Abstract
This study aimed to provide a national profile of homebound and semi-homebound older adults with depressive symptoms and to compare risk factors of depressive symptoms by home-bound status. A sample of 1,885 homebound and semi-home-bound older adults was selected from Round 1 of the National Health and Aging Trends Study (NHATS). The prevalence of depressive symptoms was 43.9% in homebound older adults and 28.1% in semi-homebound older adults, representing over 830,000 and 1.4 million individuals in the population, respectively. Nearly two-thirds of homebound and over half of semi-homebound older adults with clinically significant depressive symptoms also had significant anxiety symptoms. Results from logistic regression showed that younger age, certain medical morbidities, severity of functional limitations, and pain were common risk factors for depressive symptoms among home-bound and semi-homebound older adults. Some differences in the risk factor profile emerged between the homebound and the semi-homebound populations. Alleviating the burden of depression in the semi-homebound population may focus on early prevention that considers the diversity of this population. Home-based, integrated programs of health and mental health services that simultaneously address the medical, psychiatric, and neurologic comorbidities and disabilities of homebound older adults are needed to meet the complex needs of this population.
Keywords: Anxiety, depression, home-based care, homebound
Introduction
Homebound older adults refer to those who are confined to home or require substantial assistance leaving home. The number of homebound older adults is projected to increase as the number of older adults increases rapidly (Qiu et al., 2010). Community-based studies since the 2000s consistently documented higher rates of depression among homebound older adults compared with their community-dwelling counterparts not confined to homes (Bruce et al., 2002; Choi & McDougall, 2007; Choi, Teeters, Perez, Farar, & Thompson, 2010; Ell, Unutzer, Aranda, Sanchez, & Lee, 2005; Li & Conwell, 2007; Richardson et al., 2012; Sirey et al., 2008). Estimate of current major depressive disorders in older adults receiving home healthcare was as high as 13.5% (Bruce et al., 2002). In comparison, the prevalence of 12-month major depressive disorder was 2.7% among the general older adult population based on results from the National Epidemiologic Survey on Alcoholism and Related Conditions (Hasin, Goodwin, Stinson, & Grant, 2005). Depression complicates the management of other medical disorders, increases fall risks and hospitalization rates, and upsurges healthcare costs for homebound older adults (Qiu et al., 2010). Even mild or subthreshold depressive symptoms can cause serious functional impairment (Lyness, Chapman, McGriff, Drayer, & Duberstein, 2009). However, depression treatment, particularly specialty mental health services, is underutilized by homebound older adults (Gum, Iser, & Petkus, 2010).
Medical morbidity, disability, and social isolation are known risk factors of depression in late life (Reynolds et al., 2012), making homebound older adults at increased risk for depression. Studies examining risk factors of depression in homebound older adults reported that clinical factors such as medical morbidities, functional limitations, cognitive impairment, and pain were strong and consistent correlates of depression (Choi & McDougall, 2007; Choi et al., 2010; Ell et al., 2005; Li & Conwell, 2007; Richardson et al., 2012; Sirey et al., 2008). To a lesser degree, risks of depression differed by sociodemographic characteristics in the homebound population (Choi et al., 2010; Ell et al., 2005; Li & Conwell, 2007).
Previous studies of depression and its correlates among homebound older adults have several limitations. Most studies used regional samples and focused on older adults receiving services from Medicare home health agencies, home-delivered meal programs, and other community long-term care or aging services providers (Bruce et al., 2002; Choi & McDougall, 2007; Choi et al., 2010; Ell et al., 2005; Li & Conwell, 2007; Richardson et al., 2012; Sirey et al., 2008). This approach includes individuals who receive these services but may not meet the definition of homebound and systematically excludes those who are homebound but do not receive services. In addition, none of studies cited previously examined home-bound status as a continuum. Because homebound is a result of declining physical capacity and lack of personal assistance (Ornstein et al., 2015), the degree of confinement depends on the severity of impairment and the availability of personal assistance. A recent descriptive study using the NHATS showed significant differences in sociodemographic and clinical characteristics by the degree of confinement measured by the frequency of leaving homes (Ornstein et al., 2015). Patterns of risk factors may differ across the homebound continuum, which may require different strategies to prevent and treat depression.
This study addresses these limitations by (1) providing a national profile of homebound and semi-homebound older adults with depressive symptoms, and (2) comparing socioeconomic and clinical risk factors of clinically significant depressive symptoms between homebound and semi-homebound older adults. An ecological model of aging (Satariano, 2006) and previous studies guided the selection of risk factor, including sociodemographic characteristics, disease and comorbidities, physical functioning and impairment, cognitive status, falls and injuries.
Design and methods
Data
Data came from Round 1 (2011) of the NHATS public use files. The NHATS is a nationally representative panel study of Medicare beneficiaries aged 65 or older with oversamples of the oldest-old and African Americans. A total of 7,777 older adults who lived in the community (including traditional community settings, retirement communities, and alternative residential care) completed sample person interviews at baseline. This study examined 1,360 semi-homebound and 561 homebound older adults. After excluding cases with missing data on the depression screener (n = 36), the final analytical sample consisted of 1,885 older adults (1,341 classified as semi-homebound and 544 as homebound).
Measures
Homebound status
Participants were homebound if they “never” or “rarely” went out the home in the last month. Participants were semi-homebound if they (1) received help going out of the home and would “never,” “rarely,” or “sometimes” go outside by themselves; or (2) did not receive help going out of the home but reported “a lot,” “some,” or “a little” difficulty leaving the home by themselves. This classification of homebound status is modified based on the measures developed by Ornstein et al. (2015).
Depressive symptoms
Depressive symptoms were assessed using the Patient Health Questionnaire-2 (PHQ-2) (Löwe, Kroenke, & Gräfe, 2005). The PHQ-2 measures how often the participant has been bothered by “little interest or pleasure in doing things” and “feeling down, depressed or hopeless” over the last month. Reponses were on a four-point Likert scale, from “not at all” (0), “several days” (1), “more than half the days” (2), to “nearly every day” (3). Clinically significant depressive symptoms were defined as a PHQ-2 score of 3 and above. A cut-off score of 3 has a sensitivity of 0.87 and a specificity of 0.78 for major depressive disorder; and has a sensitivity of 0.79 and specificity of 0.86 for any depressive disorder (Löwe et al., 2005).
Sociodemographic characteristics
Sociodemographic variables included age, sex, race/ethnicity, nativity (foreign born or US born), English proficiency, education, living arrangement (live alone or live with others), and Medicare–Medicaid dual eligibility. Limited English proficiency refers to a response of “not well” or “not at all” when asked how well one understood or spoke English.
Disease and comorbidities
Participants reported whether a doctor had ever told them that they had the following conditions: heart attack/heart disease, arthritis, osteoporosis, diabetes, lung disease, stroke, and cancer. Anxiety symptoms were assessed using the Generalized Anxiety Disorder Scale (GAD-2) (Kroenke, Spitzer, Williams, & Löwe, 2010). 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 four-point Likert scale. A cut-off score of 3 has been associated with symptoms of generalized anxiety, panic, social anxiety, and post-traumatic stress disorder (Kroenke et al., 2010).
Physical functioning and impairment
We examined needs for assistance with activities of daily living (ADL; including eating, bathing, toileting, and dressing) and instrumental activities of daily living (IADL; including laundry, grocery shopping, meal preparation, banking or paying bills, and keeping track of medication). Participants were considered to have a need for assistance for an activity if they (1) received assistance with the activity for health or functioning reasons, or (2) performed the activity alone with difficulty. Other indicators of physical impairment included self-reported limitations in activities due to pain, and difficulty in hearing, seeing, and communication. Participants were classified as having hearing difficulty if with a hearing aid, they could not hear well enough to use telephone, carry on a conversation in a room with a radio or TV playing, or carry on a conversation in a quiet room. Participants were classified as having a seeing difficulty if with glasses or contacts or vision aids, they could not see well enough to recognize someone across the street, watch television across the room, or see well enough to read newspaper print. Communication difficulty referred to problems in speaking or making themselves understood during communication in the last month.
Cognitive status
NHATS classifies participants into three groups—no dementia, possible dementia, and probable dementia—based on self-reported diagnosis of dementia or Alzheimer's disease, AD8 Dementia Screening Interview, and cognitive tests (Kasper, Freedman, & Spillman, 2013).
Falls and injuries
Participants responded if they had fallen down in last month. A fall refers to any fall, slip, or trip in which one loses balances and land on the floor or ground or at a lower level.
Data analysis
Bivariate comparisons of sample characteristics by depression status stratified by homebound status were conducted using χ2 tests. Multiple logistic regression was used to identify risk factors of clinically significant depressive symptoms stratified by homebound status. Selection of independent variables in multiple logistic regression was based on the following criteria: (1) p-value cut-off point of 0.25 from χ2 tests. Traditional p-value of 0.05 can fail to identify important variables (Bendel & Afifi, 1977); (2) the inclusion of a variable leads to a significant improvement in the overall model fit as evaluated by the likelihood ratio test; (3) the variable to be included is not highly correlated with other independent variables. This process was repeated to reach the most parsimonious model in each homebound group using the same set of risk factors.
A summary score of the total count of ADL and IADL needs was used in multivariable analysis to address the collinearity between individual items and to improve statistical power. Race/ethnicity, nativity, living arrangements, and dementia status were excluded from multiple logistic regression due to collinearity and/or failing to improve model fit for both homebound and semi-homebound groups. Anxiety symptoms were not included as a covariate in multiple logistic regression models. Anxiety and depression consistently co-occur. A dimensional, rather than a categorical, classification of anxiety and depression in older adults has been suggested (Schoevers, Deeg, van Tilburg, & Beekman, 2005). Given the conceptual similarity between depression and anxiety, inclusion of anxiety in the regression model could bias the estimates of the impact of other covariates.
Because direct comparisons of odds ratios from different samples can be problematic (Mood, 2010), comparison of the effect sizes of risk factors between the homebound and the semi-homebound groups involved estimating and testing the differences of the average marginal effects (AMEs). The AME refers to the average change in the probability of depression when a predictor increases by one unit while holding other covariates constant. Analyses accounted for the sampling weights and design factors of NHATS, using the svy command of Stata 12.0, SE version (StataCorp, College Station, TX, USA). Stata margins command was used to estimate and compare AMEs across the homebound groups.
Results
Characteristics of homebound and semi-homebound older adults
The prevalence rate of clinically significant depressive symptoms was 43.9% (95% CI = 38.9–49.0%) among homebound older adults, representing 836,249 individuals in the population. The rate of clinically significant depressive symptoms was 28.1% (95% CI = 25.4–30.9%) among semi-home-bound older adults, representing 1,465,010 individuals in the population. In comparison, 1 in 10 (95% CI = 9.0–11.2%) non-homebound older adults had clinically significant depressive symptoms (p < .001).
Table 1 presents the characteristics of study sample stratified by homebound and depression status. Among homebound Medicare beneficiaries, people with clinically significant depressive symptoms were more likely to have heart disease (46.6% vs. 31.1%), osteoporosis (43.6% vs. 31.9%), limitations in activities due to pain (68.6% vs. 47.9%), difficulty in seeing (33.8% vs. 23.5%) and communication (37.3% vs. 14.1%), probable dementia (50% vs. 39.7%), and falls in the past month (30.% vs. 19.8%). People with depressive symptoms reported significantly higher rates of ADL and IADL needs across all activities, with the only exception being needs with bathing. The relative difference in the prevalence rate was the biggest for anxiety symptoms, with 65% of homebound older adults with depression reporting clinically significant anxiety symptoms as compared with 18% among those without depression.
Table 1.
Characteristics of homebound and semi-homebound medicare beneficiaries by depression status.
| Homebound (n = 544) | Semi-homebound (n = 1,341) | |||||
|---|---|---|---|---|---|---|
|
|
|
|||||
| Depressed | Not depressed | p | Depressed | Not depressed | p | |
| Socio-demographic characteristics | ||||||
| Age groups (%) | .168 | .036 | ||||
| 65–69 years | 13.1 (.03) | 8.5 (.03) | 22.3 (.03) | 18.4 (.01) | ||
| 70–74 years | 17.9 (.03) | 16.8 (.03) | 21.5 (.03) | 16.4 (.02) | ||
| 75–79 years | 12.1 (.02) | 11.5 (.02) | 20.9 (.02) | 19.4 (.02) | ||
| 80–84 years | 23.1 (.03) | 16.5 (.02) | 16.7 (.02) | 18.8 (.01) | ||
| 85–89 years | 20.9 (.03) | 28.9 (.03) | 11.9 (.01) | 16.8 (.01) | ||
| 90 years or over | 12.9 (.02) | 17.8 (.02) | 6.6 (.01) | 10.4 (.01) | ||
| Sex (%) | .837 | .007 | ||||
| Female | 74.8 (.04) | 75.7 (.03) | 59.6 (.03) | 69.3 (.02) | ||
| Male | 25.2 (.04) | 24.4 (.03) | 40.4 (.03) | 30.7 (.02) | ||
| Race/ethnicity (%) | .616 | .005 | ||||
| White, Non-Hispanic | 69.4 (.03) | 66.4 (.03) | 67.5 (.03) | 77.7 (.01) | ||
| Black, Non-Hispanic | 9.8 (.02) | 13.6 (.02) | 11.9 (.01) | 9.9 (.01) | ||
| Hispanic | 15.1 (.04) | 14.6 (.03) | 15.0 (.02) | 8.3 (.01) | ||
| Other | 5.7 (.02) | 5.4 (.01) | 5.6 (.02) | 4.1 (.01) | ||
| Foreign born (%) | 18.1 (.04) | 20.2 (.03) | .630 | 22.9 (.03) | 13.5 (.01) | <.001 |
| Limited English proficiency (%) | 13.9 (.03) | 14.3 (.03) | .907 | 15.8 (.02) | 7.4 (.01) | <.001 |
| Education (%) | .065 | |||||
| Less than high school | 43.8 (.04) | 39.9 (.03) | 45.0 (.04) | 30.3 (.02) | <.001 | |
| High school | 32.5 (.04) | 25.9 (.04) | 24.6 (.03) | 27.7 (.01) | ||
| Some college, no degree | 11.4 (.03) | 20.9 (.03) | 17.2 (.02) | 21.5 (.02) | ||
| College graduate | 12.3 (.03) | 13.3 (.02) | 13.2 (.02) | 20.5 (.02) | ||
| Live alone | 31.3 (.04) | 41.9 (.04) | .055 | 34.6 (.03) | 35.5 (.02) | .780 |
| Medicare-Medicaid dual eligibility (%) | 31.2 (.03) | 29.5 (.03) | .660 | 32.0 (.03) | 18.6 (.01) | <.001 |
| Disease and comorbidities | ||||||
| Heart disease (%) | 46.6 (.4) | 31.1 (.03) | .002 | 44.0 (.03) | 35.6 (.02) | .032 |
| Arthritis (%) | 74.4 (.04) | 69.7 (.03) | .279 | 78.5 (.03) | 67.2 (.02) | .002 |
| Osteoporosis (%) | 43.6 (.04) | 31.9 (.03) | .008 | 35.0 (.03) | 30.5 (.02) | .267 |
| Diabetes (%) | 31.6 (.04) | 32.0 (.03) | .947 | 47.9 (.03) | 27.1 (.02) | <.001 |
| Lung disease (%) | 31.3 (.04) | 26.3 (.03) | .261 | 27.7 (.03) | 20.8 (.01) | .012 |
| Stroke (%) | 24.5 (.04) | 21.6 (.03) | .518 | 30.2 (.03) | 14.6 (.01) | <.001 |
| Cancer (%) | 26.0 (.03) | 20.8 (.02) | .140 | 27.7 (.03) | 27.4 (.02) | .898 |
| Anxiety symptoms in last month (%) | 65.3 (.04) | 18.3 (.03) | <.001 | 51.9 (.03) | 14.2 (.02) | <.001 |
| Physical functioning and impairment | ||||||
| ADL needs | ||||||
| Eating | 46.7 (.05) | 27.6 (.03) | <.001 | 29.7 (.03) | 17.5 (.10) | <.001 |
| Bathing | 68.1 (.04) | 60.8 (.03) | .141 | 62.9 (.03) | 45.5 (.02) | <.001 |
| Toileting | 55.5 (.04) | 30.3 (.03) | <.001 | 42.8 (.03) | 25.3 (.02) | <.001 |
| Dressing | 69.3 (.03) | 55.2 (.04) | .005 | 65.2 (.03) | 48.4 (.02) | <.001 |
| IADL needs | ||||||
| Laundry | 79.9 (.03) | 61.6 (.03) | <.001 | 64.5 (.03) | 48.0 (.02) | <.001 |
| Shopping for groceries | 86.9 (.03) | 76.1 (.03) | .022 | 75.8 (.03) | 61.7 (.02) | <.001 |
| Meal preparation | 83.7 (.03) | 68.3 (.03) | .003 | 70.8 (.03) | 52.7 (.02) | <.001 |
| Banking or paying bills | 73.0 (.04) | 56.3 (.03) | .002 | 56.4 (.03) | 38.5 (.02) | <.001 |
| Keeping track of medication | 62.7 (.04) | 47.0 (.03) | .004 | 51.2 (.03) | 35.9 (.02) | <.001 |
| Limited in activities due to pain (%) | 68.6 (.04) | 47.9 (.04) | <.001 | 73.7 (.02) | 54.0 (.02) | <.001 |
| Difficulty in hearing (%) | 30.3 (.04) | 23.6 (.03) | .110 | 28.6 (.03) | 20.8 (.01) | .004 |
| Difficulty in seeing (%) | 33.8 (.03) | 23.5 (.03) | .032 | 23.2 (.03) | 16.6 (.01) | .039 |
| Difficulty in communication (%) | 37.3 (.04) | 14.1 (.02) | <.001 | 22.1 (.03) | 12.2 (.01) | <.001 |
| Cognitive status (%) | .136 | <.001 | ||||
| No dementia | 35.9 (.04) | 44.0 (.04) | 48.1 (.03) | 66.6 (.01) | ||
| Possible dementia | 14.1 (.03) | 16.3 (.02) | 19.5 (.02) | 13.8 (.01) | ||
| Probable dementia | 50.0 (.04) | 39.7 (.04) | 32.4 (.03) | 19.5 (.01) | ||
| Any falls in the past month (%) | 30.7 (.04) | 19.8 (.03) | .031 | 28.7 (.03) | 20.7 (.01) | .007 |
Note. ADL = activities of daily living. IADL = instrumental activities of daily living. Linearized standard error in parentheses. Sampling weights and design factors were accounted for when estimating prevalence rates and means.
Among semi-homebound Medicare beneficiaries, the following sociodemographic groups reported a higher rate of depressive symptoms: men, Hispanics, the foreign-born population, people with limited English proficiency, and people with lower levels of education. The rate of clinically significant depressive symptoms decreased as age increased. Medicare–Medicaid dual eligibles reported a higher rate of clinically significant depressive symptoms compared with the Medicare only group (32.0% vs. 18.6%). People with clinically significant depressive symptoms reported higher rates of heart disease (44.0% vs. 35.6%), arthritis (78.5% vs. 67.2%), diabetes (47.9% vs. 27.1%), lung disease (27.7% vs. 20.8%), stroke (30.2% vs. 14.6%), all ADL and IADL needs, anxiety symptoms (51.9% vs. 14.2%), limitations in activities due to pain (73.7% vs. 54.0%), hearing (28.6% vs. 20.8%), seeing (23.2% vs. 16.6%), and communication difficulties (22.1% vs. 12.2%), possible (19.5% vs. 13.8%) or probable dementia (32.4% vs. 19.5%), and falls in the past month (28.7% vs. 20.7%).
Risk factors of depressive symptoms among homebound and semi-homebound older adults
Table 2 shows the results from multiple logistic regression on the correlates of depressive symptoms stratified by homebound status. Two separate logistic regressions were estimated, one for the homebound group and the other for semi-homebound group. Among homebound older adults, significant risk factors of depression included a diagnosis of heart disease, higher number of ADL and IADL needs, limitations in activities due to pain, and difficulty in communication. Homebound older adults aged 85 years and older had lower odds of depression compared with the 65–69 years group. Some college education was associated with lower odds of depression relative to less than high school education.
Table 2.
Risk factors of depressive symptoms among homebound and semi-homebound medicare beneficiaries.
| Predictors | Homebound | Semi-homebound | AME differenceb | ||
|---|---|---|---|---|---|
|
|
|
||||
| Odds ratio | AMEa | Odds ratio | AMEa | ||
| Age groups (reference 65–69 years) | |||||
| 70–74 years | 0.80 (0.25, 2.55) | −0.04 | 1.02 (0.59, 1.78) | 0.00 | −0.05 |
| 75–79 years | 0.31 (0.08, 1.26) | −0.23 | 0.83 (0.50, 1.39) | −0.03 | −0.20 |
| 80–84 years | 0.48 (0.14, 1.63) | −0.14 | 0.75 (0.41, 1.37) | −0.05 | −0.09 |
| 85–89 years | 0.22 (0.07, 0.62)** | −0.30** | 0.59 (0.34, 1.01) | −0.08 | −0.21* |
| 90 years or over | 0.26 (0.09, 0.74)* | −0.26* | 0.60 (0.32, 1.12) | −0.08 | −0.18 |
| Male sex (reference: female) | 0.78 (0.46, 1.32) | −0.05 | 1.56 (1.11, 2.18)* | 0.08* | −0.12* |
| Limited English proficiency | 1.28 (0.64, 2.58) | 0.05 | 1.70 (1.16, 2.48)** | 0.09** | −0.06 |
| Education (reference: <high school) | |||||
| High school | 1.22 (0.77, 1.91) | 0.04 | .90 (0.54, 1.48) | −0.02 | 0.06 |
| Some college, no degree | 0.42 (0.21, 0.87)* | −0.16* | .70 (0.42, 1.15) | −0.06 | −0.10 |
| College graduate | 0.71 (0.31, 1.64) | −0.07 | .59 (0.38, 1.00) | −0.09 | 0.02 |
| Medicare–Medicaid dual eligibles | 0.98 (0.62, 1.56) | 0.00 | 1.35 (0.92, 2.00) | 0.05 | −0.05 |
| Disease and comorbidities | |||||
| Heart disease | 1.92 (1.22, 3.03)** | 0.13** | 0.94 (0.65, 1.37) | −0.01 | 0.14** |
| Arthritis | 0.92 (0.51, 1.65) | −0.02 | 1.45 (1.00, 2.09)* | 0.06* | −0.08 |
| Osteoporosis | 1.41 (0.94, 2.11) | 0.07 | 1.06 (0.73, 1.56) | 0.01 | 0.06 |
| Diabetes | 0.75 (0.45, 1.24) | −0.06 | 1.99 (1.51, 2.63)*** | 0.12*** | −0.18** |
| Lung disease | 0.87 (0.53, 1.41) | −0.03 | 1.09 (0.75, 1.58) | 0.01 | −0.04 |
| Stroke | 0.80 (0.45, 1.41) | −0.04 | 2.10 (1.50, 2.95)*** | 0.14*** | −0.18* |
| Cancer | 1.17 (0.70, 1.97) | 0.03 | 1.04 (0.73, 1.47) | 0.01 | 0.03 |
| Count of ADL and IADL needs | 1.17 (1.06, 1.30)** | 0.03** | 1.15 (1.09, 1.21)*** | 0.02*** | 0.01 |
| Limited in activities due to pain | 1.92 (1.11, 3.30)* | 0.13* | 1.80 (1.30, 2.49)** | 0.10** | 0.03 |
| Difficulty in hearing | 1.22 (0.75, 1.99) | 0.04 | 1.19 (0.84, 1.69) | 0.03 | 0.01 |
| Difficulty in seeing | 1.36 (0.72, 2.53) | 0.06 | 1.08 (0.68, 1.72) | 0.01 | 0.05 |
| Difficulty in communication | 2.58 (1.49, 4.46)** | 0.20** | 1.26 (0.80, 2.01) | 0.04 | 0.16* |
| Any falls in the past month | 1.53 (0.81, 2.91) | 0.08 | 1.06 (0.74, 1.51) | 0.01 | 0.08 |
Note. 95% confidence intervals in parentheses. Sampling weights and design factors were accounted for in regression estimates.
AME, average marginal effect. AME shows the average change in the probability of depression when the predictor increases by one unit while holding other covariates constant. For categorical predictors, one unit increase is the change from the reference category.
A significant test result indicates that the impact of the corresponding predictor differs by homebound status.
p < .05,
p < .01,
p < .001.
Among semi-homebound older adults, significant risk factors of depression included male sex, limited English proficiency, and a diagnosis of arthritis, diabetes or stroke, higher number of ADL and IADL needs, and limitations in activities due to pain. Semi-homebound older adults aged 85–89 years had lower odds of depression as compared with 65–69 years old.
Comparisons of AMEs revealed different patterns of risk factors of depression by homebound status. Relative to the 65–69 years age group, the 85–89 years age group had an average of 0.30 lower probability of depression among homebound older adults, compared with a 0.08 lower probability of depression among semi-homebound older adults (AME difference = −0.21, p < .05). Semi-homebound men had an average of 0.08 higher probability of depression than semi-homebound women, whereas no significant sex difference emerged in the homebound group (AME difference = −0.12, p < .05). Homebound older adults with heart disease had a 0.13 higher probability of depression, whereas heart disease was not a significant risk factor of depression in the semi-homebound group (AME difference = 0.14, p < .01). Semi-homebound older adults with a diagnosis of diabetes or stroke had a 0.12 and 0.14 higher probability of depression, respectively. Neither diabetes nor stroke was a significant risk factor of depression in the homebound group (AME difference for diabetes = −0.18 for both, p < .05). Communication difficulty was associated with a 0.20 higher probability of depression in the homebound group, whereas no significant effect of such was found in the semi-homebound group (AME difference = 0.16, p < .05).
Discussion
In a nationally representative sample of Medicare beneficiaries aged 65 or older, clinically significant depressive symptoms affected 44% of homebound older adults (representing more than 830,000 individuals), and 28% of semi-homebound older adults (representing more than 1.4 million individuals). Nearly two-thirds of homebound and over half of semi-homebound older adults with clinically significant depressive symptoms also had significant anxiety symptoms. Half of homebound and nearly one-third of semi-home-bound older adults had probable dementia. Younger age, certain medical morbidities, severity of functional limitations, and pain appeared as common risk factors for depressive symptoms among homebound and semi-home-bound older adults.
Study results revealed some differences in the types of risk factors and the relative impact of risk factors between the homebound and the semi-home-bound groups. Disparities in depression rates relating to race/ethnicity and nativity were pronounced in the semi-homebound group. Bivariate comparisons showed that Hispanics, the foreign-born population, and people with limited English proficiency had disproportionately higher rates of depressive symptoms among semi-homebound older adults. In multivariable analysis, English proficiency remained a significant and robust predictor of depressive symptoms in the semi-homebound group. In the homebound group, however, race/ethnicity, nativity, and English proficiency were not significant correlates of depressive symptoms. Heart disease was the only medical morbidity associated with depressive symptoms in the homebound group. In comparison, arthritis, diabetes, and stroke were independently associated with depressive symptoms in the semi-homebound group. Underlying mechanisms for these differences are unknown. A possible explanation is that the homebound group represents a homogenous group characterized by severe disability and impairment. Disability and illness become the most salient predictor of health and mental well-being in this group, leaving little variation in the residuals for sociodemographic differences.
Alleviating the burden of depression in the semi-homebound population may focus on early prevention. Reynolds et al. (2012) detailed the rationale for prevention of depression in late life. For semi-homebound older adults, depression prevention may have the added benefits of preventing or delaying the process of becoming fully homebound. Primary care and aging services agencies that reach the semi-homebound population are well positioned to deliver preventive interventions. Priority areas of depression-prevention research for older adults include demonstrating cost-effectiveness, building community partnerships, and adaptations to the language and culture of an increasingly diverse older adult population (Reynolds et al., 2012).
The substantial burden of depressive symptoms in the homebound population calls for innovative, person-centered mental health delivery models to meet their needs. A number of home-based mental health services programs have been developed that conveniently bring mental health services into clients' homes (Reifler & Bruce, 2014). Opportunities exist to integrate mental health services in home-based primary care programs as they are increasingly available to homebound older adults (Reckrey et al., 2015). For example, the Veterans Affairs has fully integrated mental health services into the VA home-based primary care team (Karlin, Karel, & Meeks, 2013). Other integrated home-based primary care programs have recently been evaluated (Reckrey et al., 2015). In addition, alternative settings such as home health and aging services settings provide tremendous opportunities for improving the recognition and treatment of depression for the homebound population (Choi, Sirey, & Bruce, 2013). Several innovative interventions integrated in these alternative settings have been developed and assessed, such as the Depression CAREPATH—a home-health integrated, collaborative depression care model (Bruce et al., 2015), and an in-home telehealth problem-solving therapy model using Meals on Wheels as the delivery platform (Choi et al., 2014). Adpatations of depression screening and treatment programs to other community long-term care settings such as in-home services, adult day care, residential care facilities, and retirement communities should be developed and assessed in future studies.
Depression in the homebound and semi-homebound population is characterized by substantial medical, psychiatric, and neurologic comorbidities, as well as various types of functional impairment and disability. However, many late-life depression prevention and treatment studies systematically exclude people with certain comorbidities (Reynolds et al., 2012). Depression interventions not specifically targeted to the homebound population that concurrently address these comorbidities are limited. Existing depression treatment studies targeted to the homebound population have rarely addressed these comorbidities (Bruce et al., 2015; Choi et al., 2014; Ell et al., 2007). The Community Aging in Place, Advancing Better Living for Elders (CAPABLE) intervention is one of the handful of interventions that simultaneously address multiple domains of functional limitations (Szanton, Leff, Wolff, Roberts, & Gitlin, 2016). Preliminary analyses showed that depressive symptoms improved in 53% of the CAPABLE participants at 5-month follow-up. However, depressive symptoms worsened in slightly less than one-third of participants. A randomized controlled trial of CAPABLE is being conducted and will provide a more-rigorous evaluation of the intervention. We would like to echo Reynolds et al.'s (2012) recommendations for treatment studies to recruit clinically representative participants with medical, psychiatric, and neurologic comorbidities and concurrently address the burden of these clinical symptoms in addition to depressive symptoms. Future studies should also examine whether simultaneously addressing these comorbidities will lead to improvement in the alleviation of depressive symptoms compared with addressing depressive symptoms alone.
This study has some limitations. All measures were self-reported and subject to recall bias and reporting errors. Estimates of the homebound population may be biased downward if non-respondents and older adults not covered by Medicare were more likely to be homebound. Only the PHQ-2 was available to assess depression. The PHQ-2 is intended as an initial screening tool and requires follow-up with the PHQ-9 to probe severity of depression.
The burden of depression among homebound older adults is even greater than previously reported. Alleviating the burden of depression in the semi-homebound population may focus on early prevention that considers the diversity of this population. Home-based, integrated programs of health and mental health services that simultaneously address the medical, psychiatric, and neurologic comorbidities and disabilities of homebound older adults are needed to meet the complex needs of this population.
Acknowledgments
Funding: Jessica Brooks is supported by a T32 postdoctoral research fellowship in geriatric mental health services from National Institute of Mental Health (grant number MH073553) (PI: Stephen Bartels).
References
- Bendel RB, Afifi AA. Comparison of stopping rules in forward “stepwise” regression. Journal of the American Statistical Association. 1977;72(357):46–53. [Google Scholar]
- Bruce ML, McAvay GJ, Raue PJ, Brown EL, Meyers BS, Keohane DJ, et al. Weber C. Major depression in elderly home health care patients. American Journal of Psychiatry. 2002;159(8):1367–1374. doi: 10.1176/appi.ajp.159.8.1367. [DOI] [PubMed] [Google Scholar]
- Bruce ML, Raue PJ, Reilly CF, Greenberg RL, Meyers BS, Banerjee S, et al. Leon AC. Clinical effectiveness of integrating depression care management into medicare home health: The Depression CAREPATH Randomized trial. JAMA Internal Medicine. 2015;175(1):55–64. doi: 10.1001/jamainternmed.2014.5835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choi NG, Marti CN, Bruce ML, Hegel MT, Wilson NL, Kunik ME. Six-Month postintervention depression and disability outcomes of in-home telehealth problem-solving therapy for depressed, low-income homebound older adults. Depression and Anxiety. 2014;31(8):653–661. doi: 10.1002/da.2014.31.issue-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choi NG, McDougall GJ. Comparison of depressive symptoms between homebound older adults and ambulatory older adults. Aging & Mental Health. 2007;11(3):310–322. doi: 10.1080/13607860600844614. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choi NG, Sirey JA, Bruce ML. Depression in homebound older adults: Recent advances in screening and psychosocial interventions. Current Translational Geriatrics and Gerontology Reports. 2013;2:16–23. doi: 10.1007/s13670-012-0032-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choi NG, Teeters M, Perez L, Farar B, Thompson D. Severity and correlates of depressive symptoms among recipients of meals on wheels: Age, gender, and racial/ethnic difference. Aging & Mental Health. 2010;14(2):145–154. doi: 10.1080/13607860903421078. [DOI] [PubMed] [Google Scholar]
- Ell K, Unutzer J, Aranda M, Gibbs NE, Lee PJ, Xie B. Managing depression in home health care: A randomized clinical trial. Home Health Care Services Quarterly. 2007;26(3):81–104. doi: 10.1300/J027v26n03_05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ell K, Unutzer J, Aranda M, Sanchez K, Lee PJ. Routine PHQ-9 depression screening in home health care: Depression, prevalence, clinical and treatment characteristics and screening implementation. Home Health Care Services Quarterly. 2005;24(4):1–19. doi: 10.1300/J027v24n04_01. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gum AM, Iser L, Petkus A. Behavioral health service utilization and preferences of older adults receiving home-based aging services. American Journal of Geriatric Psychiatry. 2010;18(6):491–501. doi: 10.1097/JGP.0b013e3181c29495. [DOI] [PubMed] [Google Scholar]
- Hasin DS, Goodwin RD, Stinson FS, Grant BF. Epidemiology of major depressive disorder: Results from the National Epidemiologic Survey on Alcoholism and Related Conditions. Archives of General Psychiatry. 2005;62(10):1097–1106. doi: 10.1001/archpsyc.62.10.1097. [DOI] [PubMed] [Google Scholar]
- Karlin BE, Karel MJ, Meeks S. National integration of mental health providers in VA home-based primary care: An innovative model for mental health care delivery with older adults. The Gerontologist. 2013;54(5):868–879. doi: 10.1093/geront/gnt142. [DOI] [PubMed] [Google Scholar]
- Kasper JD, Freedman VA, Spillman B. Baltimore, MD: Johns Hopkins University School of Public Health; 2013. Classification of persons by dementia status in the National Health and Aging Trends Study. Technical Paper #5. Retrieved from www.NHATS.org. [Google Scholar]
- Kroenke K, Spitzer RL, Williams JB, Löwe B. The patient health questionnaire somatic, anxiety, and depressive symptom scales: A systematic review. General Hospital Psychiatry. 2010;32(4):345–359. doi: 10.1016/j.genhosppsych.2010.03.006. [DOI] [PubMed] [Google Scholar]
- Li LW, Conwell Y. Mental health status of home care elders in Michigan. The Gerontologist. 2007;47(4):528–534. doi: 10.1093/geront/47.4.528. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Löwe B, Kroenke K, Gräfe K. Detecting and monitoring depression with a two-item questionnaire (PHQ-2) Journal of Psychosomatic Research. 2005;58(2):163–171. doi: 10.1016/j.jpsychores.2004.09.006. [DOI] [PubMed] [Google Scholar]
- Lyness JM, Chapman BP, McGriff J, Drayer R, Duberstein PR. One-year outcomes of minor and subsyndromal depression in older primary care patients. International Psychogeriatrics. 2009;21(01):60–68. doi: 10.1017/S1041610208007746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mood C. Logistic regression: Why we cannot do what we think we can do, and what we can do about it. European Sociological Review. 2010;26(1):67–82. doi: 10.1093/esr/jcp006. [DOI] [Google Scholar]
- Ornstein KA, Leff B, Covinsky KE, Ritchie CS, Federman AD, Roberts L, et al. Szanton SL. Epidemiology of the homebound population in the United States. JAMA Internal Medicine. 2015;175(7):1180–1186. doi: 10.1001/jamainternmed.2015.1849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Qiu WQ, Dean M, Liu T, George L, Gann M, Cohen J, Bruce ML. Physical and mental health of homebound older adults: An overlooked population. Journal of the American Geriatrics Society. 2010;58(12):2423–2428. doi: 10.1111/j.1532-5415.2010.03161.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reckrey JM, DeCherrie LV, Dugue M, Rosen A, Soriano TA, Ornstein K. Meeting the mental health needs of the homebound: A psychiatric consult service within a home-based primary care program. Care Management Journals. 2015;16(3):122–128. doi: 10.1891/1521-0987.16.3.122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reifler BV, Bruce ML. Home-based mental health services for older adults: A review of ten model programs. The American Journal of Geriatric Psychiatry. 2014;22(3):241–247. doi: 10.1016/j.jagp.2012.12.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reynolds CF, Cuijpers P, Patel V, Cohen A, Dias A, Chowdhary N, et al. Albert SM. Early intervention to reduce the global health and economic burden of major depression in older adults. Annual Review of Public Health. 2012;33:123–135. doi: 10.1146/annurev-publhealth-031811-124544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Richardson TM, Friedman B, Podgorski C, Knox K, Fisher S, He H, Conwell Y. Depression and its correlates among older adults accessing aging services. The American Journal of Geriatric Psychiatry. 2012;20(4):346–354. doi: 10.1097/JGP.0b013e3182107e50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Satariano W. Epidemiology of aging: An ecological approach. Burlington, MA: Jones & Bartlett Learning; 2006. [Google Scholar]
- Schoevers RA, Deeg D, Van Tilburg W, Beekman A. Depression and generalized anxiety disorder: Co-occurrence and longitudinal patterns in elderly patients. American Journal of Geriatric Psychiatry. 2005;13(1):31–39. doi: 10.1097/00019442-200501000-00006. [DOI] [PubMed] [Google Scholar]
- Sirey JA, Bruce ML, Carpenter M, Booker D, Reid MC, Newell KA, Alexopoulos GS. Depressive symptoms and suicidal ideation among older adults receiving home delivered meals. International Journal of Geriatric Psychiatry. 2008;23(12):1306. doi: 10.1002/gps.2070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Szanton SL, Leff B, Wolff JL, Roberts L, Gitlin LN. Home-based care program reduces disability and promotes aging in place. Health Affairs. 2016;35(9):1558–1563. doi: 10.1377/hlthaff.2016.0140. [DOI] [PubMed] [Google Scholar]