Abstract
In response to the dominant narrative that informal care episodes become increasingly heavy over time, this study empirically identifies distinct informal care trajectories in terms of developments in care demands (care receivers’ health), intensity (number of hours) and complexity (number and type of tasks) over the course of a care episode. Latent class analysis on 5,595 informal care episodes among 2,884 Dutch carers identified a ‘stable trajectory’ (41 per cent) and ‘decrease trajectory’ (22 per cent) apart from an ‘increase trajectory’ (37 per cent). The trajectories were related to care receivers’ characteristics (age), carers’ difficulty of saying ‘no’ (relationship and obligation) and changes in the care context (living situation).
Introduction
The dominant narrative about the path informal carers follow over the course of time they help or care for their loved one is one of increasing care demands, care intensity and complexity (Schulz et al, 2020). A typical example is that care provision starts with doing grocery shopping once a week for one’s frail mother but gradually develops into daily assistance with dressing and eating while one’s mother is becoming increasingly dependent. Another example is when the progressing Alzheimer’s disease of one’s spouse means that the carer who only assisted with small things in the beginning is eventually occupied with the care receiver 24/7. This study presents an empirical description of the trajectories informal carers experience to find out whether this dominant narrative is the standard path or whether other paths are common as well. In addition, we will assess in which situations which trajectories are more likely. We will differentiate trajectories on the basis of three ‘ingredients’: changes in care demands (that is, the care receiver’s health condition), care intensity (that is, hours of care provision) and care complexity (that is, number and type of tasks) between beginning and end of the period in which informal care is provided. The three ‘ingredients’ together express developments in how heavy the care situation is for the carer.
The dominant narrative is presented explicitly in the review article by Schulz and colleagues (2020). They describe the prototypical longitudinal trajectory for older adult carers: it starts with sporadic assistance with small tasks and evolves into more intensive practical help and, later, personal or medical care when the care receiver’s care needs increase. Their model emphasises the cumulative nature of the progression of tasks for the carer: ‘With increasing care recipient disability and need for care, the carer’s role becomes more labor and time intensive, more complex, and increasingly stressful’ (Schulz et al, 2020: 638). The issue of progression is also present in the highly influential stress process model (Pearlin et al, 1990). This model explains carers’ stress by developments over the course of the caregiving experience. The care receiver’s care needs and the amount and complexity of the care provided in response to those (together labelled as ‘primary stressors’) are assumed to become more serious over time and to fatigue carers in the long run. As such, after a while, primary stressors trigger secondary stressors (such as role strain and intra-psychic strains), and together they fuel carers’ stress levels. Similarly, in the well-known wear-and-tear hypothesis (Townsend et al, 1989), next to the depletion of carers’ physical and psychological resources, increasing care demands is presented as a reason why carers’ well-being deteriorates over time. In sum, the dominant view is that the typical trajectory of a carer is one in which the care burden becomes heavier over time: carers are confronted with increasingly serious health conditions in the care receiver, which raises anxiety and grief, together with increasing efforts in their care provision, both in terms of the time spent on it and the difficulty of tasks they perform.
An important reason why this trajectory is the dominant view is that much research on informal care focuses on care provided to older adults or people with Alzheimer’s disease, and the progressive nature will be particularly common in these groups. This focus can be understood from the large impact of the field of gerontology on the caregiving literature, as well as the high (and increasing) incidence of these types of care situations. Nevertheless, it captures only a selection of all care situations. Other realistic options are trajectories that end due to recovery (Uccheddu et al, 2019), such as care provision in response to temporary injuries from an accident, recovery from surgery or cancer treatment. Also, the care process may look very different if the care receiver is not an older person but, for instance, a child with lifelong disabilities who starts living independently from their caregiving family once grown up. In this study, we embrace a wider scope of informal care situations, including the provision of help and care to a partner, family, friend, neighbour, acquaintance or colleague who is in need of help because of physical, psychological or mental limitations, or problems related to old age. The care receiver can be of any age and can suffer from different kinds of conditions, and the carers are not limited to typical age or gender categories. We expect that the inclusion of a wider range of care situations – a range more accurately reflecting existing care situations in society – will show more diverse trajectories than that in the dominant narrative.
We argue that it is important to examine whether diversity in care trajectories exists. First, a sole focus on the increasingly heavy trajectory may foster a too negative view of informal caregiving. Caregiving is known to be potentially burdening (Pinquart and Sörensen, 2003), but this is likely especially the case in the increasingly heavy trajectories. Second, it is helpful to learn whether other paths also exist because more diverse informal care situations are to be expected in present-day society. Societies face a growing number of older people and chronic illnesses, both of which will increase the need of care (Colombo et al, 2011). With retrenching governments, more care provision is expected from informal carers (Verbeek-Oudijk et al, 2014). Care situations may therefore become more diverse, with a stronger call to help not only close kin but also non-kin, such as neighbours or friends. This call may also diversify the pool of informal carers in terms of their life stages and the care they can or want to provide.
On top of insights into the diversity of trajectories carers experience, we argue that it is important to find out in which situations which trajectories are more likely. Such knowledge helps to identify high-risk informal carers and is therefore useful in targeting support to carers. We will present theoretical propositions from which we derive hypotheses about the associations between the characteristics of the care receiver, carer and care context, on the one hand, and the type of trajectory, on the other. The theoretical arguments used refer to the expected health progression of the care receiver, the difficulty of saying ‘no’ for the carer and changes in the care context.
In sum, the aim of this study is to theorise and empirically show which types of trajectories can be identified in terms of developments in care demands, intensity and complexity over the course of a care episode. We will answer two research questions:
Which types of trajectories in terms of care demands, care intensity and complexity can be distinguished in informal care episodes?
To what extent do the different trajectories relate to characteristics of the care receiver, carer and care context?
We will use unique data from the Netherlands that capture the complete informal caregiving careers of 2,884 respondents, covering 5,595 care episodes (Verbakel and CentERdata, 2021). To identify distinct trajectories, latent class analysis (LCA) will be applied with input variables for developments in (1) care demands (care receivers’ memory, psychological and physical conditions), (2) intensity (hours of informal care) and (3) complexity (number of tasks and transitions towards more demanding tasks). We choose this strategy because clustering these elements together reflects the ‘total picture of heaviness’ that carers are confronted with. Care demands add to heaviness because they are associated with carers’ worries and grief; it hurts to see a loved one in ill health. Carers’ efforts add to heaviness because intensive and complex caregiving requires energy and time that cannot be spent in other life domains.
Our contributions to the literature are threefold. First, to our knowledge, an empirical description of informal care trajectories among a broad sample of carers has not been presented before, while it helps us to reflect on the dominant narrative of increasingly heavy caregiving paths. Second, we will contribute theoretically by formulating and testing theoretical propositions that help us understand why certain caregiving paths become increasingly heavy, whereas others do not. Finally, we want to add to the literature that calls for ‘moving beyond snapshots’ (Moen and DePasquale, 2017; Keating et al, 2019). In line with this call, we argue that to better understand carers and their outcomes, it is not enough to know their current situation; the process is important too. With our approach, we can show the heterogeneity of care trajectories.
Expectations on care trajectories in care demands, intensity and complexity
Besides the dominant narrative of ‘increasingly’ heavy care situations, we expect at least two other types of trajectories: those with ‘decreasing’ and those with ‘stable’ care demands, intensity and complexity. Mixed trajectories in which developments in care demands do not parallel with developments in intensity or complexity are possible too. To limit complexity, we will focus our theoretical expectations on the three baseline trajectories that can be expected. In the following, we will describe three arguments to hypothesise in which situations each of these trajectories may be expected. The arguments relate to the care receiver, carer and care context, respectively.
Health progression of the care receiver
The first argument relates to the expected health progression of the care receiver. Health deterioration is naturally related to care receivers’ age, though diversity in health among older adults is strong too (World Health Organization, 2015). Generally, deterioration of health can be expected with care receivers’ age, whereas stable health situations or even recovery from health problems are more likely at younger ages. This pattern results from: the biological ageing process, with the gradual accumulation of molecular and cellular damage leading to the impairment of body functions and lower capacities to counter diseases; the different types of health problems that manifest themselves at different ages; and the higher incidence rate of multimorbidity at higher ages (World Health Organization, 2015). Therefore, we expect to find the ‘increase trajectory’ among carers who care for older care receivers and a ‘stable trajectory’ but especially ‘decrease trajectory’ among carers of younger care receivers.
Difficulty of saying ‘no’
The second argument relates to how difficult (or easy) it is for the carer to resist responding to care demands by means of putting in extra hours (that is, higher intensity) or to accept more and more difficult tasks (that is, higher complexity). It can be argued that saying ‘no’ is more difficult the closer the relationship between the carer and care receiver and the stronger the carer’s feeling of obligation to care.
As for the relationship, commitment and loyalty are generally strong for close kin (De Klerk et al, 2021), and people feel less of a choice when a close family member is in need of care than other relatives or non-relatives (Schulz et al, 2012). A feeling of obligation partly emerges from care norms set by others (the care receiver or the wider context) which prescribe that one should (always) provide help, so-called ‘normative obligation’ (Broese van Groenou et al, 2013). The (perceived) prospect of negative social sanctions makes people conform to such norms and hence make it difficult to act against them (Bicchieri, 2006). Another source of a felt obligation to care is the absence of alternative care providers, also labelled ‘structural obligation’ (Broese van Groenou et al, 2013). Such absence increases the perceived pressure that it is the carer who should respond to care demands. Saying ‘no’ would directly imply that the care needs of the care receiver remain unmet. This feels ethically unacceptable to many carers, which feeds their feeling of obligation to keep providing (more) care.
Based on these assumptions, we expect that the ‘increase trajectory’ is more likely if the carer provides care to close kin and if they feel high levels of obligation to care. We expect the ‘decrease trajectory’ or ‘stable trajectory’ to be more likely if the carer provides care to non-kin or less-close kin and if they feel low levels of obligation.
Changes in the care context
The third and final argument relates to changes in the wider care context in which the carer provides care. Prominent in this respect are changes in the living situation of the care receiver. Such changes relate to care trajectories for two reasons: they may have direct implications for carers’ care intensity and complexity because the care context is indicative of the availability of other (formal or informal) carers; and they may reflect developments in the health condition of the care receiver.
Generally, care intensity and complexity are strongest if care receivers live in the same household as the carer (Tennstedt et al, 1993). The carer is then often the primary carer, and in extreme cases, co-residence implies a 24/7 care situation. If the care receiver moves out of the carers’ home into an independent household, a reduction in care intensity and complexity for the carer can be expected simply because the care receiver is not around all the time. In addition, a move into an independent household signals that the care receiver’s health conditions or independence have improved. Therefore, we expect that the ‘decrease trajectory’ is more likely in care episodes in which the care receiver moves into an independent household.
The same reasoning can be used for situations in which the care receiver moves to a care facility. Since 24/7 formal care is available in care facilities, the carer is relieved in terms of intensity and complexity of care (though this does not mean that no informal care is needed anymore) (Gaugler, 2005). However, since we will assess developments in care episodes in this study by comparing the beginning and end situation (and not the situation directly before and after institutionalisation), it is likely that intensity and complexity have increased over time. This is because moving to a care institution is linked to the deteriorating health of the care receiver, which also assumes increasing demands. In all, we therefore expect that care episodes in which the care receiver moves to a care facility are more likely to follow an ‘increase trajectory’.
Another possible change in living situation is when care receivers first lived in an independent household (that is, not in the same household as the carer) with others but ended up living alone (that is, because of their partner’s death or divorce). A care situation like this likely implies that the care load on the shoulders of the carer increases, not only because there is no household member present to provide (immediate) care but also because a stronger urge is felt to provide company to counter loneliness. We expect the ‘increase trajectory’ to be more likely in care episodes in which the living situation of the care receiver changes from living in an independent household with others to living alone.
If no changes in the care receiver’s living situation occur during the care period, there is relatively less need to adjust care intensity and complexity, and it is less likely that the health condition of the care receiver will have deteriorated or improved substantially. Therefore, the ‘stable trajectory’ is expected to be more likely if no changes in living situation take place.
Method
Data
We used data from the Retrospective Informal Care Careers module in the Longitudinal Internet Studies for the Social Sciences (LISS) panel collected in March 2020 (that is, 2021). The LISS panel is based on a random sample of Dutch households, with an overall response rate of around 73 per cent (see: www.lissdata.nl). The Retrospective Informal Care Careers module was fielded among respondents aged 16 to 78 who, in a screening questionnaire in January 2020, indicated providing or having previously provided informal care. The response rate was 87.5 per cent, resulting in a sample of 3,061 respondents. For a maximum of seven care episodes, the online survey posed retrospective questions about the start date of the episode, the end date (if the episode was completed) and the relationship with the care receiver.1 More extensive information about the care episodes (for example, the care receiver’s health conditions, caregiving hours and caregiving tasks at beginning and end) was asked for a maximum of three randomly selected care episodes. Respondent information (for example, age and sex) was retrieved from the background module of the LISS panel of March 2020. We selected the 5,837 care episodes of 2,999 carers for which the extensive questionnaire was completed. Episodes with missing information on any of the variables were dropped (N = 242 episodes [4.2 per cent]), resulting in a final analytical sample of 5,595 care episodes, nested in 2,884 carers. In our sample, information was available on one episode for 40.1 per cent of carers, on two episodes for 25.8 per cent and on three episodes for 34.1 per cent.
Measurements
The first step of our study is to identify care trajectories in terms of developments in care demands, intensity and complexity (see the first research question). For that, we constructed seven input variables to be used in an LCA. Since we were interested in developments over the course of a care episode, the input variables’ focus was on change. For that, we compared information retrospectively provided by the respondent referring to the care episode’s beginning and end. If the care episode was still ongoing at the moment of survey, we compared the beginning of the situation with the current situation. Our measurements were defined at the level of the care episode.
Three elements were considered to assess changes in care receivers’ health condition. First, a change in the care receiver’s memory problems was based on the respondent’s assessment of to what extent the care receiver had problems with memory, attention span and thinking (answering categories: 1 = no problems; 2 = some problems; 3 = severe problems). Comparison of this assessment at the beginning and at the end of the care episode resulted in three categories: a decrease (5 per cent), stability (68 per cent) and increase in memory problems (27 per cent) (regardless of the level of memory problems at the beginning of the episode). Second, change in the care receiver’s psychological problems was based on the respondent’s assessment of the extent to which the care receiver suffered from psychological problems, such as depression or phobia, at the beginning and end of the care episode (answering categories: 1 = no problems; 2 = some problems; 3 = severe problems). Psychological problems decreased in 8 per cent of the episodes, were stable in 76 per cent and increased in 16 per cent. Third, change in the care receiver’s physical problems was proxied by the number of (instrumental) activities of daily living ([I]ADLs) that the care receiver was unable to perform independently at the beginning and the end of the episode, according to the respondent. The list of (I)ADLs presented to the respondent included walking, getting dressed, eating, housekeeping, grocery shopping and preparing a meal. Three categories were constructed, based on the difference between the beginning and end of the episode in the sum of activities the care receiver was unable to perform independently: decreased (18 per cent), stable (47 per cent) and increased physical problems (35 per cent).
Developments in care intensity during the care episode were captured by the change in the number of informal caregiving hours. This variable was constructed with the self-reported average number of hours per week spent on informal caregiving at both the beginning and the end of the episode (rounded in whole hours, with a maximum of 168). We defined three categories: a decrease (27 per cent), no change (42 per cent) and an increase in informal caregiving hours (31 per cent). We decided that a one-hour change between beginning and end was sufficient to be defined as a change because the questions in the survey on caregiving hours at the beginning and end were graphically offered directly next to each other. Even a one-hour difference therefore likely reflected a conscious assessment of change by the respondent.
Changes in care complexity during the care episode were captured by assessing differences in both the number and the type of tasks. A higher number of tasks reflects more variety, and hence complexity, in the care provided; the type of tasks captures qualitative differences, with some tasks (for example, medical care) being more complex than others (for example, provide company). A list of caregiving tasks was presented to the respondent, and for both the beginning and end of the episode, they were asked whether they performed the particular task. Tasks included were provide company or emotional support, provide transport, accompany to doctor visits, administrative help, help with housekeeping, provide personal care (for example, showering or visiting the toilet), provide nursing care (for example, administering medicine or tending to wounds), organise and coordinate caregiving, and provide other help. Change in the number of tasks was assessed by comparing the sum of tasks at the beginning and end of the episode. We constructed three categories: a decrease (26 per cent), no change (54 per cent) and an increase in number of tasks (20 per cent). The change in type of tasks was measured by two specific transitions that arguably most clearly indicate a change in complexity and were common enough in our sample: (1) the transition from providing only emotional support and company at the beginning of the episode to doing other tasks at the end of the episode (regardless of whether the respondent was still giving emotional support or company) (3 per cent); and (2) the transition to personal and nursing care (7 per cent), meaning that at the beginning of the episode, the respondent did not provide personal and/or nursing care but did so at the end.
Finally, duration of the informal care episode was constructed by the number of months the episode had lasted. For episodes that were still ongoing, the survey date was taken as the end of the episode. Since the impact of duration likely does not follow a linear path, categories were constructed: up to six months (22 per cent), seven to 24 months (21 per cent), 25 to 60 months (21 per cent), 61 to 120 months (16 per cent) and over 120 months (20 per cent).
The second step in our study is to examine whether the identified trajectories relate to the characteristics of the care receiver, carer and care context (see the second research question). For that, we constructed four variables and three control variables. Age of care receiver at the start of the episode was measured as a continuous variable (mean = 61 years, SD = 22 years). There were 46 episodes with a negative age due to a rounding error in the construction of the year variable or minor inconsistencies in respondents’ answers. In these episodes, the age of the care receiver at the start of the episode was set to 0. For interpretability purposes, we divided the care receiver’s age by 10. The relationship with the care receiver indicated who the carer was caring for. We differentiated between partner (10 per cent), parent (43 per cent), child (6 per cent), sibling (7 per cent), other family member (12 per cent), friend, acquaintance or colleague (15 per cent), and neighbour (8 per cent). It should be noted that care for a child referred not to regular childcare but to informal care to a child with health issues. Feeling of obligation to care was based on the extent to which the reasons ‘it was expected of me’ (signalling normative obligation) and ‘there was no one else available’ (signalling structural obligation) played a role in the respondent’s care provision (answering categories: 0 = not; 1 = a bit; 2 = strongly; 3 = I do not know). ‘I do not know’ answers were set to missing. We took the average score of the two items (with a minimum requirement of one valid score), which ranged from 0 to 2 (mean = 1.86, SD = 0.65). Finally, we assessed change in the care receiver’s living situation by comparing the living situation at the beginning and end of the care episode (answering categories: 1 = in the same house as carer; 2 = in an independent household with other housemates; 3 = in an independent household without other housemates; 4 = in a care facility). Only those situations were defined that were theoretically interesting and occurred relatively frequently. The first four expressed stability: care receiver and carer remained living in the same house (14 per cent); the care receiver remained living in an independent household with others (29 per cent); the care receiver remained living in an independent household without others (26 per cent); and the care receiver remained living in a care facility (4 per cent). Next, we defined three relevant transitions: the care receiver lived in the same house as the carer at the beginning of the care episode but lived independently (with or without others) at the end (3 per cent); the care receiver did not live in a care facility at the beginning of the care episode but did so at the end (16 per cent); and the care receiver lived independently with others at the beginning of the care episode but without others at the end (4 per cent). Finally, we added a category with all transitions that were not explicitly coded as a separate category (3 per cent).
We controlled for the carer’s gender, contrasting men and women (60 per cent women). The carer’s age at the start of the informal care episode was coded into five categories to reflect different life stages in which the carer started providing care: younger than 24 years (13 per cent), 25–34 years (15 per cent), 35–44 years (18 per cent), 45–54 years (23 per cent) and 55 years or older (31 per cent). Finally, we controlled for historical time in which the care episode took place. We took the midpoint between start and end year of the care episode, which ranged from 1961 to 2020, and recoded so that score 0 referred to 1961. Table 1 shows the descriptive statistics of all measurements, which highlight the large variety of informal care situations included in our sample.
Descriptive statistics of input, independent and control variables
| Pro. | Mean | St.de. | |
|---|---|---|---|
| Input variables | |||
| Change in care receiver’s memory problems | |||
| Decrease | 0.05 | ||
| Stable | 0.68 | ||
| Increase | 0.27 | ||
| Change in care receiver’s psychological problems | |||
| Decrease | 0.08 | ||
| Stable | 0.76 | ||
| Increase | 0.16 | ||
| Change in care receiver’s physical problems | |||
| Decrease | 0.18 | ||
| Stable | 0.47 | ||
| Increase | 0.35 | ||
| Change in the number of caregiving hours | |||
| Decrease | 0.27 | ||
| Stable | 0.42 | ||
| Increase | 0.31 | ||
| Change in the number of caregiving tasks | |||
| Decrease | 0.26 | ||
| Stable | 0.54 | ||
| Increase | 0.20 | ||
| Transition to more than emotional support | 0.03 | ||
| Transition to personal/nursing care | 0.07 | ||
| Duration of care episode | |||
| 1–6 months | 0.22 | ||
| 7–24 months | 0.21 | ||
| 25–60 months | 0.21 | ||
| 61–120 months | 0.16 | ||
| 120+ months | 0.20 | ||
| Independent variables | |||
| Care receiver’s age at start episode (range: 0–105) | 60.97 | 21.96 | |
| Relationship: care receiver is … | |||
| Partner | 0.10 | ||
| Parent | 0.43 | ||
| Child | 0.06 | ||
| Sibling | 0.07 | ||
| Other family member | 0.12 | ||
| Friend, acquaintance, colleague | 0.15 | ||
| Neighbour | 0.08 | ||
| Feeling of obligation to care (range: 0–2) | 0.86 | 0.65 | |
| Change in care receiver’s living situation | |||
| Remained living with the carer | 0.14 | ||
| Remained living independently with others | 0.29 | ||
| Remained living independently without others | 0.26 | ||
| Remained living in care facility | 0.04 | ||
| From living with the carer to living independently | 0.03 | ||
| Moved to care facility | 0.16 | ||
| From living independently with others to without others | 0.04 | ||
| Other changes | 0.03 | ||
| Control variables | |||
| Female | 0.60 | ||
| Carer’s age at start episode | |||
| 24 years and younger | 0.13 | ||
| 25–34 years | 0.15 | ||
| 35–44 years | 0.18 | ||
| 45–54 years | 0.23 | ||
| 55 years and older | 0.31 | ||
| Historical time (1961 = 0, range: 0–59) | 49.92 | 10.05 |
Note: N = 5,595 episodes. Prop = proportion.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
Analytical strategy
Our analytical strategy consisted of two steps. First, we used LCA to identify different trajectories in the developments of care demands, intensity and complexity in care episodes. LCA is a statistical method to identify subgroups within populations that share certain characteristics in the response to several items that have categorical answer categories (Weller et al, 2020). With the poLCA package (Version 1.6.0) in R (Linzer and Lewis, 2011), we estimated several models with different numbers of classes (starting with a one-class model and increasing the number of classes one by one) and compared them to decide on the model that most accurately fitted the data. We assessed the model fit with different fit statistics: Bayesian Information Criterion (BIC), Akaike Information Criterion (AIC), G2 and log-likelihood (LL). We particularly focused on the BIC because this is considered to be the most reliable fit statistic (Weller et al, 2020). We determined the number of classes based on fit statistics in combination with the theoretical interpretability of the classes. To interpret the classes of the chosen model, we looked at each class’s mean values on the input variables. In addition, we reviewed classification diagnostics, that is, the average latent class posterior probability (ALCPP) and the entropy. The ALCPP reflects the average probability of the model accurately predicting the class membership for individuals. Relatedly, entropy indicates how accurately the model is in defining classes. Higher values of ALCPP and entropy reflect higher accuracy (Weller et al, 2020).2 After determination of the class solution, each respondent was assigned the class they most likely belonged to based on the highest estimated probability.
In the second step, we estimated multinomial logistic regression models (reported in Table A1 in Appendix 1) to test whether the trajectories identified in the first step were related to the characteristics of the care receiver, carer and care context. Since these characteristics were mutually related and partly led to opposite expectations, we included them simultaneously, together with the control variables. We accounted for the hierarchical data structure of care episodes nested in carers by including a random intercept in the model. To interpret effect sizes, we calculated predicted probabilities based on the regression models, with all other variables kept at their means (Mood, 2010). The predicted probabilities were contrasted to the grand mean to test for significant deviations from the mean.
Results
Identification of informal care trajectories
The fit statistics (see Table 2 and Figure 1) show that the fit of the models did not increase substantially when models contained more than four classes. For both the three-class and four-class solutions, the classification diagnostics were good. The ALCPPs amounted to 0.88 in the three-class model and 0.85 in the four-class model, reflecting high accuracy (Weller et al, 2020); the entropy was 0.76 and 0.77, respectively, which also showed sufficient accuracy (Weller et al, 2020). We choose the three-class solution over the four-class solution because the extra class in the four-class solution forms a subgroup rather than a distinctive class. We will elaborate on this further in the following.
Fit indices for latent class models with different numbers of classes
| Models | AIC | BIC | G2 | LL | Entropy | ALCPP |
|---|---|---|---|---|---|---|
| 1 class | 73661.44 | 73767.51 | 11684.61 | –36814.72 | – | – |
| 2 classes | 67584.57 | 67803.35 | 5573.75 | –33759.29 | 0.78 | 0.93 |
| 3 classes | 65576.73 | 65908.21 | 3531.90 | –32738.37 | 0.76 | 0.88 |
| 4 classes | 64733.46 | 65177.65 | 2654.64 | –32299.73 | 0.77 | 0.85 |
| 5 classes | 64488.86 | 65045.75 | 2376.03 | –32160.43 | 0.78 | 0.8 |
| 6 classes | 64299.30 | 64968.89 | 2152.47 | –32048.65 | 0.72 | 0.77 |
| 7 classes | 64138.39 | 64920.69 | 1957.56 | –31951.20 | 0.73 | 0.75 |
Note: N = 5,595 episodes.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
Model fit (BIC) for latent class models with different numbers of classes
Citation: International Journal of Care and Caring 2024; 10.1332/23978821Y2023D000000021
To interpret and label each class, we compared the mean values of the input variables between the three classes (see Table 3). For instance, the average occurrence of a decrease in memory problems was 0.22 in Class 1 versus 0.01 in both Classes 2 and 3. The three classes very clearly represented trajectories that can be characterised by, respectively, a decrease, stability and increase in care demands, intensity and complexity. We observed decreasing memory, psychological and physical problems among the care receivers (all indicators of care demands) to be over-represented in the ‘decrease trajectory’; stability in care demands was most likely in the ‘stable trajectory’, and an increase in care demands was connected to the ‘increase trajectory’. Changes in care intensity (number of hours) and task complexity (number of tasks) followed the same patterns: decreases were disproportionally found in the ‘decrease trajectory’, stability in the ‘stable trajectory’ and increases in the ‘increase trajectory’. In addition, transitions to more than emotional support or transitions to personal or medical care virtually only occurred in the ‘increase trajectory’. Shorter care episodes were over-represented in the ‘decrease trajectory’ and longer care episodes in the ‘increase trajectory’. Importantly, each class was found to be of substantial size, covering 22, 41 and 37 per cent of all care episodes, respectively.
Mean values on input variables by informal care trajectory
| Decrease trajectory | Stable trajectory | Increase trajectory | |
|---|---|---|---|
| Change in care receiver’s memory problems | |||
| Decrease | 0.22 | 0.01 | 0.01 |
| Stable | 0.72 | 0.93 | 0.37 |
| Increase | 0.06 | 0.06 | 0.62 |
| Change in care receiver’s psychological problems | |||
| Decrease | 0.31 | 0.02 | 0.01 |
| Stable | 0.66 | 0.96 | 0.61 |
| Increase | 0.03 | 0.02 | 0.38 |
| Change in care receiver’s physical problems | |||
| Decrease | 0.65 | 0.06 | 0.04 |
| Stable | 0.30 | 0.84 | 0.17 |
| Increase | 0.04 | 0.10 | 0.80 |
| Change in the number of caregiving hours | |||
| Decrease | 0.84 | 0.13 | 0.09 |
| Stable | 0.13 | 0.76 | 0.22 |
| Increase | 0.03 | 0.11 | 0.69 |
| Change in the number of caregiving tasks | |||
| Decrease | 0.76 | 0.07 | 0.17 |
| Stable | 0.22 | 0.90 | 0.33 |
| Increase | 0.02 | 0.03 | 0.50 |
| Transition to more than emotional support | 0.00 | 0.01 | 0.08 |
| Transition to personal/nursing care | 0.00 | 0.00 | 0.18 |
| Duration of care episode | |||
| 1–6 months | 0.31 | 0.31 | 0.06 |
| 7–24 months | 0.24 | 0.21 | 0.19 |
| 25–60 months | 0.16 | 0.21 | 0.25 |
| 61–120 months | 0.11 | 0.13 | 0.23 |
| 120+ months | 0.18 | 0.15 | 0.27 |
| Total | 0.22 | 0.41 | 0.37 |
Note: N = 5,595 episodes.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
In sum, care episodes could be clustered in an informative way representing highly distinct trajectories that carers follow. Some carers experience what is put forward in the dominant narrative of increasingly heavy care situations, with deteriorating health problems of the care receiver combined with more efforts in time and complexity from their side. However, this type of trajectory was ‘only’ found in 37 per cent of episodes. A slightly higher proportion of care episodes (41 per cent) were characterised by stability, in which the carers typically faced no major changes in their care situation. A non-negligible proportion of care episodes (22 per cent) actually followed a counterintuitive path in which the care receiver’s health improved and less was asked from the carer over time.
Relation between trajectories and the characteristics of the care receiver, carer and care context
We expected that certain trajectories would be more likely in certain situations based on theoretical arguments regarding the care receiver’s health progression, difficulty of saying ‘no’ by the carer and changes in the living situation of the care receiver. Table 4 shows the predicted probabilities and their deviation to the grand mean probability of that particular trajectory. The results are also presented visually in Figures 2 to 5.
Predicted probabilities of being in each of the three trajectories and their deviations from the grand mean probability of that particular trajectory
| Care receiver’s age at start episode | Decrease trajectory | Stable trajectory | Increase trajectory | ||||||||||||
| PP | se | Δ | se | PP | se | Δ | se | PP | se | Δ | se | ||||
| 20 | 0.41 | 0.02 | 0.19 | ** | 0.02 | 0.43 | 0.02 | 0.02 | 0.02 | 0.17 | 0.01 | –0.21 | ** | 0.01 | |
| 30 | 0.36 | 0.02 | 0.14 | ** | 0.01 | 0.44 | 0.02 | 0.03 | ** | 0.01 | 0.21 | 0.01 | –0.17 | ** | 0.01 |
| 40 | 0.31 | 0.01 | 0.09 | ** | 0.01 | 0.44 | 0.01 | 0.04 | ** | 0.01 | 0.25 | 0.01 | –0.12 | ** | 0.01 |
| 50 | 0.26 | 0.01 | 0.04 | ** | 0.00 | 0.44 | 0.01 | 0.03 | ** | 0.01 | 0.30 | 0.01 | –0.07 | ** | 0.01 |
| 60 | 0.21 | 0.01 | –0.01 | ** | 0.00 | 0.43 | 0.01 | 0.03 | ** | 0.00 | 0.36 | 0.01 | –0.02 | ** | 0.00 |
| 70 | 0.17 | 0.01 | –0.05 | ** | 0.00 | 0.42 | 0.01 | 0.01 | ** | 0.00 | 0.41 | 0.01 | 0.04 | ** | 0.00 |
| 80 | 0.13 | 0.01 | –0.09 | ** | 0.01 | 0.40 | 0.01 | –0.01 | 0.01 | 0.47 | 0.01 | 0.09 | ** | 0.01 | |
| Relationship: care receiver is … | |||||||||||||||
| Partner | 0.27 | 0.02 | 0.05 | ** | 0.02 | 0.30 | 0.02 | –0.11 | ** | 0.02 | 0.43 | 0.03 | 0.06 | ** | 0.03 |
| Parent | 0.22 | 0.01 | 0.00 | 0.01 | 0.37 | 0.01 | –0.04 | ** | 0.01 | 0.41 | 0.01 | 0.04 | ** | 0.01 | |
| Child | 0.24 | 0.02 | 0.02 | 0.02 | 0.41 | 0.03 | 0.01 | 0.03 | 0.35 | 0.03 | –0.02 | 0.03 | |||
| Sibling | 0.19 | 0.02 | –0.03 | 0.02 | 0.45 | 0.03 | 0.04 | 0.02 | 0.36 | 0.02 | –0.01 | 0.02 | |||
| Other family member | 0.20 | 0.02 | –0.03 | 0.02 | 0.43 | 0.02 | 0.02 | 0.02 | 0.38 | 0.02 | 0.00 | 0.02 | |||
| Friend, acquaintance, colleague | 0.21 | 0.01 | –0.01 | 0.01 | 0.53 | 0.02 | 0.12 | ** | 0.02 | 0.26 | 0.02 | –0.12 | ** | 0.02 | |
| Neighbour | 0.17 | 0.02 | –0.05 | ** | 0.02 | 0.55 | 0.02 | 0.14 | ** | 0.02 | 0.28 | 0.02 | –0.09 | ** | 0.02 |
| Feeling of obligation to care | |||||||||||||||
| 0 (lowest) | 0.19 | 0.01 | –0.03 | ** | 0.01 | 0.46 | 0.01 | 0.06 | ** | 0.01 | 0.35 | 0.01 | –0.03 | ** | 0.01 |
| 1 | 0.22 | 0.01 | 0.00 | 0.00 | 0.40 | 0.01 | 0.00 | 0.00 | 0.38 | 0.01 | 0.00 | 0.00 | |||
| 2 (highest) | 0.25 | 0.01 | 0.03 | ** | 0.01 | 0.34 | 0.01 | –0.06 | ** | 0.01 | 0.40 | 0.01 | 0.03 | ** | 0.01 |
| Change in care receiver’s living situation | |||||||||||||||
| Remained living with the carer | 0.25 | 0.02 | 0.03 | 0.02 | 0.46 | 0.02 | 0.05 | ** | 0.02 | 0.29 | 0.02 | –0.08 | ** | 0.02 | |
| Remained living independently with others | 0.20 | 0.01 | –0.02 | ** | 0.01 | 0.49 | 0.01 | 0.09 | ** | 0.01 | 0.31 | 0.01 | –0.06 | ** | 0.01 |
| Remained living independently without others | 0.21 | 0.01 | –0.01 | 0.01 | 0.46 | 0.01 | 0.05 | ** | 0.01 | 0.33 | 0.01 | –0.04 | ** | 0.01 | |
| Remained living in care facility | 0.10 | 0.02 | –0.12 | ** | 0.02 | 0.53 | 0.03 | 0.12 | ** | 0.03 | 0.38 | 0.03 | 0.00 | 0.03 | |
| From living with the carer to living independently | 0.42 | 0.04 | 0.20 | ** | 0.04 | 0.22 | 0.03 | –0.19 | ** | 0.03 | 0.37 | 0.04 | –0.01 | 0.04 | |
| Moved to care facility | 0.18 | 0.01 | –0.04 | ** | 0.01 | 0.23 | 0.02 | –0.18 | ** | 0.01 | 0.59 | 0.02 | 0.22 | ** | 0.02 |
| From living independently with others to without others | 0.23 | 0.03 | 0.01 | 0.03 | 0.31 | 0.03 | –0.09 | ** | 0.03 | 0.46 | 0.03 | 0.09 | ** | 0.03 | |
| Other changes | 0.31 | 0.03 | 0.09 | ** | 0.03 | 0.34 | 0.03 | –0.06 | ** | 0.03 | 0.35 | 0.03 | –0.03 | 0.03 | |
| Carer’s sex | |||||||||||||||
| Male | 0.19 | 0.01 | –0.03 | ** | 0.01 | 0.45 | 0.01 | 0.04 | ** | 0.01 | 0.36 | 0.01 | –0.01 | 0.01 | |
| Female | 0.24 | 0.01 | 0.02 | ** | 0.00 | 0.39 | 0.01 | –0.02 | ** | 0.01 | 0.38 | 0.01 | 0.00 | 0.01 | |
| Carer’s age at start episode | |||||||||||||||
| 24 years and younger | 0.17 | 0.01 | –0.05 | ** | 0.01 | 0.49 | 0.02 | 0.08 | ** | 0.02 | 0.34 | 0.02 | –0.03 | 0.02 | |
| 25–34 years | 0.21 | 0.01 | –0.01 | 0.01 | 0.44 | 0.02 | 0.03 | ** | 0.02 | 0.35 | 0.02 | –0.02 | 0.02 | ||
| 35–44 years | 0.21 | 0.01 | –0.01 | 0.01 | 0.43 | 0.02 | 0.02 | 0.01 | 0.37 | 0.01 | –0.01 | 0.01 | |||
| 45–54 years | 0.22 | 0.01 | 0.00 | 0.01 | 0.38 | 0.01 | –0.03 | ** | 0.01 | 0.40 | 0.01 | 0.03 | ** | 0.01 | |
| 55 years and older | 0.24 | 0.01 | 0.02 | 0.01 | 0.39 | 0.01 | –0.02 | 0.01 | 0.37 | 0.01 | 0.00 | 0.01 | |||
| Historical time | |||||||||||||||
| 1961 | 0.05 | 0.01 | –0.17 | ** | 0.01 | 0.24 | 0.03 | –0.17 | ** | 0.03 | 0.71 | 0.03 | 0.34 | ** | 0.03 |
| 1980 | 0.09 | 0.01 | –0.13 | ** | 0.01 | 0.31 | 0.02 | –0.09 | ** | 0.02 | 0.59 | 0.02 | 0.22 | ** | 0.02 |
| 2000 | 0.17 | 0.01 | –0.06 | ** | 0.01 | 0.38 | 0.01 | –0.02 | ** | 0.01 | 0.45 | 0.01 | 0.08 | ** | 0.01 |
| 2020 | 0.26 | 0.01 | 0.04 | ** | 0.01 | 0.43 | 0.01 | 0.03 | ** | 0.01 | 0.31 | 0.01 | –0.07 | ** | 0.01 |
Notes: Predicted probabilities based on multinomial logistic regression model as reported in Table A1 in Appendix 1; other variables kept at their means. N = 5,595 episodes in N = 2,884 carers. * p < 0.05; ** p < 0.01. Significance test based on deviation to the grand mean.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
Predicted probabilities of having experienced a decrease trajectory, stable trajectory or increase trajectory by care receiver’s age at start of episode
Citation: International Journal of Care and Caring 2024; 10.1332/23978821Y2023D000000021
Predicted probabilities of having experienced a decrease trajectory, stable trajectory, or increase trajectory by relationship with the care receiver
Citation: International Journal of Care and Caring 2024; 10.1332/23978821Y2023D000000021
Predicted probabilities of having experienced a decrease trajectory, stable trajectory, or increase trajectory by feeling of obligation to care
Citation: International Journal of Care and Caring 2024; 10.1332/23978821Y2023D000000021
Predicted probabilities of having experienced a decrease trajectory, stable trajectory or increase trajectory by change in the care receiver’s living situation
Citation: International Journal of Care and Caring 2024; 10.1332/23978821Y2023D000000021
In line with our expectation, we found that when care was provided to people who were aged 70 and older at the moment caregiving started, the care episode was significantly more likely to follow the ‘increase trajectory’. When the care receiver was relatively young (up until 50 years) at the moment the care episode started, the ‘decrease trajectory’ was significantly more likely. For instance, the predicted probability for a care episode to follow an ‘increase trajectory’ was 0.47 when the care receiver was 80 years old at the start of the care episode. This was 0.09 higher than the average predicted probability to be in the ‘increase trajectory’, which was a significant difference. For care episodes with a 20-year-old care receiver, the predicted probability to belong to the ‘decrease trajectory’ was 0.41, which was 0.19 higher than the average predicted probability.
We found partial support for the idea that difficulty of saying ‘no’ was related to the trajectory carers experienced. In accordance with this idea, we found that the probability of an ‘increase trajectory’ was significantly higher than average in cases of partner (Δ = 0.06) and parent care (Δ = 0.04), that is, to close kin, and when the carer felt high levels of obligation (Δ = 0.03). Lower probabilities to follow the ‘increase trajectory’, but higher probabilities to follow the ‘stable trajectory’ instead, were found if the care receiver was a neighbour (Δ = 0.14) and friend, acquaintance or colleague (Δ = 0.12), and if carers felt little obligation (Δ = 0.06). However, among two groups that were expected to have difficulties saying ‘no’, namely, carers who cared for their partner and carers who felt high levels of obligation, we also found significant higher-than-average probabilities to experience a ‘decrease trajectory’ (Δ = 0.05 and Δ = 0.03, respectively). This was not in line with our expectation. It should be noted that differences between relationship types did not reflect different levels of obligation, care receivers’ varying living situations or stages of health progression related to biological age because probabilities were predicted on the basis of a model that included all independent variables simultaneously.
Changes in the care context, particularly changes in the living situation of the care receiver, were significantly related to the type of trajectory in a way that was in line with our expectations. If the care receiver moved from the carer’s home to an independent household during the care episode, the carer was more likely than average to experience a ‘decrease trajectory’ (Δ = 0.20). Both a care receiver’s move to a care facility and a change from a shared to a single household during the care episode were related to higher-than-average probabilities of an ‘increase trajectory’ (Δ = 0.22 and Δ = 0.09, respectively). It should be noted that we cannot show whether or not care intensity and complexity decreased after the care receiver moved to a care facility. With our data, we can show that if the care receiver moved to a care facility somewhere during the care episode, the care demands, intensity and complexity were heavier at the end of the care episode than at the beginning. This does not imply that the care episode was also heavier after the move to a care facility than directly before that moment. Finally, stability in living situation often went together with the ‘stable trajectory’.
Insights from the four-class solution
In the four-class solution (see Table A2 in Appendix 1), the ‘decrease trajectory’ and ‘stable trajectory’ comprised largely the same respondents as in the three-class solution, but the respondents in the ‘increase trajectory’ were split into two classes. The substantive interpretation of one of them (class size was 19 per cent) very much resembled the ‘increase trajectory’ of the three-class solution. The extra class (class size was 21 per cent) was characterised by increased care demands as well, but this was less often accompanied by increased intensity than in the ‘increase trajectory’ (45 per cent versus 85 per cent) and not accompanied by an increased number or complexity of tasks. This ‘mixed trajectory’ stood out in one important respect: the prevalence of care receivers moving to a care facility was particularly high. We are therefore reluctant to capitalise on the lower levels of carers’ efforts in the ‘mixed trajectory’ to claim that the ‘mixed trajectory’ is truly different from the ‘increase trajectory’. We expect these trajectories to follow the same path for the major part of the care episode, both with increasing intensity and task complexity, up until the moment that care receivers in the ‘mixed trajectory’ are admitted to a care facility. However, theoretically, this ‘mixed trajectory’ is interesting to test more strictly the ‘difficulty of saying “no”’ hypothesis, as it comprises carers who were confronted with increasing care demands without (clear) increasing efforts in response. Additional analyses showed that caring for close kin (to whom it is presumably more difficult to say ‘no’) was indeed particularly likely in the ‘increase trajectory’, whereas care for less-close kin was common in the ‘mixed trajectory’. However, no different patterns were observed for the ‘increase trajectory’ and ‘mixed trajectory’ with respect to feelings of obligation (see Tables A3 and A4 in Appendix 1).
Conclusion and discussion
In response to the dominant narrative that informal care situations become increasingly heavy over time, this study set out to empirically assess whether and to what extent other informal care trajectories exist as well, and to examine in which situations which trajectories are more likely. In doing so, we explicitly considered caregiving as a process and not as a snapshot (Moen and DePasquale, 2017), and acknowledged that possibly large variation exists in the developments of care situations. We deliberately did not limit our study to care provided to family members in old age or with Alzheimer’s disease but covered a wide and more complete array of care situations. The empirical analyses were based on 5,595 informal care episodes among 2,884 carers, who were retrieved from a representative Dutch sample and who reported on their complete caregiving careers.
Our first research question was assessed with LCA. We identified three classes that represented distinct care trajectories in terms of developments in the care demands, intensity and complexity: the decrease, stable and increase trajectories. The first major conclusion is therefore that the dominant narrative of care trajectories that become increasingly demanding, intensive and complex over time paints only part of the picture. Less than four out of ten of the care episodes in our study matched this profile (37 per cent). Another four out of ten profiles followed a ‘stable trajectory’ (41 per cent) and more than one in five followed a ‘decrease trajectory’ (22 per cent), in which the care receiver’s health improved over time and the informal care task (both in terms of time and complexity) became smaller.
It is beyond the scope of this study to empirically assess the consequences of these different trajectories, but from the stress process model (Pearlin et al, 1990) and wear-and-tear hypothesis (Townsend et al, 1989), it can be argued that carers who experience an ‘increase trajectory’ will have higher levels of stress and worse mental health. Our finding that the majority of informal care episodes followed a ‘stable trajectory’ or even ‘decrease trajectory’ therefore sheds a more positive light on caregiving. Such episodes are arguably less burdensome. Possibly, in such caregiving situations, positive experiences derived from caregiving – such as feelings of gratitude or fulfilment, closer social contacts, or learning new things – which appear to be present in the majority of care situations (Pysklywec et al, 2020), may even outweigh the negative ones. We think it is important to realise that care situations evolve in very different directions and that the dominant view of increasing heaviness does not apply to all care situations. Such awareness of heterogeneity in care situations helps to respond to carers’ needs more accurately.
In answering our second question regarding the situations in which the different trajectories were especially likely, we conclude that the different trajectories were related to the characteristics of the care receiver, carer and care context. We found, first, that care to older (that is, 70+) care receivers often followed the ‘increase trajectory’, whereas care to younger (that is, until the age of 50) care receivers often followed the ‘decrease trajectory’. This may be unsurprising considering that higher ages come with worse health. Its implications, however, are important given the rapid ageing of society, which suggests that a rising proportion of care episodes will involve care to older people.
Second, we found partial evidence for the theoretical idea that the ‘increase trajectory’ would be more common when it is difficult for carers to say ‘no’, in this study, operationalised by the relationship type between carer and care receiver, as well as carers’ feeling of obligation to provide care. As expected, care for parents or a partner and higher levels of felt obligation went together with relatively high probabilities to end up in the ‘increase trajectory’, whereas care for non-kin and low obligation followed relatively often the ‘stable trajectory’. At the same time, however, partner care and strong obligation were also associated with higher-than-average probabilities to be in the ‘decrease trajectory’. Perceiving difficulties of saying ‘no’ may thus imply that carers jump in immediately when needed, both in situations with progressive and temporary health problems, adjusting flexibly to what is needed. Carers with more distant positions from the actual care situation, on the other hand, may perform their caregiving hours and tasks more independently from the total care needs. In this respect, the Dutch government’s call upon neighbours and friends to become more active in providing informal care may lead to a pool of steady extra helpers that are not highly at risk of becoming overburdened.
Finally, the living situation of the care receiver was strongly related to the developments the carer experienced. The ‘increase trajectory’ was found to be relatively likely if the care receiver moved to a care facility or ended up living alone somewhere during the care episode. The ‘decrease trajectory’, on the other hand, was relatively likely when the care receiver left the home of the carer to live independently. Although this study cannot inform about the causal order of events, these results suggest that carers whose care receiver loses their housemate(s), for instance, because of the death of their partner, may be a group that deserve more attention in carer support policies.
Reflecting on our study, we believe that the data we used offered important advantages, most notably, capturing a more realistic and complete picture of the wide array of care situations. A first reason for this was the fact that respondents were retrieved from a representative sample of Dutch households on the basis of ‘having provided care at least once during their lives’, not via care receivers with a specific disease. A second reason was that respondents were not forced to report only their current and/or major care situation but, rather, all episodes, long or short, intensive or not, and going on simultaneously or successively. Finally, care situations were not limited to family care but explicitly also included care provided to neighbours, friends, acquaintances and colleagues. However, the data also had some limitations. First, the retrospective nature of the data may have induced recall bias. Care episodes that were relatively insignificant to the carer or that took place longer ago may have been under-reported (Kjellsson et al, 2014). ‘Insignificant’ care episodes will arguably not have followed an ‘increase trajectory’, implying that our estimate of the proportion of care episodes following an ‘increase trajectory’ will be an upper bound. Second, developments in a care episode were based on a comparison between the reported beginning and end situation. However, more complex dynamics could have happened in between that we did not pick up, hampering more precise comparisons between the trajectories over time. For instance, carers in the ‘mixed trajectory’ (in the four-class solution) of increasing demands but without (clear) increasing intensity and task complexity may follow the same path as carers in the ‘increase trajectory’ for a large part of the care episode and only diverge relatively late in the episode (triggered by the admittance of the care receiver to a care facility). The lack of information between the beginning and end of the care episodes also implied that we were not able to assess the exact timing of a decrease or increase in care demands, intensity and complexity, and hence could not establish the causal order when relating care trajectories to, most notably, changes in care receivers’ living situation.
In closing, we believe that it was worthwhile to examine a wide range of care situations to get a more complete and realistic view of the processes carers experience. We have learnt that the dominant narrative of increasingly heavy care situations is incomplete and that carers often experience stable or even decreasing care demands, intensity and complexity over time.
Notes
To identify the different care episodes in carers’ lives, the questionnaire started with the question: In January you indicated in a questionnaire that you provide or provided informal care to people you know who had health issues. In this questionnaire we want to look at this more closely. We therefore want to ask you to revisit the past. First we would like to know to whom you provided informal care. Below we will ask you to indicate all the people known to you to whom you have ever provided informal care on account of their health issues. These people could be your partner, a family member, a friend, neighbour, acquaintance or colleague who needs or needed help because of physical, psychological or mental limitations or because of old age. Examples of informal care are doing household chores, helping with washing and dressing, keeping company, providing transport or performing odd jobs. You may have done so for a short period or for a long period. It could involve people known to you to whom you provided care in the past, but could also involve people known to you to whom you are providing care at present. Care provided as part of your occupation or as a volunteer does not count. Please take a moment to think about the people known to you to whom you have provided or are providing informal care on account of health issues.
The ALCPPs are available in a matrix with diagonal cells representing the average probability of a person being assigned to a class given their scores on the input variables. We reported the lowest ALCPP on the diagonal.
Funding
This work was supported by the Open Data Infrastructure for Social Science and Economic Innovations (ODISSEI) in the Netherlands (see: www.odissei-data.nl).
Conflict of interest
The authors declare that there is no conflict of interest.
References
Bicchieri, C. (2006) The Grammar of Society: The Nature and Dynamics of Social Norms, New York: Cambridge University Press.
Broese van Groenou, M.I., de Boer, A. and Iedema, J. (2013) Positive and negative evaluation of caregiving among three different types of informal care relationships, European Journal of Ageing, 10: 301–11. doi: 10.1007/s10433-013-0276-6
Colombo, F., Llena-Nozal, A., Mercier, J. and Tjadens, F. (2011) Help Wanted? Providing and Paying for Long-Term Care, OECD Health Policy Studies, Paris: OECD Publishing, doi: 10.1787/9789264097759-en
De Klerk, M., de Boer, A. and Plaisier, I. (2021) Determinants of informal care-giving in various social relationships in the Netherlands, Health and Social Care in the Community, 29(6): 1779–88. doi: 10.1111/hsc.13286
Gaugler, J.E. (2005) Family involvement in residential long-term care: a synthesis and critical review, Aging & Mental Health, 9(2): 105–18. doi: 10.1080/13607860412331310245
Keating, N., Eales, J., Funk, L., Fast, J. and Min, J. (2019) Life course trajectories of family care, International Journal of Care and Caring, 3(2): 147–63. doi: 10.1332/239788219x15473079319309
Kjellsson, G., Clark, P. and Gerdtham, U.G. (2014) Forgetting to remember or remembering to forget: a study of the recall period length in health care survey questions, Journal of Health Economics, 35: 34–46. doi: 10.1016/j.jhealeco.2014.01.007
Linzer, D.A. and Lewis, J.B. (2011) PoLCA: an R package for polytomous variable latent class analysis, Journal of Statistical Software, 42(10): 1–29. doi: 10.18637/jss.v042.i10
Moen, P. and DePasquale, N. (2017) Family care work: a policy-relevant research agenda, International Journal of Care and Caring, 1(1): 45–62. doi: 10.1332/239788217X14866284542346
Mood, C. (2010) Logistic regression: why we cannot do what we think we can do, and what we can do about it, European Sociological Review, 26(1): 67–82. doi: 10.1093/esr/jcp006
Pearlin, L.I., Mullan, J.T., Semple, S.J. and Skaff, M.M. (1990) Caregiving and the stress process: an overview of concepts and their measurements, The Gerontologist, 30(5): 583–94. doi: 10.1093/geront/30.5.583
Pinquart, M. and Sörensen, S. (2003) Differences between caregivers and noncaregivers in psychological health and physical health: a meta-analysis, Psychology and Aging, 18(2): 250–67. doi: 10.1037/0882-7974.18.2.250
Pysklywec, A., Plante, M., Auger, C., Mortenson, W.B., Eales, J., Routhier, F. and Demers, L. (2020) The positive effects of caring for family carers of older adults: a scoping review, International Journal of Care and Caring, 4(3): 349–75. doi: 10.1332/239788220x15925902138734
Schulz, R., Beach, S.R., Cook, T.B., Martire, L.M., Tomlinson, J.M. and Monin, J.K. (2012) Predictors and consequences of perceived lack of choice in becoming an informal caregiver, Aging & Mental Health, 16(6): 712–21. doi: 10.1080/13607863.2011.651439
Schulz, R., Beach, S.R., Czaja, S.J., Martire, L.M. and Monin, J.K. (2020) Family caregiving for older adults, Annual Review of Psychology, 71: 635–59. doi: 10.1146/annurev-psych-010419-050754
Tennstedt, S.L., Crawford, S. and McKinlay, J.B. (1993) Determining the pattern of community care: is coresidence more important than caregiver relationship?, Journal of Gerontology, 48(2): S74–83. doi: 10.1093/geronj/48.2.s74
Townsend, A., Noelker, L., Deimling, G. and Bass, D. (1989) Longitudinal impact of interhousehold caregiving on adult children’s mental health, Psychology and Aging, 4(4): 393–401. doi: 10.1037//0882-7974.4.4.393
Uccheddu, D., Gauthier, A.H., Steverink, N. and Emery, T. (2019) The pains and reliefs of the transitions into and out of spousal caregiving. A cross-national comparison of the health consequences of caregiving by gender, Social Science & Medicine, 240: 112517. doi: 10.1016/j.socscimed.2019.112517
Verbakel, E. and CentERdata (2021) LISS Panel – Retrospective Informal Care Careers: Main Measurement, Tilburg and Nijmegen: DANS/KNAW.
Verbeek-Oudijk, D., Woittiez, I., Eggink, E. and Putman, L. (2014) Who Cares in Europe? A Comparison of Long-Term Care for the Over-50s in Sixteen European Countries, The Hague: The Netherlands Institute for Social Research.
Weller, B.E., Bowen, N.K. and Faubert, S.J. (2020) Latent class analysis: a guide to best practice, Journal of Black Psychology, 46(4): 287–311. doi: 10.1177/0095798420930932
World Health Organization (2015) World Report on Ageing and Health, Geneva: World Health Organization.
Appendix 1
Multinomial regression analysis on informal care trajectories with random intercept, effects on log-odds
| Care receiver’s age at start episode (x 10) | Decrease-trajectory (vs stable-trajectory) | Increase-trajectory (vs stable-trajectory) | ||||
| b | se | b | se | |||
| –0.18 | ** | 0.03 | 0.22 | ** | 0.03 | |
| Relationship: care receiver is … (ref: parent) | ||||||
| Partner | 0.52 | ** | 0.19 | 0.35 | 0.19 | |
| Child | –0.07 | 0.22 | –0.36 | 0.24 | ||
| Sibling | –0.42 | * | 0.19 | –0.45 | * | 0.18 |
| Other family member | –0.34 | * | 0.16 | –0.33 | * | 0.13 |
| Friend, acquaintance, colleague | –0.54 | ** | 0.15 | –1.14 | ** | 0.14 |
| Neighbour | –0.88 | ** | 0.18 | –1.08 | ** | 0.16 |
| Feeling of obligation to care | 0.38 | ** | 0.07 | 0.31 | ** | 0.07 |
| Change in care receiver’s living situation (ref: remained living with the carer) | ||||||
| Remained living independently with others | –0.36 | * | 0.17 | –0.02 | 0.17 | |
| Remained living independently without others | –0.20 | 0.17 | 0.16 | 0.18 | ||
| Remained living in care facility | –1.26 | ** | 0.33 | 0.10 | 0.24 | |
| From living with the carer to living independently | 1.58 | ** | 0.27 | 1.28 | ** | 0.30 |
| Moved to care facility | 0.56 | ** | 0.21 | 1.84 | ** | 0.19 |
| From living independently with others to without others | 0.43 | 0.28 | 1.10 | ** | 0.25 | |
| Other changes | 0.67 | ** | 0.25 | 0.62 | * | 0.26 |
| Female | 0.49 | ** | 0.10 | 0.26 | ** | 0.09 |
| Age carer at start episode (ref: 55 years and older) | ||||||
| 24 years and younger | –0.70 | ** | 0.18 | –0.45 | ** | 0.17 |
| 25–34 years | –0.33 | * | 0.16 | –0.26 | 0.15 | |
| 35–44 years | –0.30 | * | 0.15 | –0.15 | 0.13 | |
| 45–54 years | –0.03 | 0.13 | 0.15 | 0.12 | ||
| Historical time (1961 = 0) | 0.02 | ** | 0.01 | –0.03 | ** | 0.00 |
| Constant | –0.69 | 0.42 | –0.21 | 0.37 | ||
| estimate | ||||||
| Variance level 2 | 1.23 | ** | 0.18 | |||
Notes: N = 5,595 episodes in N = 2,884 carers. * p < 0.05; ** p < 0.01.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
Mean values on input variables by informal care trajectory (four-class solution)
| Decrease trajectory | Stable trajectory | Increase trajectory | Mixed trajectory | |
|---|---|---|---|---|
| Change in care receiver’s memory problems | ||||
| Decrease | 0.24 | 0.01 | 0.01 | 0.00 |
| Stable | 0.74 | 0.95 | 0.44 | 0.32 |
| Increase | 0.01 | 0.05 | 0.55 | 0.67 |
| Change in care receiver’s psychological problems | ||||
| Decrease | 0.33 | 0.02 | 0.01 | 0.02 |
| Stable | 0.66 | 0.97 | 0.66 | 0.57 |
| Increase | 0.01 | 0.00 | 0.33 | 0.42 |
| Change in care receiver’s physical problems | ||||
| Decrease | 0.67 | 0.06 | 0.04 | 0.08 |
| Stable | 0.31 | 0.85 | 0.18 | 0.19 |
| Increase | 0.02 | 0.09 | 0.78 | 0.73 |
| Change in the number of caregiving hours | ||||
| Decrease | 0.85 | 0.12 | 0.03 | 0.21 |
| Stable | 0.13 | 0.76 | 0.12 | 0.34 |
| Increase | 0.02 | 0.11 | 0.85 | 0.45 |
| Change in the number of caregiving tasks | ||||
| Decrease | 0.75 | 0.06 | 0.01 | 0.39 |
| Stable | 0.23 | 0.90 | 0.03 | 0.61 |
| Increase | 0.02 | 0.04 | 0.96 | 0.00 |
| Transition to more than emotional support | 0.01 | 0.01 | 0.14 | 0.00 |
| Transition to personal/nursing care | 0.00 | 0.00 | 0.33 | 0.01 |
| Duration of care episode | ||||
| 1–6 months | 0.33 | 0.32 | 0.06 | 0.06 |
| 7–24 months | 0.25 | 0.21 | 0.18 | 0.18 |
| 25–60 months | 0.16 | 0.20 | 0.23 | 0.27 |
| 61–120 months | 0.10 | 0.12 | 0.22 | 0.24 |
| 120+ months | 0.16 | 0.15 | 0.31 | 0.24 |
| Total | 0.20 | 0.39 | 0.19 | 0.21 |
Note: N = 5,595 episodes.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
Predicted probabilities of being in each of the four trajectories and their deviations from the grand mean probability of that particular trajectory (four-class solution)
| Care receiver’s age at start of episode | Decrease trajectory | Stable trajectory | Increase trajectory | Mixed trajectory | ||||||||||||||||
| PP | se | Δ | se | PP | se | Δ | se | PP | se | Δ | se | PP | se | Δ | se | |||||
| 20 | 0.39 | 0.02 | 0.18 | ** | 0.02 | 0.41 | 0.02 | 0.03 | 0.02 | 0.10 | 0.01 | –0.09 | ** | 0.01 | 0.10 | 0.01 | –0.12 | ** | 0.01 | |
| 30 | 0.33 | 0.02 | 0.13 | ** | 0.01 | 0.42 | 0.02 | 0.03 | * | 0.01 | 0.12 | 0.01 | –0.07 | ** | 0.01 | 0.12 | 0.01 | –0.10 | ** | 0.01 |
| 40 | 0.28 | 0.01 | 0.08 | ** | 0.01 | 0.43 | 0.01 | 0.04 | ** | 0.01 | 0.14 | 0.01 | –0.05 | ** | 0.01 | 0.15 | 0.01 | –0.07 | ** | 0.01 |
| 50 | 0.23 | 0.01 | 0.03 | ** | 0.00 | 0.42 | 0.01 | 0.03 | ** | 0.01 | 0.17 | 0.01 | –0.03 | ** | 0.01 | 0.18 | 0.01 | –0.04 | ** | 0.01 |
| 60 | 0.19 | 0.01 | –0.01 | ** | 0.00 | 0.41 | 0.01 | 0.03 | ** | 0.00 | 0.19 | 0.01 | 0.00 | 0.00 | 0.21 | 0.01 | –0.01 | ** | 0.00 | |
| 70 | 0.15 | 0.01 | –0.05 | ** | 0.00 | 0.40 | 0.01 | 0.01 | ** | 0.00 | 0.21 | 0.01 | 0.02 | ** | 0.00 | 0.24 | 0.01 | 0.02 | ** | 0.00 |
| 80 | 0.12 | 0.01 | –0.08 | ** | 0.01 | 0.38 | 0.01 | –0.01 | 0.01 | 0.23 | 0.01 | 0.04 | ** | 0.01 | 0.27 | 0.01 | 0.05 | ** | 0.01 | |
| Relationship: care receiver is … | ||||||||||||||||||||
| Partner | 0.25 | 0.02 | 0.05 | ** | 0.02 | 0.28 | 0.02 | –0.11 | ** | 0.02 | 0.26 | 0.03 | 0.07 | * | 0.03 | 0.21 | 0.03 | –0.01 | 0.02 | |
| Parent | 0.20 | 0.01 | 0.00 | 0.01 | 0.36 | 0.01 | –0.03 | ** | 0.01 | 0.23 | 0.01 | 0.04 | ** | 0.01 | 0.21 | 0.01 | –0.01 | 0.01 | ||
| Child | 0.21 | 0.02 | 0.01 | 0.02 | 0.39 | 0.03 | 0.00 | 0.03 | 0.15 | 0.03 | –0.04 | 0.03 | 0.24 | 0.03 | 0.03 | 0.03 | ||||
| Sibling | 0.18 | 0.02 | –0.02 | 0.02 | 0.43 | 0.02 | 0.04 | 0.02 | 0.16 | 0.02 | –0.03 | 0.02 | 0.23 | 0.02 | 0.01 | 0.02 | ||||
| Other family member | 0.17 | 0.02 | –0.03 | 0.01 | 0.41 | 0.02 | 0.03 | 0.02 | 0.15 | 0.01 | –0.04 | ** | 0.01 | 0.26 | 0.02 | 0.04 | ** | 0.02 | ||
| Friend, acquaintance, colleague | 0.20 | 0.01 | 0.00 | 0.01 | 0.51 | 0.02 | 0.12 | ** | 0.02 | 0.09 | 0.01 | –0.10 | ** | 0.01 | 0.21 | 0.02 | –0.01 | 0.01 | ||
| Neighbour | 0.15 | 0.02 | –0.05 | ** | 0.02 | 0.53 | 0.02 | 0.14 | ** | 0.02 | 0.12 | 0.02 | –0.07 | ** | 0.02 | 0.20 | 0.02 | –0.01 | 0.02 | |
| Feeling of obligation to care | ||||||||||||||||||||
| 0 (lowest) | 0.17 | 0.01 | –0.03 | ** | 0.01 | 0.45 | 0.01 | 0.06 | ** | 0.01 | 0.17 | 0.01 | –0.02 | ** | 0.01 | 0.21 | 0.01 | –0.01 | 0.01 | |
| 1 | 0.20 | 0.01 | 0.00 | 0.00 | 0.39 | 0.01 | 0.00 | 0.00 | 0.19 | 0.01 | 0.00 | 0.00 | 0.22 | 0.01 | 0.00 | 0.00 | ||||
| 2 (highest) | 0.23 | 0.01 | 0.03 | ** | 0.01 | 0.33 | 0.01 | –0.06 | ** | 0.01 | 0.22 | 0.01 | 0.02 | * | 0.01 | 0.22 | 0.01 | 0.01 | 0.01 | |
| Change in care receiver’s living situation | ||||||||||||||||||||
| Remained living with the carer | 0.24 | 0.02 | 0.04 | * | 0.02 | 0.46 | 0.02 | 0.07 | ** | 0.02 | 0.17 | 0.02 | –0.03 | 0.02 | 0.14 | 0.02 | –0.08 | ** | 0.02 | |
| Remained living independently with others | 0.19 | 0.01 | –0.01 | 0.01 | 0.48 | 0.01 | 0.09 | ** | 0.01 | 0.17 | 0.01 | –0.02 | * | 0.01 | 0.16 | 0.01 | –0.06 | ** | 0.01 | |
| Remained living independently without others | 0.20 | 0.01 | 0.00 | 0.01 | 0.44 | 0.01 | 0.05 | ** | 0.01 | 0.21 | 0.01 | 0.01 | 0.01 | 0.16 | 0.01 | –0.06 | ** | 0.01 | ||
| Remained living in care facility | 0.09 | 0.02 | –0.11 | ** | 0.02 | 0.51 | 0.03 | 0.12 | ** | 0.03 | 0.14 | 0.02 | –0.05 | * | 0.02 | 0.27 | 0.03 | 0.05 | 0.03 | |
| From living with the carer to living independently | 0.37 | 0.03 | 0.16 | ** | 0.03 | 0.20 | 0.03 | –0.19 | ** | 0.03 | 0.19 | 0.03 | 0.00 | 0.03 | 0.24 | 0.04 | 0.02 | 0.04 | ||
| Moved to care facility | 0.12 | 0.01 | –0.08 | ** | 0.01 | 0.20 | 0.01 | –0.18 | ** | 0.01 | 0.24 | 0.01 | 0.04 | ** | 0.01 | 0.44 | 0.02 | 0.23 | ** | 0.02 |
| From living independently with others to without others | 0.21 | 0.03 | 0.00 | 0.03 | 0.28 | 0.03 | –0.11 | ** | 0.03 | 0.28 | 0.03 | 0.09 | ** | 0.03 | 0.23 | 0.03 | 0.01 | 0.03 | ||
| Other changes | 0.32 | 0.03 | 0.12 | ** | 0.03 | 0.33 | 0.03 | –0.06 | 0.03 | 0.21 | 0.03 | 0.02 | 0.03 | 0.14 | 0.03 | –0.07 | ** | 0.03 | ||
| Carer’s sex | ||||||||||||||||||||
| Male | 0.17 | 0.01 | –0.03 | ** | 0.01 | 0.43 | 0.01 | 0.04 | ** | 0.01 | 0.18 | 0.01 | –0.02 | ** | 0.01 | 0.22 | 0.01 | 0.00 | 0.01 | |
| Female | 0.22 | 0.01 | 0.02 | ** | 0.00 | 0.37 | 0.01 | –0.02 | ** | 0.01 | 0.20 | 0.01 | 0.01 | 0.00 | 0.21 | 0.01 | –0.01 | 0.01 | ||
| Carer’s age at start of episode | ||||||||||||||||||||
| 24 years and younger | 0.16 | 0.01 | –0.04 | ** | 0.01 | 0.46 | 0.02 | 0.08 | ** | 0.02 | 0.19 | 0.02 | 0.00 | 0.02 | 0.19 | 0.02 | –0.03 | 0.02 | ||
| 25–34 years | 0.19 | 0.01 | –0.01 | 0.01 | 0.42 | 0.02 | 0.04 | * | 0.02 | 0.19 | 0.01 | 0.00 | 0.01 | 0.19 | 0.01 | –0.03 | 0.01 | |||
| 35–44 years | 0.19 | 0.01 | –0.01 | 0.01 | 0.41 | 0.02 | 0.02 | 0.01 | 0.18 | 0.01 | –0.01 | 0.01 | 0.21 | 0.01 | –0.01 | 0.01 | ||||
| 45–54 years | 0.20 | 0.01 | 0.00 | 0.01 | 0.36 | 0.01 | –0.03 | * | 0.01 | 0.21 | 0.01 | 0.01 | 0.01 | 0.23 | 0.01 | 0.01 | 0.01 | |||
| 55 years and older | 0.22 | 0.01 | 0.02 | * | 0.01 | 0.37 | 0.01 | –0.02 | 0.01 | 0.18 | 0.01 | –0.01 | 0.01 | 0.22 | 0.01 | 0.01 | 0.01 | |||
| Historical time | ||||||||||||||||||||
| 1961 | 0.03 | 0.01 | –0.17 | ** | 0.01 | 0.21 | 0.03 | –0.17 | ** | 0.03 | 0.39 | 0.04 | 0.20 | ** | 0.04 | 0.36 | 0.04 | 0.14 | ** | 0.04 |
| 1980 | 0.07 | 0.01 | –0.13 | ** | 0.01 | 0.29 | 0.02 | –0.10 | ** | 0.02 | 0.32 | 0.02 | 0.13 | ** | 0.02 | 0.31 | 0.02 | 0.10 | ** | 0.02 |
| 2000 | 0.14 | 0.01 | –0.06 | ** | 0.01 | 0.37 | 0.01 | –0.02 | ** | 0.01 | 0.23 | 0.01 | 0.04 | ** | 0.00 | 0.25 | 0.01 | 0.04 | ** | 0.01 |
| 2020 | 0.24 | 0.01 | 0.04 | ** | 0.01 | 0.42 | 0.01 | 0.03 | ** | 0.01 | 0.15 | 0.01 | –0.04 | ** | 0.01 | 0.18 | 0.01 | –0.03 | ** | 0.01 |
Notes: Predicted probabilities based on multinomial logistic regression model as reported in Table A4 in Appendix 1; other variables kept at their means. N = 5,595 episodes in N = 2,884 carers. * p < 0.05; ** p < 0.01. Significance test based on deviation to the grand mean.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
Multinomial regression analysis on informal care trajectories with random intercept (four-class solution), effects on log-odds
| Decrease trajectory (vs stable trajectory | Increase trajectory (vs stable trajectory | Mixed trajectory (vs stable trajectory | |||||||
|---|---|---|---|---|---|---|---|---|---|
| b | se | b | se | b | se | ||||
| Care receiver’s age at start episode (×10) | –0.19 | ** | 0.03 | 0.19 | ** | 0.03 | 0.23 | ** | 0.03 |
| Relationship: care receiver is … (ref: parent) | |||||||||
| Partner | 0.60 | ** | 0.19 | 0.44 | * | 0.21 | 0.32 | 0.23 | |
| Child | –0.07 | 0.23 | –0.58 | * | 0.29 | 0.00 | 0.27 | ||
| Sibling | –0.37 | 0.19 | –0.70 | ** | 0.21 | –0.22 | 0.20 | ||
| Other family member | –0.38 | * | 0.17 | –0.65 | ** | 0.16 | 0.00 | 0.15 | |
| Friend, acquaintance, colleague | –0.50 | ** | 0.15 | –1.58 | ** | 0.18 | –0.65 | ** | 0.16 |
| Neighbour | –0.87 | ** | 0.19 | –1.32 | ** | 0.19 | –0.69 | ** | 0.18 |
| Feeling of obligation to care | 0.39 | ** | 0.07 | 0.36 | ** | 0.08 | 0.29 | ** | 0.07 |
| Change in care receiver’s living situation (ref: remained living with the carer) | |||||||||
| Remained living independently with others | –0.32 | 0.17 | 0.00 | 0.19 | 0.13 | 0.21 | |||
| Remained living independently without others | –0.14 | 0.18 | 0.33 | 0.20 | 0.28 | 0.21 | |||
| Remained living in care facility | –1.24 | ** | 0.34 | –0.29 | 0.29 | 0.60 | * | 0.27 | |
| From living with the carer to living independently | 1.57 | ** | 0.27 | 1.31 | ** | 0.33 | 1.70 | ** | 0.34 |
| Moved to care facility | 0.28 | 0.23 | 1.64 | ** | 0.21 | 2.45 | ** | 0.22 | |
| From living independently with others to without others | 0.49 | 0.29 | 1.32 | ** | 0.28 | 1.30 | ** | 0.30 | |
| Other changes | 0.78 | ** | 0.25 | 0.71 | * | 0.30 | 0.54 | 0.33 | |
| Female | 0.50 | ** | 0.10 | 0.36 | ** | 0.10 | 0.17 | 0.10 | |
| Age of carer at start of episode (ref: 55 years and older) | |||||||||
| 24 years and younger | –0.72 | ** | 0.19 | –0.32 | 0.20 | –0.53 | ** | 0.19 | |
| 25–34 years | –0.35 | * | 0.16 | –0.18 | 0.17 | –0.38 | * | 0.17 | |
| 35–44 years | –0.32 | * | 0.15 | –0.16 | 0.15 | –0.22 | 0.15 | ||
| 45–54 years | –0.07 | 0.13 | 0.17 | 0.13 | 0.09 | 0.13 | |||
| Historical time (1961=0) | 0.02 | ** | 0.01 | –0.04 | ** | 0.00 | –0.03 | ** | 0.01 |
| Constant | –0.89 | * | 0.43 | –0.50 | 0.43 | –1.10 | ** | 0.42 | |
| estimate | se | ||||||||
| Variance level 2 | 1.18 | 0.17 | |||||||
Notes: N = 5,595 episodes in N = 2,884 carers. * p < 0.05; ** p < 0.01.
Source: Retrospective Informal Care Careers, 2020 (Verbakel and CentERdata, 2021).
