This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation

Stress in Caregivers of Hospitalized Oldest-Old Patients

^a University of Tennessee College of Medicine—Chattanooga Unit, Department of Medicine, Tennessee
^b Marshfield Medical Research Foundation, Wisconsin
^c Division of Health Management & Policy, School of Public Health, University of Minnesota, Minneapolis
^d Center to Improve Care of the Dying, Washington, DC

Norman A. Desbiens, Department of Medicine, University of Tennessee College of Medicine—Chattanooga Unit, 975 East Third Street, Box 94, Chattanooga, TN 37403 E-mail: desbiena{at}erlanger.org.

Decision Editor: John E. Morley, MB, BCh

	Abstract

Top Abstract Methods Results Discussion References

 
Background. Stress in caregivers of elderly patients is a well-recognized health care problem. However, little has been published about the stress in caregivers of the oldest-old patients, the most rapidly growing segment of our population.

Methods. A prospective cohort study was conducted in four teaching hospitals. Questionnaires were administered to patients 80 years of age and older and their surrogates (the person who would make decisions if the patient were unable to—usually a family member) who identified themselves as the primary caregivers for the patients. Data were abstracted from medical records.

Results. Caregivers tended to be female and 50 years of age or older. About one in five described her own health as fair or poor; nearly half of them lived with the patient. About one quarter spent at least 8 h/d caring for the patient, and they had few persons available to help them with care. Most of the caregivers reported mild-to-moderate levels of stress. After adjustment, higher stress scores were associated with female caregivers, poorer caregiver health, more hours per day spent caring for the patient, and the presence of patient depression and hearing impairment.

Conclusion. Stress is common in caregivers of the hospitalized oldest-old patients. Women who are in poor health and spend 8 or more hours every day caring for relatives aged 80 and over are at high risk for caregiver stress. Treatment of patient depression and hearing impairment may ameliorate caregiver stress.

THE aging of our population, the ability to prolong life without curing underlying chronic illness, and the shifting of care from hospitals to outpatient venues necessitate that the needs of elderly persons for assistance with instrumental and basic activities of daily living will continue to burgeon well into the twenty-first century ⁽¹⁾. Recognition of this growing need has increased emphasis on support of the informal caregivers of elderly patients who provide most of the care. Our ability to nurture these nurturers requires identification of the most burdensome aspects of caregiving ⁽²⁾.

Although recent studies have looked at caregiving needs in selected patient populations ⁽³⁾⁽⁴⁾, the caregivers of the hospitalized oldest-old patients have not been adequately studied ⁽⁵⁾. These caregivers are an important population to study because the extra burdens of the convalescent period often fall upon the person who was providing care prior to hospitalization. Although hospitalization most commonly returns patients to the level of function that they had prior to the illness that led to the acute hospitalization, it usually does not improve the function that the patient had prior to the acute illness ⁽⁶⁾. Therefore, on average, caregiver stress would be expected to increase in the convalescent period only to decrease to its baseline level afterwards.

As part of the Hospitalized Elderly Longitudinal Project (HELP), a national study of outcomes and decision making in hospitalized patients 80 years of age and older, we interviewed caregivers and studied factors associated with the stress they perceived in caring for these patients. On the basis of the literature and our judgment, we hypothesized that caregiver gender, burden of care, and patient factors (e.g., presence of dementia and number of dependencies in activities of daily living [ADLs]) would be associated with increased caregiver stress.

	Methods

Top Abstract Methods Results Discussion References

 
Study Description
HELP is a study of patients 80 years of age and older consecutively hospitalized at four medical centers (Beth Israel Hospital, Boston, MA; MetroHealth Center, Cleveland, OH; Marshfield Clinic/St. Joseph's Hospital, Marshfield, WI; and the University of California at Los Angeles Hospital, Los Angeles, CA) in 1994 ⁽⁷⁾. Patients were excluded if they had elective surgery, had AIDS, were discharged within 48 hours of admission, or had multiple traumas. In order to even the study burden across sites, all patients admitted at the Ohio and California sites were enrolled, while at the Wisconsin and Massachusetts sites 40% random samples of consecutively admitted patients were chosen.

Data Collection
To protect patients' safety, approval for the study was independently obtained from the institutional review board at each site. Data were abstracted from patients' medical records and obtained from interviews with patients who had passed cognitive screening and their surrogates (the person who would make decisions if the patient were unable—usually a family member). Data obtained from the medical record included patient age, gender, study hospital, number of comorbid conditions, laboratory and physical exam data needed to calculate the acute physiologic score (APS) of APACHE III ⁽⁸⁾, the number of medications the patient was taking upon admission, and the presence of any of the following conditions: hearing impairment, vision impairment, previous dementia, prolonged bed rest or being bed-bound, fecal incontinence, urinary incontinence, frailty or "failing," cancer, depression, and diabetes.

An early hospital interview was conducted with patients (median day after admission: 3; interquartile range: 2–4) and surrogates (median day after admission: 4; interquartile range: 3–6). Patients who had passed a cognitive screen [the revised version of the information and orientation subset of the Wechsler Memory Scale ⁽⁹⁾] and surrogates were asked about patients' education, income, race, hearing and vision impairment, dependencies in ADLs 2 weeks prior to admission, falling at home in the last 6 months, and number of hospital admissions in the previous year. Surrogates were also asked if they were primary caregivers for the patients and, if so, for their gender, age, relationship to the patient, the number of hours per day they cared for the patient, the number of persons inside and outside the home who could help care for the patient, and a 5-point estimate of their own health (poor...excellent).

Adopted from the work of Silliman ⁽¹⁰⁾ and on the basis of a conceptual framework by Pearlin and colleagues ⁽¹¹⁾, surrogates who identified themselves as primary caregivers were also asked the following six questions, each answered using a 5-point Likert scale as "not at all," "a little," "somewhat," "a lot," and "a great deal": Has caring for the patient: caused your health to suffer?, made you feel isolated and alone?, added tension in your life?, caused you to feel tired all the time?, reduced time for yourself?, made you feel that you have lost control of your life? A caregiver stress score was calculated by adding the points for each of the six questions; therefore, the possible range of scores was 6 (no stress) to 30 (maximal stress).

Analyses
Descriptive comparisons.-- Characteristics of patients with and without caregivers were compared using chi-square tests for categorical variables and Wilcoxon tests for continuous variables.

Stress scale evaluation.-- Internal consistency of the stress score was assessed using Cronbach's alpha ⁽¹²⁾ (SPSS, Inc., Chicago, IL) and factorial composition ⁽¹³⁾ was studied using principal components analysis (SAS Institute, Cary, NC).

Association between stress score and other variables.-- The general linear model (GLM; SAS Institute, Cary, NC) was used to study the association between the stress score and the following characteristics: patient demographics (age, gender, and income); patient disease burden (patient estimate of number of hospital admissions in the past 6 months, patient estimate of whether falls occurred at home during the past 3 months, and APS score assessed at Day 3 in the hospital); care resources (whether the caregiver lives with the patient and number of persons inside and outside the home who could help care for the patient); burden of care (patient depression, hearing impairment if indicated in the chart or during the patient or surrogate interview, dementia, number of dependencies in ADLs as assessed by the caregiver, and caregiver assessment of hours per day caregiver gives care to the patient), caregiver demographics (relationship to patient, rating of own health, and gender) and study hospital. The relationship between the stress score and each continuous independent variable was assessed for nonlinearity using restricted cubic splines ⁽¹⁴⁾. The association between the stress score and number of hours per day that the caregiver took care of the patient was significantly nonlinear. To facilitate the presentation of results, we repeated the analyses using this variable categorized by quartiles. These changes had only small effects on the regression coefficients in our model. Model assumptions were checked by residual and normal probability plots. A two-sided alpha level of .05 was used as the cutoff for determining statistical significance.

Missing values.-- There were 395 cases without any missing values for the variables in the GLM model. To reduce the amount of bias from missing responses for the independent variables, an imputation and substitution strategy was developed. For the number of patient's dependencies in ADLs, the surrogate response was used; if missing, patient responses calibrated to the surrogate response were substituted (6 cases, 1.4%). For missing caregiver gender, the gender opposite the patient was used if the caregiver was the spouse (5 cases, 1.2%). To obtain a value for the number of hours per day that the caregiver gave care to the patient for 24 cases (5.6%), a stepwise linear regression model (SAS Institute, Cary, NC) with a p value for variable inclusion of .15 was employed. The independent variables in this model were caregiver gender and co-residency; patient ADL dependencies; presence of dementia, diabetes, or vision impairment; number of medications; and study hospital. Using these strategies, 428 cases (95.1% of the 450 cases who completed all stress questions) were able to be used in the multivariable model.

Characteristics of Stress Score
The correlations of each question with each of the five others ranged from .47 to .68. The questions had acceptable internal consistency, with Cronbach's alpha equal to .90. Principal component analysis showed that three components of the six accounted for 85% of the variance of the questions. The first component accounted for 68% of the variance; its eigenvector had approximately equal loadings for all questions. The second component, which explained 9% more of the total variance, had high positive loadings for the questions pertaining to caregiver health and isolation and negative loadings for the remainder of the questions. The third component, which explained 8% more of the total variance, had high positive loadings for questions pertaining to caregiver isolation and loss of control, near neutral loading for the question pertaining to reduced personal time, and negative loadings for other questions.

	Results

Top Abstract Methods Results Discussion References

 
Participants
Twelve hundred sixty six patients were enrolled in HELP. Of the 921 surrogates interviewed, 463 (50.3%) identified themselves as primary caregivers; 450 answered all six questions about caregiver stress and are the subject of this report. Compared with other patients, patients whose surrogates were caretakers were more often non-white, had less education and slightly higher yearly income, had more comorbidities and vision and hearing impairment, and were diabetic (Table 1 ).

Caregiver Stress
In decreasing order of frequency, surrogates who were caregivers responded "a lot" or "a great deal" to the following issues in caring for the patient: added tension to their lives, 26.1%; reduced time for themselves, 23.2%; always made them feel tired, 15.7%; made them feel isolated, 11.0%; and made them feel a loss of control of their own lives, 10.4% (Table 3 ). Their stress scores ranged from 6 to 30, with 5th, 25th, 50th, 75th, and 95th percentiles of 6, 7, 10, 15, and 24, respectively.

After adjustment for markers of patient demographics, patient disease burden, care resources, burden of care, caregiver demographics, and study hospital (see Methods and Table 3 ), two patient characteristics were associated with greater caregiver stress: depression and hearing impairment. Caregivers of patients who were depressed had scores that were on average 1.7 (95% CI: 0.0–3.5) points higher than those of patients who were not depressed. Caregivers of patients who had hearing impairments had scores that were on average 1.7 (95% CI: 0.4–2.9) points higher than those of patients without such impairments.

	Discussion

Top Abstract Methods Results Discussion References

 
The octogenarians and nonagenarians who comprised our unselected group of hospitalized patients had the types of characteristics, diseases, and sensory defects that would be expected in this population. They had less dementia and fewer ADL dependencies than would be seen in a comparably aged nursing home population.

As expected, the caregivers were more often women, and, on average, about 30 years younger than the patients. About one in five reported their own health as fair or poor, and nearly half of them lived with the patient. About a quarter spent 8 or more hours per day caring for the patient and had few persons inside or outside the house who would be available to help them with care. Twenty-five percent had scores greater than 15 on our 30-point stress scale.

We hypothesized that caregiver gender would be associated with increased stress. After adjustment for other factors, gender was associated with caregiver stress, with women caregivers reporting stress scores that were, on average, slightly more than 2 points higher than the men. Some ^(15)(16) but not all ⁽¹⁷⁾ studies have also suggested that women tend to be more stressed than men, but these studies have been small and not entirely persuasive ⁽¹⁶⁾.

Burden of care was also associated with caregiver stress, but after adjustment, the number of hours per day spent caring for the patient and not the patient's dependencies in ADLs was significant. Stress scores increased, on average, four points when caregivers reported caring for the patient 8 or more hours per day. Caregivers who spend long hours with patients may be less able to meet their own needs for socialization. Indeed, the principal component analysis of the stress score indicated that isolation was a prominent variable in the first three factors that explained 85% of the total variance of the score. Our findings are consistent with those published in a recent British study of the caregivers of hospitalized patients ⁽⁵⁾. In that study, nearly half reported that caregiving had affected their social lives. Recent U.S. studies also document restriction of social activity or isolation caused by caregiving ⁽¹¹⁾ and the importance of a social network in caregiver health.

We also demonstrated a strong effect of caregiver health on perceived stress. Caregivers who rated their health as poor had stress scores that were on average nearly 5 points higher than those who rated their health as excellent. Several studies have also reported an association between physical and psychological health in caregivers ^(10)(18). Although caregiving by those in poor health might be more burdensome and lead to greater stress, it is also possible that stress affects caregivers' health ⁽¹⁹⁾.

We found that patient depression documented in the medical record was associated with higher stress scores. Although many studies have investigated depression in caregivers, and several studies have looked at adjustment of caregivers of depressed older adults ^(20)(21), none has studied depression in the patient as a risk factor for stress in the caregiver. This finding deserves further study using a validated depression scale.

After adjustment for other factors, hearing impairment was associated with increasing caregiver stress, with, on average, 1.6 points higher stress scores if present. Although it is certainly plausible that continued repetition or attempts to communicate could increase caregiver stress, to our knowledge this association had not been previously reported. It bears confirmation, however, because patient hearing amplification devices might decrease caregiver stress.

Our study has several limitations. Although we studied many variables associated with caregiver stress, we did not study other causes of stress, such as caregiver life events ⁽²²⁾, nor did we study factors that might modulate caregiver stress, such as coping abilities ⁽²³⁾ and previous psychiatric history. While not refuting our findings, consideration of such factors would refine our understanding of caregiver stress. In addition, we did not examine whether there is an association between our stress scores and illness in caregivers; although such an association is plausible, its demonstration will require other studies. While our study emphasizes that there is stress involved in caring for the oldest old patients, we did not address the counterbalancing joys that caregivers often derive in caring for these patients ⁽²⁴⁾.

Ours is the largest study of stress in caregivers of unselected hospitalized patients. Women who spent more than 8 hours each day caring for their patients and who identified themselves as being in poor health were at high risk for caregiver stress, with stress scores (range: 6–30) that were over 11 points higher, on average, than male caregivers who spent less than one hour caring for patients and who perceived their health to be excellent. Although the majority of caregivers in our unselected group were not seriously stressed by caring for their older relatives, 25% of caregivers had scores of 15 or greater, levels that might be associated with clinical depression or anxiety. Several of the factors that we identified as associated with caregiver stress are potentially modifiable. For example, assistance can be found for caretakers, and patients' depression and hearing impairment can be treated.

In summary, in caregivers of the oldest-old patients, caregiver gender, the time that the caregiver spent caring for the patient, the caregiver's health, and patient depression and hearing impairment were associated with caregiver stress. Treatment of modifiable factors might lessen caregiver stress.

	Acknowledgments

 
We thank all the patients, family members, and physicians who participated in this study, the entire HELP staff that collected and entered the data presented in this paper, the HELP investigators, and the Robert Wood Johnson Foundation who funded this study. (The opinions and findings contained in this manuscript are those of the authors and do not necessarily represent the views of The Robert Wood Johnson Foundation or its Board of Trustees.)

Received March 28, 2000

Accepted March 29, 2000

	References

Top Abstract Methods Results Discussion References

 

1. Brody EM, 1995. Prospects for family caregiving: responses to change, continuity, and diversity. Kane RA, Penrod JD, , ed.Family Caregiving in an Aging Society 15-28. Sage Publications, Thousand Oaks, CA.
2. Kane RA, Penrod JD, 1995. In search of family caregiving policy: general considerations. Kane RA, Penrod JD, , ed.Family Caregiving in an Aging Society 1-14. Sage Publications, Thousand Oaks, CA.
3. Barrett JJ, Haley WE, Powers RE, 1996. Alzheimer's disease patients and their caregivers: medical care issues for the primary care physician. South Med J. 89:1-9. [Medline]
4. Boucher L, Renvall MJ, Jackson JE, 1996. Cognitively impaired spouses as primary caregivers for demented elderly people. J Am Geriatr Soc. 44:828-831. [Medline]
5. Liston R, Mann L, Banerjee A, 1995. Stress in informal carers of hospitalized elderly patients. J R Coll Physicians Lond. 29:388-391. [Medline]
6. Wu AW, Damiano A, Lynn J, et al. 1995. Predicting future functional status for seriously ill hospitalized adults: the SUPPORT model. Ann Intern Med. 122:342-350. [Abstract/Free Full Text]
7. Lynn J, Teno JM, Phillips RS, et al. 1997. Perceptions by family members of the dying experience of older and seriously ill patients. Ann Intern Med. 126:97-106. [Abstract/Free Full Text]
8. Knaus WA, Wagner DP, Draper EA, et al. 1991. The APACHE III prognostic system: risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 100:1619-1636. [Abstract/Free Full Text]
9. Wechsler D, 1987. Wechsler Memory Scale (revised) The Psychological Corporation, New York.
10. Silliman RA, 1993. Predictors of family caregivers' physical and psychological health following hospitalization of their elders. J Am Geriatr Soc. 41:1039-1046. [Medline]
11. Pearlin LI, Mullan JT, Semple SJ, Skaff MM, 1990. Caregiving and the stress process: an overview of concepts and their measures. Gerontologist. 30:5835-5894.
12. Nunnally JC, 1978. Assessment of reliability. Psychometric Theory 2nd ed. 225-255. McGraw-Hill, St. Louis, MO.
13. Nunnally JC, 1978. Validity. Psychometric Theory 2nd ed. 86-113. McGraw-Hill, St. Louis, MO.
14. Harrell FE, Jr Lee KL, Pollock BG, 1988. Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst. 80:1198-1202. [Abstract/Free Full Text]
15. Lutzky SM, Knight BG, 1994. Explaining gender differences in caregiver distress: the roles of emotional attentiveness and coping styles. Psychol Aging. 9:513-519. [Medline]
16. Miller B, Cafasso L, 1992. Gender differences in caregiving: fact or artifact?. Gerontologist. 32:498-507. [Abstract]
17. Pot AM, van Dyck R, Deeg DJ, 1995[Perceived stress caused by informal caregiving. Construction of a scale]. Ervaren druk door informele zorg. Constructie van een schaal. Tijdschr Gerontol Geriatr. 26:214-219. [Medline]
18. Franks MM, Stephens AP, 1996. Social support in the context of caregiving: husbands' provision of support to wives involved in parent care. J Gerontol Psychol Sci 51B:P43-P52.
19. Kiecolt-Glaser JK, Glaser R, Gravenstein S, Malarkey WB, Sheridan J, 1996. Chronic stress alters the immune response to influenza virus vaccine in older adults. Proc Natl Acad Sci USA. 93:3043-3047. [Abstract/Free Full Text]
20. Hinrichsen GA, 1991. Adjustment of caregivers to depressed older adults. Psychol Aging. 6:631-639. [Medline]
21. Hinrichsen GA, Hernandez NA, Pollack S, 1992. Difficulties and rewards in family care of the depressed older adult. Gerontologist. 32:486-492. [Abstract]
22. Russo J, Vitaliano PP, 1995. Life events as correlates of burden in spouse caregivers of persons with Alzheimer's disease. Exp Aging Res. 21:273-294. [Medline]
23. Solomon P, Draine J, 1995. Subjective burden among family members of mentally ill adults: relation to stress, coping, and adaptation. Am J Orthopsychiatry. 65:419-427. [Medline]
24. Walker AJ, Martin SS, Jones LL, 1992. The benefits and costs of caregiving and care receiving for daughters and mothers. J Gerontol Soc Sci. 47:S130-S139.

This article has been cited by other articles:

				M. Pinquart and S. Sorensen Gender Differences in Caregiver Stressors, Social Resources, and Health: An Updated Meta-Analysis J. Gerontol. B. Psychol. Sci. Soc. Sci., January 1, 2006; 61(1): P33 - P45. [Abstract] [Full Text] [PDF]

				M. W. Rabow, J. M. Hauser, and J. Adams Supporting Family Caregivers at the End of Life: "They Don't Know What They Don't Know" JAMA, January 28, 2004; 291(4): 483 - 491. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation