Why do spouses provide personal care? A study among care-receiving Dutch community-dwelling older adults.

This study investigates under what conditions older spouses receive personal care from their spouse. Whether spousal care is provided is determined by individual and societal factors related to informal and formal care provision. Individual factors concern the need for care (the care recipient's health status), the spouse's ability to provide care (the spouse's health status) and the quality of the marital bond. Societal factors reflect changing policies on long-term care (indicated by the year in which care started) and gender role socialisation (gender). From the Longitudinal Aging Study Amsterdam, which completed eight observations between 1996 and 2016, we selected 221 independently living married respondents, aged 59-93, who received personal care for the first time and had at least one previous measurement without care use. The results show that if an older adult received personal care, the likelihood of receiving that care from the spouse decreased over the years: from 80% in 1996 to 50% in 2016. A husband or wife was less likely to receive spousal care when the spouse was unable to provide care or the quality of the relationship was low. No gender differences were found in either the prevalence of spousal care use or in the factors associated with that use. Thus, individual factors and the societal context seem to determine whether one receives personal care from their spouse. The decline in the likelihood of personal care provision from a spouse over the years may indicate a crumbling of family solidarity, an unmeasured and growing inability of the older spouse to provide care or an increasing complexity of care needs that requires the use of formal care. As care-giving can be a chronic stressor and most spouses provide care without assistance from others, attention from policy makers is needed to sustain the well-being of older couples.

What is known about this topic

Due to welfare state retrenchments, there is an increased need for informal care in many Western European countries.

Couples are growing older together; often spousal care is provided.

Spousal care‐giving arrangements vary with welfare policy.

What this paper adds

Spouses merely provide personal care when they are able to do so.

Both individual and societal factors determine whether the spouse provides personal care.

If an older adult received personal care, the likelihood of receiving that care from the spouse decreased over the last twenty years and no gender differences were found.

INTRODUCTION

Due to increasing longevity (Eurostat, 2017), many older people will become care‐dependent in the years to come. This creates an increasing demand for public long‐term care services, challenging the healthcare systems in many countries (Calvó‐Perxas et al., 2018). Most of the care received in later life is informal care provided by the family (Zigante, 2018), and many governments rely on the family to provide support in times of welfare state retrenchment (Pavolini & Ranci, 2008). The increasing longevity also increases the number of older couples and thus the potential for spousal care‐giving. Spouses are the most important source of care in old age (Bertogg & Strauss, 2020). In the Netherlands, intensive informal care for older adults is often provided by the spouse: 47% of care‐giving spouses provide at least eight hours of care per week, compared to 14% of care‐giving children (De Klerk et al., 2015). Although many spouses readily take on the role of caregiver, this cannot be expected from all spouses of those in need of care (Cash et al., 2013). As spousal informal care‐giving is one of the most important sources of care provision to the growing proportion of the population in need of care (Calvó‐Perxas et al., 2018), it is important to know under what conditions spousal care is not received.

As the spouse is the most important source of care provision to an older person (Bertogg & Strauss, 2020), we searched for conditions that exempt this practicality. One of these is the replacement of spousal care by care from other types of helpers, such as professional care providers or adult children. However, it appears that care from these types of helpers is more often a supplement to the spouse's care than a replacement, particularly when the health of the care‐receiving spouse is declining (Bertogg & Strauss, 2020; Feld et al., 2006). A second condition is that the spouse is not able to provide care due to having his or her own health problems or living in residential care (Feld et al., 2006). Third, due to poor relationship quality, the spouse may not want to provide care (Lawler et al., 2008). Finally, as husbands are less likely to provide care to their spouse compared to wives (Jang & Kawachi, 2019), gender roles in care‐giving may also explain why some spouses do or do not provide care.

Spouses generally provide all types of care, that is, instrumental and emotional support, household help and personal care. Most research on spousal care‐giving investigates household help and personal care. In line with Bertogg and Strauss (2020), we only studied personal care, and not household help, because the boundaries between sharing usual household chores and providing household help as a carer are blurred when studying couples. Furthermore, household chores may be perceived and defined differently according to gender (Bertogg & Strauss, 2020). Additionally, from a policy perspective, a focus on personal care provision is most relevant, as a lack of spousal personal care most likely necessitates publicly provided formal care.

In the current study, health, relationship characteristics, gender and the year in which care started are used to explain whether spousal care was received. Based on longitudinal data from the Netherlands, we selected community‐residing married older adults. The design has three characteristics. First, the focal observation in the longitudinal trajectory is when the use of care began, and then we focus on whether the spouse took on this care task. Second, we include information on the marital relationship before the care‐giving began. Third, the data spans multiple years and we are able to study the trend of spousal care‐giving between 1996 and 2016 and gender differences therein.

Conceptual framework and hypotheses

The Informal Care Model discusses several drivers of informal care use and provision (Broese van Groenou & De Boer, 2016). This model distinguishes (1) the care receivers’ need for care, (2) the disposition of the caregiver and (3) the societal context of long‐term care provision. We apply this model to spousal care use.

First, health problems may make it necessary for many older adults to rely on others for care (Miller et al., 2009; Potter, 2019). These so‐called need factors can be physical as well as cognitive. As all our respondents need care at a certain moment, we distinguish between the type (and the severity) of health problems to predict whether the spouse provides care. A formal helper may intervene when difficult or specific care is needed. Those in poor health may require nursing assistance at home, for example for injections and the use of an infusion pump. In these situations, professional care is required (Reinhard et al., 2014). Therefore, it is likely that formal helpers take over the provision of personal care to those in poor health while the spouse continues to provide other, less specialised types of care. We hypothesise that having more complex health needs increases the likelihood of the use of formal home care and thus decreases the likelihood of spousal personal care receipt (H1).

Second, whether the spouse provides personal care depends on his or her ability and willingness to provide care. When both spouses are old, they may experience alternating and overlapping health issues over time (Meyler et al., 2007). The accumulation of health issues in old age may make it necessary for both spouses to rely on others for care (Lu & Shelley, 2019; Potter, 2019). If the spouse himself or herself has health problems, he or she is probably unable to provide personal care. We hypothesise that spousal personal care is not received when the spouse is not healthy (H2a). Next, we assume that the willingness to provide care is dependent on the quality of the marital relationship. It is likely that a spouse in a high‐quality relationship wants to invest time and effort into providing spousal care (Braun et al., 2009; Lawler et al., 2008; Lima et al., 2008). Caring for a sick or frail spouse can be seen as an extension of the love and support that has been exchanged for many years (Bertogg & Strauss, 2020; Hong & Coogle, 2016). We hypothesise that an older adult in a high‐quality marital relationship is more likely to receive personal care from the spouse compared to one in a low‐quality relationship (H2b).

Third, on a societal level, we assume that the availability of publicly provided formal home care and gender roles in care‐giving are related to individual care use. Since approximately 1990, social policies in many Western countries emphasised choice and individual control over personal decisions, in contrast with the traditional roles and responsibilities of family and society (Fine, 2013). Home care was the mode of care preferred by recipients (Genet et al., 2013). This focus on individual choice also translated into government austerities in residential care; people were encouraged (and required) to stay longer in their own home and residential care facilities were reduced (Maarse & Jeunissen, 2016). Home care was promoted as a potentially cost‐effective way of providing care while at the same time maintaining people's independence. These processes of policy priority and choices, combined with the increase in available professional home care, may have reduced the traditional role of partners, children and other family in providing informal care in two ways. On the one hand, the lack of residential care forced older people to use home care, and the spouse, if present, would be the first one to provide care. This leads to a possible increase in the use of spousal care over the years. On the other hand, austerity measures required home care organisations to provide personal care and nursing care in particular and leave other types of care (household help for example) to informal caregivers or privately paid caregivers. This may have contributed to a specialisation of tasks between spouses and formal caregivers, and may imply a decrease in the likelihood of spousal personal care use when care is needed over the years. Also, older persons nowadays are healthier and more self‐sufficient than in the past (Mathers et al., 2015), so we can assume that if there is personal care need, professional formal care is more preferred.

Previous research among disabled older Americans showed a decrease in the use of informal care: spouses’ and children's propensity to provide care declined from 1982 to 2004 (Janus & Doty, 2018). Also, Dutch research showed that informal care use decreased and formal care use increased between 1992 and 2012 (Swinkels et al., 2016). The Dutch study used the same data as analysed in the current study, but differs from the current study in that all older adults were compared, regardless of whether or not they used care and whether or not they had a spouse. The current study adds to this that it examines trends in spousal care when personal care is used. We hypothesise that, if personal care is used, the likelihood of receiving care from the spouse has decreased in the period since about 1990 (H3a).

We see the gendered care‐giving role as a societal influence on the likelihood of receiving personal care from the husband or the wife. Following the socialisation model with regard to gender roles, women bear the responsibility for care (Cash et al., 2013; Eagly & Wood, 1991; Finley, 1989) and are more committed to care‐giving than men (Calasanti, 2010; Hong & Coogle, 2016; Pavalko & Woodbury, 2000; Savundranayagam & Montgomery, 2010). Wives are more likely than husbands to receive formal home and residential care than spousal care (Russell, 2007). As gender norms have changed in the last decades, due to various developments such as women's liberation in general and increasing female labour participation specifically, the gender difference in care‐giving may have narrowed (Glauber, 2017). This is corroborated by recent studies showing gender equality in spousal care (Langner & Furstenberg, 2020; Sharma et al., 2016). In a meta‐analysis including studies published in the years 1983 to 2005, small differences are found in the male and female take up of care‐giving tasks when it concerns spouses (Pinquart & Sörensen, 2011). We expect that men are more likely to receive spousal care than women, but that the gender difference has decreased over the years (H3b).

METHOD

Sample

This study employs data from the Longitudinal Aging Study Amsterdam (LASA). The samples are drawn from the registers of nine municipalities varying in urbanity in three geographic regions in the Netherlands (Huisman et al., 2011). Respondents were born between 1908 and 1957. The first interviews were carried out in 1992–1993 among respondents aged 55–85 (N = 3,107). The response rate at baseline was 63%. Observations are conducted every three or four years. In 2002–2003, a new sample aged 55–65 was added (N = 1,002), followed by a third sample aged 55–65 in 2012–2013 (N = 1,023). For each follow‐up, an average of 82% of respondents were re‐interviewed, 11% had died, 2% were too ill or too cognitively impaired to be interviewed, 5% refused to be re‐interviewed and fewer than 1% could not be contacted due to a relocation to another country or an unknown destination. For this study, we used data until 2016; the first observation in which the initiation of personal care use could be assessed was completed in 1996.

For our purposes, we selected respondents with at least two measurements, one or more without personal care use and a follow‐up with personal care use. At baseline (N = 5,132), we selected respondents living independently (excluding 123 institutionalised respondents) who shared their household with their spouse of the opposite sex (excluding 1,478 respondents not living with a spouse; including four respondents who were not married but lived with a partner of the opposite sex) and who did not receive personal care (excluding 100 respondents receiving personal care for whom it was not known when they received personal care for the first time, and six respondents with missing data). We then selected the first follow‐up observation in which a respondent received personal care. For 440 respondents, follow‐up observations were not available. In the course of the study, before they possibly received personal care at home, 194 respondents were institutionalised, for 62 the spouse was institutionalised and for 512 the marriage ended. Among the remaining respondents, 1,996 did not receive personal care at any of the follow‐up observations. We analysed the data of 221 respondents (132 husbands and 89 wives). At the observation that they received personal care for the first time, their age was on average 77.9 years (ranging from 59 to 94) and their spouse's age was on average 76.2 years (ranging from 52 to 91). For the 221 respondents, 914 observations were available: 221 observations with personal care receipt and on average 3.1 previous observations without personal care receipt. We used the previous observations for measurement of the marital quality and to impute missing data if applicable.

Measurements

Dependent variable

The respondent was asked whether he or she received personal care at the time of the interview. Our focus is on personal care when help with activities of daily living is needed, that is, washing, bathing or showering, dressing, going to the toilet, getting up and sitting down. After an affirmative answer, the respondent was asked to identify one or more sources of the personal care out of twelve options: spouse, resident child, resident other, non‐resident child, non‐resident other family, neighbour or friend or acquaintance, volunteer, publicly provided home carer, privately hired carer or an employee from a hospital. The responses were converted into two categories: ‘spousal care received’ (with or without care from other sources) or ‘care received but no spousal care’ (receiving care from other informal, formal and private sources).

Explanatory variables

The need factors of the respondent are measured by, first, the age of the respondent at the time of the interview. Second, the physical functioning of the respondent, which is measured with six questions about activities of daily living based on Katz et al., (1963). An example is ‘Can you walk up and down stairs?’ The five possible answers varied between (1) not at all to (5) without difficulty; sum scores range between 6 and 30. Third, the cognitive functioning of the respondent, measured using the Mini‐Mental State Examination scale (MMSE; Folstein et al., 1975). The scale scores range from 0 to 30, with higher scores indicating better cognitive functioning. Fourth, the number of chronic diseases of the respondent, which counts the presence of seven diseases. Possible scores range from 0 to 7.

To indicate whether the spouse is able to provide care, we used the age of the spouse at the time of the interview, and whether the spouse receives personal care himself or herself for which we asked the respondent whether the spouse currently needs help with his or her personal care.

For the quality of the relationship, we have two measurements available. Respondent's emotional loneliness is derived from the 11‐item loneliness scale (De Jong Gierveld & Van Tilburg, 1999) and coded as lonely (agreeing with one or more of the six emotional loneliness items) or not lonely. We further assessed whether the spouse was a confidant for the respondent. The respondent was asked who was a confidant among the personal network members (Van Tilburg, 1998). For both measurements, we used the observation before personal care is received. Therefore, the answers regarding martial quality are not influenced by the care that the wife or husband provided. The relationship quality can change as one spouse becomes dependent on the other (Monin et al., 2019).

To assess societal circumstances, we used the year of the interview in which the personal care is reported (year in which care started). The gender of the respondent distinguishes men and women.

Procedure

After conducting the descriptive statistics and testing for gender differences in our sample, we applied logistic regression analysis to test the hypotheses. The number of parameters in a multivariate logistic model should not exceed 10% of the lowest number of respondents in the two categories of the dependent variable (Peduzzi et al., 1996). Therefore, we tested our hypotheses with bivariate models and thereafter with a parsimonious multivariate model. To build the parsimonious model, we included only explanatory variables significant in the bivariate models. To test Hypothesis 3b, we added an interaction term of gender with year. We checked for problems with tolerance and influential cases. We used age of the respondent as a proxy for what drives care receipt, reflecting disability and poor health (Werblow et al., 2007) and age of the spouse to determine the ability to deliver the care. Due to a lack of tolerance, we controlled age of the respondent for spouse's age by including the residuals of the regression of respondent's age on the spouse's age. We chose the age of the spouse to use in our parsimonious model, as we have more variables available to measure the health of the respondent. We computed predicted probabilities for categories of explanatory variables.

The two categories of our dependent variable are not homogeneous, that is, both include respondents using care from one source and from more sources. To see whether this influenced our results, an additional analysis included respondents only receiving care from the spouse or receiving formal care.

The data were collected through face‐to‐face interviews and, when that was not possible, through telephone interviews with a shortened version of the face‐to‐face questionnaire. Of the 221 observations with personal care receipt, data in 62 were gathered in telephone interviews. This necessitated the imputation of missing values for cognitive functioning, number of chronic diseases and whether the spouse received care, with the value of the previous observation. For the spouse being a confidant, 46 respondents had missing data and for them we specified an additional category.

RESULTS

Of the 221 respondents, 142 received spousal care at the first follow‐up observation in which a respondent received personal care. For 115 respondents, this was the only source of care, and 27 used both spousal care and care from other sources (twelve respondents received formal care; ten other informal care; three formal and other informal care; one formal and private care and one private care). Among the 79 respondents who did not receive spousal care, 71 received formal care only, six received care from other informal caregivers and two received both formal and non‐spousal informal care. The mean values of the variables are shown in Table 1. Additional analyses showed gender differences in the age of the respondent at the moment the first care was received: for women the mean age was 75 and for men 80. No age differences were found for their care‐giving spouses.

TABLE 1

Descriptive statistics for study variables (N = 221)

	Minimum	Maximum	Mean	SD
Spousal care use (vs. not)			0.64
Age of the respondent	59.0	94.0	77.9	8.1
Physical functioning (poor‐good)	6	30	18.6	6.3
Cognitive functioning (poor‐good)	8	30	25.2	4.4
Number of chronic diseases	0	5	1.8	1.2
Age of the spouse	52.5	91.8	76.2	8.4
Spouse receives care (vs. not)			0.12
Emotional lonely (vs. not)			0.38
Spouse is confidant (no)			0.36
Spouse is confidant (yes)			0.43
Spouse is confidant (missing)			0.21
Year of start personal care use	1995.8	2016.7	2004.4	6.4
Respondent is female (vs. male)			0.40

Bivariate analyses show that age of both the respondent and the spouse, health of the spouse, marital quality and year in which care started explain whether spousal care is used (Table 1). The results from the bivariate analyses are forwarded to the parsimonious model (Table 2). The parsimonious model included four explanatory variables which explained 30% of the variance in spousal care. There are no influential cases; for all models the maximum Cook's distance is smaller than one. The results of the additional analysis among respondents only receiving care from one source, that is, spousal or formal care, are presented in the Supplementary Material, Table S1. Comparing the results of this analysis with the analysis among all respondents shows that the conclusions remain the same. Therefore, we continue with the discussion of the results from the analyses among all respondents.

TABLE 2

Logistic regression of spousal care use versus formal care use (N = 221)

	Bivariate models			Parsimonious multivariate model
	OR	95% CI	Nagelkerke R2	OR	95% CI
Need factors
Age of the respondent (59–93)	0.92***	0.89–0.96	0.11
Age of the respondent (controlled for spouse's age)a	0.87***	0.81–0.94	0.09	0.96	0.88–1.05
Physical functioning (poor‐good) (6–30)	1.00	0.98–1.07	0.00
Cognitive functioning (poor‐good) (8–30)	0.99	0.92–1.05	0.00
Number of chronic diseases (0–5)	1.04	0.82–1.32	0.00
Ability spouse
Age of the spouse (52–91)	0.90***	0.86–0.94	0.19	0.92***	0.87–0.96
Spouse receives care (vs. not)	0.16***	0.07–0.41	0.10	0.20**	0.08–0.54
Relationship quality
Emotional lonely (vs. not)	0.44**	0.25–0.77	0.05	0.52*	0.45–0.06
Spouse is confidant (no; reference)	0.02
Spouse is confidant (yes)	1.37	0.74–2.53
Spouse is confidant (missing)	1.82	0.83–3.98
Societal factors
Year in which care began (1995–2016)	0.94**	0.90–0.98	0.05	0.95*	0.90–0.99
Respondent is female (vs. male)	0.84	0.48–1.46	0.00
Model with interaction terms
Year in which care began (1995–2016)b	0.93**	0.88–0.96
Respondent is femaleb	0.64	0.17–2.45
Female * yearb	0.02	0.94–1.12	0.05
Nagelkerke R 2				0.30

Residual age from regression of age respondent on age spouse.

Main and interaction effects added together.

p < .05

p < .01

p < .0.01.

The predicted probability of receiving spousal care is 0.68. We did not find support for Hypothesis 1, which stated that the respondent's need affects the likelihood of spousal personal care use. Hypothesis 2a, regarding the effect of the spouse's ability to provide care, was supported. The likelihood of receiving spousal care decreases with spousal age. For a respondent with a spouse aged 60, the average predicted probability of receiving spousal care is 0.90, compared to 0.79 for those with a spouse aged 70, 0.61 for those aged 80 and 0.40 for those with spouses aged 90. For the second indicator, when the spouse receives care, the respondent's likelihood of receiving spousal care decreases. The average predicted probability for receiving spousal care is 0.35 if the spouse receives care, compared to 0.73 if the spouse does not receive care. For Hypothesis 2b, regarding the marital quality, we found no support for whether the spouse was a confidant, but we did find support for emotional loneliness. The average predicted probability of receiving spousal care is 0.73 for a respondent who was not emotionally lonely, compared to 0.59 for someone who was lonely. Hypothesis 3a, regarding care receiving as a socially influenced process, was supported: the results show that if an older adult received personal care, the likelihood of receiving that care from the spouse decreased over the years. When we consider respondents who received care for the first time as assessed in 1996, the average predicted probability of receiving spousal care is 0.78, compared to 0.53 in 2016. We did not find a gender difference in the likelihood of receiving personal care, even when we examined this in interaction with the year (Hypothesis 3b).

CONCLUSION AND DISCUSSION

This study investigates older adults who received personal care from their spouse and contrasted them with those who received care from other sources. The majority of the older adults received spousal care. In most other cases, only formal home care is received. Both individual and societal factors related to informal care provision predicted whether spousal care was used. In particular, health of the spouse, marital quality and year in which care started were associated with the use of spousal personal care.

The health status of older husbands and wives did not predict whether the spouse provides personal care. That spouses provide care regardless of the recipient's care need could be because any level of care‐giving makes spouses feel needed, useful and good about themselves (Tarlow et al., 2004). Caring for the spouse may promote positive feelings, while interacting with formal caregivers may be frustrating for the older couple or may feel too impersonal (Poulin et al., 2010). Previous research has shown that those who care for their spouse experience more care‐giving gains and fewer emotional care‐giving difficulties compared to caregivers who have many interactions with formal care‐giving agencies (Polenick et al., 2017). Thus, it seems that many older spouses prefer to provide care themselves, even if the care is very difficult, until the situation is so severe or specific that formal care must be used exclusively.

We used two indicators of marital quality. Emotional loneliness was associated with spousal care use, but the spouse being identified as a confidant was not associated. Cash et al., (2019) reported, based on a qualitative study, that Australian spousal caregivers identified moral obligations or a lack of alternative sources of care provision as the reasons they undertook spousal care. This implies that spouses may automatically assume and continue in caregiver roles in later life, regardless of the relationship quality and willingness to provide care. In a social‐democratic welfare state regime, of which the Netherlands has characteristics, the state adopts responsibility and home care is available as an alternative for spousal care, which may give spouses greater freedom of choice in taking up care‐giving tasks (Warburton & Jeppsson Grassman, 2011). If there is a choice because formal home care is available, relational characteristics are of importance for whether the older person chooses to provide care or not.

Besides individual characteristics of the spouse, national societal circumstances also played a role in whether spousal care was received. Our data covered a period of twenty years. It is noticeable that, if an older adult received personal care, the likelihood of receiving that care from the spouse decreased (estimated as 78% in 1996 and 53% in 2016) and increasingly care was provided by other sources, mainly formal care. This decline in the likelihood of spousal care‐giving can be associated with the greater availability of publicly or privately paid services for older adults living at home. We did not empirically study whether the decline in spousal care reflects macro‐level changes, such as less family solidarity or a more generous public provision of home care, but these are possible causes of the decline. The decrease in spousal care may be explained by more free choice and less obligation to care in modern societies (Cash et al., 2013; Fine, 2013). Another possible explanation is that because older persons nowadays are healthier and more self‐sufficient than in the past (Mathers et al., 2015), the group that does need care in recent years may need more complicated levels of care, and thus requires professional home care. However, the need factors included in this study do not support this explanation.

Except for the finding that women using spousal care were on average younger than their male counterparts, gender differences were not found in the explanatory variables in this study. This contrasts with many previous studies on care‐giving (e.g., Feld et al., 2006; Pinquart & Sörensen, 2006). However, as women and men age, the gender gap in care use decreases because after retirement men are also more likely to care for their wives (Glauber, 2017). Lagner and Furstenberg (2020) found in a longitudinal study that gendered behavioural patterns can change in later life. Their results also showed that gender differences in care response to an ill spouse are not as evident as many studies previously reported.

A strong point of our research was that we had a random sample of older adults, which we followed to the moment of the beginning of care use. We overcame the sampling limitations of many previous studies, that is, sampling of the already sick, sampling of employed spouses only and sampling of those with a specific disease (Langner & Furstenberg, 2020). We observed that only a small number of the older adults who were eligible for our study received personal care. To increase the power in studies on this topic, and to allow for more explanatory variables, there is a need for studies with a larger sample size than we had available.

A limitation of this study was that we lacked more precise information on the care need factors (we assessed age and health‐related limitations). We have no particular information on what type of personal care task (e.g., helping with dressing or with showering) was performed, which may be important for those who provide care, that is, the spouse or formal home services. Similarly, for relationship quality, we would have liked to have had a better measurement of whether the couple does not want to be involved in a ‘care relationship’ and thus would rather use formal home care, or the spouses want to help each other and thus favour spousal care. Regarding the spouse's physical capacity to provide care, we lacked information on his or her health limitations.

Although our study concerns spouses in the Netherlands, we suggest our results are generalizable to some European countries. For example, as in the Netherlands, in Scandinavian countries formal home care services are widely available, and informal and family care is provided less often (Alber & Kohler, 2004). In Southern and Eastern European countries, on the other hand, more informal care‐giving occurs due to a lack of formal care facilities. Outside Europe, for example, Asia and the global south, children and other family members are preferred as caregivers (Awuviry‐Newton et al., 2020). Thus, our findings may be particularly generalizable to countries with similar norms on family responsibility and long‐term care policies.

The current study showed that spousal care is likely to be received under the condition that the spouse has no severe disabilities and a good spousal relationship is maintained. The decline in spousal personal care use over the years possibly reflects the ageing of older couples as well as the choice spouses seem to have because of the (still) generous long‐term care policy in the Netherlands. If welfare states want to extend ageing at home, they must keep up a high‐quality system of formal home care, in order to avoid pressure on the often old and frail spouse to provide informal care. Spouses were and remain an important source of care in later life. This is important as rising numbers of older couples are living into their eighties and nineties and experiencing health impairments (Carr & Utz, 2020). Attention from policy makers as well as researchers is needed to promote and facilitate the development of practices to sustain the well‐being of older couples needing personal care.

CONFLICT OF INTEREST

The authors have no conflicts of interest.

Table S1

Click here for additional data file.