rely on features that a prototype does not provide.
Also, system reliability in a prototype is very important in cases when human memory failure is possible. Loss of Family-Link data during the study would have been disastrous as the information stored in the calendar relate to medical and social events that are important to PwAs and their family members. Storing this information was important during Baseline phases too. Backing up data to the PC was major concern for Family 3 and 6. To address this, I installed the commercial PC software onto their computers to enable backups of the Palm Calendar. This proved complicated as the commercial software had bugs and issues that needed to be resolved, such as not supporting Windows Vista.
7.4.1.2 Converting Events Between Phases
There were times when it was necessary to transfer information from one calendar software to another as the study shifted between phases. Initially, manual creation of events was done, but this was very time consuming and prone to error. Instead, I developed conversion tools to transfer the data from one calendar application to the other. I met with the families between each phase and used these conversion tools. However, manual validation of every event was also needed to address a few issues I will discuss below.
If the family was transitioning to the I1 or I2 phase, events were copied from Palm Calendar to Family-Link. I met with each participant in the study before any intervention phase, and manually disabled all alarms in the Palm Calendar that were scheduled to occur during the intervention phase. This prevented the Alarm Manager of the Palm device from activating and indicating to the users that events were missed. I also turned off the sound and vibrations for the Palm Calendar. I also hid the icon to the default Palm Calendar to prevent the PwAs from adding or editing entries in the Palm Calendar.
164
If the family was transitioning to B1 or B2, I disabled alarms in Family-Link. I copied events in Family-Link into the Palm Calendar, but only if the participant was involved in that event. This had the potential to yield inconsistent data when merging from Palm Calendar back to Family- Link in the I2 phase. For example, shared events in Family-Link that are copied to multiple Palm calendars may be modified individually, resulting in many versions of the same event. For example, consider a birthday party that is created in Family-Link during the I1 phase and involves everyone. At B2, this event is copied to everyone’s Palm Calendar. One person changes the time of this event in their Palm Calendar. This is not reflected in any of the devices since it is a Baseline phase. When merging the data for I2, it is not clear if the modified event is the same event as the others or if the event is a new event as the Palm Calendar does not track history. I overlooked these inconsistencies by clearing old Family-Link events for the duration of I2 and just copying the events back from the Palm Calendar as independent personal events (not involving others). This resulted in multiple copies of shared events (where each copy is on a separate calendar). Unfortunately, this solution only partially solves the issue because the result is that the birthday event is no longer shared in Family-Link. I performed manual validation and discussed any events in question with family members to resolve such inconsistencies.
Since participants would not have access to Family-Link after the study, they were concerned about how they could look at past information retrospectively. The loss of Family-Link data was not acceptable in these contexts, so I developed a method to transfer all Family-Link events into the Palm Calendar at the end of the study. This data needed to be kept for organizational, social, and medical purposes.
7.4.1.3 Length of Early Phases
It seemed that it took some time for PwAs to become accustomed to new hardware at start of study. The length of phases was informed by my pilot work, but varied because each family handled change differently. Caregivers reported that the amount of time for getting used to the Palm Calendar on the Treo was adequate, but this included training time during which PwAs used the Palm Calendar. Some caregivers (C3, C4) felt that more time was needed in the first Intervention phase for PwAs to get used to the new software. One caregiver suggested at least two months. In light of this, perhaps more time should be added to the early phases of an ABAB study in which new technology is introduced for use by individuals with severe memory
165 impairments. Lengthening some of these phases could help stabilize results and remove learning effects. This would also allow more time for PwAs to gain experience with the new hardware in real settings.
7.4.2 Deployment Issues
7.4.2.1 Network Connectivity and Remote Troubleshooting
The Internet Service Providers (ISP) that I chose was Rogers for Families 1 and 2, Bell for Family 3, and Telus for Family 4. This was because at the time that I began work with Families 1 and 2, the Centro unit was our only choice for the hardware platform. However, I later discovered that the Centro units were not well suited for general use with PwAs (see Section 5.5.1), and decide to use Palm Treo products for the other families. I settled on the Treo 700p for Family 3 and Treo 750p for Families 4, 5, and 6.
Both Treo devices support EV-DO wireless connectivity, which both Bell and Telus, offer on their cellular network. However, I learned that despite being within the Greater Toronto Area, not all areas have EV-DO access. I had not anticipated that this may be a problem because testing was done downtown and near Baycrest where EV-DO connectivity was not a problem. The first two families did not experience this problem, but Families 3 and 4 did. When EV-DO connectivity is not available, the Palm devices back off and use CDMA2000 1xRTT to communicate data. This is noticeably slower than EV-DO. Rather than taking a few seconds, the data communication exchange with Family-Link could take minutes. This significant delay was an issue because users could not use the device during this time (as it is single threaded).
To minimize the effects of this issue, I redesigned the software so that fewer updates were sent in each synchronization session (e.g. if the time was changed twice, only the second time change would be updated).
Due to the connectivity problem described above, if a device was communicating with the server for a long time, clients sometimes thought the device froze and then reset the device. The reset would result in the server rolling back any changes in the update session. As well, Family-Link would try to send all the updates on the next communication attempt. When a Palm device is reset (this is also known as a soft reset), third-party programs are not loaded into memory and so
166 the Family-Link synchronization process is not automatically scheduled. Thus, a device that was reset does not communicate with the server automatically.
I added a progress bar which increments with each item that is transferred to give more feedback to the user. I also built into Family-Link a trigger that would activate the scheduling for the synchronization process as long as the user tapped the Family-Link icon. If long delays were noticed in the log, I called the participant and asked them to tap the Family-Link icon. This allowed me to do some remote troubleshooting.
7.4.2.2 Long Delays Between Synchronizations
Although I provided instruction otherwise, family members who did not use the device on a daily basis often turned off their devices for days at a time or let their Palm run out of battery charge. This resulted in long time lapses between when Family-Link connected with the server, which led to having a lot of updates (e.g., change in the title of an event, notifications, new events). The server stored these events in a queue. When these users turned their devices back on and accessed Family-Link, the software connected to the server and began the update process. However, this process was lengthy because the software downloaded every update that was missed when the device was offline. This was exacerbated by poor network connectivity.
I resolved this by modifying Family-Link to only get updated information for the current and upcoming day when connecting after a prolonged period of time, and then grabbing only a handful of the pending updates every time the system reconnected with the server until all pending updates were sent. This spread the number of updates over multiple synchronization sessions.
7.4.3 Training Issues
Visual changes that do not change functionality were handled better by PwAs than visual changes that removed functionality. For example, there was a minor change in the Palm Calendar software between Zire 72 and Treo. In the Zire 72, the order of the buttons of the bottom screen are: New, Details, and Goto. In the Centro and Treo, the order of the buttons are: Details, New, and Goto. This ordinal change did not seem to impact users during training, but PwAs experienced issues when visual access to functionality was modified.
167
Most of the trouble with training was the need to bring up the digital keyboard. A1, A3, A5, and A6 preferred using the digital keyboard over the physical keyboard. On the Zire 72, bringing up the digital keyboard was done using a dedicated button of the device that the user could tap with the stylus. However, this dedicated button was not available on the Centro or Treo. It seemed that the lack of visual cues for bringing up the digital keyboard lengthened training times. When the digital keyboard was needed, PwAs would typically move their stylus downwards (where the dedicated button would have been located on the Zire), but would stop and not know the next step.
7.5 Individual Differences Affecting the Usability of Collaborative Memory Aids
If one could identify factors that increase the likelihood of successful adoption of cognitive technology, one might be able to screen new families with a PwA, and prescribe strategies and technology that have maximal applicability and chance of success.
In our study, two families (Families 4 and 5) were able to successfully adopt and utilize Family- Link during Intervention phases. However, Families 3 and 6 struggled in their use of Family- Link. It is useful to think about why Families 4 and 5 were more readily able to adopt the collaborative technology then were Families 3 and 6.
Evans at al (2003) found four factors that predicted use of memory aids in people with brain trauma:
1. Age: The younger the person, the more likely to use memory aids.
2. Time since injury: The shorter the time since injury, the more memory aids were being used.
3. Number of aids used before the brain trauma: There was a relationship between the number of memory aids use before the brain trauma and the number of memory aids use after the brain trauma.
4. Attention skills: Attention and speed of memory processing was strongly related to the likelihood that persons used memory aids.
168
Reviewing our results and study data, five factors emerged that might indicate whether particular families would derive benefit from Family-Link or other collaborative memory aids. The factors that may increase likelihood of successful uptake of new technology in families are: high severity of memory impairment in PwA, little proactive interference in PwA, low caregiver burden, willingness to develop collaborative practices, and distributed family members.
1. High severity of memory impairment in PwA : A4 and A5 scored lower than A3 and A6 in the following tests: Verbal Memory Delay, Visual memory learning, and Visual memory delay. Surprisingly, A4 and A5 were better able to use Family-Link. This may be because PwAs with high severity of memory impairment would have greater need for a memory tool. They might rely on an external aid more than others with less severe member impairment, and therefore may have greater familiarity with similar tools.
2. Little proactive interference in PwA : Proactive interference a type of interference in which already learned information interferes with the acquisition of newer information (Underwood , 1957). This inefficiency is characteristic of normal memory function and is often heightened in amnesia. It appeared that adapting to Family-Link was difficult for PwAs because of their extensive experience and past training with other hardware and software. A2 used different hardware (PocketPC). A3 had difficulty learning to bring up the digital keyboard and was also reliant on specific aspects of the repeat functionality in the Palm Calendar. A5 relied on another calendar tool (Datebook5) and really missed particular features that he had learned through much practice on his own. Proactive interference was observed in all PwAs, although to differing degrees. For example, C5 mentioned that A5 had a habit of turning off his cell phone while charging it because she encouraged that at the beginning, but he could not remember to turn on his cell when he went out so she could not reach him. Despite her attempts to get A5 to charge his cell without turning it off, he does not do so.
3. Low caregiver burden : Low caregiver burden may be a factor influencing successful use of collaborative technology. Throughout the study, the caregivers of Families 4 and 5 exhibited less caregiver burden than caregivers of Family 3 and 6. Low caregiver burden may suggest that caregivers are more willing to test and refine new memory strategies.
169
Caregivers with low burden may also have greater tolerance when things go wrong (e.g., computer glitches).
4. Willingness to develop new methods for collaboration : When the new devices were used by Family 4, C4, D4 and E4 developed new electronic calendar practices, such as beaming events, and tended to use the Treos for keeping track of events. In comparison, D3 and E3 did not maintain paper calendars prior to the study and did not develop routine calendar practices during the study. This limited sharing of events and awareness in Family 3. A3 himself noted that Family-Link was only as useful as the information that was input. This seems to suggest that families willing to develop collaborative practices would benefit most from collaborative technology.
5. Distributed family members : When PwAs are apart from caregivers during the day, PwAs are more prone to getting off track and caregivers are more worried about their PwAs. Family 4 was the only family in which members were apart for long portions of the day during the study and they seemed to have the greatest need for the collaborative memory aid. While other families were together more often, primary caregivers of Families 3, 5, and 6 mentioned that families with distributed caregivers had the greatest potential to benefit from Family-Link. This may be factor that could identify which families could benefit most from collaborative technology.
This information is summarized in Table 7.2.
170
Severity of Memory Willingness to Proactive Location of Impairment in PwA Caregiver Develop New Family Interference Family Burden Collaboration Verbal Visual in PwA Members Memory Memory Methods Minimal Caregivers development of on strike or Impaired- Mild to 3 Average Moderate collaboration vacation; Borderline Moderate methods lives in same house Minimal Caregiver development of works at Moderate to 6 Average Average High collaboration home; lives Severe methods in same house Developed Caregivers numerous all work strategies such outside as beaming home; events with the secondary Low Little or No Palm Calendar caregivers 4 Impaired Low Average Burden and developing live in a coordination different policy about location who would add Family-Link events Caregiver had Caregiver PwA enter her did not events into his work; lives Impaired- Mild to device so that in same 5 Impaired Low Borderline Moderate he could house remind her when the alarm rang
Table 7.2 : Summary of individual differences affecting the usability of Family-Link in the study participants.
171
8 Conclusions and Future Work 8.1 Summary
The purpose of this thesis was to learn about the collaborative needs of families with a PwA, and to design technology to support such families.
To gain a greater understanding in this domain, I carried out an ethnographic study with ten PwAs and their families and found that PwAs were extensively supported by family members who provided effective memory support. I applied distributed cognition theory to the data analysis and learned that participant families worked very closely together to coordinate and accomplish everyday activities, such as planning a doctor’s appointment. At the same time, however, these activities were often undermined by memory issues. This led to the notion that memory aids could be viewed as collaboration technology, rather than aids designed solely for individuals.
To a design collaborative memory aid to support PwAs and their families, I assembled a participatory design team that included PwAs and also family members on occasion. The design outcome was a collaborative technology called Family-Link, which I implemented for Palm Treo devices. I evaluated Family-Link with six families over a six-month period. Two families participated in the pilot study and four in the actual study. Results suggest while Family-Link
172 was roughly equivalent to the commercially available Palm Calendar (in terms of PwAs anxiety, PwAs completeness, PwAs timeliness, caregiver confidence, caregiver effort, caregiver frustration, and caregiver burden), participants shared significantly more events in their calendar when using Family-Link. Family-Link also increased awareness of family members’ schedules, which all participants valued. In particular, caregivers found that awareness enabled them to track their PwA, giving them a sense of security. As well, some caregivers reported that Family- Link saved them time and reduced the amount of effort needed to coordinate.
8.2 Contributions
This thesis contributed:
1. Evidence from an ethnographic study that members of families with a PwA work together as a distributed cognitive unit to combat the effects of memory impairment in the PwA. This led to the notion that memory aids might be designed as collaborative tools rather than solely as individual aids. This also suggests that distributed cognitive processes can be impaired, which was not previously articulated in DC theory.
2. A design of a collaborative memory aid for families with a PwA, based on participatory design with PwAs, family members, and neuropsychologists.
3. A set of study procedures and analysis methods for evaluating a collaborative memory aid in the field.
4. Results from a field study that suggest that Family-Link increased sharing of events in users’ calendars, increased awareness for caregivers, and was found to be useful by users.
8.3 Implications for the Design of Collaborative Memory Aids
This research has resulted in a number of implications for the design of collaborative memory aids.
8.3.1 Allow PwAs opportunity to interact with critical information
Our results suggest that PwAs can play an integral role in their own rehabilitation. They should be allowed to actively manage, change, and interact with information relevant to their care,
173 rather than only being a recipient of such information. Repetition and repeated exposure to critical information can also help PwAs become more familiar with such information.
8.3.2 Allow for personalization of shared resources
Events shared in Family-Link have the same textual title across multiple devices. However, titles that may be meaningful to one person may not be meaningful to another. For example, consider that “Doctor’s appointment” is a shared event that appears on two family members’ calendars because one member is attending the appointment while another is providing transportation. This title may lead to confusion as it does not indicate who has the actual appointment. Family-Link attempts to address this issue by adding the event creator’s name in brackets in the title of shared appointments that are created by others. This enables family members to distinguish such ambiguous events, but sometimes these events still caused confusion. Ideally, the system should allow shared events to be renamed in personal views to be more meaningful to the user while retaining the original title for others. Adding this feature for personal views of shared resources can help prevent confusion.
Personalization could also potentially help PwAs become more familiar with shared resources, as suggested by the levels-of-processing model of memory (Craik and Lockhart, 1972). This model suggests that the deeper the level of processing of information (e.g, form a relationship, make an association, attach a meaning to the information), the greater the retention of memory.
8.3.3 Support resource sharing without limiting PwAs’ independence
Most collaborative assistive technologies rely on caregivers to create all shared resources for persons with cognitive disabilities. Rather than taking such an approach, Family-Link explored a more symmetric interaction where PwAs and caregivers were able to create and edit shared resources. We observed an interesting dilemma; while caregivers found that adding events to PwAs’ calendars was useful, they did not want to take away the sense of independence and responsibility from PwAs. Thus, collaborative memory aids should support methods of resource sharing in a way that does not limit independence of persons the aids are designed to support. One way to do this in may be to provide PwAs with the ability to explicitly approve new events that are added to their calendars.
174
8.3.4 Avoid deviating significantly from existing user interface designs in each iteration of the design cycle
One way to minimize the negative effects associated with high proactive interference in PwAs is to develop new technology that closely resembles older technology that PwAs are accustomed to using. Designers should keep the user interface as close to the existing user interface as possible. This may limit designs and reduce the rate at which the user interface design changes, but appears to be necessary for memory impaired populations that suffer from significant proactive interference. Avoiding major deviations in user interface design between design iterations may also reduce the amount of training needed.
8.3.5 Maintain orientation of users by providing access to history
An issue that many PwAs face is withstanding distractions. Often distractibility can result in the PWA forgetting to enter an event in their calendar. Family-Link attempted to maintain the orientation of the user by using simple linear wizards rather than having a complex user interface, providing more confirmations of actions, and providing access to history with “back” and “forward” buttons. Although linear wizards seemed to decrease the amount of training needed, more stylus taps were needed to progress from one screen to the next. The extra confirmations also required stylus taps. PwAs found this level of explicit prompting unnecessary, but benefited from seeing the history of their actions when they became disoriented.
8.3.6 Provide different user interfaces and functionality for different stakeholders in DC systems even when the task is the same
I set out to design a collaborative system for a tightly-knit group where each member was performing the same task. From a DC standpoint, it was surprising that different users liked different functionality. PwAs ranked features supporting their memory highly, while caregivers valued sharing and awareness features. One could imagine designing two user interfaces (UI), each providing access to separate functionality. Carmien and Kintsch (2006) previously proposed the idea of using one interface for persons with cognitive impairments and one interface for caregivers to simplify the complexity of systems. Simplification of the UI would result in decreased training time and ease of use for PwAs. Deciding on the right amount of functionality to expose to the user should be based on individual preference and skill.
175
8.4 Future Directions
8.4.1 Further Use of Distributed Cognition Theory as a Framework
Distributed cognition was a useful theory that guided the data analysis of the Ethnographic Study. At a high level, the theory helped to show that families living with a PwA acted as cognitive units, which revealed that collaborative aids may support memory. Future work in the use of this theory in cognitive rehabilitation follows five areas.
The first area of future research is exploring the ability of DC to influence design. In my research, I designed a collaborative system for a tightly-coupled group and presented the same interface to each family member because they each had the same task to do. However, my findings demonstrated that different features were liked depending on if the user was a PwA or caregiver; this suggesting that different UIs could be useful. From the perspective of DC, this is a surprising finding as my results revealed that there were two unique perspectives on the exact same task that demanded attention. Additional research in this area could additionally explore how designs resulting from DC relate back to theory.
The second area of future work is the domain of applicability of DC. I have shown in this thesis that DC can be applied in the domain of PwAs and their family members. The theory positively influenced and how I made sense of my field data and my designs. In the past, research has focused more on structured teams that have been trained to perform a task. Applying DC to PwAs and their family members is perhaps more challenging, as elements emotion, anxiety, stress, relationship, and spontaneity come into play. It would be interesting to learn if DC can be applied to other domains similar to PwAs and their families. Therefore a potential area of research is exploring whether families coping with other cognitive disorders, such as Alzheimer’s Disease (AD), also act as distributed cognitive groups. In the case of AD, do caregivers coordinate their care of the person with AD? How are their strategies the same or different to the themes presented in this thesis?
The third area of future research is learning more about the composition of the DC system. How many units are needed for a DC system to be successful in their goals? I would hypothesize that the more units that are part of the DC system, the more helpful the system is in providing care to
176 the person with the cognitive impairment. However, there may be issues and challenges to communication and coordination among larger groups. This is an interesting area of future work.
The fourth area of future research is assessing and measuring the strength of distributed cognitive units. Rather than only assessing individual components separately, a metric that assessed the strength of the entire system could be very valuable. As Wilson (1999) has noted PwAs can vary considerably in their constellation of memory and cognitive impairment; this in turn could affect the cohesion and coping ability of families. Our study results in Chapter 6 show that families with a PwA did vary in a number of ways. Results from the evaluation showed that some families have a quality that enables them to more easily adapt to changes. This might be considered the strength of the DC unit, or how tightly knit the distributed cognitive unit works together to combat the effects of memory loss. One hypothesis might be that DC units with greater strength may adapt to new technology more easily. Further study would be needed to determine this.
Finally, more research is needed from DC theorists to integrate the idea that DC systems can have impaired cognitive processes that are distributed over artifacts and people. Can DC explain phenomena as distributed cognitive deficits and if so, what are the broader implications for DC theory? The research presented in this thesis suggests that thinking about DC systems in this way can be fruitful.
8.4.2 Participatory Design Method
I applied similar participatory design techniques in my earlier research with PwAs (Wu, 2004), with one notable change: the involvement of caregivers at the design phase.
An area of future work is to consider the composition and order of when to introduce various stakeholders in the design of collaborative aids when one stakeholder has cognitive impairments. Although I had difficulty scheduling family members to participate in the design sessions due to their work schedules and limited free time, their input was invaluable. However, it was not always ideal to have all the stakeholders together in one place. PwAs appeared to offer less feedback when caregivers were around, perhaps allowing caregivers to take the lead. Caregivers also tended to be very careful when speaking about memory issues in front of PwAs. In her research on a picture-based remote communication system for persons with cognitive
177 impairments, Dawe (2007a) involved caregivers in participatory design sessions, and suggested that “the motivations of the caregivers must be understood and balanced with the individual’s motivations during the design process” (2007b). More research is needed to understand this dynamic.
Another potential area of research is where to draw the line between group design and one-on- one design sessions. Customization (in my work, see Section 5.3.4) is an important process when working with PwAs (Cole and Dehdashti, 1998; Cole and Matthews, 1999) because of the degree of variation in the memory impairment that can lead to having particular needs. The group design sessions facilitated high-level design, while individual meetings with PwAs facilitated customizing and software tailoring to fit individual needs. Some PwAs offered a lot of feedback during training sessions. In some cases, the PwAs were unable to complete the ten trials within an hour and a half because they would suggest interface customizations at each step along the way. While customization is important for PwAs, one might imagine deviations from core system concepts if there were more individual design sessions over group design sessions. Future research should be carried out to determine where best to switch from group design to individual customization.
One might imagine that participants who were involved in participatory design sessions would have the most positive responses to Family-Link, but some negative feedback was received from them. It is possible that the relationship I created and maintained with my design partners enabled them to be open about their feedback regardless of if it were positive or negative. Further study of this effect would be interesting, as it lends support for participatory action research as a viable approach for designing for people with cognitive impairments.
Finally, in some instances, I persuaded the design team to follow my intuitions. For example, the design team envisioned advanced location-aware functionality for Family-Link but I convinced everyone to focus on the core functionality for the initial electronic prototype before discussing advanced features in detail (although I still discussed their ideas and recorded them for future iterations). Each member of the team was an expert in their own field – the neuropsychologists understood psychology theory, the caregivers knew about how to provide care, and the PwAs knew what it was like to live with severe memory impairments. I considered myself to be knowledgeable in technology and design, and so at various points throughout the design process,
178
I needed to persuade, refocus, or guide our team in the matters of technology and design. I was able to persuade the design team when such situations occurred. However, situations in which the design team disagrees with the technology designer could arise in the future, so a potential area of future research is determining a systematic way to handle or avoid such disagreements.
8.4.3 Collaborative Memory Aids
The study uncovered some limitations and areas of improvement for Family-Link, including repeated events and notifications. The next steps for this research are to evolve the design further through iterative design.
1. Family-Link’s method of handling repeated events needs to be redesigned and the underlying data structures need to be modified to accommodate general repeated events. The interface for editing and changing repeated events should also be modified to be more similar to the Palm Calendar.
2. Notifications were helpful for caregivers to know about important changes to events in Family-Link, but PwAs often set too many events to notify their caregivers, resulting in caregivers sometimes feeling overwhelmed by the number of notifications they received. While caregivers were able to work around this by having PwAs notify caregivers about fewer events, the notification system of Family-Link could be vastly improved by including a notification manager. The study revealed that some notifications were more important than others. Having a way to specify priority in Family-Link (rather than a binary option of providing notification or not) would be very helpful. As well, the interface could offer a view that presents the most important notifications first, or highlights them in the chronological view.
The feedback from participatory design sessions and our deployment experiences from the study have also pointed me toward several areas of future work more generally with regards to collaborative aids.
1. Integration of a Global Positioning System (GPS) would be very interesting and future work in this area is very promising. Caregivers reported wanting to know the location of their PwA at all times. Of the six families participating in the study, A4 experimented with a standalone GPS device and found it useful for getting around town. Sending the
179
GPS data to the caregiver could increase their sense of security, knowing that their PwA was safely at home or was at the right place at the right time. This also raises privacy implications that need to be further explored. A4 mentioned that it would be nice to have the GPS built into his Palm device so that he carries one fewer item whenever he leaves his home. This was important because he sometimes forgets items (currently he carries his keys, wallet, Palm device, GPS, and cell phone every time he leaves his house).
2. In our design sessions, PwAs also envisioned an extension for Family-Link that could use GPS data, task lists, and schedules to intelligently suggest tasks for PwAs, based on where they were, what needed to be done by the end of the week, and how much time they had. This would enable them to more efficiently plan their days on a week-long basis, instead of making multiple trips to the same location. Some of this planning is done by the caregivers at the moment, but an intelligent planner for PwAs and their family members may lessen the planning load of caregivers. Some aspects that need to be considered for such a system are the energy levels of PwAs during the day, the locations and transportation routes with which that they are familiar, and the willingness of PwAs to record what they need to do in advance.
3. In my ethnographic study, I saw that PwAs and caregivers differed in their prioritization of tasks and information. In the evaluation study, they differed in their opinions of which features were useful. One idea is to have multiple interfaces that expose different functionality of a collaborative memory aid based on the user. More research is needed to determine if this would be a better approach than the approach of equal access to a collaborative memory aid.
4. Finally, it would be interesting to know whether other user groups, such as persons with Alzheimer’s Disease and their caregivers, could benefit from collaborative memory aids such as Family-Link.
8.5 Conclusion
The Memory-Link program has shifted from paper-based memory aids to utilizing commercially available handheld devices for cognitive rehabilitation, and while this technology is promising, designers of such tools have often assumed that (a) users have full memory capacity, and (b)
180 memory aids are only for individual use. This thesis explores the consequences resulting from both of these assumptions being challenged at the same time. From this perspective, distributed cognition theory proved to be a useful framework. Technology can be designed to support family collaboration and involvement in a PwAs’ cognitive rehabilitation. This thesis takes a first step in designing and evaluating a collaborative memory aid, and shows that this research path has potential. The prototype I developed, Family-Link, was not significantly different from the Palm Calendar in measures that evaluate the system in general, which is an achievement given that the Palm Calendar is a commercial product that has had years of development and refinement. At the same time, in measures of collaboration, Family-Link significantly increased sharing of events in participants and provided greater awareness of other family members’ schedules. Qualitative evidence showed the importance of this in terms of reduced amount of effort needed to coordinate, increased caregiver time, and increased sense of security – important factors in improving the lives of families with a PwA, and easing their collaboration when not physically co-located.
181
9 Bibliography Ackerman, M. and Halverson, C. (1998). Considering an Organization's Memory. Proceedings of the ACM 1998 Conference on Computer Supported Cooperative Work , New York, NY: ACM, 39-48. Adair, G. (1984). The Hawthorne effect: A reconsideration of the methodological artifact. Journal of Appl. Psychology , 69 (2), 334-345. Allen, M., McGrenere, J., and Purves, B. (2007). PhotoTalk: The design and evaluation of a digital image based communication tool for people who have aphasia. Proceedings of ACM ASSETS 2007 , 187-194. Allen, M., McGrenere, J., and Purves, B. (2008). The field evaluation of a mobile digital image communication application designed for people with aphasia. ACM Transactions on Accessible Computing, 1(1), Article 5, 1-26. Bandura, A. (2006). Guide for Constructive Self-Efficacy Scales. In F. Pajares and T. Urdan, (eds.), Self-Efficacy Beliefs of Adolescents . Information Age Publishing. 307-337. Baecker, R.M., Marziali, E., Chatland, S., Easley, K., Crete, M., and Yeung, M. (2007). Multimedia Biographies for Individuals with Alzheimer's Disease and Their Families, Proceedings of 2nd International Conference on Technology and Aging , June 2007. Barlow, D., and Hersen, M. (1984). Single-case experimental designs: Strategies for studying behavior change , New York: Pergamon Press. Beard, D. and Palanlappan, M. (1990). A Visual Calendar for Scheduling Group Meetings. Proceedings of Computer Supported Cooperative Work (CSCW 1990). ACM Press, 279- 290. Beck, A.T. and Steer, R.A. (1993). Beck Anxiety Inventory Manual. San Antonio, TX: The Psychological Corporation Harcourt Brace & Company. Beck, A.T., Epstein, N., Brown, G., and Steer, R.A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology , 56, 893–897. Bedard, M., Molloy, D.W., Squire, L., Dubois, S., Lever, J.A., andDonnell, M. (2001).The Zarit Burden Interview: A New Short Version and Screening Version. Gerontologist , October 1, 2001; 41(5): 652-657. Bederson, B., Clamage, A., Czerwinski, M., Robertson, G. (2002).DateLens: A fisheye calendar interface for PDAs. ACM Transactions on Computer-Human Interaction (ToCHI) , 11, 1 (Mar. 2004), 90-119. Benidict, R.H.B. (1997). Brief Visuospatial Memory Test –Revised. Odessa, Fla.: Psychological Assessment Resources. Bjerknes, G., Ehn, E., and Kyng, M. (1987). Computers and democracy: A Scandinavian challenge. Gower, Brookfield, VT: 1987.
182
Bødker, S. (1991). Through the interface: A human activity approach to user interface design . Erlbaum, Hillsdale, N.J.: 1991. Boyd-Graber, J., Nikolova, S., Moffatt, K., Kin, K., Lee, J., Mackey, L., Tremaine, M., and Klawe, M. (2006). Participatory design with proxies: developing a desktop-PDA system to support people with aphasia. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (Montréal, Québec, Canada, April 22 - 27, 2006). R. Grinter, T. Rodden, P. Aoki, E. Cutrell, R. Jeffries, and G. Olson, Eds. CHI '06. ACM, New York, NY, 151-160. Caprani, N., Greaney, J. and Porter, N. (2006). A Review of Memory Aid Devices for an Ageing Population. PsychNology Journal , 4(3), 205-243. Carmien, S. (2004). MAPS: creating socio-technical environments in support of distributed cognition for people with cognitive impairments and their caregivers. In CHI '04 Extended Abstracts on Human Factors in Computing Systems (Vienna, Austria, April 24 - 29, 2004). CHI '04. ACM, New York, NY, 1051-1052. Carmien, S., Dawe, M., Fischer, G., Gorman, A., Kintsch, A., and Sullivan, J. F. (2005) Socio- Technical Environments Supporting People with Cognitive Disabilities Using Public Transportation, Transactions on Human-Computer Interaction (ToCHI) , 12(2), 233 - 262. Carmien, S., DePaula, R., Gorman, A., and Kintsch, A. (2004) Increasing Workplace Independence for People with Cognitive Disabilities by Leveraging Distributed Cognition among Caregivers and Clients. The Journal of Collaborative Computing, Special Issue of the Journal of Computer Supported Cooperative Work (CSCW), 13(5-6) Dec. 2004, 443-470 Carmien, S., and Fischer, G. (2008). Design, adoption, and assessment of a socio-technical environment supporting independence for persons with cognitive disabilities. In Proceedings SIGCHI Conference on Human Factors in Computing Systems (Florence, Italy, April 05 - 10, 2008). CHI '08. ACM, New York, NY, 597-606. Carmien, S., and Kintsch, A (2006). Dual User Interface Design As Key To Adoption For Computationally Complex Assistive Technology, in Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) Conference , June 2006 Atlanta, GA. Carmien, S. (2005). End User Programming and Context Responsiveness in Handheld Prompting Systems for Persons with Cognitive Disabilities and Caregivers, in ACM Conference on Computer-Human Interaction CHI05 , Portland Oregon USA, 1252-1255 Carmien, S. (2006). Socio-Technical Environments Supporting Distributed Cognition for Persons with Cognitive Disabilities . PhD dissertation, Department of Computer Science, University of Colorado. October 2006. Clement, A. (n.d.). Participatory Design . Retrieved June 22, 2010, http://cpsr.org/issues/pd/introInfo
183
Cohene, T., Baecker, R.M., and Marziali, E. (2005). Designing Interactive Life Story Multimedia for a Family Affected by Alzheimer’s Disease: A Case Study, Proceedings of ACM CHI 2005 , Portland, OR, April 2-7, 2005. Cole, E. (1999). Cognitive Prosthetics: An overview to a method of treatment, NeuroRehabilitation , 12(1), 39-51. Cole, E. (2006). Patient-centered design as a research strategy for cognitive prosthetics: Lessons learned from working with patients and clinicians for 2 decades, Keynote address to the Workshop on Designing Technology for People with Cognitive Impairments, ACM SIGCHI Conference 2006 , Montreal, April 22-3, 2006. Retrieved April 22, 2010, http://www.brain-rehab.com/biblio3.htm Cole, E. and Dehdashti, P. (1998). Computer-based cognitive prosthetics: Assistive technology for the treatment of cognitive disabilities, Proceedings of ASSETS 1998, Third International ACM Conference on Assistive Technologies , Marina del Rey, CA, 11-18. Cole, E., Dehdashti, P., Petti, L., and Angert, M. (1994a). Design and outcomes of computer- based cognitive prosthetics for brain injury: A field study of three subjects, NeuroRehabilitation , 4(3), 174-186. Cole, E., Dehdashti, P., Petti, L., and Angert, M. (1994b). Participatory design for sensitive interface parameters: contributions of traumatic brain injury patients to their prosthetic software. In Conference Companion on Human Factors in Computing Systems (Boston, Massachusetts, United States, April 24 - 28, 1994). C. Plaisant, (ed). CHI '94. ACM, New York, NY, 115-116. Cole, E., and Matthews Jr., M.K. (1999). Cognitive Prosthetics and Telerehabilitation: Approaches for the Rehabilitation of Mild Brain Injuries, in J. Murner, and T.M. Ettlin (eds) HSW-Distorsion & Leichte Traumatische Hirnverletzung Behandlungskonzepte , Basel, Switzerland, June 1999. 111-120. Consolvo, S., Roessler, P., and Shelton, B.E. (2004) The CareNet Display: Lessons Learned from an In Home Evaluation of an Ambient Display, Proceedings of the 6th Int'l Conference on Ubiquitous Computing: UbiComp '04 , (Sep 2004), 1-17. Consolvo, S., Roessler, P., Shelton, B.E., LaMarca, A., Schilit, B. and Bly, S. (2004) Technology for Care Networks of Elders, IEEE Pervasive Computing Mobile and Ubiquitous Systems: Successful Aging , Vol. 3, No. 2, (Apr-Jun 2004), 22-29. Craik, F.I.M., and Lockhart, R.S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior , 11, 671-684. Craik, F.I.M., and Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General , 104, 268-294. Creswell, J.W., and Zhang, W. (2009). The application of mixed methods designs to trauma research, Journal of Traumatic Stress , 22(6), 612-621. Curran, T and Schacter, D (2000) Cognitive neuropsychological issues, in M. Farah and T. Feinberg (eds) Patient-Based Approaches to Cognitive Neuroscience . The MIT Press, 291-299.
184
Davies, R., Marcella, S., McGrenere, J., andPurves, B. (2004). The ethnographically informed participatory design of a PDA application to support communication. Proceedings of ACM ASSETS 2004, 153-160. Dawe, M. (2006). Desperately seeking simplicity: how young adults with cognitive disabilities and their families adopt assistive technologies. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (Montréal, Québec, Canada, April 22 - 27, 2006). CHI '06. ACM, New York, NY, 1143-1152. Dawe, M. (2007a). Let me show you what I want: engaging individuals with cognitive disabilities and their families in design. In CHI '07 Extended Abstracts on Human Factors in Computing Systems (San Jose, CA, USA, April 28 - May 03, 2007). CHI '07. ACM, New York, NY, 2177-2182. Dawe, M. (2007b). Reflective Design-in-Use: Co-Designing an Assistive Remote Communication System with Individuals with Cognitive Disabilities and their Families . PhD dissertation, Department of Computer Science, University of Colorado. 2007. Decortis, F., Noirfalise, S., and Saudelli, B. (n.d.) Distributed Cognition as a Framework for Cooperative Work. Retrieved November 1, 2005, http://www- sv.cict.fr/cotcos/pjs/TheoreticalApproaches/DistributedCog/DistCognitionpaperDecortis. htm Degeneffe, C (2001) Family Caregiving and Traumatic Brain Injury. Health and Social Work . 26(4). Delis, D.C., Kramer, J.H., Kaplan, E., and Ober, B.A. (1987). California Verbal Learning Test: Adult Version. San Antonio, TX: The Psychological Corporation. Easterbrook, S., Singer, J., Storey, M.-A., Damian, D. (2008). Selecting empirical methods for software engineering research, Chapter 11 in Shull F, Singer J, Sjøberg D (eds) Guide to Advanced Empirical Software Engineering . Springer-Verlag: London. P. 285-311. Emerson, R.M., Fretz, R.I., and Shaw, L.L. (1995). Writing Ethnographic Fieldnotes. Chicago: University of Chicago Press. Epstein, N. B., Baldwin, L. M., and Bishop, D. S. (1983). The McMaster Family Assessment Device. Journal of Marital and Family Therapy , 9, 171-180. Evans J.J, Wilson B.A., Needham, P., and Brentnall, S.J. (2003). Who makes good use of memory aids? Results of a survey of people with acquired brain injury. Journal of the International Neuropsychological Society (2003) , 9, 925-935. Fischer, G., and Sullivan, J. (2002). Human-centered public transportation systems for persons with cognitive disabilities. In Proceedings of the Participatory Design Conference , 194- 198. Flor, N.V. and Hutchins, E.L. (1992). Analyzing Distributed Cognition in Software Teams: A Case Study of Team Programming During Perfective Software Maintenance. In J. Koenemann-Belliveau, T. G. Moher, and S. P. Robertson, (eds) Empirical Studies of Programmers: Fourth Workshop . 1991: Ablex Publishing Corporation. 36-64. Garbis, C. and Waern, Y. (1999). Team co-ordination and communication in a rescue command staff - The role of public representations. Le Travail Humain , Special issue on Human- Machine Co-operation, 62 (3), 273-291.
185
Glisky, E.L., Schacter, D.L., and Tulving, E. (1986). Computer learning by memory-impaired patients: Acquisition and retention of complex knowledge. Neuropsychologia , 24, 313- 328. Gowans, G., Campbell, J., Alm, N., Dye, R., Astell, A., and Ellis, M. (2004). Designing a multimedia conversation aid for reminiscence therapy in dementia care environments. Extended Abstracts on Human Factors in Computing Systems , 825-836. Greenbaum, J., and Kyng, M., (1992). Design at work: cooperative design of computer systems , Lawrence Erlbaum Associates, Inc., Mahwah, NJ, 1992. Halverson, C.A. (1995). Inside the cognitive workplace: new technology and air traffic control. PhD Thesis, UC San Diego. Halverson, C.A. (1994). Distributed Cognition as a theoretical framework for HCI: Don't throw the Baby out with the bathwater - the importance of the cursor in Air Traffic Control. Tech Report No. 94-03, Dept. of Cognitive Science, University of California, San Diego. Hawkey, K., Inkpen, KM, Rockwood, K, McAllister, M, and Slonim, J (2005). Requirements gathering with Alzheimer's patients and caregivers. Proceedings of SIGACCESS 2005 , 142-149. Hayes, G.R. and Abowd, G.D. (2006). Tensions in designing capture technologies for an evidence-based care community. Proceedings of CHI 2006 . 937-946. Hayes, G., Truong, K., Abowd, G., and Pering, T. (2005) Experience buffers: a socially appropriate, selective archiving tool for evidence-based care. Extended Abstracts on Human Factors in Computing Systems , 1435-1438. Heaton, R.K. (1981). Wisconsin Card Sorting Test . Odessa, FL: Psychological Assessment Resources. Hersh, N, and Treadgold, L (1994). NeuroPage: The rehabilitation of memory dysfunction by prosthetic memory and cueing. Neurorehabilitation , 4, 187-197. Hill, WC, Hollan, JD, Wroblewski, D, and McCandless, T (1992). Edit wear and read wear. CHI 1992 , 3-9. Hodges, S., Williams, L., Berry, E., Izadi, S., Srinivasan, J., Butler, A., Smyth, G., Kapur, N., and Wood, K. (2006). SenseCam: A retrospective memory aid. Proceedings of Ubicomp , 177-193. Hollan, J., Hutchins, E. and Kirsh, D. (2000). Distributed Cognition: Toward a New Foundation for Human Computer Interaction Research. ACM Transactions on Computer-Human Interaction (ToCHI) . 7(2), 174-196. Hutchins, E. and Klausen, T. (1996). Distributed cognition in an airline cockpit. In Y. Engeström and D. Middleton (eds) Cognition and communication at work . New York: Cambridge University Press. Hutchins, E. (1995). Cognition in the Wild. MIT Press, Cambridge, Mass. Hutchins, E. (1995). How a cockpit remembers its speed. Cognitive Science . 19, 265-288. Inglis, E.A, Szymkowiak, A., Gregor, P., Newell, A.F., Hine, N., Shah, P., Wilson, B.A., and Evans. J. (2003). Issues Surrounding the User-Centred Development of a new Interactive Memory Aid, Universal Access in the Information Society (2003) 2 (3), 226-234.
186
Kapur, N., Glisky, E., and Wilson, B. (2004). Technological memory aids for people with memory deficits. Neuropsychological Rehabilitation , 14(1/2): 41-60. Kennedy, T., and Wellman, B. (in press). It takes a network. In L. Rainie and B. Wellman (eds) Networked: The New Social Operating System . MIT Press, 2011. Kim, H.J., Burke, D.T., Dowds Jr, M.M., Boone, K.A.R., and Park, G.J. (2000) Electronic memory aids for outpatient brain injury: follow-up findings. Brain Injury , 14(2), 187- 196. Kraut, R., Steinfield, C., Chan, A.P., Butler, B., and Hoag, A. (1999). Coordination and virtualization: The Role of Electronic Networks and Personal Relationships. Organization Science , 10(6), 722-740. Kreutzer, J.S., Marwitz, J.H., Hsu, N., Williams, K., and Riddick, A. (2007). Marital stability after brain injury: an investigation and analysis. NeuroRehabilitation , 22(1), 53-59. Kincaid, C. and DuPont, P. (1985). Electronic Calendars in the Office: An Assessment of User Needs and Current Technology. Transactions on Office Information Systems , 3(1). ACM Press, 89-102. Lamming, M., Brown, P., Carter, K., Eldridge, M., Flynn, M., Louie, G., Robinson, P., and Sellen, A. (1994). The Design of a Human Memory Prosthesis, The Computer Journal , 37(3), 153-163. Langa, K.M., Chernew, M.E., Kabeto, M.U., Herzog, A.R., Ofstedal, M.B., Willis, R.J., Wallace, R.B., Mucha, L.M., Straus, W.L., and Fendrick, A.M. (2004). National Estimates of the Quantity and Cost of Informal Caregiving for the Elderly with Dementia. Journal of General Internal Medicine , 16 (11), 770-778. LoPresti, E.F. Mihailidis, A., and Kirsch, N. (2004). Assistive technology for cognitive rehabilitation: State of the art. Neuropsychological Rehabilitation , 14 (1/2), 5-39. Madsen, M., el Kaliouby, R., Eckhardt, M., Hoque, M. E., Goodwin, M. S., and Picard, R. (2009). Lessons from participatory design with adolescents on the autism spectrum. In Proceedings of the 27th international Conference Extended Abstracts on Human Factors in Computing Systems (Boston, MA, USA, April 04 - 09, 2009). CHI '09. ACM, New York, NY, 3835-3840. McGrenere, J., Davies, R., Findlater, L., Graf, P., Klawe, M., Moffatt, K., Purves, B., and Yang. S. (2003). Insights from the aphasia project: Designing technology for and with people who have aphasia. Proceedings of ACM Conference on Universal Usability , 112-118. Mihailidis, A., Barbenel, J.C., and Fernie, G. (2004). The efficacy of an intelligent cognitive orthosis to facilitate handwashing by persons with moderate to severe dementia. Neuropsych. Rehabilitation , 14(1/2), 135-171. Moffatt, K, McGrenere, J, Purves, B, and Klawe, M (2004). The participatory design of a sound and image enhanced daily planner for people with aphasia. Proceedings of CHI 2004 , 407-414. Morris, M., Lundell, J., Dishman, E., and Needham, B. (2003). New Perspectives on Ubiquitous Computing from Ethnographic Study of Elders with Cognitive Decline. Proceedings of Ubicomp 2003 , 227-242.
187
Morrison K., Szymkowiak A., and Gregor P., (2004) Memojog - An Interactive Memory Aid Incorporating Mobile Based Technologies, Proceedings of MobileHCI , Lecture Notes in Computer Science (LNCS) 3160, 481-485, Springer-Verlag. Mueller, E. (2000). A Calendar with Common Sense. Proceedings of Intelligent User Interfaces 2000 . ACM Press, 198-201. Muller, M.J., Kuchinskaya, O., Minassian, S.O., Tang, J.C., Danis, C., Zhao, C., Harrison, B., and Moran, T.P. (2005). Shared landmarks in complex coordination environments. In CHI '05 Extended Abstracts on Human Factors in Computing Systems (Portland, OR, USA, April 02 - 07, 2005). CHI '05. ACM, New York, NY, 1681-1684. Muller, M.J. and Kuhn, S. (1993). Participatory design. Communications of the ACM . 36(6), June 1993, 24-28. Mynatt, E.D., Rowan, J., Craighill, S., and Jacobs, A. (2001). Digital family portraits: supporting peace of mind for extended family members. Proceedings of the Conference on Human Factors in Computing Systems , 333-340. Neuropage (n.d.) Retrived March 1, 2008 from http://www.neuropage.nhs.uk Neustaedter, C. and Brush, A.J. (2006): "LINC-ing" the family: the participatory design of an inkable family calendar. In Proceedings of ACM CHI 2006 Conference on Human Factors in Computing Systems 2006 . 141-150. Neustaedter, C., Brush, A.J. and Greenberg, S. (2007). A Digital Family Calendar in the Home: Lessons from Field Trials of LINC. Proceedings of Graphics Interface , May 28-30, 199- 206. Neustaedter, C., Brush, A.J., and Greenberg, S., (2009). The Calendar is Crucial: Coordination and Awareness through the Family Calendar, In ACM Transactions on Computer-Human Interaction (ToCHI) , Vol. 16, Issue 1, April 2009, ACM Press. Newell, A.F. and Gregor, P., (1997). Human computer interfaces for people with disabilities, in M. Helander, T.K. Landauer, and P. Prabhu (eds) Handbook of HCI , 813-824. Nygaard, K. (1990). Program development as a social activity. In Information Processing 86 , H.- J. Kugler, Ed., Elsevier, Amsterdam. Reprinted in PDC'90: Participatory Design Conference Proceedings , A. Namioka and D. Schuler (eds.) Seattle, Wash. Computer Professionals for Social Responsibility (March-Apr. 1990). Oddy, M. and Cogan, J. (2004). Coping with severe memory impairment. Neuropsych. Rehabilitation , 14(5), 281-494. Olson, D.H., Portner, J., and Lavee, Y. (1985). Faces-III: Family Adaptability and Cohesion Evaluation Scales. In D. H. Olson, H. I. McCubbin, H. Barnes, A. Larsen, M. Muxen, and M. Wilson (eds) Family inventories: Inventories used in a national survey of families across the family life cycle. St. Paul: Family Social Science, University of Minnesota. 1– 42. Olson, D.H., Sprenkle, D.H., and Russell, C. (1979). Circumplex model of marital and family systems: I. Cohesion and adaptability dimensions, family types, and clinical applications. Family Process , 18, 3-28.
188
Palm™ DualDate™ 1.1 Getting Started Guide (n.d.) Retrieved January 28, 2010 from http://www.palm.com/us/support/handbooks/dualdate_ug.pdf Park, E., Bell, J.D., and Baker, A.J. (April 2008). Traumatic brain injury: Can the consequences be stopped? Canadian Medical Association Journal , 178 (9): 1163-1170. Parks, S.M., and Novielli, K.D. (2000). A Practical Guide to Caring for Caregivers. American Family Physician . December 15, 2000. Perry, M. (2003) Distributed Cognition. In J.M. Carroll (Ed.) HCI Models, Theories, and Frameworks: Toward an Interdisciplinary Science . San Francisco: Morgan Kaufmann Publishers. 193-223. Plaisant, C., Clamage, A., Hutchinson, H.B., Bederson, B.B., and Druin, A. (2006). Shared family calendars: Promoting symmetry and accessibility. ACM Transactions on Computer-Human Interaction (ToCHI) , 13, 3 (Sep. 2006), 313-346. Rhodes, B. (1997). The wearable remembrance agent: A system for augmented memory, in Proc. Int. Symp. on Wearable Computers (ISWC), Boston, MA, 1997, 123. Richards, B. (2007). Use of external memory aids in the rehabilitation of severe memory impairment. Proceedings of the 17th Annual Rotman Research Institute Conference. Brain and Cognition. Richards, B., Leach, L., and Proulx, G. (1990). Memory rehabilitation in a patient with bilateral dorsomedial thalamic infarcts. Journal of Clinical and Experimental Neuropsychology , 12: 395. Richards, S.B., Taylor, R.L., Ramasamy, R., and Richards, R. (1999). Single subject research: Applications in education and clinical settings . San Diego, CA: Singular Publishing. Rising Tide: The Impact of Dementia on Canadian Society (2009) Alzheimer Society, http://www.alzheimer.ca/english/rising_tide/rising_tide_summary.htm Robertson, G., Mackinlay, J.D., Card, S. (1991). The Perspective Wall: Detail And Context Smoothly Integrated. In Proceedings of CHI '91 Conference , New Orleans, Louisiana, June 1991. Association for Computing Machinery, 173-179. Rogers, Y. (1997) An Introduction to Distributed Cognition. Retrieved January 27, 2010, http://mcs.open.ac.uk/yr258/papers/dcog/dcog-brief-intro.pdf Rogers, Y. and Ellis, J. (1994) Distributed Cognition: an alternative framework for analysing and explaining collaborative working. Journal of Information Technology , 1994, 9(2), 119- 128. Rogers, Y. and Scaife, M. Distributed Cognition. (n.d.). Retrieved November 2005, http://www- sv.cict.fr/cotcos/pjs/TheoreticalApproaches/DistributedCog/DistCognitionpaperRogers.ht m Rogers, Y. (2005). An Distributed Cognition and Communication. To appear in The Encyclopedia of Language and Linguistics 2nd Edition. Edited by Keith Brown Elsevier: Oxford, 731-733. Rogers, Y. (2006). Distributed Cognition and Communication, in K. Brown (ed) The Encyclopedia of Language and Linguistics . Elsevier: Oxford. 181-202. Sachs, P.R. (1991). Treating families of brain-injury survivors . New York: Springer.
189
Scaife, M. and Rogers, Y. (1996). External cognition: how do graphical representations work? Int. J. Hum.-Comput. Stud. 45, 2 (Aug. 1996), 185-213. Schacter, L. (1996). Implicit memory: a new frontier for cognitive neuroscience. In Gazzaniga, M.S. The Cognitive Neurosciences , MIT Press, 815-824. Schulze, H. (2004). MEMOS: an interactive assistive system for prospective memory deficit compensation-architecture and functionality. Proceedings of the 6th international ACM SIGACCESS conference on Computers and accessibility. 79-85. Schulze, H., Voinikonis, A., Hoffmann, T., and Irmscher, K. (2003). Modeling a Mobile Memory Aid System, Proceedings of 13. ITG/GI-Fachtagung "Kommunikation in Verteilten Systemen -KiVS2003" Springer-Verlag, Informatikaktuell. Berlin, Heidelberg, New York, 143-153. Sohlberg, M.M., and Mateer, C.A. (2001) Cognitive Rehabilitation: An Integrative Neuropsychological Approach . New York: Guilford Press. Stapleton, S., Adams, M., and Atterton, L. (2007). A mobile phone as a memory aid for individuals with traumatic brain injury: A preliminary investigation. Brain Injury, 21(4), 401-411. Stull, D.E. (1996). The Multidimensional Caregiver Strain Index (MCSI): Its measurement and structure. Journal of Clinical Geropsychology , 2, 175-196. Svoboda, E., Richards, B. (2009). Compensating for anterograde amnesia: A new training method that capitalizes on emerging smartphone technologies . Journal of the International Neuropsychological Society , 15, 629-638. Szymkowiak, A., Morrison, K., Inglis, E.A., Gregor, P., Shah, P., Evans, J.J., and Wilson, B.A. (2004). Memojog - an interactive memory aid with remote communication. Paper presented at the Workshop on Universal Access and AssistiveTechnology (CWUAAT), Cambridge, UK. Szymkowiak, A., Morrison, K., Gregor, P., Shah, P., Evans, J., and Wilson, A. (2005). A memory aid with remote communication using distributed technology. Personal Ubiquitous Comput. 9, 1 (Jan. 2005), 1-5. Tan, D.S., Stefanucci, J.K., Proffitt, D.R., and Pausch, R. (2001). The Infocockpit: providing location and place to aid human memory. In Proceedings of the 2001 Workshop on Perceptive User interfaces. PUI '01 , vol. 15, 1-4. Teasdale, T.W. Emslie, H., Quirk, K., Evans, J.J., Fish, J., and Wilson, B.A. (2009). Alleviation of carer strain during the use of the NeuroPage device by people with acquired brain injury. Journal of Neurology Neurosurgery and Psychiatry , 80 (7). 781-783. Tee, K., Moffatt, K., Findlater, L., Macgregor, E., McGrenere, J., Purves, B., and Fels, S. (2005). A visual recipe book for persons with language impairments. Proceedings of ACM CHI 2005, 501-510. Tullio, J., Goecks, J., Mynatt, E., and Nguyen. (2002). Augmenting Shared Personal Calendars. Proceedings of User Interface Software and Technology (UIST 2002) . ACM Press, 11- 20.
190
Underwood, B.J. (1957). Interference and forgetting. Psychol. Rev . 64:49-60. Psychological Research. New York: Appleton-Century-Crofts. Vicente, K.J., and Torenvliet, G.L. (2000). The earth is spherical (p < 0:05): Alternative methods of statistical inference. Theoretical Issues in Ergonomics Science , 1(3): 248-271. Voinikonis, A., Irmscher, K., and Schulze, H. (2005). Distributed processing of reminding tasks within the mobile memory aid system, MEMOS. Personal and Ubiquitous Computing , 9(5), 284-290. Walther, K., Schulze, H., and Thöne-Otto, A. (2004). An interactive memory aid designed for patients with head injury: Comparing MEMOS with two commercially available electronic memory aids. Poster at First Congress of the European Neuropsychological Societies , April 2004. Wang, Q. (2005). Disability and American families: Census 2000 Special Reports. United State Census 2000 . Retrieved from http://www.census.gov/prod/2005pubs/censr-23.pdf Wechsler, D. (1999). Wechsler Abbreviated Scale of Intelligence . San Antonio, TX: The Psychological Corporation. Wilson, B.A. (1999). Case Studies in Neuropsychological Rehabilitation : Oxford University Press. Wilson, B.A. (2009). Brain injury: recovery and rehabilitation. Wiley Interdisciplinary Reviews: Cognitive Science . 1(1), 108-118. Wilson, B.A., Baddeley, A., Evans, J.J., and Shiel, A. (1994). Errorless learning in the rehabilitation of memory impaired people. Neuropsychological Rehabilitation, 4 , 307- 326. Wilson, B.A., Emslie, H., Evans, J., Quirk, K., Watson, P., and Fish, J. (2009). The NeuroPage system for children and adolescents with neurological deficits. Developmental Neurorehabilitation , 12(6), 421-426. Wilson, B.A., Emslie, H., Quirk, K., and Evans, J., (2001). Reducing everyday memory and planning problems by means of a paging system: a randomized control crossover study. Journal of Neurology, Neurosurgery and Psychiatry , 2001 April; 70(4): 477-482. Wilson, B.A., and Evans, J.J. (1996). Error-free learning in the rehabilitation of people with memory impairments. Journal of Head Trauma Rehabilitation , 11: 54-64. Wilson B.A., Evans J.J., Emslie H., Malinek V. (1997) NeuroPage: Evaluation of a new. memory aid. Journal of Neurology, Neurosurgery and Psychiatry , 63: 113-115. Wilson, B.A., J.C., and Hughes, E. (1997). Coping with amnesia: The natural history of a compensatory memory system. Neuropsychological Rehabilitation , 7(1), 43-56. Wu, M., Baecker, R., and Richards, B. (2005). Participatory design of an orientation aid for amnesics. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (Portland, Oregon, USA, April 02 - 07, 2005). CHI '05. ACM, New York, NY, 511-520. Wu, M., Baecker, R., and Richards, B. (2007). Designing a Cognitive Aid for and with People who have Anterograde Amnesia, in (ed) Jonathan Lazar, Universal Usability , John Wiley & Sons. 317-356.
191
Wu, M., Baecker, R., and Richards, B. (in press). Field Evaluation of a Collaborative Memory Aid for Persons with Amnesia and their Family Members, Proceedings of SIGACCESS 2010. Wu, M., Birnholtz, J., Richards, B., Baecker, R., and Massimi, M. (2008). Collaborating to remember: a distributed cognition account of families coping with memory impairments. In Proceeding of the Twenty-Sixth Annual SIGCHI Conference on Human Factors in Computing Systems (Florence, Italy, April 05 - 10, 2008). CHI '08. ACM, New York, NY, 825-834. Wu, M., Richards, B., and Baecker, R. (2004). Participatory design with individuals who have amnesia. In Proceedings of the Eighth Conference on Participatory Design: Artful integration: interweaving Media, Materials and Practices - Volume 1 (Toronto, Ontario, Canada, July 27 - 31, 2004). PDC 04. ACM, New York, NY, 214-223. Wu, M. (2004) The Participatory Design of an Orientation Aid for People with Amnesia . M.Sc. dissertation, Graduate Department of Computer Science, University of Toronto. Yasuda, K., Misu, T., Beckman, B., Watanabe, O., Ozawa, Y., and Nakamura, T. (2002). Use of an IC recorder as a voice output memory aid for patients with prospective memory impairment. Neuropsychological Rehabilitation , 12, 155-166. Zarit, S.H. (1985). The hidden victims of Alzheimer’s disease: families under stress. New York University Press. viii, 218, 84-85. Zarit, S.H., Reever, K. E. and Bach-Peterson, J. (1980), Relatives of the impaired elderly, correlates of feeling of burden. Gerontologist , 20(6), 649-655. Zukewich, N. (2003) Unpaid informal caregiving. Canadian Social Trends , Autumn 2003. Statistics Canada. Catalogue No. 11-006.
192
10 Appendices 10.1 Consent Form for Exploratory Study
Project Title: Information Sharing Between Amnesic Individuals and Their Families
Investigators: Dr Brian Richards, Department of Psychology, Baycrest Centre for Geriatric Care
Dr Ron Baecker, Department of Computer Science, University of Toronto
Mike Wu, Department of Computer Science, University of Toronto
We agree to participate in the study to investigate information sharing between amnesic individuals and their families.
I have been given the attached information sheet. I understand that all information obtained by the investigators will be kept confidential and my name will not be used in any publication. I understand the risks and benefits of participating in the study as described in the information sheet. I understand that I may withdraw from the study at any time and for any reason. I also understand that I may ask questions at any time about the procedures.
Client ______
(Print name) ______
Family member/partner/spouse ______
(Print name) ______
Relationship ______
Dated the ______day of ______, 20__
Witness ______
(Print name) ______
193
10.2 Information Sheet for Exploratory Study
Project Title: Information Sharing Between Amnesic Individuals and Their Families
Investigators: Dr Brian Richards, Department of Psychology, Baycrest Centre for Geriatric Care
Mike Wu, Department of Computer Science, University of Toronto
Dr Ron Baecker, Department of Computer Science, University of Toronto
We are asking you and your spouse/partner/family member to take part in a study to help us better understand the kinds of information that might be shared between individuals who have amnesia and their families. This will, in turn, help guide us in the development of technological aids and services that support this information sharing.
You will be participating in either a diary study, or a participant observation study. You will be made aware of which study you will be participating in when you receive this document.
(1) The diary study will last for 1 month and will involve you recording experiences in which you have information someone gave you that you want to remember or thatr you want to give someone else. We will provide a paper-based booklet that you can carry around throughout the study. We would like to know what kind of information is recorded, and if relevant, to whom it is intended, when it was shared, where it was shared, and how you shared it. The booklet will be collected at the end of the study for analysis and will be returned if you want. If you own a PDA, you may use it instead of the booklet to record this data. With your help, we will collect the appropriate files for data analysis.
(2) The participant observation will involve one researcher spending 30-40 hours in your home during the day with you and your spouse, spread over 3 to 5 days. The information of interest will be the same information as that gathered by the booklet in the diary study, just with more details. You will be given the chance to see the information collected at the end of each observation day.
Interviews will be conducted before and after these studies. Each will last roughly one hour and will involve you and your spouse/partner/family member. The interview will be conducted at
194 your home or at Baycrest Centre Psychology Department if you prefer. With your permission, the interview will be audio taped and later transcribed for us to analyze the information you provide. All information gathered from the study will be strictly confidential. Your name will not appear on transcripts; instead, your records will be assigned an ID number. At the conclusion of the study, all audio tapes will be destroyed. Written transcripts and data from the diary study will be locked in a secure area. Neither your name nor anything that could identify you will ever be published.
There is no risk to your physical or mental health. During the interview you may want to take a break – opportunities for a break will be provided.
Participation in this project will be of no immediate direct benefit to you. The development of our technological aids and services is in the early stages.
Your decision to participate or not participate in the study will not affect the service that you, or your family member, are currently receiving. It will not affect any services you may require in the future at Baycrest. If you decide to participate but subsequently change your mind, you may withdraw from the study at any time. This also will not affect current or future treatment.
The interview can be conducted in the Psychology Department at Baycrest or in your home if you prefer.
If you wish to contact someone not connected to the project about your rights as a research participant, feel free to call the Chair of the Research Ethics Board at (416) 785 – 2500 ext 2190.
If you have any questions about the current study, feel free to contact Dr. Brian Richards at (416) 785 – 2500 ext 2764. Also, if you have any questions in the future about the progress of our research including the study results, please contact Dr Richards at the above number.
195
10.3 Interview Questions for Exploratory Study
Daily Life What kind of information do you desire to remember or think you might need sometime later? (spend time exploring each of the cases) Recording Information Have you developed any methods or strategies to help you remember this information? Is this information recorded in any way (eg, writing it down, having someone else write it down, using a tape recorder)? What is most important to record? Why? Involving Others Are there ever times when you would like to remember some message so that you can tell someone else (for example telephone messages)? What is recorded when a message is taken down for someone else? What details are most important in this case? How was this done (using what method/materials)? From time to time when I was back home in Vancouver, I forgot to deliver a message to one of my sisters. Have you had any similar experiences? What happened in those cases? How often does this occur? Sharing Experiences Are there ever times when you share your memory strategies or experiences with others? Who was involved? How was this done (what was recorded, using what method/materials)? How often does this occur? Completion Did I miss anything that you would like to tell me about?
196
10.4 Diary Booklet for Exploratory Study BOOKLET: EXPERIENCES DEALING WITH CRITICAL INFORMATION ID: #______This is for use in a study being conducted at Baycrest Centre for Geriatric Care. If found, please contact Dr. Brian Richards at (416) 785 – 2500 ext 2764. Thank you.
197
10.5 Codes from Analysis of Exploratory Study
1. Families functioning as cognitive systems
a. Information storage and management
i. Use of external artifacts for scheduling
ii. Use of external artifacts for reminding
iii. Personal and collaborative storing of information (e.g. diary writing)
iv. Capturing images and storytelling
b. Redundancy in information and communication
i. Caregivers providing reminders (cueing)
ii. Updating caregivers with information about completeness of tasks (immediate, routine)
iii. Redundancy of information in multiple external artifacts
c. Coordination processes
i. Communication protocols
ii. Collaborative planning and scheduling
iii. Running errands with family members
2. Impacts of memory impairment on cognitive systems
a. Amount of effort required
b. Differences in information or task prioritzation
c. Stress, tension, frustration
198
10.6 Family-Link Screenshots
Progression of screens goes from left-to-right, then top-to-bottom.
199
200
201
202
203
204
205
206
207
208
209
210
211
212
10.7 Consent Form for Evaluation
A Study of a Collaborative Memory Aid for Families with a PwA
Investigators: Dr. Brian Richards, Mike Wu, and Dr. Ron Baecker
Study description
The purpose of this study is to evaluate an electronic memory aid that has been designed to help families with a PwA. Evaluative research at Memory Link is an important part of ensuring that we offer individuals with amnesia and their families the best available memory interventions.
The study will last approximately 5 months and will evaluate the effectiveness of new memory aid software called Family-Link that we have designed for Palm Treo devices. Over the next 5 months your participation in the form of training, interviews and questionnaire completion will be approximately 16 hours, or 3.3 hours per month. We will provide a Palm Treo for you and each of your family members for the entire duration of the study. You will be able to use the Treo as you live your daily lives. The study is divided into four 3-week-long phases. In the first and third months you will use the default Palm calendar tool. In the second and forth months you will use the new collaborative memory aid. At no point will you be without the support of the Treo or a calendaring application. At the end of each phase and also 3 weeks after the end of the last intervention phase, a researcher will visit you to interview you and ask you to fill out a paper questionnaire (overall, 5 visits total). 30 minute individual phone interviews will also be conducted. These will be audio recorded so that we have a transcript of what was said. The audio tape will be kept in a locked cabinet at Baycrest. If you do not wish to be audio recorded, you can decline it with no negative consequences and still continue to participate in the study. Paper questionnaires assessing anxiety and caregiver burden will be administered.
During the baseline and intervention phases, a researcher will make brief phone calls to you (5-minutes in duration) once a week to check up and make sure everything is running smoothly. The primary caregiver will be asked to complete an electronic
213 questionnaire using the Palm Treo for every day of the baseline and intervention phases. This questionnaire consists of multiple-choice questions and is estimated to take about 10 minutes to complete.
Prior to the first baseline and first intervention phases, you will go through training and an ease-of-use assessment. The duration of training will vary based on individual trainability. The ease-of-use assessment (1 hour) will be done with your family. In this assessment, we will ask you and your family to comment on ease-of-use issues about the software that arise from the training sessions. Video taping will be done during this ease-of-use assessment to record how the aid is used for the purposes of improving the design. The videotape will be stored in a locked cabinet at Baycrest. If you do not wish to be videotaped, you can decline it with no negative consequences and still continue to participate in the study.
For an overview and visual representation of the above schedule see “Study schedule” below.
Confidentiality
All personal and clinical information is kept confidential. Only researchers directly involved in the study will have access to your files. Files are stored in a locked filing cabinet in the Baycrest Psychology Department. During the entire study, we will electronically log the information that you store into the calendaring tools. A number will be assigned to each participant and this number will be used in the transmission of data. Transmitted information will be encoded before it is sent wirelessly. Information will be stored in password-protected files on a password-protected computer at the University of Toronto, behind locked doors. Only researchers involved with this study will have access to this information. Should you consent to your information being used for the purpose of this research, no identifying information will be disclosed in any resulting publication or presentation. If you are interested, we would be happy to provide you with the final results of the study.
Benefits and risks
214
Any benefits that you receive from your research involvement will be collective in that evaluative research at Memory Link contributes to our goal to provide clients and their families with the best available memory interventions. Upon the conclusion of the study, you will be able to keep the Palm Treo for personal use. Data plans for the Palm Treo will be provided for the entire duration of the study. There are no additional risks involved in participating in this study. If new information related to the benefits or risks of the study is obtained, you will be informed.
Consent to research
Your decision whether or not to participate in this research study is completely voluntary. You may withdraw at any time. Your decision will not affect the care that you, or your family members, receive at Memory Link or Baycrest.
Study schedule key
Paper questionnaires (10 minutes for participant, 10 minutes for caregiver)
Electronic questionnaire (10 minutes for caregiver)
Phone interviews (30 minutes with caregiver, 30 minutes with participant)
Checkup phone call (5 minutes with participant, 5 minutes with caregiver)
Ease-of-use test (1 hour with participant and caregiver)
Training session (5-6 hour-long sessions with participant; only 1 hour-long session with caregiver)
See the next page for the study schedule.
215
Study schedule
Pha We Sun Mon Tue Wed Thu Fri Sat se ek
1
2
Training
3
4
5
PalmCalendar 6
7
8
Training
9
10
11
Family-Link
12
13 Palm Calen
216
14
15
16
17
Family-Link
18
19
20
21
217
Signatures
I agree to participate in the study: “A Study of a Collaborative Memory Aid for Families with a PwA”.
I have read the information above and I understand the purpose of my participation, the procedures involved and the risks and benefits of participating in the study. I understand that my consent to participate in this research is voluntary and that I may withdraw my consent at any point in time with no consequence to the care that I receive at Memory Link or Baycrest.
The benefits and risks of this study have been explained to me. I understand that I will be audio and/or video recording may be used in a small portion of the study, but I may decline those options with no consequences.
It has been explained to me that the results of the study are confidential. Neither my identity nor any personal information will be available to anyone other than the researchers of this study. No identifying information will be disclosed in any resulting publication or presentation. I have been given a copy of this informed consent form which includes information about the study. After the study, I can request for the final results of the study from the researchers.
In no way does signing this consent form waive my legal rights nor does it relieve the investigators, sponsors or involved institutions from their legal and professional responsibilities.
Participant:
Name: ______Age: ______
Signature: ______
218
Investigator:
Name: ______
Signature: ______
Date: ______
If you have any further questions regarding this study or other research studies at Memory Link please feel free to call Dr. Brian Richards at (416) 785-2500 ext. 2764 or Eva Svoboda ext. 3194. If you wish to contact someone not connected with this project about your rights as a research participant, feel free to call Dr. Angie Troyer, Chair of the Research Ethics Board at (416) 785-2500 ext. 2190.
219
10.8 Electronic Questionnaire for Evaluation – Screenshots
Progression of screens goes from left-to-right, then top-to-bottom.
220
221
222
223
224
225
10.9 Electronic Questionnaire for Evaluation – List of Questions Question Response Options 0=Don’t know 5=All of it Do you believe this was done? Question 1 4=Most of it
(Completeness) 3=Half
2=Some of it 1=Not at all 0=Don’t know 1=Very early Question 2 2=A little early Was this done on time? (Timeliness) 3=On time 4=A little late 5=Very late 0=Don’t know 5=Strongly Agree Question 3 4=Agree I have confidence my family member was able to do this. (Confidence) 3=Neither 2=Disagree 1=Strongly Disagree 0=Don’t know 5=Strongly Agree Question 4 4=Agree This required effort from me. (Effort) 3=Neither 2=Disagree 1=Strongly Disagree 0=Don’t know 5=Strongly Agree Question 5 4=Agree This led to frustration for me. (Frustration) 3=Neither 2=Disagree 1=Strongly Disagree
Table. Questions in the Electronic Questionnaire and their responses. Values were assigned in scoring the results.
226
10.10 Training Form – Add an Event to Palm Calendar
227
10.11 Training Form – Add an Event to Family-Link
228
10.12 Training Form – Mark an Event Completed in Family-Link
229
10.13 Training Form – Add a Note to an Event
230
10.14 Training Form – Add a Chat to an Event
231
10.15 Palm Device Setup for Evaluation
• Set date, time, and location for the device to automatically sync with the cellular network
• Install Filez application ( http://www.freewarepalm.com/utilities/filez.shtml ) and remove the backup of pmTraceDatabase, which is an older Palm database that causes HotSync problems
• Install ApplicationsPanel ( http://mytreo.net/downloads/applications-panel,1257.html ) to hide unused icons from the PalmOS launcher
• Set the Palm calendar settings to have no default alarm, which is the original setting on the Zire devices
• Turn off the ring tone and remove the vibration setting for phone calls
• Apply label on back of device so participants can determine who the devices belong to.
232
10.16 Beck Anxiety Inventory
Participant Initials: ___
Date: ______
Below is a list of common symptoms of anxiety. Please carefully read each item in the list. Indicate how much you have been bothered by that symptom during the past month, including today, by circling the number in the corresponding space in the column next to each symptom. Not At All Mildly but Moderately Severely – it didn’t - it it bother me much. wasn’t bothered pleasant at me a lot times Numbness or tingling 0 1 2 3 Feeling hot 0 1 2 3 Wobbliness in legs 0 1 2 3 Unable to relax 0 1 2 3 Fear of worst happening 0 1 2 3 Dizzy or lightheaded 0 1 2 3 Heart pounding/racing 0 1 2 3 Unsteady 0 1 2 3 Terrified or afraid 0 1 2 3 Nervous 0 1 2 3 Feeling of choking 0 1 2 3 Hands trembling 0 1 2 3 Shaky / unsteady 0 1 2 3 Fear of losing control 0 1 2 3 Difficulty in breathing 0 1 2 3 Fear of dying 0 1 2 3 Scared 0 1 2 3 Indigestion 0 1 2 3 Faint / lightheaded 0 1 2 3 Face flushed 0 1 2 3 Hot/cold sweats 0 1 2 3
233
10.17 The Zarit Burden Interview
Participant Initials: ___
Date: ______
Please circle the response the best describes how you feel. Never Rarely Someti Quite Nearly mes Freque Alway ntly s 1. Do you feel that your relative asks for 0 1 2 3 4 more help than he/she needs? 2. Do you feel that because of the time you spend with your relative that you don’t have 0 1 2 3 4 enough time for yourself? 3. Do you feel stressed between caring for your relative and trying to meet other 0 1 2 3 4 responsibilities for your family or work? 4. Do you feel embarrassed over your 0 1 2 3 4 relative’s behaviour?
5. Do you feel angry when you are around 0 1 2 3 4 your relative? 6. Do you feel that your relative currently affects our relationships with other family 0 1 2 3 4 members or friends in a negative way? 7. Are you afraid what the future holds for 0 1 2 3 4 your relative?
8. Do you feel your relative is dependent on 0 1 2 3 4 you?
9. Do you feel strained when you are around 0 1 2 3 4 your relative? 10. Do you feel your health has suffered because of your involvement with your 0 1 2 3 4 relative? 11. Do you feel that you don’t have as much privacy as you would like because of your 0 1 2 3 4 relative?
234
12. Do you feel that your social life has suffered because you are caring for your 0 1 2 3 4 relative? 13. Do you feel uncomfortable about having 0 1 2 3 4 friends over because of your relative? 14. Do you feel that your relative seems to expect you to take care of him/her as if you 0 1 2 3 4 were the only one he/she could depend on? 15. Do you feel that you don’t have enough money to take care of your relative in 0 1 2 3 4 addition to the rest of your expenses? 16. Do you feel that you will be unable to 0 1 2 3 4 take care of your relative much longer?
17. Do you feel you have lost control of your 0 1 2 3 4 life since your relative’s illness?
18. Do you wish you could leave the care of 0 1 2 3 4 your relative to someone else?
19. Do you feel uncertain about what to do 0 1 2 3 4 about your relative?
20. Do you feel you should be doing more 0 1 2 3 4 for your relative?
21. Do you feel you could do a better job in 0 1 2 3 4 caring for your relative?
22. Overall, how burdened do you feel in 0 1 2 3 4 caring for your relative?
235
10.18 Interview Questions for Evaluation – Introduction
Participant Initials: ___
Date: ______
(in person)
Ask participant with amnesia to fill out Beck Anxiety Inventory and family members to fill out Zarit Burden Interview.
Ask the following questions for this semi-structured interview, follow-up with additional questions as necessary. 1. Schedule: when at home. Ideal times for meeting up between study phases. 2. Occupation, if any. 3. When got amnesia? 4. When attended Memory-Link? 5. Is anyone else living in your household? Do they participate in helping to remember important scheduling information? 6. What do you consider a personal event? What about family event? a. How do you keep track of family events? b. Do family members help remember personal events? c. Do family members help remember family events? 7. Can you give me some examples of events that may change over time (ie. changed dentist appointment)? a. What is a particularly volatile event that changes? b. What happens when there are changes to a family event? What is impact on anxiety? What is your emotional reaction? 8. Are others notified of changes? How are they notified? 9. Do you use paper calendars or memory books? (photo them if ok) a. How often do you use it? b. With whom do you use it? 10. Do you have any long term goals or tasks that you would like to improve upon in the new few months? a. Memory? b. Social? c. Hobbies? d. Are there events that you think you may plan to help you attain that goal? 11. What is the impact of the Palm since you began using it? a. On anxiety? b. On effort?
236
10.19 Interview Questions for Evaluation – End of Phase
Participant Initials: ___
Date: ______
(over phone with each family member separately)
TO EVERYONE
Ask participant with amnesia to fill out Beck Anxiety Inventory and family members to fill out Zarit Burden Interview.
Ask the following questions for this semi-structured interview, follow-up with additional questions as necessary. 1. Did you use the device? For what did you use the device? 2. What features of the system did you like? Dislike? 3. Can you think of any features that are important for us to add? 4. How often did you check the device (once a week, daily, many times a day)? 5. How informed do you feel about what’s going on in the family? 6. Can you give me examples of how the software assisted you? 7. Can you give me examples of how the software got in the way? (perhaps various situations: at home, on the road, etc) 8. How did your family member find using the device? 9. Did you notice improvements or problems with how your relative used it? 10. Has using the software changed with how he was using it before? 11. Can you tell me about any Events that were missed? 12. Have you made any progress on your long term goals? 13. What has been the impact of the device on your life over the last phase? In terms of burden (family member) or anxiety (individual with amnesia)? Has it made life harder/easier? 14. How is it the same or different from other tech that you use? Are you still using the wall calendar?
TO PARTICIPANTS WITH AMNESIA
Anxiety 1. For each of the symptoms you rated as Moderately or Severely, was it related to your use or experiences with the Palm device over the last few weeks? 2. Do you feel anxious about using the software?
Family dynamics
237
1. Do you feel that you are playing an active role in family affairs?
TO FAMILY MEMBERS
Stress 1. What aspects of the software that you have used over the past few weeks are a source of stress? 2. Were there any interesting events over the last few weeks in which changes in the event were not known to you or your relative? What was the result? Did it add stress? Did it require additional effort to resolve? 3. Does learning the software add to your stress? 4. Does maintaining/fixing the device add to your stress?
Family dynamics 1. Do you feel that your relative with amnesia has an active role in family affairs?
Reminding 2. Do you remind your relative with amnesia about person events?
TO EVERYONE
Are any events particularly volatile? As in, they change often or need updates frequently. Would you need to tell your family about these changes?
Is there anything that we missed in this interview that you would like to tell us?
238
10.20 Interview Questions for Evaluation – End of Study
Participant Initials: ___
Date: ______
TO EVERYONE
Ask participant with amnesia to fill out Beck Anxiety Inventory and family members to fill out Zarit Burden Interview. Then ask everyone to fill out the appropriate version of the comparison questionnaire and the usefulness questionnaire.
Ask the following questions for this semi-structured interview, follow-up with additional questions as necessary. 1. What positive experiences did you have while using Family-Link? a. What negative experiences did you have while using Family-Link? 2. Can you comment on aspects of Family-Link that worked well? a. Can you comment on aspects of Family-Link that need improvement? 3. What feature of Family-Link did you find most useful? a. What feature of Family-Link did you find least useful? b. What features would be important for us to add? 4. Can you give examples of events that were handled better by the Palm Calendar than Family-Link? a. Can you give examples of events that were handled better by Family-Link than the Palm Calendar? 5. How has using the Palm Calendar affected your lives? a. How has using Family-Link affected your lives? 6. Have you made any progress on your long term goals? a. Has the Palm Calendar or Family-Link made reaching those goals easier? 7. Would you continue to use Family-Link if it was available? a. Why or why not? i. If not, would you use it in future iterations if we addressed the system’s shortcomings? b. Would you use it given the associated wireless data costs?
TO PARTICIPANTS WITH AMNESIA
Anxiety 1. For each of the symptoms you rated as Moderately or Severely, was it related to your use or experiences with the Palm device over the last few weeks?
239
Family dynamics 1. Do you feel that you are playing an active role in family affairs?
TO FAMILY MEMBERS
Stress 1. What aspects of the software that you have used over the past few weeks is a source of stress? 2. Were there any interesting events over the last few weeks in which changes in the event were not known to you or your relative? What was the result? Did it add stress? Did it require additional effort to resolve?
Family dynamics 1. Do you feel that your relative with amnesia has an active role in family affairs?
TO EVERYONE Is there anything that we missed in this interview that you would like to tell us?
240
10.21 Interview Questions for Evaluation – Follow-up for Primary Caregivers
Participant Initials: ___
Date: ______
FOLLOW-UP INTERVIEW
Anxiety
Did your family member experience anxiety at any point during the study?
If yes, did you notice more anxiety when using the Palm Calendar, Family-Link, or was it about the same?
Caregiver Burden
Did you experience caregiver burden (ie. stress) at any point during the study?
If yes, did you notice more caregiver burden when using the Palm Calendar, Family-Link, or was it about the same?
Completeness of Events
Were events more fully completed by your family member when using Palm Calendar, Family- Link, or was it about the same?
Is it burdensome when things are not done?
Timeliness of Events
Were events done more on time by your family member when using Palm Calendar, Family- Link, or was it about the same?
Confidence
241
Were you more confident that events were done by your family member when using Palm Calendar, Family-Link, or was it about the same?
Effort
Was more effort required by you when using Palm Calendar, Family-Link, or was it about the same?
Frustration
Did you experience frustration at any point during the study?
If yes, did you feel more frustration when using Palm Calendar, Family-Link, or was it about the same?
Awareness
Were you more aware of your family member’s schedule when using Palm Calendar, Family- Link, or was it about the same?
Sharing
Did you share more events when using Palm Calendar, Family-Link, or was it about the same?
Did your family member shared more events when using Palm Calendar, Family-Link, or was it about the same?
Integration
Do you feel that your family member was more integrated with the family when using the Palm Calendar, Family-Link or was it about the same?
Change
Does your family member still use the Treo? Or Zire 72? Have you noticed any change in how your family member uses the Treo?
242
10.22 Questionnaire for Evaluation – Comparison of Systems for Family Member Participant Initials: ______Date: ______Please check off the most appropriate option for the following questions. Palm Family- Roughly Calendar Link equivalent Which software is better at helping your relative plan family events? Which software is better at helping your relative complete their responsibilities in a timely fashion? Which software is better at reminding your relative to do something? Which software is better at helping you know what your relative is doing? Which software is better at helping your relative attend to and carry out their responsibilities? Which software is better at helping you know that your relative has completed or not completed their responsibilities? Which software allows greater participation from your relative in making important family decisions? Which software is better at helping your relative stay on track during the day? Which software is better at helping your relative know what needs to be done in a given day? Which software is better at helping you know what events change during the day? Which software do you prefer to use if someone forgets and something goes wrong as a result? Overall, which software do you prefer? Palm Calendar Family-Link No preference Would you use Family-Link if it was available? No Yes Undecided
243
10.23 Questionnaire for Evaluation – Comparison of Systems for Participant Participant Initials: ______Date: ______Please check off the most appropriate option for the following questions. Palm Family- Roughly Calendar Link equivalent Which software is better at helping you plan family events? Which software is better at helping you complete your responsibilities in a timely fashion? Which software is better at reminding you to do something? Which software is better at helping you know what your family is doing? Which software is better at helping you attend to and carry out your responsibilities? Which software is better at helping you know that you have completed or have not completed your responsibilities? Which software allows greater participation from you in making important family decisions? Which software is better at helping you stay on track during the day? Which software is better at helping you know what needs to be done in a given day? Which software is better at helping you know what events change during the day? Which software do you prefer to use if someone forgets and something goes wrong as a result? Overall, which software do you prefer? Palm Calendar Family-Link No preference Would you use Family-Link if it was available? No Yes Undecided
244
10.24 Questionnaire for Evaluation - Usefulness of Family-Link Participant Initials: ______Date: ______
Please circle the most appropriate rating for the following features in Family-Link .
Not Somewhat Very Uncertain Useful Useful Useful Useful
Viewing other family members’ 1 2 3 4 5 calendars
Editing other family members’ 1 2 3 4 5 calendars
Setting alarms for other people 1 2 3 4 5
Having the changes in calendars automatically sync with other 1 2 3 4 5 calendars
Being notified of changes in events 1 2 3 4 5
Attaching notes to an event 1 2 3 4 5
Attaching chat messages to an event 1 2 3 4 5
Searching for events in your calendar 1 2 3 4 5
Being continually reminded until the 1 2 3 4 5 event is done
Seeing who created events on your 1 2 3 4 5 calendar