Anxiety and Unmet Needs for Assistance With Daily Activities Among Older Adults

Abstract

Objective: To examine the potentially bidirectional relationship between anxiety symptoms and unmet needs for assistance with daily activities among older adults. Method: Data came from the National Health and Aging Trends Study, 2011 through 2016 surveys. The study sample consisted of 3,936 Medicare beneficiaries with activity limitations at baseline, aged 65 or older. Cox proportional hazards regression was used to test the proposed relationship between anxiety symptoms and unmet needs for assistance with daily activities. Result: Having unmet needs increased the risk of the onset of anxiety symptoms, and elevated anxiety symptoms increased the risk of incident unmet needs for assistance with daily activities. Conclusion: Anxiety symptoms and unmet needs form a bidirectional relationship. Integrated mental health and community-based long-term care services may help reduce the burden of late-life anxiety disorders and stressful life incidents contributing to disability.

Keywords

adverse consequences psychological distress personal assistance disability

Introduction

An increasing number of community-dwelling older adults live with disabilities and require personal assistance to safely age in place. Between 8.7 and 12 million Medicare beneficiaries report needs for assistance with activities of daily living (ADL; such as dressing and toileting), instrumental activities of daily living (IADL; such as laundry and shopping), and mobility tasks (such as going out of the home) (Allen, Piette, & Mor, 2014). However, not everyone has access to sufficient informal and formal long-term care and supports. As a result, a large number of older adults with needs for personal assistance experience adverse consequences due to unmet needs, such as wearing soiled clothing, going without a meal, and involuntarily staying indoors (Allen & Mor, 1997; Allen et al., 2014; Desai, Lentzner, & Weeks, 2001; LaPlante, Kaye, Kang, & Harrington, 2004). The percentages of older adults experiencing these adverse consequences range from 7% to 50% depending on the types of activities involved, affecting millions of older adults (Davey, Takagi, Sundström, & Malmberg, 2013; Freedman & Spillman, 2014). Given the nature of the activities involved, having unmet needs has a direct negative impact on the quality of life of older adults with disabilities (Allen et al., 2014). Moreover, reports of adverse consequences due to unmet needs have been found to increase the risk of hospitalization, emergency room visits, nursing home admissions, and premature mortality (Gaugler, Kane, Kane, & Newcomer, 2005; Hass, DePalma, Craig, Xu, & Sands, 2017; Xu, Covinsky, Stallard, Thomas, & Sands, 2012).

Despite the significance of having unmet needs for assistance with daily activities in older adults’ well-being and survival, we have a limited understanding of the causes and outcomes of these unmet needs. Studies that consider populations at high risk of adverse consequences due to unmet needs have focused on sociodemographic and physical risk factors, revealing functional limitations as important risk factors for reporting unmet needs (Allen et al., 2014; Desai et al., 2001; LaPlante et al., 2004). Studies that consider outcomes of unmet needs have focused on physical health and health services utilization (Gaugler et al., 2005; Hass et al., 2017; Xu et al., 2012). Less is known regarding the psychosocial causes and outcomes of unmet needs. Several studies have reported a cross-sectional association between indicators of mental health and unmet needs (Allen & Mor, 1997; Otero, Yébenes, Rodríguez-Laso, & Zunzunegui, 2003; Quail, Wofson, & Lippman, 2011). A regional study of older women in Montreal, Canada, found that unmet needs for IADL were significantly associated with elevated psychological distress (Quail et al., 2011). The Montreal study adopted a cross-sectional design, focused on women, and used a regional sample, limiting its ability to make generalizations and causal speculations. In a recent nationally representative study of older adults in the United States, elevated depressive symptoms predicted higher rates of adverse consequences during the follow-up (Xiang, An, & Heinemann, 2018). These studies suggest that psychiatric symptoms may be both an outcome and a cause of unmet needs. However, this potential bidirectional relationship has not yet been evaluated in a single study.

Anxiety is a common but understudied mental disorder in older adults. Estimates of the prevalence of anxiety disorders meeting clinically diagnostic criteria range from 2.4% to 15% in older adults, with a greater percentage of older adults reporting feelings of anxiety (Bryant, Jackson, & Ames, 2008; Jayasinghe, Rocha, Sheeran, Wyka, & Bruce, 2013). Anxiety symptoms have been linked to increased risk of physical illnesses such as diabetes and hyperlipidemia (Mackenzie, Reynolds, Chou, Pagura, & Sareen, 2011; Porensky et al., 2009), decreased quality of life (Diefenbach, Tolin, & Gilliam, 2011; Porensky et al., 2009), and reduced life expectancy among older adults (He et al., 2015). Emerging evidence suggests that generalized anxiety disorder causes the same or even greater level of impairment in daily activities as depression and other physical illnesses (Mackenzie et al., 2011; Porensky et al., 2009). The disabling effect of anxiety may increase the demand for personal assistance, which increases the risk of having unmet needs. Therefore, persons with anxiety may experience a greater burden of unmet needs.

Older adults with anxiety may also face greater barriers to community-based long-term services and supports, decreasing their supply of personal assistance. Older adults with mental illness are a population often characterized by multiple concurrent chronic conditions, functional and cognitive impairments, mental health challenges, and social vulnerability (Barnett et al., 2012). Persons with complex care needs face continued challenges in accessing person-centered, coordinated care in the fee-for-service models of reimbursement for health care system, (Kuluski, Ho, Hans, & Nelson, 2017). Moreover, because health care is historically oriented to deliver acute and episode care under biomedical and single-disease frameworks, our current health care system rarely addresses needs for long-term social care services, such as assistance with ADL and IADL (Kuluski, Peckham, Williams, & Upshur, 2016). For example, as of 2018, Medicare does not cover long-term nonskilled home care, which provides personal assistance to older adults with ADL and IADL needs. A growing number of older adults in need of personal assistance pay out-of-pocket for private duty home care to get the personal assistance needed to remain safely at home. Persons with mental illness may have fewer resources to pay for these services, and when faced with limited resources, many prioritize paying acute care to address their immediate medical needs over long-term care services and supports. In sum, anxiety may increase the risk of experiencing adverse consequences due to increased demand for personal assistance and decreased supply of personal assistance.

However, experiencing chronic strain due to unmet needs may lead to elevated anxiety symptoms through the stress process. Studies have linked stressful life events with anxiety outcomes (Hannaford, Moore, & Macleod, 2017; Richardson, Simning, He, & Conwell, 2011). Incidences of these hassles such as soiling clothes, hunger, and home confinement have direct implications for the survival and safety of older adults with disabilities and can cause significant stress to these older adults. These acute hassles may repeatedly occur due to insufficient personal assistance and become chronic strains, leading to incident and persistent anxiety symptoms.

Our study expands the literature by examining the bidirectional relationship between anxiety symptoms and unmet needs. To our best knowledge, the potential bidirectional relationship between anxiety and unmet needs for personal assistance has not yet been tested in a single study. The present study uses a nationally representative sample of older adults in the United States to explore the relationship between anxiety symptoms and unmet needs for assistance with daily activities over an extended period. We hypothesize that elevated anxiety symptoms significantly increase the risk of reports of adverse consequences due to unmet needs with personal assistance and that reports of adverse consequences due to unmet needs increase the risk of elevated anxiety symptoms.

Methods

Data

Data came from Round 1 (2011) through Round 6 (2016) of the National Health and Aging Trends Study (NHATS). The NHATS is a nationally representative panel study of Medicare beneficiaries aged 65 and older. NHATS adopts a complex survey design with oversampling of persons in older age groups and African Americans. At baseline (Round 1), 7,609 older adults who lived in the community and not nursing homes completed sample person interviews. These people received annual follow-up interviews regardless of their subsequent residential status. The present study included 3,936 Medicare beneficiaries who reported needs for assistance with any daily activities in the ADL, IADL, or mobility domains at baseline. The study sample was restricted to participants with needs for assistance with daily activities because need is a prerequisite for experiencing adverse consequences. NHATS methodology involves the use of proxy, who was most often a close family, when a respondent was unable to answer for themselves. We included data from proxy respondents because excluding these observations would have caused selection bias (Skolarus et al., 2010).

Measures

Needs for assistance with daily activities

NHATS assesses limitations associated with performing a variety of daily activities, including ADL (including eating, bathing, toileting, and dressing), IADL (including laundry, shopping for groceries or personal items, meal preparation, banking or paying bills, and keeping track of medication), and mobility tasks (going outside the home, getting around inside the home, and getting out of bed). Participants were asked whether they performed each activity with assistance or by themselves in the last month. For participants who performed an activity with assistance, a follow-up question asked whether the reason for assistance was related to health or functioning. For participants who performed an activity alone, a follow-up question asked about the difficulty involved in performing the activity. Participants were classified as having a need for assistance if they received assistance with an activity due to health or functioning or if they reported difficulty performing an activity alone. Two sets of summary indicators were created. One set included three dichotomous indicators of any need in the ADL, IADL, and mobility domains. The other set included three count variables representing the number of needs in the ADL, IADL, and mobility domains.

Adverse consequences due to unmet needs

Questions about adverse consequences related to a specific need were asked only if a participant reported that need. For example, for participants with a need for assistance with toileting, they were asked if they ever wet or soiled themselves during the last month because no one was there to help them or because they had difficulty toileting alone. Adverse consequences in relation to ADL needs included “went without eating,” “went without taking a bath,” “wet or soiled clothes,” and “went without getting dressed.” Adverse consequences in relation to IADL needs included “went without clean laundry,” “went without groceries or personal items,” “went without a hot meal,” “went without handling bills and banking matters,” and “made a mistake in taking prescribed medicines.” Adverse consequences in relation to mobility needs included “had to stay inside,” “did not go to places inside one’s home,” and “had to stay in bed.” Three ADL, IADL, and mobility indicators were created for data analysis. For the purpose of the study, we use “unmet needs” and “adverse consequences due to unmet needs” interchangeably. Older adults have unmet needs if they report at least one adverse consequence due to unmet needs.

Elevated anxiety symptoms

The Generalized Anxiety Disorder–2 (GAD-2) Scale is a validated screener for anxiety disorder. The GAD-2 measures how often participants “felt nervous, anxious, or on edge” and “been unable to stop or control worrying” over the past month on a 4-point Likert-type scale: that is, “not at all” (0), “several days” (1), “more than half the days” (2), and “nearly every day” (3). The GAD-2 score ranges from 0 to 6, with a higher score indicating more severe anxiety symptoms. A cutoff score of 3 is recommended for detecting generalized anxiety, panic, social anxiety, and posttraumatic stress disorder. We created a dichotomous indicator of elevated anxiety symptoms (GAD-2 ≥3) for our analyses.

Covariates

Time-invariant covariates included sex (female or male), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, or other race/multi-race), and education (less than high school, high school graduate, some college but no degree, or college graduate). The time-varying sociodemographic indicator is age group (65-69, 70-74, 75-79, 80-84, 85-89, or 90+). A time-varying indicator of social isolation was included to represent psychosocial status. developed a social isolation measure based on six questions from NHATS, with a composite score ranging from 0 to 6. A score of 0 to 1 indicates “not isolated,” a score of 2 to 3 indicates “somewhat isolated,” and a score of 4 to 6 indicates “socially isolated.” Time-varying indicators of physical health included dementia status, count of self-reported physical illnesses (including heart disease, arthritis, osteoporosis, diabetes, lung disease, stroke, and cancer), and an indicator of past-year hospitalization. NHATS researchers developed a classification of dementia status (no dementia, possible dementia, and probable dementia) based on a self-reported diagnosis of dementia or Alzheimer’s disease, an AD8 Dementia Screening Interview, and cognitive tests (Kasper, Freedman, & Spillman, 2013). In addition, time-varying counts of ADL, IADL, and mobility needs were included in the analysis. Finally, we included a dichotomous indicator of proxy respondent in each survey round to adjust for the potential bias associated with proxy responses (Skolarus et al., 2010).

Data Analysis

The proposed relationship between anxiety symptoms and unmet needs for personal assistance was tested using the Cox proportional hazards regression model. The regression model that estimated the onset of unmet needs (i.e., an incident report of an adverse consequence) as a function of anxiety symptoms was performed on participants free of adverse consequences in the respective domain at baseline. The regression that estimated the onset of anxiety symptoms due to unmet needs (i.e., report of an adverse consequence) was performed on participants free of elevated anxiety symptoms at baseline. In these Cox proportional hazards regression models, a subject was considered a “survivor” until a “failure event” (i.e., adverse consequences due to unmet needs in ADL, IADL, and mobility domains and elevated anxiety symptoms) was reported, if ever. Subjects who died or were lost to follow-up without ever reporting these conditions were censored during their last interview. Subjects who were alive and remained free from a failure event throughout the study period were censored at Round 6. We used baseline NHATS survey design factors (PSU, strata, and sampling weights) to generate nationally representative estimates using the Taylor linearization for variance estimation in Stata 15.0 SE Version (StataCorp, TX).

Results

Table 1 presents weighted descriptive statistics for baseline sample characteristics. About a quarter of Medicare beneficiaries with a need for assistance with ADL, IADL, or mobility tasks were over 85 years of age. The majority of them were female (63.2%) and non-Hispanic White (77.9%). Less than half went to college and 22.7% had a college degree. Half of the study sample were somewhat isolated on the social isolation scale, whereas 20.8% were socially isolated. One in five (20.8%) of the study sample had elevated anxiety symptoms. Dementia was prevalent in the sample: 14.3% had possible dementia and 18.8% had probable dementia. The average count of physical illnesses was 2.9. Nearly one third of the sample (30.6%) were hospitalized during the past 12 months. Half of the study sample had at least one ADL need, nearly three quarters of the sample (73.9%) had at least one IADL need, and two thirds of the sample (66.9%) had at least one mobility need. Among individuals with ADL needs, nearly a quarter (23.9%) reported at least one adverse consequence due to unmet needs. Among individuals with IADL needs, the prevalence of at least one adverse consequence due to unmet needs was 17.3%. Among individuals with mobility needs, 29.6% reported at least one adverse consequence due to unmet needs. About one in 10 participants in the study sample used proxy respondents.

Table 1.

Baseline Sample Characteristics (N = 3,936).

	Weighted statistics
Sociodemographic characteristics
Age groups (%)
65-69 years	20.8 [19.1, 22.6]
70-74 years	20.5 [19.1, 22.0]
75-79 years	19.0 [17.8, 20.3]
80-84 years	18.0 [16.8, 19.2]
85-89 years	14.2 [13.1, 15.4]
≥90 years	7.6 [6.8, 8.4]
Sex (%)
Female	63.2 [61.3, 65.0]
Male	36.8 [35.0, 38.7]
Race/ethnicity (%)
White, non-Hispanic	77.9 [75.5, 80.0]
Black, non-Hispanic	9.5 [8.5, 10.6]
Hispanic	8.3 [7.0, 9.8]
Other	4.4 [3.4, 5.7]
Education (%)
Less than high school	27.6 [25.5, 29.8]
High school	28.5 [26.8, 30.3]
Some college, no degree	21.2 [19.6, 22.9]
College graduate	22.7 [20.5, 25.0]
Psychosocial status
Social isolation (%)
Not socially isolated	13.1 [11.8, 14.5]
Somewhat isolated	54.4 [52.6, 56.3]
Socially isolated	32.5 [30.7, 34.5]
Elevated anxiety symptoms (GAD ≥3) (%)	20.8 [19.1, 22.6]
Physical health
Dementia status (%)
No dementia	66.9 [64.9, 68.9]
Possible dementia	14.3 [13.0, 15.7]
Probable dementia	18.8 [17.4, 20.2]
Chronic physical illness count (mean)	2.9 [2.8, 2.9]
Hospitalized in past 12 months (%)	30.6 [28.7, 32.5]
Proxy respondents (%)	11.1 [9.9, 12.4]
Needs for assistance with daily activities
Had any ADL limitations	54.0 [52.4, 55.6]
Had any adverse consequences associated with unmet ADL needs	23.9 [22.0, 26.1]
Had any IADL limitations (%)	73.9 [72.0, 75.7]
Had any adverse consequences associated with unmet IADL needs (%)	17.3 [15.6, 19.2]
Had any mobility limitations (%)	66.9 [65.0, 68.8]
Had any adverse consequences associated with unmet mobility needs	29.6 [27.4, 31.9]

Note. Estimates presented in the table were mean unless otherwise noted. The complex survey design of NHATS was adjusted in estimates. 95% confidence interval in parentheses. GAD = generalized anxiety disorder; ADL = activities of daily living; IADL = instrumental activities of daily living; NHATS = National Health and Aging Trends Study.

Table 2 presents adjusted hazard ratios for adverse consequences due to unmet needs in the ADL, IADL, and mobility domains. Among persons with ADL needs but initially free of ADL-related adverse consequences at baseline, those with elevated anxiety symptoms were 24% more likely to experience an adverse consequence during the follow-up period (hazard Ratio [HR]=1.24, 95% confidence interval [CI] = [1.02, 1.52], p < .05). A greater ADL and IADL limitation count significantly increased the risk of ADL-related adverse consequences. The risk of the onset of ADL-related adverse consequences was lower among persons aged 90 or over compared with those aged 65 to 90 years (HR = 0.61, 95% CI = [0.38, 0.99], p < .05). None of the other covariates in the regression model were significantly associated with the risk of the onset of ADL-related adverse consequences.

Table 2.

Adjusted Hazard Ratios for Adverse Consequences Due to Unmet ADL, IADL, and Mobility Needs in Cox Proportional Hazards Regressions.

Predictors	Domains of adverse consequences of unmet needs
Predictors	ADL	IADL	Mobility
Elevated anxiety symptoms	1.24 [1.02, 1.52]*	1.31 [1.05, 1.63]*	1.70 [1.34, 2.15]***
Age groups (%)
65-69 years	Reference	Reference	Reference
70-74 years	0.76 [0.41, 1.39]	0.97 [0.65, 1.43]	1.07 [0.76, 1.51]
75-79 years	0.68 [0.39, 1.19]	0.73 [0.50, 1.08]	0.89 [0.58, 1.35]
80-84 years	0.74 [0.46, 1.17]	0.81 [0.54, 1.20]	0.88 [0.61, 1.25]
85-89 years	0.71 [0.43, 1.18]	0.80 [0.54, 1.18]	1.05 [0.78, 1.42]
≥90 years	0.61 [0.38, 0.99]*	0.62 [0.41, 0.94]*	0.71 [0.46, 1.12]
Male sex	1.08 [0.83, 1.41]	1.28 [1.03, 1.58]*	1.25 [0.99, 1.58]
Race/ethnicity
White, non-Hispanic	Reference	Reference	Reference
Black, non-Hispanic	0.82 [0.65, 1.05]	0.89 [0.67, 1.17]	1.10 [0.85, 1.43]
Hispanic	0.81 [0.56, 1.19]	1.07 [0.69, 1.66]	1.27 [0.90, 1.80]
Other	0.99 [0.36, 2.69]	0.57 [0.25, 1.29]	0.71 [0.32, 1.60]
Education
Less than high school	Reference	Reference	Reference
High school	1.06 [0.81, 1.39]	0.84 [0.62, 1.14]	1.19 [0.88, 1.60]
Some college, no degree	1.18 [0.85, 1.65]	0.93 [0.67, 1.31]	1.29 [0.96, 1.74]
College graduate	1.09 [0.74, 1.59]	1.14 [0.83, 1.55]	1.58 [1.19, 2.08]**
Social isolation
Not isolated	Reference	Reference	Reference
Somewhat isolated	0.92 [0.58, 1.46]	0.98 [0.62, 1.54]	0.87 [0.65, 1.16]
Socially isolated	1.07 [0.67, 1.71]	0.96 [0.61, 1.53]	0.71 [0.49, 1.05]
Dementia status
No dementia	Reference	Reference	Reference
Possible dementia	0.91 [0.64, 1.30]	0.92 [0.66, 1.30]	1.05 [0.76, 1.44]
Probable dementia	1.04 [0.77, 1.41]	0.94 [0.70, 1.27]	0.91 [0.70, 1.18]
Chronic physical illness count	1.05 [0.97, 1.13]	0.99 [0.91, 1.08]	1.00 [0.93, 1.08]
Past-year hospitalization	1.03 [0.83, 1.29]	1.22 [0.92, 1.63]	0.97 [0.78, 1.21]
ADL limitation count	2.34 [2.14, 2.56]***	0.95 [0.87, 1.04]	0.96 [0.86, 1.08]
IADL limitation count	1.15 [1.03, 1.30]*	1.69 [1.53, 1.86]***	1.79 [1.64, 1.95]***
Mobility limitation count	1.06 [0.93, 1.21]	1.06 [0.94, 1.20]	1.10 [0.94, 1.28]***
Proxy respondents	0.74 [0.53, 1.04]	0.22 [0.15, 0.33]***	0.38 [0.23, 0.62]***
Sample size	1,674	2,511	1,878
No. of person-years	5,029	7,599	5,776

Note. Cox proportional hazards regressions were performed to estimate the adjusted hazard ratios for adverse consequences due to unmet needs for assistance with ADL, IADL, and mobility tasks among NHATS participants free from ADL-, IADL-, and mobility-related adverse consequences at baseline, respectively. 95% confidence intervals in parentheses. Estimates were weighted adjusting for the complex survey design of NHATS. Covariates were time-varying with the exception of sex, race, and education. ADL = activities of daily living; IADL = instrumental activities of daily living; NHATS = National Health and Aging Trends Study.

p < .05. **p < .01. ***p < .001.

Among persons with IADL needs but initially free of IADL-related adverse consequences at baseline, those with elevated anxiety symptoms were 31% more likely to experience an adverse consequence during the follow-up period (HR = 1.31, 95% CI = [1.05, 1.63], p < .05). The risk of the onset of IADL-related adverse consequences was higher among men compared with women (HR = 1.28, 95% CI = [1.03, 1.58], p < .05) and among persons with greater IADL limitations (HR = 1.69, 95% CI = [1.53, 1.86], p < .001). The risk of the onset of IADL-related adverse consequences was lower among persons aged 90 or over compared with those aged 65 to 69 years (HR = 0.62, 95% CI = [0.41, 0.94], p < .05 (Table 2).

Among persons with mobility needs but initially free of mobility-related adverse consequences at baseline, those with elevated anxiety symptoms were 70% more likely to experience an adverse consequence during the follow-up period (HR = 1.70, 95% CI = [1.34, 2.15], p < .001). The risk of the onset of mobility-related adverse consequences was higher among persons with greater IADL limitations (HR = 1.79, 95% CI = [1.64, 1.95], p < .001), as well as for those with an associates or college degree (HR = 1.58, 95% CI = [1.19, 2.08], p < .01). (Table 2).

Unmet ADL, IADL, and mobility needs were significantly associated with a higher risk of elevated anxiety symptoms. Specifically, the risk of elevated anxiety symptoms during follow-up increased by 34% for individuals with ADL-related adverse consequences (HR = 1.34, 95% CI = [1.01, 1.77], p < .05), by 43% for individuals with IADL-related adverse consequences (HR = 1.43, 95% CI = [1.15, 1.79], p < .01), and by 34% for those with mobility-related adverse consequences (HR = 1.34, 95% CI = [1.05, 1.72], p < .05), compared with individuals without adverse consequences in the respective domain (Table 3).

Table 3.

Adjusted Hazard Ratios for Anxiety in Relation to ADL-, IADL-, and Mobility-Related Consequences in Cox Proportional Hazards Regressions.

Predictors	Domains of adverse consequences of unmet needs
Predictors	ADL	IADL	Mobility
Any ADL-related consequences	1.34 [1.01, 1.77]*	—	—
Any IADL-related consequences	—	1.43 [1.15, 1.79]**	—
Any mobility-related consequences	—	—	1.34 [1.05, 1.72]*
Age groups (%)
65-69 years	Reference	Reference	Reference
70-74 years	1.21 [0.70, 2.09]	1.02 [0.65, 1.59]	0.73 [0.47, 1.14]
75-79 years	1.14 [0.66, 1.98]	1.09 [0.72, 1.66]	0.86 [0.56, 1.32]
80-84 years	1.14 [0.62, 2.08]	1.00 [0.68, 1.46]	0.76 [0.51, 1.13]
85-89 years	1.25 [0.63, 2.49]	0.96 [0.61, 1.52]	0.81 [0.50, 1.33]
≥90 years	0.88 [0.49, 1.58]	0.76 [0.50, 1.15]	0.64 [0.42, 0.98]*
Male sex	1.00 [0.82, 1.22]	0.86 [0.68, 1.10]	0.96 [0.80, 1.17]
Race/ethnicity
White, non-Hispanic	Reference	Reference	Reference
Black, non-Hispanic	0.99 [0.72, 1.35]	0.99 [0.78, 1.26]	0.96 [0.76, 1.22]
Hispanic
Hispanic	1.27 [0.73, 2.23]	1.26 [0.83, 1.92]	1.18 [0.75, 1.86]
Other	1.32 [0.66, 2.64]	1.40 [0.78, 2.50]	1.66 [0.89, 3.11]
Education
Less than high school	Reference	Reference	Reference
High school	1.05 [0.80, 1.38]	0.96 [0.74, 1.26]	0.92 [0.76, 1.13]
Some college, no degree	1.00 [0.71, 1.41]	0.88 [0.68, 1.15]	0.87 [0.66, 1.14]
College graduate	0.69 [0.48, 1.00]*	0.67 [0.50, 0.89]**	0.63 [0.48, 0.83]**
Social isolation
Not isolated	Reference	Reference	Reference
Somewhat isolated	1.30 [0.80, 2.11]	1.30 [0.89, 1.91]	1.25 [0.84, 1.87]
Socially isolated	1.53 [0.96, 2.42]	1.46 [1.01, 2.10]*	1.37 [0.92, 2.05]
Dementia status
No dementia	Reference	Reference
Possible dementia	1.11 [0.79, 1.56]	1.28 [0.91, 1.79]	1.08 [0.77, 1.52]
Probable dementia	1.55 [1.18, 2.03]**	1.66 [1.29, 2.15]**	1.62 [1.27, 2.07]***
Chronic physical illness count	1.16 [1.06, 1.26]**	1.16 [1.06, 1.26]**	1.14 [1.05, 1.25]**
Past-year hospitalization	1.04 [0.86, 1.25]	1.04 [0.85, 1.28]	1.05 [0.86, 1.28]
ADL limitation count	0.94 [0.84, 1.06]	1.07 [0.98, 1.16]	1.10 [1.00, 1.20]
IADL limitation count	1.13 [1.01, 1.25]*	1.09 [0.97, 1.23]	1.08 [0.99, 1.18]
Mobility limitation count	1.27 [1.14, 1.43]***	1.24 [1.12, 1.37]***	1.09 [0.95, 1.26]
Proxy respondents	0.76 [0.57, 1.02]***	0.80 [0.61, 1.04]	0.75 [0.57, 0.99]*
Sample size	1,426	1,932	2,030
No. of person-years	4,689	6,577	5,679

Note. Cox proportional hazards regressions were performed to estimate the adjusted hazard ratios for anxiety in relation to unmet needs for assistance with ADL, IADL, and mobility tasks among NHATS participants with needs for these tasks but free from anxiety at baseline. Estimates were weighted adjusting for the complex survey design of NHATS. Covariates were time-varying with the exception of sex, race, and education, which were time-invariant. ADL = activities of daily living; IADL = instrumental activities of daily living; NHATS = National Health and Aging Trends Study.

p < .05. **p < .01. ***p < .001.

Discussion

One in five older adults with needs for assistance with daily activities had elevated anxiety symptoms. Older adults with elevated anxiety symptoms were more likely to experience incident adverse consequences due to unmet needs for assistance with ADL, IADL, and mobility tasks; unmet needs increased the risk of the onset of anxiety symptoms. These findings suggest that anxiety symptoms and unmet needs form potentially bidirectional relationship.

Although little research exists to evaluate the bidirectional relationship between anxiety and unmet needs, our primary findings are in line with previous studies that have examined the relationship between anxiety and disabilities (de Beurs, Beekman, van Dyk, van Balkom, & van Tilburg, 1999; Hendriks et al., 2014; Lenze et al., 2001; Norton et al., 2012; Sareen et al., 2006). A community-based study of French seniors found that baseline anxiety disorder was associated with an increased risk of incident IADL limitations and that trait anxiety was associated with an increased risk of incident confinement to the bed, home, or neighborhood (Norton et al., 2012). Anxiety has also been consistently linked to increased likelihood of disability (Hendriks et al., 2014; Lenze et al., 2001; Sareen et al., 2006). The disabling effect of anxiety increases the demand for personal assistance, which increases the risk of having unmet needs.

Our primary findings are also consistent with previous studies that have investigated the relationship between stressful life events and anxiety disorders in older adults (Allen & Mor, 1997; Allen et al., 2014; Desai et al., 2001; Hannaford et al., 2017; Richardson et al., 2011). Hassles such as wetting or soiling clothes, making a mistake in taking medications, or not being able to go outside or leave the house are sources of acute and chronic strains that can cause significant stress (Allen & Mor, 1997; Allen et al., 2014; Desai et al., 2001). There is abundant literature documenting the link between stressful and incident and prevalence anxiety symptoms in older adults (Hannaford et al., 2017; Richardson et al., 2011).

Unexpectedly, older adults with probable dementia, despite having greater ADL and IADL limitations in general, were less likely to report incident adverse consequences of unmet needs. This may be due to the sufficient assistance persons with dementia receive. Older adults with dementia tend to receive more attention from caregivers and have more adequate personal assistance from different sources due to the unique characteristics and demands of the disease (Brodaty & Donkin, 2009). Adequacy of personal assistance has been suggested to moderate the relationship between illness severity and unmet needs (Allen et al., 2014). Another possible explanation is that persons with dementia are less aware of their circumstances and less likely to report adverse consequences. That being said, the survey design allows a proxy respondent to answer for the interviewee in cases where they may be unable to answer for themselves. Whether proxy respondents are reliable substitutes for self-reports is questionable (Magaziner, Simonsick, Kashner, & Hebel, 1988). Proxies may have been more apt to underreport unmet needs due to social desirability bias. Because dementia was a common reason why a proxy respondent was needed, responses (including self-reports and proxy reports) from persons with dementia may therefore systematically underreport incidences of unmet needs. That being said, we included an indicator of proxy respondent in multivariable analyses, which should mitigate the reporting bias associated with proxies. Although persons with dementia may have a lower risk of experiencing unmet needs, we found that dementia increased the risk of the onset of anxiety symptoms by 40% to 50%. This finding is consistent with existing literature, which also notes that dementia, anxiety, and depression often have many overlapping symptoms (Seignourel, Kunik, Snow, Wilson, & Stanley, 2008). Measuring anxiety in those who have dementia is complicated, and as such, the finding of increased risk of the onset of anxiety symptoms in those with dementia should not be overemphasized.

Our findings also suggested differential risk patterns for the onset of adverse consequences by age groups. Older age groups tended to have higher risks for the onset of adverse consequences of unmet needs, although not all of these estimates were statistically significant. Particularly, persons aged 90 years or over had significantly lower risk for the onset of IADL-related (p =.026) adverse consequences, and their risk for the onset of ADL-related adverse consequences also had statistical significance (p = .045). Several potential mechanisms may explain this paradoxical finding. Conceptually, the oldest old who remained in independent living at such an advanced age may have good health outcomes overall and a lower demand for personal assistance. It is also possible their advanced age attracts more attention from people, providing them with higher levels of informal support to meet their needs. Statistically speaking, this paradoxical finding may reflect the collinearity between age and activity limitations. It is also not uncommon that including certain covariates can absorb some of the residual variability in the dependent variable and thus increasing the power of the statistical test for the age effect. We further examined this pattern by running hierarchical regression, adding other covariates one by one to a previous model with age. Hierarchical regression showed that only after adding IADL and mobility limitation counts into the regression model did we see a statistically significant difference between the 90+ age group and the 65 to 69 age group. We then tested regression models including interaction terms between age and activity limitations in each domain. These additional regression analyses suggested potential moderation effects of age such that activity limitations had a relatively smaller impact on the risk of adverse consequences in older age groups.

This study has several limitations. First, all measures were self-reported and subject to social desirability bias and reporting errors. For example, adverse consequences due to unmet needs may be underreported because of feelings of shame and embarrassment, particularly in a society that values self-sufficiency. Second, anxiety symptoms were assessed using a GAD-2 screener. We used a dichotomous indicator of elevated anxiety symptoms in our analyses, which may have dampened the obtained effect sizes reported in this study. Despite having satisfactory psychometric properties, GAD-2 is a brief screener with only two questions and not designed to diagnose anxiety disorders. Moreover, anxiety and depression symptoms often co-occur. In our study sample, 10% of all person-years of observations had co-occurring anxiety (GAD-2 ≥3) and depression (Patient Health Questionnaire: PHQ-2 ≥3) symptoms. The prevalence of co-occurrence was 35% among person-years of observations marked with either elevated depression or anxiety symptoms. We examined the potential influence of co-occurring depression symptoms by adding a dichotomous indicator of depression (PHQ-2 ≥3) in our main regression models (for the models set up for our main models, see Tables 2 and 3). Results from the models adjusting for depression symptoms provided similar parameter estimates for the main effects of anxiety symptoms on the onset of adverse consequences as the estimates from our main models. These additional analyses suggested that co-occurrence of depression and anxiety symptoms had a trivial impact on our parameter estimates. Third, it should be noted that while we performed longitudinal analysis, given that the effective samples used in the statistical models were different from each other, we could not conclude that the relationship between anxiety symptoms and unmet needs is truly bidirectional or reciprocal. Fourth, due to a skip pattern of the survey questions, adverse consequences were examined in relation to a specific daily activity and only probed if a respondent reported a need with that activity. A person may not perceive having a need but experience adverse consequences. Moreover, the list of daily activities and associated adverse consequences examined in our study was not exhaustive. There may be other daily activities and adverse consequences associated with impaired ability to perform these activities that are meaningful for the well-being of older adults. Fifth, adverse consequences were categorized into three domains due to the low occurrence of some consequences, preventing us from examining the nuances within each domain. Finally, limitations in the ADL, IADL, and mobility domains often co-occur. This collinearity may cause somewhat unstable parameter estimates, interfering the determination of the precise effect of activity limitations in each domain. The parameter estimates regarding the effect of activity limitations from our analyses should be interpreted with a grain of salt. Nevertheless, the collinearity of the activity limitations has little effect on the parameter estimates regarding the effects of anxiety symptoms and the overall fit of our models.

Overall, our findings supported our hypotheses that anxiety symptoms and unmet needs for personal assistance have a bidirectional relationship such that anxiety symptoms can increase the risk of the onset of adverse consequences due to unmet needs in ADL, IADL, and mobility domains, and adverse consequences in these domains can increase the risk of the onset of elevated anxiety symptoms. The bidirectional relationship between anxiety symptoms and unmet needs for personal assistance suggests that addressing one condition may alleviate the burden of the other. Efforts aimed at preventing and treating anxiety symptoms may improve efficiency by targeting older adults with high risk of unmet needs or recent experiences of adverse consequences due to unmet needs, given their increased risk of the onset of anxiety symptoms. Future intervention studies are needed to examine the effectiveness and cost-effectiveness of anxiety prevention and treatment targeted at this high-risk population. Improving informal and formal support to better address ADL, IADL, and mobility limitations may have added benefits to reducing the occurrence or severity of anxiety symptoms among older adults. For example, caregiver interventions supporting family caregivers of older adults with anxiety disorders and respite care programs can consider adding adverse consequences as a secondary outcome measure to examine whether improving informal and formal support leads to improved anxiety symptoms. Our findings reinforce the notion that needs for mental health services and community-based long-term services and supports are interconnected (Kuluski et al., 2017; Kuluski et al., 2016). In addition to testing interventions focused on individual- and family-level risk factors, policies supporting the integration of health, mental health, and social services (e.g., personal assistance) may yield improved mental health and quality-of-life outcomes at a larger scale. Several models aimed at integrating mental health care in aging services such as meals-on-wheels and senior centers have shown promising results (Choi et al., 2014; Ciechanowski et al., 2004; Gitlin et al., 2013; Quijano et al., 2007). However, most of these new models focus on addressing depression; the effectiveness of these programs on alleviating anxiety symptoms is unclear. Future studies should expand the current focus on depression to include late-life anxiety disorders.

Conclusion

Anxiety symptoms and unmet needs for personal assistance have a bidirectional relationship among community-dwelling older adults with disabilities. Improving access to community-based long-term care services and supports may reduce the burden of late-life anxiety disorders, whereas access to effective mental health treatments may reduce stressful incidents associated with unmet needs for personal assistance. This calls for innovative care delivery models that better coordinate and integrate mental health and long-term care services.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by a grant from the University of Michigan School of Social Work Gerontology Learning Community (AZ) and a grant from the National Institutes of Health, University of Michigan Older Americans Independence Center Research Education Core (Grant number: AG024824).

ORCID iD

Ashley Zuverink

References

Allen

S. M.

Mor

(1997). The prevalence and consequences of unmet need: Contrasts between older and younger adults with disability. Medical Care, 35, 1132-1148. doi:10.1097/00005650-199711000-00005

Allen

S. M.

Piette

E. R.

Mor

(2014). The adverse consequences of unmet need among older persons living in the community: Dual-eligible versus Medicare-only beneficiaries. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, 69, S51-S58. doi:10.1093/geronb/gbu124

Barnett

Mercer

S. W.

Norbury

Graham

Wyke

Guthrie

(2012). Epidemiology of multimorbidity and implications for health care, research, and medical education: A cross-sectional study. The Lancet, 380, 37-43. doi:10.1016/S0140-6736(12)60240-2

Brodaty

Donkin

(2009). Family caregivers of people with dementia. Dialogues in Clinical Neuroscience, 11, 217-228. Available from https://www.dialogues-cns.org

Bryant

Jackson

Ames

(2008). The prevalence of anxiety in older adults: Methodological issues and a review of the literature. Journal of Affective Disorders, 109, 233-250. doi:10.1016/j.jad.2007.11.08

Choi

N. G.

Marti

C. N.

Bruce

M. L.

Hegel

M. T.

Wilson

N. L.

Kunik

M. E.

(2014). Six-month postintervention depression and disability outcomes of in-home telehealth problem-solving therapy for depressed, low-income homebound older adults. Depression and Anxiety, 31, 653-661. doi:10.1002/da.22242

Ciechanowski

Wagner

Schmaling

Schwartz

Williams

Diehr

. . . LoGerfo

(2004). Community-integrated home-based depression treatment in older adults: A randomized controlled trial. Journal of the American Medical Association, 291, 1569-1577. doi:10.1001/jama.291.13.1569

Davey

Takagi

Sundström

Malmberg

(2013). (In)formal support and unmet needs in the National Long-Term Care Survey. Journal of Comparative Family Studies, 44, 437-453. Retrieved from https://soci.ucalgary.ca/jcfs/

de Beurs

Beekman

A. T. F.

van Balkom

A. J. L. M.

Deeg

D. J. H.

van Dyck

D. J. H. R.

van Tilburg

. (1999). Consequences of anxiety in older persons: Its effect on disability, well-being, and use of health services. Psychological Medicine, 29, 583-593. doi:10.1017/S0033291799008351

10.

Desai

M. M.

Lentzner

H. R.

Weeks

J. D.

(2001). Unmet need for personal assistance with activities of daily living among older adults. The Gerontologist, 41, 82-88. doi:10.1093/geront/41.1.82

11.

Diefenbach

G. J.

Tolin

D. F.

Gilliam

C. M.

(2011). Impairments in life quality among clients in geriatric home care: Associations with depressive and anxiety symptoms. International Journal of Geriatric Psychiatry, 27, 828-835. doi:10.1002/gps.2791

12.

Freedman

V. A.

Spillman

B. C.

(2014). Disability and care needs among older Americans. The Milbank Quarterly, 92, 509-541. doi:10.1111/1468-0009.12076

13.

Gaugler

J. E.

Kane

R. L.

Kane

R. A.

Newcomer

(2005). Unmet care needs and key outcomes in dementia. Journal of the American Geriatrics Society, 53, 2098-2105. doi:10.111/j.1532-5415.2005.00495.x

14.

Gitlin

L. N.

Harris

L. F.

McCoy

M. C.

Chernett

N. L.

Pizzi

L. T.

Jutkowitz

. . . Hauck

W. W.

(2013). A home-based intervention to reduce depressive symptoms and improve quality of life in older African Americans: A randomized trial. Annals of Internal Medicine, 159, 243-252. doi:10.7326/0003-4819-159-4-201308200-00005

15.

Hannaford

Moore

Macleod

F. J.

(2017). What a difference a year makes: Comparing relationships between stressful life events, mood, and life satisfaction among older adults, and their working-age counterparts. Aging & Mental Health, 22, 1658-1665. doi:10.1080/13607863.2017.1387761

16.

Hass

DePalma

Craig

B. A.

Sands

L. P.

(2017). Unmet need for help with activities of daily living disabilities and emergency department admissions among older Medicare recipients. The Gerontologist, 57, 206-210. doi:10.1093/geront/gnv142

17.

Craig

B. A.

Covinsky

K. E.

Stallard

Thomas

III . . . Sand

L. P.

(2015). Unmet need for ADL assistance is associated with mortality among older adults with mild disability. Journals of Gerontology, Series A: Biological Sciences & Medical Sciences, 70, 1128-1132. doi:10.1093/gerona/glv028

18.

Hendriks

S. M.

Spijker

Licht

C. M. M.

Hardeveld

de Graaf

Batelaan

N. M.

. . . Beekman

A. T. F.

(2014). Long-term disability in anxiety disorders. BMC Psychiatry, 16, Article 248. doi:10.1186/s12888-016-0946-y

19.

Jayasinghe

Rocha

L. P.

Sheeran

Wyka

Bruce

M. L.

(2013). Anxiety symptoms in older home health care recipients: Prevalence and associates. Home Health Care Services Quarterly, 32, 163-177. doi:10.1080/01621424.2013.8813885

20.

Kasper

J. D.

Freedman

V. A.

Spillman

(2013). Classification of persons by dementia status in the National Health and Aging Trends Study (Technical Paper #5). Baltimore, MD: Johns Hopkins University School of Public Health. Available from www.NHATS.org

21.

Kuluski

J. W.

Hans

P. K.

Nelson

M. L. A.

(2017). Community care for people with complex care needs: Bridging the gap between health and social care. International Journal of Integrated Care, 17(4), 2-12. doi:10.5334/ijic.2944

22.

Kuluski

Peckham

Williams

A. P.

Upshur

R. E. G.

(2016). What gets in the way of person-centred care for people with multimorbidity? Lessons from Ontario, Canada. Healthcare Quarterly, 19, 17-23. doi:10.12927/hcq.2016.24694

23.

LaPlante

M. P.

Kaye

H. S.

Kang

Harrington

(2004). Unmet need for personal assistance services: Estimating the shortfall in hours of help and adverse consequences. The Journals of Gerontology, Series B, Psychological Sciences & Social Services, 59, S98-S108. doi:10.1093/geronb/49.2.s98

24.

Lenze

E. J.

Rogers

J. C.

Martire

L. M.

Muslant

B. H.

Rollman

B. L.

Dew

M. A.

. . . Reynolds

C. F.

(2001). The association of late-life depression and anxiety with physical disability: A review of the literature and prospects for future research. American Journal of Geriatric Psychiatry, 9, 113-135. doi:10.1097/001942-200105000-00004

25.

Mackenzie

C. S.

Reynolds

Chou

Pagura

Sareen

(2011). Prevalence and correlates of generalized anxiety disorder in a national sample of older adults. American Journal of Geriatric Psychiatry, 19, 3015-3315. doi:10.1097/JGP.0b013e318202bc62

26.

Magaziner

Simonsick

E. M.

Kashner

T. M.

Hebel

J. R.

(1988). Patient-proxy response comparability on measures of patient health and functional status. Journal of Clinical Epidemiology, 41, 1065-1074. doi:10.1016/0895-4356(88)90076-5

27.

Norton

Ancelin

M. L.

Stewart

Berr

Ritchie

Carriere

(2012). Anxiety symptoms and disorder predict activity limitations in the elderly. Journal of Affective Disorders, 141, 276-285. doi:10.1016/j.jad.2012.04.002

28.

Otero

Yébenes

M. J. G.

Rodríguez-Laso

Zunzunegui

M. V.

(2003). Unmet home care needs among community-dwelling elderly people in Spain. Aging Clinical and Experimental Research, 15, 234-242. doi:10.1007/BF03324504

29.

Porensky

E. K.

Dew

M. A.

Karp

J. F.

Skidmore

Rollman

B. L.

Shear

M. K.

Lenze

E. J.

(2009). The burden of late-life generalized anxiety disorder: Effects on disability, health-related quality of life, and healthcare utilization. American Journal of Geriatric Psychiatry, 17, 473-482. doi:10.1097/JGP.0b013e31819b87b2

30.

Quail

J. M.

Wofson

Lippman

(2011). Unmet need for assistance to perform activities of daily living and psychological distress in community-dwelling elderly women. Canadian Journal on Aging, 30, 591-602. doi:10.1017/S0714980811000493

31.

Quijano

L. M.

Stanley

M. A.

Petersen

N. J.

Casado

B. L.

Steinberg

E. H.

Cully

J. A.

. . . Wilson

N. L.

(2007). Healthy IDEAS: A depression intervention delivered by community-based case managers serving older adults. Journal of Applied Gerontology, 26, 139-156. doi:10.1177/0733464807299354

32.

Richardson

T. M.

Simning

Conwell

(2011). Anxiety and its correlates among older adults accessing aging services. International Journal of Geriatric Psychiatry, 26, 31-38. doi:10.1002/gps.2474

33.

Sareen

Jacobi

Cox

B. J.

Belik

Clara

Stein

M. B.

(2006). Disability and poor quality of life associated with comorbid anxiety disorders and physical conditions. Archives of Internal Medicine, 166, 2109-2116. doi:10.1001/archinte.166.19.2109

34.

Seignourel

P. J.

Kunik

M. E.

Snow

Wilson

Stanley

(2008). Anxiety in dementia: A critical review. Clinical Psychology Review, 28, 1071-1082. doi:10.1016/j.cpr.2008.02.008

35.

Skolarus

L. E.

Sánchez

B. N.

Morgenstern

L. B.

Garcia

N. M.

Smith

M. A.

Brown

D. L.

Lisabeth

L. D.

(2010). Validity of proxies and correction for proxy use when evaluating social determinants of health in stroke patients. Stroke, 41, 510-515.

36.

Xiang

Heinemann

(2018). Depression and unmet needs for assistance with daily activities among community-dwelling older adults. Gerontologist, 58, 428-437. doi:10.1093/geront/gnw262

37.

Covinsky

K. E.

Stallard

Thomas

III. Sands

L. P.

(2012). Insufficient help for activity of daily living disabilities and risk of all-cause hospitalization. Journal of the American Geriatrics Society, 60, 927-933. doi:10.1111/j.1532-5415.2012.03926.x