The Impact of Assistive Technologies on Formal and Informal Home Care

Abstract

Purpose of the Study:

Assistive technologies help people with disabilities compensate for their impairments. This study assessed which of 5 categories of assistive technologies—indoor/outdoor mobility, bed transfer, bathing, toileting, and telephone assistance—were substitutes or complements for human personal assistance by differentiating between total and formal personal assistance service (PAS) hours.

Design and Methods:

The study analyzed 2004 National Long-Term Care Survey community-dwelling respondents receiving assistance with activities of daily living. Ordinary least squares (OLS) on total PAS hours was estimated on the entire sample, and logit and OLS models were estimated on the likelihood and hours of formal PAS, respectively.

Results:

Assistive technology for indoor/outdoor mobility, bed transfer, and bathing was found to be substitutes for total PAS, whereas assistive technology for bed transfer and toileting was found to be complements for the use of formal PAS. Telephone assistance was not significant for either total or formal PAS hours.

Implications:

The use of some assistive technologies by older people with disabilities appears to reduce the amount of informal care provided, but not the amount of paid PAS. Thus, this study does not provide support for the hypothesis that the use of assistive technologies will reduce use of paid care and, therefore, spending for long-term care.

Older and younger people with disabilities living at home compensate for their functional limitations by receiving personal assistance from human caregivers and through assistive technologies. These assistive technologies range from common devices such as canes and ramps to more high-tech devices such as electric wheelchairs and devices to monitor and prompt people with disabilities to take medicine and perform other tasks like eating. With the increasing demand for long-term care services (Johnson, Toohey, & Wiener, 2007; Organization for Economic Co-operation and Development, 2006) and a likely long-run shortage of personal care workers (Stone & Harahan, 2010), policymakers are interested in whether assistive technology can be a substitute for paid long-term care services or whether it is a complementary support. If assistive technology is a substitute for personal assistance, such as when a cane is used rather than relying on a human caregiver, then cost savings may be possible by promoting its use. Alternatively, if assistive technology is a complement, meaning both are used together, such as when a mechanical lifting device is used by a personal assistant for transferring a person to and from bed, then providing such technology will increase costs to the extent that it is provided, although it may reduce unmet need (Agree & Freedman, 2003).

Between the early 1980s and the early 2000s, changes occurred in both the U.S. disability rate and how underlying needs of people with disabilities were addressed. Prevalence of any type of disability in the older population dropped from 26.5% in 1982 to 19.0% in 2005 (Manton, Gu, & Lamb, 2006). More recent studies have shown that the disability rate has been relatively constant over the last decade (Freedman et al., 2013).

This decrease in disability rates was coupled with a decline in the need for personal assistance services (PAS) (Manton, Corder, & Stallard, 1993). From 1992 to 2001, in the older population living in the community experiencing difficulty in self-care activities, underlying difficulty in performing various tasks declined from 30% to 26% and the use of PAS declined from 11% to 9% (Freedman, Agree, Martin, & Cornman, 2005). These declines were on the order of 1%–2% per year (Freedman et al., 2004).

One potential driver in the decline in the need for PAS is the greater use of assistive technologies over the last few decades (Kaye, Kang, & LaPlante, 2000; LaPlante, Hendershot, & Moss, 1992; Manton et al., 1993). Between 1984 and 1999, the proportion of older chronically disabled community residents using equipment, with or without human help, for all activities that were analyzed doubled to nearly 30%; in the same population, the proportion relying solely on human help fell to 14% (Spillman, 2005). The use of assistive devices with no human help for any deficits in the activities of daily living (ADLs) showed an upward trend throughout the 1990s that continued through 2004 (Spillman, 2012).

The policy implications of these changes are potentially far reaching. With the aging of the population, especially among persons aged 85 and older who are more likely to have disabilities, demand for long-term support services will increase dramatically (Vincent & Velkoff, 2010), but the country may face an even greater shortage of long-term care workers to meet that need than exists today (Institute of Medicine, 2008). Therefore, understanding what role assistive technologies may play in substituting for PAS is important. Although a major benefit of assistive technologies is promoting independence for people with disabilities, offsetting the need for formal (paid) care is also important given the current budget environment. Moreover, reducing the need for informal care (unpaid care by family and friends) may also increase the labor supply of taxpaying workers (Institute of Medicine, 2008).

Prior research sought to determine whether assistive technologies and PAS are complements or substitutes. The limited research literature differs on this point, with some studies showing they are complements (Agree, Freedman, Cornman, Wolf, & Marcotte, 2005), some substitutes (Agree & Freedman, 2000; Hoenig, Taylor, & Sloan, 2003; Mortenson et al., 2012), and some both (Allen, Foster, & Berg, 2001; Allen, Resnik, & Roy, 2006). The differences across studies seem to depend on contextual factors, data, and analysis methodologies.

A critical issue in understanding the relationship between assistive technologies and PAS use is controlling for differences in disability levels of recipients using these services. If the underlying need is relatively light, some assistive technologies may accommodate most or all of one’s need for help. On the other hand, if underlying need is great, then substantial amounts of assistive technologies and human personal assistance may be needed. Controlling for differences in case mix in analyses is important if policy makers are to accurately assess whether there is a substitution effect between assistive technologies and PAS (Hoenig et al., 2003). Otherwise, higher estimates in regression analyses of the effect of assistive technologies on PAS would likely be obtained, reflecting the uncontrolled higher level of functional limitations.

To address these issues, this study identified three research questions: (a) Are assistive technologies and PAS complements, substitutes, or both? (b) Which assistive technologies serve as complements and substitutes? and (c) Which assistive technologies have the greatest impact on paid formal care? We hypothesize that after controlling for potential case mix differences, sociodemographic and other factors, all assistive technology types identified in the analyses will substitute for PAS.

This study differs from prior research in three ways. First, not all prior studies distinguished the effects of different types of assistive technologies and their effect on PAS. Second, this study uses more recent data, a wider range of data to control for case mix differences, and has fewer missing data than prior studies. Third, we employ a two-part model to control for downward bias in estimates of assistive technology on formal care use.

Design and Methods

Conceptual Framework

For persons living at home with functional impairment, PAS and assistive technologies are used to address limitations in performing ADLs and other tasks required for independent living. To the extent that persons in need cannot obtain sufficient help from assistive technologies or informal care, formal care is sought for any remaining needs. Given that formal care often is paid at taxpayer expense (Kaye, Harrington, & LaPlante, 2010; Wiener, 2006), differentiating between formal and informal care use is important, as well as assessing the amount of formal care needed (Agree et al., 2005). Overall though, from both practical and research perspectives, it is important to understand the extent to which assistive technologies can reduce total PAS, regardless of whether PAS is paid or not (Hoenig et al., 2003). This study analyzes the effects of assistive technology types on three related outcomes: total PAS, the probability of using any formal PAS, and the amount of formal PAS used if any.

Because many state Medicaid programs provide home- and community-based services depending on the degree to which an individual cannot perform ADLs (O’Keeffe, Maier, & Freiman, 2010), we grouped an extensive list of more than 100 assistive technologies in our data by the ADL domains that they addressed. This effort was aided by the survey questionnaire, which had solicited responses about assistive technologies used by ADL domain. Prevalence of use was high enough for various assistive technologies for us to create four ADL domains: indoor/outdoor mobility, bed transfer, bathing, and toileting. Prevalence was too low to form categories for assistive technologies addressing eating and dressing because doing so would have produced spurious regression estimates for these low prevalence categories. Finally, we also sought to create at least one assistive technology category designed to meet instrumental activities of daily living (IADLs), but prevalence of these types of assistive technologies was too low to support more than one group—providing assistance with using the telephone.

We included a broad range of functional impairment and health measures as case mix controls in multivariate analyses (Hoenig et al., 2003). The degree of functional impairment may be indicated by how many ADL and IADL deficits a person has. In addition, whether they have dementia may affect the ability to use assistive technology (O’Keeffe et al., 2010). The need for supervision among persons with Alzheimer’s disease may limit the ability of assistive technology to substitute for formal PAS. Health factors potentially affecting the need for PAS are hypothesized to include the number of chronic conditions (Hoenig et al., 2003) and self-reported health status. In addition, whether someone actively participated in the community may serve as an indicator of one’s overall ability, which can be measured by whether they participated in events outside the home at least once in the past month.

A range of demographic and socioeconomic characteristics also are hypothesized to affect PAS receipt (Agree et al., 2005; Hanley & Wiener, 1991; Johnson & Wiener, 2006; Kaye et al., 2010; Spillman, 2005). These characteristics were considered as confounders that needed to be controlled rather than key issues of interest for this study. Considering demographic characteristics, increasing age may result in greater need for PAS because of the greater risk of disability and the declining availability of care from spouses. Women have been found to have a higher need for PAS than men because of greater longevity and the tendency of men to marry women who are younger than them. Whites when compared with people of other races may have greater use of PAS because of possible discrimination against other races and greater informal supports among older people in ethnic communities. In terms of socioeconomic status, persons with less education and income may be more likely users of PAS because of the greater willingness to receive informal support. Finally, persons with Medicaid are more likely to receive PAS because home- and community-based services are covered under the program.

Data

This study analyzed the 2004 National Long-Term Care Survey (NLTCS), which was sponsored by the National Institute on Aging and conducted by the U.S. Census Bureau, which includes a nationally representative sample of the community and institutionalized populations aged 65 and older. Sample members are selected from a list of eligible Medicare beneficiaries provided by the Centers for Medicare & Medicaid Services. The survey contains information about ADL and IADL disabilities, helpers and hours of help provided, equipment use, medical conditions, cognitive function, and demographic and socioeconomic characteristics (National Archive of Computerized Data on Aging, n.d.).

The NLTCS began with a screener interview to determine a person’s functional and residential status. Individuals who are not functionally impaired and living in the community were given an abbreviated interview. Persons with functional impairment and living at home, the sample of interest for this study, were administered a detailed community interview.

Proxies knowledgeable about the sample person’s health conditions and physical activities were interviewed instead of or along with the sample member if the sample person was unavailable or unable to respond to survey items. The proportion of proxies participating varied across sections of the survey instrument, but comprised about one fifth of all respondents.

The 2004 NLTCS contained 5,201 persons who completed a community interview. To construct our analysis sample, we excluded people who were not in need of long-term care services or were missing certain data. Respondents were excluded from the analysis sample if they were part of the new Medicare cohort that had no functional impairment (729 people), had no ADL or IADL deficits expected to last 3 months or more (1,175 people), had ADL or IADL deficits but no helpers (1,143 people), and had unknown educational status (59 people) or unknown overall health status (14 people). The final sample size was 2,081 respondents. Thus, the analysis sample population consisted of older people with ADL or IADL disabilities.

Three dependent variables were created for regression analyses. The first two, total PAS hours in the past week for use in an ordinary least squares (OLS) regression and whether any formal PAS was used in the past week for use in a logistic regression, were analyzed using the full sample of 2,081 respondents. The third dependent variable, number of formal PAS hours in the past week, was analyzed with the 686 respondents with such care using an OLS regression.

The NLTCS includes questions about the hours of help received because of a disability or health problem as perceived by the sample person/proxy. Respondents were asked “During the past week, how much time did someone help you because of a disability or health problem?” These questions were asked about both formal (paid) and informal (unpaid) caregivers.

The NLTCS includes information on more than 100 assistive technologies. Five variables representing assistive technology use were created as the key policy variables of interest for use in regressions. Each respondent was characterized as using or not using each of five domains of assistive technology. The five domains were technologies intended to address indoor/outdoor mobility (e.g., wheelchairs, walkers, canes, railings, crutches, elevators, ramps, orthopedic shoes, leg or back braces, chairlifts on stairs, and prostheses), bed transfer (e.g., bed lift, wheelchair, and walker), bathing (e.g., shower seat, tub stool, handle bar, hand held shower, and rubber mat), toileting (e.g., raised or portable toilet, grab bar, and special underwear), and the telephone (e.g., amplifier and enlarged dialer).

A variable was created for informal PAS hours in the past week and used as an independent variable in the logistic and OLS regressions on formal PAS use. The remaining independent variables served largely as controls in all regressions. These variables included a range of functional status, health status, demographic, socioeconomic, and health insurance characteristics. In particular, we created a count variable for five ADL deficits (movement, eating, transferring, bathing, and toileting) and an IADL deficit count variable for nine IADLs (e.g., doing housework and laundry, managing money, shopping, and using the telephone). About 15% of the sample had missing values for income, which was imputed using the demographic variables in the analysis.

Methods

Five analytic techniques were used in the analyses. First, we used the NLTCS weights to estimate descriptive statistics. Second, because PAS hours in the past week are skewed to the right with few respondents having large numbers of hours, we created a more normal-shaped distribution by logging the PAS hours measures when used as a dependent variable. Third, when making postestimation predictions of a logged dependent variable, a smearing factor was used when transforming a logged distribution back to a normal distribution (Duan, 1983). Fourth, the assistive technology variables were interacted with the variable for informal PAS hours but were not statistically significant and had small coefficients. As a result, these interactions are not included in the final equations. Finally, marginal effects were calculated to estimate the effects of separately using each individual assistive technology. To do so, predictions were first estimated by making all sample members have a given assistive technology, then again with no one having that assistive technology. After making those estimates, the difference between the two predictions for each sample member was calculated. Each prediction was estimated with confidence intervals so as to determine whether the mean difference in the sample for using versus not using each assistive technology was statistically significant.

Results

Table 1 presents descriptive statistics for the full sample of 2,081 respondents receiving some type of PAS. The proportions for dummy variables in the table are reported as percentages in the text. The mean of the first three dependent variables, total (formal and informal) PAS hours in the past week, was 30.2hr (2.4hr when logged), but many respondents reported far fewer hours. Approximately 25.1% of the sample received fewer than four total PAS hours per week, and 38.3% received fewer than eight total PAS hours per week (not shown). On the other hand, 11.2% of respondents received 80 or more total PAS hours per week.

One third (33.0%) of respondents reported using any formal (paid) PAS, the second dependent variable. The mean of informal PAS hours across the sample regardless of any use was 22.2hr, much higher than the mean for formal care. Approximately 85.7% of all respondents reported receiving some informal PAS in the past week, with a mean of 25.9hr if any informal PAS was used (not shown).

Assistive technology use ranged from 68.4% of respondents using assistive technologies for mobility-related needs to 9.3% using assistive technology for the telephone. In addition, about 61.2% used assistive technology for bathing, 46.3% used assistive technology for toileting, and 40.5% used assistive technology for getting in and out of bed. Approximately 8.7% of respondents reported no assistive technology use.

Regarding functional status and health characteristics, respondents averaged 2.9 ADL deficits and 4.9 IADL deficits. Most respondents (85.7%) reported having at least one ADL deficit (not shown). Few (14.3%) respondents reported having only IADL deficits.

Approximately 32.6% of respondents/proxies reported the presence of dementia such as Alzheimer’s disease or mild cognitive impairment. On average, respondents reported having five chronic conditions based on a list of conditions identified in the NLTCS. For self-reported health status, 13.7% reported excellent health, 34.7% good health, 33.8% fair health, and 17.7% poor health. Only 13.5% of respondents reported being active in the community. Several demographic characteristics are also reported in the table. Almost half of the sample (47.7%) was age 85 and older. In terms of health insurance, 10.5% of the sample had Medicaid coverage.

The third dependent variable, the natural log of formal PAS hours in the past week if any formal PAS hours were used, had a mean of 24.5hr (2.2hr when logged; Table 2). Respondents with formal PAS use reported receiving an average of 14.0hr of informal PAS in the past week. Table 2 presents results on remaining analytic variables, differing from Table 1 by the samples involved in regressions.

Table 3 presents cross-tabulations of the proportion of respondents using various types of assistive technologies with regard to their use of formal PAS. Respondents with formal PAS hours in the past week usually had much higher proportions of assistive technology use than respondents without formal PAS hours. For example, in the overall sample, 75.7% of respondents with any formal PAS in the past week used assistive technology for mobility issues, whereas only 64.8% of respondents with no formal PAS did so. A similar relationship holds for each assistive technology type examined.

Of the five assistive technology variables in each of the three regressions (Tables 4–6), only one was statistically significant in each OLS regression (indoor/outdoor mobility for the total PAS regression and telephone assistive technology for the formal PAS regression). Three assistive technology variables (bed transfer, bathing, and toileting) were statistically significant in the logistic regression of receiving any formal PAS with coefficients of 0.28, 0.59, and 0.42, respectively. In all three regressions, coefficients of independent variables showing statistical significance at p < .05 had the expected positive or negative sign based on the conceptual framework. For example, regarding demographic, socioeconomic, and functional status characteristics in logistic analysis of receiving any formal PAS, people who were women, older, had Medicaid coverage, and people who were more ADL and IADL impaired had a greater likelihood of using any formal PAS. In the OLS regression on formal PAS hours, there was approximately a one-for-one trade-off with informal care hours. That is, an increase in 1hr of formal PAS hours was associated with a roughly 1hr decrease in informal care hours. R² estimates in the two OLS regressions were .28 and .34, showing good model fit and potentially good control of case mix differences across respondents as indicated by statistically significant case mix coefficients.

The marginal effects calculations of the changes in total and formal PAS hours in the past week that were performed after the regressions were run indicate that assistive technology use is associated with lower total PAS hours, but modestly higher formal PAS hours (Table 7). Assistive technology for indoor/outdoor mobility, bed transfer, and bathing was each associated with approximately 8-hr reductions in total PAS hours in the past week. Neither assistive technology for toileting nor telephone was statistically significant. In contrast, assistive technology for bed transfer and toileting was each associated with about a 4-hr increase in formal PAS hours in the past week. Assistive technologies for indoor/outdoor mobility, bathing, and phone were not statistically significant predictors of formal PAS hours.

Discussion

With the aging of the population, policymakers face three interrelated strains concerning the financing and delivery of long-term care (Commission on Long-Term Care, 2013). First is the rapidly growing need for additional long-term care services. If current use rates are held constant, the number of people needing informal care, home care, and nursing home care will roughly double between 2000 and 2040 (Johnson et al., 2007). The second factor is a likely substantial increase in public spending on long-term care that will accompany the increase in use. The Organization for Economic Co-operation and Development projects that public long-term care expenditures in the United States, which were about 1% of gross domestic product in 2005, will climb to 2%–3% in 2050 (Organization for Economic Co-operation and Development, 2006).

The third factor is that the number of people likely to need long-term care services will likely increase much faster than the working age population, which may create worker shortages that threaten the ability to provide needed services even if financing is available (Institute of Medicine, 2008; Stone & Harahan, 2010). For example, the ratio of the working age population (people aged 20–64) to the population most likely to need long-term care (people aged 85 and older) is projected to decline from 32.3 in 2010 to 15.3 in 2050 (author’s calculations of data in Vincent & Velkoff, 2010).

Faced with these pressures, policymakers are seeking strategies that can reduce the upward demand for services. One hypothesis is that the use of assistive technologies can reduce the need for paid and unpaid PAS, thereby reducing the need for public spending and long-term care workers and the strains on informal caregivers. Previous research has produced inconsistent findings as to whether assistive technology is a substitute or complement to PAS (Agree & Freedman, 2000; Agree et al., 2005; Allen et al., 2001, 2006; Hoenig et al., 2003; Mortenson et al., 2012). In addition, by reducing the need for human assistance, assistive technologies may provide greater independence and improved quality of life for people with disabilities.

This study found a complex relationship between assistive technology and PAS use. In our analysis of the 2004 NLTCS, across the total sample, assistive technology for indoor/outdoor mobility, bed transfer, and bathing each resulted in about an 8-hr decrease in total PAS hours during the past week. Given that the mean for total PAS hours in the past week was about 30hr, these three assistive technology types resulted in about a 25% reduction each in total PAS hours. These reductions in total PAS hours appear to result in only lower informal care hours because assistive technology use either had no or only modestly positive effects on formal care use. Only assistive technology for bed transfer and toileting had statistically significant effects on formal PAS hours in the past week. These two assistive technology types appear to increase, not decrease, formal care hours, although this finding may be confounded by higher functional impairment if the control variables used did not adequately account for it. Thus, they served as complements for formal PAS rather than substitutes. In other words, some assistive technologies appear to reduce informal care, but not paid care, and others are associated with modestly higher levels of paid care. This study did not find evidence that assistive technologies reduced paid care.

Why use of various assistive technologies would substantially reduce informal care hours but modestly increase formal care hours is uncertain. One possibility is that both informal and formal caregivers also benefit from assistive technology use by care recipients. More obviously, informal caregivers such as family and friends effectively decrease their care time and responsibility if assistive technology is used. But formal caregivers may also benefit from assistive technology use if they would have been providing even more care had it not been for assistive technology use by their care recipients. Formal care providers may be instrumental in recommending assistive technology for heavy-care recipients to keep care hours under the authorized amount of care by third-party payers and to potentially prevent injury to workers, both of which are benefits that accrue largely to public payers. Given that most assistive technology types had no effect on formal care hours, some assistive technology may make the job of formal caregivers easier even if it modestly increases formal care hours.

Another possibility is that assistive technology cannot completely address the total needs of people with disabilities and that PAS are still needed. When comparing the mean estimates of assistive technology use for respondents with formal PAS (Table 2) versus the entire sample (Table 1), respondents with formal PAS were greater users of assistive technology than those without formal PAS. Even after controlling for case mix, it is likely that respondents who were less severely disabled might have been able to use assistive technology for most or all of their functional needs and therefore may have used fewer PAS hours as result. On the other hand, more disabled respondents might have been able to use assistive technology for only some of their needs (or a given assistive technology type may not have completely compensated for their entire need for help for the task) and so had to use more hours of PAS than those with less disability to make up for their additional need for help.

Given the benefits of assistive technologies for reducing total PAS hours, it is notable that assistive technology is not used by everyone. One distinct difference of policy significance is that assistive technology use rates are approximately 25%–50% higher among those using formal PAS compared with those not using formal PAS. Even so, among persons with no formal care use, except for telephone assistive technology, assistive technology use rates range from only 35% to 65%. One possibility is that assistive technologies may not be as accessible to those relying only on informal care because they do not know about them, do not know how to obtain or pay for them, or cannot afford them. Wider adoption of assistive technologies by those with only informal care use would promote greater independence and increased quality of life among such users and potentially allow caregivers either respite time or a greater opportunity to work. It may also be possible that they may not want to use assistive technology if it means less personal interaction with family and friends. The technology also may not work equally well for all individuals.

The potential reduction in PAS hours is only one measure of the benefits of assistive technology. In addition, assistive technologies may also reduce the intensity of formal care activities or stretch informal care resources in ways that our formal and informal PAS measures did not capture. These additional benefits may offset initial costs of acquiring and learning to use assistive technologies for which this study did not account.

Although this study provides new information about the relationship between assistive technology and the use of PAS, it has a number of limitations. The potential inadequacy of the case mix control variables is the primary study limitation. The marginal effects calculations derived from the regressions are unbiased only if regression models adequately control for case mix differences in care need across recipients. Otherwise, we would overstate the effect of assistive technologies on PAS and not observe the substitution effect found in total PAS hours. High R² values for model fit (meaning we found an adequate set of case mix adjusters), and the statistically nonsignificant findings, or modest positive effects of assistive technology use on formal care use, suggest that case mix differences are largely controlled.

This study used measures for functional and cognitive status, the number of chronic conditions, self-reported health status, and community participation to control for case mix differences. Using older NLTCS data, Hoenig and colleagues (2003) found that assistive technology use and PAS are correlated with frailty, but once frailty is controlled for, a substitution effect was found, resulting in four fewer total PAS hours per week if any assistive technologies were used. The current study also found a substitution effect for total PAS hours, subject to modestly higher formal care use for two assistive technology types.

In addition to any potential failure of case mix controls, this study had four other limitations. First, cross-sectional data were used, whereas panel data may have provided better detection of changes in formal and informal PAS over time. Moreover, some assistive technologies have changed substantially over time and with those changes the dynamics between assistive technologies and PAS may have changed as well.

Second, formal and informal care and assistive technology use may be simultaneously determined in practice, suggesting potential endogeneity between PAS and assistive technology use. No plausible instrumental variable was available to allow for potential correction with two-stage least squares estimation, and such estimation is known to introduce measurement error in estimates. This approach is consistent with the work of other researchers (Agree et al., 2005; Allen et al., 2001; Hoenig et al., 2003). Third, although this study addresses PAS utilization, it does not assess needs for services; thus, it cannot address whether people using assistive technology have fewer unmet needs. Fourth, the study only examines the impact of assistive technologies on the use of PAS in the community. Assistive technologies also may help prevent or minimize significant care needs with high costs (e.g., emergency room visits, hospital admissions, or nursing home use), although some studies have found that home care use does not substantially affect acute care use (Anderson, Wiener, & Khatutsky, 2006). Fifth, our categorical measures of assistive technology treat each device within each ADL category as equal, whereas different devices within an ADL category may not have the same importance or value in decreasing PAS hours.

Understanding the relationship between assistive technology and PAS use, and how both affect the disablement process, can lead to better policies to promote independent living. Future research on this topic could focus on the impact of the newer “high-tech” assistive technologies and on the effectiveness of assistive technologies for certain subpopulations, such as people with Alzheimer’s disease (O’Keeffe et al., 2010). Improvements in assistive technology devices and policies to promote assistive technology adaptation are needed to promote greater use. The goals of increased use of assistive technologies should be to increase the independence of people with disabilities, reduce the strain on informal caregivers, and reduce the need for services by paid caregivers.

Funding

This work was supported, in part, by the National Institute for Disability and Rehabilitation Research, U.S. Department of Education, through grant no. H133B080002 to the University of California, San Francisco, with a subcontract with RTI International.

Acknowledgments

We gratefully acknowledge useful comments by Charlene Harrington, H. Steven Kaye, and Mitchell LaPlante on an earlier presentation of these findings. The authors also gratefully acknowledge the programming support of Richard Pickett. The views expressed in this paper are those of the authors and do not necessarily represent those of the National Institute for Disability and Rehabilitation Research, the University of California, San Francisco, or RTI International.

References

Author notes