Expecting Relocation to a Nursing Home: Longitudinal Links with Functional Limitations, Self-Rated Health, and Life Satisfaction

With advancing old age, individuals are often confronted with increased states of dependency, sometimes leading to a need for formal care provided in a nursing home [1]. Relocation to a nursing home can be considered an age-related challenge, with potential negative impacts such as emotional distress and depression [2].

Psychosocial models of successful aging emphasize the potential of older adults to age well, even when confronted with challenges [3, 4]. These models propose that older adults employ strategies directed toward maintenance and growth, as well as strategies to minimize the impact of losses [3, 4]. The lifespan approach to successful aging [5] suggests that the strategies used to age successfully change across the lifespan. In advancing old age, the balance between gains and losses shifts, wherein losses oftentimes outweigh or hinder opportunities for maintenance and growth. As a result, strategies to age successfully become more directed toward the minimization of the impact of losses [3, 6, 7]. Developing realistic expectations of future old age constitutes an adaptational strategy which facilitates the anticipation of and adjustment to potentially inevitable challenges [8, 9], such as relocation to a nursing home.

The present study focuses on the specific adaptational strategy of developing an expectation to relocate to a nursing home. Is the development of this expectation in older adulthood triggered by declines in health and well-being? And when relocation to a nursing home takes place, do individuals who expected this transition beforehand show better adjustment afterward?

Relocation to a nursing home is usually preceded by health declines, medical burdens, increases in functional limitations, and cognitive impairment [10]. In comparison to older adults who experience other types of relocations, older adults who move to nursing homes show the lowest mean levels and steepest declines in self-rated health (SRH) and functional abilities before relocation [1]. Research including measures of well-being before relocation to a nursing home is scarce. Cross-sectionally, a negative association was found between well-being and future relocation to a nursing home [11, 12]. Longitudinally, well-being can be expected to decrease in the years before relocation to a nursing home as accumulating risks of health issues have been suggested to negatively impact well-being [13, 14]. This leads us to the question of whether older adults develop their expectations to relocate to a nursing home in line with experienced changes in health and well-being.

Especially when losses are inevitable, forming realistic expectations of future challenges can be seen as an adaptive strategy [7‒9]. Based on personal experiences older adults actively construct scenarios of future old age [15], by which they imagine what life in a nursing home would look like. Forming these expectations can be seen as a psychological preparation for the transition to a nursing home, which facilitates acceptance of corresponding future challenges [8, 9, 16].

Excluding instances in which relocation is related to sudden events (e.g., hospitalization), relocation is often considered an outcome of a longer period of anticipation, preparation, and decision-making [17, 18]. Hence, older adults who experience stronger declines in health and well-being may expect relocation to a nursing home in the coming years to become more likely. Research suggests that expectations to relocate are rationally based as these expectations are linked to personal characteristics and circumstances. Previous work showed that individuals with worse health status (e.g., lower SRH, more health-related adverse events, more functional limitations, more frequent falls, and worse cognitive functioning) also indicated higher expectations to relocate to a nursing home [19‒21]. Although measures of well-being were not included in the aforementioned studies, accumulating health issues before relocation might adversely affect well-being [13, 14, 22], and hence be related to expectations to relocate to a nursing home. As previous work was mainly cross-sectional, it remains unclear whether changes in health status or well-being are related to changes in the expectation of relocation over time.

Expecting future challenges also serves an adaptational purpose after a specific challenge takes place. Someone who forms expectations is more likely to accept this future scenario, even before the challenge occurs [8, 9, 16]. After relocation, these expectations potentially minimize the negative impact of the event, by diminishing the shock, stress, or disappointment that often accompany such challenges [15].

Emerging evidence indicates that having more realistic expectations is associated with better health and well-being. Overestimations of future health or well-being were associated with less positive and more negative affect [23, 24], as well as a higher risk of disease [25]. Yet, more realistic expectations of future old age were associated with higher subsequent psychological well-being [15] and a lower risk of future disability and mortality [8]. Because our study focuses specifically on those individuals who eventually relocate to a nursing home, not expecting this relocation may imply that they overestimated their ability to continue living at home in the coming years. In contrast, expecting relocation would indicate a more realistic view of their future, leading to better adaptation after moving to a nursing home and potentially positively impacting outcomes after this event.

Few longitudinal studies have examined changes in health and well-being among nursing home residents. Functional limitations and SRH were found to increase and decline, respectively, among older adults up to 2 years after a residential relocation [26, 27]. Yet, these patterns were similar to within-person changes before the event and among older adults who did not relocate across a 6-year time interval [26‒28]. Regarding well-being, relocation to a nursing home was associated with short-term decreases up to 2 years after admission [29, 30]. Afterward, well-being among nursing home residents increased, yet, not completely returning to pre-admission levels across a 20-year follow-up [30]. In addition to these mean-level changes, people likely differ in the direction and degree of change in health and well-being. We contribute to previous work by examining whether the expectation to relocate to a nursing home may partly explain these individual differences, and buffer against decreases in health and well-being after this transition.

In line with the lifespan approach to successful aging, realistic expectations of future old age potentially constitute adaptive strategies to anticipate and adjust to future challenges. Yet, this framework has not yet been applied in the specific context of relocation to a nursing home. This pre-registered study contributes to previous work by answering the following two research questions. First, are changes in functional limitations (i.e., ADL and IADL), SRH, and life satisfaction associated with changes in the expectation of relocation to a nursing home? Based on previous research, we expected functional limitations (H1a) to show mean-level increases in the years before moving to a nursing home and SRH (H1b) and life satisfaction (H1c) to show mean-level decreases in the years before moving to a nursing home. Moreover, we hypothesized that higher initial levels of (and increases in) functional limitations (H1d; H1g) and lower levels of (and decreases in) SRH (H1e; H1h) and life satisfaction (H1f; H1i) before the transition are associated with higher initial expectations of moving to a nursing home. Second, is the expectation of relocation associated with functional limitations, SRH, and life satisfaction after the transition to a nursing home? We hypothesized that a higher level of expectation of relocation is associated with more positive changes in SRH and life satisfaction (i.e., increases, less steep declines, or stability) after relocation (H2a; H2b). Since people are less likely to be able to change their functional limitations by adaptive strategies, we did not expect expectations to relocate to be associated with changes in functional limitations after relocation.

By answering these research questions, we aim to inform theories on successful aging and adaptive strategies by providing insights into long-term dynamic changes in expectations to relocate to a nursing home, as well as the associated changes in health and life satisfaction before and after this important life transition. These insights can function as a starting point for future research (e.g., intervention studies) on more specific adaptive strategies among older adults who are likely to move to a nursing home in the near future, to facilitate adjustment after the transition [31, 32].

We used data from the Health and Retirement Study (HRS) [33, 34], a biennial panel survey of a US nationally representative sample of adults aged 50 years and older. Data collection was initiated in 1992 and is ongoing. Interviews take place in person or by telephone on topics such as demographics, health, psychosocial characteristics (e.g., well-being, life events, personality), and expectations. New cohorts are added every 6 years. The HRS applies multi-stage sampling based on geographical stratification and clustering and oversampling of particular demographic groups. The HRS was approved by the Institutional Reviewing Board at the University of Michigan and the National Institute on Aging (HUM00061128). The study is assisted by the National Institute on Aging (NIA U01AG009740) and the Social Security Administration and is conducted by the University of Michigan [35].

We selected older adults who were not living in a nursing home at their first assessment and relocated permanently to nursing homes over time. Individuals younger than 65 years old were excluded from our analyses because the question regarding the expectation to relocate to a nursing home was worded differently for them. The data were restructured such that relocation to a nursing home took place at the same time for all respondents (see online suppl. material Fig. S1; for all online suppl. material, see https://doi.org/10.1159/000541336). Two samples were created to answer our research questions. (a) The pre-relocation sample (n = 1,048) consisted of individuals with at least two assessments of each indicator across a maximum of four waves (7 years) before relocation. (b) The relocation sample (n = 307) consisted of individuals with information on all of our indicators in the wave before relocation and with at least one assessment for functional limitations, SRH, and life satisfaction across a maximum of three waves (5 years) after relocation. In this way, we captured within-person changes from the last wave before relocation up to the last wave after relocation.

In the last wave before relocation, the pre-relocation and relocation samples included respondents aged 71–107 and 69–98 years old, with an average of 84.65 (SD = 6.27) and 84.70 (SD = 6.24) years of age, respectively. The pre-relocation sample included 405 males and 643 females and the relocation sample included 87 males and 220 females (38.6% and 28.3% males, respectively). The samples had an average of 12.47 (SD = 2.94) and 12.81 (SD = 2.66) years of education (see online suppl. material S1 for details on sample selectivity, attrition, and response rates).

Participants were asked to rate their chances of moving to a nursing home in each survey wave if they were not living in a nursing home at that time: “What is the percent chance that you will move to a nursing home in the next 5 years?” Nursing homes were defined for the respondents as institutions for individuals who are unable to live independently and need non-stop nursing supervision. Individuals rated this question from 0% to 100%, with 0 indicating absolutely no chance and 100 indicating absolute certainty of relocation within the next 5 years. Since all older adults in our study relocate to a nursing home over time, this indicator informs us about the extent to which they realistically expect this event to take place in the future. Expectations to relocate to a nursing home were related to the probability of actual entry and decisions about long-term care insurance [36]. Measures of subjective probabilities have been incorporated and studied frequently in numerous household surveys (e.g., HRS, ELSA, SHARE) and have considerable predictive power [37].

Based on the Katz ADL and Lawton IADL scales [38, 39], ADL and IADL limitations were assessed using six and five items, respectively. To measure ADL limitations, participants were asked whether or not they had “any difficulty” with the activities of getting dressed, walking across a room, bathing, eating, getting in and out of bed, and using the toilet. For IADL limitations questions referred to preparing a meal, shopping for groceries, making phone calls, taking medications, and managing money. By definition, impairments in ADL indicate a more progressed level of disability, while limitations in IADL signal earlier decline in the performance of daily activities [40]. Initial answering categories included (1 = yes, 5 = no, 6 = can’t do, 7 = don’t do). The responses were dichotomized into having 0 = no difficulty and 1 = having difficulty, where the original score of can’t do was appended to having difficulty. Mean scores were calculated only over the activities that people generally would perform, disregarding activities they reported as don’t do (e.g., some men might have never performed traditionally female tasks such as preparing a meal) [41]. The mean scores were multiplied by 10 for convenience of interpretation later on. Since these measures of functional limitations reflect a formative construct (i.e., a composite score of various limitations) rather than a reflective one (i.e., a latent disability factor), we did not calculate Cronbach’s alpha coefficients for these measures [42].

SRH was used as a subjective indicator of health [43], based on self-reported answers to the question: “Would you say your health is excellent, very good, good, fair, or poor?” Scores were reverse coded to make higher scores reflect higher SRH (0 = poor to 4 = excellent).

Global life satisfaction, which focuses on how people evaluate their lives as a whole based on broad life domains and general life circumstances [44], was included as a measure of well-being. Based on self-reported answers to the question: “Please think about your life as a whole. How satisfied are you with it?”, life satisfaction was assessed in every wave except for 2006. Answering categories were reverse coded, making higher scores reflect higher life satisfaction (0 = not at all satisfied to 4 = completely satisfied). Previous research indicates that this measure has good psychometric properties (e.g., good reliability and validity) and correlates substantially with multiple-item measures. Moreover, the one- and multiple-item assessments similarly correlated with theoretically relevant factors such as subjective health and affect [45].

The pre-registration, supplementary materials, and analyses scripts are available in an OSF repository: https://osf.io/gt9cb/. Data preparation was done in R version 4.2.1. All analyses were performed using structural equation modeling (SEM) in Mplus [46]. Models were estimated using full information maximum likelihood (FIML) to account for missingness in the variables of interest. Below, our statistical procedure is documented for both research questions separately. For each question, we have subsequently modeled univariate and bivariate changes.

For the first research question, levels and changes in health and life satisfaction were linked to levels and changes in expectations of relocation to a nursing home before relocation. We corrected the non-normal distribution in the expectation measure (see Table 1) by estimating zero-inflated negative binomial growth curve models (ZINB LGMs; see online suppl. material S2).

First, we fitted univariate LGMs for each indicator separately. We compared models with no change, linear change, and quadratic change using the Akaike Information Criterion (AIC) (see online suppl. material S3). Based on the best-fitting models, we assessed the mean-level changes of functional limitations, SRH, and life satisfaction before relocation (H1a, H1b, H1c). Thereafter, bivariate LGMs were estimated to explore associations between initial levels of functional limitations, SRH, and life satisfaction with initial levels of expectations to relocate to a nursing home (H1d, H1e, H1f), as well as correlated changes of functional limitations, SRH, and life satisfaction with the expectation of relocation (H1g, H1h, H1i).

For the second research question, the expectation to relocate (i.e., 0–100%) in the last wave before relocation was modeled as a stable predictor of changes in health and life satisfaction after relocation. To correct for the zero-inflation of expectations to relocate, we included a dichotomous control variable of expectations (i.e., zero vs. positive expectations).

We fitted univariate LGMs for the indicators of health and life satisfaction after relocation using the same approach as for the first research question (see online suppl. material S3). Changes in health and life satisfaction were assessed from the last wave before relocation up to the last wave after relocation included in this study. Subsequently, we built on the best-fitting univariate models, and estimated bivariate LGMs in which we examined the associations between the level of expectations and changes in the functional limitations, SRH, and life satisfaction (H2a, H2b).

Table 1 displays descriptive information on expectations to relocate to a nursing home. A substantial proportion of older adults reported absolute certainty to not relocate to a nursing home in the next 5 years even in the wave right before relocation (i.e., around 36% and 30% for the pre-relocation and relocation samples, respectively). Moreover, expectations to relocate were generally low across all waves before relocation, as shown by the median and third quartile scores. Table 2 includes descriptive information of the variables of interest in the last wave before relocation. On average, both samples reported few limitations in ADL and IADL and a fair-to-good SRH. Moreover, in both samples, older adults were somewhat-to-very satisfied with their lives as a whole.

Table 3 shows the parameter estimates for the best-fitting univariate LGMs. Parameter estimates indicated linear mean-level increases in expectations to relocate to a nursing home across the years before relocation. Hence, expectations to relocate to a nursing home increased in the years before relocation, and these changes were found to be similar across older adults (see online suppl. material Fig. S5). Although the linear model with individual differences in change provided the best fit, we fixed the variance around the linear slope because it was not significant and caused non-convergence in subsequent bivariate LGMs.

Both measures of functional limitations showed quadratic mean-level increases in the years before relocation, demonstrating that limitations in basic and instrumental ADLs increased, with steeper increases closer to relocation (see online suppl. material Fig. S6). SRH and life satisfaction showed linear mean-level decreases in the years before relocation (see online suppl. material Fig. S7). Individual variability in change was observed for both domains of functional limitations and SRH, while changes in life satisfaction were consistent across older adults.

Bivariate LGMs were used to assess associations of initial levels of functional limitations, SRH, and life satisfaction with initial levels of expectations before relocation (see Table 4; online suppl. material Fig. S8). Since we found no individual differences in change in expectations to relocate to a nursing home, correlated changes were not assessed. More initial limitations in ADL and IADL were associated with higher initial expectations to relocate to a nursing home. In addition, poorer initial SRH was associated with higher initial expectations to relocate to a nursing home. In contrast, initial levels of life satisfaction were not found to be related to initial levels of expectations to relocate to a nursing home.

Table 3 shows the parameter estimates for the best-fitting univariate LGMs. The best-fitting models for ADL and IADL showed quadratic mean-level increases and linear mean-level increases after relocation, respectively (see online suppl. material Fig. S6). For ADL, the negative quadratic slope indicated less steep increases as time elapsed following relocation. SRH and life satisfaction showed linear mean-level decreases after relocation to a nursing home (see online suppl. material Fig. S7). Individual differences in change were found for functional limitations as well as for SRH but not for life satisfaction.

Bivariate LGMs were used to assess the associations between levels of expectations with changes in functional limitations and SRH after relocation (see Table 4; online suppl. material Fig. S9). Since there were no individual differences in change in life satisfaction, we could not asses the corresponding bivariate association. The expectation to relocate to a nursing home was not significantly related to changes in ADL, IADL, or SRH after relocation.

This study examined the longitudinal associations of expectations to relocate to a nursing home with health and well-being from up to 7 years before until up to 5 years after relocation. To do so, bivariate LGMs were estimated among older adults who relocated to a nursing home using data from the Health and Retirement Study (HRS).

In the 7 years before relocation to a nursing home, functional limitations increased, while SRH and life satisfaction decreased. These findings are in line with previous studies [1]. Moreover, expectations to relocate to a nursing home were generally low and increased across the years before relocation. As hypothesized and in accordance with cross-sectional findings [19, 20], more functional limitations and poorer SRH were related to higher expectations of relocation.

However, we did not observe interindividual differences in change in expectations to relocate to a nursing home. Due to this restricted variability, no external factors, such as changes in health or well-being, could be linked to changes in expectations to relocate. Hence, contrary to previous literature [15], developing an expectation to relocate to a nursing home does not seem to constitute an adaptive process in response to changing personal circumstances (i.e., declines in health and well-being). The uniform trend that older adults perceive relocation to a nursing home to become more likely over time might be related to internalized age stereotypes. Age stereotypes encompass dominant views of the aging process and characteristics of older adults in general, which can impact expectations of future old age [47]. As such, age stereotypes which reflect increased states of dependency may lead older adults to judge the likelihood of relocation to a nursing home to increase over time. Moreover, demographic population changes including declining fertility rates, increased female workplace participation, and geographic proximity of informal caregivers [48] may put restrictions on informal caregiving networks. As a result, older adults may uniformly expect relocation to become more likely as they age.

Across the years after relocation, the patterns of change in health and well-being mirrored those observed before the move. That is, across a timespan of up to 5 years, functional limitations increased, while SRH and life satisfaction decreased. Surprisingly, and contrary to studies that highlighted the adaptational purpose of expecting future challenges [15, 49], expecting relocation to a nursing home was not associated with changes in health or well-being after relocation.

The absence of an association between expecting relocation and post-relocation health and well-being may suggest that this aspect of future orientation is not pivotal for adapting to relocation to a nursing home. Pro-active engagement with future old age is a process encompassing multiple stages of anticipation [6]. Even though thoughts about future events (e.g., expecting them to occur) are an important first step, they represent a relatively early stage in the anticipatory process. Later stages in this process, involving more instrumental forms of preparation, such as the exploration of various supportive housing options [1, 17], might have a bigger impact on how well older adults adapt to their new living situation in the nursing home.

In addition to these main findings, our study highlights the mismatch of older adults’ perception of the likelihood and imminence of relocation with the actual chances and timing of relocation to a nursing home. Across the years before relocation, a substantial proportion of older adults did not expect to relocate to a nursing home. While this proportion declined closer to the relocation event, this proportion was still substantial in the last wave before relocation (i.e., around 36% and 30% for the pre-relocation and relocation samples, respectively). The unexpectedness of relocation to a nursing home has been documented before [19]. Yet, this finding is particularly striking in our study which exclusively included older adults who relocated to a nursing home, unlike other studies. Older adults may be hindered for considering this type of long-term care arrangement in advance since relocation to a nursing home is often dreaded or prompted by a crisis event such as hospitalization [17, 50].

Our study has contributed to the field by examining the adaptive potential of having realistic expectations of future old age (i.e., expecting relocation to a nursing home). By use of large-scale panel data, we were able to model changes from up to 7 years before until up to 5 years after relocation. As a result, we were able to provide insights into long-term developments in expectations, health, and well-being before and after the transition to a nursing home. In addition, by using a variety of measures of health and well-being we provided a more comprehensive overview of how different aspects of health and well-being were associated with expectations to relocate to a nursing home across the years before and after relocation. Another strength of our work is the use of advanced statistical modeling techniques, which enabled us to adequately assess developments (i.e., between-person variability and within-person change) in the expectations to relocate to a nursing home.

Yet, some limitations of this study have to be taken into account when putting current findings into perspective. First, we should be cautious with generalizing our findings to the general population of older adults who relocate to a nursing home. As shown in previous work, older adults who experience worse health and well-being participate in surveys less often and for shorter periods of time [51]. Average descriptive statistics indicated that our samples were relatively healthy (i.e., high SRH and few functional limitations) and higher response rates were found to be related to less functional limitations, better SRH, better cognitive functioning and younger age. Especially because relocation to a nursing home is characterized by increasing states of frailty and dependency [1, 10], we may have therefore underestimated declines in health and well-being and increases in expectations to relocate. Yet, only small differences between our samples were revealed for IADL and cognitive functioning, indicating that the impact of selection effects of the relocation sample was limited. Moreover, FIML estimations corrected for potential selective attrition bias.

Further, the time interval of 2 years between measurement occasions limited our ability to assess dynamic and sudden changes in health before relocation (e.g., health crises), short-term changes in well-being after the event, and their associations with expecting relocation across this intricate process. The largest impact of relocation to a nursing home is generally observed within the first 2 years after the relocation event [26, 27, 29, 30]. Therefore, future study designs could benefit from a more fine-grained time scale, including weeks or months. This is not only crucial when considering the potential negative impacts of relocation to a nursing home [2] but also when examining the potential beneficial outcomes of the anticipation of age-related challenges [15, 49].

Future studies should consider more factors in the anticipation period of relocation to a nursing home to further deepen our understanding of anticipatory and adaptational processes before and after this age-related challenge. Although subjective probability measures have considerable predictive power, they are also known for their potential methodological limitations [37]. One limitation is the tendency for people to report an expectation of zero. We corrected for these excess zeros by using ZINB LGM. Another limitation is that this general one-item measure does not inform us about more specific adaptational strategies used by older adults. Future work can include more detailed longitudinal measures to better explain and understand the anticipatory process before relocation and the outcomes thereof. More instrumental forms of preparation for older age hold particular significance for future research, including, for example, the pro-active exploration of different types of care provision (e.g., moving in closer proximity to children; arranging formal care at home) and weighing alternative supportive housing options (e.g., assisted living facilities or nursing homes) [1, 17]. Moreover, future research could assess other factors that may relate to expectations of relocation to a nursing home. Relocation to a nursing home is an intricate process, expectations of which are also affected by proximity to potential informal caregivers, having (had) parents in a nursing home, educational level, marital status, and homeownership [18‒21].

The minimization of the impact of losses is imperative to age successfully, especially in advancing old age. Among older adults who relocated to a nursing home, a substantial proportion did not expect to do so. As such, older adults did not have a realistic expectation of the likelihood and imminence of a future relocation to a nursing home. Although expectations to relocate increased across the years before relocation, these changes were not related to changes in health and life satisfaction. Moreover, current findings indicate that expectations to relocate to a nursing home did not impact changes in health or life satisfaction after relocation. The absence of a link with changes in health and well-being suggests that the expectation of a future relocation does not constitute an anticipatory or adaptational process before or after relocation. Future studies should assess more pro-active strategies for this transition in old age to further facilitate successful aging in the face of age-related challenges.

The Health and Retirement Study was approved by the Institutional Reviewing Board at the University of Michigan and the National Institute on Aging (HUM00061128). Written informed consent was obtained from all participants.

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Emmie A. M. Verspeek played the leading role in writing – original draft, methodology, formal analysis and an equal role in conceptualization, and writing – review and editing. Yvonne Brehmer played an equal role in conceptualization and writing – review and editing and a supporting role in supervision. Joran Jongerling played a supporting role in methodology, formal analysis, and writing – review and editing. Alexandra Hering played an equal role in conceptualization and a supporting role in writing – review and editing and supervision. Manon A. van Scheppingen played a leading role in supervision, an equal role in conceptualization and writing – review and editing, and a supporting role in methodology and formal analysis.

OSF link: https://osf.io/gt9cb/

Data from the Health and Retirement Study 2006–2018 are publicly available at https://hrs.isr.umich.edu/about. The pre-registration and scripts of this study are available in an online OSF repository: https://osf.io/gt9cb/.