Predictors of Life Satisfaction in Frail Elderly doc

This study examined the relationship between life satisfac-tion and physical status, emosatisfac-tional health, social support and locus of control in the frail elderly.. Multiple regre

Predictors of Life Satisfaction

in Frail Elderly

Soleman H Abu-Bader, PhD Anissa Rogers, PhD Amanda S Barusch, PhD

ABSTRACT This study examined the relationship between life

satisfac-tion and physical status, emosatisfac-tional health, social support and locus of control in the frail elderly A random sample of 99 low-income, frail el-derly living in the community was interviewed Almost 40% of partici-pants reported high levels of life satisfaction Multiple regression analysis identified four significant predictors of life satisfaction: Perceived physical health, social support, emotional balance, and locus of control Physical health emerged as the most significant predictor of life satisfaction ac-counting for 14% of the variance Social support, emotional balance and lo-cus of control each accounted for an additional 6% of the variance in life

satisfaction All four predictors explained 32% (R = 57) of the total

vari-ance in life satisfaction Implications for practice and recommendations are discussed.[Article copies available for a fee from The Haworth Document Deliv-ery Service: 1-800-HAWORTH E-mail address: <getinfo@haworthpressinc.com> Website: <http://www.HaworthPress.com> © 2002 by The Haworth Press, Inc All rights reserved.]

Soleman H Abu-Bader is Assistant Professor, School of Social Work, Howard Uni-versity.

Anissa Rogers is Assistant Professor, Department of Social and Behavioral Sciences, University of Portland.

Amanda Barusch is Professor, Graduate School of Social Work, University of Utah Address correspondence to: Soleman H Abu-Bader, PhD, Howard University, 601 Howard Place, NW, Washington, DC 20059 (E-mail: sabu-bader@howard.edu) This study was funded by the Goodwill Family Foundation.

Paper presented at the 4th Annual Conference of the Society for Social Work & Re-search, Charleston, South Carolina, January 29-31, 2000.

Journal of Gerontological Social Work, Vol 38(3) 2002

http://www.haworthpressinc.com/store/product.asp?sku=J083

Ó 2002 by The Haworth Press, Inc All rights reserved 3

KEYWORDS Frail, elderly, life satisfaction, physical health, social

sup-port, cognitive status, emotional balance, locus of control, life events, ac-tivities of daily livings

The life satisfaction of the elderly has been widely researched and dis-cussed The construct is particularly important for professionals, such as social workers, whose work aims to enhance the quality of life of the el-derly Much research in this area has analyzed how life satisfaction is in-fluenced by factors such as social support, financial status, physical health, and locus of control Most of this work has been conducted with healthy, community-dwelling elders No studies to date have considered the life satisfaction of elders who have significant physical limitations (possibly because of a tendency to assume that physical problems will re-sult in low life satisfaction) The purpose of the current study is to expand

on prior studies by analyzing factors that may influence life satisfaction among frail elders

FACTORS INFLUENCING LIFE SATISFACTION

AMONG THE ELDERLY

Factors that influence life satisfaction include environmental character-istics, such as the availability of social support, and personal traits, such as self-esteem, physical health, financial resources, a sense of connectedness, and locus of control

Social Support and Life Satisfaction

Several studies have examined the relationship between social sup-port and life satisfaction among the elderly Most of this literature has indicated a positive relationship between social support and life satis-faction One study conducted by Aquino, Russell, Cutrona, and Altmaier (1996), found that social support was significantly related to life satisfaction Aquino et al surveyed 301 community-dwelling elders aged

65 years old and over to determine how demographic variables such as fi-nancial status, educational level, and work patterns affect life satisfaction Results from face-to-face interviews indicated that elders who were work-ing or volunteerwork-ing showed higher life satisfaction than those who were not working or volunteering Further, these authors found that participants who engaged in volunteer work had more social supports than those who

were not engaged in volunteer work, which in turn led to higher levels of life satisfaction The findings also indicated that participants who reported low education and socioeconomic levels and who had poor physical health indicated that they had few social supports and low life satisfaction Consequently, participants who were not functioning well enough to work

or volunteer had fewer opportunities to build social networks, which af-forded fewer opportunities to engage in satisfying relationships outside of the workplace than participants who were working or volunteering Though many of the measures used in the aforementioned study were standardized, particularly those measuring social support and life satisfac-tion, it is unclear whether these instruments are appropriate for use with older adults

In another study conducted by Newsome and Schulz (1996), 5,201 people aged 65+ were randomly selected from Medicare lists Partici-pants were surveyed to gather information regarding their social net-works, level of functioning, perceived social supports, and life satisfaction Results indicated that participants who reported decreased physical functioning also perceived their social supports as poor Further, participants who perceived their social supports as poor reported low life satisfaction Thus, participants who reported physical difficulties also per-ceived their social supports to be poor, which may have affected their level of life satisfaction

At this point the theoretical connection between physical functioning and social support remains unclear While some argue that social support enhances physical health or buffers an individual from the deleterious ef-fects of stress, these connections have not been empirically demonstrated

Personal Traits

Other studies have indicated that factors such as self-esteem, per-ceived physical health, and locus of control are associated with life sat-isfaction (Girzadas, Counte, Glandon, & Tancredi, 1993; Rogers, 1999) Still other literature posits that financial security and a sense of closeness and connectedness with others predict life satisfaction (Fisher, 1995; Girzadas et al., 1993; Gray, Ventis, & Hayslip, 1992; Kahana et al., 1995; Levitt, Antonucci, Clark, Rotton, & Finley, 1986; McGhee, 1984; Revicki & Mitchell, 1986; Wing-Leung Lai & McDon-ald, 1995) Indeed, Kahana et al (1995) found that short-term problems such as those caused by financial difficulties and changes in relationships through retirement or death may have a significant impact on life satisfac-tion

Locus of control has been another widely studied construct in relation

to life satisfaction among the elderly Most research has focused on the relationship between internal, external, and chance locus of control and life satisfaction, and conclusions as to the nature of this relationship have been mixed In an exploratory study conducted by Girzadas et al (1993), 258 community-dwelling individuals aged 55+ were selected from a larger study that examined the relationship between health status, locus of control, and life satisfaction The larger study recruited partici-pants from the rolls of Health Maintenance Organizations and private physicians Results from face-to-face interviews with participants indi-cated that functional health status was positively associated with life satis-faction Further, participants who scored high on chance locus of control also scored low on life satisfaction Specifically, participants who re-ported poor physical health and who demonstrated a tendency toward be-lieving their health outcomes were based on chance also showed relatively low life satisfaction

Results from other studies suggest that individuals with a tendency to-ward internal locus of control, particularly with regard to physical health, show higher levels of life satisfaction than those who show a tendency to-ward external or chance locus of control (e.g., Haber, 1994; Searle, M S., Mahon, M J., & Iso-Ahola, S E., 1995; Wing-Leung Lai & McDonald, 1995) It follows that older adults who are not internally focused may show a tendency toward low life satisfaction For instance, Park and Vanderberg (1994) found from a sample of 154 individuals aged 58+ that those who demonstrated low levels of personal autonomy and high levels

of dependency tended to be more negatively affected by poor health and showed a need for social support in the form of a confidant than more au-tonomous and independent individuals Moreover, Vallerand, O’Connor, and Blais (1989) found that older adults living in nursing homes that did not allow for personal autonomy or self-determination showed lower life satisfaction than older adults living in nursing homes that allowed for more personal independence or those living independently in the commu-nity

Conversely, some studies have indicated that older adults who demon-strate a tendency towards external locus of control have higher life satis-faction than those with an internal or chance locus of control (e.g., Haber, 1994; Rogers, 1999) It may be that older individuals who are externally focused and who have developed trust in their health care provider actu-ally demonstrate higher levels of life satisfaction than those who rely on themselves or even chance for health care decisions Specifically, the lat-ter group may experience more guilt or feelings of hopelessness when

faced with health problems, poor treatment, or poor decision making with regard to health care (Haber, 1994)

In summary, studies of life satisfaction among elderly have identified several important constructs that may influence this measure These structs include the following: social support, physical health, locus of con-trol (internal, external, and chance), financial status, and life events Using multiple regression analysis, the current study incorporated these and other measures to develop a model that best describes factors associated with life satisfaction among frail elderly

METHODOLOGY Sample

Recruiting Methods.Participants for this study were randomly selected from clients of The Alternatives Program (TAP) in Salt Lake County, Utah TAP provides services such as adult day care and homemakers to low-income adults whose physicians certify that they are at risk of nursing home placement Of the 182 TAP clients contacted, 102 (56%) were inter-viewed, of whom 99 were 60 years old and over and thus included in the final analysis Of the 99 subjects, 87 (88%) were women Program staff provided access to demographic information on non-participants, and independent t-tests and chi-square analyses were conducted to identify differences in age, race, gender, income, and living status (whether or not the respondent lived alone) between participants and

non-partici-pants These analyses yielded significant differences in age (t = 2.42, p =

.017) Older TAP clients refused to participate in the study more fre-quently Analyses revealed no significant differences on other variables The most common reason for non-participation, given by 43% of those who declined, related to poor health and cognitive difficulties Others (26%) indicated they just did not want to be interviewed, while the re-mainder cited other reasons

Participant Characteristics.Participants in this study were predomi-nantly white (94%) Their mean age was 78 years (SD = 8.5) with a range from 60 to 101 years The median monthly income was $575 with a range from $257 to $1,584 per month The vast majority (81%) were either wid-owed (53%) or divorced (28%) The majority (70%) lived alone The modal level of education was high school (34%)

Respondents were frail, needing assistance with an average of 2.6 Ac-tivities of Daily Living (ADL’s) and 5.3 Instrumental AcAc-tivities of Daily

Living (IADL’s) Respondents reported their health to be “fair” (38%) to

“good” (24%)

Over one third (36%) had a diagnosis relating to the musculoskeletal system, primarily a form of arthritis The next most common diagnostic category was neurological, with 24% of the sample experiencing an ill-ness of this kind This was followed by cardiovascular illill-ness, experi-enced by 19% of the sample The mean duration of illness was 14 years, with a range from 1 to 70 Half (53%) of the sample reported a steady de-cline associated with their illness Others reported that their conditions were stable (16%) or had variable “up and down” trajectories (14%) A few respondents (10%) reported their condition had recently improved

Data Collection

Interviews. Trained undergraduate and graduate MSW students con-ducted in-depth interviews in respondents’ homes Interviews lasted one

to three hours When participants became fatigued, interviewers com-pleted the interview in a second home visit While most of the protocol was highly structured, interviews closed with a “guided conversation” de-signed to explore respondents’ understanding of the impact of physical ill-ness on their lives and emotional well-being For individuals with significant cognitive impairment, interviews were conducted with visual aids and additional time was allowed

Instruments.A general demographic section recorded age, gender, race, religion, education, marital status, participant’s occupation, spouse’s occupation, primary diagnosis, length of residency in home, al-cohol and prescription drug use, home ownership, and living arrange-ment

The dependent variable, life satisfaction, was measured using the LSI-Z

(Wood, Wylie, & Sheafor, 1969) The LSI-Z is a shorter form of the LSI-A (Neugarten, Havighurts, & Tobin, 1961) The LSI-Z contains 14

items that respondents rate agree, disagree, or uncertain Possible total

scores range between 1 and 14 with higher scores indicating higher life satisfaction The LSI-Z is a self-report measure that has been normed on

100 elderly subjects with a mean life satisfaction score of 11.6 and a stan-dard deviation of 4.4 (Sauer & Warland, 1982) The LSI-Z has a Kuder-Richardson reliability1coefficient of 79

The Iowa Self-Assessment Inventory (Morris & Buckwalter, 1988) is

56-item scale that consists of seven subscales including economic re-sources, emotional balance, physical health, trusting others, mobility, cognitive status, and social support Items are rated on a scale from 1

(usu-ally or always true) to 4 (usu(usu-ally or always false) The scores for each item are summed to comprise a score that ranges from 8 to 32 This inventory primarily has been normed on adults 65+ years old (Morris, Buckwalter, Cleary, Gilmer, & Andrews, 1992) The internal consistency reliability coefficients for the seven subscales range from 74 to 86 (Morris, Buckwalter, Cleary, Gilmer, Hatz, & Studer, 1990) Construct validity has been shown between each subscale and similar measures (Morris, Buckwalter, Cleary, Gilmer, Hatz, & Studer, 1989)

The Geriatric Scale of Recent Life Events (Kiyak, Liang, & Kahana,

1976) is a 55-item scale that measures the number of life events that have occurred within the previous three years Subjects circle all events that they have experienced, and the number of events circled is summed to comprise a total score Item examples include: “Death of a Close Friend”;

“Minor Illness”; “Gain a new family member”; and “Grandchild mar-ried.” This scale was normed on 248 individuals 60+ years old and has shown good concurrent validity (.51 to 84) (Kiyak, Liang, & Kahana, 1976) The analysis reported here was based on a subscale of negative events constructed for the study It included 32 items

The Multidimensional Health Locus of Control Scale (MHLC) (Wallston,

Wallston, & DeVellis, 1978) is an 18-item instrument that rates respon-dents according to internal health locus of control (IHLC), powerful oth-ers health locus of control (PHLC), and chance health locus of control (CHLC) Respondents answer questions on a 4-point, Likert-type scale (1

= strongly disagree to 4 = strongly agree) Scores on each subscale are

summed, and higher scores indicate more external beliefs in locus of con-trol Typical questions include: “No matter what I do, if I am going to get sick, I will get sick” and “Most things that affect my health happen to me

by accident” (Wallston et al., 1978) Inter-item reliability has been esti-mated to range from 67 to 86 Concurrent validity with similar locus of control measures has been estimated to range from 51 to 73 This scale was normed on the general population, with a median age range from 35 to

44 years

The Index of Activities of Daily Living (ADL) (Katz, Ford, Moskowitz,

Jackson, & Jaffee, 1963) is a widely used measure that assesses a person’s ability to carry out daily tasks such as walking, getting out of bed, climb-ing stairs, groomclimb-ing, bathclimb-ing, dressclimb-ing, toiletclimb-ing, and feedclimb-ing Items are rated “yes” or “no” in terms of dependence or independence for each ac-tivity This index has shown good interrater reliability, as well as strong correlations with measures of mobility and house confinement (Kane & Kane, 1981)

Finally, perceived health in general was measured by a 5-point Likert scale item “In general, would you say your health is excellent, very good, good, fair, or poor.”

Data Analysis

Two multiple regression methods, hierarchical and stepwise methods, were conducted to determine which psychosocial measures were signifi-cantly associated with life satisfaction among respondents Measures that have shown significant correlations (p < 05) with life satisfaction were entered in the regression equation These variables were number of nega-tive life events, emotional balance, physical health, cogninega-tive status, so-cial support, powerful others locus of control, and perceived health in general The dependent variable was life satisfaction, as measured by the LSI-Z

Prior to these analyses, frequencies and histograms were generated for all variables to assess distributions and outlying cases All variables were normally distributed and had no outliers For the regression analysis, diag-nostics were performed to ensure multivariate assumptions were met Bivariate correlation matrices, variance inflation factor (VIF) values, and tolerance criteria indicated no multicollinearity among the independent variables

FINDINGS Descriptive Findings

Life Satisfaction.To describe the life satisfaction of participants in this study, we computed descriptive statistics using the dependent variable, life satisfaction The mean score on the Life Satisfaction Index was 9.6 (SD = 2.5), slightly below the cutoff score (11.6) for a normal population

of elderly as reported by Sauer and Warland (1982) Scores on the life sat-isfaction ranged from 3 to 14, with 39% of participants reporting a score of

11 or higher (61% below the cutoff score)

Cognitive Status.Cognitive status was gauged using the Mini Mental State Exam, as well as the cognitive status sub-scale from the Iowa Self-Assessment Inventory The mean score on the Mini Mental State Ex-amination was 25.8, with a range from 13 to 30 Folstein, Folstein, and McHugh (1975), suggest using a criterion of 21 or less to gauge cognitive impairment Using this cut-off, only 14% of respondents showed

signifi-cant cognitive impairment Further evidence of cognitive capacity was provided by the cognitive status subscale from the Iowa Self-Assessment Inventory The mean score on this measure was 21.7, with a range from 8-32

Locus of Control.Participants in this study showed a clear tendency to-ward external causal attributions Their mean score on the Internal Locus

of Control sub-scale was 5.6, with a range from 2 to 8 The mean score on external (“powerful others”) sub-scale was 14.7, with a range from 6 to

24 Respondents’ mean on the chance sub-scale was 12.4, with a range from 6 to 19 Mean scores obtained when this instrument was normed were 25.1 on the Internal Locus of Control sub-scale; 19.9 on the Powerful Others sub-scale, and 15.6 on the chance sub-scale (Wallston et al., 1978) Table 1 displays results for these descriptive analyses

Life Events.The average number of negative life events participants ex-perienced in the year prior to this study was 7.5, with a range from 2 to 17 The most common event reported was difficulty walking, experienced by 85% of the sample This was followed by minor illness (84%), reduced recreation (76%), illness of a family member (60%), hearing and vision problems (56%), and less church activity (49%) Nearly half (48%) of the sample had experienced death of a family member Table 2 displays the negative life events included in this study

Predicting Life Satisfaction

Hierarchical and stepwise multiple regression methods were conducted

to estimate a model predicting life satisfaction As was previously men-tioned, variables that were significantly correlated with life satisfaction were entered in the regression equation Table 3 presents the correlations between the dependent variable (life satisfaction) and the independent measures

The results of both hierarchical and stepwise multiple regression methods were consistent, and therefore, the results of the stepwise are reported in this study These results are presented in Table 4 These re-sults show that life satisfaction is a function of physical health (beta =

.26, p < 0001), social support (beta = 19, p = 007), emotional balance (beta = 34, p = 039), and powerful others locus of control (beta = 28, p =

.003) The results show that physical health has the strongest contribution

to the variance of life satisfaction It contributes 14% of the variance in life satisfaction, while each one of the other variables contributes 6% of the

variance All four variables combined contribute 32% (R = 57) of the total

variance of life satisfaction

This sample of frail elders reported levels of life satisfaction that were somewhat lower than those observed in studies of more healthy elders In this study, the mean score of life satisfaction (9.6) was slightly below scores reported in studies of healthy seniors In these studies, mean life satisfaction scores have ranged from a low of 11.6 to a high of 15.39 (Kahana et al., 1995; Wood, Wylie, & Sheafor, 1969; Rao & Rao, 1981; Gray, Ventis, & Hayslip, 1992; Adams, 1969) This difference is small, and a significant proportion of respondents in this study reported levels of satisfaction that were well within the range of those reported by more healthy elders Indeed, these results underscore what practitioners in the

field already know Namely, that not all frail elders experience low life

satisfaction In this sample, nearly half reported high scores on this mea-sure

It is tempting to attribute the sample’s lower mean life satisfaction to health and functional limitations But results of the multivariate analysis

do not support this interpretation The subjective measure, perceived physical health was an important predictor of life satisfaction, whereas more objective health measures were not In bivariate analyses reported in Table 3, more objective measures of functional ability (IADL and ADL)

TABLE 1 Respondents’ Cognitive and Emotional Status (N = 99)