Introduction

Informal caregivers provide an enormous amount of care to ill family members, often at a considerable cost to their own well-being [1–3]. The point prevalence of clinically relevant psychological distress in individuals providing care to family members at the end of life is about 30 %; most frequent are depression and anxiety [4]. Informal caregivers of palliative care (PC) patients often report equivalent or even higher levels of burden than the patients themselves [3]. The burden of caregivers in the final phase of their relative's life and in early bereavement is comparable [5]. In palliative care, family caregivers are also care recipients—they are an integral part of the unit of care [6]. While there is abundant research about the needs and burden of caregivers, there is a dearth of information about conceptual models for preventive approaches and effective interventions [2, 3, 7, 8]. Reviews about existing interventions criticise the lack of stringent designs and many interventions show only small or no effects [9–11].

In the treatment of palliative patients, clinicians and researchers have recognised that existential concerns are just as important to address as well-known clinical syndromes such as depression or anxiety [12, 13]. Newer interventions in end-of-life care have been aimed at increasing the feeling of dignity [14] and meaning in patients' lives [15–17]; however, they have mostly neglected promotion of meaning in life (MiL) in relatives and caregivers so far.

MiL is related to well-being in the general population [18, 19], as well as in samples of caregivers [20, 21], and the bereaved [22, 23]. MiL has been defined as “the cognizance of order, coherence, and purpose in one's existence, the pursuit and attainment of worthwhile goals, and an accompanying sense of fulfillment” ([24], p. 41); thus, it comprises a cognitive and a motivational component (providing a framework for meaning), and an emotional component (fulfillment). A large number of questionnaires have been developed to assess the various aspects of MiL [25]. Idiographic MiL measures assess the subjective meaning a person experiences by allowing them to list applicable meaning areas [25].

Meaning making [26, 27] and finding meaning through caregiving [28–30] reflect coping strategies used when faced with adverse events [31, 32] and must be differentiated from general MiL. The latter can be conceptualised as “meanings-made” [31]; thus, MiL can reflect an outcome of the aforementioned coping strategies. Several studies have examined general MiL in the context of caregiving and bereavement. The experience of MiL may be a buffering mechanism against negative outcomes in caregiving [21] and in bereavement [23, 33]. Higher MiL was associated with better health in spousal caregivers of patients with advanced cancer [21]. Widows were shown to have higher MiL than psychiatric patients, but lower MiL than a mixed sample of healthy adults and non-psychiatric patients [34]. In grieving parents, higher MiL was related to less grief [33]. In newer studies, perceiving one's life as worth living was associated with lower burden in caregivers of disabled elderly [35]. MiL significantly predicted psychological well-being in older adults after spousal loss, even after controlling for social support and physical health [22]. These studies were restricted to investigate the relationship between MiL fulfillment and caregiver well-being. However, no studies have reported on how the framework for MiL in bereaved family carers differs from other populations.

The objective of this study is to explore MiL in bereaved caregivers and compare it to a representative population sample [36]. Using the individual assessment afforded by the Schedule for Meaning in Life Evaluation (SMiLE; [37]), we explore the extent to which bereaved caregivers experience fulfillment of MiL, as well as look at how they construct their MiL framework (number and type of listed MiL sources; ascribed importance). Additionally, we investigate how MiL fulfillment relates to quality of life and psychological distress in bereaved carers.

Methods

Study design

Participants were informal carers of PC patients who agreed to take part in a randomised supportive group intervention trial [38]. They were recruited from two palliative care wards and a radiooncology department in Munich, and completed a battery of self-report questionnaires including the ones used in this study. Questionnaires were either mailed or handed out in person and were completed by respondents on their own. To preclude intervention influences, respondents completed self-report questionnaires before the intervention started. At this point, 70 % of the randomised controlled trial participants were bereaved. To achieve reasonable sample homogeneity, only data from the bereaved subgroup are reported. This group's MiL is compared to MiL data from a German nationwide representative sample, gathered via computer-assisted telephone interviews [36]. The study has been approved by the Research Ethics Committee of the University Hospital Munich, Germany (project no. 120-06).

Measures

MiL was assessed with the SMiLE [37]. Respondents list up to seven areas relevant for their MiL, and rate their current level of satisfaction (−3 to +3) and importance of each area (0–7). From the importance ratings, an overall index of weighting is calculated (IoW, 0–100), the satisfaction ratings yield a total index of satisfaction (IoS, 0–100) and a combination of the importance and satisfaction ratings yields the overall index of weighted satisfaction (IoWS, 0–100). In contrast to previous publications [36, 39, 40], an 8-point Likert scale was introduced for the importance ratings to avoid ceiling effects (previously 5-point) [39]. Importance ratings of the previously assessed representative sample were transformed using proportional transformations [41]. Furthermore, to control for individual differences in response scale use, we calculated relative importance ratings following an established method from personal values research [42]. MiL areas were subsumed under 13 categories found in earlier studies [36, 37]. Satisfaction ratings (overall and by content category) indicate the extent of MiL fulfillment an individual experiences, while the content of listed MiL areas, the overall index of weighting and the relative importance ratings reflect a person's individual construction of MiL framework. The SMiLE has shown satisfactory psychometric properties (retest reliability, convergent and discriminant validity) [37].

Psychological distress was assessed with the Global Severity Index of the 53-item Brief Symptom Inventory [43]. Raw scores were converted into gender-adjusted T values (mean = 50, SD = 10). Quality of life was assessed with the 26-item WHOQOL-BREF [44], which comprises four dimensions: physical, psychological, social and environmental quality of life. An overall score (0–100) was used in this study. Additionally, overall quality of life was assessed using a single item numerical rating scale (NRS, 0–10; “How do you rate your quality of life at the moment”). Life satisfaction was assessed using the five-item Satisfaction with Life Scale (SWLS, 1–7) [45].

Statistical analysis

Separate linear regression model analyses were used to test for group differences in MiL between PC caregivers and the representative sample. All analyses were controlled for gender and age (categorical, up to 19, 20–29, 30–39, etc.). Loglinear quasi-Poisson regression analyses were applied to test for differences in the number of listed MiL areas. Binary logistic regression analyses were performed to identify differences in the likelihood of listing a specific MiL area. In these, the number of remaining MiL areas listed by each participant was entered as an additional covariate. The same analyses were used in the preliminary analyses to examine the effects of relationship with the patient and time since loss. For all analyses, Bonferroni corrections were employed, and differences were considered to be statistically significant at the adjusted level for p < .05 for multiple comparisons. Results significant at p < .05 without alpha-error adjustments are reported for exploratory purposes. To evaluate the relationship between MiL (IoWS), psychological distress and well-being in the bereaved PC caregiver sample, partial age- and gender-adjusted Pearson-correlation coefficients were calculated. Statistical tests were performed with the Statistical Package for Social Sciences (SPSS), version 19.

Results

Preliminary analyses

Of the participants, 69.0 % were partners of the patient, 23.8 % adult children, 3.6 % parents and 3.6 % had other relationships. Time since death of patient was 68.0 ± 47.0 days (range, 4–183 days). To examine whether our sample of bereaved PC caregivers is homogenous with regard to their experience of MiL, we performed preliminary analyses to test for effects of relationship to the patient and bereavement length of time within the sample of caregivers. The inclusion of the variable time since loss (in days) in all analyses revealed only one effect at p < .05: The relative importance of friends increased with time (days) since loss. The inclusion of a variable coding the effect of being the partner of the patient versus having a different relationship revealed no effect on the SMiLE indices or number of provided areas. However, partners of patients were more likely to include the areas partner and spirituality as meaning providing areas, they ascribed more importance to partner, and were more satisfied with family, while meaning areas concerning their house/garden were less important and yielded less satisfaction (all p < .05, except lower satisfaction with house/garden: p < .01). Since none of these effects reached significance after Bonferroni correction, data from all bereaved caregivers were analysed as a single group.

Respondents' characteristics

Table 1 provides the respondents' characteristics for the bereaved PC caregivers as well as the representative sample.

Table 1 Respondents' characteristics of bereaved PC caregivers (n = 84) and a German representative sample (n = 977)
Full size table

Mean age of PC caregivers was 55.5 ± 12.9 years, 81.0 % were caregivers of cancer patients. Other diagnoses included neurological disorders (13.0 %), chronic obstructive pulmonary disease (3.6 %) and other (2.4 %).

With regard to sociodemographic characteristics, significant group differences emerge with regard to age [χ 2(6) = 52.78, p < .001], gender [χ 2(1) = 4.38, p < .05] and marital status [χ 2(3) = 214.37, p < .001], but not with regard to education [χ 2(2) = 0.48, n.s.]. We included age and gender as covariates; marital status was not included because it is confounded with group membership for the 69.0 % of PC caregivers who had just experienced the loss of their spouse at the time of assessment.

Overall MiL

Table 2 shows descriptive statistics of the overall SMiLE indices and the number of listed MiL areas for the two groups.

Table 2 Descriptive statistics of overall MiL indices by group
Full size table

Table 3 shows linear regression results examining group differences in these indices. PC caregivers had lower overall satisfaction (IoS), overall importance (IoW) and weighted satisfaction (IoWS) than the representative sample (all p < .001). With a B ranging from −9.57 to −15.52 points, effect sizes are medium for the SMiLE indices (possible range from 0 to 100). As an interpretation example, a B of −15.52 for the IoWS indicates that when comparing a respondent of the PC caregiver group to a person with the same age and gender in the representative sample, the former's overall SMiLE index is 15.52 points lower. On the other hand, PC caregivers listed on average more MiL areas than the representative sample [exp(B) = 1.32, SE = 0.05; CI, 1.22–1.42, p < .001]. The exp(B) of 1.32 indicates that a person in the PC caregiver group would list 1.32 times more MIL areas than a person of similar age and gender in the representative sample (multiplicative effect).

Table 3 Linear regression model testing for group differences between bereaved PC caregivers and the German representative sample in overall MiL indicators (controlled for age and gender, N = 1061)
Full size table

Frequency of specific MiL areas

PC caregivers listed overall 424 MiL areas. These listings were assigned to 13 categories derived from the representative analyses [36].

Table 4 presents percentages of participants who named MiL areas in the respective category along with descriptive statistics of satisfaction (s) and relative importance (w rel) ratings for PC caregivers and the representative sample. Kendall's rank correlation coefficient was used to compare the order of MiL areas (N = 13) in the two groups. The groups showed similar rank orders with regard to frequency of MiL types (r = .59, p = .006), but only a trend towards a similar rank order with regard to satisfaction (r = .37, p = .09) and relative importance scores of MiL areas (r = .40, p = .07) emerged.

Table 4 Descriptive statistics by MiL area: frequencies of listing, satisfaction and importance ratings by group
Full size table

Results of the binary logistic regression analyses testing for significant group differences in the likelihood to list specific MiL areas are provided in Fig. 1.

Fig. 1
figure 1

Odds ratios [Exp(B)] with low and high confidence intervals for comparison of frequencies of MiL areas between bereaved PC caregivers and the German representative sample (controlled for age, gender, and remaining number of listed MiL areas). Numbers below 1 indicate a lower, numbers above 1 a higher likelihood in PC caregivers to list a specific MiL area. * p < .004 (significant), ^p < .05 (trend)

Full size image

Compared to the representative population sample, PC caregivers are more likely to list MiL areas falling within the categories friends, leisure, nature/animals and altruism (p < .004, see Fig. 1). In addition, PC caregivers tend to list spirituality more often and health and finances less often (p < .05). The largest effects are observed with respect to friends (OR, 5.26), altruism (OR, 3.94) and nature/animals (OR, 2.91).

Satisfaction with MiL areas

Compared to the representative sample, bereaved PC caregivers showed significantly lower satisfaction in the MiL areas of family, friends, leisure, nature/animals, partner, and altruism (p < .004, see Table 5). Effect sizes are medium to large (B ranges from −0.81 to −1.70, response scale range −3 to +3). Additionally, there is a trend towards lower satisfaction in the area of work, spirituality, home/garden and well-being (p < .05).

Table 5 Linear regression model testing for group differences between bereaved PC caregivers and the German representative sample in satisfaction and relative importance ratings by MiL area (controlled for age and gender)
Full size table

Relative importance of MiL areas

Compared to the representative sample (see Table 5), PC caregivers ascribed higher relative importance to family (p < .004) with a small to medium effect size and showed trends towards higher ratings for partner and health and lower ratings for pleasure (p < .05).

Relationship between MiL and well-being in PC caregivers

Only small correlations were observed between MiL (IoWS, M = 68.33, SD = 19.12) and overall symptom severity (M = 65.14, SD = 11.42, r = −.26, p < .02) and quality of life (NRS; M = 4.12, SD = 1.96, r = .23, p < .05). However, moderate correlations were found between MiL and life satisfaction (SWLS, M = 3.79, SD = 1.15, r = .45, p < .001) and quality of life assessed with the WHOQOL-BREF (M = 63.78, SD = 10.50, r = .41, p < .001).

Discussion

This study evaluates the experience of individual MiL and its relationship to well-being in bereaved informal caregivers.

Overall MiL in caregivers

Bereaved caregivers had lower overall importance (IoW), satisfaction (IoS) and weighted satisfaction (IoWS) with regard to MiL than a representative sample of the German population. This result concurs with Levinson's findings of lower MiL in widows compared to a mixed adult (non-psychiatric patient) sample [34]. In studies comparing palliative patients to the representative sample [39, 46], PC patients also had lower satisfaction and weighted satisfaction scores (fulfillment), but did not differ from the general population in terms of the overall importance they ascribed to their meaning-providing areas (framework). This finding underscores the importance of support for informal carers during palliative care [1–3, 9–11] and into bereavement [47, 48], and confirms the notion that they experience often as much or even higher psychological maladjustment than the patients [3].

Differences in MiL fulfillment

Bereaved PC caregivers' overall experience of satisfaction in MiL is considerably lower than in the representative sample [36]. It is worth noting that caregivers have the absolute lowest IoS and IoWS observed in any of the samples so far, their average indices being between 6 and 10 points lower than in samples of patients with amyotrophic lateral sclerosis (ALS) [46], cancer patients [40] and students [37]. The preliminary analyses indicate that these effects similarly affect partners and non-partners, and are homogenous throughout the first 6 months of bereavement.

All areas that 15 % or more of PC caregivers list as meaning providing are impaired: Their satisfaction is significantly (family, friends, leisure, nature/animals, partner and altruism) or by trend lower (work and spirituality) than in the representative sample. Among the affected are all areas that caregivers are more likely to list. These consistently lower satisfaction ratings indicate a pervasive impact of the caregiving role and the recent loss of a loved one on the experience of MiL. While PC patients [39] succeeded in keeping their satisfaction unimpaired for at least some MiL areas (family, partner and spirituality), caregivers do not seem to have such protected areas. Furthermore, comparing descriptive statistics shows that caregivers seem to be less satisfied with family than patients [39]. This may reflect conflict potential and unfulfilled expectations with regard to support from one's family [48, 49]. Preliminary analyses indicate that caregivers who are partners of the patient are somewhat more satisfied with family than non-partner caregivers. Possibly, when one's partner is affected, immediate and extended family may rally together more to support the surviving partner, than when the patient is not a partner. Attention should be paid to whether non-partner carers receive sufficient informal social support during caregiving and bereavement. The lower satisfaction expressed by partners with house/garden when compared to non-partners could arise from the fact that the surroundings they have shared with their partner before the death are filled with permanent reminders of their loss and the necessity to adapt to the new situation.

Differences in MiL framework

The unique MiL framework of caregivers is expressed (1) by the specific MiL areas mentioned, as well as (2) by the importance ratings associated with them.

First, the categories mentioned most often by bereaved PC caregivers (in descending order) were family, friends, leisure, work, nature/animals and partner. Caregivers are equally likely to list family or work as relevant for MiL as the representative sample. However, they list friends, leisure, nature/animals and altruism much more frequently. Leisure and nature/animals were also listed more often by PC patients [39]. These areas may be important for both patients and caregivers as ways to counteract the respective burden experienced in their illness or caregiving role. Friends may be listed more often either as part of a social support network in relation to caregiving, or as a niche where carers are able to occasionally escape from the illness or loss of their relative. Altruism was listed by 16 % of the caregivers; this is about four times more often than in the representative sample. This may reflect that caregivers experience meaning in their role as caregiver of a terminally ill patient [3, 29]. The trend (p = .004) that caregivers list health less often parallels findings from ALS patients [46], and may be explained by a possible response shift [50]. Experiencing the terminal illness of a loved one and the helplessness associated with it may lead to perceiving health as something brittle that is mostly outside one's control. Thus, building one's MiL around other areas may be an adaptive response to the experience of caregiving [32]. Relatives who had lost their partner were more likely to name the areas partner and spirituality as MiL providing, possibly indicating that the loss of one's partner leads more often to questions concerning the transcendent.

Second, only one significant difference in importance of specific MiL areas was observed: Family was rated as more important in caregivers; combined with the fact that this area is relevant for most caregivers (82.3 %) and that they experience distinctly lower satisfaction in it, this is a prominent target for intervention, such as family focused grief therapy [49] or counselling targeted at improving family relationships [48]. The lower overall index of weighting may reflect a general insecurity or even “shattering” of assumptive worlds [31] in the bereaved with regard to their MiL framework. Referring to Park's meaning-making model [31], it seems that bereaved caregivers are in a transition phase. They have experienced loss of MiL and have started to change their framework of how to experience MiL (as evidenced by the different types of MiL areas and the overall lower certainty about what is important to them), but have not yet achieved satisfactory fulfillment within their new framework [51].

Relationship of MiL to other constructs

In line with Fry [22], MiL in bereaved PC caregivers was moderately related to well-being in the expected direction. Thus, MiL appears to be a relevant construct for psycho-oncology and end-of-life research, and may mediate the relationship between caregiver burden and well-being [18, 35]. Within the current cross-sectional design, it is not possible to discern a plausible causal direction of these relationships. Longitudinal studies are needed to investigate what mechanisms underlie these associations.

Limitations of the study

One limitation is the mixed sample of bereaved caregivers: They had various relationships to the deceased family member, and time since loss varied from a few days to 6 months. However, preliminary analyses indicated that these subgroups are sufficiently similar with regard to their experience of MiL. The cross-sectional design of the study does not allow assessment of changes within respondents over time. Differences between samples could have existed prior to the beginning of caregiving. Only caregivers who decided to participate in an intervention study and had lost their ill relative prior to the pre-treatment assessment were included. It is likely that participating caregivers differ from those who declined with regard to the degree of burden and other variables (e.g. time constraints) that play a role in the uptake of such offers [1, 52]. We have introduced a new response-scale format to prevent ceiling effects [39], which may have led to slight distortions in the IoW and IoWS comparisons [41].

Conclusion

In conclusion, both the framework and the fulfillment of MiL appear to be substantially impaired in PC caregivers. The observed differences in framework point to a response shift [50]. The loss of MiL fulfillment, comparable in extent or even higher than the loss observed in ALS [46] and PC [39] patients, underscore the need for specific interventions for informal caregivers in PC. Such interventions should not only provide practical support but also focus on promoting emotional well-being [1, 53]. As MiL was related to well-being, promoting the reconstruction of caregivers' MiL may be a promising focus for caregiver interventions. Existential behavioural therapy, for example, was shown to improve quality of life and reduce distress of caregivers immediately after treatment and at a 1-year follow-up [38]. To our knowledge, this is the first empirical investigation of the specific areas of MiL important to bereaved caregivers. Further research is needed to examine whether these differences represent a general phenomenon in caregivers.