Trajectories of negative and positive experiences of caregiving for older adults with severe dementia: application of group-based multi-trajectory modelling

Dementia is a public health priority affecting more than 55 million people worldwide [1]. Globally, it is one of the major causes of disability and dependency among older adults and is the seventh leading cause of death [2]. Community-dwelling older adults with severe dementia (henceforth referred to as older adults in this manuscript) are highly dependent on their family caregivers [3]. These family caregivers not only report negative experiences of caregiving including adverse impacts on their schedule, health, and finances [4‐6], but also positive experiences of caregiving such as personal growth, gain in relationships, and spiritual growth as a result of caregiving [7, 8]. It is thus imperative that they be jointly examined to fully understand the holistic experience of caregiving for older adults, and guide interventions to improve the caregiving experience.

A plethora of literature exists regarding the experiences of caregivers of older adults, however with limitations. First, most of these studies are cross-sectional surveys [5, 6, 9, 10] or one-time qualitative interviews [11‐13]. Given the progressive nature of severe dementia [9, 14], longitudinal rather than cross-sectional studies can help understand the dynamic nature of the caregiving experience. Second, some longitudinal studies conducted with caregivers of older adults have described an “average” trajectory for either the ‘negative’ [15‐17], or the ‘positive’ caregiving experience [15, 16]. There has also been an assessment of heterogeneity between caregivers in their trajectories of negative experiences of caregiving [11, 14, 18], but scant literature exists in the context of heterogeneous trajectories for positive experiences of caregiving [19, 20]. Third, most of these studies have focused on persons with mild or moderate dementia, with either no or limited numbers of caregivers of older adults with severe dementia [14, 21]. A focus on older adults with severe dementia is relevant given that their caregivers will have more negative experiences of caregiving [14, 22]. Last, most existing studies have considered negative and positive experiences of caregiving separately. A joint longitudinal assessment of heterogeneity in trajectories for negative and positive experiences of caregiving is yet to be conducted, despite the evidence that such experiences co-exist throughout caregiving. A previous study conducted by members of our team among caregivers of patients with metastatic cancer delineated the trajectories for negative and positive experiences of caregiving and found that, though these trajectories varied among caregivers, they remained largely constant with time [20].

One approach to jointly assess the trajectories for negative and positive experiences of caregiving is the group-based multi-trajectory model (GBMTM) [23]. GBMTM is a statistical approach for finite mixture modelling, i.e., it is designed to identify a finite number of groups of individuals following similar trajectories over time, and allows the analysis of the interrelationship of multiple indicators for an outcome of interest - in this case, the negative and positive experiences of caregiving [24].

Existing literature provides some indication of sub-groups of caregivers at risk of experiencing higher levels of negative and positive experiences of caregiving. Caregivers of older adults with functional and cognitive limitations [25‐27], behavioural and neuropsychiatric symptoms [14, 26, 28, 29], lack of ability to communicate [30], and those with low socio-economic status [29], greater financial burden [31] or provide longer hours on caregiving [32] report more negative experiences of caregiving. Furthermore, caregivers co-residing with older adults report both negative and positive experiences of caregiving [10, 31]. Although there is little empirical literature linking resilience with negative and positive experiences of caregiving, some qualitative studies suggest that more resilient caregivers are likely to experience positive emotions during caregiving, to consider caregiving as part of their duty, and to maintain their social relationships despite their caregiving activities [33]. On the other hand, caregivers receiving additional support in caregiving activities and having a strong family network have lower levels of negative but higher levels of positive experiences of caregiving [25, 31].

Experiences of caregiving also influence caregivers’ outcomes after the death of the older adult. When death is preceded by a prolonged and intense period of caregiving, as is the case in dementia caregiving, many caregivers experience an end to the stressors of caregiving and may report improved psychosocial outcomes (relief or stress reduction theory) after the death of the older adult [34]. As caregiving burden ends, caregivers may experience relief and they may be able to pursue their previously neglected social and employment roles, self-care, and leisure activities, thus allowing them to adjust better to a life without the older adult [35, 36]. In contrast, the resource depletion theory suggests that some caregivers may instead fare worse after the death of the older adult [37]. Exposure to caregiving stress over a long period of time may diminish caregivers’ coping resources over time, and their ability to have a sense of closure following older adults’ death. When coping resources are diminished, caregivers may have difficulties adjusting during bereavement, resulting in distress and complicated grief [35]. The association between negative experiences of caregiving and caregivers’ outcomes after the death of the older adult thus remains unclear. Literature also shows that caregivers reporting higher levels of positive caregiving experiences may have worse outcomes after the death of the older adult [34, 38]. These caregivers may be deriving a sense of satisfaction from their caregiving experience and the loss of this role may lead to a loss of purpose for them. However, most studies are limited to examining caregivers’ grief and psychological distress after the death of the older adult [34, 37, 38]; overall quality of life and spiritual well-being, have been less examined [39, 40]. None have examined the relationship between trajectories of negative and positive experiences of caregiving and caregivers’ outcomes after the death of the older adult. Greater clarity regarding this relationship would enable early identification of caregivers likely to experience worse outcomes after the death of the older adult.

To address these gaps, our first aim was to use GBMTM to delineate joint trajectories for negative and positive experiences of caregiving in a sample of family caregivers of community-dwelling older adults with severe dementia in Singapore. Our second aim was to identify the risk factors associated with membership of each joint trajectory. We hypothesized (H1) that caregivers of older adults with more behavioural symptoms, and functional and cognitive impairment will be more likely to follow joint trajectories representing higher levels of negative and lower levels of positive experiences of caregiving. We also hypothesized (H2) that caregivers co-residing with older adults and paying for older adults’ treatments will be more likely to follow joint trajectories representing higher levels of negative and positive experiences of caregiving; while those receiving caregiving help from a migrant domestic worker (live-in, full-time domestic workers, mostly women from neighbouring low-income countries), receiving emotional support from family, having greater resilience and being spousal caregivers will be more likely to follow joint trajectories representing lower levels of negative and higher levels of positive experiences of caregiving. Our third aim was to assess the association between the joint trajectories and caregivers’ outcomes six months after the death of the older adult. We hypothesized (H3) that caregivers following joint trajectories representing higher levels of negative and positive caregiving experience would report worse outcomes after the death of the older adult including higher grief and distress, and poorer overall quality of life and spiritual well-being.

A total of 215 caregivers of older adults participated in the study. Of the 80 older adults (37%) who died within the study period, caregivers of 76 older adults (95%) were surveyed 6 months after older adults’ death.

Table 1 shows the sample characteristics measured at baseline. Older adults were aged 82.9±8.1 years, and the majority were females (77%). Among the caregivers, 83% were adult children of the older adults, nearly two-thirds (74%) co-resided with older adults, and the majority received caregiving support from a migrant domestic worker (79%). On an average 5.2±4.4 hours were spent daily on caregiving. Nearly half the caregivers (48%) used their own Medisave to pay for older adults’ treatment expenses and 58% of the caregivers reported receiving strong emotional support from their family.

GBMTM model was used to identify the 4-group trajectory model describing the positive and negative experiences of the caregivers. Model fit indices are provided in the supplement (Supplementary Table 1). The 5-group joint trajectory model had the lowest BIC value; however, we selected the 4-group model as the percentage change in the BIC from a 4 to 5-group model was less than 1%, therefore, a parsimonious 4-group model was preferred.

Based on the relative levels of positive and negative experiences of caregivers across the delineated four trajectories as shown in Fig. 1, we named these joint trajectories as ‘very high positive, low negative’, ‘high positive, moderate negative’, ‘very high positive, moderate negative’, and ‘high positive, high negative’ (Table 2). The trajectory ‘very high positive, low negative’ (22.7% of caregivers) had the highest positive and lowest negative experience of caregiving and was used as a reference category in our analysis. The average posterior probability of being in each trajectory was >0.8.

Using prospective data, our study is the first to assess heterogeneity in trajectories for negative and positive experiences of caregiving over 2 years among caregivers of older adults with severe dementia. We identified four joint trajectories – “very high positive, low negative”, “high positive, moderate negative”, “very high positive, moderate negative” and “high positive, high negative”.

Consistent with a previous study, we found that experiences of caregiving, both negative and positive remained stable over time [20]. It is possible that caregivers in our study sample had been caring for older adults for the past several years and since the onset of dementia. Therefore, these caregivers were unlikely to make changes in how they appraise their experience after several years of caregiving, explaining why reports of both negative and positive experiences of caregiving remained stable over time.

Notably, our results showed that although negative experiences of caregiving varied more widely among caregivers, caregivers’ positive experiences remained generally high. This may reflect the specific cultural context in which this study was conducted. In Asian cultures, adult children (constituting 83% of our sample) are bound by the tradition of filial piety and consider caring for older family members to be their responsibility [13]. As a result, despite experiencing high levels of stressors, these caregivers may derive a sense of purpose and satisfaction in caring for the older adult.

As hypothesized, caregivers of older adults with more behavioural symptoms were more likely to belong to the trajectories representing moderate negative experiences of caregiving. Behavioural symptoms may be difficult for caregivers to manage, potentially resulting in embarrassing or abusive situations [59], physical and psychological morbidity, and social isolation for the caregiver [4]. Not surprising, the literature has been largely consistent in showing that behavioural symptoms increase negative experiences of caregiving [14, 26, 28, 29].

Consistent with the previous studies [10, 31], we found that caregivers co-residing with older adults reported more negative but very high positive experiences of caregiving. Further, caregivers spending more time on caregiving reported moderate to high negative experiences of caregiving, as shown previously [32]. We also found that caregivers who paid for treatment using their savings were more likely to follow “high positive, moderate negative”, and “very high positive, moderate negative” versus “very high positive, low negative” trajectory. While being able to help older family members with dementia in receiving treatments may improve caregivers’ self-esteem and satisfaction from caregiving, given the high out-of-pocket costs in Singapore and the long treatment period for dementia, many caregivers may experience financial burden and loss of their healthcare savings, thereby reporting more negative experiences of caregiving [12].

Our findings show that caregivers with greater resilience were more likely to follow a trajectory with a higher positive and lower negative caregiving experience. Research shows that more resilient caregivers are able to adapt and recover from the physical and psychological demands of caring for older adults [60]. Studies show higher levels of resilience to be associated with the use of more positive coping strategies, greater self-efficacy, and lower burden and stress [33, 61]. Given the positive association between resilience and caregivers’ experiences, future research should aim to develop and evaluate interventions to increase caregivers’ resilience [62, 63].

Study results suggest that caregivers need both emotional and instrumental support to cope with caregiving activities. While emotional support comes from having other supportive family members, instrumental support is primarily provided by a migrant domestic worker. Having a migrant domestic worker enables caregivers to have lower negative experiences of caregiving albeit with lower positive caregiver experiences. It is remarkable that 79% of caregivers in our sample received help from a migrant domestic worker and reflects the high contribution of these workers in caregiving activities. This arrangement allows the older adults to be cared for at home while enabling the informal caregiver to pursue employment and engage in other activities.

Our results showed that compared to the very ‘high positive, low negative’, trajectory, caregivers belonging to trajectories representing more negative caregiving experiences (‘very high positive, moderate negative’, and ‘high positive, high negative’) experienced worse bereavement outcomes. These results are consistent with the ‘resource depletion theory’, which suggests that a high and sustained level of caregiver stress accumulated through negative experiences of caregiving may diminish caregivers’ coping resources over time [37]. Depletion of coping resources may leave the caregiver more vulnerable after the death of the older adult and interfere with grief resolution. Resultantly, these caregivers experience greater distress, lower spiritual well-being, and lower quality of life at 6 months after the death of the older adult. Notably, we find that these worse bereavement outcomes may happen despite caregivers experiencing high or very high positive caregiving experiences. This is consistent with previous studies suggesting that many caregivers with complicated grief report having positive caregiving experiences [38].

The main strength of our study is that it is a longitudinal prospective study with multiple assessments conducted over 2 years. Our findings add to the growing body of literature on caregivers’ negative and positive experiences, examining them jointly and longitudinally, and assessing their impact on caregivers’ outcomes after the death of the person. Our study also has limitations. Firstly, there was missing data at each time point. It is possible that missing data was not at random and may represent caregivers who did not complete the study because of older adults’ sickness or death. However, the trajectory analysis employed uses full-information maximum likelihood to handle missing data, which is more efficient than other ways of handling missing data [64‐66]. Data from all caregivers regardless of whether they had missing information were used to estimate the model [67]. Secondly, the scales used to measure negative and positive experiences of caregiving do not have cut-offs for clinically meaningful values, therefore we are unable to comment on that. Thirdly, our results on the association between experiences of caregiving and caregivers’ outcomes after the death of the older adult are based on a sub-sample of caregivers whose older adult had died during the study duration. Lastly, given the unique cultural context in Singapore, the generalizability of our findings needs to be tested in different settings.

Our study results provide some indications on how to improve caregivers’ experiences. Although positive experiences of caregiving were generally high, there is potential to further improve caregivers’ sense of mastery and enable them to find meaning in their caregiver role. Training caregivers to anticipate and manage older adults’ behavioural symptoms can reduce negative experiences of caregiving. Evidence also suggests that training family caregivers to use non-pharmacologic strategies is more effective in reducing these symptoms than pharmacologic interventions [68]. Additionally, training caregivers to be resilient may prevent worse outcomes. Evidence suggests that interventions such as resilience training, cognitive behavioural therapy, mindfulness, and promoting the use of digital resilience monitoring tools in clinical settings could help strengthen the resilience and well-being of the caregivers of older adults [69, 70]. Further, social support interventions such as peer-support groups, befriending schemes, family support, and remote internet and technological support may reduce the psychological burden of the caregivers by protecting them against social isolation and loneliness experienced in the process of caregiving [71].

The experience, over time, of caregiving for older adults with severe dementia is heterogenous, with varying extent of negative and positive experiences. Health and/or social service practitioners working with family caregivers of older adults with severe dementia should be mindful of this heterogeneity. Modifiable risk factors for trajectories involving negative experiences of caregiving identified in this study, like more behavioural symptoms of older adults with severe dementia, co-residence, and financial and emotional support available to caregivers, can be considered as targets in future interventions. With caregivers having a “high positive, high negative” trajectory (vs “very high positive, low negative”) expressing greater grief and distress, and lower spiritual well-being and quality of life six months after the death of the older adult, it is important to continue to care for (past) family caregivers even after the end of their caregiving role due to the death of the older adult.