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 http://aging.jmir.org, as well as this copyright and license information must be included.
Wearable technology for fall alerts among older adult care recipients is one of the more frequently studied areas of technology, given the concerning consequences of falls among this population. Falls are quite prevalent in later life. While there is a growing amount of literature on older adults’ acceptance of technology, less is known about how caregivers’ attitudes toward technology can impact care recipients’ use of such technology.
The objective of our study was to examine associations between caregivers’ attitudes toward technology for caregiving and care recipients’ use of fall alert wearables.
This study examined data collected with an online survey from 626 caregivers for adults 50 years and older. Adapted from the technology acceptance model, a structural equation model tested the following prespecified hypotheses: (1) higher perceived usefulness of technologies for caregiving would predict higher perceived value of and greater interest in technologies for caregiving; (2) higher perceived value of technologies for caregiving would predict greater interest in technologies for caregiving; and (3) greater interest in technologies for caregiving would predict greater use of fall alert wearables among care recipients. Additionally, we included demographic factors (eg, caregivers’ and care recipients’ ages) and caregiving context (eg, caregiver type and caregiving situation) as important predictors of care recipients’ use of fall alert wearables.
Of 626 total respondents, 548 (87.5%) with all valid responses were included in this study. Among care recipients, 28% used fall alert wearables. The final model had a good to fair model fit: a confirmatory factor index of 0.93, a standardized root mean square residual of 0.049, and root mean square error of approximation of 0.066. Caregivers’ perceived usefulness of technology was positively associated with their attitudes toward using technology in caregiving (b=.70,
This study underscores the importance of caregivers’ attitudes in care recipients’ technology use for falls management. Raising awareness and improving perception about technologies for caregiving may help caregivers and care recipients adopt and better utilize technologies that can promote independence and enhance safety.
By 2035, adults 65 years and older in the United States are projected to outnumber children (under 18 years), mostly due to the continued aging of the Baby Boomer generation [
According to the American Association of Retired Persons 2020 Report: Caregiving in America [
Wearable technology for fall alerts among older adult care recipients is one of the more frequently studied areas of technology, given the concerning consequences of falls among this population. Falls are quite prevalent in later life; approximately 1 in 4 community-dwelling older adults fall each year, and 20% of falls result in injury [
In recent years, falls have become viewed as preventable with evidence-based programs helping older adults prevent and better manage risk factors associated with falling [
The objective of this paper is to better understand associations between caregivers’ attitudes toward technology for caregiving and care recipients’ use of fall alert wearables.
Based on an adapted framework of the Technology Acceptance Model (TAM) [
We further based our analyses on specific demographic factors and caregiving contexts available from the data in our survey. In addition, our analyses were based on the following subhypotheses, supported in the literature: (1) younger age among caregivers would predict greater perceived usefulness, perceived value, and interest in using technology; (2) more demanding caregiving situations such as longer caregiving hours and dementia among care recipients would increase caregivers’ interest in technology; (3) older age among care recipients would predict greater health care needs and fall risks, hence more need for and use of fall alert–related technology [
For this study, we adapted a validated model of technology acceptance by users in organizations, based on the TAM and an updated version (TAM2) [
In this study, we adapted TAM and TAM2 to build and test a framework (
Initially hypothesized model predicting care recipient’s use of fall alert wearables. CG: caregiver; CR: care recipient.
This study used a cross-sectional online survey collected from 626 paid and unpaid caregivers for adults 50 years and older. The caregivers were recruited through an internet panel (Qualtrics XM) in November 2019. Survey respondents were eligible to be included in this study if they were aged 18 years or older, were either paid or unpaid, and provided at least 8 hours of care per week for at least one person who was over 50 years of age and who lived in a home environment. The recruited sample was targeted to resemble the population distribution across 4 US regions (eg, northwest: 17.2%; midwest: 20.9%; west: 23.8%; south: 38.1%) based on 2018 census data [
Caregiver’s perceived usefulness of technologies in caregiving was measured using 6 items on the extent technology helps with (1) reducing the caregiving burden in the future; (2) enabling the care recipient to live more independently; (3) enabling caregiver to have a better quality of life; (4) improving the caregiver’s relationship with their care recipient; (5) improving communication with the care recipient’s family and friends; and (6) improving communications with the care recipient’s health care team. Each item was measured on a 0-to-100-point slider, with higher scores indicating greater perceived usefulness. For the 6 items, Cronbach α=.92. The Kaiser-Meyer-Olkin measure was 0.89, and the Bartlett test of sphericity (χ215=2458.77,
Caregiver’s attitudes toward various safety-related technology for caregiving was assessed by asking perceived value of (1) watches and wearables that enable emergency calls and provide easy to use communications with family members; (2) cameras and alerts to make the house safe; (3) wearable technology to track care recipient health conditions (eg, breathing, pulse, and blood pressure); (4) watches and wearable sensors to monitor and send emergency alerts about falls; (5) watches and sensors that provide care recipient's location; and (6) wearables and sensors that alert if care recipients are at risk for falls. The survey respondents rated perceived value of each technology on a 0-to-100-point slider, with higher scores indicating greater perceived value of the technology in caregiving. For the 6 items, Cronbach α was .91. The Kaiser-Meyer-Olkin was 0.90, and Bartlett test of sphericity was statistically significant (χ215=2130.27,
Two items were used to measure caregiver’s interest in using technology for tracking their care recipient’s location and providing alerts if their care recipient is at risk for a fall. The valid response range for the 2 items was 0 to 100 points, using a slider with higher scores indicating greater interests in using the technology. The Spearman-Brown reliability estimate for the 2 items was 0.75.
The online survey collected sociodemographic characteristics of caregivers and caregiving context, as well as the caregiver's oldest care recipient’s age, dementia diagnosis status, and use of fall alert wearables (eg, pendant or other wearable to alert others that a fall has occurred). Sociodemographic characteristics of caregivers included age in years, gender, race/ethnicity, place of residence (rural vs urban), education (associate degree or less education vs bachelor degree or higher education), employment status (employed for wages or self-employed vs other), previous year’s household income (<US $50,000 vs ≥$50,000), and financial stress (ie, “In general, how do your finances usually work out at the end of the month? Do you find that you usually: end up with some money left over/have just enough money to make ends meet/not have enough money to make ends meet?”). Self-reported zip codes were approximated to the census tract–based rural-urban commuting area codes [
Characteristics of the study’s caregivers, their care recipients, and caregiving contexts, as well as caregivers’ attitudes toward using technology in caregiving, were described using mean and standard deviation or frequency and percentage. Independent group comparison (eg, 2-tailed independent
Revised model predicting care recipient’s use of falls alert wearables. CG: caregiver; CR: care recipient. *
Fewer than 28% (153/548) of the study’s care recipients used a fall alert wearable. In a bivariate analyses comparing caregivers for those who do not use fall alert wearables to those who do use fall alert wearables found that the caregivers of those who used fall alert wearables were significantly younger (
Characteristics of the study respondents and caregiving context and caregivers’ attitudes toward using technology in caregiving.
Characteristic | All (N=548) | Care recipients using fall alert wearables (n=153) | Care recipients not using fall alert wearables (n=395) | |||
Age (years), mean (SD) | 58.1 (14.07) | 53.2 (16.58) | 59.8 (12.90) | <.001 | ||
|
.71 | |||||
Female | 417 (76.2) | 115 (75.2) | 302 (76.6) | |||
Male | 131 (23.8) | 38 (24.8) | 93 (23.4) | |||
|
.005 | |||||
Non-Hispanic White | 354 (65.0) | 82 (53.6) | 272 (69.4) | |||
Non-Hispanic Black | 93 (17.1) | 38 (24.8) | 55 (14.0) | |||
Non-Hispanic Asian | 35 (6.4) | 10 (6.5) | 25 (6.4) | |||
Non-Hispanic other races | 9 (1.7) | 2 (1.3) | 7 (1.8) | |||
Hispanic | 54 (9.9) | 21 (13.7) | 33 (8.4) | |||
|
.87 | |||||
High school or lower | 128 (23.4) | 35 (22.9) | 93 (23.5) | |||
Some college or higher | 420 (76.6) | 118 (77.1) | 302 (76.5) | |||
|
<.001 | |||||
Employed for wages or self-employed | 237 (43.2) | 96 (62.7) | 141 (35.7) | |||
Not employed for wages, not self-employed | 311 (56.8) | 57 (37.3) | 254 (64.3) | |||
|
.72 | |||||
Less than US $50,000 | 279 (50.9) | 76 (49.7) | 203 (51.4) | |||
More than US $50,000 | 269 (49.1) | 77 (50.3) | 192 (48.6) | |||
|
.003 | |||||
End up with some money left over | 246 (45.4) | 79 (52.3) | 167 (42.7) | |||
Have just enough money to make ends meet | 212 (39.1) | 61 (40.4) | 151 (38.6) | |||
Not have enough money to make ends meet | 84 (15.5) | 11 (7.3) | 73 (18.7) | |||
|
.29 | |||||
Rural | 47 (8.6) | 10 (6.5) | 37 (9.4) | |||
Urban | 500 (91.4) | 143 (93.5) | 357 (90.6) | |||
|
||||||
Age (years), mean (SD) | 74.5 (11.93) | 77.2 (12.21) | 73.5 (11.95) | <.001 | ||
|
.01 | |||||
Yes | 128 (23.4) | 47 (30.7) | 81 (20.5) | |||
No | 420 (76.6) | 106 (69.3) | 314 (79.5) | |||
|
||||||
|
<.001 | |||||
Paid caregiver | 116 (21.2) | 68 (44.4) | 48 (12.2) | |||
Unpaid caregiver | 432 (78.8) | 85 (55.6) | 347 (87.8) | |||
Weekly hours of caregivingb, mean (SD) | 37.5 (28.98) | 31.3 (23.83) | 39.3 (30.00) | .002 | ||
|
<.001 | |||||
Yes | 311 (56.8) | 53 (34.6) | 258 (65.3) | |||
No | 237 (43.2) | 100 (65.4) | 137 (34.7) | |||
|
||||||
Perceived usefulness | 58.3 (25.57) | 68.2 (21.94) | 54.5 (25.86) | <.001 | ||
Perceived value | 63.5 (27.22) | 73.6 (20.48) | 59.5 (28.48) | <.001 | ||
Interest | 59.2 (30.40) | 72.6 (26.02) | 54.0 (30.40) | <.001 |
aResults from unadjusted independent group comparison between the group, in which care recipients use fall alert wearables, and another group, in which care recipients do not use fall alert wearables.
bTotal weekly hours of caregiving was capped at 100 hours.
cValues ranged from 0 to 100, with a higher value indicating greater perceived usefulness, greater perceived value, or more interest in using technology in caregiving.
Goodness of fit, of the model shown in
Direct, indirect, and total effects of each predictor on care recipient’s use of fall alert wearables.
Variable | Direct effects | Indirect effects | Total effects | |||
ba (SE) | b (SE) | b (SE) | ||||
Caregivers’ perceived usefulness of technology in caregiving | 0 | N/Ab | .18 (0.03) | <.001 | .18 (0.03) | <.001 |
Caregivers’ attitudes toward using technology in caregiving | 0 | N/A | .17 (0.03) | <.001 | .17 (0.03) | <.001 |
Caregivers’ interests in using technology in caregiving | .27 (0.04) | <.001 | 0 | N/A | .27 (0.04) | <.001 |
Caregivers’ age | 0 | N/A | –.03 (0.009) | .003 | –.03 (0.009) | .003 |
Unpaid caregiver (vs paid caregiver) | –.33 (0.04) | <.001 | 0 | N/A | –.33 (0.04) | <.001 |
Hours of caregiving | 0 | N/A | –.02 (0.01) | .046 | –.02 (0.01) | .046 |
Care recipient having dementia | 0 | N/A | .03 (0.01) | <.001 | .03 (0.01) | <.001 |
Care recipients’ age | .11 (0.04) | .004 | 0 | N/A | .11 (0.04) | .004 |
aStandardized estimates.
bN/A: not applicable.
As shown in
The model fit among paid caregiver was fair (CFI 0.93, SRMR 0.076, and RMSEA 0.062) and was comparable to the comprehensive model (CFI 0.93, SRMR 0.049, and RMSEA 0.066). The 3 prespecified hypotheses remained statistically significance, and corresponding path coefficients were comparable to the comprehensive model (
The model fit among unpaid caregivers was good to fair (CFI 0.93, SRMR 0.051, and RMSEA 0.069) and was comparable to that of the comprehensive model (CFI 0.93, SRMR 0.049, and RMSEA 0.066). All path coefficients remained statistically significant, and path coefficients were comparable to the comprehensive model (
For both paid and unpaid caregivers, caregiver’s interest in using technology for caregiving had the strongest positive effects on care recipient’s use of fall alert wearables (b=.21,
From our analyses, we have demonstrated that the adapted TAM2 concepts of caregivers provide support for our hypotheses about care recipients’ use of fall alert wearables, which is reflective of previous literature [
Our results demonstrated that the strongest predictor of care recipients’ use of fall alert wearable was the type of caregiver and that care recipients with paid caregivers were more likely to use this type of technology than care recipients with unpaid caregivers. While not expected, this may reflect the scenario where the path of caregiving for older adults typically begins with a family member or unpaid caregiver who lives in close proximity to the care recipient and provides human monitoring. Concerns for falls often results in investment in fall alert wearables for older adults living independently [
Our subhypothesis that more demanding caregiving situations, including longer hours of caregiving and instances of dementia among care recipients, was partially supported in this study. As hypothesized, dementia among care recipients positively predicts their use of fall alert wearables. However, contrary to our hypothesis, fewer caregiving hours was associated with care recipient’s use of fall alert wearables. A potential interpretation may be that caregivers providing fewer hours of care could be more inclined to use wearables to compensate for longer durations of nonsupervised time. According to the subgroup analyses based on caregiver’s payment status, the statistical significance of the subhypotheses are likely to be largely driven by unpaid care recipients, who constituted almost 79% (432/548) of the total analytic sample. While caregiver’s attitude toward technology in caregiving were significantly associated with care recipient’s use of fall alert wearables in both paid and unpaid caregivers, caregivers’ and care recipients’ age, and caregiving situations were significantly associated with care recipients’ use of fall alert wearables only among unpaid caregiver participants. The smaller sample size of paid caregivers may have limited the statistical power of the model. Another potential explanation is the differential involvement of paid and unpaid caregivers in caregiving decisions [
There is relatively little research that examines how caregivers and their care recipients (either paid or unpaid) actually use fall alert technology in their everyday lives or how such experiences may affect their safety and well-being. The little research that exists is limited in scale, often focused on care recipients with dementia and on cross-sectional interview methodologies focused on the adoption of the wearable fall alert technology [
There were some limitations to our study. First, our caregiver population in the panel-based survey may not be representative of the caregiver population across the United States, despite our best efforts. While we have used quota sampling to match the distribution of key characteristics (eg, geographical region, age, gender, and race), this online sample excludes caregivers without access to internet and related technology (eg, computer, smartphone, or tablets). We assume that respondents were more willing to sign up to participate because they are comfortable with technology. Thus, caregivers who do have online access but are not as comfortable with technology may have elected not to participate. We also excluded caregivers who might have had online access but who had limited English proficiency. While we asked participants to self-identify as either paid or unpaid, there was no way to tell if there were subsets of unpaid caregivers who received some sort of stipend or benefit. With our cross-sectional design, it was not possible to draw conclusions about the causality between attitudes, caregiving contexts, and use of fall alert wearables. Additionally, the proposed model is limited by lack of potential factors, such as perceived ease of use for specific technology, fall history, and interpersonal relationships between caregivers and care recipients. In addition, the care recipients’ use of fall alert wearables were proxy-reported by caregivers, a further study using direct observation or self-reported measure by care recipients could supplement the proxy-reported evidence. More information on the types of technology and how the specific technologies are used would help establish circumstantial data to set out recommendations for practice and policy. Future research using in-depth interviews with caregivers to explore the nuances of technology adoption would be instructive for understanding more about the context driving our quantitative research findings. Despite these limitations, we believe our data and analyses provide important new information on how caregivers’ attitudes and values about technology influence adoption about the use of fall alert wearables for the protection and safety of their care recipients.
With this study, we have taken a small step in addressing the knowledge gap about how caregiver attitudes affect adoption of assistive intelligent technology such as wearable fall alert technologies in caregiving, but much remains to be learned. With the growth of the aging population over the forthcoming years, and the anticipated rise of the occurrence of falls and related injuries based on the increasing numbers of older Americans, the caregiving workforce will benefit from advanced and effective technologies used in caregiving. It will continue to be crucial for public health researchers to keep pace with the advances of technology and maintain an advocacy role for both caretakers and care recipients in the adoption and use of technology to support their health and wellbeing.
Characteristics of the study respondents based on caregivers’ paid status.
Revised model predicting paid care recipient’s use of fall alert wearables.
Revised model predicting unpaid care recipient’s use of fall alert wearables.
confirmatory factor index
root mean square error of approximation
standardized root mean square residual
Technology Acceptance Model
We thank all of the caregivers who participated in our online survey which formed the basis of this research. The caregiver survey was funded by contributions from DVD Associates LLC, Clairvoyant Networks Inc, and The Texas A&M Center for Population Health and Aging.
DVD was employed by DVD Associates LLC and is an editor for JMIR Cancer. SP is employed by Clairvoyant Networks Inc. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.