This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Assistive technology (AT) is rapidly emerging within dementia care and support. One area of AT application is support of people with dementia in compensating for cognitive symptoms and thereby promoting their self-management. There is, however, little evidence for the applicability, usability, and effectiveness of AT for people with dementia, and there is a need to identify factors that can promote adoption.
This study aimed to (1) evaluate the applicability and usability of an app, tailor-made for people with dementia; (2) explore factors affecting adoption; (3) explore the possible influence of caregiver involvement; and (4) contribute to process evaluation of the intervention.
The ReACT (
Data from 112 participants and 98 caregivers were included. Shorter time from diagnosis (U=595;
For participants who became adopters, the ReACT app and the methods for self-applied implementation were applicable. However, the results were also in accordance with the well-known challenges of nonadoption and nonadherence to digital health interventions. The study provided insight into the importance of timely introduction and caregiver support for adoption of AT among people with dementia. It also underlined the high complexity of personal and contextual factors that influence adoption. These complex factors need to be considered when designing and implementing AT for people with dementia.
Globally, the number of people living with dementia is increasing rapidly [
AT comprises a variety of solutions, ranging from basic everyday low-technology devices to advanced technology devices [
The optimism on the potential of AT to support people with dementia is unfortunately not based on strong evidence. There is a great need for research addressing applicability, usability, and effectiveness of AT for people with dementia [
This paper has presented data from the third substudy of the research project,
Accordingly, the aims of the study were to (1) investigate the applicability and usability of the ReACT app to a mixed population of people with dementia, (2) investigate user patterns and factors influencing adoption and nonadoption, (3) explore the possible influence of caregiver involvement on participants’ adoption of the app, and (4) contribute to process evaluation of the app and methods used for deployment and adoption to guide future adaption of the app and methods of implementation.
Participants were recruited from 9 Danish memory clinics. The aim was to recruit a broad variety of people with dementia, and therefore, few inclusion criteria were specified. Participants were eligible if they (1) were patients in the memory clinics, (2) were motivated to try using the app, and (3) had access to a tablet where the app could be installed (iPad). It was not mandatory to have a caregiver as coparticipant, but if they were accompanied by a family caregiver when visiting the clinic, the caregiver was invited as coparticipant. There were no inclusion criteria related to age, language, or other personal or disease-related characteristics, and participants were not required to have had a final diagnosis at the time of inclusion. As the intervention addressed people with dementia as primary participants, they are referred to as participants in this paper, and the coparticipating caregivers are referred to as caregivers.
Information about the study was presented to patients in the memory clinics. Posters and flyers presenting the app and the study were available in waiting areas and introduced by staff. Eligible participants who showed interest in the study were given a detailed oral introduction and additional written material describing the details of the study. Participants and caregivers were then given an opportunity to deliberate before deciding whether to participate in the study.
Participants were recruited from June 2017 to February 2018, and during this period, 116 participants and 98 caregivers were enrolled.
The regional scientific ethical committees of the Capital Region of Denmark (protocol number H-15005558) evaluated the study protocol and decided that the study did not need approval because it was not considered to be within the framework of biomedical research. All participants received oral and written information about the study objectives and methods, and all participants gave written informed consent.
Participants and caregivers were given access to the ReACT app after being included in the study. The specific features of the ReACT app are illustrated and specified in
Features of the ReACT app.
As described in
Participants and caregivers were provided written material to support their self-applied implementation of the app. This written material had been validated as applicable for a person with early stage dementia during the second substudy of the ReACT research project [
A help feature was also built into the app, as illustrated in
Demographic information for both participants and caregivers was collected at the time of inclusion. Data included participant’s age, gender, and education; caregiver’s gender; and the relation between participant and caregiver. In addition, data from participants’ medical records were included to document diagnosis, time of diagnosis, and the most recent score on the Mini-Mental State Examination (MMSE) [
App usage was monitored through data logs from each participant’s and caregiver’s use of the app. The log files provided data on participant’s and caregiver’s actions in the app and provided information on action types and timestamps for these actions. The action types that were logged were activating the app; using adaptive features; and activities related to appointment, diary note, memos, checklist, and search features.
Adoption of the app was defined as a minimum period of 90 days between the first and last use of the app. This criterion was set based on results from the previous substudy [
A Web-based survey was conducted to collect additional background information and to collect feedback on the app. It was distributed via email 3 to 4 months after inclusion in the study. In cases where email correspondence was unsuccessful, a printed version of the survey was sent out by mail.
Two versions of the survey were distributed: one for participants and another by-proxy version for caregivers. It was constructed in an adaptable manner, enabling specific questions being directed at those who had used the app and those who had not. The survey included questions with a fixed set of possible answers. For nonusers, the questions addressed the level of skills to use a tablet and reasons for not using the app. For users, the questions addressed methods used when learning how to use the app and the level of skills to use a tablet. Moreover, 2 optional text boxes were also included, allowing additional comments on reasons for not using the app and general feedback on the app. The general feedback mainly addressed specific technical and functional issues of the app, which is not within the scope of this paper.
In those cases where the participant had tried using the app, the USEdem questionnaire [
Quantitative data were analyzed using IBM SPSS Statistics version 22. Baseline characteristics and log data from adopters were explored with descriptive statistics. Possible between-group differences on baseline characteristics, log data, and data from the surveys were analyzed using nonparametric chi-square tests for categorical variables, and for continuous variables, Kruskal-Wallis tests, Mann-Whitney U tests, or Fisher exact tests were used as appropriate. Logistic regression was conducted to explore whether baseline characteristics predicted adoption status.
Tests of significance were performed 2-tailed, with a significance level of .05. Imputed values for missing data on background characteristics were calculated following standard procedures for multiple regression modeling.
Qualitative data from surveys, from the textbox allowing feedback on reasons for not using the app, were processed and summarized in themes, as outlined in constant comparison analysis [
Data from 112 participants and 98 caregivers were included in data analysis. No participants or caregivers withdrew their consent to participate in the study, but because of insufficient background information from 4 participants, these were excluded from the original sample of 116 participants.
Log data monitoring the use of the app were collected for all participants and caregivers for a maximum of 90 days; hence, the intervention period for this study was 90 consecutive days after activating the app or 90 days from the inclusion of those who did not activate the app. Additional follow-up data are not within the scope of this paper.
Data from the surveys were obtained from 35 participants and 30 caregivers, and 19 of these cases were overlapping, with a reply from both. Of 35 participants, 14 had support from a caregiver when answering the survey, and in 2 cases, a caregiver had answered the participant survey on behalf of the participant; these 2 were excluded, leaving 33 participant and 30 caregiver replies for data analysis. Included in this were data on the USEdem questionnaire from 14 participants and 9 caregivers from cases where the participant had tried using the app were included.
The characteristics of participants and caregivers are summarized in
Characteristics of participants and caregivers and differences between adopters and nonadopters.
Characteristics | All participants (N=112) | Adoptersa (N=18) | Nonadoptersa (N=94) | |||
Age (years), mean (SD; range) | 68 (8.8; 39-86) | 69 (10; 52-82) | 68 (8.6; 39-86) | .86 | ||
|
|
|||||
|
Women | 49 (43) | 9 (50) | 40 (43) | .51 | |
|
.36 | |||||
|
≤10 | 16 (14) | 3 (17) | 13 (14) |
|
|
|
11-12 | 23 (20) | 3 (17) | 20 (21) |
|
|
|
13-14 | 25 (22) | 2 (11) | 23 (25) |
|
|
|
15-16 | 29 (25) | 8 (44) | 21 (22) |
|
|
|
≥17 | 19 (16) | 2 (11) | 17 (18) |
|
|
|
.29 | |||||
|
Alzheimer disease | 65 (58) | 12 (67) | 53 (56) |
|
|
|
Vascular dementia | 2 (1) | 0 | 2 (2) |
|
|
|
Dementia with Lewy bodies | 1 (0) | 1 (6) | 0 |
|
|
|
Frontotemporal dementia | 3 (2) | 0 | 3 (3) |
|
|
|
Mild cognitive impairment | 9 (8) | 2 (11) | 7 (8) |
|
|
|
Otherb | 27 (24) | 2 (11) | 25 (26) |
|
|
|
Unresolvedc | 5 (4) | 1 (6) | 4 (4) |
|
|
Months since diagnosis, mean (SD; range) | 12 (15; 0-73) | 6 (7; 0-25) | 16 (14; 0-73) | .046 | ||
Mini-Mental State Examination scored, mean (SD; range) | 25 (4.2; 11-30) | 25 (5; 11-30) | 25 (4; 11-30) | .69 | ||
|
|
|||||
|
Yes | 98 (87) | 15 (83) | 83 (88) | .70 | |
|
|
|||||
|
Woman | 58 (59) | 8 (44) | 50 (60) | .78 | |
|
.30 | |||||
|
Spouse | 81 (83) | 11 (73) | 70 (84) |
|
|
|
Son or daughter | 13 (13) | 4 (27) | 9 (11) |
|
|
|
Other | 4 (4) | 0 | 4 (5) |
|
|
|
|
|||||
|
Adopter | 7 (6) | 3 (17) | 4 (4) | .08 | |
|
|
|||||
|
Yes | 21 (19) | 8 (44) | 13 (14) | .02 |
aAdoption was defined as the use of the app for ≥90 days.
bFor example, stroke, Huntington disease, and unspecified dementia diagnosis.
cParticipants who were not diagnosed at the time of inclusion.
dHigher scores indicate higher attainment. The scores on Mini-Mental State Examination (MMSE) are the most recent score documented in the participants’ medical record. Data on months between the latest MMSE score and study inclusion showed MMSE were, on average, conducted 9 months before inclusion (SD 10; range 0-47), and 89 (89/112, 79.4%) of the MMSEs were conducted less than 12 months before inclusion.
eIncludes all levels of caregiver adoption status.
The details of adoption status and period of adherence are specified in
Caregivers had also activated the app in 44% of the 18 cases where the participants became adopters, and in 3 of these cases, both the participant and caregiver were adopters. In 4 additional cases, the caregiver became adopter without the participant becoming adopter.
As summarized in
An exploratory logistic regression analysis was performed to assess the impact of baseline characteristics and caregiver’s app activities on participant adopter status. As shown in
Data from the survey provided additional information on adopters and nonadopters. As summarized in
The survey gave opportunity to further explore reasons why participants did not use the app or become an adopter. As outlined in
Adoption status and period of adherence among participants and caregivers.
Participants and caregivers | Participants (N=112) | Caregivers (N=98) | |
|
|||
|
Never used the app (0 days) | 47 (42) | 78 (80) |
|
|||
|
Short use (1-10 days) | 19 (17) | 7 (7) |
Early abandonment (11-31 days) | 11 (10) | 2 (2) | |
Late abandonment (32-89 days) | 17 (15) | 4 (4) | |
Adopter (≥90 days) | 18 (16) | 7 (7) |
Logistic regression: Impact of baseline characteristics and caregiver’s app activities on participant adoption status.
Included | Beta (SE) | Odds ratio (95% CI) | |
Participant’s age | −.03 (0.03) | .29 | 0.97 (0.91-1.03) |
Participant’s gender | −.27 (0.60) | .65 | 0.76 (0.24-2.47) |
Months since diagnosis | −.01 (0.02) | .74 | 0.99 (0.96-1.03) |
Mini-Mental State Examination score | −.09 (0.06) | .17 | 0.92 (0.81-1.04) |
Caregiver adoption status | .13 (1.01) | .90 | 1.1 (0.16-8.24) |
Caregiver activated the app | 1.6 (0.70) | .02 | 5.1 (1.29-19.99) |
Constant | 2.5 (2.83) | .29 | 12.4 |
Baseline characteristics of participants who replied to the survey compared with participants who did not reply to the survey.
Characteristics | Participants who replied to survey (N=35) | Participants who did not reply to survey (N=77) | |||
Age (years), mean (SD), range | 68 (9.7), 52-86 | 69 (8.5), 39-83 | .65 | ||
|
.84 | ||||
|
Women | 15 (46) | 33 (43) |
|
|
|
.92 | ||||
|
≤10 | 6 (18) | 10 (13) |
|
|
11-12 | 6 (18) | 16 (21) |
|
||
13-14 | 6 (18) | 19 (25) |
|
||
15-16 | 9 (27) | 19 (25) |
|
||
≥17 | 6 (18) | 13 (17) |
|
||
|
.33 | ||||
|
Alzheimer disease | 21 (64) | 43 (56) |
|
|
|
Vascular dementia | 0 | 2 (3) |
|
|
|
Dementia with Lewy bodies | 1 (3) | 0 (0) |
|
|
|
Frontotemporal dementia | 0 | 3 (4) |
|
|
|
Mild cognitive impairment | 4 (12) | 5 (7) |
|
|
|
Othera | 6 (15) | 21 (27) |
|
|
|
Unresolvedb | 2 (6) | 3 (4) |
|
|
Months since diagnosis, mean (SD), range | 9 (12), 0-61 | 14 (15), 0-73 | .06 | ||
Mini-Mental State Examination scorec, mean (SD), range | 26 (4.1), 11-30 | 24 (4.4), 11-30 | .12 |
aFor example, Huntington disease or stroke.
bParticipants who were not diagnosed at the time of inclusion.
cHigher scores indicate higher attainment. The scores on the Mini-Mental State Examination are the most recent scores documented in participants’ medical records.
Data from survey: participant’s and caregiver’s rating of the participant’s level of experience, skills, and need for help when using a tablet.
Participants’ and caregivers’ by-proxy ratinga | Adopterb | Nonadopterb | ||||||
|
||||||||
|
|
.49 | ||||||
|
|
Much experience | 2 (15) | 5 (28) |
|
|||
Some experience | 7 (54) | 5 (28) |
|
|||||
Little experience | 1 (8) | 4 (22) |
|
|||||
Novel user | 3 (23) | 4 (22) |
|
|||||
|
|
.15 | ||||||
|
|
Uncomplicated | 7 (54) | 7 (44) |
|
|||
A little difficult | 5 (38) | 2 (12) |
|
|||||
Quite difficult | 1 (8) | 5 (31) |
|
|||||
Very difficult | 0 | 2 (12) |
|
|||||
|
|
.54 | ||||||
|
|
No help | 5 (38) | 6 (33) |
|
|||
A little help | 6 (46) | 5 (28) |
|
|||||
Some help: | 1 (8) | 3 (17) |
|
|||||
A lot of help | 1 (8) | 4 (22) |
|
|||||
|
||||||||
|
|
.41 | ||||||
|
|
Much experience | 3 (38) | 4 (19) |
|
|||
Some experience | 3 (38) | 7 (33) |
|
|||||
Little experience | 1 (12) | 7 (33) |
|
|||||
Novel user | 1 (12) | 3 (14) |
|
|||||
|
|
.38 | ||||||
|
|
Uncomplicated | 3 (37) | 6 (30) |
|
|||
A little difficult | 3 (37) | 3 (15) |
|
|||||
Quite difficult | 2 (35) | 6 (30) |
|
|||||
Very difficult | 0 | 5 (25) |
|
|||||
|
|
.28 | ||||||
|
|
No help | 2 (25) | 3 (14) |
|
|||
A little help | 4 (50) | 8 (36) |
|
|||||
Some help | 2 (25) | 4 (18) |
|
|||||
A lot of help | 0 | 7 (32) |
|
aRatings cannot be compared between participants and caregivers because participant’s and caregiver’s rating does not refer to parallel cases.
bAnswers were not complete to all questions; hence, the total number of answers on each question varied for both participants and caregivers.
In those cases where the app had not been activated by the participant or the person had stopped using it, the survey gave opportunity to comment on reasons for not using or adopting the app. These comments are summarized in themes (
Theme: participant intends to start using the app
Participant quotes
I want to use the app, I just need to learn how to use it.
I will try to install the app.
I think I will start using it, the calendar would be nice to have.
Theme: using the app was found premature
Participant quotes
It’s not relevant for me...yet.
Caregiver quotes
It was too simple for him
I think we will wait until it is the right time to use it.
Theme: participant does not use the app because of dementia symptoms
Participant quotes
I forget to use it.
I don’t think about using it...I forget.
Frankly, I have forgotten that I have ever heard of it.
Caregiver quotes
He is not able to use the iPad anymore.
She has opened the app a few times, she but does not understand how to use it. It is too late to introduce it.
The problem is my wife forgets to use it.
Our dad probably got the app too late.
She feels under surveillance when using the app.
Theme: participant prefer to use other technology solutions
Participant quotes
I am happy with the things I use already.
I use a computer for e-mails, calendar and things like that...I have an iPad, but I haven’t started using it yet.
When I read more about what the app could do I realised that I was covered by the things I already use, and I like using the things that I already know.
I prefer to use the calendar that is installed on my phone.
When I read more about what the app could do I realised that I was covered by the things I already use, and I like using the things that I already know.
I prefer to use the calendar that is installed on my phone.
Caregiver quotes
He prefers to use the calendar which is installed on the iPad.
Theme: participant prefer to use nontechnology solutions
Participant quotes
It is easier for me to use a paper diary.
I prefer my paper diary.
Caregiver quotes
She finds it easier to use an ordinary paper diary.
To enable a more detailed analysis of use patterns of those participants who activated the app (N=65), they were split into 4 groups based on the number of days they had been users of the app (number of days from the first to last activity in the app, with a maximum of 90 days): short use (1-10 days), early abandonment (11-31 days), late abandonment (32-90 days), and adopter (≥90 days). These are summarized in
The number of activities in the app for the 4 groups of participants who had activated the app is summarized in
To further assess if the content of the app was relevant to users or if adaptions were needed, data from adopters were further analyzed. The number of times each functionality was used by a participant or caregiver is illustrated in
Number of activities in the app (maximum 90 days) for all participants who activated the app.
Log data from adopters illustrating what app functionalities were used.
The level of satisfaction with the app could be analyzed based on data from the USEdem questionnaire. Analysis of results (not illustrated) showed an overall average score of 40 (SD 9.4; range 21-55) for participants and 34 (SD 12; range 18-51) for caregivers, which indicated a generally positive rating of the app with regard to usefulness, satisfaction, and ease of use, but with large variation. For the participants, the score on the USEdem questionnaire was significantly higher for adopters (median 46) compared with nonadopters (median 38; U=5.5;
The aims of this study were to investigate the applicability and usability of the ReACT app in a mixed group of people with dementia and to investigate user patterns, factors influencing adoption, and possible impact of caregiver involvement. The study also served as a process evaluation of both the app and methods for implementation.
The overall adoption rate was not high in this study, and whether it can be considered a successful adoption rate or not is hard to estimate since, to the best of the authors’ knowledge, no previous comparable studies have been conducted within the field of dementia research. The relatively low adoption rate is in line with the well-known challenge of nonadoption and nonadherence for digital health interventions across all medical fields, referred to as
According to the study objectives, a heterogeneous sample of people referred to a memory clinic were included in the study. As expected from a mixed population of people with dementia, most of the participants had a diagnosis of Alzheimer disease; however, the prevalence was higher among participants in this study compared with a general Danish population of people referred to a memory clinic [
The relatively young age of participants could reflect an age or generation-related higher frequency of using tablets and apps among middle-aged and younger seniors compared with older generations [
The MMSE scores, indicating the level of cognitive function, were generally higher among participants than in a general Danish population of people referred to a memory clinic [
As illustrated in
In general, research is needed to explore how long-term adherence to technology can be supported among people with dementia and to define realistic goals for long-term adherence among this group. It is also important to consider whether progressive cognitive symptoms might imply that aiming for long-term adherence is too ambitious. Various models and frameworks have been proposed for successful innovation, design, and implementation of AT [
Our results clearly indicated that caregiver involvement in using the app could influence participants’ adoption of it. These results are in line with other studies finding caregiver engagement important for the everyday use of technology among people with dementia [
The benefits of caregiver involvement also highlight the need to provide information and guidance to caregivers on how they can best support people with dementia in using AT. In this study, this was done by providing written and Web-based material giving advice on caregiver involvement. The need to consider methods to support caregiver involvement has also been discussed by others [
It is important to acknowledge, however, that in some cases, participants adopted the app and were high-frequency users with minimal or no caregiver involvement, again indicating a large variation among adopters and stressing that the person with dementia should be addressed as the main user of AT for people with dementia.
The regression analysis showed that none of the participants’ background characteristics could predict participants’ adoption status, and overall, the regression to predict adoption was relatively low in predictive power (Nagelkerke
Data from the surveys revealed that there were no significant differences between adopters and nonadopters when it came to how much experience they had using a tablet, their skills when using it, and how much help they needed to use it. Some of the adopters were even characterized as novel users, indicating that the app is applicable for a varied group of users and can be used despite having a low level of tablet skills, which has also been demonstrated in a previous pilot study [
Use patterns among all participants who activated the app are illustrated in
The results from the USEdem questionnaire revealed a relatively high satisfaction with the app among participants and caregivers, but with some variation, and participants who became adopters rated it significantly higher than nonadopters. Results from caregivers were generally less positive, and this should, of course, be investigated further. However, data quality on this questionnaire was limited because of the small number of replies.
Log data showed that all functionalities in the app were used, as illustrated in
This study has several limitations that should be acknowledged. Participants were recruited from memory clinics, and consequently, only people who have sought an examination and had contact with a memory clinic were included. This limitation will be addressed in a subsequent study with open access to the app. In addition, staff could have been biasing inclusion. Although information on the study was generally available in the clinics, inclusion was also promoted by staff, and they could have been directing information to subgroups of participants, based on common ideas of who can benefit from using AT (eg, younger participants or those with mild cognitive symptoms).
In addition, a number of participants did not have a dementia diagnosis; hence, describing participants as a group of people with dementia could be considered imprecise. There was also a quite large proportion of participants who were categorized as
Another limitation was that the app could only be used on a series of tablets (iPads) during this study, and this, of course, excluded potential participants who had access to other kinds of tablets. For practical and financial reasons, a specific type of tablet had to be selected when designing the app for this study, and the selection was based on various factors (eg, the iPad was the most common tablet in Denmark) [
Data included in this study were generally rich and provided interesting results and observations, but there were also limitations in relation to data quality and quantity. Disease-related information was only obtained from medical records, and no pre-post measures were applied. This could be changed in future studies. However, as shown by previous studies [
The study was designed to provide separate log data from participants and caregivers; however, this does not reveal all details on how the app is used in real life, for example, caregiver support could be more intense than revealed by current data. In future studies, richer data can be obtained by including more qualitative data (eg, by interviewing participants and caregivers). The aim of the study was, however, to obtain detailed data from a large group of potential users of the app, and this could be obtained by the mixed method design, which is also used to evaluate the use of apps in other fields of health care [
The results from this study were in line with the general well-known challenges of nonadoption and nonadherence to digital health interventions. However, for those who adopted the app, results showed that the ReACT app was applicable and useful for a mixed population of people with dementia and that the methods used for deployment and self-applied implementation were applicable for this group of end users. The study clearly demonstrated the benefits of applying mixed methods when accessing applicability, usability, and effectiveness of AT. The analysis of detailed log data contributed valuable insights into use patterns and allowed a detailed analysis of factors influencing adoption. In addition, it provided detailed data used in the process evaluation and validation of the ReACT app. To the best of the authors’ knowledge, this study is the first to include such a large and rich dataset from the everyday use of AT among people with dementia and their family caregivers.
The results from the study revealed factors that could influence the adoption of AT among people with dementia. Timely introduction of AT and support from caregivers had significant influence on whether participants adopted the ReACT app. However, data also revealed great variation among adopters when it came to personal, disease-related, and contextual factors, and the predictive value of caregiver involvement was small. This underlines that adoption of AT among people with dementia is influenced by a complex set of personal and contextual factors, which is made even more complex by the changing needs imposed by living with a progressive dementia disease. This complexity and variability restrain the extent to which adoption and adherence to AT can be predicted and emphasize the importance of incorporating this wide range of changeable factors when designing and implementing AT for people with dementia.
assistive technology
electronic health
information and communication technology
Mini-Mental State Examination
Rehabilitation in Alzheimer's disease using Cognitive support Technology
The authors are grateful for the cooperation and engagement of all participants, family caregivers, and professionals involved in the study, including staff from the memory clinics at Aalborg University Hospital, Aarhus University Hospital, Odense University Hospital, Roskilde University Hospital, Slagelse Hospital, Svendborg Hospital, and Rigshospitalet, where participants were recruited.
Danish Dementia Research Centre is supported by the Danish Ministry of Health.
None declared.