- Split View
-
Views
-
Cite
Cite
Hans-Werner Wahl, Vera Heyl, Philipp M. Drapaniotis, Karl Hörmann, Jost B. Jonas, Peter K. Plinkert, Klaus Rohrschneider, Severe Vision and Hearing Impairment and Successful Aging: A Multidimensional View, The Gerontologist, Volume 53, Issue 6, December 2013, Pages 950–962, https://doi.org/10.1093/geront/gnt013
- Share Icon Share
Abstract
Purpose: Previous research on psychosocial adaptation of sensory-impaired older adults has focused mainly on only one sensory modality and on a limited number of successful aging outcomes. We considered a broad range of successful aging indicators and compared older adults with vision impairment, hearing impairment, and dual sensory impairments and without sensory impairment. Design and Methods: Data came from samples of severely visually impaired (VI; N = 121), severely hearing-impaired (HI; N = 116), dual sensory-impaired (DI; N = 43), and sensory-unimpaired older adults (UI; N = 150). Participants underwent a wide-ranging assessment, covering everyday competence, cognitive functioning, social resources, self-regulation strategies, cognitive and affective well-being, and 4-year survival status (except the DI group). Results: The most pronounced difference among groups was in the area of everyday competence (lowest in VI and DI). Multigroup comparisons in latent space revealed both similar and differing relationship strengths among health, everyday competence, social resources, self-regulation, and overall well-being, depending on sensory status. After 4 years, mortality in VI (29%) and HI (30%) was significantly higher than in UI (20%) at the bivariate level, but not after controlling for confounders in a multivariate analysis. Implications: A multidimensional approach to the understanding of sensory impairment and psychosocial adaptation in old age reveals a complex picture of loss and maintenance.
Baltes and Baltes (1990), Rowe and Kahn (1998), and Ryff (1989) have argued that successful aging requires a multidimensional view, which should include health and longevity, everyday competence, cognitive performance, social engagement, coping, and well-being-oriented indicators. Baltes, Lindenberger, and Staudinger (2006) emphasized that human development contains elements of loss and gain across the life span, although loss increases as people age. This paper examines whether and how chronic conditions, such as severe vision and hearing impairment (Heyl & Wahl, 2012; Horowitz, Brennan, & Reinhardt, 2005; Li-Korotky, 2012), are threats to successful aging.
Previous Research
Evidence that both hearing and vision impairments are accompanied by heightened multimorbidity and an increased rate of mortality is inconsistent. For example, although Anstey, Luscz, Giles, and Andrews (2001) and Appollonio, Carabellese, Magini, Frattola, and Trabucchi (1995) have found that visual and hearing declines over time predict death in very old age, other studies have detected no significant relationship after adjusting for confounders such as age, sex, and education (Ostbye, Steenhuis, Wolfson, Walton, & Hill, 1999).
Age-related vision impairment has been found to be closely associated with significantly lower everyday competence, as visual capacity is a critical prerequisite for such behaviors (Burmedi, Becker, Heyl, Wahl, & Himmelsbach, 2002a; Heyl & Wahl, 2001). Hearing loss has not been found to have a major affect on behavioral functioning (Rudberg, Furner, Dunn, & Cassel, 1993), although some have observed its negative long-term consequences on everyday functioning (Dalton et al., 2003).
Reduced sensory function is accompanied by a decrease in cognitive performance in older adults, according to Wahl and Heyl (2003). Differences in the relationship of vision and hearing impairments with cognitive performance have not been identified, although there is some evidence of a stronger link with vision loss (Anstey, Luszcz, & Sanchez, 2001).
Existing evidence supports the view that social relations of older adults with severe vision and hearing loss are not fundamentally different from those of sensory-unimpaired (UI) individuals. Reinhardt (1996) found that visually impaired (VI) older adults named an average of 5.4 persons with whom they had intimate relationships within their family network and 3.5 persons within their friendship network, similar to UI older adults (Wagner, Schütze, & Lang, 1999). Socioemotional selectivity theory (Carstensen, 2006) posits that older adults—including sensory-impaired individuals—invest a great deal in maintaining close relationships because of a more limited future-time perspective in old age. Neither vision nor hearing impairment are thought to affect the experience of loneliness dramatically (Burmedi, Becker, Heyl, Wahl, & Himmelsbach, 2002b; Wahl & Tesch-Römer, 2001), although a recent study found an increased risk of loneliness (Nachtegaal et al., 2009). Hearing impairment negatively impacts social communication and carries a strong stigma, however (Dalton et al., 2003; Wahl & Tesch-Römer, 2001; Wallhagen, 2010).
Successful aging also includes self-regulation skills, such as the ability to work toward life goals tenaciously and adjust and adapt goals flexibly (e.g., Brandtstädter, 2009; Brandtstädter & Renner, 1990). Evidence is emerging of flexible goal adjustment for mental health outcomes in those with vision impairment (Boerner, 2004; Wahl, Becker, Schilling, Burmedi, & Himmelsbach, 2005). However, the combination of tenacity and flexibility in pursuing life goals has never been examined in comparisons among those with vision impairment, hearing impairment, and no sensory impairment.
VI older adults have shown evidence of diminished well-being compared with UI older adults (Horowitz & Reinhardt, 2000; Nachtegaal et al., 2009; Wahl, Schilling, Oswald, & Heyl, 1999), although a related meta-analysis found only small effect sizes (Pinquart & Pfeiffer, 2011). Differences in well-being between hearing-impaired (HI) and unimpaired older adults appear to be small or non-existent (Tesch-Römer, 2001). The “well-being paradox” in old age, assuming that adaptive resources to maintain well-being are pronounced and efficient even under adverse conditions (Kunzmann, Little, & Smith, 2000), may also apply to sensory-impaired older adults (Schilling & Wahl, 2006; Schilling, Wahl, Horowitz, Reinhardt, & Boerner, 2011).
Finally, the overall psychosocial situation of those with dual sensory impairment, which affects about 20% of those 70 years and older, has been found to be worse than for those with single impairments, particularly in everyday functioning (Brennan, Horowitz, & Su, 2005).
Research Questions and Expectations
We examine the impact of severe vision and hearing impairments, as well as dual sensory impairment, on a range of successful aging indicators, covering health and everyday functioning, cognitive performance, social integration, self-regulation, and well-being. Given Baltes and coworkers (2006) view that development in old age always contains elements of loss and gain, we expect sensory impairment to be accompanied by both losses and maintenance. Regarding loss, we predict that sensory-impaired older adults will rate low in health-related and everyday competence indicators, and that vision impairment may have even greater negative impact than hearing impairment. In terms of cognitive functioning, we expect lower performance in severely sensory-impaired older adults as compared with UI older adults. Regarding self-regulation, we predict lower tenacious goal pursuit in those with vision impairment, assuming that deficient everyday functioning blocks goal attainment. In terms of maintenance, socioemotional selectivity theory suggests that sensory-impaired older adults will demonstrate social relationships and loneliness similarly to UI older adults. With respect to well-being, the well-being paradox and the potential for late-life adaptation may limit differences. We also test major predictors of an overall index of well-being in latent data space, influenced by classic work of Campbell (1976) and Smith et al. (1999), and we examine possible differences among those with vision impairment, hearing impairment, and no sensory impairment. We exclude cognitive function, not generally regarded as an important predictor of well-being (Kotter-Grühn et al., 2010; Smith et al., 1999). Finally, we examine survival status 4 years after baseline.
Method
Sampling Strategy and Sample Description
The ethics commission of the German Psychological Society approved this study in 2004. Eligible participants were in advanced old age (75–94 years), community-dwelling, and without major cognitive impairment; those with sensory impairments had been impaired for at least 2 years prior to enrollment. Best-corrected distance and/or near visual acuity of participants in the VI sample also had to be no more than .30 in the better eye (approximately 20/70 in the U.S. metric). Regarding the HI sample, the average hearing loss in decibels (dB HL) at frequencies of 500, 1000, and 2000 Hz had to be at least 35 dB HL in the better ear. The dual sensory-impaired (DI) sample met both the VI and HI inclusion criteria. UI participants were expected to have adequate vision and hearing.
See Table 1 for a detailed description of the recruitment process. Letters of invitation were sent to outpatients at university eye and ear, nose, and throat clinics in Heidelberg and Mannheim, Germany. Of those who contacted the project, 121 comprised the final VI sample (49.2% of 833) and 116 comprised the final HI sample (53.7% of 413); 22 of the eligible VI patients and 21 of the eligible HI patients were assigned to the DI sample due to dual sensory impairment. The UI sample was randomly drawn from addresses in the Heidelberg city register; 150 comprised the final UI sample (15.8% of 950).
. | VI . | HI . | DI . | UI . |
---|---|---|---|---|
Potential participant pool | 833 outpatients | 413 outpatients | — | 950 community-dwelling adults |
Recruitment source | University Eye Clinic (Heidelberg, Germany) | University Eye and Ear, Nose, and Throat Clinic (Mannheim, Germany) | — | Heidelberg city register |
Agreed to participate/eligible for inclusion | 246 | 216 | — | 907 |
Reason for drop-out | ||||
Too ill | 23 | 20 | — | 206 |
No interest | 11 | 21 | — | 181 |
Too effortful | 10 | 5 | — | 10 |
No time | 2 | 6 | — | 33 |
No reason | 8 | 5 | — | 195 |
Deceased | 14 | 10 | — | — |
Did not meet criteria | 30 | 8 | — | — |
Unable to contact | — | — | — | 124 |
Reassigned due to dual sensory impairmenta | 22 | 21 | — | — |
Excluded due to cognitive impairmentb | 5 | 4 | — | 8 |
Final study sample | 121 | 116 | 43a | 150 |
. | VI . | HI . | DI . | UI . |
---|---|---|---|---|
Potential participant pool | 833 outpatients | 413 outpatients | — | 950 community-dwelling adults |
Recruitment source | University Eye Clinic (Heidelberg, Germany) | University Eye and Ear, Nose, and Throat Clinic (Mannheim, Germany) | — | Heidelberg city register |
Agreed to participate/eligible for inclusion | 246 | 216 | — | 907 |
Reason for drop-out | ||||
Too ill | 23 | 20 | — | 206 |
No interest | 11 | 21 | — | 181 |
Too effortful | 10 | 5 | — | 10 |
No time | 2 | 6 | — | 33 |
No reason | 8 | 5 | — | 195 |
Deceased | 14 | 10 | — | — |
Did not meet criteria | 30 | 8 | — | — |
Unable to contact | — | — | — | 124 |
Reassigned due to dual sensory impairmenta | 22 | 21 | — | — |
Excluded due to cognitive impairmentb | 5 | 4 | — | 8 |
Final study sample | 121 | 116 | 43a | 150 |
Notes: VI = visually impaired; HI = hearing impaired; DI = dual sensory impaired; UI = sensory unimpaired.
aDiverted from original VI and HI study samples and assigned to DI due to dual sensory impairment.
bCut-off score of 7 based on the Cognitive Impairment Test (Brooke & Bullock, 1999; German version in Schneekloth & Leven, 2003).
. | VI . | HI . | DI . | UI . |
---|---|---|---|---|
Potential participant pool | 833 outpatients | 413 outpatients | — | 950 community-dwelling adults |
Recruitment source | University Eye Clinic (Heidelberg, Germany) | University Eye and Ear, Nose, and Throat Clinic (Mannheim, Germany) | — | Heidelberg city register |
Agreed to participate/eligible for inclusion | 246 | 216 | — | 907 |
Reason for drop-out | ||||
Too ill | 23 | 20 | — | 206 |
No interest | 11 | 21 | — | 181 |
Too effortful | 10 | 5 | — | 10 |
No time | 2 | 6 | — | 33 |
No reason | 8 | 5 | — | 195 |
Deceased | 14 | 10 | — | — |
Did not meet criteria | 30 | 8 | — | — |
Unable to contact | — | — | — | 124 |
Reassigned due to dual sensory impairmenta | 22 | 21 | — | — |
Excluded due to cognitive impairmentb | 5 | 4 | — | 8 |
Final study sample | 121 | 116 | 43a | 150 |
. | VI . | HI . | DI . | UI . |
---|---|---|---|---|
Potential participant pool | 833 outpatients | 413 outpatients | — | 950 community-dwelling adults |
Recruitment source | University Eye Clinic (Heidelberg, Germany) | University Eye and Ear, Nose, and Throat Clinic (Mannheim, Germany) | — | Heidelberg city register |
Agreed to participate/eligible for inclusion | 246 | 216 | — | 907 |
Reason for drop-out | ||||
Too ill | 23 | 20 | — | 206 |
No interest | 11 | 21 | — | 181 |
Too effortful | 10 | 5 | — | 10 |
No time | 2 | 6 | — | 33 |
No reason | 8 | 5 | — | 195 |
Deceased | 14 | 10 | — | — |
Did not meet criteria | 30 | 8 | — | — |
Unable to contact | — | — | — | 124 |
Reassigned due to dual sensory impairmenta | 22 | 21 | — | — |
Excluded due to cognitive impairmentb | 5 | 4 | — | 8 |
Final study sample | 121 | 116 | 43a | 150 |
Notes: VI = visually impaired; HI = hearing impaired; DI = dual sensory impaired; UI = sensory unimpaired.
aDiverted from original VI and HI study samples and assigned to DI due to dual sensory impairment.
bCut-off score of 7 based on the Cognitive Impairment Test (Brooke & Bullock, 1999; German version in Schneekloth & Leven, 2003).
Table 2 indicates that study groups were of comparable age, marital status, and household composition, with minor differences in gender, education, and income. As expected, VI, HI, and DI appeared unique in both objective and perceived vision and hearing performance, compared with the remaining groups. No significant differences were observed in the cognitive screening results.
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) or % . | M (SD) or % . | M (SD) or % . | M (SD) or % . | |||
Age in years | 82.60 (4.63) | 82.69 (5.08) | 83.37 (4.51) | 82.26 (4.50) | .575 | f = .07 |
Gender (% women) | 58.68A | 41.38B | 39.53B | 49.33A,B | .032 | w = .14 |
Education in years | 9.21A,B (1.84) | 9.74A (2.21) | 8.77B (1.70). | 9.65A,B (2.03) | .014 | f = .16 |
Income (% >2000 Euro) | 40.34A,B | 45.05A,B | 29.27A | 50.67B | .045 | w = .13 |
Marital status (% married) | 39.67 | 52.59 | 37.21 | 51.33 | .072 | w = .13 |
Living arrangements (% living alone) | 57.02 | 51.72 | 58.14 | 47.33 | .366 | w = .09 |
Distance visiona | .13A (.10) | .69B (.19) | .14A (.11). | .72B (.19) | .000 | f = .86 |
Near visiona | .11A (.13) | .62B (.19) | .11A (.11) | .64B (.19) | .000 | f = .83 |
Subjective visiona,b | 4.13A (.80) | 2.45B (.93) | 4.24A (.79) | 2.39B (.82) | .000 | f = .71 |
Duration of vision impairment in years | 6.57 (10.29) | — | 8.72 (14.66) | — | — | — |
Hearing impairment (dB HL)c | 22.31A (7.75) | 47.95B (10.22) | 45.49B (9.86) | 20.65A (7.86) | .000 | f = .82 |
Subjective hearingb,c | 2.32A (.74) | 3.33B (.95) | 3.02B (.87) | 2.33A (.82) | .000 | f = .47 |
Duration of hearing impairment in years | — | 11.78 (13.27) | 9.00 (12.97) | — | — | — |
Cognitive status/6 cognitive impairment testb | 2.49 (2.76) | 2.42 (2.55) | 2.42 (2.85) | 2.19 (2.35) | .492 | f = .05 |
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) or % . | M (SD) or % . | M (SD) or % . | M (SD) or % . | |||
Age in years | 82.60 (4.63) | 82.69 (5.08) | 83.37 (4.51) | 82.26 (4.50) | .575 | f = .07 |
Gender (% women) | 58.68A | 41.38B | 39.53B | 49.33A,B | .032 | w = .14 |
Education in years | 9.21A,B (1.84) | 9.74A (2.21) | 8.77B (1.70). | 9.65A,B (2.03) | .014 | f = .16 |
Income (% >2000 Euro) | 40.34A,B | 45.05A,B | 29.27A | 50.67B | .045 | w = .13 |
Marital status (% married) | 39.67 | 52.59 | 37.21 | 51.33 | .072 | w = .13 |
Living arrangements (% living alone) | 57.02 | 51.72 | 58.14 | 47.33 | .366 | w = .09 |
Distance visiona | .13A (.10) | .69B (.19) | .14A (.11). | .72B (.19) | .000 | f = .86 |
Near visiona | .11A (.13) | .62B (.19) | .11A (.11) | .64B (.19) | .000 | f = .83 |
Subjective visiona,b | 4.13A (.80) | 2.45B (.93) | 4.24A (.79) | 2.39B (.82) | .000 | f = .71 |
Duration of vision impairment in years | 6.57 (10.29) | — | 8.72 (14.66) | — | — | — |
Hearing impairment (dB HL)c | 22.31A (7.75) | 47.95B (10.22) | 45.49B (9.86) | 20.65A (7.86) | .000 | f = .82 |
Subjective hearingb,c | 2.32A (.74) | 3.33B (.95) | 3.02B (.87) | 2.33A (.82) | .000 | f = .47 |
Duration of hearing impairment in years | — | 11.78 (13.27) | 9.00 (12.97) | — | — | — |
Cognitive status/6 cognitive impairment testb | 2.49 (2.76) | 2.42 (2.55) | 2.42 (2.85) | 2.19 (2.35) | .492 | f = .05 |
Notes: VI = visually impaired; HI = hearing impaired; DI = dual sensory impaired; UI = sensory unimpaired.
Different subscripts indicate statistically significant means.
aDistance vision assessment based on a classic clinical screening distance visual acuity chart in the Snellen tradition (Snellen, 1862). Near vision assessment based on Radner and coworkers (1998). Subjective vision capacity based on a 5-point Likert-type scale ranging from 1 (very good) to 5 (very bad).
bHigher values indicate lower functioning.
cBased on data from VI: n = 89, HI: n = 106, UI: n = 99; DI: n = 41. Audiometric assessment of both ears was done in the lab-like environment of a commercial hearing-aid provider, based on a common pure-tone audiometer (manufactured by GN Otometrics). Missing data resulted from the inability or unwillingness of some participants to visit our cooperating acoustic specialist. Subjective hearing capacity based on a 5-point Likert-type scale ranging from 1 (very good) to 5 (very bad).
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) or % . | M (SD) or % . | M (SD) or % . | M (SD) or % . | |||
Age in years | 82.60 (4.63) | 82.69 (5.08) | 83.37 (4.51) | 82.26 (4.50) | .575 | f = .07 |
Gender (% women) | 58.68A | 41.38B | 39.53B | 49.33A,B | .032 | w = .14 |
Education in years | 9.21A,B (1.84) | 9.74A (2.21) | 8.77B (1.70). | 9.65A,B (2.03) | .014 | f = .16 |
Income (% >2000 Euro) | 40.34A,B | 45.05A,B | 29.27A | 50.67B | .045 | w = .13 |
Marital status (% married) | 39.67 | 52.59 | 37.21 | 51.33 | .072 | w = .13 |
Living arrangements (% living alone) | 57.02 | 51.72 | 58.14 | 47.33 | .366 | w = .09 |
Distance visiona | .13A (.10) | .69B (.19) | .14A (.11). | .72B (.19) | .000 | f = .86 |
Near visiona | .11A (.13) | .62B (.19) | .11A (.11) | .64B (.19) | .000 | f = .83 |
Subjective visiona,b | 4.13A (.80) | 2.45B (.93) | 4.24A (.79) | 2.39B (.82) | .000 | f = .71 |
Duration of vision impairment in years | 6.57 (10.29) | — | 8.72 (14.66) | — | — | — |
Hearing impairment (dB HL)c | 22.31A (7.75) | 47.95B (10.22) | 45.49B (9.86) | 20.65A (7.86) | .000 | f = .82 |
Subjective hearingb,c | 2.32A (.74) | 3.33B (.95) | 3.02B (.87) | 2.33A (.82) | .000 | f = .47 |
Duration of hearing impairment in years | — | 11.78 (13.27) | 9.00 (12.97) | — | — | — |
Cognitive status/6 cognitive impairment testb | 2.49 (2.76) | 2.42 (2.55) | 2.42 (2.85) | 2.19 (2.35) | .492 | f = .05 |
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) or % . | M (SD) or % . | M (SD) or % . | M (SD) or % . | |||
Age in years | 82.60 (4.63) | 82.69 (5.08) | 83.37 (4.51) | 82.26 (4.50) | .575 | f = .07 |
Gender (% women) | 58.68A | 41.38B | 39.53B | 49.33A,B | .032 | w = .14 |
Education in years | 9.21A,B (1.84) | 9.74A (2.21) | 8.77B (1.70). | 9.65A,B (2.03) | .014 | f = .16 |
Income (% >2000 Euro) | 40.34A,B | 45.05A,B | 29.27A | 50.67B | .045 | w = .13 |
Marital status (% married) | 39.67 | 52.59 | 37.21 | 51.33 | .072 | w = .13 |
Living arrangements (% living alone) | 57.02 | 51.72 | 58.14 | 47.33 | .366 | w = .09 |
Distance visiona | .13A (.10) | .69B (.19) | .14A (.11). | .72B (.19) | .000 | f = .86 |
Near visiona | .11A (.13) | .62B (.19) | .11A (.11) | .64B (.19) | .000 | f = .83 |
Subjective visiona,b | 4.13A (.80) | 2.45B (.93) | 4.24A (.79) | 2.39B (.82) | .000 | f = .71 |
Duration of vision impairment in years | 6.57 (10.29) | — | 8.72 (14.66) | — | — | — |
Hearing impairment (dB HL)c | 22.31A (7.75) | 47.95B (10.22) | 45.49B (9.86) | 20.65A (7.86) | .000 | f = .82 |
Subjective hearingb,c | 2.32A (.74) | 3.33B (.95) | 3.02B (.87) | 2.33A (.82) | .000 | f = .47 |
Duration of hearing impairment in years | — | 11.78 (13.27) | 9.00 (12.97) | — | — | — |
Cognitive status/6 cognitive impairment testb | 2.49 (2.76) | 2.42 (2.55) | 2.42 (2.85) | 2.19 (2.35) | .492 | f = .05 |
Notes: VI = visually impaired; HI = hearing impaired; DI = dual sensory impaired; UI = sensory unimpaired.
Different subscripts indicate statistically significant means.
aDistance vision assessment based on a classic clinical screening distance visual acuity chart in the Snellen tradition (Snellen, 1862). Near vision assessment based on Radner and coworkers (1998). Subjective vision capacity based on a 5-point Likert-type scale ranging from 1 (very good) to 5 (very bad).
bHigher values indicate lower functioning.
cBased on data from VI: n = 89, HI: n = 106, UI: n = 99; DI: n = 41. Audiometric assessment of both ears was done in the lab-like environment of a commercial hearing-aid provider, based on a common pure-tone audiometer (manufactured by GN Otometrics). Missing data resulted from the inability or unwillingness of some participants to visit our cooperating acoustic specialist. Subjective hearing capacity based on a 5-point Likert-type scale ranging from 1 (very good) to 5 (very bad).
Follow-up survival status of the VI, HI, and VI groups was assessed after 4 years; the DI group was excluded because of its small size. Information sources, such as family members and official city registries, enabled tracking of almost all study participants except for four individuals (3%) in VI, five (4%) in HI, and five (3%) in UI.
Measures
Health.—
The “number of chronic conditions,” such as cardiovascular disease, diabetes, cancer, rheumatism, or thyroid disease, from the Duke University Older Americans Resources and Services questionnaire (Pfeiffer, 1978) was used as a target variable. Subjective health was assessed with a 5-point Likert-type scale ranging from 1 (very good) to 5 (very bad).
Everyday Functioning.—
Ten items that are pursued out-of-home and somewhat challenging for older adults with chronic conditions, such as using public transportation, shopping, and walking for 2 km, were selected from classic scales (e.g., Lawton & Brody, 1969). Items were scored as follows: 2 (without difficulty), 1 (only with difficulty), and 0 (only possible with the help of others); the total score ranged from 0 to 20; Cronbach’s α (total sample, N = 430) α = .88. In addition, 10 more out-of-home items were selected from classic leisure activity scales, for example, engagement in sports, church attendance, and travel in the previous year. Answers of 1 (yes) or 0 (no) resulted in a score ranging from 0 to 10; Cronbach’s α (total sample) α = .65. Subjective autonomy in daily life was assessed by one item from the Neuropsychological Aging Inventory (Oswald, 2005), with ratings ranging from 0 (completely dependent) to 10 (completely independent).
Cognitive Functioning.—
Four internationally established tests of mental functioning were administered by auditory means. First, counting backwards from 100 for 30 s (score = 100 minus achieved number minus errors; Tun & Lachman, 2006) measured processing speed, a fluid ability. Working-memory performance was tested with the digit-span backwards subtest of the German version of the revised Wechsler Adult Intelligence scale (WAIS-R; Tewes, 1991); it consists of two trials of seven sets of digits, which are repeated backwards, with digit length increasing from two to eight numbers. The score indicates performance before failure in both trials and ranged from 0 to 14. For a measure of semantic fluency, which relies on experience and indicates crystallized cognitive functioning, participants named as many animals as possible within 60 s (score = number of correct mentions of animal names). Finally, the similarities subtest of the German WAIS-R was used to assess general abstraction capacity, another indicator of crystallized abilities. Participants were asked to detect similarity in 16 pairs of apparently dissimilar items (e.g., orange and banana = fruit); a maximum score of 32 was possible.
Social Relations and Loneliness.—
Social relations were assessed with Kahn and Antonucci’s (1980) classic circle-diagram method. Participants differentiated the importance of their social network members by means of three circles: Circle 1 includes persons to whom one feels most close emotionally; Circle 2 includes persons who are not as close as those in the first circle, but who are still very important; and Circle 3 includes persons with whom one feels less close, but who are important for other reasons. Scores indicate the number of persons nominated in each circle.
Loneliness was assessed with the 20-item University of California Loneliness scale (German translation: Döring & Bortz, 1993; Russell, Peplau, & Cultrona, 1980). Answers on a 5-point Likert-type scale from 1 (completely wrong) to 5 (completely right) lead to a score ranging from 20 to 100; Cronbach’s α (total sample) α = .76.
Self-Regulation.—
Brandtstädter and Renner’s (1990) scales address tenacious goal pursuit versus flexible goal adjustment, each assessed with 15 items on a 5-point Likert-type scale from 1 (do not agree at all) to 5 (do fully agree). Scores can range from 5 to 75 for each scale and then are divided by the number of items; Cronbach’s α (total sample) α = .65 and α = .67 for tenacity and flexibility, respectively.
Well-being.—
Cognitive well-being was assessed with the 5-item Satisfaction with Life scale (Diener, Emmons, Larsen, & Griffin, 1985; German version by Schumacher, 2003). Responses on a 7-point Likert-type scale from 1 (do not agree at all) to 7 (do fully agree) lead to scores ranging from 5 to 35; Cronbach’s α (total sample) α = .80. The Environmental Mastery scale (Ryff, 1989) assesses perceived mastery of one’s environment and day-to-day life. These nine items are answered using a 5-point Likert-type format from 1 (does not apply) to 5 (does completely apply), with scale scores ranging from 9 to 45; Cronbach’s α (total sample) α = .80.
For affective well-being, positive and negative affects were assessed with the Positive and Negative Affect Schedule (Watson, Clark, & Tellegen, 1988), which consists of 10 positive and 10 negative affect items rated in reference to the month before assessment. Answers on a 5-point Likert-type scale from 1 (not at all) to 5 (very frequently) lead to scores ranging from 10 and 50 for each scale; Cronbach’s α (total sample) α = .70 (positive affect) and α = .78 (negative affect).
Procedure
Although most data collection took place in the research lab, some interviews were done at home (30.0% in the UI group; 26.4% in the VI group; 2.3% in the DI group, and 7.8% in the HI group). All interviews were conducted in a very quiet room by intensively trained research assistants with a background in psychology. They were instructed to speak slowly, clearly, and loudly, and to ask participants prior to testing: “Do you understand me well or should I speak louder?” Hearing-impaired participants were encouraged to wear their hearing aid(s). Testing sessions lasted about 90min and participants were paid 20 Euro.
Data Analytic Design
The Bonferroni correction was applied to avoid Type I error accumulation, as the same study construct was evaluated several times. For example, four group comparisons were made using four indicators of cognitive functioning. Thus, to preserve a significance level of p = .05 for testing the global study expectation, the significance level for each individual comparison has to be p = .05/4 = .0125. We also report effect sizes for all comparisons. Overall differences among group means were tested with subsequent pair-wise comparisons (Scheffé test; p < .05). Because of its potential importance but small size, the DI group was included in bivariate but not multivariate analyses. Multinomial regression analysis was used to examine which variables best predicted group status, and χ2 testing and logistic regression analyses examined differences in mortality status among the groups (except DI) after 4 years.
Structural equation modeling was used to examine the overall model of well-being. Data were analyzed using the full information maximum likelihood (FIML) estimation procedure of Amos 19 (Arbuckle, 1996; 2010). FIML is a state-of-the-art missing data estimation approach for structural equation modeling. Parceling was used to develop the latent constructs of tenacious goal pursuit and flexible goal adjustment, reducing the number of parameters to be estimated in the measurement model (West, Finch, & Curran, 1995). Models were specified separately for each sample, but model fit was evaluated simultaneously (multiple group modeling approach) based on the root mean-squared error of approximation (RMSEA). As suggested by Hu and Bentler (1999), only models with RMSEA values less than .06 were accepted.
Results
Descriptive information for all successful aging indicators and bivariate statistical contrasts can be found in Table 3. Regarding health indicators (α level adjustment: p = .025), all sensory-impaired groups consistently reported poorer subjective health compared with the unimpaired group, but only the VI–UI and DI–UI contrasts were statistically significant. The mean DI score was significantly lower than HI. The number of chronic conditions was not significantly different, although the mean values for all sensory-impaired groups were consistently higher than those of UI. Everyday functioning (p = .017) was consistently lower in VI and DI as compared with HI and UI, which did not differ. The greatest effect sizes were found for everyday functioning, particularly out-of-home daily activities and subjective autonomy, but were still in the range of small effects (Cohen, 1988).
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) . | M (SD) . | M (SD) . | M (SD) . | |||
Health indicators | ||||||
Subjective health | 2.78A,C (.86) | 2.62A,B (.84) | 3.05C (.88) | 2.49B (.79) | .000 | f = .20 |
Number of chronic conditionsa | 3.73 (2.62) | 3.51 (2.33) | 4.00 (2.17) | 3.16 (2.30) | .086 | f = .12 |
Everyday functioning | ||||||
Out-of-home daily activities | 12.59A (4.70) | 15.55B (3.92) | 11.74A (4.68) | 16.63B (4.01) | .000 | f = .41 |
Out-of-home leisure activities | 4.31A (2.07) | 5.20 B (2.10) | 4.37A (2.05) | 5.47B (2.09) | .000 | f = .24 |
Subjective autonomy | 6.80A (2.53) | 8.39B (2.07) | 7.14A (2.39) | 8.39B (1.87) | .000 | f = .32 |
Cognitive resources | ||||||
Counting backwards | 28.82 (7.34) | 28.16 (6.67) | 27.95 (6.74) | 29.82 (6.56) | .211 | f = .11 |
Digit-span backwards | 5.17 (2.09) | 5.41 (1.87) | 5.49 (2.05) | 5.66 (1.74) | .212 | f = .10 |
Naming animals | 19.23 (5.56) | 19.02 (5.85) | 17.53 (4.80) | 19.74 (5.37) | .170 | f = .11 |
Finding similarities | 20.89 (7.44) | 20.70 (7.44) | 19.56 (7.70) | 22.16 (6.72) | .251 | f = .11 |
Social resources and loneliness | ||||||
Number of persons in Circle 1 | 4.71 (3.38) | 4.70 (3.55) | 4.50 (3.42) | 5.17 (3.39) | .415 | f = .07 |
Number of persons in Circle 2 | 2.57 (2.35)A,B | 1.95 (2.31)A | 2.56 (2.40)A,B | 3.32 (3.17)B | .001 | f = .21 |
Number of persons in Circle 3 | 1.36 (1.86)A,B | 1.07 (1.64)A | .55 (.97) A,C | 1.80 (2.0)B | .000 | f = .24 |
Loneliness | 39.50 (12.85) | 39.50 (12.58) | 41.11 (11.61) | 38.49 (11.47) | .688 | f = .06 |
Self-regulation | ||||||
Flexible goal adjustment | 3.84 (.46) | 3.83 (.48) | 3.86 (.44) | 3.85 (.40) | .965 | f = .00 |
Tenacious goal pursuit | 2.89A (.45) | 3.07B (.52) | 2.93A,B (.37) | 3.06B (.48) | .008 | f = .18 |
Well-being indicators | ||||||
Satisfaction with life | 25.42 (5.36) | 25.01 (5.85) | 24.05 (5.89) | 26.32 (4.43) | .050 | f = .13 |
Environmental mastery | 37.18 (5.37) | 37.57 (6.20) | 35.77 (6.35) | 38.04 (5.05) | .074 | f = .11 |
Positive affect | 32.67 (6.24) | 33.54 (6.03) | 32.00 (7.87) | 34.24 (5.53) | .059 | f = .12 |
Negative affect | 19.25 (5.87) | 18.90 (5.90) | 20.60 (5.06) | 19.49 (5.76) | .425 | f = .08 |
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) . | M (SD) . | M (SD) . | M (SD) . | |||
Health indicators | ||||||
Subjective health | 2.78A,C (.86) | 2.62A,B (.84) | 3.05C (.88) | 2.49B (.79) | .000 | f = .20 |
Number of chronic conditionsa | 3.73 (2.62) | 3.51 (2.33) | 4.00 (2.17) | 3.16 (2.30) | .086 | f = .12 |
Everyday functioning | ||||||
Out-of-home daily activities | 12.59A (4.70) | 15.55B (3.92) | 11.74A (4.68) | 16.63B (4.01) | .000 | f = .41 |
Out-of-home leisure activities | 4.31A (2.07) | 5.20 B (2.10) | 4.37A (2.05) | 5.47B (2.09) | .000 | f = .24 |
Subjective autonomy | 6.80A (2.53) | 8.39B (2.07) | 7.14A (2.39) | 8.39B (1.87) | .000 | f = .32 |
Cognitive resources | ||||||
Counting backwards | 28.82 (7.34) | 28.16 (6.67) | 27.95 (6.74) | 29.82 (6.56) | .211 | f = .11 |
Digit-span backwards | 5.17 (2.09) | 5.41 (1.87) | 5.49 (2.05) | 5.66 (1.74) | .212 | f = .10 |
Naming animals | 19.23 (5.56) | 19.02 (5.85) | 17.53 (4.80) | 19.74 (5.37) | .170 | f = .11 |
Finding similarities | 20.89 (7.44) | 20.70 (7.44) | 19.56 (7.70) | 22.16 (6.72) | .251 | f = .11 |
Social resources and loneliness | ||||||
Number of persons in Circle 1 | 4.71 (3.38) | 4.70 (3.55) | 4.50 (3.42) | 5.17 (3.39) | .415 | f = .07 |
Number of persons in Circle 2 | 2.57 (2.35)A,B | 1.95 (2.31)A | 2.56 (2.40)A,B | 3.32 (3.17)B | .001 | f = .21 |
Number of persons in Circle 3 | 1.36 (1.86)A,B | 1.07 (1.64)A | .55 (.97) A,C | 1.80 (2.0)B | .000 | f = .24 |
Loneliness | 39.50 (12.85) | 39.50 (12.58) | 41.11 (11.61) | 38.49 (11.47) | .688 | f = .06 |
Self-regulation | ||||||
Flexible goal adjustment | 3.84 (.46) | 3.83 (.48) | 3.86 (.44) | 3.85 (.40) | .965 | f = .00 |
Tenacious goal pursuit | 2.89A (.45) | 3.07B (.52) | 2.93A,B (.37) | 3.06B (.48) | .008 | f = .18 |
Well-being indicators | ||||||
Satisfaction with life | 25.42 (5.36) | 25.01 (5.85) | 24.05 (5.89) | 26.32 (4.43) | .050 | f = .13 |
Environmental mastery | 37.18 (5.37) | 37.57 (6.20) | 35.77 (6.35) | 38.04 (5.05) | .074 | f = .11 |
Positive affect | 32.67 (6.24) | 33.54 (6.03) | 32.00 (7.87) | 34.24 (5.53) | .059 | f = .12 |
Negative affect | 19.25 (5.87) | 18.90 (5.90) | 20.60 (5.06) | 19.49 (5.76) | .425 | f = .08 |
Notes: VI = visually impaired; HI = hearing impaired; DI = dual sensory impaired; UI = sensory unimpaired.
Different subscripts indicate statistically significant means.
aHigher values indicate lower functioning.
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) . | M (SD) . | M (SD) . | M (SD) . | |||
Health indicators | ||||||
Subjective health | 2.78A,C (.86) | 2.62A,B (.84) | 3.05C (.88) | 2.49B (.79) | .000 | f = .20 |
Number of chronic conditionsa | 3.73 (2.62) | 3.51 (2.33) | 4.00 (2.17) | 3.16 (2.30) | .086 | f = .12 |
Everyday functioning | ||||||
Out-of-home daily activities | 12.59A (4.70) | 15.55B (3.92) | 11.74A (4.68) | 16.63B (4.01) | .000 | f = .41 |
Out-of-home leisure activities | 4.31A (2.07) | 5.20 B (2.10) | 4.37A (2.05) | 5.47B (2.09) | .000 | f = .24 |
Subjective autonomy | 6.80A (2.53) | 8.39B (2.07) | 7.14A (2.39) | 8.39B (1.87) | .000 | f = .32 |
Cognitive resources | ||||||
Counting backwards | 28.82 (7.34) | 28.16 (6.67) | 27.95 (6.74) | 29.82 (6.56) | .211 | f = .11 |
Digit-span backwards | 5.17 (2.09) | 5.41 (1.87) | 5.49 (2.05) | 5.66 (1.74) | .212 | f = .10 |
Naming animals | 19.23 (5.56) | 19.02 (5.85) | 17.53 (4.80) | 19.74 (5.37) | .170 | f = .11 |
Finding similarities | 20.89 (7.44) | 20.70 (7.44) | 19.56 (7.70) | 22.16 (6.72) | .251 | f = .11 |
Social resources and loneliness | ||||||
Number of persons in Circle 1 | 4.71 (3.38) | 4.70 (3.55) | 4.50 (3.42) | 5.17 (3.39) | .415 | f = .07 |
Number of persons in Circle 2 | 2.57 (2.35)A,B | 1.95 (2.31)A | 2.56 (2.40)A,B | 3.32 (3.17)B | .001 | f = .21 |
Number of persons in Circle 3 | 1.36 (1.86)A,B | 1.07 (1.64)A | .55 (.97) A,C | 1.80 (2.0)B | .000 | f = .24 |
Loneliness | 39.50 (12.85) | 39.50 (12.58) | 41.11 (11.61) | 38.49 (11.47) | .688 | f = .06 |
Self-regulation | ||||||
Flexible goal adjustment | 3.84 (.46) | 3.83 (.48) | 3.86 (.44) | 3.85 (.40) | .965 | f = .00 |
Tenacious goal pursuit | 2.89A (.45) | 3.07B (.52) | 2.93A,B (.37) | 3.06B (.48) | .008 | f = .18 |
Well-being indicators | ||||||
Satisfaction with life | 25.42 (5.36) | 25.01 (5.85) | 24.05 (5.89) | 26.32 (4.43) | .050 | f = .13 |
Environmental mastery | 37.18 (5.37) | 37.57 (6.20) | 35.77 (6.35) | 38.04 (5.05) | .074 | f = .11 |
Positive affect | 32.67 (6.24) | 33.54 (6.03) | 32.00 (7.87) | 34.24 (5.53) | .059 | f = .12 |
Negative affect | 19.25 (5.87) | 18.90 (5.90) | 20.60 (5.06) | 19.49 (5.76) | .425 | f = .08 |
Variable . | VI (n = 121) . | HI (n = 116) . | DI (n = 43) . | UI (n = 150) . | p value . | Effect size . |
---|---|---|---|---|---|---|
M (SD) . | M (SD) . | M (SD) . | M (SD) . | |||
Health indicators | ||||||
Subjective health | 2.78A,C (.86) | 2.62A,B (.84) | 3.05C (.88) | 2.49B (.79) | .000 | f = .20 |
Number of chronic conditionsa | 3.73 (2.62) | 3.51 (2.33) | 4.00 (2.17) | 3.16 (2.30) | .086 | f = .12 |
Everyday functioning | ||||||
Out-of-home daily activities | 12.59A (4.70) | 15.55B (3.92) | 11.74A (4.68) | 16.63B (4.01) | .000 | f = .41 |
Out-of-home leisure activities | 4.31A (2.07) | 5.20 B (2.10) | 4.37A (2.05) | 5.47B (2.09) | .000 | f = .24 |
Subjective autonomy | 6.80A (2.53) | 8.39B (2.07) | 7.14A (2.39) | 8.39B (1.87) | .000 | f = .32 |
Cognitive resources | ||||||
Counting backwards | 28.82 (7.34) | 28.16 (6.67) | 27.95 (6.74) | 29.82 (6.56) | .211 | f = .11 |
Digit-span backwards | 5.17 (2.09) | 5.41 (1.87) | 5.49 (2.05) | 5.66 (1.74) | .212 | f = .10 |
Naming animals | 19.23 (5.56) | 19.02 (5.85) | 17.53 (4.80) | 19.74 (5.37) | .170 | f = .11 |
Finding similarities | 20.89 (7.44) | 20.70 (7.44) | 19.56 (7.70) | 22.16 (6.72) | .251 | f = .11 |
Social resources and loneliness | ||||||
Number of persons in Circle 1 | 4.71 (3.38) | 4.70 (3.55) | 4.50 (3.42) | 5.17 (3.39) | .415 | f = .07 |
Number of persons in Circle 2 | 2.57 (2.35)A,B | 1.95 (2.31)A | 2.56 (2.40)A,B | 3.32 (3.17)B | .001 | f = .21 |
Number of persons in Circle 3 | 1.36 (1.86)A,B | 1.07 (1.64)A | .55 (.97) A,C | 1.80 (2.0)B | .000 | f = .24 |
Loneliness | 39.50 (12.85) | 39.50 (12.58) | 41.11 (11.61) | 38.49 (11.47) | .688 | f = .06 |
Self-regulation | ||||||
Flexible goal adjustment | 3.84 (.46) | 3.83 (.48) | 3.86 (.44) | 3.85 (.40) | .965 | f = .00 |
Tenacious goal pursuit | 2.89A (.45) | 3.07B (.52) | 2.93A,B (.37) | 3.06B (.48) | .008 | f = .18 |
Well-being indicators | ||||||
Satisfaction with life | 25.42 (5.36) | 25.01 (5.85) | 24.05 (5.89) | 26.32 (4.43) | .050 | f = .13 |
Environmental mastery | 37.18 (5.37) | 37.57 (6.20) | 35.77 (6.35) | 38.04 (5.05) | .074 | f = .11 |
Positive affect | 32.67 (6.24) | 33.54 (6.03) | 32.00 (7.87) | 34.24 (5.53) | .059 | f = .12 |
Negative affect | 19.25 (5.87) | 18.90 (5.90) | 20.60 (5.06) | 19.49 (5.76) | .425 | f = .08 |
Notes: VI = visually impaired; HI = hearing impaired; DI = dual sensory impaired; UI = sensory unimpaired.
Different subscripts indicate statistically significant means.
aHigher values indicate lower functioning.
Cognitive resources (p = .0125) were consistently similar among groups. With respect to social resources and loneliness (p = .0125), there were no differences among groups in the number of persons nominated for Circle 1. For Circles 2 and 3, however, the mean number in all sensory-impaired groups was lower compared with UI; interestingly, DI had the fewest persons in Circle 3 among all groups. No differences in loneliness were found. Self-regulative strategies (p = .025) revealed no group differences in terms of flexible goal adjustment. Lower tenacious goal pursuit was observed in both VI and DI compared with the remaining groups, whereas only the VI–HI and VI–UI contrasts were significant.
Finally, the four well-being indicators (p = .0125) did not differ among the three impaired groups, although there was a tendency toward lower values in the DI group (except for negative affect).
To reduce the number of predictors in relation to sample size, certain indicators of successful aging were selected to represent each domain in a multinomial regression analysis (Table 4). The VI–UI contrast resulted in the most pronounced difference in everyday functioning (p < .001). Explained variance was in an acceptable range (Cox & Snell, 1989), indicating a possible meaningful group contrast. In contrast, although everyday functioning was significantly different (p < .05) in the HI–VI comparison, the overall model was not acceptable, with low explained variance. Thus, multinomial analysis confirmed substantial differences between VI and UI, but only minor differences between HI and UI.
Predictors . | Odds ratio . | 95% Confidence interval . |
---|---|---|
VI versus UI | ||
Subjective health | 1.089 | .720–1.649 |
Out-of-home daily activities | .821*** | .751–.896 |
Subjective autonomy | .864 | .731–1.021 |
Counting backwards | 1.014 | .968–1.063 |
Naming animals | 1.009 | .951–1.071 |
Persons in Circle 1 | .972 | .906–1.041 |
Loneliness | .991 | .960–1.023 |
Satisfaction with life | .978 | .916–1.045 |
Flexible goal adjustment | .978 | .436–2.198 |
Tenacious goal pursuit | .653 | .322–1.325 |
Positive affect | 1.010 | .947–1.077 |
R² | ||
Cox & Snell | .257 | |
Nagelkerke | .338 | |
HI versus UI | ||
Subjective health | 1.033 | .697–1.531 |
Out-of-home daily activities | .900* | .822–.984 |
Subjective autonomy | 1.169 | .975–1.402 |
Counting backwards | .978 | .933–1.025 |
Naming animals | .986 | .935–1.040 |
Persons in Circle 1 | .969 | .901–1.042 |
Loneliness | 1.011 | .980–1.042 |
Satisfaction with life | .963 | .905–1.024 |
Flexible goal adjustment | 1.408 | .705–2.815 |
Tenacious goal pursuit | 1.478 | .805–2.713 |
Positive affect | .999 | .938–1.064 |
R² | ||
Cox & Snell | .072 | |
Nagelkerke | .097 |
Predictors . | Odds ratio . | 95% Confidence interval . |
---|---|---|
VI versus UI | ||
Subjective health | 1.089 | .720–1.649 |
Out-of-home daily activities | .821*** | .751–.896 |
Subjective autonomy | .864 | .731–1.021 |
Counting backwards | 1.014 | .968–1.063 |
Naming animals | 1.009 | .951–1.071 |
Persons in Circle 1 | .972 | .906–1.041 |
Loneliness | .991 | .960–1.023 |
Satisfaction with life | .978 | .916–1.045 |
Flexible goal adjustment | .978 | .436–2.198 |
Tenacious goal pursuit | .653 | .322–1.325 |
Positive affect | 1.010 | .947–1.077 |
R² | ||
Cox & Snell | .257 | |
Nagelkerke | .338 | |
HI versus UI | ||
Subjective health | 1.033 | .697–1.531 |
Out-of-home daily activities | .900* | .822–.984 |
Subjective autonomy | 1.169 | .975–1.402 |
Counting backwards | .978 | .933–1.025 |
Naming animals | .986 | .935–1.040 |
Persons in Circle 1 | .969 | .901–1.042 |
Loneliness | 1.011 | .980–1.042 |
Satisfaction with life | .963 | .905–1.024 |
Flexible goal adjustment | 1.408 | .705–2.815 |
Tenacious goal pursuit | 1.478 | .805–2.713 |
Positive affect | .999 | .938–1.064 |
R² | ||
Cox & Snell | .072 | |
Nagelkerke | .097 |
Notes: VI = visually impaired; HI = hearing impaired; UI = sensory unimpaired.
*p < .05. ** p < .01. ***p < .001
Predictors . | Odds ratio . | 95% Confidence interval . |
---|---|---|
VI versus UI | ||
Subjective health | 1.089 | .720–1.649 |
Out-of-home daily activities | .821*** | .751–.896 |
Subjective autonomy | .864 | .731–1.021 |
Counting backwards | 1.014 | .968–1.063 |
Naming animals | 1.009 | .951–1.071 |
Persons in Circle 1 | .972 | .906–1.041 |
Loneliness | .991 | .960–1.023 |
Satisfaction with life | .978 | .916–1.045 |
Flexible goal adjustment | .978 | .436–2.198 |
Tenacious goal pursuit | .653 | .322–1.325 |
Positive affect | 1.010 | .947–1.077 |
R² | ||
Cox & Snell | .257 | |
Nagelkerke | .338 | |
HI versus UI | ||
Subjective health | 1.033 | .697–1.531 |
Out-of-home daily activities | .900* | .822–.984 |
Subjective autonomy | 1.169 | .975–1.402 |
Counting backwards | .978 | .933–1.025 |
Naming animals | .986 | .935–1.040 |
Persons in Circle 1 | .969 | .901–1.042 |
Loneliness | 1.011 | .980–1.042 |
Satisfaction with life | .963 | .905–1.024 |
Flexible goal adjustment | 1.408 | .705–2.815 |
Tenacious goal pursuit | 1.478 | .805–2.713 |
Positive affect | .999 | .938–1.064 |
R² | ||
Cox & Snell | .072 | |
Nagelkerke | .097 |
Predictors . | Odds ratio . | 95% Confidence interval . |
---|---|---|
VI versus UI | ||
Subjective health | 1.089 | .720–1.649 |
Out-of-home daily activities | .821*** | .751–.896 |
Subjective autonomy | .864 | .731–1.021 |
Counting backwards | 1.014 | .968–1.063 |
Naming animals | 1.009 | .951–1.071 |
Persons in Circle 1 | .972 | .906–1.041 |
Loneliness | .991 | .960–1.023 |
Satisfaction with life | .978 | .916–1.045 |
Flexible goal adjustment | .978 | .436–2.198 |
Tenacious goal pursuit | .653 | .322–1.325 |
Positive affect | 1.010 | .947–1.077 |
R² | ||
Cox & Snell | .257 | |
Nagelkerke | .338 | |
HI versus UI | ||
Subjective health | 1.033 | .697–1.531 |
Out-of-home daily activities | .900* | .822–.984 |
Subjective autonomy | 1.169 | .975–1.402 |
Counting backwards | .978 | .933–1.025 |
Naming animals | .986 | .935–1.040 |
Persons in Circle 1 | .969 | .901–1.042 |
Loneliness | 1.011 | .980–1.042 |
Satisfaction with life | .963 | .905–1.024 |
Flexible goal adjustment | 1.408 | .705–2.815 |
Tenacious goal pursuit | 1.478 | .805–2.713 |
Positive affect | .999 | .938–1.064 |
R² | ||
Cox & Snell | .072 | |
Nagelkerke | .097 |
Notes: VI = visually impaired; HI = hearing impaired; UI = sensory unimpaired.
*p < .05. ** p < .01. ***p < .001
All available indicators were considered for structural equation model testing so that relevant latent variables were represented as strongly as possible. First, an unconstrained model was specified for the VI, HI, and UI groups (Model 1), with five predictor variables (one observed and four latent variables)—number of persons in Circle 1, health, everyday competence, tenacious goal pursuit, and flexible goal adjustment—which are assumed to predict the latent outcome variable overall well-being. This model, which assumes a correlation between health and everyday competence, had a good model fit (RMSEA = .046). Subsequently, in Model 2, error variances and factor loadings were constrained to equality across groups to test for invariance of measurement models. This model fit the data well (RMSEA = .043) and did not differ significantly from Model 1 (Δχ2 = 56.71; df = 46; p < .134). Thus, the measurement models are considered equal across groups. In the HI and UI groups, the path between everyday competence and overall well-being was not statistically significant, nor was the path between number of persons in Circle 1 and overall well-being among the HI. Thus, in Model 3, these paths were constrained to zero for the respective groups, added to the constraints from Model 2. This model had a good fit (RMSEA = .043) and also did not differ significantly from Model 1 (Δχ2 = 60.51; df = 49; p < .125). Figure 1 depicts the accepted Model 3. Within groups, the model explains 60%–70% of the variance in overall well-being, and most of the resources show similarly robust relationships with the outcome variable across groups. However, neither the number of persons in Circle 1 nor everyday functioning contributed significantly to overall well-being among the HI; the latter was a significant predictor only in the VI.
Finally, regarding mortality 4 years after baseline, 35 (28.9%) of the 121 persons in the VI group died; 35 (30.2%) of the HI group died; and 22 (20%) of the 150 persons in the UI group died. Although a χ2 analysis showed significant mortality differences in VI versus UI (p < .01) and HI versus UI (p < .01), these differences were not significant in logistic regression analysis, while controlling for age, gender, education, subjective health, out-of-home daily activities, and marital status (Table 5). Aside from age and gender, out-of-home daily activities were an important predictor of mortality, but only in the VI group.
Variable . | UI–VI group (n = 271) . | UI–HI group (n = 266) . |
---|---|---|
Odds ratio (95% confidence interval) . | Odds ratio (95% confidence interval) . | |
Age in years | .887** (.820–.959) | .865*** (.798–.937) |
Gender (men = 0; women = 1) | 3.916** (1.808–8.482) | 3.381** (1.444–7.920) |
Subjective health | .789 (.514–1.210) | .718 (.467–1.147) |
Education in years | 1.146 (.960–1.367) | 1.139 (.964–1.346) |
Living arrangements (% living alone) | .883 (.414–1.880) | 1.720 (.756–3.912) |
Out-of-home daily activities | 1.098* (1.012–1.191) | 1.080 (.984–1.186) |
Group membership (UI = 0; VI/HI = 1) | 1.348 (.934–1.946) | 1.937 (.983–1.986) |
Variable . | UI–VI group (n = 271) . | UI–HI group (n = 266) . |
---|---|---|
Odds ratio (95% confidence interval) . | Odds ratio (95% confidence interval) . | |
Age in years | .887** (.820–.959) | .865*** (.798–.937) |
Gender (men = 0; women = 1) | 3.916** (1.808–8.482) | 3.381** (1.444–7.920) |
Subjective health | .789 (.514–1.210) | .718 (.467–1.147) |
Education in years | 1.146 (.960–1.367) | 1.139 (.964–1.346) |
Living arrangements (% living alone) | .883 (.414–1.880) | 1.720 (.756–3.912) |
Out-of-home daily activities | 1.098* (1.012–1.191) | 1.080 (.984–1.186) |
Group membership (UI = 0; VI/HI = 1) | 1.348 (.934–1.946) | 1.937 (.983–1.986) |
Notes: VI = visually impaired; HI = hearing impaired; UI = sensory unimpaired.
*p < .05. **p < .01. ***p < .001.
Variable . | UI–VI group (n = 271) . | UI–HI group (n = 266) . |
---|---|---|
Odds ratio (95% confidence interval) . | Odds ratio (95% confidence interval) . | |
Age in years | .887** (.820–.959) | .865*** (.798–.937) |
Gender (men = 0; women = 1) | 3.916** (1.808–8.482) | 3.381** (1.444–7.920) |
Subjective health | .789 (.514–1.210) | .718 (.467–1.147) |
Education in years | 1.146 (.960–1.367) | 1.139 (.964–1.346) |
Living arrangements (% living alone) | .883 (.414–1.880) | 1.720 (.756–3.912) |
Out-of-home daily activities | 1.098* (1.012–1.191) | 1.080 (.984–1.186) |
Group membership (UI = 0; VI/HI = 1) | 1.348 (.934–1.946) | 1.937 (.983–1.986) |
Variable . | UI–VI group (n = 271) . | UI–HI group (n = 266) . |
---|---|---|
Odds ratio (95% confidence interval) . | Odds ratio (95% confidence interval) . | |
Age in years | .887** (.820–.959) | .865*** (.798–.937) |
Gender (men = 0; women = 1) | 3.916** (1.808–8.482) | 3.381** (1.444–7.920) |
Subjective health | .789 (.514–1.210) | .718 (.467–1.147) |
Education in years | 1.146 (.960–1.367) | 1.139 (.964–1.346) |
Living arrangements (% living alone) | .883 (.414–1.880) | 1.720 (.756–3.912) |
Out-of-home daily activities | 1.098* (1.012–1.191) | 1.080 (.984–1.186) |
Group membership (UI = 0; VI/HI = 1) | 1.348 (.934–1.946) | 1.937 (.983–1.986) |
Notes: VI = visually impaired; HI = hearing impaired; UI = sensory unimpaired.
*p < .05. **p < .01. ***p < .001.
Discussion
In the multidimensional consideration of successful aging indicators, most comparisons did not find significant differences between sensory-impaired groups and UI groups, despite somewhat lower health indicator scores in the sensory-impaired groups. However, the everyday functioning scores of VI and DI individuals were significantly lower compared with both HI and UI, confirming earlier research (Burmedi et al., 2002a; Heyl & Wahl, 2001, 2012; Horowitz et al., 2005). Similarly, although tenacious goal pursuit scores were lower in VI and DI compared with the remaining groups, only the VI–HI and VI–UI contrasts reached statistical significance.
Consistent with socioemotional selectivity theory (Carstensen, 2006), all four groups nominated a comparable mean number of socially most meaningful persons (Circle 1), despite sensory impairment. In addition, all sensory impairment groups nominated fewer persons in Circles 2 and 3 compared with the UI group. This could be interpreted as a lack of motivation and/or capability among VI, HI, and DI to engage in maintaining more marginal social relationships compared with UI. Difficulty in maintaining relationships that are less close could be due a variety of challenges: using the phone for HI persons, writing cards or letters for the VI, and simple face-to-face conversations for both; communication problems for those with hearing impairment and mobility problems for those with visual impairment may make it difficult to meet in person. Surprisingly, significant differences among the groups were not found in cognitive resources. This could be due to the possibility that our samples of VI and/or HI older adults were positively selected. Because we only studied adults older than 75 years, more vulnerable individuals may have already passed away or relocated to long-term care, thereby equalizing the overall quality of life of our subgroups. Perhaps, we also exerted more effort than have others previously to accommodate very old, sensory-impaired adults and test them fairly.
Consistent with the well-being paradox (Kunzmann et al., 2000; Schilling & Wahl, 2006), our groups showed no pronounced differences in well-being-related indicators, demonstrating that older adults, including those with chronic conditions, have considerable psychological resilience for countering functional loss. Nevertheless, the DI group achieved the lowest scores regarding satisfaction with life, environmental mastery, and positive affect. Together with the health and everyday competence findings, it appears that this group carries the highest psychosocial burden of all the groups (Brennan et al., 2005; Crews & Campbell, 2004). Multigroup testing of the covariance structure among major predictors of well-being and an overall indication of well-being in VI, HI, and UI at the latent space level, added significantly to our findings, given that both similarities and differences emerged. Health, tenacious goal pursuit, and flexible goal adjustment were significant predictors of overall well-being, similarly among all groups (with flexible goal adjustment being more important than tenacious goal pursuit). Yet, everyday competence was only important in the VI group, and the number of most emotionally important persons was unimportant in the HI group, pointing to the critical importance of maintaining everyday functional competence in the VI. Perhaps, the number of most emotionally meaningful persons is less important for maintaining well-being in the HI group; this may be due to communication problems and issues related to stigma, even in the context of family and friends (Wahl & Tesch-Römer, 2001; Wallhagen, 2010). Obviously, a causal interpretation is not possible with our data, and therefore, longitudinal data are needed in future research.
Regarding mortality findings, VI and HI showed a higher mortality rate after 4 years compared with the UI at the bivariate level, but not after potential confounders were added in multivariate analyses. These findings exemplify persistent inconsistencies in the sensory impairment and old age literature. Some large studies have found mortality differences, where those with sensory impairment are at a disadvantage (e.g., Anstey, Luscz, Giles et al., 2001; Karpa et al., 2009; Lee et al., 2007), although others have found no differences (Ostbye et al., 1999) or only partial confirmation (e.g., Knudtson, Klein, & Klein, 2006). It seems that both calendar age and the various confounders considered are important. For example, the increased mortality risk in the Karpa and coworkers (2009) study was only observed for those younger than 75 years (and thus not in our study population). Also, everyday functioning was the only significant predictor variable besides age and gender in our analysis, but only for the VI group. Other studies have reported pronounced disadvantages for those with vision and hearing impairment in terms of successful aging (also known as quality of life) indicators, such as well-being (Heyl & Wahl, 2001; Wahl et al., 1999), loneliness, and social role involvement (Crews & Campbell, 2004; Nachtegaal et al., 2009). In addition, our inclusion criteria focused on moderate vision and hearing impairments and thus may have missed significant numbers of those with very severe impairment.
In policy and practical terms, we clearly confirmed that loss of everyday functioning is a major characteristic of age-related vision impairment (but not hearing impairment), relating to survival. Therefore, investing in maintaining everyday functioning in vision impairment-specific rehabilitation programs seems absolutely crucial. Hearing impairment in advanced old age, at first glance not “as bad as” vision impairment in terms of everyday functioning, appears to be more of a vulnerability in social communication (Wallhagen, 2010) and may be negatively influenced by negative-age stereotyping in the longer term (Levy, Slade, & Gill, 2006). Therefore, social communication training and reductions in negative-age stereotyping in the social environment of those with hearing impairment are critical interventions. In the dually impaired, both of these types of interventions could probably enhance successful aging trajectories.
Several limitations of this study must be noted. First, although sample sizes were considerable in terms of the groups of individuals with severe impairment, samples sizes overall were rather small, particularly the DI group, thus limiting the robustness and generalizability of our findings. Second, our health indicators were limited, due to their subjective quality. Third, we did not assess depressed mood in this study. Depression may have been an important health indicator, particularly for the VI group, which has demonstrated a higher rate of depression than the general population (Horowitz et al., 2005). Fourth, we collected cross-sectional data and thus were not able to support any causal conclusions. Fifth, given response rates, community residence, and selective mortality, it is likely that our sample is biased toward the positive end; this may also explain our finding of no major group differences in cognitive functioning and limit the external validity of the study.
In conclusion, a multidimensional approach to the understanding of sensory impairment and psychosocial adaptation in old age reveals a complex picture of loss and maintenance.
Funding
German Research Foundation (WA 809/7-1 awarded to H-W.W.).
Acknowledgments
We are very thankful for support from the KIND Hörgeräte Company (Heidelberg, Germany) for audiometric assessment. We would also like to thank our project staff, particularly Nadine Langer, Christina Hunger, Katharina Hager, Anna Sawall, and our study participants, who invested their time and energy in our research.
References
Author notes
Decision Editor: Rachel Pruchno, PhD