Gender differences among Canadian spousal caregivers at the end of life

Transcription

1 Health and Social Care in the Community (2009) 17(2), doi: /j x Blackwell Publishing Ltd Gender differences among Canadian spousal caregivers at the end of life Kevin Brazil PhD 1,2, Lehana Thabane PhD 1,3, Gary Foster PhD 4 and Michel Bédard PhD 5,6 1 Department of Clinical Epidemiology and Biostatistics & Division of Palliative Care, Faculty of Health Sciences, McMaster University, Hamilton, ON, Canada, 2 St. Joseph s Health System Research Network, Hamilton, ON, Canada, 3 Biostatistics Unit, Father Sean O Sullivan Research Centre & Centre for Evaluation of Medicines, St Joseph s Healthcare, Hamilton ON, Canada, 4 Biostatistics Unit, Father Sean O Sullivan Research Centre & Centre for Evaluation of Medicines, St Joseph s Healthcare, Hamilton ON, Canada, 5 Public Health Program, Lakehead University, Thunder Bay, ON, Canada, and 6 Division of Human Sciences, Northern Ontario School of Medicine, Thunder Bay, ON, Canada Correspondence Kevin Brazil Department of Clinical Epidemiology and Biostatistics 105 Main St., E., Level P1 Hamilton, ON Canada L8N 1G6 brazilk@mcmaster.ca Abstract The purpose of this study was to examine gender differences in spousal caregiving at the end of life. The primary research question was to determine gender differences in caregiver strain among spousal caregivers. Secondary research questions investigated included (i) the presence of gender differences among spousal caregivers in the duration of care provided; (ii) gender differences among spousal caregivers in formal service use and unmet service needs; and (iii) whether support to care recipients in activities of daily living varied according to the gender of the spousal caregiver. The study was conducted over a 2-year period ( ) in south-central Ontario, Canada. The study sample included 283 informal spousal caregivers (198 females, 85 males) each of whom were caring for a terminally ill spouse at the time they participated in a cross-sectional telephone survey. The analysis showed that females reported a significantly greater level of caregiving strain than males (t = 2.12, d.f. = 281, P = 0.035). When considering source of support in activities of daily living for the care recipient, differential assistance was noted on the basis of caregiver gender. Female caregivers had almost twice the odds of providing support in toileting-related tasks than male caregivers (odds ratio (OR) = 1.98, 95% confidence interval (CI) = , P = 0.044), while male caregivers had approximately twice the odds of providing support in mobility-related tasks (OR = 0.41, 95% CI = , P = 0.011). Care recipients who had a female caregiver had lower odds of receiving support from family and friends in tasks associated with personal care (OR = 0.17, 95% CI = , P = 0.002). To address gender differences in caregiving, a realistic home-based palliative care approach must take into account the importance of informal caregivers. Keywords: end-of-life care, gender differences, home-based palliative care, palliative care, spousal caregiving Accepted for publication 27 July 2008 Introduction In the last decade, policy-makers in industrialised countries have increasingly become interested in home-based palliative care and death at home (Senate Committee on Social Affairs 2000, Wilson et al. 2002). Several factors have encouraged the re-emergence of home death, including preference of the terminally ill and their families, recognition of the limits of medical care, improvements in home-based medical technology 2008 The Authors, Journal compilation 2008 Blackwell Publishing Ltd 159

2 K. Brazil et al. and pharmacology, and hospital restructuring with the resulting closure of hospital beds (Gott et al. 2004, Thomas et al. 2004). This growing interest in home-based palliative care poses many complex issues that must be taken into consideration. A key issue is the impact that home-based care has on informal caregivers who bear much of the responsibility in caring for the terminally ill (Stajduhar & Davies 1998, Payne et al. 1999). Caregiving at the end of life is often seen among informal caregivers as a moral responsibility, obligation or duty that they must uphold for their loved ones (Aranda & Peerson 2001). The research literature reports that informal caregivers commonly experience physical, emotional, social and financial burden throughout the caregiving process (Smith 1990, Covinsky et al. 1994, Brazil et al. 2003). It has been well documented that family caregivers often convey feelings of confinement, isolation, exhaustion, anxiety and depression (Smith 1990, Covinsky et al. 1994, Hodgson et al. 1997, Grunfeld et al. 1997, Emanuel et al. 2000, Steinhauser et al. 2000, Brazil et al. 2003, Teno et al. 2004). Informal caregivers have been described as lacking the experience and knowledge in coping with the intensity required for the care of a terminally ill individual (Sankar 1993, Stajduhar & Davies 1998, Ross et al. 2002). This lack of knowledge and inadequate experience of caregiving along with numerous overwhelming tasks required of the caregiver poses a threat to not only the health and well-being of the care recipient, but also to that of the caregiver. It is evident that the complex nature of the caregiving experience presents challenges for both the informal caregiver and the care recipient. The consequences of the policy shift from institutional to home and community care are different for women and men. Informal caregiving is mostly provided by women (Dahlberg et al. 2007), who spend more hours caring for sick relatives than men (Keating 1999). Studies have also shown that women s health is more likely to be adversely affected by caregiving than men s; women caregivers report greater burden, stress, anxiety, depression, and adverse health outcomes (Yee & Schulz 2000, Navaie-Waliser et al. 2002, Gallicchio et al. 2002, Bedard et al. 2005). While there has been extensive research on home and community care, there is a paucity of gendersensitive analysis on informal caregiving in community care for caregiving of terminally ill individuals (Morris 2001). As the experience of informal caregiving may reflect gender differences, it is important that such differences be examined and understood in the context of healthcare delivery. To meet the needs of family caregivers, it is important to acknowledge the variations between different groups of caregivers. The purpose of this study is to extend the research on gender differences in informal caregiving into the domain of palliative care. The primary research question was: Are there gender differences among spousal caregivers in caregiver strain as measured by the Robinson scale? Secondary research questions included: (i) Are there gender differences among spousal caregivers in the duration of care provided? (ii) Are there gender differences among spousal caregivers in formal service use and unmet needs? (iii) Does support to care recipients in activities of daily living vary according to the gender of the spousal caregiver? Methods Participant recruitment and procedure This report is based on data from a longitudinal study investigating care of the terminally ill. The study was located in south-central/western Ontario, Canada. To be eligible for participation in the study, the caregivers had to meet the following inclusion criteria: the caregiver spoke English, the care recipient was 50 years or older, received in-home services delivered through a Community Care Access Centre (CCAC), and was designated as palliative by the CCAC. In the province of Ontario, CCACs are publicly funded agencies designed to provide a unified service access point to community care services. CCAC case managers are responsible for client service planning, case management and act as a broker of community-based services. Case managers or visiting nurses were asked to identify all eligible participants (based on the inclusion criteria described above) using agency records over a 2- year period ( ). Agency staff contacted eligible participants, described the study to them, and obtained consent to release their name and contact number to the study investigators. An information letter was also left with the family or the investigators mailed the letter 7 days before making telephone contact. The letter informed the potential participant about the study, and advised them to expect a telephone call unless they contacted the investigators to decline participation. Participants contacted by phone were reminded of the letter that was sent to them and invited to participate in the study, providing informed consent. If a respondent agreed to participate, a convenient time was arranged for a telephone interview. Through repeated call-back procedures (up to six attempts) all potential participants were eventually contacted. This study was conducted in compliance with the Ethical Conduct for Research Involving Human Tri-Council Policy (Statement ethics/english/policy.htm). Ethics approval was obtained from the St. Joseph s Healthcare Ethics Review Board, Hamilton, Canada The Authors, Journal compilation 2008 Blackwell Publishing Ltd

3 Gender and caregiving at the end of life Instruments Interview variables included in this analysis covered care recipient and caregiver characteristics. Care recipient characteristics included sex, age and living arrangements (number of people the care recipient lived with). The functional status of the care recipient was measured using the original 10-item version of the Barthel Index (Collin et al. 1988). The index measures functional independence in personal care, toileting and mobility as reported by the informal caregiver. Each item is rated in terms of whether the individual can perform the task independently, with some assistance, or is dependent on help. An overall score is formed by adding scores in each rating. Total scores range from 0 to 20, with higher scores indicating greater independence. In activities where the care recipient was reported as needing assistance, sources of support were identified (spousal support, informal support (friends and family) and formal support (healthcare worker)). Caregiver characteristics that were collected included sex, age, education, income, self-reported health status, number of weeks caregiving and hours a week spent on caregiving. Caregivers also completed the Robinson Caregiver Strain Scale, a 13-item, yes/no response instrument that measures various domains of strain related to caregiving (e.g. physical, personal, family, financial). The total scale scores ranges from 0 to 13, with higher scores indicating greater caregiver strain. The internal consistency of the instrument has been identified in previous research as high (Cronbach s alpha = 0.86) (Robinson 1983). Construct validity of the measure is also supported by correlating total scale scores with care recipient characteristics, caregivers subjective views of the caregiving situation, and emotional well-being of caregivers (Gaugler et al. 2000). To catalogue service use, a detailed list of home support services was developed. Several case managers who were responsible for coordinating client services within the study locations reviewed the service list to ensure that all possible services available within their catchment area were included. The final list consisted of 24 services. Recording service use was modelled after the interview protocol used in the Canadian Study of Health and Aging (Canadian Study of Health & Aging Working Group 1994). If a respondent reported using a specific service within the 4 weeks prior to the interview, the interviewer completed a structured interview that recorded the perceptions of that service. After reviewing possible services, spousal caregivers were asked to identify services of which they would have liked to have had or received more of while caregiving. Analysis Socio-demographic characteristics, Barthel Index and Robinson Scale were compared by gender. Results were summarised using descriptive measures expressed as mean (standard deviation (SD)) for continuous variables and number (percent) for categorical variables. We used histograms to check the symmetry of the distribution for continuous variables prior to adopting mean (SD) for summary statistics. Unadjusted differences between groups were tested using the chi-squared test for categorical variables and the t-test for continuous variables. The Mann Whitney U-test was used when the assumption of normality was violated. Barthel Index score (and its components) and Robinson score were modelled to determine the impact of caregiver gender controlling for the number of hours per week spent helping, number of weeks of this amount of care, general health status of the caregiver, number of services used and caregiver age. For the purpose of examining supports provided to care recipients, three composite scores were developed by the authors for the Barthel Index. The composite binary score for personal grooming was obtained by combining responses to the items bathing, grooming, dressing and personal care in such a way that if any one of these items was a yes, then the composite would be a yes (code = 1, presence of spousal support, informal, or formal support) or no (code = 0, there was an absence of support). Toileting was calculated by combining responses to the items for bowel, bladder and toilet use in such a way that if any one of these items was a yes, then the composite binary score would be a yes (code = 1, presence of spousal support, informal or formal support) or no (code = 0, there was an absence of support). A composite binary score for mobility was developed by combining responses to the three items: stairs, mobility and transfer in such a way that that if any one of these items was a yes, then the composite would be a yes (code = 1, presence of spousal caregiver, informal, or formal support) or no (code = 0, there was an absence of support). To examine the third research question, that caregiver gender is independently associated with support for activities of daily living, we conducted a series of regression analyses. We used multiple regression to analyse the total Barthel score (the sum of the individual items). The next set of models used logistic regression to analysis the individual composite binary outcomes/ scores for personal toileting, and mobility described above. We also investigated potential interactions between caregiver gender and age, and gender and general health status. However, these interactions were not statistically significant and therefore were left out of the final models. The association between gender and 2008 The Authors, Journal compilation 2008 Blackwell Publishing Ltd 161

4 K. Brazil et al. support for ADL for the Barthel total is expressed as a regression coefficient with the corresponding 95% confidence interval (CI) and associated P-value. The association between gender and the individual Barthel items based on the logistic regression models are expressed as odds ratios (OR) with corresponding CI. All P-values are reported to three decimal places with P-values less than reported as P < The criterion for statistical significance was set a priori at alpha = Statistical software used for all analyses was SAS 9.1. We assessed multicollinearity using the variance inflation factor, which measures the extent to which the variance of the model coefficients will be inflated (because of the correlation of the variable with other predictor variables) if that variable is included in the model. Variables with variance inflation factor > 10 were considered collinear and were excluded from the analysis. We examined the residuals to assess model assumptions. In general, the Hosmer Lemeshow fit was acceptable (Hosmer & Lemeshow 1989). For all regression analyses we conducted, we had a sufficiently large sample size (Harris 1985, Green 1991). For logistic regression models, simulation studies show that one requires events per predictor to produce stable estimates (Peduzzi et al. 1996, Babyak 2004). Note that even with the inclusion of the interaction terms our sample size of 283 was sufficient to develop multivariable models. Results A total of 648 potential participants were identified as eligible to participate in the study; 530 consented to have the researchers contact them to describe the study and solicit their participation. Of these 530, 102 caregivers reported that the care recipient had died and did not complete the interview. Of the remaining 428, 55 declined to participate in the study, leaving 373 participating caregivers for an 87% consent rate. Of these interviewees, 294 were spouses actively caregiving. The spousal status was confirmed by the CCAC staff. Due to incomplete data, 11 spousal interviews were removed, leaving 283 spousal interviews to form the basis of the analyses presented in this report. Characteristics of study sample As shown in Table 1, most spousal caregivers were female and were more likely to care for an older care recipient. The two groups identified by gender did not differ with regard to any other demographic characteristics. Care recipients needs for assistance in activities in daily living did not differ between the two groups. Gender differences in the amount of caregiving and caregiver strain Examining for gender differences in the duration and amount of caregiving revealed that the mean number of caregiving hours spent in a week did not differ significantly between females and males, 63.6 (SD = 60.7) and 70.4 (SD = 62.4), respectively (t = 0.86, d.f. = 281, P =0.389). Females reported a mean of 27.2 weeks as active caregivers (SD = 48.6), while males reported a mean of 15.3 weeks (SD = 16.3). Due to the skewed distribution of responses on the length of time caregiving variable, the Mann Whitney U-test was calculated, revealing a non-statistically significant difference in length of time caregiving between male and female caregivers (Z = 1.09, P = 0.276). However, female caregivers reported a significantly greater level of caregiving strain (mean = 6.3, SD = 3.2) than males (mean = 5.4, SD = 3.2) (t = 2.12, d.f. = 281, P = 0.035). Gender differences in service use and need Table 2 shows community services used during the 4 weeks prior to the interview. The most frequently used services in both groups included home visiting nurses, case management, religious support, homemaking and occupational therapy. Female caregivers reported statistically significant greater use of transportation services while male caregivers reported statistically significant greater use of pain and symptom management consultants. Overall, the two groups reported receiving the same number of services during the 4-week period prior to the telephone interview. While the identified service needs for both groups was low, Table 3 identifies the most frequently reported service needs among both male and female caregivers as homemaking, respite and in-home nursing care. Gender differences in patterns of support to the care recipient Supports provided to care recipients was examined by type of care needs (personal grooming, toileting and mobility) and by source of support (spousal support, informal support (family, friends) and formal support (healthcare worker)). Table 4 provides the results of the univariable and multivariable analyses for all outcomes adjusting for number of hours per week spent helping, number of weeks of this amount of care, general health status, number of services used and caregiver age. Here we report the main findings based on the multivariable analyses. The Barthel Index score revealed that between male and female caregivers there was no statistically significant difference on the functional The Authors, Journal compilation 2008 Blackwell Publishing Ltd

5 Gender and caregiving at the end of life Table 1 Characteristics of the study sample (n =283) Gender of caregiver Characteristics Male (n =85) Female (n = 198) Care receiver Age (years): M (SD) 65.8 (8.2) 68.7 (9.0) Living arrangements: n (%) Live with one person 69 (81.2) 156 (78.8) Live with two or more persons 16 (18.8) 42 (21.2) Barthel score: M (SD) 13.6 (4.8) 13.5 (4.9) Spousal caregiver Age (years): M (SD) 68.3 (8.2) 65.3 (9.4) Annual family income (Canadian dollars): n (%) (6.6) 21 (12.8) (27.6) 52 (31.7) (17.1) 28 (17.1) (11.8) 16 (9.8) (7.9) 24 (14.6) (6.6) 8 (4.9) (22.4) 15 (9.1) Education: n (%) Elementary completed 37 (43.5) 78 (39.4) High school completed 21 (24.7) 85 (42.9) College completed 11 (12.9) 15 (7.6) University completed 14 (16.5) 17 (8.6) Would you say that in general your health is: n (%) Excellent 11 (12.9) 21 (10.6) Very good 27 (31.8) 50 (25.3) Good 29 (34.1) 86 (43.4) Fair 14 (16.5) 27 (13.6) Poor 4 (4.7) 14 (7.1) Caregiver strain: M (SD) 5.4 (3.2) 6.3 (3.2) M, mean; SD, standard deviation. Table 2 Service use by gender of caregiver (n =283) Gender of caregiver Services Male (n =85): number (%) Female (n =198): number (%) In-home nursing care 82 (96.5) 195 (98.5) Case management 50 (58.8) 120 (60.6) Religious support 38 (44.7) 80 (40.4) Homemaking 27 (31.8) 48 (24.2) Occupational therapy 24 (28.2) 44 (22.2) Nutritionist/dietician 20 (23.5) 47 (23.7) Physiotherapy 16 (18.8) 39 (19.7) Personal support worker 14 (16.5) 34 (17.2) Transportation 7 (8.2) 38 (19.3)* Pain and symptom management 14 (16.5) 15 (7.6) Day hospital 8 (9.4) 19 (9.6) Counselling 5 (5.9) 20 (10.1) Self-help/support group 7 (8.2) 16 (8.1) Podiatry 2 (2.4) 12 (6.1) Number of services used (mean (standard deviation)) 3.7 (1.6) 3.8 (1.6) * χ² = 5.41, d.f. = 1, P = χ² = 5.11, d.f. = 1, P = The Authors, Journal compilation 2008 Blackwell Publishing Ltd 163

6 K. Brazil et al. Gender of caregiver Table 3 Caregiver report on additional services needs (n =283) Services Male: n =85: number (%) Female: n =198: number (%) Homemaking 7 (8.2) 19 (9.6) Respite services 4 (4.7) 18 (9.1) In-home nursing care 6 (7.1) 13 (6.6) Self-help/support group 4 (4.7) 7 (3.5) Podiatry 3 (3.5) 8 (4.0) Personal support worker 3 (3.5) 6 (3.0) Counselling 1 (1.2) 7 (3.5) Case management 1 (1.2) 2 (1.0) Pain and symptom management 1 (1.2) 1 (0.5) Physiotherapy 1 (1.2) 1 (0.5) Transportation 4 (2.0) Speech therapy 1 (1.2) Hospice palliative care volunteers 2 (1.0) Nutritionist/dietician 2 (1.0) Outcome* Univariable results Caregiver gender effect (95% CI); P Multivariable results Caregiver gender effect (95% CI); P Table 4 Caregiver gender differences in patterns of support to the care recipient Barthel overall sum 0.17 ( 1.06, 1.41); ( 0.38, 1.57); Personal Spousal support 1.32 (0.79, 2.21); (0.93, 3.08); Informal support 0.17 (0.06, 0.52); (0.05, 0.53); Formal support 1.11 (0.62, 2.00); (0.61, 2.54); Any supports 0.95 (0.55, 1.64); (0.60, 2.20); Toileting Spousal support 1.40 (0.81, 2.43); (1.01, 3.85); Informal support 0.21 (0.02, 2.36); (0.00, 3.02); Formal support 1.32 (0.50, 3.44); (0.53, 4.13); Any supports 1.51 (0.88, 2.59); (1.14, 4.30); Mobility Spousal support 0.41 (0.22, 0.77); (0.21, 0.81); Informal support 0.80 (0.40, 1.62); (0.43, 1.83); Formal support 1.90 (0.69, 5.21); (0.67, 5.99); Any supports 0.42 (0.22, 0.79); (0.21, 0.83); *Barthel overall sum items: regression coefficient (95% CI); P-value. Other items (Spousal, Informal, Formal, Any support): odds ratio (95% CI); P-value. Coding: male = 0, female = 1. Controlling for the following predictors (no interaction terms included in the model): number of hours per week spent helping; number of weeks of this amount of care; general health status; number of services used; and spousal age (year). status of care recipients (P=0.234). When considering source of support in tasks associated with daily living for the care recipient, differential assistance was noted on the basis of caregiver gender. Spousal support was the main avenue for support in toileting and mobility with female caregivers having almost twice the odds of providing support in toileting-related tasks than male caregivers (OR = 1.98, P = 0.044). Compared to female caregivers, male caregivers had approximately twice the odds of providing supports in mobility-related tasks (OR = 041, P = 0.011). It was observed that care recipients who had a female caregiver had lower odds of receiving informal supports from family and friends in tasks associated with personal care (OR = 0.17, P = 0.002). Discussion The majority of studies on informal caregiving in palliative care have not taken gender into account. The findings from this study demonstrate that men and The Authors, Journal compilation 2008 Blackwell Publishing Ltd

7 Gender and caregiving at the end of life women have different experiences when caring for a terminally ill spouse. Consistent with what has been reported in other study populations, most caregivers in our study were women (Ross et al. 2002), who, compared to male caregivers, reported greater strain as a result of caregiving (Keating 1999, Yee & Schulz 2000, Navaie- Waliser et al. 2002, Bedard et al. 2005). Overall, the use of community services was similar between male and female caregivers. However, it was noted that female caregivers made greater use of transportation services while male caregivers made greater use of pain and symptom management consultants. The reason for differential use in these two services is unknown and requires more in-depth investigation than data collection strategies in the present study allowed. Differential assistance to the care recipient was also noted on basis of caregiver gender. Female caregivers had lower odds of receiving assistance from extended family and friends in personal care. Female caregivers were twice as likely to provide assistance in toiletingrelated tasks whilst male caregivers were more than twice as likely to provide assistance in mobility-related tasks. Explanation for this observed difference may be due to gendered nature of caregiving. Women are socialised and viewed by society as natural caregivers regardless of individual ability or willingness to assume the caregiving role (Brody 2004). Meanwhile, in the case of male caregivers, it is assumed by others that assistance is required. In our study, it was observed that others were more likely to assist male caregivers in personal support tasks. Maintaining care of the terminally ill in the community is dependent on the availability, willingness and capacity of informal caregivers, typically a female family member. While caregiving remains a typical female role, many changes in recent years have placed even greater pressures on women to care for the terminally ill person. Meanwhile, inadequately funded home care programmes have forced caregivers to assume greater responsibilities in response to a limited community care system (Aronson & Neysmith 1997). Equating informal caregiving as a natural part of family relations may create the perception that informal caregiving is simply part of family responsibilities and not necessarily a policy concern. Informal caregivers represent an essential feature of community care and require formal supports in their caregiving responsibilities. Participants in our study identified respite care as one support option. This would include more hours of service, more flexibility and ensuring that caregivers are aware of the availability of respite care (Chappell 1993). Study participants also identified the need for more instrumental support that would alleviate caregiver strain, namely nursing care and homemaking. Numerous options relevant to supporting the informal caregiver in the community have been identified in the literature. Training for caregivers has been identified as one strategy to relieve burden and strain. Skills training could include transferring the care recipient from a bed to a chair, lifting, bathing, toileting and other personal care activities. Informal caregivers may also benefit from information support where counselling, medical, and legal assistance, as well as help in navigating the health/social care system, can be obtained (Ostwald et al. 1999). Additional considerations include improved training and education in palliative care for community service providers (Institute of Medicine 1997). Despite the strengths of our study, its limitations should be noted. While the intent of the investigators was to recruit all eligible candidates as they came on service with the participating agencies, it is possible that some eligible participants were missed because of the challenges experienced by case managers recruiting participants while providing front-line care, although we have no means of determining to what extent this may have occurred. Furthermore, it is uncertain as to the presence of a selection bias among those who did participate in the study (e.g. those who are most busy caring, and perhaps most under strain, may not have the capacity to participate in the study). However, given that the literature suggest that women are typically under greater strain than men, it is possible that our results represent an underestimate of the gender differential. While the above service development options are directed to improving community services for female caregivers, they hold more profound implications, by enhancing the recognition of informal caregivers within policy circles (Gallicchio et al. 2002). To address gender differences in caregiving, a realistic home-based palliative care approach must take into account the importance of informal caregivers, the gendered nature of caregiving and the unique support needs of caregivers on the basis of gender. Improvements in home-based palliative care services will benefit both men and women who assume the caregiving role. Acknowledgement The Canadian Health Services Research Foundation provided funding for this research project (RC ). References Aranda S. & Peerson A. (2001) Caregiving in advanced cancer: lay decision making. Journal of Palliative Care 17, Aronson J. & Neysmith S.M. (1997) The retreat of the state and long term care provision: implications for frail elderly 2008 The Authors, Journal compilation 2008 Blackwell Publishing Ltd 165

8 K. Brazil et al. people unpaid family carers and paid home care workers. Studies in Political Economy 53, Babyak M.A. (2004) What you see may not be what you get: a brief, nontechnical introduction to overfitting in regressiontype models. Psychosomatic Medicine 66, Bedard M., Kuzik R., Chambers L., Molloy D.W., Dubois S. & Lever J.A. (2005) Understanding burden differences between men and women caregivers: the contribution of care-recipient problem behaviors. International Psychogeriatric 17, Brazil K., Bedard M., Willison K. & Hode M. (2003) Caregiving and its impact on families of the terminally ill. Aging and Mental Health 7, Brody E. (2004) Women in the Middle: Their Parent Care Years. Springer Publishing, New York. Canadian Study of Health and Aging Working Group (1994) Canadian study of health and aging: study methods and prevalence of dementia. Canadian Medical Association Journal 150, Chappell N.L. (1993) Implications of shifting health care policy for caregiving in Canada. Journal of Aging and Social Policy 5, Collin C., Wade D.T., Davies S. & Horne V. (1988) The Barthel ADL Index: a reliability study. International Disability Studies 10, Covinsky K.E., Goldman L., Cook E.F., et al. (1994) The impact of serious illness on patients families. SUPPORT Investigators. Study to understand prognoses and preferences for outcomes and risks of treatment. Journal of the American Medical Association 272, Dahlberg L., Demack S. & Bambra C. (2007) Age and gender of informal carers: a population-based study in the UK. Health and Social Care in the Community 15, Emanuel E.J., Fairclough D.L., Slutsman J. & Emanuel L.L. (2000) Understanding economic and other burdens of terminal illness: the experience of patients and their caregivers. Annals of International Medicine 132, Gallicchio L., Siddiqi N., Langenberg P. & Baumgarten M. (2002) Gender differences in burden and depression among informal caregivers of demented elders in the community. International Journal of Geriatric Psychiatry 17, Gaugler J.E., Kane R.A. & Langlois J. (2000) Assessment of family caregivers of older adults. In: R.L. Kane & R.A. Kane (Eds) Assessing Older Persons. Measures, Meaning, and Practical Applications, pp Oxford University Press, New York. Gott M., Seymour J., Bellamy G., Clark D. & Ahmedzai S. (2004) Older people s views about home as a place of care at the end of life. Palliative Medicine 18, Green S.B. (1991) How many subjects does it take to do a regression analysis? Multivariate Behavioral Research 26, Grunfeld E., Glossop R., McDowell I. & Danbrook C. (1997) Caring for elderly people at home: the consequences to caregivers. Canadian Medical Association Journal 157, Harris R. (1985) A Primer of Multivariate Statistics, 2nd edn. Academic Press, New York. Hodgson C., Higginson I., McDonnell M. & Butters E. (1997) Family anxiety in advanced cancer: a multicentre prospective study in Ireland. British Journal of Cancer 76, Hosmer D. & Lemeshow S. (1989) Applied Logistic Regression. John Wiley & Sons, New York. Institute of Medicine (1997) Approaching Death: Improving Care at the End of Life. National Academy Press, Washington, DC. Keating N.C. (1999) Eldercare in Canada: Context, Content and Consequences. Statistics Canada, Housing, Family and Social Statistics Division, Ottawa, ON. Morris M. (2001) Gender-sensitive Home and Community Care and Caregiving Research: A Synthesis Paper. Women s Health Bureau, Health Canada, Ottawa, ON. Navaie-Waliser M., Spriggs A. & Feldman P.H. (2002) Informal caregiving: differential experiences by gender. Medical Care 40, Ostwald S.K., Hepburn K.W., Caron W., Burns T. & Mantell R. (1999) Reducing caregiver burden: a randomized psychoeducational intervention for caregivers of persons with dementia. Gerontologist 39, Payne S., Smith P. & Dean S. (1999) Identifying the concerns of informal carers in palliative care. Journal of Palliative Medicine 13, Peduzzi P., Concato J., Kemper E., Holford T.R. & Feinstein A.R. (1996) A simulation study of the number of events per variable in logistic regression analysis. Journal of Clinical Epidemiology 49, Robinson B.C. (1983) Validation of a caregiver strain index. Journal of Gerontology 38, Ross M.M., MacLean M.J., Cain R., Sellick S. & Fisher R. (2002) End of life care: the experience of seniors and informal caregivers. Canadian Journal on Aging 21, Sankar A. (1993) Images of home death and the elderly patient: romantic versus real. Generations 17, Senate Committee on Social Affairs (2000) Quality End-of-Life Care: The Right of Every Canadian. Parliament of Canada, Ottawa, ON. Smith N. (1990) The impact of terminal illness on the family. Journal of Palliative Medicine 4, Stajduhar K.I. & Davies B. (1998) Death at home: challenges for families and directions for the future. Journal of Palliative Care 14, Steinhauser K.E., Christakis N.A., Clipp E.C., McNeilly M., McIntyre L.M. & Tulsky J. (2000) Factors considered important at the end of life by patients, family, physicians, and other care providers. Journal of the American Medical Association 284, Teno J.M., Clarridge B.R., Casey V., et al. (2004) Family perspectives on end-of-life care at the last place of care. Journal of the American Medical Association 291, Thomas C., Morris S.M. & Clark D. (2004) Place of death: preferences among cancer patients and their carers. Social Science Medicine 58, Wilson D.M., Smith S.L., Anderson M.C., et al. (2002) Twentiethcentury social and health-care influences on location of death in Canada. Canadian Journal of Nursing Research 34, Yee J.L. & Schulz R. (2000) Gender differences in psychiatric morbidity among family caregivers: a review and analysis. Gerontologist 40, The Authors, Journal compilation 2008 Blackwell Publishing Ltd