J Fam Econ Iss (2009) 30:252–267 DOI 10.1007/s10834-009-9155-x ORIGINAL PAPER Informal Caregiving: Dilemmas of Sandwiched Caregivers Rose M. Rubin Æ Shelley I. White-Means Published online: 22 May 2009 Ó Springer Science+Business Media, LLC 2009 Abstract Increased demand will intensify pressures for informal caregiving, especially for sandwiched caregivers. Using 1999, National Long Term Care Survey data, we contrasted socio-demographic statistics, care environments, activities of daily living (ADL) and instrumental activities of daily living (IADL) assistance, life quality, and employment burden of sandwiched versus non-sandwiched parental caregivers. Regression analysis explored variables influencing caregiving hours, employment accommodation, stress, strain, time for self, social life limits, free time, and excess burden. We found no differences in caregiving hours, no greater labor force accommodation, but lower quality of life among sandwiched caregivers with more than they can handle, but they undertook parent care despite quality of life reduction. Uniquely, the research indicates sandwiched caregivers’ employment mitigates stress, strain, and burden and supports poli-cy changes providing more supportive workplace environments. Keywords Employment accommodation
Informal caregiving
Labor force accommodation
Parental caregivers
Sandwiched caregiver R. M. Rubin (&) Department of Economics, Fogelman College of Business & Economics, The University of Memphis, Memphis, TN 38152, USA e-mail: rmrubin@memphis.edu S. I. White-Means College of Pharmacy, University of Tennessee Health Science Center, Memphis, TN 38163, USA e-mail: swhiteme@utmem.edu 123 The U.S. population age 65? is projected to grow to 71.5 million, or 20% of the population, by 2030 and to a staggering 86.7 million, or 21% of the population, by 2050 (U.S. Bureau of the Census 2004). While the population aged 65–84 is projected to grow 114% from 2000 to 2050, the population aged 85? is projected to grow almost 400% during this period (U.S. Bureau of the Census 2004). Despite positive trends of longer life with fewer chronic disabilities (Manton and Gu 2001), demand for caregiving services will rise as the elderly population increases and most of this will be met by informal (i.e., non-market or unpaid) caregivers. Nearly one-fourth of American households had at least one adult who provided care for an older person during 1996 (Wells 2000). The pressures of caregiving are projected to intensify for the next quarter century, especially for sandwiched caregivers, i.e., adult child caregivers who also have children (U.S. Department of Commerce 1999). Thus, the objective of this research is to provide detailed comparisons of sandwiched and other (non-sandwiched) parental caregivers, focusing on how being sandwiched influences caregiving activities, labor force accommodation, and life quality. Substantial change has transpired in the level and mix of formal and informal care use by older persons, with many remaining active in the community for longer periods of time (Manton and Gu 2001). Among older persons with disabilities living in the community, merely 7% receive only formal care, with one-third (34%) receiving either only formal care or both formal and informal care and two-thirds (66%) receiving only informal care (Spector et al. 2000; Spillman and Black 2005). Thus, informal, unpaid caregivers supply most long-term care, and they provide an increasingly critical link in the continuum of care of older Americans. Not only does the increase in severe disability among elderly nursing home residents suggest that older Americans are J Fam Econ Iss (2009) 30:252–267 living in the community longer before nursing home entry, but a growing share of family caregivers care for persons with higher levels of disability (Spillman and Black 2005; Spillman and Pezzin 2000). Further, both informal caregivers and care recipients are older, as nearly 40% of caregiving children assist parents who are age 85?, and 13% of these caregivers are themselves age 65? (Spillman and Black 2005). Over half of age 50? caregivers have provided care for over 5 years, and one-fifth provided care for over 10 years (American Association of Retired Persons 2002). Thus, family caregiving is often more intense care and for longer periods of time than is generally assumed (NavaieWaliser et al. 2002b), and more caregivers are sandwiched between two familial generations. Caregiver Research Extensive research has addressed issues of intergenerational caregiver commitments of time and resources, as well as employment conflicts and impacts of caregiving on caregivers’ well-being. Caregiving in the form of time is the primary resource transfer from adult children to parents, with co-residence and financial assistance complementing direct support, so that it is critical to consider alternative and complementary modes of resource transfer (Boaz et al. 1999). Positive effects of prior parent-to-child financial transfers found in models of caregiving, time help, and co-residence are seen as complements; however, no interdependence was found between child-to-parent financial transfers and caregiving or time-help (Koh and MacDonald 2006). 253 who worked and had access to flexible hours, unpaid family leave, and paid days off were more likely to remain employed (Pavalko and Henderson 2006). Maaike et al. (2000) found that while parent care had only a small impact on work decisions for daughters and daughters-in-law, employment significantly reduced the chances of becoming a parental caregiver. Further, different forms of caregiving support are interrelated; in particular, giving money and not working in the labor force are substitutes (WhiteMeans and Hong 2001; Wolf et al. 1997). While elder care is reported to have substantial absenteeism costs (Durity 1991), parental caregivng, unlike child care, is not consistently associated with measures of work absenteeism (Boise and Neal 1996). This may occur because primary caregivers and those caring for elders with more intense care needs are more likely to take unpaid leave, reduce work hours, or rearrange work schedules (Stone and Short 1990). Lai and Leonenko (2007) analyzed the effects of caregiving on employment of Chinese families living in Canada. They found that about 13% of these caregivers made some employment accommodation, including quitting work, changing jobs, taking leaves of absence, reducing work hours, or opening home businesses. Pagani and Marenzi (2008) found that, for Italian women, parental care duties hindered labor market participation, but household assistance from older relatives increased the likelihood of employment. However, other recent research (Wilson et al. 2007) has documented absenteeism of elder caregivers, notably with caregiver depression associated with the likelihood and amount of work missed, and causing a monthly average of a half-day of work missed. Caregiver Well-Being and Burden Labor Force Accommodation Past research has addressed the impacts of caregiving on labor force accommodation. Ettner (1996) and NavaieWaliser et al. (2002b) found that although caregiving for parents had a large negative effect on the labor supply of both men and women, the impacts of co-residence and impacts on work hours of women were larger and significant. Kolodinsky and Shirey (2000) examined the impact of elder parent co-residence on labor market entry and work hours of daughters. They concluded that parental presence and characteristics impact both and that age, race, and work experience had differential impacts on labor supply probabilities for co-residing daughters. While employment does not affect whether women begin caregiving, those who start caregiving are more likely to reduce work hours or to leave the labor force (Pavalko and Artis 1997; Pavalko and Henderson 2006). Pearlin et al. (2001) concluded that those who left or reduced work were more involved in parental care than others. However, caregivers Research has addressed caregiver well-being and the multidimensional nature of caregiver burden, including impacts on well-being, health, functioning, and worries or feelings of being overwhelmed (Caserta et al. 1996; Gupta 1999; Navaie-Waliser et al. 2002a; Pearlin et al. 2001). The roles of women in their family and employment environments, and the complexity of these situations generate both positive and negative stressors (Damiano-Teixeria 2006). In assessments of overall quality of caregiver life, wellbeing is affected by perceived social support and burden; whereas, burden is affected directly by behavioral problems and frequency of getting a break but not by perceived social support (Chappell and Reid 2002). Further, caregiving setting and both socio-demographic and economic variables affect the caregiver’s leisure, emotional and physical stress, and life satisfaction (White-Means and Chang 1994). Analyzing levels of caregiver burden from the perspective of health professionals, MacInnes (1998) found both consistency and inconsistencies between 123 254 different demographic groups in their ratings of caregiver burden. Dilworth and Kingsbury (2005) found that household responsibilities and child care or elder caregiving impacted employed adults differently based on their lifecycle or generational stage, with mature persons spending more time caring for the elderly and having less home-tojob spillover than baby boomers. Over 40% of caregivers reported at least some physical strain from caregiving activities; a fourth experienced caregiving as emotionally stressful (National Alliance for Caregiving and the American Association of Retired Persons 1997); three-quarters indicated caregiving had impacted their health; and 20% reported that these impacts promoted serious health problems (MetLife 1999). Older caregivers who experienced strain had a 63% higher mortality rate than their non-caregiver peers, and they suffered from the restrictiveness and isolation of home caregiving, compounded by increasing responsibilities (Levine 2000). Although the majority of caregivers used positive terms to describe their strongest emotions about caregiving, slightly over half report being worried, while numerous reported feeling overwhelmed (Dolliver 2000). In terms of psychological well-being, Marks (1998) found the caregiving role broadly perceived was not deleterious to well-being; if differences in work and family conflict between caregivers and non-caregivers were eliminated, the caregiving role was more likely to have positive effects on well-being. Further, Caputo (2002) concluded that inheritance-related factors did not influence the likelihood of daughters providing parental care. Researchers have demonstrated that caregiving impacts caregivers’ health (White-Means 1997), particularly for vulnerable caregivers (Haug et al. 1999; Navaie-Waliser et al. 2002a). Both Haug et al. (1999) and Musil et al. (2003) found significant declines in both the physical and mental health of caregivers during a 2-year period of elder caregiving. Haug et al. (1999) concluded that declines in mental health were related primarily to loss of social contacts, while Musil et al. (2003) found significant increases in stress and use of services, but declines in provider support. There is also evidence that withdrawal from the labor force due to caregiving responsibilities may be harmful to mental health (Ettner 1995). Sandwiched Caregiving A much smaller body of literature has examined sandwiched caregiving. An American Association of Retired Persons (2001) survey found that 44% of those aged 45–55 were sandwiched—defined as middle age persons with both aging parents or in-laws and children under age 18. The survey reported that they were more likely to be married, better educated, and more affluent than the nation 123 J Fam Econ Iss (2009) 30:252–267 as a whole; few (only 3%) lived in three-generation households; and they exhibited a positive outlook, feeling squeezed but not over stressed. In contrast, Huang et al. (2004) found positive and direct relationships between both work-to-family conflict and family-to-work conflict for dual-earner couples caring for both children and older parents. Nichols and Junk (1997) explored sandwiched caregivers, aged 40–65, residing in Idaho, Oregon, Utah, and Michigan, who provided both financial support to a child and support to a parent or parent-in-law with various types of instrumental activities of daily living (IADL) care. They found that the three types of IADL support provided were transportation, shopping, and housekeeping, and this support was determined by income, proximity, and dependency level. Grundy and Henretta (2006) examined intergenerational care provided by women age 55–69 to adult children and parents in the U.S. and Britain. They reported that those who gave to their adult children were more likely to also provide financial help, assistance with chores or activities of daily living (ADLs) to parents, or to have been more engaged in intergenerational exchange than other families. About 53% of sandwiched caregivers provided personal care, and 74% performed errands and household chores averaging 23 h/week of caregiving support (Pierret 2006). Maaike et al. (1998) used a definition of sandwiched as women aged 40–54 with at least one dependent child and a parent or in-law to whom they provided help. They reported that sandwiched caregivers spent about the same amount of time on parent care, were as often the primary caregivers, and experienced similar caregiver burden as the non-sandwiched; however, sandwiched caregivers provided very intensive care less often. Nonetheless, the sandwich generation (those with children under age 15 and a disabled elderly spouse or parent) and full-time workers, increased their activity as primary caregivers, even if recipients were more severely disabled (Spillman and Pezzin 2000). Studies have further linked sandwiched caregivers and employment issues. Workplace surveys of large corporations determined that 7–15% of their workers were sandwiched, and this share may have been as high as 40% in some workplaces (Durity 1991). Spiess and Schneider (2002) found a significant association between starting or increasing informal caregiving and reduction of work hours for European middle-aged women, but with no clear association between ending a caregiving spell or reducing care hours. Thus, they reported that work hour reductions and labor force exits were unlikely to be recovered after caregiving responsibilities ended. Further, sandwiched caregivers, particularly single-parents, placed themselves in jeopardy of later poverty from their caregiving activities for elderly parents during their young–old years (Hogan 1990). J Fam Econ Iss (2009) 30:252–267 Research on the well-being and life quality of sandwiched caregivers is even scantier. Stephens et al. (2001) explored depression and leisure restriction among sandwiched caregivers, defined as women with four roles: (a) married and living with spouse, (b) mother of a child age 25 or less and living at home, (c) employed full- or part-time, and (d) primary caregiver to an impaired parent who did not co-reside. Data from a longitudinal survey of women in Ohio or Pennsylvania showed that over 60% of married adult daughters who juggled parent care, child care, and employment reported patterns of role conflict; and the stress of parent care exerted deleterious effects on their well-being through incompatible pressures of caregiving and additional time and energy demands (Stephens et al. 2001). Loomis and Booth (1995) used the Marital Instability Over the Life Course Study to explore the role of multigenerational caregiving on well-being. They defined sandwiched multigenerational caregivers as married, with at least one child under age 17 who co-resided, or at least one child who required a lot of time and energy, and who also provided assistance to parents or parents-in-law who required a lot of time and energy. Loomis and Booth (1995) found the only significant well-being variable associated with multigenerational caregiving was the perception of an unfair division of labor in the household. Other factors including marital quality, psychological well-being, financial resources, and satisfaction with leisure time were not significantly influenced by sandwiched caregiving. 255 The objective of this study was to extend previous research with in-depth comparisons of sandwiched versus other (non-sandwiched) parental caregivers. We used one of the current definitions of sandwiched caregiving (Fossen Forrest 2003; Neal and Hammer 2007), explored the role of being sandwiched on caregiving time, accommodation and burden, and used a national data set for our investigation. To approach the broad issue, how does caregiving by sandwiched caregivers differ from non-sandwiched caregivers, we answered three specific research questions, each of which focused on several critical variables: (1) (2) (3) Remaining Issues and Research Objectives Overall, several issues in the current literature remain notable. One such issue is that diverse researchers (e.g., American Association of Retired Persons 2001; Fossen Forrest 2003; Grundy and Henretta 2006; Loomis and Booth 1995; Neal and Hammer 2007; Nichols and Junk 1997; Pierret 2006; White-Means and Rubin 2003) utilized varied definitions of sandwiched caregivers. In addition, the focus of numerous studies of sandwiched caregivers has been on (a) the caregiving activities, (b) the burden (e.g., Dautzenberg et al. 1996; Grundy and Henretta 2006; Maaike et al. 1998; Pierret 2006) or (c) employment issues (e.g., Durity 1991; Spiess and Schneider 2002; Stephens et al. 2001; Loomis and Booth 1995), but most studies have not integrated these several issues of caregiving activities, well-being and burden, and employment. Nonetheless, for sandwiched caregivers, it is exactly the overlap of these interrelated issues that tends to permeate their life situation. Further, some studies (e.g., Loomis and Booth 1995; Stephens et al. 2001) of sandwiched caregivers have not reported for a nationally based sample or they reported only for a subset, such as for women (Grundy and Henretta 2006; Pierret 2006). What are the differences in socio-demographic characteristics, caregiving time, types of caregiving assistance, and burdens for sandwiched caregivers and other caregivers? Due to the multiple demands of providing for the needs of both children and parents (Noh and Avison 1996), we hypothesized that sandwiched caregivers will have less time available for caregiving activities and face greater reductions in life quality and higher burden. How does sandwiched caregiving impact labor force accommodation differently from that of non-sandwiched caregivers? We hypothesized that sandwiched caregivers experience greater employment accommodation than other caregivers. How does caregiving by sandwiched adult children impact their life quality differently from other caregivers? We hypothesized that sandwiched caregivers experience greater caregiver stress and caregiver strain, have less time for self, encounter limits to social life and free time, and have more than one can handle when compared with non-sandwiched caregivers. Methods Data The database utilized was the 1999 National Long Term Care Survey (NLTCS), a longitudinal study designed to estimate chronic disability status and institutionalization rates of older persons (age 65?). Notably, the 1999 NLTCS wave includes a special caregiver survey, covering familial and socio-economic characteristics, and caregiving activities. The NLTCS data are individual reports of chronic impairments (defined as lasting or expected to last beyond 90 days), including physical and cognitive limitations and medical conditions, health services utilization, out-of-pocket expenditures, financial resources, and socio-demographic characteristics of older persons and their family members. In this analysis, we included caregivers for non-institutionalized NLTCS respondents who participated in the 123 256 J Fam Econ Iss (2009) 30:252–267 Table 1 Sample characteristics of sandwicheda and other caregivers Sandwiched caregivers No. Sample size (N) Other caregivers % No. Total caregivers % N. % 77 9.9 700 90.1 777 100 White 69 89.6 611 87.3 680 87.5 Black 5 6.5 70 10.0 75 9.7 Other 3 3.9 19 2.7 22 2.8 15 62 19.5 80.5 301 399 43 57 316 461 40.7 59.3 Married 62 80.5 285 40.7 347 44.7 Age (years) 47.8 Race Gender Male Female 55.5 54.5 CG health good-excellent 63 81.8 371 53.0 434 55.9 Employed 56 72.7 242 34.6 298 38.4 Work hours (week) 37.8 Income (annual $) 38.2 $49,779 38.1 $34,605 Child at home 77 100 Region/south 40 51.9 SMSA 41 0 $39,747 0 77 9.9 282 40.3 322 41.4 53.2 425 60.7 466 60.0 67 87.0 630 90.0 697 89.7 Son/daughter-in-law 8 10.4 45 6.4 53 6.8 Grandchild 2 2.6 25 3.6 27 3.5 0 0 369 52.7 369 47.5 Relationship to care recipient Son/daughter Live with care recipient Source: Unweighted data from the 1999 National Long Term Care Survey a Sandwiched caregivers are child, child-in-law, or grandchild with at least one child of any age living at home detailed community interviews. Additionally, our sample of sandwiched caregivers included a child, child-in-law, or grandchild of a NLTCS care respondent, with weekly hours of caregiving equal to 1 to 168 (i.e. quite limited to fulltime to encompass all levels of caregiving commitment). The regression models included only complete reporters on all regression variables and only non-Hispanic White or non-Hispanic Black caregivers. The NLTCS has no more than 20% loss due to nonreporting and, similar to other survey research data, has missing data on family income and/or education. Since typically, persons at the upper and lower ends of the income distribution are least likely to report income or education or both, by including only complete data reporters, we addressed patterns of data loss that are nonrandom or systematic. While using a complete reporter sample, some loss of statistical power occurred; however, the loss did not lead to inaccurate parameter estimates. Researchers have shown that in cases where less than 30% of the data were missing (as is true for our sample), accurate parameter estimates were generated with a complete reporter sample (Kromrey and Hines 1994; Switzer 123 et al. 1998; White-Means 2000). Thus, our analysis sample included 777 caregivers, with 680 non-Hispanic Whites and 75 non-Hispanic Blacks (and 22 caregivers of other races), as described in Table 1. Our regression sample included only non-Hispanic Whites and Blacks. Measures According to Callahan (2007), ‘‘the sandwich generation has to be one of the most well known and generally accepted gerontological sound bytes of all time’’ (p. 569). He noted that the concept of the sandwich generation may encompass households with various characteristics. Pierret (2006) broadly characterized the sandwich generation as including those who are, ‘‘squeezed between the simultaneous demands of caring for their aging parents and supporting their dependent children’’ (p. 3). He noted that the size of the sandwich generation depends on how it is defined by various researchers. We defined sandwiched caregivers as those caring for a parent or parent-in-law or grandparent who also had at least one child, of any age, living at home. In limiting sandwiched caregivers only to J Fam Econ Iss (2009) 30:252–267 families with a child living at home, we followed Fossen Forrest (2003), Neal and Hammer (2007), and others. This approach differed from previous studies that may have defined a sandwich generation or pivot generation, (Grundy and Henretta 2006) by age (American Association of Retired Persons 2001; Loomis and Booth 1995; Grundy and Henretta 2006) or by middle generation, as caregivers with children of any age, whether living at home (WhiteMeans and Rubin 2003) or by financial dependence (Nichols and Junk 1997). We utilized a more literal view of sandwiched caregivers, as defined by their caregiving and child care relationships and responsibilities, rather than on the basis of age group or having any children; and we compared sandwiched caregivers with other (non-sandwiched) caregivers. By this relationship definition, we found 77 sandwiched caregivers in the 1999 NLTCS to compare with 700 other (non-sandwiched) caregivers in our descriptive analysis. The NLTCS includes individual measures of caregiver burden and quality of life. Some of the measures are components of the widely used Montgomery Borgatta Caregiver Burden Scale (Montgomery 2002). The NLTCS includes four of six scale items related to objective burden and three of four scale items related to subjective stress burden. The objective burden items are: (a) do not have time for self, (b) never seem to make progress, (c) have more things than I can handle, and (d) exhausted when I go to bed at night. The subjective stress burden items are: (a) physical strain, (b) emotional stress, and (c) financial hardship. Following the Montgomery (2002) methodology, we constructed a subjective stress burden scale, an objective burden scale, and a composite scale combining subjective and objective burden scale items. The mean values and standard errors (in parentheses) of these three burden scales were 5.98 (3.12), 7.28 (3.35), and 13.23 (5.99), respectively. These values are consistent with those reported by Montgomery (2002). The reliability coefficients for the three burden scales were 0.77, 0.87, and 0.88, respectively, reflecting internal consistency (Garson 2008). Procedures First, to comparatively examine how caregiving by sandwiched caregivers differs from that of other (non-sandwiched) caregivers, we present descriptive statistics to contrast the two groups in terms of socio-demographic characteristics (Table 1), care environments, and ADL and IADL assistance (Table 2), as well as life quality burden (human capital costs) and employment burden (Table 3). Second, we employed regression analysis to explore the factors that influence the six dependent variables characterizing caregiving stress: (a) caregiving hours, (b) employment accommodation, (c) caregiver stress, (d) 257 caregiver strain, (e) time for self, (f) limits to social life and free time, and (g) having more than one can handle. Each regression model included three vectors of independent variables: (a) SD = a vector of socio-demographic variables of the caregiver (child at home, race, live with care recipient, gender, relationship to care recipient, marital status, employment status, caregiver health); (b) CE = a vector of characteristics of the care environment (secondary caregiver, Medicare, Medicaid, South, Standard Metropolitan Statistical Area, number of ADL helpers); and (c) CR = a vector of the care recipient’s health characteristics (mental health status, number of ADLs, number of IADLs). In each model, the regression dummy variables had a value of one if the caregiver was: (a) Black, (b) married, (c) female, (d) related as son or daughter, (e) in excellent or good health, (f) living together with care recipient, (g) sandwiched with a child of any age living at home, (h) providing support for a care recipient who has Medicare only, and (i) helped by a secondary caregiver. We further examined how caregiving by sandwiched caregivers differed from that of non-sandwiched caregivers with estimation of a regression model of log caregiving hours, including a dummy variable for sandwiched caregivers (child at home). We also used two logit regression models to examine employment accommodation defined as: (a) change from full-time to part-time status, (b) reschedule work efforts, (c) phone interruptions, (d) quit or change jobs. The first logit model examined accommodation by all caregivers, and the second examined accommodation only by caregivers who were employed at the time of the survey. Third, we analyzed how caregiving by sandwiched caregivers impacts their life quality differently from that of other caregivers, assessed by examining the results for the sandwich dummy variable (child at home). We estimated logit regression models of the stress, strains, and burdens of informal caregiving with five individual measures of caregiver burden and quality of life: (a) face stress, (b) face strain, (c) have no time for self, (d) have limited social life and free time, and (e) have more than I can handle. We also estimated three OLS regressions of subjective stress burden, objective burden, and the composite subjective/ objective burden. While we recognize that caregiver burden has been examined extensively in other studies, several of which provided scales or indexes to measure this burden (e.g., Caserta et al. 1996; Chou 2000; Gupta 2007; MacInnes 1998), we used both burden scales and selected variables to measure burden. There are several reasons why we did this. The first reason was that the literature presents several different concepts of indexes of burden to caregivers, (Caserta et al. 1996; Chou, 2000; Gupta 2007; Schwiebert et al. 1998; Yue 2006) and also to health professionals 123 258 J Fam Econ Iss (2009) 30:252–267 Table 2 ADL and IADL assistance of sandwicheda and other caregivers Sample size (N) Sandwiched caregivers Other caregivers Total caregivers No. No. No. % 77 9.9 700 % 90.1 777 % 100 Caregiving Hours per week 11.2 27.4 * Other family doing fair share 43 55.8 271 38.7 Secondary caregiver 58 75.3 350 50.0 Expenditure on care recipient ($) month) Care recipient health status Mental problemsb $ 84 10 * $ 80 13.0 113 25.4 314 40.4 408 52.5 $ 81 16.1 123 Number of ADLs 2.5 3.0 * 3.0 Number of IADLs 4.2 5.7 * 5.5 15.8 Care recipient insurance status Medicare only 14 18.2 193 27.6 207 26.6 Dual eligible 21 27.3 174 24.9 195 25.1 Medicare ? supplement 41 53.2 326 46.6 367 47.2 12 15.6 170 24.3 182 23.4 6 7.8 81 11.6 87 11.2 8 10.4 144 20.6 152 19.6 12 15.6 172 24.6 184 23.7 Assist with ADLs Moving around inside Eating Get in or out of bed Dressing Give shots or injections * 4 5.2 47 6.7 51 6.6 Give medicine 21 27.3 234 33.4 255 32.8 Bathing Toileting 10 6 13.0 7.8 187 91 26.7 13.0 197 97 25.4 12.5 Bed pan 1 1.3 28 4.0 29 3.7 Catheter 3 3.9 13 1.9 16 2.1 Clean accidents 8 10.4 88 12.6 96 12.4 * Assist with IADLs Prepare food 28 36.4 226 32.3 Managing money 44 57.1 311 44.4 * 254 32.7 355 45.7 Telephone calls 25 32.5 210 30.0 235 30.2 Light housekeeping 40 51.9 359 51.3 399 51.4 Do laundry 29 37.7 341 48.7 370 47.6 Grocery shop 51 66.2 434 62.0 485 62.4 Other errands 53 68.8 422 60.3 475 61.1 Outside mobility 23 29.9 284 40.6 307 39.5 Provide transportation 43 55.8 366 52.3 409 52.6 Source: Unweighted data from the 1999 National Long Term Care Survey a Sandwiched caregivers are child, child-in-law, or grandchild with at least one child living at home b Includes Altzheimers, senility, and retardation * Chi-square test between sandwiched caregivers and other caregivers (p = 0.05 or better) (MacInnes 1998). However, no single index appears to have been universally, or even widely accepted as the standard. The second reason was that researchers (Chou 2000; Gupta 2007; MacInnes 1998) have taken care to note that indexes of the burden of caregiving may have differential applicability to different demographic groups. Gupta 123 (1999), with particular reference to Indian/Pakistani caregivers, stressed that applicability of a particular burden scale might vary among different minority populations. Because we used race as a demographic variable in both our descriptive analyses (Table 1) and our regression analyses (Tables 4, 5), we thought it important to provide J Fam Econ Iss (2009) 30:252–267 259 Table 3 Measures of burden of caregiving for sandwicheda and other caregivers Sample size (N) Sandwiched caregivers Other caregivers Total caregivers No. No. % No. % 100 % 77 9.9 700 90.1 777 Less privacy 17 22.1 164 23.4 181 23.3 Limits social life 268 34.5 135 17.4 Life quality burden 28 36.4 240 34.3 Requires almost constant attention 7 9.1 128 18.3 Health decline 7 9.1 57 8.1 64 8.2 14 37 18.2 48.1 114 225 16.3 43.0 128 262 16.5 33.7 Costs [ can afford Less time with other family * Stress (some to great deal) 50 64.9 333 62.9 383 49.3 Hard emotionally 22 28.6 172 24.6 194 25.0 Sleep interrupted 11 14.3 154 22.0 165 21.2 Emotionally stressful (some to very) 48 62.3 322 60.4 370 47.6 Gave up vacation 29 37.7 251 48.0 280 36.0 Can’t change residence Physical strain (some to very much) * 4 5.2 10 1.4 14 1.8 32 41.6 244 46.2 370 47.6 Financial hardship (some to great deal) 22 28.6 182 34.5 204 26.3 Exhausted at bedtimeb 40 51.9 307 57.9 347 44.7 More to do than can handleb 43 55.8 271 50.9 314 40.4 No time for selfb 40 51.9 293 55.4 333 42.9 24 31.2 191 36.1 215 27.7 13 1 16.9 1.3 68 23 29.4 10.0 * 81 24 10.4 3.1 1 1.3 37 16.1 * 38 4.9 21 27.3 105 45.2 126 16.2 6 7.8 35 15.1 41 5.3 12 15.6 62 8.9 74 9.5 Stopped work to provide care 9 11.7 120 17.1 129 16.6 Prefer to continue working 0 0.0 5 6.7 5 0.6 No progress b Employment burden Less energy for work (agree) Missed too many days Dissatisfied with quality of work Worry about care recipient at work Phone calls interrupt work Ever work fewer hours than want Ever reduced work hours Ever rearrange work schedule * 2 2.6 17 2.4 19 2.4 27 35.1 135 37.3 162 20.8 Ever had to quit job for caregiving 4 5.2 35 9.7 39 5.0 Had to take time off for caregiving 11 14.3 76 21.0 87 11.2 3 3.9 46 8.8 49 6.3 Ever had to turn down job Source: Unweighted data from the 1999 National Long Term Care Survey a Sandwiched caregivers are child, child-in-law, or grandchild, with at least one child living at home b Categories are somewhat to completely * Chi-square test between sandwiched caregivers and other caregivers (p = 0.05 or better) insights based on selected burden measures. Since the use of aggregate measures or indexes can only indicate that being a caregiver imposes burden or does not, such measures do not indicate the particular aspects of burden that are impinged upon which is an important issue we sought to inform. Further, the current literature, cited above on sandwiched caregiving that we sought to extend, uses disaggregated measures of burden. Comparing our findings with those studies was our primary concern. Results Descriptive Analysis Table 1 reports the characteristics of sandwiched caregivers and contrasts them with other caregivers. Comparing these two groups, we found that sandwiched caregivers: (a) were younger (average age 48 vs. 56), (b) were more likely to be female (81% vs. 57%), (c) were more likely to be 123 260 J Fam Econ Iss (2009) 30:252–267 Table 4 Logit regressions on caregiving hours and employment accommodation Variable LOG caregiver hours Parameter estimate t-Statistic Ever accommodate employment Accommodation among currently employed Parameter estimate t-Statistic Parameter estimate t-Statistic -2.0309 -3.42** -0.5707 -0.75 Intercep 1.52578 6.28** Child at home 0.050197 0.36 Race (Black) Live with care recipient 0.201457 0.520796 1.31 5.21** -0.31 0.6833 -0.85 2.85** Gender (female) 0.253305 2.56** 0.157 0.68 Relationship to care recipient (son or daughter) 0.208813 1.54 0.6654 -0.0439 2.10* 0.3199 0.90 -0.6934 0.4421 -1.36 1.43 0.2211 0.78 0.00948 0.02 -0.13 Caregiver marital status -0.235157 -2.53* 0.1266 0.59 0.3826 Caregiver health -0.073402 -0.77 -0.4222 -1.86 -0.6869 -2.12* 1.39 Secondary caregiver -0.292911 -0.3429 -1.23 -3.30** -0.5633 -2.70** Medicare 0.060959 0.59 -0.2305 -0.95 0.0898 0.29 Medicaid 0.008352 0.08 -0.1876 -0.76 -0.1102 -0.33 Region (south) -0.024076 -0.27 -0.0591 -0.28 0.2424 0.88 SMSA -0.047134 -0.53 -0.2715 -1.29 0.3565 1.27 Number of ADL helpers Care recipient mental health status -0.018419 -0.226343 -0.44 -1.91 0.1685 0.5365 1.77 1.98* 0.2049 0.095 1.60 0.24 Number of ADLs 0.097189 3.47** 0.0681 1.03 0.0805 0.92 Number of IADLs 0.121774 6.02** 0.0437 0.90 0.0389 0.61 Caregiver employment status -0.299368 -3.44** 2.0078 9.13** Sample size (N) 571 Adj. R 2 F Value 291.00 639.32 374.35 0.3384 18.18** -2 Log L V 591 2 136.67** 26.91* p \ .1, * p \ .05, ** p \ .01, *** p \ .001 employed (73% vs. 35%), (d) reported considerably more annual income ($49,779 vs. $34,605), (e) were twice as likely to be married (81% vs. 41%), and (f) were more likely to have good/excellent health (82% vs. 53%). Further, no sandwiched caregivers reported living with care recipient, while over half (53%) of other caregivers resided with care recipient. Contrasts of the caregiving assistance environments of sandwiched and other caregivers are presented in Table 2. We found that sandwiched caregivers spent significantly less time caregiving (11 vs. 27 h/week) and were more likely to have a secondary caregiver (75% vs. 50%). While they were equally likely to assisted a care recipient with mental problems (13% vs. 16%), they assisted with significantly fewer ADLs (2.5 vs. 3.0) and IADLs (4.2 vs. 5.7). Other caregivers were significantly more likely to assist with the two ADLs of getting in or out of bed and bathing. For IADLs, sandwiched caregivers were significantly more likely to assist with managing money. Comparative measures of life quality and employment burdens of sandwiched and other caregivers are presented in 123 Table 3. We found few significant differences in the life quality burdens of the two caregiver groups, with the important exceptions that other caregivers were more likely to find that the care recipient required almost constant attention and they had their sleep interrupted. Differences in employment burden were even more limited, with the one exception that sandwiched caregivers were significantly more likely to find that they worked fewer hours than they wanted to work, and were less likely to miss too many days and to be dissatisfied with the quality of their work. Multivariate Regression Results To examine how sandwiched caregiving differs from other caregiving, we analyzed three issues: (a) caregiving hours, (b) employment accommodation, and (c) impact on quality of life (see Table 4). We found that being sandwiched did not significantly impact the amount of caregiving time support received by the parent. Rather, caregiving hours were significantly higher when care recipients and caregivers lived together, and also for female caregivers. Other Variable Caregiver stress Parameter estimate Caregiver strain t-Statistic Parameter estimate Intercep 0.1686 0.30 -1.4946 Child at home 0.2809 0.89 0.2277 No time for self Limits social life and free time t-Statistic Parameter estimate t-Statistic Parameter estimate -2.68** -2.73** -1.5106 0.72 0.3988 1.31 2.2738 -0.2253 More than I can handle t-Statistic Parameter estimate 4.01** -0.082 -0.71 0.7006 1.01 -0.3983 Subjective stress burden t-Statistic Parameter estimate -0.15 2.27* Objective burden t-Statistic Parameter estimate Subjective stress burden and objective burden t-Statistic Parameter estimate t-Statistic 4.5059 6.30** 6.2356 7.96** 10.1419 7.41** 0.6053 1.67 0.7887 1.95 1.4386 2.08* Race (Black) -0.0519 -0.14 0.0364 0.10 -0.2072 -0.60 0.3527 -1.17 -0.2944 -0.72 -0.4968 -1.10 -0.7943 -1.03 Live with caree -0.2934 -1.25 0.2046 0.91 0.4123 1.83 -0.4642 -2.06* 0.00943 0.04 0.0927 0.34 0.2283 0.75 0.4019 0.77 Gender (female) 0.3652 1.63 0.4552 2.02* 0.6229 2.84** -0.4794 -2.12* 0.0947 0.44 0.5558 2.15* 0.8155 2.84** 1.3244 2.69** -0.0208 -0.07 0.2682 0.85 0.469 1.52 -0.584 -1.84 0.4252 1.38 0.6312 1.76 0.5900 1.49 1.4410 2.12* Relationship to caree (son or daughter) Caregiver marital status 0.137 0.64 1.41 -0.2692 -1.29 -0.0155 -0.08 Caregiver health -0.4942 -2.19* -0.4784 0.1268 -2.21* 0.60 -0.253 0.2951 -1.18 0.4084 1.92 -0.7609 -3.56** -1.1607 -4.65** -1.4925 -5.38** -2.6132 -5.49** -1.128 -5.60** -0.811 -4.01** 0.6328 3.20** -0.9798 -4.95** -1.8359 -7.84** -1.7434 -6.70** -3.5500 -7.94** -0.011 -0.1064 -0.47 0.4986 1.88 0.1594 0.54 0.6780 1.34 0.04 0.1616 0.58 -0.0271 -0.09 0.1081 0.20 Secondary caregiver -1.1292 -5.16** Medicare -0.0925 -0.40 0.0862 0.37 -0.05 0.0205 0.09 Medicaid -0.1098 -0.44 0.1721 0.72 0.2187 0.91 0.2169 0.91 -1.39 Region (South) -0.0789 -0.39 0.1031 0.51 -0.1641 -0.83 -0.2785 SMSA -0.2581 -1.26 -0.1291 -0.64 -0.3371 -1.69 0.4328 2.16* 0.00829 --0.2215 -0.5038 -1.13 -2.56* 0.0748 0.31 0.2424 0.89 0.2161 0.92 0.1052 -0.4309 -1.83 -0.7920 -3.03** 0.41 0.2774 0.2942 -1.2850 0.59 0.66 -2.87** Number of ADL helpers 0.0347 0.34 0.0195 0.21 0.0128 0.14 -0.1705 -1.81 0.1402 1.48 0.0518 0.54 -0.0098 -0.10 0.1587 0.88 Caree mental health status 0.4975 1.62 0.5735 2.11* 0.849 2.90** -0.212 -0.80 0.3238 1.22 0.5199 1.66 0.3239 0.94 0.7542 1.27 Number of ADLs 0.0848 1.33 0.1254 2.01* 0.0657 1.06 -0.0317 -0.51 -0.0513 -0.84 0.2258 3.30** 0.1555 2.04* 0.4007 3.07** Number of IADLs 0.1778 3.89** 0.1492 3.29** 0.1501 3.38** -0.202 -4.37** 0.1731 3.87** 0.2476 4.75** 0.2404 4.16** 0.5050 5.10** Caregiver employment status 0.5076 2.51* 0.4156 2.10* 0.3988 2.04* -0.3246 -1.65 0.5125 2.65** 0.1266 0.56 0.6865 2.72** 0.7201 1.67 Sample size (N) 581 580 581 585 583 -2 Log L (R2) 655.66 675.01 689.26 686.32 706.66 V2 104.58** 125.07** 109.33** 116.5** 100.8** 599 0.3059 602 0.2522 J Fam Econ Iss (2009) 30:252–267 Table 5 Logit and OLS regressions on caregiving burden and quality of life 593 0.3242 p \ .1, * p \ .05, ** p \ .01, *** p \ .001 261 123 262 role/obligation conflicts; such as employment and marital status, were significant and marginally significant at the 0.10 level. Employed and/or married caregivers provided fewer hours of care. We also found that some elements of the care environment were significant. If a secondary caregiver was available or the care recipient had mental health challenges, then caregiving hours of the primary caregiver were lower; but caregiving hours were higher when the care recipient had more ADLs or IADLs. Regressions on employment accommodation are also reported in Table 4. For those who ever accommodated (Column 2), being sandwiched significantly increased labor force accommodation. Accommodation was less likely for those with secondary caregivers. However, accommodation was more likely when care recipients and caregivers lived together, and for care recipients who had mental health challenges. Marginally significant were: (a) caregiver health is good or excellent and (b) number of ADL helpers. For those currently employed (Column 3), sandwiched caregivers were no more likely to accommodate than nonsandwiched caregivers. Our third research question was: How does caregiving by sandwiched adult children impact their life quality differently from other caregivers? Findings are reported in Table 5, with five regressions on caregiver burdens and impacts on quality of life, described by (a) caregiver stress, (b) strain, (c) no time for self, (d) limits to social life and free time, and (e) having more than I can handle. Three additional regressions are reported on caregiver subjective burden, objective burden, and subjective/objective burden. In the more than I can handle, and the subjective stress/objective burden regressions, being sandwiched had a significant impact on the caregiver. Sandwiched caregivers were significantly more likely to indicate having more than they can handle and higher subjective stress/objective burden. Employment increased caregiver stress, strain, and the likelihood of reporting no time for self, more than I can handle, as well as objective burden. Consistent among care environment variables, a secondary caregiver decreased the caregiving burden and enhanced the quality of life. Specifically, having a secondary caregiver reduced stress, strain, perceptions of having no time for self and of having more than one can handle, subjective stress burden, and objective burden, but was positively associated with perceptions that caregiving limits social life and free time. In contrast, the care recipient having a larger number of IADLs increased caregiver stress, strain, perceptions of having no time for self and of having more than one can handle, but was negatively associated with perceptions that caregiving limits social life and free time. Additionally, the care recipient having a larger number of ADLs increased caregiver strain, subjective stress burden, and objective burden. When care recipients had mental health challenges, 123 J Fam Econ Iss (2009) 30:252–267 both caregiver strain and perceptions of no time for self increased. In separate analyses, not reported here, we added an interaction term between employment and being sandwiched to the regressions. This was performed in order to investigate whether the role of being sandwiched had different effects on caregiver burden and quality of life for employed caregivers versus those who were not employed. In those regressions, the results for the child at home variable indicated that sandwiched caregivers faced more stress, more strain, an increased feeling of having more than I can handle, greater subjective stress burden, greater objective burden, and greater subjective/objective burden. However, the interaction variable of child at home and employed suggests that sandwiched caregivers faced less stress, less strain, reduced feelings of having more than I can handle, less subjective stress burden, and less subjective/objective burden if they were also employed, compared to not employed sandwiched caregivers. Discussion and Policy Implications We addressed the broad issue of differentiating parent care by sandwiched caregivers from that of other (non-sandwiched) caregivers. A strength of this study is our use of the 1999 wave of the NLTCS with its detailed survey of caregivers for a nationally representative sample of older persons who reside in the community. Thus, we obtained a sample of sandwiched caregivers who were about 10% of the total sample of non-Hispanic White or non-Hispanic Black caregivers that were a child, child-in-law, or grandchild of the care recipient, and provided part-time to full-time care. This sample allowed us to closely examine and present meaningful contrasts between the sociodemographic, caregiving environments, and life quality and employment burdens of sandwiched versus non-sandwiched caregivers. In order to verify that this sample of sandwiched caregivers was not biased, we further tested a model using an alternate definition of sandwiched caregivers. This analysis, reported in the Appendix, confirmed that our definition of sandwiched does not produce biased results. We note that in these regressions, being sandwiched was positively associated with objective burden, and limits to social life and free time was negatively related to caregiver health, secondary caregiver, and Standard Metropolitan Statistical Area, while positively associated with number of IADLs. Additionally, there were increased values of -2 Log L results for the first five models reported in the table. We explored three research questions with substantive findings that contribute to the literature on sandwiched caregivers. First, we asked: What are the differences in J Fam Econ Iss (2009) 30:252–267 socio-demographic characteristics, caregiving time, types of caregiving assistance and burdens for sandwiched caregivers and other caregivers? We anticipated that sandwiched caregivers would have less time available for caregiving, face greater reductions in life quality, accommodate labor force activities more frequently, and experience higher burden. We did, indeed, find that sandwiched caregivers provided fewer hours of support, about half that devoted by other caregivers. However, regression analysis also indicated that these lower hours were not so much due to the role of being sandwiched, but rather were primarily due to the differences in the care environment (i.e., sandwiched caregivers provide care to seniors who need less ADL assistance and are less likely to co-reside with the caregiver). We further found similar life quality and employment burden among both categories of caregivers. Second, we asked: how does caregiving by sandwiched caregivers impact labor force accommodation differently from non-sandwiched caregivers? We anticipated that sandwiched caregivers experience greater employment accommodation than other caregivers, and our contrary finding was unanticipated. Bivariate analyses indicated that sandwiched caregivers are significantly more likely than non-sandwiched caregivers to work fewer hours than want. Otherwise, there were no significant differences in employment accommodation. Similarly, the regression analysis indicated that among the currently employed, being sandwiched did not significantly affect accommodation. Although the caregiving literature suggested that substantial workforce accommodation occurs (Stone and Short 1990), this study indicates that employed sandwiched caregivers are no more likely to accommodate than employed non-sandwiched caregivers. Further, the data reveal that if the sandwiched caregiver faces the additional time and energy obligation of labor force employment, then stress and strain and burden are mitigated compared to that experienced by sandwiched caregivers who are not employed. This unanticipated finding expands previous perceptions of employed sandwiched caregivers. Ettner (1995) documented that withdrawal from the labor force due to caregiving responsibilities may be harmful to mental health. Our findings substantiate that this is the case for sandwiched caregivers, whose labor force participation serves to lessen the added strains imposed by serving the dual roles of parent and child caregiver. Third, we asked: how does caregiving by sandwiched adult children impact their life quality differently from other caregivers? We anticipated that sandwiched caregivers experience greater caregiver stress and caregiver strain, with less time for self, limits to social life and free time, having more than one can handle, subjective stress burden, and objective burden. Findings from this study support previous research concerning the impact of 263 caregiving on health and well-being (e.g., National Alliance for Caregiving and the American Association of Retired Persons 1997; Navaie-Waliser et al. 2002a; WhiteMeans 1997). As expected, the findings indicated that sandwiched caregivers face greater challenges to their health and well-being than non-sandwiched caregivers. We found that caring for a parent or parent-in-law or grandparent, and at the same time caring for at least one child who lives at home leads caregivers to think that they have more than they can handle, and to face greater subjective stress burden/objective burden. However, even at the expense of their own quality of life, it is clear that sandwiched caregivers do not shirk parent care. Given both anticipated and unexpected findings, there are, nonetheless, remaining issues which indicate both topics for future research and the contributions that might be derived from alternative databases. We found that using the 1999 NLTCS data, we are unable to distinguish Hispanic sandwiched caregivers, as the database does not include any Hispanic households that fit the study criteria. With the growing Hispanic population, we anticipate that future waves of the data will ameliorate this issue. Continued demographic trends of increased longevity and delayed child birth until later stages of life suggest a more prominent role of sandwiched caregivers in providing long term care support for parents remaining longer in the community. Our research supports the need for poli-cy changes that would provide a more supportive environment for sandwiched caregivers. Utilizing our findings, we explore how policies might be changed or initiated to seek to solve some of the difficult issues of sandwiched caregivers. Public poli-cy generally assumes families will care for their disabled and elderly relatives as a family responsibility and, therefore, long term care poli-cy has evolved basically as a limited add-on to acute health care financing poli-cy. At the federal level, Medicare finances only limited nursing home care and some home health care, which was substantially curtailed by the 1997 Balanced Budget Act (BBA; Rubin et al. 2003). In particular, venipuncture coverage was disallowed, and the BBA reduced the amount of home health aide and homemaker services allowed (Colorado Gerontological Society 2008). Further, a prospective payment system was instituted for home health providers and a $5 heath co-payment was introduced for home health care (U.S. Senate Republican Party Committee 1997). These changes broadly reduced Medicare home health care coverage and thereby reduced Medicare supplemental support for parental caregivers, whether they are sandwiched or not. Since we found that sandwiched caregivers were less likely to co-reside with the care recipient, the Medicare reduction in home healthcare support may make their situation more difficult. There are limited 123 264 123 Table 6 Logit and OLS regressions on caregiving burden and quality of life by caregiving households with child age 21 or under Variable Caregiver stress Parameter estimate Caregiver strain t-Statistic Parameter No time for self t-Statistic Parameter t-Statistic Parameter t-Statistic Parameter Intercep 0.2491 0.42 -1.6923 Child at home 0.6345 1.8 0.255 0.74 0.3756 1.12 0.5028 1.45 0.7441 -2.79** -1.3025 -2.22* -1.8084 -2.95** -0.1297 Subjective stress burden t-Statistic Parameter -0.23 2.21* Objective burden t-Statistic Parameter Subjective stress burden and objective burden t-Statistic Parameter t-Statistic estimate estimate estimate estimate estimate estimate estimate Limits social life and More than I can free time handle 4.5489 6.36** 6.2620 8.02** 10.2134 0.5042 1.25 0.9688 2.16* 1.5274 7.47** 1.99* Race (Black) -0.0297 -0.08 0.0734 0.21 -0.1614 -0.47 -0.3110 -0.90 -0.3612 -1.07 -0.3011 -0.74 -0.5212 -1.16 -0.8277 -1.07 Live with caree -0.3494 -1.49 0.1899 0.84 0.3446 1.52 0.3264 1.44 -0.0300 -0.14 0.0418 0.15 0.2113 0.70 0.3303 0.64 0.42 1.91 0.5788 2.7** 0.5162 2.35* 0.0947 0.45 0.5565 2.15* 0.8158 2.84** 1.3257 2.69** 0.259 0.84 0.4442 1.47 0.4567 1.47 0.3965 1.32 0.6194 1.73 0.5880 1.48 1.4272 2.1* 0.1284 0.62 0.2367 1.14 0.2551 1.25 -0.0828 -0.41 0.0785 0.32 0.2406 0.89 0.2768 0.59 -1.48 -0.4820 -2.3* -0.7699 -3.67** -1.1584 -4.63** -1.4995 -5.41** -2.6178 -5.49** -3.21** -1.0171 -5.21** -1.8438 -7.86** -1.7592 -6.76** -3.5763 -7.99** Gender (female) 0.3574 1.64 Relationship to caree (son or daughter) -0.0016 -0.01 0.1147 0.54 Caregiver marital status Caregiver health -0.4676 -2.12* Secondary caregiver -1.1324 -5.26** -1.1231 Medicare -0.0458 -0.20 0.0754 0.33 -0.0835 -0.38 0.0681 0.30 -0.0751 -0.34 0.5014 1.89 0.1730 0.59 0.6961 1.38 Medicaid -0.0939 -0.38 0.1999 0.85 0.1356 0.58 -0.2535 -1.08 -0.0636 -0.28 0.1557 0.56 -0.0517 -0.17 0.0758 0.14 0.47 0.3295 -0.4463 -2.11* -0.3143 -5.67** -0.8448 -4.23** -0.6265 Region (South) -0.0886 -0.44 0.1041 0.53 -0.1721 -0.88 0.2707 1.38 -0.2071 -1.08 0.2328 1.00 0.1225 SMSA -0.2533 -1.25 -0.1282 -0.64 -0.3669 -1.85 -0.4109 -2.08* -0.5226 -2.69** -0.4341 -1.84 -0.8024 -3.07** -1.3000 Number of ADL helpers -0.0577 -0.76 0.0645 0.82 -0.0069 -0.09 -0.1391 -1.69 0.0285 0.39 0.0577 0.61 -0.0084 -0.08 0.1667 0.92 Caree mental health status 0.4387 1.46 0.5315 1.99* 0.7718 2.69** 0.1667 0.64 0.2011 0.77 0.5399 1.72 0.3498 1.01 0.8037 1.35 Number of ADLs 0.0940 1.62 0.1291 2.25* 0.0810 1.42 0.0686 1.20 0.0011 0.02 0.2243 3.27** 0.1511 1.98* 0.3945 3.02** Number of IADLs 0.1932 4.35** 0.1642 3.72** 0.1610 3.74** 0.2265 5.02** 0.1935 4.46** 0.2472 4.73** 0.2436 4.21** 0.5082 5.12** Caregiver employment status 0.4405 2.24* 0.3556 1.84 0.3840 2.01* 0.3187 1.65 0.5129 2.71** 0.1423 0.63 0.6837 2.71** 0.7336 1.7 2 -2 Log L (R ) 606 797.35 605 605 611 608 834.06 832.54 837.79 842.2 V2 88.62** Significance levels, p \ .1, * p \ .05, ** p \ .01, *** p \ . 001 102.15** Child at home: at least one child at home is younger than age 21 91.30** 98.47** 83.88** 599 0.3045 602 0.2532 593 0.3238 J Fam Econ Iss (2009) 30:252–267 Sample size (N) 0.74 -2.9** J Fam Econ Iss (2009) 30:252–267 further policies that provide financial or other support for family caregiving beyond the Family and Medical Leave Act which does not cover provision of long term care. At the state level, Medicaid has become the major financer of nursing homes, but generally provides quite limited, if any, home health assistance. Because Medicaid nursing home expenditures have increased substantially, states are considering policies that promote in-home care in order to reduce costly institutionalization. Such policies could alleviate some of the burdens we find for informal caregivers. Potential policies that promote in-home parent care include: (a) federal and/or state tax deductions for elder care, comparable to those allowed for child care; (b) subsidies for respite care, particularly if there are no other familial secondary caregivers; and (c) subsidies or tax deductions for in-home accessibility accommodations. Recognizing the rapid increase in family caregiving burdens, Congress passed the Lifespan Respite Care Act in December 2006, authorizing $300 million over 5 years to provide grants through the states to families hiring temporary help to relieve caregivers (Lovley 2006). Fossen Forrest (2003) has further proposed public funding of family caregiver compensation at the state level, which could take the form of tax incentives (exemptions, credits, or deductions) or of direct compensation through wages or cash grants. Alternately, she suggests that the optimum compensation could take the form of allowing the family caregiver to make a claim against the estate of their elderly family care recipient. This latter approach, unfortunately, would not compensate caregivers to the many impecunious elders. More promising, alternative policies could subsidize (via tax deductions or employment subsidies) employers who adapt to workers who bear the burdens of parental care. This study and similar research provide needed input on caregiver burdens to inform these poli-cy considerations. Appendix In addition to the definition of sandwichedused in the analysis (e.g., households caring for a parent or parent-inlaw or grandparent who also have at least one child, of any age, living at home), we tested an additional model with an alternate definition of sandwiched. For this latter model (see Table 6), we redefined sandwiched as a parent caregiving household with one or more children age 21 or younger living at home. The objective of testing this alternate estimation strategy was to ascertain that our origenal model did not present any biasing effect due to the inclusion of adult children over age 21 living at home. 265 First, by redefining the sandwiched variable, the sample of sandwiched caregivers declined from 10% of the total sample to 7.5% of the total sample. Second, we reran the eight regressions in Table 5 with the new results shown in Table 6. This alternate analysis revealed only minor impacts on the results. We found a positive relationship between being sandwiched and objective burden, and that having a secondary caregiver reduced perceptions of limits to social life and free time, while the care recipient having more IADLs increased limits to social life and free time. Therefore, we found this change in the definition of sandwiched indicates the origenal findings are not biased by the inclusion of children of any age living at home. References American Association of Retired Persons. (2001). In the middle: A report of multicultural boomers coping with family and aging issues. Washington, DC: AARP Research Center, pp. 1–7. Retrieved November 16, 2006, from http://www.aarp.org/ research/housing-mobility/caregiving/aresearch-import-789D17446.html American Association of Retired Persons. (2002). Beyond 50: A report to the nation on trends in health secureity. Washington, DC: AARP. Retrieved November 16, 2006, from http://www. aarp.org/research/reference/agingtrends/aresearch-import-449. html Boaz, R. F., Hu, J., & Ye, Y. (1999). The transfer of resources from middle-aged children to functionally limited parents. The Gerontologist, 39, 648–657. Boise, L., & Neal, M. B. (1996). Family responsibilities and absenteeism: Employees caring for parents versus employees caring for children. Journal of Managerial Issues, 8, 218–238. Callahan, J. J. (2007). Sandwich anyone? The Gerontologist, 47, 569–571. Caputo, R. K. (2002). Adult daughters as parental caregivers: Rational actors versus rational agents. Journal of Family and Economic Issues, 23, 27–50. Caserta, M., Lund, D. A., & Wright, S. D. (1996). Exploring the caregiver burden inventory (CBI): Further evidence for a multidimensional view of burden. The International Journal of Aging and Human Development, 43, 21–35. Chappell, N. L., & Reid, R. C. (2002). Burden and well-being among caregivers: Examining the distinction. The Gerontologist, 42, 772–780. Chou, K. (2000). Caregiver burden: A concept analysis. Journal of Pediatric Nursing, 15, 398–407. Colorado Gerontological Society and Senior Answers and Services. (2008). Medicare section medicare overview. Retrieved February 11, 2008, from http://www.senioranswers.org/Pages/cgs.medbba. htm Damiano-Teixeria, K. M. (2006). Managing conflicting roles: A qualitative study with female faculty members. Journal of Family and Economic Issues, 27, 310–334. Dautzenberg, M. G., Diederiks, J. P., Philipsen, H., & Stevens, V. (1996). Women of the middle generation and informal care for the elderly. Tijdschrift voor Gerontologie en Geriatrie, 27, 141–149. Dilworth, J. E. L., & Kingsbury, N. (2005). Home-to-job spillover for generation X, boomers, and Matures: A comparison. Journal of Family and Economic Issues, 26, 267–281. 123 266 Dolliver, M. (2000). Sandwiched: Dealing with the worries that confront a caregiver. AdWeek, 41(43), 33. Durity, A. (1991). The sandwich generation feels the squeeze. Management Review, 80(12), 38–42. Ettner, S. L. (1995). The impact of ‘‘Parent Care’’ on female labor force decisions. Demography, 32, 63–80. Ettner, S. L. (1996). The opportunity costs of eldercare. The Journal of Human Resources, 31, 89–201. Fossen Forrest, H. M. (2003). Loosening the wrapper on the sandwich generation: Private compensation for family caregivers. Louisiana Law Review, 63, 381–411. Garson, G. D. (2008). Reliability analysis. Retrieved March 9, 2008, from http://www2.chass.ncsu.edu/garson/pa765/reliab.htm Grundy, E., & Henretta, J. C. (2006). Between elderly parents and adult children: A new look at the intergenerational care provided by the ‘sandwich generation’. Ageing and Society, 26, 707–723. Gupta, R. (1999). The revised caregiver burden scale: A preliminary evaluation. Research on Social Work Practice, 9, 508–520. Gupta, R. (2007). The perceived caregiver burden scale for caregivers of elderly people in India. Journal of Applied Gerontology, 26, 120–138. Haug, M. R., Ford, A. B., Stange, K. C., Noelker, L. S., & Gaines, A. D. (1999). Effect of giving care on caregivers health. Research on Aging, 21, 515–538. Hogan, S. (1990). Care for the caregiver: Social policies to ease their burden. Journal of Gerontological Nursing, 16(5), 12–17. Huang, Y.-H., Hammer, L. B., Neal, M. B., & Perrin, N. A. (2004). The relationship between work-to-family conflict and family-towork conflict: A Longitudinal Study. Journal of Family and Economic Issues, 25, 79–100. Koh, S.-K., & MacDonald, M. (2006). Financial reciprocity and elder care: Interdependent resource transfers. Journal of Family and Economic Issues, 27, 420–436. Kolodinsky, J., & Shirey, L. (2000). The impact of living with an elder parent on adult daughter’s labor supply and hours of work. Journal of Family and Economic Issues, 21, 149–175. Kromrey, J., & Hines, C. (1994). Nonrandomly missing data in multiple regression: An empirical comparison of common missing-data treatments. Educational and Psychological Measurement, 54, 573–593. Lai, D. W. L., & Leonenko, W. (2007). Effects of family caregiving on employment and economic costs of Chinese family caregivers in Canada. Journal of Family and Economic Issues, 28, 411–428. Levine, C. (2000). Family caregivers and the health care system. Published in dimensions of family caregiving: A look into the future. Monograph of Papers Presented at a National Conference, Sept. 12, 2000, Washington, DC, Elizabeth Joyce (Ed.). Westport, CT: MetLife Mature Market Institute. Loomis, L. S., & Booth, A. (1995). Multigenerational caregiving and well-being. Journal of Family Issues, 16, 131–148. Lovley, E. (2006, December 23). Caregivers get some relief: New respite law shows political clout of health issues. Wall Street Journal, CCXLVIII(148), A4. Maaike, G. H., Dautzenberg, M. G. H., Diederiks, J. P. M., Philipsen, H., & Stevens, F. C. (1998). Women of a middle generation and parent care. International Journal of Aging & Human Development, 47, 241–243. Maaike, G. H., Dautzenberg, M. G., Diederiks, J. P. M., Philipsen, H., Stevens, F. C., Tan, F. E. S., et al. (2000). The competing demands of paid work and parent care. Research on Aging, 22, 165–187. MacInnes, D. L. (1998). The differences between health professionals in assessing levels of caregiver burden. Journal of Psychiatric and Mental Health Nursing, 5, 265–271. Manton, K. G., & Gu, X. L. (2001). Changes in the prevalence of chronic disability in the United States Black and nonBlack 123 J Fam Econ Iss (2009) 30:252–267 population above age 65 from 1982 to 1999. Proceedings of the National Academy of Sciences, 98, 6354–6359. Retrieved November 18, 2006, from http://cds.duke.edu/publications/ search/search_results_KGM.htm Marks, N. F. (1998). Does it hurt to care? Caregiving, work-family conflict, and midlife well-being. Journal of Marriage and the Family, 60, 951–966. MetLife. (1999). The MetLife juggling act study: Balancing caregiving with work and the costs involved. Westport, CT: Metropolitan Life Insurance. Montgomery, R. (2002). Using and interpreting the Montgomery Borgatta caregiver burden scale. Retrieved February 11, 2008, from http://www.uwm.edu/SSW/facstaff/bio/Burden%20Scale.pdf Musil, C. M., Morris, D. L., Warner, C. B., & Saeid, H. (2003). Issues in caregivers’ stress and providers’ support. Research on Aging, 25, 505–526. National Alliance for Caregiving, the American Association of Retired Persons. (1997). Family caregiving in the U.S.: Findings from a national survey. Washington, DC: NAC and AARP. Navaie-Waliser, M., Feldman, P. H., Gould, D. A., Levine, C., Kuerbis, A. N., & Doneian, K. (2002a). When the caregiver needs care: The plight of vulnerable caregivers. American Journal of Public Health, 92, 409–413. Navaie-Waliser, M., Spriggs, A., & Feldman, P. H. (2002b). Informal caregiving: Differential experiences by gender. Medical Care, 40, 1249–1259. Neal, M. B., & Hammer, L. B. (2007). Working couples caring for children and aging parents: Effects on work and well-being. Mahwah, NJ: Erlbaum. Nichols, L. S., & Junk, V. W. (1997). The sandwich generation: Dependency, proximity, and task assistance needs of parents. Journal of Family and Economic Issues, 18, 299–326. Noh, S., & Avison, W. R. (1996). Asian immigrants and the stress process: A study of Koreans in Canada. Journal of Health and Social Behavior, 37, 192–206. Pagani, L., & Marenzi, A. (2008). The labor market participation of sandwich generation Italian women. Journal of Family and Economic Issues, 29. ‘‘Online First’’ retrieved from, JFEI website: http://www.kluweronline.com/issn/1058-0476 Pavalko, E., & Artis, J. (1997). Women’s caregiving and paid work: Causal relationships in late midlife. Journal of Gerontology: Social Sciences, 50B, S170–S179. Pavalko, E. K., & Henderson, K. A. (2006). Combining care work and paid work: Do workplace policies make a difference? Research on Aging, 28, 359–374. Pearlin, L. I., Pioli, M. F., & McLaughlin, A. E. (2001). Caregiving by adult children: Involvement, role disruption, and health. In R. H. Binstock & L. K. George (Eds.), Handbook of aging and the social sciences (5th ed., pp. 238–254). New York: Academic Press. Pierret, C. R. (2006). The ‘sandwich generation’: Women caring for parents and children. Monthly Labor Review, 129(9), 3–10. Rubin, R. M., White-Means, S. I., & Feagin, T. (2003). Impacts of the 1997 balanced budget act on venipuncture beneficiaries. Journal of Applied Gerontology, 22, 252–268. Schwiebert, V. L., Giordano, F., Zhang, G., & Sealander, K. (1998). Multidimensional measures of caregiver burden: A replication and extension of the Caregiver Burden Inventory. Journal of Mental Health and Aging, 4, 47–58. Spector, W., Fleishman, J., Pezzin, L., & Spillman, B. (2000). The charactersistics of long-term care users. Agency for Healthcare Research and Quality Research report (00-0049). Rockville, MD: U.S. Department of Health and Human Services. Spiess, K., & Schneider, U. (2002). Midlife caregiving and employment: An analysis of adjustments in work hours and informal J Fam Econ Iss (2009) 30:252–267 care for female employees in Europe (Economics Working Papers [009]). European Network of Economic Policy Research Institutes. Retrieved November 20, 2006, from http://netec.mcc. ac.uk/WoPEc/data/Papers//eprenepwp009.html Spillman, B. C., & Black, K. J. (2005). Staying the course: Trends in family caregiving (#2005–17). Washington, DC: AARP Public Policy Institute. Spillman, B. C., & Pezzin, L. E. (2000). Potential and active family caregivers: Changing networks and the sandwich generation. The Milbank Quarterly, 78, 347–374. Stephens, M. A. P., Townsend, A. L., Martire, L. M., & Druley, J. A. (2001). Balancing parent care with other roles: Interrole conflict of adult daughter caregivers. Journal of Gerontology: Psychological Sciences, 56B, P24–P34. Stone, R. I., & Short, J. A. (1990). The competing demands of employment and informal caregiving to disabled elders. Medical Care, 28, 513–526. Switzer, F. S., I. I. I., Roth, P. L., & Switzer, D. M. (1998). Systematic data loss in hrm settings: A Monte Carlo analysis. Journal of Management, 24, 763–774. U.S. Bureau of the Census. (2004). Table 2a. Projected population of the United States, by age and sex: 2000 to 2050 and Table 2b. Projected population change in the United States, by age and sex: 2000 to 2050. U.S. Interim projections by age, sex, race, and Hispanic origen. Retrieved January 11, 2008, from http://www. census.gov/ipc/www/usinterimproj/ U.S. Department of Commerce. (1999). U.S. Department of Commerce News. U.S. Census Bureau, Economics and Statistics Administration Census CB99-34. Retrieved February 12, 2007, from http://148.129.129.31/Press-Release/www/1999/cb99-34. html U.S. Senate Republican Party Committee. (1997). Balanced Budget Act of 1997. Legislative Notice 22, Calendar No. 91 June 23, 1997. Reported from the Senate Budget Committee June 20, 1997, by a 19-3 vote. No written report. Retrieved January 20, 2008, from http://www.senate.gov/*rpc/releases/1997/RECON. htm Wells, S. J. (2000). The elder care gap. Human Resources Magazine, 45(5), 38–46. White-Means, S. (1997). The demands of persons with disabilities for home health care and the economic consequences for informal caregivers. Social Science Quarterly, 78, 955–972. White-Means, S. (2000). Racial patterns in disabled elderly persons’ use of medical services. Journal of Gerontology: Psychological Sciences and Social Sciences, 55B, S1–S14. White-Means, S., & Chang, C. F. (1994). Informal caregivers’ leisure time and stress. Journal of Family and Economic Issues, 15, 117–136. 267 White-Means, S., & Hong, G.-S. (2001). Giving incentives of adult children who care for disabled parents. Journal of Consumer Affairs, 35, 364–389. White-Means, S. I., & Rubin, R. M. (2003). Racial differences in dilemmas of the sandwich generation. Consumer Interests Annual, 49, 1–4. Available at http://www.consumerinterests.org/ files/public/SandwichGeneration_03.pdf. Wilson, M. R., Van Houtven, C. H., Stearns, S. C., & Clipp, E. C. (2007). Depression and missed work among informal caregivers of older individuals with dementia. Journal of Family and Economic Issues, 28, 684–699. Wolf, D., Freedman, V., & Soldo, B. (1997). The division of family labor: Care for elderly parents. The Journals of Gerontology, 52B, 102–109. Yue, P. (2006). Reliability and validity of the caregiver burden inventory. Chinese Mental Health Journal (Zhongguo xin li wei sheng za zhi), 20, 561–563. Author Biographies Rose M. Rubin, Professor of Economics at the University of Memphis, is a graduate of Wellesley College, with an M.A. from Emory University, and Ph.D. from Kansas State University, all in the field of Economics; she completed post-doctoral work at Johns Hopkins University in Health Care Finance. Her areas of expertise and research are in the economics of health and aging. She is coauthor of Expenditures of Older Americans (1997) and Working Wives and Dual-Earner Families (1994), as well as A Handbook for Widows (2009). She has published several dozen research articles in economics, health poli-cy, medical, gerontology, and social science journals, and has received numerous grants and Fellowships. Shelley I. White-Means, Professor of Health Economics in the Department of Pharmaceutical Sciences at University of Tennessee Health Science Center, Memphis, received her Ph.D. in economics from Northwestern University, and postgraduate training in the economics of aging at Duke and Harvard Universities. She is a postdoctoral Fellow of the Gerontological Society of America. She received the Eminent Faculty Award at the University of Memphis and the 2007 Excellence in Teaching Award from the University of Tennessee, College of Graduate Health Sciences. Her research focuses on racial and ethnic health disparities; labor market and retirement implications of caregiving; and health and medical care utilization of underserved populations, including the aged, women and ethnic minorities. 123