The linked lives perspective is one of the core principles of life-course theory and it is useful for explaining the relationship between the health of elderly couples and their satisfaction with their sex lives (Angela, 1996; Lodge & Umberson, 2012). According to this perspective, the couples’ health is closely related and the results of their interactions are accumulated over the life course because the family members of a family share daily life and exchange care (Elder, 1994). This perspective suggests that physical and cognitive health, and the sexual life of couples are closely related and the pattern is more distinctive for elderly couples who have shared their daily life and sexual life for decades (Biddle & Yap, 2010; Elder, 1998; Franks, Lucas, Stephens, Rook, & Gonzalez, 2010). When one half of an elderly couple has restrictions on the activities of daily living (ADL) or the instrumental activities of daily living (IADL), the spouse becomes a primary caregiver (Wolff & Kasper, 2006), and it becomes harder to maintain a sexual life in the same way as before (Lee, 2014). Consequently, deterioration in physical or cognitive health and the restrictions of independent living can hinder both husband and wife’s satisfaction in their sexual life (Kim & Hong, 2016; Lee, 2014; Yoo, 2017).

It is possible to understand the sexual life of the elderly couples better by integrating the linked lives perspective and the perspective of intersectionality (Lodge & Umberson, 2012). The concept of intersectionality is primarily used in women’s studies; however, the main idea of this perspective is to respect the diversity of sub-groups that are sub-groups whose identities may be suppressed by belonging to a wider group. From this perspective, a husband and a wife may give different meanings and values to their sex life (Hirayama & Walker, 2011; Karraker & DeLamater, 2013; Karraker, DeLamater, & Schwartz, 2011) owing to the gender ideologies of masculinity and femininity, which have been constructed in their contemporary manner for several decades (Lodge & Umberson, 2012; Waite, Laumann, Das, & Schumm, 2009). Although this relationship between exogenous and endogenous variables can be explained logically according to the two theories (i.e., linked lives and intersectionality), there is no empirical study that evaluates how the limitations in physical and cognitive health affect a couple’s satisfaction in their sexual life in Korea. Our aim is to provide empirical evidence for this relationship in an East Asian context.

This section will examine what is known about this topic by reviewing previous studies on the physical and cognitive health problems, limitations in independent living, and sexual life satisfaction of elderly couples. Moreover, it will identify the issues to be evaluated further in this study. Some studies have argued that sexual life satisfaction varies according to the age, gender, and household income of the elderly husbands and wives (Kim, 2018; Kim & Lee, 2013; Yoo, 2017). However, other studies have reported contradictory results, namely that sexual life satisfaction can be maintained at a high level although the pattern of sexual relations changes and frequency decreases as the couple ages (Kim & Lee, 2013; Lee, 2014; Lindau et al., 2007; Lodge & Umberson, 2012). Some research has claimed that deteriorating health rather than biological age influences the sexual life directly and negatively (Karraker, DeLamater, & Schwartz, 2011; Shin & Park, 2014). However, there are arguments that take an opposing perspective. For example, Waite et al. (2009) reported that mental health is more closely related to dissatisfaction with sexual life than physical health. Previous studies have also revealed that the prescribed medicine for chronic diseases causes a loss of libido only in female patients and that diagnosed chronic diseases do not affect various sexual activities including petting, oral sex, and masturbation (DeLamater & Moorman, 2007). Moreover, physical sexual dysfunctions can be overcome by using sexual function enhancers (Kingsberg, 2002). Overall, many researchers agree that the elderly individual’s gender is related to their sexual functioning, libido, and sexual life satisfaction. However, different opinions exist on whether the main factor hindering sexual life satisfaction is biological age, physical health, or mental health. It is difficult to find a dyadic-structure study that evaluates the effects of the husband or wife’s health on their own sexual life satisfaction in elderly Korean couples.

The results of previous studies did not agree on the relationships among age, physical health, cognitive functioning, and the sexual life satisfaction of the elderly. It is possible that the limitations in independent living, which occur when age, physical health issues, or deteriorated cognitive functions exceed a certain level, may affect the sexual life satisfaction of a partner. This is because the limitations in independent living due to a physical health issue or worsened cognitive functions can have adverse effects on the marital relationship between an elderly person with a health problem and a spouse providing care. The elderly, who have health issues due to chronic illness or who use long-term care services, also have a desire for a sexual life (Doll, 2013). An elderly person, who needs long-term care, and his or her spouse have more severe difficulties in engaging in a sexual life than those who live independently, thus their sexuality is neglected and excluded from the subject of sexual life support (Doll, 2013; Harris, Adams, Zubatsky, & White, 2011). Studying the sexual lives of the elderly who have limitations in independent living and identifying measures to support them is therefore particularly a timely issue (Bentrott & Margrett, 2011; Frankowski & Clark, 2009).

We examined whether the linked life perspective actually explained the relationship between the health of the elderly couple and sexual life satisfaction; our aim was to complement and develop several tasks identified by previous studies conducted on the sexual life satisfaction of the elderly. First, most previous studies conducted surveys or in-depth interviews with the elderly using a specific facility in a specific area and adopting convenience sampling; these studies recommended that follow-up studies should use data representing the whole population (Hwang & Shin, 2012; Yoo, 2017). Therefore, our intention was to verify the results of previous studies by using the Survey on Elderly Life Conditions (SELC) in 2011, which is representative of the elderly in South Korea.

Second, the sexual life and satisfaction of a married person should be considered as a bilateral relationship, but previous studies analyzed data at the individual level and could not adequately reflect the characteristics of the spouse (Kim & Lee, 2013). Therefore, we analyzed the physical and cognitive health of the husband and the wife, their limitations in independent living, and how their individual sexual life satisfaction influenced their partner by using the Actor-Partner Interdependence Model (APIM) (Cook & Kenny, 2005).

Third, sexual life satisfaction in the later period of life is not solely determined by biological age or the level of physical and cognitive functioning, and it may differ because of the occurrence of limitations in independent living. However, it is difficult to identify previous studies examining the effects of the limitations caused by the deteriorated physical and cognitive health of a husband or a wife and how they affect sexual life satisfaction. Therefore, we evaluated how the limitations in the couple’s independent living affected each other’s sexual life satisfaction by focusing on the mediating effects.

Lastly, our aim was to conduct analyses emphasizing intersectionality whereby the elderly were divided into various sub-groups having diverse and different attributes, because it was not clear how the factors affecting their sexual life satisfaction differed specifically according to physical health, cognitive function, actor, partner, and gender.

Our objective was both to seek ways to improve the sexual life satisfaction of the elderly and guarantee the sexual rights of those who had poor physical and cognitive health conditions and limitations in independent living. To accomplish this objective, we analyzed the data of 1,930 couples over 65 years old from the SELC in 2011, which represented the elderly of South Korea, and conducted an extended version of the APIM analysis.

The specific research questions were as follows.

• Question 1: How do the physical health problems and worse cognitive functions of elderly couples directly affect the sexual life satisfaction of the husband and the wife?
• Question 2: What are the mediating effects of the limitations in independent living due to physical health problems and the worsened cognitive functions on the individuals’ sexual life satisfaction?
• Question 3: Are there direct effects of the physical health problems and worsened cognitive function on sexual life satisfaction and do the indirect effects mediated by independent living differ by actor, partner, or gender?

The subjects in this study were 1,930 couples representative of elderly couples older than 65 years living in South Korea. We used the SELC (2011) conducted by the Ministry of Health and Welfare and the Korea Institute for Health and Social Affairs to analyze information regarding sexual life satisfaction, cognitive functioning, independent living, age, and household income. The SELC in 2011 was carried out according to Article 5 of the Welfare of the Aged Act; it targeted 10,674 elderly people (65 years old or older) residing in 3,142 survey districts created by applying a stratified two-stage cluster sampling method to cities and provinces. Sixty-eight professionally trained investigators visited the houses and interviewed subjects; the information obtained was subsequently verified by phone interviews with over 91% of respondents. We obtained consent from participants before collecting sensitive information on sexual life satisfaction and informed participants that they could voluntarily refuse to answer an uncomfortable question or end their participation in the study at any time. Furthermore, investigators were trained on these issues before conducting the interviews. We also received Institutional Review Board (IRB) approval from our institution so as to comply with established research ethics before conducting the study and analyzing the data obtained. The SELC data collected through this process in 2011 are not only representative and credible but also include a variety of material about the sexual life satisfaction, physical and cognitive health, independent living, and social and demographic characteristics of the couples thus it provided data suitable for the analytic purposes of this study.

The elderly couple in a dyadic structure (rather than an individual elderly person) was used as a unit of analysis because we examined the bilateral relationship between a husband and a wife. To analyze the couple unit, we excluded 2,508 subjects who did not have a spouse, who had a spouse but one member of the couple was younger than 65, or those where only one member participated in the SELC in 2011. Most of the subjects excluded from the study had either had their spouses die or had spouses who were too under 65. Moreover, 68 subjects were excluded from the study because their spouses were very ill (e.g., dementia and long-term hospitalization) and they could not answer the SELC questions directly. An additional 77 subjects were excluded from the analysis because they denied answering questions related to sexual life satisfaction, a dependent variable of this study. Finally, we used 1,930 couples (total 3,860 individual persons: 1,930 husbands and 1,930 wives), who were over 65 years old and responded to the survey as study subjects.

Figure 1 shows the conceptual framework of the study.

Figure 1. 
The analytical framework for the study.

This conceptual framework assumed that a decrease in the physical and cognitive health of a husband or a wife would cause limitations in independent living and these factors could affect the sexual life satisfaction of either the husband or wife. We used an extended version of APIM for this analysis. The strength of APIM is that it can quantify the effects of an actor’s attributes on a partner and vice versa by using multiple people having close interactions as a group (Cook & Kenny, 2005). The application of APIM, which considers the characteristics of both people, is appropriate because the sexual life of elderly couples involves the physical and emotional interaction between a husband and a wife; a spouse becomes a primary caregiver when the other person has physical and cognitive limitations in independent living. The extended version of APIM was used to reflect the mediating effects in the interdependent relationship between the husband and the wife because the physical health problem and the decline in cognitive function could induce limitations to independent living, and these factors could, at the same time, influence the sexual life satisfaction of the couple. It also has several advantages as it is based on full information maximum likelihood method (FIML), which includes missing cases of exogenous variables, minimizes measurement error by using the latent variables of physical health problems, and can consider multiple endogenous variables, including parameters, at the same time (Duncan, Duncan, & Strycker, 2013; Enders & Bandalos, 2001).

We conducted the following intensified process according to the research questions. To evaluate the first research question, we estimated the direct path actor-partner effect of the physical health problem and the worse cognitive functions of the couple on the sexual life satisfaction for each partner and compared the estimated values. To address the second research question, we evaluated the mediating effects of the limitations in independent living on the path from worse physical health problems and cognitive functions of the couple to sexual life satisfaction. Lastly, the third research question constrained each path of the direct and the indirect effects of the limitations in independent living that were estimated for previous questions, and compared the differences in model fitness x² between the free model, which set all paths free, and the other models. We conducted identity verification and compared the specific differences in physical health, cognitive function, actor, partner, and gender.

Table 1 shows the results of the descriptive statistics and the correlation analysis of the variables.

The mean sexual life satisfaction of the husband and that of the wife were similar to each other at approximately 3.0. The mean number of chronic diseases was 2.2 and 2.8 for the husband and wife, and the mean number of medications taken was 1.8 and 2.3, respectively. The results showed that the wife had more physical health problems. Approximately 10% of subjects had limitations in ADL and IADL: the percentage was higher for the wife (husband = 11% and wife = 13.0%). The mean cognitive function scores were 25 for the husband and 24 for the wife. When the cognitive functions of the subjects were determined based on the MMSE-KC scores, 20% of the elderly showed deteriorated cognitive functions. However, people who are admitted to a hospital or a nursing home owing to severely deteriorated cognitive function are excluded from the SELC, which examines the elderly living in local communities. Therefore, only 3% of the total subjects were confirmed by doctors as having dementia. The mean age of the husband was 75.5 years and that of the wife was 72.1 years. The results of the correlation analysis between variables showed that the sexual life satisfaction of the husband and that of the wife were closely related (r = .53, p < .001). The correlations between chronic diseases and taking medication were also very high (r = .83 for wife, r = –.89 for husband, p < .001) and the results revealed that they were appropriately used as observational variables to constitute the latent variables of physical health problems. The results indicated that the age of the couple, the worse cognitive function, and the limitations in independent living were highly correlated (p < .001).

Table 2 and Figure 2 show the results of the structural equation modeling of the physical health problem of each subject, the direct effects of the worse cognitive function, and the indirect effects of limitations in independent living, affecting sexual life satisfaction. The root mean squared error of approximation (RMSEA) for the model fit was .06, comparative fit index (CFI) was .98, Tucker-Lewis Index (TLI) was .93, and standardized root mean square residual (SRMR) was .01, which were all excellent.

Figure 2. 
Actor-partner interdependence model results on sexual life satisfaction among Korean elderly couples.

Notes. Significant paths are solid lines, and non-significant pathways are dotted lines. Husband’s age, wife’s age, and household income are controlled in these results. Path coefficients in this figure are standardized values.

*p < .05. **p < .01. ***p < .001.

As a first research question, we evaluated the results of actor-partner effects of pathways that directly connected sexual life satisfaction with physical health problems and worse cognitive function. First, having more physical health problems decreased the sexual life satisfaction of the husband, and showed significant direct actor effects (β = -.09, p < .001), but the sexual life satisfaction of the wife was not significantly influenced by own physical health problems. Second, the sexual life satisfaction of the husband and of the wife did not show significant differences because of the direct physical health problems of the partner. Third, a more severe decline in the cognitive function of the husband lowered his sexual life satisfaction and revealed significant and direct actor effects (β = -.09, p < .01) while the more severe decline in the cognitive function of the wife decreased her sexual life satisfaction and indicated significant and direct actor effects (β = -.10, p < .01). Fourth, the sexual life satisfaction of the husband was not significantly affected by the worse cognitive function of the wife directly but that of the wife significantly decreased with the worse cognitive function of the husband, and showed direct partner effects (β = -.06, p < .05). The age of the husband, the age of the wife, and the household income, which were used as control variables, did not have significant effects on sexual life satisfaction for either actor effects or partner effects.

The second research question examined the results of actor-partner effects in the pathway indirectly connecting worse cognitive function to sexual life satisfaction mediated by the limitations of independent living. First, the husband’s physical health problems (β = .17, p < .001) and worse cognitive function (β = .32, p < .001) began to limit his independent living. The limitations in the husband’s independent living had significant negative effects on both husbands’ sexual life satisfaction (β = -.07, p < .01) and wives’ sexual life satisfaction (β = -.06, p < .05). Like the husband, the wife’s physical health problems (β = .22, p < .001) and worse cognitive function (β = .29, p < .001) also led to limitations in her independent living, and the limitations in independent living negatively influenced her husband’s sexual life satisfaction (β = -.10, p < .001) and her sexual life satisfaction (β = -.05, p < .05). Although the husband’s old age mediated the limitation in independent living of the husband, and indirectly affected the sexual life satisfaction of the husband and the wife negatively, other control variables did not have even indirect significant effects.

In addition, we conducted identity verification for each pathway to evaluate if the directs effects of the physical health problem, the worse cognitive function on sexual life satisfaction, and the mediating effects of the limitations in independent living differed by actor, partner, and gender. First, the husband’s sexual life satisfaction was similarly influenced by his physical health problem and worse cognitive function [△x² (1) = .62], but it was not significantly affected by the wife’s physical health problem and worse cognitive function. Therefore, there was no direct partner effect. Secondly, the wife’s sexual life satisfaction was similarly affected by her own worse cognitive function and the husband’s worse cognitive function [△x² (1) = .58], but there was no significant difference between actor and partner effects because neither party was significantly affected by own or partner’s physical health problems. In summary, the husband’s sexual life satisfaction was directly affected more by his own health condition than by his partner’s health condition regardless of physical health and worse cognitive function. In contrast, the wife’s sexual life satisfaction was only influenced directly by her own worse cognitive function and that of her spouse.

The results of identity verification on mediator effects revealed that both husband and wife showed significantly larger effects of worse cognitive function than of physical health problems on the limitations in independent living [△x² (2) = 57.73, p < .001], and the actor effects were significantly larger than the partner effects [△x² (4) = 246.42, p < .001]. However, the effects of the limitations on independent living, a parameter, or on the husband’s and the wife’s sexual life satisfaction were not significantly different for actor, partner, or gender. The results indicated that worse cognitive function rather than physical health problems, and own rather than partner’s physical and cognitive levels induced the limitations in independent living. However, after occurring, the limitations in independent living had a negative effect on sexual life satisfaction regardless of actor, partner, or gender.