Examining Social Determinants in Use of Assistive Technology for Race/Ethnic Groups of Older Adults

Abstract

Objective:

Assistive technologies (AT) can compensate for activity limitations and loss of physical functioning. Little is known regarding how minority older adults differ in AT use as they age. This study examined race and ethnic differences in AT use among a nationally representative sample of older adults in the United States.

Design:

Weighted logistic regression analyses were conducted using the 2012 Behavioral Risk Factor Surveillance System (BRFSS), collected annually by the Centers for Disease Control and Prevention (CDC). The study sample included 282,825 non-Hispanic White, African American, Asian, and non-White Hispanic older adults. Activity limitation, health care access, overall health status and sociodemographic characteristics were included as variables in the analysis. Interaction analyses were conducted to examine the moderating effect of race/ethnicity on social determinants with AT use.

Results:

Results indicated that 13.5% of older adults reported use of an AT. African American older adults had the highest percentage of AT use (21.0%), and Asian older adults had the lowest (5.1%). Those who were 85 years and older, reported an activity limitation, were unmarried and in poor health were most likely to use an AT. Having health insurance was significantly associated with higher AT use for non-Hispanic Whites (OR = 1.66, p < 0.001) and non-White Hispanics (OR = 1.98, p < 0.01), but not African Americans and Asians.

Conclusion:

Health professionals can promote access and address barriers in AT use, particularly in regard to accessibility and acceptability among minority older adults.

With older adults living longer, there is a need to identify resources and supports for those who acquire a disability as they age (Verbrugge, Brown, & Zajacova, 2017). Assistive technology (AT) is an important resource for older adults with a disability, and has been defined in the 2004 Assistive Technology Act as ‘any item, piece of equipment or system… that is commonly used to increase, maintain or improve functional capabilities of individuals with disabilities” (Assistive Technology Act, 2004, p. 2). ATs can prevent the weakening of health, improve quality of life, support social interactions, and compensate for loss of functioning for individuals with disabilities (Loggins, Alston, & Lewis, 2014). Although past research suggested that older adults reported use in a wide array of ATs (Hartke, Prochascka, & Furner, 1998), little is known regarding AT use among non-White older adults.

Past research indicated that approximately one-quarter of older adults used ATs, with greater use in recent decades (Gell, Wallace, LaCroix, Mroz, & Patel, 2015). AT use can help reduce barriers in the physical and social environments, thereby reducing social isolation through connecting individuals with disabilities to their communities (Agree, 1999). For example, ATs (e.g., canes, walkers) can help increase mobility and prevent falls (Agree, 1999), which is one of the leading causes of fatal and nonfatal injuries among older adults (Nicklett & Taylor, 2014). In addition, ATs that enhance communication through the use of technology can be an effective tool in reducing social isolation among older adults with a disability (Chen & Schulz, 2016). By increasing independence of older adults, ATs can decrease workload for formal and informal caregivers (Marasinghe, 2016).

Although ATs have shown potential to assist growing populations of older adults who acquire a disability (Khosravi & Ghapanchi, 2016), it is unclear how much they are used among non-White older adults and what some enabling factors and barriers to its adoption. Past research has highlighted that minority older adults may be less likely to utilize health care resources due to lower quality communication with their doctors (Rhea, Marottoli, Van Ness, & Levy, 2019). Findings from previous research have suggested that the presence of chronic conditions can motivate AT use, though for African Americans, socioeconomic factors such as income can influence the accessibility and feasibility of home modifications or other supports needed to augment the benefits of AT (Rubin & White-Means, 2001). Predictors of AT use also can include older age, female gender, higher education, greater income, and having health insurance coverage, though these factors were stronger in magnitude of association for African Americans and Hispanics (Loggins et al., 2014). Conversely, socio-cultural beliefs and normative expectations from a person’s cultural background can negatively impact the use of AT in public and private settings (Gitlin, Luborsky, & Schemm, 1998), which may result in lower adoption for certain older adult cultural groups.

The Centers for Disease Control and Prevention (CDC, 2018) estimated that there are 61 million individuals with a disability, with larger percentages for older adults. From birth to older age, the U.S. Census (2019) estimated that non-Hispanic White Americans have similar rates of disability compared to African Americans (Whites, 13.8%; Blacks: 13.9%), which is different when compared to Hispanics (8.7%) and Asians (6.9%). There is a noteworthy increase when examining the prevalence of disability among Hispanics and Asians who are 65 and older (Hispanic: 40.4%; Asian: 32.3%), whose with rates are similar to older Whites and African Americans (Whites: 35.8; African Americans: 41.4%; U.S. Census, 2019).

Disability Framework and Andersen Model of Utilization

Disability has been defined by seminal scholars as difficulty performing and engaging in activities due to a health or physical problem, and can be understood as the gap between personal capability and environmental demand (Verbrugge & Jette, 1994). Chronic conditions associated with the aging process can result in difficulties performing activities of daily living (ADLs) and participating in work and social activities. However, interventions such as ATs can help to compensate for physical decline associated with later life (Verbrugge et al., 2017; Verbrugge & Jette, 1994).

The Andersen Model was used as a conceptual framework to understand the association of race, ethnicity disability and other social determinants associated with AT use in this study (Aday & Andersen, 1974; Andersen, 1995). AT use is conceptualized as utilization of a health care resource, which can be influenced by individual factors and policies that promote equitable access. The Tech Act of 2004 aimed to maximize the ability of individuals with disabilities to access technology-related resources and address physical limitations “that affect their ability to see, hear, communicate, reason, or perform other basic life functions” (Assistive Technology Act, 2004, p. 1). By design, the policy considered all disabilities equally, though access might vary for underserved populations. There may be specific factors that can have a greater influence on accessibility of ATs among minority older adults.

Despite the potential of ATs, they require flexibility, adaptation, and learning from users (Harrington & Harrington, 2000; Melkas, 2013). Among minority older adults, socioeconomic issues and other factors can lead to further disparities in access, which can present barriers in addressing physical limitations through the use of resources to meet needs in the environment (Ozawa & Yeo, 2008). Social determinants can play an important role in whether ATs can be appropriately used by minority older adults to lessen the impact of their disabilities (Rubin & White-Means, 2001). Past studies have found that African Americans were more likely to use an AT for mobility if they had resources and access to home modifications (Gell et al., 2015; Kaye, Yeager, & Reed, 2010; Rubin & White-Means, 2001). Research has found that minority status and lower education were associated with shortcomings in the delivery of home modifications, which are needed to increase safety and enhance AT use (Meucci, Gozalo, Dosa, & Allen, 2016). To date, no study has examined the association of social determinants with AT use which included Hispanic and Asian older adults.

Purpose of Study

This study examined race and ethnic differences in the association of social determinants with AT use (e.g., wheelchairs, canes, special bed, or a special telephone) for a nationally representative sample of non-Hispanic White, Black, Asian and non-White Hispanic older adults. The variable for AT use is drawn from the Assistive Technology Act of 2004, which aimed to increase accessibility of all technology-related resources for individuals with disabilities. We hypothesized that (1) the odds for AT use would vary among older adults from different race groups, even when controlling for activity limitations, overall health, health care access, lifestyle behaviors and sociodemographic variables. Based on previous research, we anticipated that the prevalence and odds for AT use will be higher for non-White older adults after controlling for all other variables in the analysis. We further hypothesized that (2) the associations of activity limitation and social determinants such as socioeconomic status and health care access on AT use will be stronger for minority older adults (African Americans, non-White Hispanics and Asians) compared to non-Hispanic Whites.

Design and Methods

Data Source and Sample

This study used data from the 2012 Behavioral Risk Factor Surveillance System (BRFSS), collected annually by the Centers for Disease Control and Prevention (CDC). BRFSS is a system of health-related telephone surveys using random digit dialing (RDD) which collected state data from U.S. residents regarding their health-related risk behaviors, chronic health conditions, and use of preventive services. Survey design weights were created using raking weighting methodology, which included information to account for differences in the probability of selection among strata calculated from geographic counties and census tracts and their density. Analysis from this data is used to provide basic population statistics on functional disabilities in the U.S.

The 2012 survey was chosen because the interview questions regarding activity limitations and assistive technology use were aligned with the definition of AT from the Tech Act of 2004. In subsequent waves of the BRFSS, AT use was no longer captured in this manner (National Center on Birth Defects and Developmental Disabilities, 2018). The study sample included non-Hispanic White, Black, Asian and non-White Hispanic older adults 50 and older. The final sample in the analysis included 282,083 respondents in the analysis.

Dependent Variable

The dependent variable, assistive technology (AT) use, was captured by the question, “Do you now have any health problem that requires you to use special equipment, such as a cane, a wheelchair, a special bed, or a special telephone?” A response of yes was coded “1” and a response of no was coded “0.” This variable is drawn directly from the language of the Tech Act of 2004 and has been used to understand resource delivery by the CDC for disability policy and practice implementation in the U.S. (National Center on Birth Defects and Developmental Disabilities, 2018).

Independent Variables

Race.

For all categories of race, “0” was coded for non-Hispanic White as the reference group in the analysis. Black was coded with “1” for non-Hispanic Black. Asian was coded as “1” for non-Hispanic Asians. Hispanic was coded as “1” for non-White Hispanic. Categorical variables for each race group were constructed to make comparisons with different race and ethnic groups of older adults.

Activity limitation.

The variable for activity limitation was captured by the question, “Are you limited in any way in any activities because of physical, mental, or emotional problems?” Although this variable lacked specificity in type of disability, it was constructed in this way with the intent to examine all disabilities equally. A response of yes was coded “1” and a response of no was coded “0.”

Health access variables.

Three variables (health insurance, having a personal doctor, and medical cost issue) were used to capture health care access, which is relevant for examination through the Andersen Model. Health insurance coverage was captured by the question “Do you have any kind of health care coverage, including health insurance, prepaid plans such as HMOs, government plans such as Medicare, or Indian Health Service?” The variable for having a personal doctor was captured by the question, “Do you have one person you think of as your personal doctor or health care provider?” Medical cost issue was captured by the question “Was there a time in the past 12 months when you needed to see a doctor but could not because of cost?” All health access variables were coded as “1” for yes and “0” for no.

Health status variables.

Three variables (overall health, at least 1 poor physical health day, at least 1 poor mental health day) were included to capture health status. Overall health was measured with the question “Would you say that in general your health is,” and a response of poor was coded as “1,” and fair, good and excellent health were coded as “0.” This recoding was performed to capture poor health as a predisposing factor for AT use. The variable for having at least 1 day in the month of poor physical health was measured with the question, “Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?” The variable for having at least 1 day in the month of poor mental health was measured with the question, “Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” Physical health day and mental health day were recoded as two separate variables, with having at least 1 day in the month coded as “1” and no days coded as “0.”

Lifestyle behavior.

Lifestyle behavior was captured through the following question, “During the past month, other than your regular job, did you participate in any physical activities or exercises such as running, calisthenics, golf, gardening, or walking for exercise?” Yes responses were coded as “1,” and “0” for no.

Sociodemographic Variables

Gender.

Gender was coded “1” for female and “0” for male. Male gender was used as the reference group. Male gender was coded “0” and female gender was coded “1.”

Marital status.

Marital status was coded as “1” for married, and “0” for divorced, widowed, separated, never married, or a member of an unmarried couple.

Age categories.

Age was coded in categories as separate dummy variables to reflect cohorts among older adults. Those who were 50 to 64 were chosen as the reference group and coded as “0” for all age categories. Those who were 65 to 74 as the “youngest-old” were coded “1” as a separate age cohort. Those between ages 75 and 84 were coded as “1” as the “middle-old.” Those aged 85 and older were coded as “1” to capture the “oldest-old” age category.

Education.

Education was coded as “0” for no post-secondary degree, and “1” for having a post-secondary degree. Education has been highlighted as a factor in health care access and utilization, particularly for minority older adults (Rubin & White-Means, 2001).

Low income.

The variable for annual household income was recoded to capture low income, and $0 to $15,000 was coded as “1” and other income categories coded as “0.” Low income has been identified as a factor in utilization of resources such as ATs among older adults, in particular for African Americans (Rubin & White-Means, 2001).

Employment.

The variable for employment was coded as “1” for having any employment or self-employment. Those who were out of work, homemakers, students, retired, and/or unable to work were coded as “0.”

Data Analysis

Weighted multivariate logistic regression was conducted to examine the effect of race and ethnic differences with assistive technology (AT) use when controlling for activity limitation, socioeconomic status, health care access and health status variables. Descriptive analyses were first conducted to examine the prevalence of AT use and activity limitations among non-Hispanic White, African American, Asian, and non-White Hispanic older adults. Weighted logistic regression analysis was conducted to examine the differences in the odds for AT use, stratified by race and ethnic groups. Wald tests were conducted to statistically determine differences in the association of race/ethnicity with AT use across subgroups. Interaction analyses were conducted to examine the moderating effect of race with having an activity limitation and health access variables. Subgroup analyses were conducted to examine differences in the association of activity limitation, health care access, health status and lifestyle behavior with AT use for different race and ethnic groups of older adults. Stata 15.0 was used for analysis in this study (StataCorp, 2016).

Results

The independent variables of interest were race and activity limitation with respect to the dependent variable assistive technology (AT) use. The analysis included 43,418 AT users and 239,407 non-users. Weighted analysis of the study sample indicated that 13.5% of older adults reported AT use, and 28.1% reported an activity limitation because of physical, mental, or emotional problems. Among those with an activity limitation, 36.8% reported AT use, compared to 4.4% of those who did not.

Statistically different rates of use were found across race groups, where Black older adults reported the highest AT use (21.0%), followed by Whites (13.0%), Hispanics (12.6%), and Asians (5.1%; see Table 1). Black older adults reported the most activity limitations (31.0%), followed by Whites (28.8%), Hispanics (24.5%) and Asians (16.0%). The majority of the overall sample was married (60.1%). Mean age was 63.8 years of age, with Whites being on average older (66.3 years old) compared to other race groups (Black: 64.1; Hispanic: 63.8; Asian: 64.1). A higher percentage of females used an AT (14.5%) compared to males (12.4%). Almost twice the percentage of unmarried older adults used an AT (18.9%) compared to those who were married (10.0%). The reference group of 50 to 64 reported the lowest AT use (10.1%), with higher reported use among the ‘youngest-old’ who were 65 to 74 (13.4%), ‘middle-old’ between 75 to 84 (21.3%), and highest among the ‘oldest-old’ who were 85 and older (38.0%). Lower AT use was found for those with a college degree (8.6%) compared to those who did not have a college degree (15.2%). Twice the percentage of those in poverty used an AT (21.5%) compared to those not in poverty (11.1%). Those who were employed reported very low AT use (3.8%), compared to those who were unemployed (20.5%).

Table 1.

Weighted Descriptive Variables Characteristic of White, Black, Asian and Hispanic 50 and Over from 2012 BRFSS (n=282,825)

	White (n=242,083)	Black (n=22,653)	Asian (n=3,591)	Hispanic (n=14,498)	p-value
Assistive Technology (AT) Use
No	87.0%	79.0%	94.9%	87.4%	p<0.0001
Yes	13.0%	21.0%	5.1%	12.6%
Activity Limitation
No	71.2%	69.0%	84.0%	75.5%	p<0.0001
Yes	28.8%	31.0%	16.0%	24.5%
AD Use for Persons with an Activity Limitation	35.4%	50.0%	17.6%	36.7%	p<0.0001
Demographics
Gender
Male	46.1%	44.5%	46.9%	49.3%	p=0.01
Female	53.9%	55.5%	53.1%	50.7%
Married
No	37.3%	61.8%	27.9%	41.3%	p<0.0001
Yes	62.7%	38.2%	72.1%	58.7%
Age Mean (SD)	64.4 (0.04)	62.2 (0.15)	61.9 (0.49)	61.5 (17.3)	p<0.0001
Age Categories
50 to 64	56.1%	66.0%	66.1%	67.7%	p<0.0001
65 to 74	24.2%	21.0%	21.5%	20.7%
75 to 84	15.4%	10.3%	10.6%	10.0%
85 and older	4.3%	2.7%	1.9%	1.8%
Socioeconomic Status
Education
No Post-Secondary Degree	72.6%	82.0%	51.0%	87.2%	p<0.0001
Post-Secondary Degree	27.4%	18.0%	49.0%	12.8%
Poverty Level
Not In Poverty	79.0%	69.5%	78.9%	62.5%	p<0.0001
In Poverty	21.0%	30.5%	21.1%	37.5%
Employment
Not Employed	57.8%	63.4%	47.2%	58.4%	p<0.0001
Employed	42.2%	36.6%	52.8%	41.7%
Health Care Access
Have Health Insurance
No	7.2%	14.7%	9.1%	24.6%	p<0.0001
Yes	92.8%	85.3%	90.9%	75.4%
Have Personal Doctor
No	0.2%	0.3%	0.1%	0.3%	p=0.460
Yes	99.8%	99.7%	99.9%	99.7%
Medical Cost Issue
No	91.0%	82.2%	87.5%	78.7%	p<0.0001
Yes	9.0%	17.8%	12.5%	21.3%
Health Variables
Poor Health
No	93.3%	90.5%	96.3%	87.6%	p<0.0001
Yes	6.7%	9.5%	3.7%	12.4%
At Least 1 Poor Physical Health Day in Month
No	63.6%	56.6%	68.1%	55.5%	p<0.0001
Yes	36.4%	43.4%	32.9%	44.5%
At Least 1 Poor Mental Health Day in Month
No	72.4%	67.9%	76.6%	66.9%	p<0.0001
Yes	27.6%	32.1%	23.4%	33.1%
Exercise in Past 30 Days
No	27.0%	34.4%	26.2%	36.9%	p<0.0001
Yes	73.0%	65.6%	73.8%	63.1%

A higher percentage of those with health insurance used an AT (14.0%) compared to those without coverage (9.5%). Statistically similar rates for AT use were found among those with a personal doctor (13.5%) and those without a personal doctor (15.6%). Contrary to expectations, those who reported being unable to see a doctor because of cost in the past 12 months indicated higher AT use (19.9%) compared to those without medical cost issues (12.7%). Respondents who reported poor health (49.6%) reported five times the amount of AT use compared to those with fair, good and excellent health (10.6%). Respondents who reported at least one poor physical health day in the past month (25.1%) had four times the percentage of AT use compared to those who did not report any poor physical health days (6.5%). Similarly, respondents who reported at least one poor mental health day had twice the percentage of AT use (20.3%) compared to those who did not report any poor mental health days (10.8%). Those who did not exercise in the past 30 days had three times the percentage of AT use (25.0%) compared to those who exercised (8.9%).

Multivariate Logistic Regression Results for AT Users and Non-users

We hypothesized that (1) the odds for AT use would vary among older adults from different race groups, even when controlling for activity limitations, overall health, health care access, lifestyle behaviors and sociodemographic variables. We anticipated that the odds for AT use would be higher for non-White older adults. Results from Table 2 indicated that Black older adults had 83% higher odds for AT use compared to Whites (OR = 1.83, p < 0.001). AT use was statistically similar for Hispanic and White older adults (OR = 1.00). Asians, however, had 35% lower odds for AT use compared to White older adults (OR = 0.65, p < 0.05). Wald tests were used to statistically determine differences in the odds of AT use across non-White subgroups. Results indicated that Asians had lower odds for AT use compared to all other race groups (Black: Wald = 20.56, p < 0.0001; Hispanic older adults: Wald = 4.80, p < 0.05). Black older adults had higher odds for AT use compared to non-White Hispanics (Wald = 48.54, p < 0.0001).

Table 2.

Weighted Logistic Regression Results for Assistive Technology Use with All Groups (n=282,825)

Variables	Odds Ratio (SE)	95% Confidence Intervals
Ethnicity †
White (Ref)	1.00
Black, non-Hispanic	1.83 (0.08)***	(1.68; 1.99)
Asian, non-Hispanic	0.65 (0.13)*	(0.43; 0.96)
Hispanic, non-White	1.00 (0.06)	(0 .89; 1.12)
Have Activity Limitation	7.10 (0.22)***	(6.69; 7.54)
Demographics
Gender
Male (Ref)	1.00
Female	0.95 (0.03)	(0.90; 1.01)
Married
No	1.00
Yes	0.76 (0.02)***	(0.72; 0.80)
Age
50 to 64 (Ref)	1.00
65 to 74	1.24 (0.04)***	(1.16; 1.33)
75 to 84	2.01 (0.08)***	(1.87; 2.17)
85 and older	4.76 (0.24)***	(4.31; 5.25)
Socioeconomic Status
Education
No Post-Secondary Degree (Ref)	1.00
Post-Secondary Degree	0.97 (0.03)	(0.92; 1.03)
Low Income
Not Low Income (Ref)	1.00
Low Income	1.21 (0.04)***	(1.14; 1.28)
Employment
No (Ref)	1.00
Yes	0.42 (0.02)***	(0.39; 0.45)
Health Care Access
Have Health Insurance
No (Ref)	1.00
Yes	1.59 (0.10)***	(1.41; 1.80)
Have Personal Doctor
No (Ref)	1.00
Yes	1.45 (0.35)	(0.90; 2.33)
Medical Cost Issue
No (Ref)	1.00
Yes	1.05 (0.05)	(0.96; 1.14)
Health Variables
Poor Health
No (Ref)	1.00
Yes	2.12 (0.08)***	(1.96; 2.29)
At Least 1 Poor Physical Health Day in Month
No (Ref)	1.00
Yes	1.93 (0.05)***	(1.82; 2.04)
At Least 1 Poor Mental Health Day in Month
No (Ref)	1.00
Yes	1.05 (0.03)	(0.99; 1.11)
Exercise in Past 30 Days
No (Ref)	1.00
Yes	0.57 (0.02)***	(0.54; 0.60)
Pseudo R 2	31.1%

^*p<0.05,

^**p<0.01,

^***p<0.001

Pseudo R² value cannot be calculated with complex sampling weights. Therefore, the reported value is calculated without sampling weights.

^†Post-estimation Wald tests found statistically significant differences for Blacks vs. Asian (p<0.0001), Asians vs. Hispanics (p<0.05), and Blacks vs. Hispanic (p<0.001)

Having an activity limitation because of physical, mental, or emotional problems was associated with over 7 times the odds for AT use (OR = 7.10, p < 0.001). In terms of other sociodemographic variables, gender was not significantly associated with AT use. Married older adults had lower odds for AT use (OR = 0.76, p < 0.01) compared to those who were unmarried. Age was associated with higher odds for AT use, with ‘oldest-old’ associated with the highest odds for AT use compared to the reference group of 50 to 64 (65 to 74: OR = 1.28, p < 0.001; 75 to 84: OR = 2.01, p < 0.001; 85 and older: OR = 4.76, p < 0.001). In regard to socioeconomic status, education was not statistically significant. Poverty was associated with higher odds for AT use (OR = 1.21, p < 0.001). Being employed was associated with lower odds for AT use (OR = 0.42, p < 0.001).

In terms of health care access, having health insurance was associated with 59% higher odds for AT use (OR = 1.59, p < 0.001). Having a personal doctor and experiencing medical cost issues were not statistically significant with AT use. Regarding health status variables, poor health was associated with more than twice the odds of AT use (OR = 2.12, p < 0.001). Having at least 1 poor physical health day in a month was associated with almost twice the odds for AT use (OR = 1.93, p < 0.001). Having at least 1 poor mental health day was statistically significant when controlling for all other variables. Exercise was associated with 43% lower odds of AT use (OR = 0.57, p < 0.001).

Interaction of Race and Activity Limitation, Age, and Health

Interaction analysis indicated no statistical differences in the effect of activity limitation on AT use across race/ethnic groups. Similarly, no interaction effects were observed for sociodemographic variables such as gender, marital status, age, education, low income status, and employment. Among health access variables, the interaction effect of race/ethnicity with having insurance was statistically significant, indicating that insurance coverage was associated with higher odds of AT use for non-Hispanic Whites (marginal effects = 0.13, SE = 0.001, p < 0.001) and non-White Hispanics (marginal effects = 0.14, SE= 0.01, p< 0.001). No other health access variables were statistically significant in their interaction with race/ethnicity and AT use.

Subgroup Logistic Regression Results for White, Black, Asian and Hispanic Older Adults

We hypothesized that (2) the associations of social determinants such as socioeconomic status and health care access on AT use will be stronger for minority older adults (African Americans, Hispanics and Asians) compared to non-Hispanic Whites, even when controlling for all other variables in the analysis. From Table 3, subgroup analyses indicated that having an activity limitation increased the odds for AT use among non-Hispanic White (OR = 7.38, p < 0.01), Black (OR = 6.19, p < 0.01), Asian (OR = 9.76, p < 0.01), and non-White Hispanic (OR = 6.28, p < 0.01) older adults. Being married was associated with lower odds for AT use among non-Hispanic Whites (OR = 0.76, p < 0.001), Blacks (OR = 0.71, p < 0.001), and Asians (OR = 0.44, p < 0.001) but not non-White Hispanics. Older age was associated with higher odds for AT use among all race and ethnic groups, and the odds were highest for Asians and non-White Hispanics among the ‘oldest-old’ cohort of 85 and older (Whites: OR = 4.77, p < 0.001; Blacks: OR = 4.13, p < 0.001; Asians: OR = 6.43, p < 0.01; Hispanics: OR = 6.03, p < 0.001). Poverty increased the odds for AT use among non-Hispanic Whites (OR = 1.19, p < 0.001), Blacks (OR = 1.28, p < 0.001), and non-White Hispanics (OR = 1.25, p < 0.001), but not Asians.

Table 3.

Multivariate Logistic Regression Results of Assistive Technology Use Stratified by White, Black, Asian and Hispanic Older Adults

	White (n=242,083)		Black (n=22,653)		Asian (n=3,619)		Hispanic (n=14,498)
Variables	Odds Ratio (SE)	95% Confidence Intervals	Odds Ratio (SE)	95% Confidence Intervals	Odds Ratio (SE)	95% Confidence Intervals	Odds Ratio (SE)	95% Confidence Intervals
Have Activity Limitation	7.38 (0.22)***	(6.97; 7.82)	6.19 (0.59)***	(5.13; 7.47)	9.76 (4.58)***	(3.90; 24.43)	6.28 (0.99)***	(4.61; 8.57)
Demographics
Gender
Male (Ref)	1.00		1.00		1.00		1.00
Female	0.95 (0.03)	(0.90; 1.00)	1.12 (0.11)	(0.92; 1.35)	0.58 (0.22)	(0.28; 1.24)	0.90 (0.13)	(0.68; 1.19)
Married
No	1.00		1.00		1.00		1.00
Yes	0.76 (0.02)***	(0.72; 0.81)	0.71 (0.08)**	(0.57; 0.87)	0.44 (0.18)*	(0.20; 0.98)	0.84 (0.11)	(0.65; 1.10)
Age
50 to 64 (Ref)	1.00		1.00		1.00		1.00
65 to 74	1.23 (0.04)***	(1.11; 1.31)	1.42 (0.15)***	(1.15; 1.76)	1.73 (1.31)	(0.39; 7.61)	1.09 (0.17)	(0.80; 1.50)
75 to 84	1.93 (0.07)***	(1.79; 2.08)	2.22 (0.29)***	(1.73; 2.86)	11.50 (7.17)***	(3.39; 39.07)	1.80 (0.34)**	(1.24; 2.61)
85 and older	4.77 (0.24)***	(4.32; 5.28)	4.13 (0.86)***	(2.74; 6.22)	6.43 (4.28)**	(1.74; 23.71)	6.03 (2.01)***	(3.14; 11.58)
Socioeconomic Status
Education
No Post-Secondary Degree (Ref)	1.00		1.00		1.00		1.00
Post-Secondary Degree	0.99 (0.03)	(0.93; 1.05)	0.88 (0.11)	(0.69; 1.12)	1.00 (0.47)	(0.39; 2.54)	1.02 (0.19)	(0.70; 1.48)
Low Income
Not Low Income (Ref)	1.00		1.00		1.00		1.00
Low Income	1.19 (0.04)***	(1.13; 1.26)	1.28 (0.12)*	(1.03; 1.54)	0.98 (0.40)	(0.45; 2.16)	1.25 (0.18)*	(0.95; 1.65)
Employment
No (Ref)	1.00		1.00		1.00		1.00
Yes	0.45 (0.02)***	(0.42; 0.49)	0.25 (0.03)***	(0.20; 0.32)	0.92 (0.60)	(0.26; 3.29)	0.39 (0.08)***	(0.25; 0.59)
Health Care Access
Have Health Insurance
No (Ref)	1.00		1.00		1.00		1.00
Yes	1.66 (0.11)***	(1.47; 1.88)	1.15 (0.19)	(0.84; 1.59)	3.32 (2.36)	(0.82; 13.40)	1.98 (0.47)**	(1.24; 3.14)
Have Personal Doctor
No (Ref)	1.00		1.00		1.00		1.00
Yes	1.09 (0.26)	(0.69; 1.73)	4.40 (2.95)*	(1.18; 16.35)	--	--	1.61 (1.17)	(0.39; 6.71)
Medical Cost Issue
No (Ref)	1.00		1.00		1.00		1.00
Yes	1.09 (0.05)	(0.99; 1.19)	0.95 (0.12)	(0.74; 1.23)	1.15 (0.71)	(0.34; 3.84)	1.02 (0.18)	(0.72; 1.45)
Health Variables
Poor Health
No (Ref)	1.00		1.00		1.00		1.00
Yes	2.05 (0.08)***	(1.90; 2.22)	2.04 (0.27)***	(1.58; 2.64)	3.51 (1.99)*	(1.15; 10.67)	2.60 (0.45)***	(1.85; 3.64)
At Least 1 Poor Physical Health Day in Month
No (Ref)	1.00				1.00		1.00
Yes	1.99 (0.06)***	(1.88; 2.10)	2.22 (0.21)***	(1.85; 2.68)	0.64 (0.33)	(0.23; 1.75)	1.48 (0.21)**	(1.12; 1.96)
At Least 1 Poor Mental Health Day in Month
No (Ref)	1.00		1.00		1.00		1.00
Yes	1.03 (0.03)	(0.97; 1.10)	1.07 (0.11)	(0.88; 1.31)	0.67 (0.28)	(0.30; 1.51)	1.18 (0.18)	(0.87; 1.60)
Exercise in Past 30 Days
No (Ref)	1.00		1.00		1.00		1.00
Yes	0.56 (0.02)***	(0.53; 0.59)	0.63 (0.06)***	(0.53; 0.74)	0.97 (0.41)	(0.42; 2.24)	0.56 (0.08)***	(0.43; 0.73)
Pseudo R 2	31.0%		30.3%		27.6%		30.2%

^*p<0.05,

^**p<0.01,

^***p<0.001

Pseudo R² value cannot be calculated with complex sampling weights. Therefore, the reported value is calculated without sampling weights.

In terms of health access variables, having health insurance was associated with higher odds for AT use among non-Hispanic Whites (OR = 1.66, p < 0.001) and non-White Hispanics (OR = 1.98, p < 0.001), but not for Blacks and Asians. Having a personal doctor was associated four times the odds for AT use among Black older adults (OR = 4.40, p < 0.05). Having a personal doctor was not significantly associated with AT use for non-Hispanic Whites, Asians and non-White Hispanics. Among health status variables, poor health was associated with higher odds for AT use for all older adult race and ethnic groups and was highest among Asians (OR = 3.51, p < 0.05), followed by non-White Hispanics (OR = 2.60, p < 0.001), non-Hispanic Whites (OR = 2.05, p < 0.001) and Blacks (OR = 2.04, p < 0.001). Having at least one poor physical health day was associated with AT use for non-Hispanic Whites (OR = 1.99, p < 0.001), Blacks (OR = 2.22, p < 0.001), and non-White Hispanics (OR = 1.48, p < 0.001), but not Asians. Having at least 1 mental health day was not statistically significant for AT use for any race and ethnic group. Exercising in the past 30 days was associated with lower odds for AT use among non-Hispanic Whites (OR = 0.56, p < 0.001), Blacks (OR = 0.63, p < 0.001), and non-White Hispanics (OR = 0.56, p < 0.001), but not Asians.

Discussion

The present study sought to examine the differences in the relationship of race/ethnicity and social determinants with assistive technology (AT) use among older adults. Findings indicated that AT use varied across non-Hispanic White, African American, non-White Hispanic and Asian older adults. Having an activity limitation was significantly associated with higher odds for AT use, even when controlling for health care access, health status, and sociodemographic variables. African Americans were more likely to use an AT compared to other race groups in this study. Being married was associated with lower AT use, which may be explained by having a spouse who can offer informal caregiver support. Although personal assistance can function as a social resource to increase AT use (Hoenig et al., 2006), findings from this study suggested ATs may be a resource of last resort for older adults who lack support from a spousal caregiver. Similarly, low income was associated with higher AT use among older adults in this analysis. Public housing units were mandated by the Americans with Disabilities Act of 1990 to provide accommodations for disability access, which made it possible for older adults residing in low-income housing to more fully use ATs in their home settings. This may explain higher AT use among low income older adults.

Race and AT Use

Findings from this study suggest that sociodemographic factors and health care access variables had different effects on AT use among race/ethnic groups of older adults. Older age was associated with higher AT use, and this association was strongest for Asians and non-White Hispanics who were in the ‘oldest-old’ cohort of 85 years and older. This finding supports previous research which indicated that among Asians and non-White Hispanics, older age was associated with higher rates of disability and need for services (Markides & Rote, 2015; Mehta, Sudharsanan, & Elo, 2014). Similarly, having an activity limitation and reporting poor overall health were associated with higher prevalence and odds for AT use, especially among Asian and non-White Hispanic ‘oldest-old’ adults.

In terms of health access variables, having insurance was associated with higher AT use, though this effect varied across race/ethnic groups. Having health insurance was associated with higher AT use for non-White Hispanic and non-Hispanic White older adults only. This is consistent with past research which suggested that having health insurance can increase access to health care resources, which is an enabling factor for AT use. Past research has highlighted the significance of having a personal connection to health providers which can increase service utilization for minority populations (Institute of Medicine (US) Committee on Understanding and Eliminating Racial and Ethnic Disparities in Health Care, 2003). Consistent with this, having a personal doctor was associated with higher AT use for African American older adults. Among Asians, health care access variables were not associated with AT use. It is unclear if there may be underlying cultural factors which may further explain this finding.

Among Asian older adults, having an activity limitation, being older, unmarried and in poor health were associated with higher AT use in this study. This suggests that Asian older adults may delay AT use until their physical limitations increase in severity and they do not/no longer have support from a spousal caregiver. Previous research has highlighted that Asians tended to have a pattern of low service utilization (Kim, Chen, & Spencer, 2012; Mutchler, Prakash, & Burr, 2007). It also is possible that Asian older adults may accept frailty and loss of physical capabilities as part of their natural aging process, thereby not seeking out supports and services. More research is needed to examine if there is an unmet need among Asian older adults, and how they can potentially be engaged earlier in use of ATs to delay physical decline and prevent further loss of functioning.

Strengths and Limitations

This study informs the gap in literature addressing racial differences in the effect of health access, health status and other social determinants on assistive technology (AT) use for non-Hispanic White, African American, Asian and non-White Hispanic older adults. Results from this study come from nationally-representative, population-based data. Although analysis from this data has important strengths due to its generalizability, there are certain limitations. Activity limitation and AT use variables from the BRFSS data are self-reported. Although self-reported data can be inaccurate in small sample studies, analysis from large, population-based data can account for most inaccuracies through rigorous weighting design methods such as those employed in this study. The data also did not include specific information in regard to the type of limitation experienced by older adults or the technology-related resource that was used. This information can provide an understanding of how a specific activity limitation such as difficulty seeing, hearing, communicating, walking, remembering, or performing activities of daily living can be compensated by different types of ATs. It is important to note, however, that the Tech Act of 2004 was non-specific in its language in large part because it aimed to increase access and address all disabilities equally. Findings from this study suggest that regardless of typology, race/ethnicity was a significant predictor of AT use as a form of health care utilization. Predisposing factors and health care access as they relate to the Andersen Model can have different effects on health care utilization for non-Hispanic White, African American, Asian and non-White Hispanic older adults.

Another study limitation is that the data used was cross-sectional and not longitudinal. Participants were interviewed at one point in time. It is beyond the limits of the data to definitively determine if variables, such as low income or poor health, were causally prior to AT use or vice versa. Older adults with pre-existing chronic health conditions living in poverty may be more likely to experience physical decline and use an AT as a resource of last resort. However, having an activity limitation, being in poor health, and having a poor physical or mental health day in the past month were used as proxies to rule out a pre-existing chronic health condition. These variables were included as controls in the multivariate logistic regression analysis. This provides further evidence on the statistically significant association of race and ethnicity and other social determinants with AT use in this study.

Despite these limitations, findings from this study offer important insights into differences in the relationship of race and ethnicity with AT use for older adults using large-scale, population health data. Findings from the analysis can inform practice, policy and future research on enabling factors and barriers to utilization of resources such as ATs for older adults. This is particularly important for at-risk and under-served minority ‘oldest-old’ who may be insufficiently engaged in the health care system.

Recommendations

Demographic shifts have highlighted that older adults are becoming more racially diverse in the United States. Findings from this study suggested that minority older adults may have different enabling factors and barriers as they relate to the utilization of health care resources such as ATs. Although AT use can reduce falls and improve overall quality of life (Khosravi & Ghapanchi, 2016), many older adults do not use them until they are very old or their needs have increased in severity. Past research has highlighted complexities in the felt need and impact on quality of life in consideration of the acceptability of AT use among older adults (McCreadie & Tinker, 2005). Clinicians can discuss the pros and cons of AT use with their older adult patients, along with enabling factors or potential barriers so ATs can be used safely and effectively in their homes (Wilson, Mitchell, Kemp, Adkins, & Mann, 2009). There is a need to increase awareness among older adults regarding the benefits of technology-related resources. ATs can be promoted through public service announcements (PSAs) and social media campaigns in different languages to increase engagement for older adults from different communities. Policies aimed at increasing utilization of health care resources must include considerations for equitable access for minority older adult populations.

It is recommended that health care professionals who work with older adults conduct more comprehensive screenings in relation to their health and service needs, paying particular attention to those who are older, unmarried and have limited support networks. Health professionals also can play a role in providing crucial linkages to appropriate resources, services, and treatment supports tailored to address health concerns specific to minority older adult populations. Policies can include provisions aimed at increasing access and utilization of prevention services such as ATs targeted to the needs of minority older adult populations.