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 Aging, is properly cited. The complete bibliographic information, a link to the original publication on https://aging.jmir.org, as well as this copyright and license information must be included.
Older adults with disabilities such as loss of autonomy face the decision of whether to stay at home or move to a health care facility such as a nursing home. Therefore, they may need support for this difficult decision.
We assessed the intention of Canadian older adults to use an electronic decision aid (eDA) to make housing decisions and identified the factors that influenced their intention.
We conducted a cross-sectional study using a web-based survey targeting older adults across 10 Canadian provinces and 3 territories. We included respondents from a web-based panel who were aged ≥65 years, understood English or French, had access to an electronic device with an internet connection, and had made a housing decision over the past few months or were planning to make a decision in the coming year. We based the web-based survey on the Unified Theory of Acceptance and Use of Technology (UTAUT). We adapted 17 UTAUT items to measure respondents’ intention to use the eDA for housing decisions, as well as items measuring 4 intention constructs (performance expectancy, effort expectancy, social influence, and facilitating conditions). We also assessed eHealth literacy using both subjective and objective scales. We used descriptive statistics and multivariable linear regression analyses to identify the factors influencing the intention to use the eDA.
Of the 11,972 invited panelists, 1176 (9.82%) met the eligibility criteria, and 1000 (85.03%) respondents completed the survey. The mean age was 72.5 (SD 5.59) years. Most respondents were male (548/1000, 54.8%), White (906/1000, 90.6%), English speakers (629/1000, 62.9%), and lived in Ontario or Quebec (628/1000, 62.8%) and in urban areas (850/1000, 85%). The mean scores were 27.8 (SD 5.88) out of 40 for subjective eHealth literacy and 3.00 (SD 0.97) out of 5 for objective eHealth literacy. In our sample, the intention score was 4.74 (SD 1.7) out of 7. The mean scores of intention constructs out of 7 were 5.63 (SD 1.28) for facilitating conditions, 4.94 (SD 1.48) for performance expectancy, 5.61 (SD 1.35) for effort expectancy, and 4.76 (SD 1.59) for social influence. In the final model, the factors associated with intention included mother tongue (β=.30;
Findings from this pan-Canadian web-based survey on Canadian older adults suggest that their intention to use the eDA to make housing decisions is similar to the findings in other studies using UTAUT. The factors identified as influencing intention were mother tongue, objective eHealth literacy, performance expectancy, social influence, and facilitating conditions. These will guide future strategies for the implementation of the eDA.
As in many other countries, older adults in Canada (ie, persons aged ≥65 years) are a rapidly growing segment of the population [
To manage loss of autonomy, meet health care and social services needs, and ensure their safety and well-being, many Canadian older adults consider receiving home care, which typically includes nursing care, therapy (physical, occupational, and speech-language), and medical and social services [
In Canada, housing decisions are considered the most frequent and difficult decisions for older adults receiving home care as well as for their caregivers [
Housing decisions and transitioning to long-term care can be experienced differently, depending on the sociocultural context. For instance, in Western cultures, some residential care facilities try to create a homelike atmosphere by allowing older adults to bring their furniture, pets, and family pictures to help ease the transition [
Amid the COVID-19 pandemic, housing decisions have become not only more frequent but also more painful and complicated for older Canadians [
To help older adults make informed decisions regarding the most appropriate housing option, shared decision-making (SDM) is advocated. SDM is the process of making a health care choice that involves patients, their relatives or family or both, and one or more health care professionals [
eHealth refers to health services and information delivered through the internet and related technologies [
In a clinical setting, a DA is usually presented before the clinical encounter to prepare the patient for SDM with a health professional or during the encounter to prepare for a subsequent encounter. Few health professionals have the time to work through a DA with the patient and come to a conclusion on the spot [
We hypothesized that older adults would find the eDA to be useful. To our knowledge, no study has yet investigated whether older adults would be willing to use the eDA for housing decisions. Therefore, our aim was to assess Canadian older adults’ intention to use the eDA to make housing decisions and to identify the factors influencing their intention to use it.
We conducted a cross-sectional web-based survey across Canada (including the 10 provinces and 3 territories) with older adults who had either made a housing decision in the past 12 months or were planning to make a housing decision the next year. We used the consensus-based Checklist for Reporting of Survey Studies (CROSS) to guide the reporting of our results [
This project was approved by the Ethical Review Board of the Integrated University Health and Social Services Centre of the Capitale-Nationale, Quebec, Canada (#MP-13-2019-1519, 2019-1519_SPPL).
Health-related behaviors are correlated with intention, which is defined as an individual’s planned and rationalized decision to perform the behavior [
Adapted version of Unified Theory of Acceptance and Use of Technology.
Respondents were eligible if they were Canadian adults aged ≥65 years, understood English or French, had access to an electronic device with an internet connection, and had made a housing decision in the past few months or were planning to make one in the coming year.
We recruited respondents through Leger Marketing, a market research and analytics company in Montreal, Canada. Leger Marketing is the largest private Canadian web-based panel (400,000 individuals) and claims to be representative of the entire population [
Each respondent from the web-based panel received a personalized email invitation containing a unique URL link to access the nonopen survey. Respondents were then asked to answer the first questions about language preference, province or territory of residence, and eligibility. Leger Marketing sent reminders via email once a week, until the survey was closed. As respondents logged into the survey using their panel membership account, we had a unique response per member because it was not possible for the same member to have multiple submissions.
A minimum of 829 participants were required. The sample size was estimated using the central limit theorem formula [
Because no validated instruments that assess older adults’ intention to use an eDA for housing decisions have been identified, we created a self-administered questionnaire based on the adapted UTAUT items. We measured our main outcome, intention, and its 4 determining constructs using the 17 UTAUT-based items. Each UTAUT construct (intention, performance expectancy, effort expectancy, social influence, and facilitating conditions) was measured using 3 or 4 items. Respondents indicated their agreement or disagreement levels with the corresponding items on a 7-point Likert scale ranging from 1 (
The survey also collected sociodemographic characteristics (ie, age, sex, gender, education, province or territory of residence, postal code, ethnicity, marital status, number of people in the household, mother tongue, and family income) using items based on Statistics Canada’s 2021 census questionnaire [
We evaluated eHealth literacy using 2 scales. The first was the Electronic Health Literacy Scale (eHeals) [
After completing the 2 eHealth literacy scales, the respondents were shown a 6-minute video vignette showing the use of the eDA in context. As mentioned by Godin et al [
The survey was 48 web pages long, took approximately 30 minutes to complete, and consisted of 50 closed-ended questions that were not randomized and appeared in the same order for all respondents. Each page included a “next” button for moving forward and a button with a list of older adult helplines for talking to a specialist who could support them mentally or emotionally if they were uncomfortable with any of the survey questions. Respondents could not move to the next page unless they had completed all the questions on the current page. Surveys were labeled as complete only if respondents had clicked on the “finish” button located at the end of the survey. Both English and French versions were pretested with a sample of 76 respondents to identify any possible ambiguity or technical problems, validate the clarity of the questions, and estimate the average completion time. No major revisions were made following the pretest.
We determined the distribution of our population for sociodemographic variables, levels of eHealth literacy, UTAUT constructs, and decision-making process variables using descriptive statistics (means, SDs, and percentages). Because intention scores could vary between 1 and 7, we interpreted intention as a continuous variable. There is no definitive threshold for a clinically significant intention score in the literature. We used the Shapiro-Wilk test to verify whether the distribution of the dependent variable was normal.
We considered the “prefer not to answer” choice as missing data (1.8%, 18/1000) for bivariate and multivariable analyses, except for the income variable. We calculated the age of the respondents by considering their date of birth and date of survey completion. We performed a mixed linear regression model including all the independent variables, that is, age, sex, gender, education, province or territory of residence, postal code, ethnicity, marital status, number of people in the household, mother tongue, family income, eHealth literacy (objective and subjective), performance expectancy, effort expectancy, social influence, and facilitating conditions, using stepwise selection with the Bayesian Information Criterion [
Of the 11,972 panelists who were invited to participate, 3789 (31.65%) panelists clicked on the survey link received by email; 1176 (31.04%) met the eligibility criteria; and 1000 (85.03%) respondents completed the entire survey and were included in the analysis (
The included respondent’s characteristics are listed in
Flow of respondents.
Respondents’ characteristics (n=1000).
|
Respondents | ||
Age (years), mean (SD) | 72.5 (5.6) | ||
|
|||
|
Male | 548 (54.8) | |
|
Female | 452 (45.2) | |
|
|||
|
Man | 546 (54.6) | |
|
Woman | 454 (45.4) | |
|
|||
|
A university certificate, diploma, or degree (eg, bachelor’s degree, degree in medicine, dentistry, and veterinary medicine) | 420 (42) | |
|
A college, CEGEPa, or other nonuniversity certificate or diploma (other than trade certificates or diplomas) | 264 (26.4) | |
|
A high school (secondary school) diploma or equivalent, a registered apprenticeship, or other trade certificate or diploma | 286 (28.6) | |
|
Lower than a high school (secondary school) diploma or equivalent (eg, primary school) | 25 (2.5) | |
|
I prefer not to answer | 5 (0.5) | |
|
|||
|
Ontario | 377 (37.7) | |
|
Quebec | 251 (25.1) | |
|
Western Canada: British Columbia, Alberta, Saskatchewan, Manitoba, and Yukon | 295 (29.5) | |
|
Eastern Canada: New Brunswick, Nova Scotia, Prince Edward Island, and Newfoundland and Labrador | 77 (7.7) | |
|
|||
|
Urban | 850 (85) | |
|
Rural | 141 (14.1) | |
|
I prefer not to answer | 9 (0.9) | |
|
|||
|
White | 906 (90.6) | |
|
Non-White | 73 (7.3) | |
|
Indigenous peoples of North America (First Nations, Métis, or Inuk [Inuit]) | 18 (1.8) | |
|
I prefer not to answer | 3 (0.3) | |
|
|||
|
Legally married (and not separated) | 516 (51.6) | |
|
Divorced | 152 (15.2) | |
|
Widowed | 138 (13.8) | |
|
Never legally married | 93 (9.3) | |
|
In a common-law union | 81 (8.1) | |
|
Separated, but still legally married | 19 (1.9) | |
|
I prefer not to answer | 1 (0.1) | |
Number of people in the household, mean (SD) | 1.80 (0.81) | ||
|
|||
|
English | 629 (62.9) | |
|
French | 283 (28.3) | |
|
Otherb | 88 (8.8) | |
|
Aboriginal languagesc | 0 (0) | |
|
|||
|
CAD $100,000 (US $76.923) or more | 172 (17.2) | |
|
CAD $75,000-$99,999 (US $57.700-US $76.922) | 153 (15.3) | |
|
CAD $50,000-$74,999 (US $38.461-US $57.692) | 221 (22.1) | |
|
CAD $25,000-$49,999 (US $19.230-US $38.460) | 262 (26.2) | |
|
<CAD $25,000 (US $19.230) | 114 (11.4) | |
|
I prefer not to answer | 78 (7.8) | |
eHealth literacy (subjective)d, mean (SD) | 27.8 (5.88) | ||
eHealth literacy (objective)e, mean (SD) | 3.00 (0.97) |
aCEGEP: Collège d'enseignement général et professionnel.
bOther: Spanish, Mandarin, Arab, Cantonese, Dutch, Flemish, German, Greek, Gujarati, Hindi, Hungarian, Igbo, Indo, Italian, Lithuanian, Polish, Portuguese, Punjabi, Romanian, Russian, Serbian, Slovak, Slovenian, Tamil, Ukrainian, and Urdu.
cAboriginal languages in Canada: Algonquian languages (eg, Cree, Ojibway, Innu or Montagnais, and Mi’kmaq), Inuit languages, Athabaskan languages, Salish languages, Siouan languages, Iroquoian languages, Tsimshian languages, Wakashan languages, Michif, Haida, Tlingit, and Kutenai.
dSum of 8 items on a 1 to 5 Likert scale (1=strongly disagree and 5=strongly agree). Scores range from 8 to 40.
eSum of 5 items (score 0 if wrong answer, score 1 if correct answer). Scores range from 0 to 5.
The UTAUT construct scores are shown in
Cronbach α values are presented for each assessed construct and are a measure of internal consistency for each construct. This is considered to be a measure of scale reliability.
The intention scores associated with each decision-making process variable are listed in
Unified Theory of Acceptance and Use of Technology (UTAUT) construct scores (n=1000).
UTAUT constructa | Scores, mean (SD) | Cronbach α |
Intention | 4.74 (1.70) | .95 |
Performance expectancy | 4.94 (1.48) | .94 |
Effort expectancy | 5.61 (1.35) | .95 |
Social influence | 4.76 (1.59) | .95 |
Facilitating conditions | 5.63 (1.28) | .90 |
aAveraging the scores of the corresponding items on a Likert scale from 1 to 7. The scores range from 1 to 7.
Intention scores associated with decision-making process variables (n=1000).
|
Respondents, n (%) | Intention scores, mean (SD) | ||||
|
||||||
|
|
|||||
|
|
Stay in your home | 736 (73.6) | 4.64 (1.7) | ||
|
|
Move to a family member’s home | 17 (1.6) | 4.29 (1.8) | ||
|
|
Move to a private seniors’ residence | 78 (7.8) | 5.25 (1.5) | ||
|
|
Move to a public residential or long-term care center | 16 (1.6) | 5.37 (1.7) | ||
|
2 chosen options | 88 (8.8) | 4.87 (1.7) | |||
|
3 chosen options | 14 (1.4) | 5.30 (1.6) | |||
|
4 chosen options | 2 (0.2) | 6.00 (0.0) | |||
|
Other option, specify | 49 (4.9) | 4.8 (2.0) | |||
|
||||||
|
Stay in your home | 843 (84.3) | 4.67 (1.7) | |||
|
Move to a family member’s home | 25 (2.5) | 4.79 (1.6) | |||
|
Move to a private seniors’ residence | 75 (7.5) | 5.2 (1.6) | |||
|
Move to a public residential or long-term care center | 17 (1.7) | 5.43 (1.8) | |||
|
Other option, specify | 40 (4) | 4.9 (1.9) | |||
|
||||||
|
|
|||||
|
|
Someone else thinks you should move | 26 (2.6) | 4.93 (1.7) | ||
|
|
You are concerned about your health | 149 (14.9) | 5.11 (1.5) | ||
|
|
You are less able to walk or move around | 44 (4.4) | 3.95 (1.4) | ||
|
|
You do not feel safe | 8 (0.8) | 5.08 (1.8) | ||
|
|
You do not have enough help at home | 21 (2.1) | 5.23 (1.7) | ||
|
|
You feel lonely | 31 (3.1) | 4.83 (1.6) | ||
|
|
You have trouble doing your groceries, getting to the pharmacy, getting to the physician’s office, etc | 28 (2.8) | 4.87 (1.5) | ||
|
|
Your relatives can no longer give you the support you need | 22 (2.2) | 5.0 (1.8) | ||
|
More than 1 reason | 253 (25.3) | 5.05 (1.6) | |||
|
Other option, specify | 418 (41.8) | 4.42 (1.8) | |||
|
||||||
|
Spouse | 271 (27.1) | 4.44 (1.8) | |||
|
Children | 181 (18.1) | 4.83 (1.6) | |||
|
Grandchildren | 5 (0.5) | 4.13 (2.3) | |||
|
Other family member | 53 (5.3) | 4.83 (1.6) | |||
|
Friends | 46 (4.6) | 4.63 (1.6) | |||
|
Physician | 26 (2.6) | 4.14 (1.7) | |||
|
Social worker | 11 (1.1) | 4.82 (2.2) | |||
|
Family and friends only | 211 (21.1) | 4.9 (1.6) | |||
|
Health care team only | 10 (1) | 6.0 (0.9) | |||
|
Both (family, friends, and health care team) | 186 (18.6) | 5.02 (1.6) | |||
|
||||||
|
Active (“I make the decision alone, I make the decision alone but consider the opinion of my relatives and/or health care providers, we decide together with my relatives and/or health care providers, equally”) | 973 (97.3) | 4.73 (1.7) | |||
|
Passive (“My relatives and/or health care providers make the decision but consider my opinion, my relatives and/or health care providers make the decision alone”) | 27 (2.7) | 5.01 (1.6) |
The alternative variable selection approach (ie, the selection of independent variables in the bivariate analyses using the threshold of 0.1 before conducting the multivariable analysis) resulted in the same final model. In total, 9 variables were retained in the bivariate analyses (
Multivariable factors significantly associated with older adults’ intention to use the electronic decision aid.
Variable | Respondents, n (%) | β (95% CI)a | ||
|
||||
|
English (ref)b | 629 (62.9) | N/Ac | N/A |
|
French | 283 (28.3) | .30 (0.17 to 0.43) | <.001 |
|
Other | 88 (8.8) | .06 (−0.11 to 0.28) | .57 |
eHealth literacy (objective) | 1000 (100) | −0.06 (−0.1 to −0.005) | .03 | |
Performance expectancy | 1000 (100) | .55 (0.49 to 0.61) | <.001 | |
Social influence | 1000 (100) | .37 (0.32 to 0.43) | <.001 | |
Facilitating conditions | 1000 (100) | .15 (0.10 to 0.21) | <.001 |
aThe estimated β for each variable and its 95% CI are presented in the table.
bRef: reference category for the analysis.
cN/A: not applicable.
The final proposed model.
To the best of our knowledge, this is the first web-based survey across the 10 Canadian provinces and 3 territories to investigate older adults’ intention to use an eDA for housing decisions. The mean intention score was moderate. In addition, we found that older adults’ most chosen and preferred housing option was to stay in their homes. Most participants had multiple reasons for this preference, which were largely related to their health conditions. Older adults were mostly supported by spouses or children in making their housing decisions, and the majority preferred to play an active role in the decision-making. We also found that intention varied across Canada according to the respondents’ mother tongue. French native speakers were more likely to use the eDA for housing decisions than those with other mother tongues. In addition, objectively evaluated eHealth literacy was negatively associated with intention (ie, a lower level of eHealth literacy was associated with higher intention scores), whereas subjectively evaluated eHealth literacy was not. Finally, the UTAUT constructs of performance expectancy, social influence, and facilitating conditions were significantly and positively associated with intention. In other words, respondents with higher scores for performance expectancy, social influence, and facilitating conditions had a greater intention to use the eDA for housing decisions. These results allowed us to make the observations elaborated in the following sections.
First, scores representing older adults’ intention to use the eDA in this study were positive and similar to the scores in 3 studies using UTAUT model in the context of digital health care, which ranged from 2.8 to 4.42 [
Second, our results suggest that older adults who are supported in their decision-making process by their family, friends, and health care team are more inclined to use the eDA to make housing decisions. Other studies have confirmed the importance of relatives in the decision-making process regarding housing options [
Third, contrary to our expectations, of the 11 sociodemographic variables in the study, only the mother tongue remained in the final model. Our results suggest that francophone Canadians are more inclined to use the eDA than anglophones. This might be because the province of Quebec, where most Canadian French native speakers live, has the highest percentage of older adults living in residential care in the country [
Quebec is considered as a “distinct society” whose culture and social values are different from those in English Canada [
Fourth, even though eHealth literacy, whether measured objectively or subjectively, was associated with intention in the bivariate analysis, only the objective measure of eHealth literacy remained in the multivariable model and seemed to have had a stronger influence on intention. This result confirms the importance of measuring eHealth literacy both objectively and subjectively. Believing oneself to have high literacy levels is not sufficient and needs to be completed with objective performance measurements, which count more in terms of assessing behavioral intentions related to health [
Fifth, as expected, we found that the 3 UTAUT constructs (performance expectancy, social influence, and facilitating conditions) were significantly associated with intention. In other words, the more respondents believed that the eDA would improve the quality of their decision-making, that their social circle would approve of the use of the eDA, and that they had the necessary assistance for using web-based resources, the more they intended to use the eDA to decide about housing. Only the construct effort expectancy was excluded from the final model. Our results are congruent with those of other studies related to eHealth, except for effort expectancy, which was included in their models and not in ours [
Finally, our findings suggest that UTAUT constructs and behavior change methods [
The strength of our study was that this was a rigorous theory-based analysis of the intentions of older adults across Canada, a country that stretches 4700 miles coast to coast, to better support them in making one of their most difficult decisions. Furthermore, Leger Marketing, the survey firm, balanced our recruited sample across age, sex, gender, and socioeconomic status. In addition, the response rate in our study (3789/11,972, 31.65%) was higher than the average for web surveys, which usually ranges from 10% to 20% [
Our study has a few limitations. First, our sample cannot be considered representative of all Canadian older adults because we excluded those with no internet access and most of our respondents were White, English speaking, highly educated, and male. Respondents may have been in a more privileged position than the average Canadian in terms of decisions about housing, that is, they could hire private home care workers or pay for private residential care [
Our study is the first to assess Canadian older adults’ intention to use an eDA to help them make housing decisions. This study makes both empirical and conceptual contributions to the field of eHealth behavior. We were able to provide a better understanding of the relationships between intention and its constructs and examine the effects of various variables on intention. In addition, we propose a modified parsimonious theoretical framework based on UTAUT, involving additional relevant concepts such as eHealth literacy. Research on older adults’ decision-making about housing (eg, eDA development, assessment of intention to use it, and eventually its implementation and integration into various care trajectories) has become increasingly relevant. This study is a step forward toward facilitating eDA implementation and integration initiatives. Our findings and conclusions can be applied in similar sociodemographic contexts where older people are an increasingly large proportion of the population and need support to play an active decision-making role throughout their care continuum.
Details of the conversion of the paper-based decision aid to electronic decision aid.
Consensus-Based Checklist for Reporting of Survey Studies.
Correlation matrix of the continuous variables (age, number of people in the household, eHealth literacy, performance expectancy, social influence, and facilitating conditions).
Factors associated with older adults’ intention to use the electronic decision aid in bivariate analyses.
Checklist for Reporting of Survey Studies
decision aid
Digital Health Literacy Instrument
electronic decision aid
Electronic Health Literacy Scale
shared decision-making
Unified Theory of Acceptance and Use of Technology
We acknowledge Louisa Blair for her precious work and kind editorial help with the manuscript. We also thank Sergio Cortez Ghio for his support with statistical analysis.
This research is part of the COORDINATEs study (project number 9003037412), which is funded by the Joint Programming Initiative More Years, Better Lives, represented by ZonMw in the Netherlands, the Canadian Institutes of Health Research, and Forte. FL holds a tier 1 Canada Research Chair in Shared Decision Making and Knowledge Translation. MF received a Fonds stratégique de développement de la recherche VITAM-centre de recherche en santé durable. The financial providers were not involved in the project.
The data sets used and analyzed during this study are available from the corresponding author upon reasonable request.
MF, KVP, VB, and FL contributed to study design. MF was responsible for data collection. MF, SG, AG, KVP, VB, and FL analyzed the data. MF, VB, AG, and FL contributed to manuscript writing. All authors were responsible for manuscript revision.
FL holds the Canada Research Chair in Shared Decision Making and Knowledge Translation.