Detecting pre-death grief in family caregivers of persons with dementia: measurement equivalence of the Mandarin-Chinese version of Marwit-Meuser caregiver grief inventory

With the rising prevalence of dementia globally [1], more family members will inevitably be drawn into the care of their relatives with dementia. Providing care to a person with dementia (PWD) is often protracted, stressful [2] and not uncommonly associated with physical and emotional burden [2‐4]. Caregiver burden, in turn, has been shown to produce less desirable outcomes such as premature nursing home placement of the PWD [5] and greater risk of mortality in the PWD [6]. For this reason, much research has targeted caregiver burden and the appropriate interventions to alleviate burden [3]. However, the loss and grief that caregivers often experience while providing care is less commonly recognised.

Caregivers of PWD may begin their bereavement long before the physical death of the PWD [7]. The caregivers’ responses to perceived losses in the pre-death context – which we refer to as pre-death grief (PDG) – include the anticipation of future loss related to the physical death of PWD and the mourning of present loss related to the psychological death of PWD (a unique phenomenon in dementia whereby the PWD is still physically present but emotionally disconnected from the caregiver) [8]. In a recent systematic review from the British Journal of Psychiatry [9], PDG has been highlighted as a key challenge faced by caregivers, with the loss of emotional bond with PWD accentuating caregivers’ unmet needs for nurture, companionship and emotional security. PDG has also been described as the central experience of spousal caregivers of PWD in another recent systematic review of the qualitative studies [10]. The prevalence of PDG has been reported in a few studies and ranged between 47 and 86% [8, 11].

PDG in dementia caregiving can be more complicated than the usual experience of loss and grief [8, 12]. The protracted trajectory of dementia prolongs the caregivers’ sense of uncertainty about the loss. During this protracted period of uncertainty, the communication with the PWD also becomes disrupted and there is no longer any opportunity to reaffirm the relationship. To make the matter worse, such experience of grief before the physical death of PWD may not be socially sanctioned, leaving the caregivers with the sense of isolation as they find it difficult to share their feelings with the people around them. When the experience of loss and grief is not recognized, caregivers may have difficulty accepting any changes that have occurred in the PWD and may attempt to fight the inevitable decline in the PWD. To that end, they may be more inclined to cope using control-based strategies, become more paternalistic in their communication with the PWD and adopt more authoritarian decision-making styles in caregiving. Eventually, this can affect the well-being of both the PWD and the caregivers, with the PWD feeling the sense of loss of autonomy and the caregivers feeling more helpless over what they cannot control [13]. Not surprisingly, high levels of PDG have been associated with caregiver burden [14] and depression [11], and increase the risk of caregivers considering nursing home placement for the PWD [15]. However, caregivers with high PDG may not be easily recognised in clinical practice because PDG is commonly disenfranchised [12] and not readily talked about by caregivers. They may also be missed by other routinely-used scales such as the caregiver burden scale as demonstrated by a recent study [8], thus indicating the need for a scale that specifically measures PDG.

The Marwit-Meuser Caregiver Grief Inventory (MM-CGI) [16] and its abbreviated MM-CGI-Short-Form (MM-CGI-SF) [17] are among the few empirically-developed scales that measure PDG [11, 12]. However, it has not had a Mandarin-Chinese version even though the Mandarin-Chinese language is among the most widely-used language worldwide [18], and countries with Chinese-speaking populations have among the largest number of PWD (with 9.5 million in China alone) [1]. Although there was a recent study that validated the Chinese MM-CGI-SF, it only focused on the Cantonese-Chinese version [19]. Cantonese-Chinese is a dialect of the Chinese language that is used in only 5% of the Chinese-speaking community [18], in contrast to the Mandarin-Chinese which is used in 70% of the community [18] and is recognised as the national language in countries such as China and Taiwan. While the Cantonese-Chinese and Mandarin-Chinese share similarities in the written characters, they are distinct and may not be mutually intelligible.

In this study, we produced the Mandarin-Chinese version of MM-CGI for use among Chinese-speaking caregivers. As the language of a scale can have prominent influence [20] on individuals’ responses to culturally-sensitive questions (such as those related to grief) [21], we are uncertain whether the new Mandarin-Chinese version will be equivalent in its measurement scores to those of the original English version. This is an important issue for future research that endeavours to compare the scores between the Chinese- and English-speaking populations in cross-cultural studies, or pool data from the two language versions to improve statistical power and representativeness of a study. As such, we sought to evaluate whether the Chinese versions of MM-CGI and its abbreviated MM-CGI-SF are equivalent in their measurement scores to the original English version. As a secondary aim, we also sought to evaluate whether the Chinese and English versions demonstrate similar patterns in reliability and validity assessment.

The demographic information of the 394 participants is shown in Table 1. The participants comprised multiple ethnicities, with a predominance of Chinese (86.6%). Spousal caregivers constituted 13.7%, with the rest being children caregivers. Participants who completed the Chinese version were slightly older (p = 0.015), less educated (p < 0.001) and less likely to be primary caregivers (p = 0.045).

The Chinese and English versions showed similar mean scores: 144.1 (SD 28.0) and 140.5 (SD 35.6) respectively for MM-CGI; and 51.9 (SD 10.6) and 51.5 (SD 13.5) respectively for MM-CGI-SF. The mean score-difference between the Chinese and English versions are illustrated in Fig. 1. In the adjusted model, the score-difference was 1.2 (90% CI -5.6 to 7.9) for MM-CGI and − 0.4 (90% CI -2.9 to 2.1) for MM-CGI-SF. The 90% CI for the adjusted score-differences fell within our predefined margins of equivalence (±8 for MM-CGI and ± 3 for MM-CGI-SF) and indicated equivalence of the scores. The adjusted score-differences for the subscales of MM-CGI and MM-CGI-SF are presented in Table 2. Although the subscales did not meet the 5% equivalence margin, all of them had the 90% CI falling within the less-stringent 10% margin which indicated possible equivalence.

The two language versions demonstrated similar patterns in internal consistency reliability, test-retest reliability, known-group validity and construct validity. The internal consistency and test-retest reliability of the two versions are presented in Table 3, with the 95% CI of all the reliability indices ≥0.70. For known-group validity, both versions showed similar characteristics of higher MM-CGI scores in spouses and in later stage of dementia (see Fig. 2 for MM-CGI; and Fig. 3 for MM-CGI-SF). For construct validity (Table 4), both language versions correlated strongly (Spearman’s ρ> 0.50) [44] with the total scores of caregiver burden (ZBI) and depression (CES-D) (ρ=0.61–0.77); and less strongly (ρ≤0.50) [44] with Finance subscale of ZBI, Positive Affect subscale of CES-D and Interpersonal Problems subscale of CES-D (ρ=0.16–0.49).

The findings from this study have several implications. With the availability of the Mandarin-Chinese version of MM-CGI, we can now expand the reach of PDG detection to family caregivers who use Mandarin-Chinese as their primary language. This is noteworthy as the Mandarin-Chinese language is used by approximately one-sixth of the world’s population [18], and Chinese-speaking populations have among the largest number of people living with dementia [1]. Detection of PDG provides an added dimension to supporting the emotional needs of family caregivers who are often the primary persons that provide most of the informal care to PWD [45] and thus can benefit from appropriate interventions for PDG.

The availability of the Mandarin-Chinese MM-CGI opens the way for further research to evaluate the efficacy of PDG interventions in dementia caregiving. To date, no intervention-studies have been conducted in the Chinese-speaking populations to address those with high PDG, possibly hindered by the lack of a relevant Chinese scale to assess PDG. Although prior studies have demonstrated the benefits of PDG interventions [7, 46, 47], they were primarily conducted in Caucasian populations. We remain uncertain whether PDG interventions are equally effective in the Chinese community, and whether there is a need to adapt the interventions to ensure that they can be culturally-relevant and sensitive to Chinese caregivers. In designing future intervention-studies for the Chinese community, researchers may possibly take reference from prior literature in this area [46‐49] and include some of the commonly employed strategies such as: (1) encouraging caregivers to tell the story of the PWD through which they are helped to identify and process the painful emotions associated with the loss; (2) normalizing their emotions related to the loss and findings ways to remain connected with other caregivers; (3) educating caregivers on new ways to remain connected to the PWD such as through spiritual practices, celebrations, humor, life review and therapeutic touch; and (4) focusing on anticipatory grief and preparing the caregivers for potential future losses.

Measurement equivalence between the Chinese and English MM-CGI attests to the cross-cultural compatibility of the scale. Even though the Chinese and English languages are vastly different in their writing system (logogram- versus alphabet-based) and sentence structure, it was still possible to establish the measurement equivalence of the two versions. This supports the universal experience of PDG in caregivers of PWD [8, 38], and affirms the applicability of the PDG construct in the Chinese population. Indirectly, the measurement equivalence also provides evidence of the linguistic and cognitive equivalence between the Chinese and English versions [50]. It lends credence to the qualitative methods that were used to translate the Chinese version in this study.

Establishing measurement equivalence lays the ground for future research on the cross-cultural aspects of PDG in dementia caregiving. Although prior literature has suggested the influence of culture on the expression of grief [21, 38], we have limited understanding to date on how the characteristics of PDG vary across ethnic groups to affect dementia caregiving. It is now possible to use MM-CGI as a valid measure to compare PDG between the English- and Chinese-speaking populations because differences in PDG scores between the populations can be attributed to true variations in cultures per se, excluding scale language as the cause of differences. This is pertinent in the light of recent evidence demonstrating the influence of scale language on individuals’ responses to culture-related questionnaires [20].

Finally, evidence of equivalence justifies the pooling of data from the two language versions to improve statistical power and representativeness of future research on PDG. This can be especially relevant in multi-ethnic populations such as ours where the use of vernacular prevails in different subgroups of the population. Using this study as an illustration, as seen in Table 1, the English version captured a subpopulation which was younger and more educated, while the Chinese version comprised those who were older and less educated. Combining data from both language versions is thereby more representative of the overall population of caregivers.

Some limitations of the study should be noted. First, the participants were recruited from tertiary dementia services and may not fully represent those in the community. However, this is less likely a problem considering that most of the PWD in Singapore still receive their dementia care from tertiary centers, and the two recruitment centres in this study are the only two dementia services that serve the population in the North-East of Singapore. Second, the proportion of spousal caregivers in this study were much lower than that of the children caregivers (Table 1). Although it is possible that spousal caregivers may have been under-represented, this is less likely considering our proportion of spousal caregivers (13.7%) were not much different from the 16.0% reported in a separate study based on nationally representative samples [51]. Third, we had a smaller sample size for participants who completed the Chinese version of MM-CGI (n = 103). At the point of recruitment, our participants were given the choice to complete either the Chinese MM-CGI (from the current study) or the English MM-CGI (from our separate study [8]). Only about one-third of our participants chose to complete the Chinese version of MM-CGI and these participants were more likely to be older and less educated. This is not inconsistent with what we would have expected of the population in Singapore [52] – the younger population is typically more educated and bilingual (competent in English and Chinese), and has a preference for the English language; while the older population is less educated and usually more competent in Chinese than English. However, the smaller sample size in the Chinese version (n = 103) resulted in some limitations in our analyses. Ideally, we would have evaluated the factor structure of Chinese MM-CGI. However, such analysis is less recommended when the sample size is around 100 or less because it may produce improper solutions and unstable parameter estimates [53]. Fourth, as the scales were self-administered, it is possible that caregivers with lower literacy were under-represented. Fifth, the stage of dementia was assessed only based on self-reports by family caregivers, which may lack the precision of other staging instruments that also encompass objective assessments by the healthcare professionals. Sixth, our definitions on the equivalence-margins required some subjectivity in deciding on the values that would be small enough to reasonably claim equivalence, as there were no previous studies to provide guidance. Regardless, we have stated the rationale for our choice of the equivalence-margins to allow readers to judge whether these values are reasonable. Last, the MM-CGI-SF in this study was administered as part of MM-CGI and it might have different properties had it been administered separately on its own.

In summary, this study has produced the Mandarin-Chinese version of MM-CGI and demonstrated its measurement equivalence, as compared to the English version, in detecting PDG. It opens the way for PDG assessment and intervention in Chinese-speaking caregivers, as well as cross-cultural research on PDG in dementia caregiving.