Skip to main content
NCBI home page
VA Author Manuscripts logoLink to VA Author Manuscripts
. Author manuscript; available in PMC: 2024 Dec 16.
Published in final edited form as: J Am Geriatr Soc. 2024 Aug 13;72(10):3098–3108. doi: 10.1111/jgs.19115

Is more care recipient time at home also a family caregiver-centered quality of life measure?

Courtney H Van Houtven 1,2,3, Karen M Stechuchak 1, Paul A Dennis 1,2, Kasey Decosimo 1, Chelsea L Whitfield 1, Nina R Sperber 1,2, S Nicole Hastings 1,2,5,6,7, Megan Shepherd-Banigan 1,2,3, Brystana G Kaufman 1,2,3, Valerie A Smith 1,2,4
PMCID: PMC11647572  NIHMSID: NIHMS2035634  PMID: 39136596

Abstract

Background

Time in healthcare facilities is associated with worse patient quality of life (QoL); however, impact on family caregiver QoL is unknown. We evaluate care recipient days not at home—days in the emergency department (ED), inpatient (IP) care, and post-acute care (PAC)—to understand how care recipient days not at home correspond to family caregiver QoL.

Methods

Secondary data were linked to care recipient utilization data. Elastic net machine learning models were used to evaluate the impact of a single day of utilization in each setting on binary QoL outcomes. We also compared composite weighted and unweighted “days not at home” variables. Two time periods, 6 and 18 months, were used to predict three caregiver QoL measures (self-rated health, depressive symptoms, and subjective burden).

Results

In the 6-month timeframe, a single day of ED utilization was associated with increased likelihood of poor QoL for all three assessed outcomes (range: 1.4%–3.2%). A day of PAC was associated to a modest degree with increased likelihood of caregiver burden (0.2%) and depressive symptoms (0.1%), with a slight protective effect for self-rated health (−0.1%). An IP day had a slight protective effect (−0.2 to −0.1%). At 18 months, ED and IP had similar, albeit more muted, relationships with caregiver burden and depressive symptoms. PAC had a slight protective effect for caregiver burden (−0.1%). Cumulative days in all settings combined generally was not associated with caregiver QoL.

Conclusion

Whereas total care recipient time away from home had some negative spillovers to family caregivers, the countervailing effects of unique settings on caregiver QoL may mask net QoL effects. This finding limits the utility of a single care recipient home time measure as a valid caregiver-centered measure. Considering cumulative care recipient time in individual settings separately may be needed to reveal the true net effects on caregiver QoL.

Keywords: family caregiver, caregiver-centered outcomes, veterans, quality of life, utilization

INTRODUCTION

Older adults prefer to remain at home, but is this good for their family caregivers? We and others have established that more time in acute and post-acute healthcare settings for older adults is associated with their own worse quality of life (QoL).1–3 Measures of home time, or days alive and not in such settings, are therefore patient-centered. But we do not know if this measure is also caregiver-centered. If caregiver QoL is affected by recipient time at home, it means that this measure reflects the intertwined QoL of the dyad. Capturing the QoL of both dyad members is vital to understand the whole health of caregivers and care recipients.

Health payment models are increasingly considering caregiver QoL, such as the Guiding and Improved Dementia Experience (GUIDE) CMMI model,4 and measures like home time, with low reporting burden, are needed to track these phenomena. Since approximately 75% of older adults with disability have a family caregiver,5 interventions that affect care recipient time at home likely affect their caregivers. Understanding whether less care recipient time in healthcare settings increases caregiver QoL is important to quantify positive spillovers of patient interventions that support more home time.6–8 If more care recipient time in settings by contrast increases the caregiver’s QoL by providing respite, it would indicate that interventions that hasten discharge may have negative spillovers to caregivers. Positive and negative caregiver spillovers should be considered in evaluating the net benefits of interventions because caregiver QoL is essential to the provision of high quality of care and sustainment of caregiving. Finally, it is unknown whether different settings have differential impacts on caregivers.

We evaluate features of a care recipient days not at home measure—days in the emergency department, inpatient care, and post-acute care—to understand how care recipient days not at home correspond to family caregiver QoL. If related to caregiver QoL, these measures could capture spillovers of interventions to family caregivers and suggest targets for future interventions concerned with caregiver and care recipient QoL.

METHODS

Overview

We take a similar approach to how we examined the patient-centeredness of a home time measure,1 turning the focus onto family caregivers of older adult veteran care recipients. We first evaluate short- and long-term measures of healthcare utilization—days in the emergency department (ED), inpatient (IP) care, and rehabilitation in a post-acute care (PAC) facility—to understand how care recipient days not at home correspond to caregiver-reported QoL outcomes. These settings were selected based on the settings included in patient measures of home time and supporting evidence about these settings for patients.9 Clinical studies document negative impacts of time spent in acute settings and PAC settings and that they signify loss of control to patients.9–11 Because caregivers also care about care recipient QoL, we surmised that the negative impacts for care recipients could harm caregiver QoL. And yet, if one of these settings is not related to any domain of caregiver QoL, including that setting may not add value to a caregiver-centered “home time” measure. The two different time frames of veteran healthcare utilization—6- and 18-months—were selected to determine whether short- or long-term utilization is most closely associated with caregiver QoL. We then used the model results to produce weighted measures of utilization at both 6 and 18 months, which we compared to corresponding unweighted raw counts of cumulative days in any of the three settings. The goal was to establish whether different settings merit weighting, such that a day away from home in some settings differently affects caregiver QoL compared to other settings. Finally, to complete the picture of whether and which care recipient home-time measure is the most caregiver-centered, we then estimated 6- and 18-month models, with both weighted and unweighted cumulative days separately (four total models for each outcome), to establish whether there is a quantifiable relationship between a care recipient’s cumulative days not at home and caregiver QoL. This study was approved as exempt research by the Institutional Review Board of the Durham Veterans Affairs (VA) Health Care System.

Study cohort and caregiver-reported outcomes

We identified a cohort of caregivers from two existing studies: Helping Invested Families Improve Veteran’s Experiences Study (HI-FIVES, Clinical Trials, NCT01777490) and iHI-FIVES (implementation of HI-FIVES, Clinical Trials, NCT03474380). We pooled these two datasets to create a combined dataset of unique caregiver-level data. Pooling was possible because common measures existed across studies and the studies used common inclusion and exclusion criteria. HI-FIVES was a single-site randomized controlled trial (RCT) that evaluated a caregiver skills training offered to friend or family member caregivers of veterans referred to home- and community-based services.12 iHI-FIVES was a multi-site RCT evaluating the effectiveness and implementation of HI-FIVES for similar caregivers at eight VA medical centers.13 For both studies, baseline telephone surveys prior to intervention start collected caregiver QoL measures. These caregiver measures were selected for comparability with other caregiver training/support interventions as they are the most common measures14 of QoL in caregiver interventions and also represent those most sensitive to change. Details on eligibility and recruitment of HI-FIVES12 and iHI-FIVES15 have been previously published. For the current study, we removed 7 iHI-FIVES participants, as they had also enrolled in HI-FIVES, and 32 iHI-FIVES participants with incomplete veteran healthcare utilization at the time of dataset construction.

Healthcare utilization

To examine home time associations with caregiver-reported QoL outcomes, veterans’ utilization prior to caregivers’ participation in the telephone-based surveys was defined as the separate counts of the number of days in ED (ED visits where one or multiple visits on the same day count as 1 day), IP (e.g., days in an observation, medical, or surgical unit in a hospital setting), and PAC (facility-based short-term nursing home days and inpatient rehabilitation days) settings using administrative claims data from VA health records (Corporate Data Warehouse and Observational Medical Outcomes Partnership), VA-purchased community care (Fee Basis and Program Integrity Tool data sources), and Centers for Medicare and Medicaid Services [CMS] (Medicare Provider and Analysis Review, Outpatient, and Minimum Data Set files). Days in these settings were compiled for two time periods: 6- and 18-months. CMS files provided utilization for fee-for-service and Medicare Advantage beneficiaries for IP and PAC (30% of the sample had at least 1 month of Medicare Advantage enrollment). We were unable to capture ED visits paid for by Medicare Advantage alone, though those provided by or with payment from VA were included. Inpatient respite stays, which are a VA-covered service that provides short-term care when family caregivers need a break to run errands or travel without their veteran care recipient, were excluded from IP days because these days are often planned and do not reflect a health-driven patient event. When combined for the cumulative days not at home measure, any given calendar day was considered not at home if spent in any of the three settings; no days were double-counted. Because the utilization occurred before study enrollment, all care recipients lived for the full utilization window; thus, we have complete capture for the intervals of interest.

Measures

The primary outcomes were three caregiver-reported measures of QoL collected at the time of the survey. Self-rated health has been mapped directly to health utility measures, which reflect global health-related QoL.16–19 Depressive symptoms and caregiver subjective burden, the latter of which reflects the strain and stress of the caregiving role,20 are each closely linked to global QoL for caregivers.21–28

Caregiver self-rated health.

Self-rated health29 was collected from both studies: “In general, how would you say your health is now?” Responses of Fair or Poor were coded as “1” and “0” if Excellent, Very Good, or Good.

Caregiver depressive symptoms.

We created a binary variable for positive for depressive symptoms since the two studies employed different validated depressive symptom screening measures. CESD-10 was the depressive symptoms measure in HI-FIVES.30 Scores ≥10 were coded as “1” (positive for depressive symptoms) and “0” if <10.30 The PHQ-2 depressive symptoms screening measure was used in iHI-FIVES.31 Scores ≥3 were coded as “1” (probable depressive symptoms) and “0” if ≤2.31

Caregiver subjective burden.

The Zarit four-item measures caregiver subjective burden as the level of stress felt by a caregiver.32 Scores range from 0 to 16, with higher scores reflecting higher burden. Across both studies, we created a binary threshold using the four-item Zarit burden measure, with scores ≥8 coded as high burden.32

Model covariates

Aside from the utilization variables, the following additional caregiver covariates were included for adjustment in each model: gender, age at time of the survey, ethnicity, race, whether highest level of education is high school or less, whether the caregiver was caring for a spouse or a non-spouse, number of deficits (unable to do or needing some help) for Activities of Daily Living (ADL) and Instrumental Activities of Daily Living (IADL)33 for the care recipient, and whether the care recipient lived in an urban area 18 months prior to the survey.

Data analysis

To assess how components of home time, ED, IP, and PAC utilization were associated with the three measures of caregiver QoL, we used elastic net regression.34 Elastic net uses a combination of LASSO regularization, which limits the number of variables in a given model, and ridge regularization, which limits the size of coefficients. We used this approach because it is robust to collinearity and deters overfitting to the sample at the cost of slight bias in estimated effects. Due to this bias, standard errors are not reported for elastic net regression. We partitioned the data into a training and test set along a 70/30 split. Within the training set, we used 10 repetitions of 5-fold cross-validation to determine optimal hyperparameters, with the dataset randomly re-ordered between repetitions. The combination of hyperparameters that provided the highest level of agreement between predicted and observed outcome values, measured as Cohen’s kappa, were retained for the final model, which was estimated using the entire training set. The resulting coefficients were then used to generate predictions for the test set along with indices of model fit (i.e., Cohen’s kappa). In sum, six primary models were analyzed: 6- and 18-month utilization models for each of the three QoL outcomes.

After completing the above models, we conducted a second set of analyses comparing an unweighted count of days in any healthcare setting with a weighted count. To generate a weighted count, we transformed the coefficients from the elastic net models to reflect the impact of a single day in each healthcare setting. We then averaged the resulting transformed coefficients across the three outcomes to derive weights specific to each type of utilization. Once completed, we calculated a single weighted utilization score for each veteran care recipient via a linear combination of their three raw utilization day counts and the corresponding weights. To determine how this weighted utilization score compared to the unweighted count of number of days spent in any of the ED, IP, and PAC settings, we used unpenalized logistic regression. Specifically, for each QoL outcome and timescale (i.e., 6- and 18-month) we modeled the effects of the two variables (weighted and unweighted cumulative days in these settings) in separate models, using the same adjustment covariates used in the previous models. To facilitate interpretation of the weighted utilization score, we scaled it so that its variance matched that of the corresponding unweighted utilization score. We then used a two-sided z-test to compare the two resulting effects.35

Missing values resulted in case-wise deletion. Primary analyses were conducted using R, version 4.1.2. Elastic net logistic regression and repeated k-fold cross-validation were performed using the “glmnet” and “caret” packages, respectively. Estimated marginal effects (EME) are reported representing the percentage point difference in the predicted probability of an outcome level (vs the reference level) associated with a one-unit change in the predictor variable (in its original scale), with baseline outcome probability levels derived from observed rates. Thus, EMEs represented the average effect of 1 day of utilization on the probability of a negative QoL outcome.

RESULTS

Descriptive results

Caregivers were predominantly female and spouses of the veteran, had a mean age of 62, and approximately 27% had a high school education or less (Table 1). Thirty-five percent of both caregivers and veterans reported Black race. Veterans were predominantly male and White, with a mean age of 75, and around one third lived in a rural area. The average number of ADL and IADL deficits for veterans was 4.2 and 5.6, respectively. Just under one third of caregivers reported fair or poor health, around 41% had high subjective burden, and just under 30% screened positive for depressive symptoms.

Table 1.

Cohort Characteristics (N = 522) of Caregivers and their Care Recipient Veterans.

Caregiver Veteran
Female gender, n (%) 470 (90.0) 18 (3.4)
Hispanic/Latino ethnicity1, n (%) 23 (4.4) 16 (3.1)
Race2, n (%)
 White 303 (58.0) 317 (60.7)
 Black 183 (35.1) 185 (35.4)
 Other 36 (6.9) 20 (3.8)
Age in years9, mean (SD) 62.3 (12.9) 75.3 (12.2)
Married or living as married, n (%) 395 (75.7) 348 (66.7)
Highest level of education is high school or less9, n (%) 139 (26.6) 227 (44.1)
Caregiver/Veteran relationship, n (%)
 Spouse/significant other 317 (60.7)
 Caregiver is child/grandchild (including in-law and step) 150 (28.7)
 Caregiver is parent/parent-in-law 8 (1.5)
 Sibling 18 (3.4)
 Caregiver is niece or nephew 8 (1.5)
 Other familial relationship or friend or 21 (4.0)
Rural (vs. suburban or urban) status, n (%) 182 (34.9)
No. ADL deficits3, mean (SD) 4.2 (2.1)
No. IADL deficits4, mean (SD) 5.6 (1.7)
No. ADL and IADL deficits5, mean (SD) 9.8 (3.4)
Caregiver-centered outcomes
 Self-Rated health fair/poor6,9, n (%) 171 (32.9)
 High caregiver burden7,9, n (%) 212 (40.8)
 Positive depressive symptoms 8,9, n (%) 153 (29.6)
Open in a new tab

Note. SD = standard deviation

1

Three veterans and 3 caregivers who reported “don’t know,” refused, or had missing data for Hispanic/Latino ethnicity are categorized as not being of Hispanic/Latino ethnicity for analytical purposes.

2

Race was categorized into 3 levels; White race (only racial category selected), Black race (can be only category selected or have additional racial categories selected), Other race (includes those selecting Asian, Native Hawaiian or Alaskan Native, Native American, Other race, multiple race categories selected [except for those who also selected Black race]). One veteran and 3 caregivers had missing for race and was included in the “Other” race category for analytical purposes.

3

Number of deficits (unable to do or needing some help) for 7 activities of daily living (ADL).

4

Number of deficits (unable to do or needing some help) for 7 instrumental activities of daily living (IADL).

5

Number of deficits (unable to do or needing some help) for 7 ADL and 7 IADL.

6

Self-rated health collected on a 5-point scale, with 1=Excellent, 2=Very Good, 3=Good, 4=Fair, 5=Poor.

7

Using the 4-item Zarit burden measure collecting scores ranging from 0 to 16, with a binary threshold where a cut-off of ≥8 indicates presence of high burden.

8

Positive depressive symptoms screen was scores of ≥10 for CESD or a score of ≥3 for PHQ-2.

9

Missing data (n): Caregiver age (4), veteran education (7), self-rated health (3), caregiver burden (2), depressive symptoms screener (5).

Almost 62% of the veterans spent at least 1 day in the ED in the 6 months prior to the caregiver survey, and 83% did so in the prior 18 months (Table 2). IP utilization was lower, with around 46% and 65% spending at least a day in an IP setting in the previous 6 months and previous 18 months, respectively. Nearly 20% of veterans spent at least 1 day in PAC in the prior 6 months, and 31% did so in the prior 18 months.

Table 2.

Summary of Veterans’ Healthcare Utilization

Veterans with ≥1 Days of Utilization
day of utilization (%) Mean (SD) Q1-Q3
6 Months
 ED 323 (61.88%) 1.44 (1.77) 0–2
 IP 238 (45.59) 5.62 (12.71) 0–6
 PAC 104 (19.92) 5.25 (14.75) 0–0
 Any 364 (69.73) 11.61 (20.91) 0–13
18 Months
 ED 434 (83.14) 3.57 (3.69) 1–5
 IP 341 (65.33) 12.29 (22.60) 0–15
 PAC 161 (30.84) 16.09 (43.46) 0–16
 Any 456 (87.36) 30.39 (51.85) 3–36
Open in a new tab

Note. SD = standard deviation, Q1-Q3 = range from 1st quartile to 3rd quartile, ED = emergency department, IP = inpatient hospitalization, PAC = post-acute care facility. Any = any utilization of ED, IP, or PAC. Multiple types of care (ED, hospitalization, PAC) can occur on a given day.

Multiple types of care (ED, IP, PAC) can occur on a given day. If an ED visit leads to an IP admission, the ED visit counted in the ED visit total. If the IP started on the day of the ED visit, that day also counted as a day of IP care. Similarly, ‘transition days’ between IP and SNF counted towards both IP and SNF totals. The “Any utilization variable” does not double-count days (as it is only looking at if any of our selected utilization happened on that day). Nursing home hospice was not counted as a day away from home.

Modeled results

Figure 1 (left panel) displays results from the elastic net logistic regression models of 6-month healthcare utilization. According to the indices of model fit (Cohen’s kappa values) derived from applying the final models to the test set, all three models demonstrated little to very slight agreement among predicted and observed outcomes (range: 0.02–0.15) (Appendix Table 1). However, ED utilization was associated with worse caregiver QoL in all three models and was the strongest indicator among the utilization variables based on coefficient size in all three models (Appendix Table 1). More time in IP was modestly protective, associated with decreased probability of caregiver self-reported health being fair or poor, decreased subjective burden, and decreased depressive symptoms. PAC utilization was associated with increased caregiver burden and depressive symptoms but had a slight protective association with self-rated health.

Figure 1.

Figure 1.

Open in a new tab

Estimated marginal effects for 6-month and 18-month care utilization with caregiver QoL (elastic net logistic regression models). Estimated marginal Effects (EME) represent the average effect of 1 day of utilization on the increased (or decreased) probability of a negative quality of life (QoL) outcome.

Figure 1 (right panel) displays results from the 18-month utilization models. Overall, the magnitude of the associations between care setting and caregiver measures is smaller in the 18-month models. All three 18-month models produced little to no agreement between predicted and observed outcome values (Cohen’s kappa range: 0.00–0.13) (Appendix Table 2). None of the utilization variables were associated with caregiver self-reported health. As in the 6-month models, ED was associated with increased caregiver burden and depressive symptoms, whereas IP was associated with modestly decreased burden and depressive symptoms. PAC utilization demonstrated slight protective effects for caregiver burden but no association with the other outcomes.

To compare weighted utilization scores (derived from the elastic net models) to unweighted counts of any utilization, we first calculated weights, representing the mean effect of each type of utilization across the three outcome variables. We used these as multipliers for days in ED, days in IP, and days in PAC. For the 6-month utilization variables, these weights were 0.1103 for ED, 0.0035 for PAC, and −0.0074 for IP. For the 18-month utilization variables, the weights were 0.0315 for ED, −0.0002 for PAC, and −0.0051 for IP.

In the 6-month models, there was no significant difference between the weighted and unweighted models according to the “Difference test p-value” (Table 3). The 18-month model suggested weaker associations. However, there was a significant difference between the weighted and unweighted models for caregiver burden (Table 3). In the weighted model the strength of association between cumulative “days not at home” and caregiver burden was statistically significant.

Table 3.

Association of Weighted and Unweighted Care Utilization on Caregiver Quality of Life (Logistic Regression Models) (N = 522)

Caregiver Self-Rated Health Caregiver Burden Caregiver Depressive Symptoms
Fair/Poor (32.9%) High (40.8%) High (29.6%)
(missing = 7) (missing = 6) (missing = 9)
6-Month Utilization of Care Recipient
Weighted Utilization Day Count
 EME (95% CI) 0.2% (−0.0–0.3%) 0.2% (−0.0–0.4%) 0.1% (−0.0–0.3%)
 p-value .083 .057 .091
Unweighted Utilization Day Count
 EME (95% CI) −0.0% (−0.2–0.2%) 0.0% (−0.2–0.2%) −0.1% (−0.2–0.1%)
 p-value .90 .85 .52
Difference test p-value .19 .22 .10
18-Month Utilization of Care Recipient
Weighted Utilization Day Count
 EME (95% CI) 0.1% (−0.0–0.1%) 0.1% (0.0–0.2%) 0.1% (−0.0–0.1%)
 p-value .19 .040 .15
Unweighted Utilization Day Count
 EME (95% CI) 0.0% (−0.1–0.1%) −0.0% (−0.1–0.0%) 0.0% (−0.1–0.1%)
 p-value .78 .29 .99
Difference test p-value .43 .026 .29
Open in a new tab

Note. Model coefficients (coef.) and estimated marginal effects (EME) derived from analysis on full training set (n = 365), minus cases with missing data. EMEs represent the average effect of one day of utilization on the increased (or decreased) probability of a negative Quality of Life (QoL) outcome. For continuous utilization variables, EME were calculated based on original unstandardized scales. Missing values resulted in row-wise deletion. ED = emergency department, IP = inpatient hospitalization, PAC = post-acute care facility.

a

Female caregivers represent reference group.

b

Caregivers 18–55 years represent reference group.

c

Non-Hispanic/Latino(a) caregivers represent reference group.

d

Caregivers caring for someone other than spouse represent reference group.

e

White caregivers represent reference group.

f

Caregivers with more than high school education represent reference group.

g

Rural veterans represent reference group.

DISCUSSION

This study asked a simple question: Is a care recipient “home time” measure reflective of caregiver QoL? If so, it could be a powerful quality measure readily garnered in claims to capture caregiver spillovers from interventions that affect time in healthcare settings. We showed that cumulatively more care recipient time in acute and post-acute settings combined generally was not associated with caregiver QoL. More cumulative days not at home were only associated with worse caregiver subjective burden in the 18-month model; the effect was similar but not significant in the 6-month model. Individual settings were associated with at least two of three QoL outcomes in general. Among the individual settings, care recipient ED time was the most consistent predictor of worse caregiver QoL; thus, it should be included in a caregiver-centered home time measure. Interestingly, we had an unexpected finding that time in IP setting—both in the 6- and 18-month time frames—was protective against caregiver burden and depressive symptoms. PAC had some protective and negative associations with the caregiver QoL measures; even across time-windows associations changed, with more PAC days related to slightly worse caregiver burden at 6 months but slightly better burden at 18 months. It is unclear why inpatient care would be protective while ED care would be destructive and why PAC care would be destructive and protective depending on time window. Perhaps ED care represents a crisis moment, one that settles with a longer duration inpatient stay, providing some respite to the caregiver for a sustained period. And yet, if the patient has to go to PAC instead of home after an IP stay, the relationship turns, and PAC is harmful to caregiver QoL in the short term, possibly because the total time away from home becomes just too long or the transition to PAC specifically is difficult but proffers gains in the longer term.

The magnitude of the relationship with QoL was low for cumulative days across all settings, possibly explained by the variation in magnitudes by setting. That is, there were negative associations of ED days on QoL outcomes that were magnitudes stronger than the protective associations of IP days or PAC days with QoL outcomes. Specifically, considering 1 day of utilization in the 6 month models, ED care was associated with a 2.6 percentage point increase in the probability of high caregiver burden, whereas IP care was associated with a 0.2 percentage point decrease in the probability of high caregiver burden. Thus considering a home time measure overall, there appear to be countervailing effects that are dominated by one setting when there are differing signs, leading to lack of associations between cumulative days and QoL overall (the exception being the 18-month caregiver burden).

In addition, in the caregiver burden 18-month model, there was a significant difference between the weighted and unweighted models; the weighted utilization day count was significantly associated with higher burden, whereas the unweighted utilization day count variable was not. For caregivers, neither self-rated health nor depressive symptoms were sensitive to either weighted or unweighted tallies of veteran total time away from home. Interestingly, these measures may be more reflective of global QoL, whereas subjective burden relates to the caregiver role. Subjective burden may be an important screening measure for caregivers when their care recipient is in healthcare settings, to be able to appropriately guide intervention or supportive services.

There are study limitations. This study was based on secondary survey data; thus, participants may not represent national veteran-caregiver characteristics, and caregivers were primarily spouses, who are often the most intensive caregivers. Veteran care recipients were overwhelmingly male, and caregivers were predominantly female. Additionally, racial diversity was present in the sample by Black race but not other minoritized races/ethnicities, potentially limiting generalizability of results. Although the VA has a relatively high-complexity patient population with high social care needs, such as in our study, it is able to provide long-term care and support services compared to non-veteran populations, which could overall impact family caregiver burden. We were not able to control for caregiver QoL levels prior to the healthcare utilization occurring, which introduces concerns of reverse causality or that the QoL measures were already poor for the caregivers before the veteran utilization occurred and therefore not sensitive to change. Also, with high levels of functional impairments of the veteran patients, time away from home may be very different than in a less impaired dyad population. We did not examine transitions between settings and transitions may partly explain the negative associations of time away from home and caregiver subjective burden and not time away from home itself. In general, many things affect caregiver QoL, not just recent care recipient utilization, which was our study’s focus, and, while we controlled for factors we could observe, such as a veteran’s level of need in functional supports (e.g., ADLs, IADLs), such unobserved factors could be masking the true relationships. Finally, whereas the QoL measures map well to global QoL (e.g., self-rated health) and are associated with global QoL (e.g., subjective burden, depressive symptoms), we did not have additional measures of caregiver QoL (e.g., Euroqol17 or SF-3636) to examine because we used existing secondary data from past trials.

We have a final important caution from our results that we believe severely limits the utility of using a single care recipient home time measure to reflect caregiver QoL. We found protective effects on caregiver burden and depressive symptoms in the IP setting and some of the PAC setting models and negative effects in the ED and some PAC setting models. This argues that spillover effects to caregivers are different based on the utilization setting, so one single home time measure has limited utility for reflecting net caregiver QoL. The weighted and unweighted cumulative results arose because IP and ED days effects cancel each other out. So, someone who enters the hospital via the ED may end up with a weighted count closer to zero because of the positive ED weight and negative IP weight. Then in the unweighted count, their effects cancel each other out in the cumulative regression. Based on the substantial variation in the relationship with QoL by setting, we conclude that a single home time measure is not reflective of the QoL impacts to caregivers. In fact, anything that combines the settings may fail to capture the spillovers because of the protective and destructive relationships with QoL cancel each other out. Instead, more work should examine separately how care recipient time in the ED setting and care recipient time in IP setting affect caregiver QoL.

CONCLUSION

While effects of a single utilization day were strongest within 6 months prior, more cumulative veteran care recipient days in healthcare settings over 18 months was related to worse caregiver subjective burden in a weighted model, suggesting that care recipient time away from home has some negative spillovers to family caregivers. Countervailing relationships with caregiver QoL by setting (e.g., ED vs IP and PAC) substantially limits the ability of a single cumulative measure of care recipient home time to accurately reflect caregiver QoL. Measuring days in these individual settings separately but in tandem would likely be more caregiver-centered than focusing on cumulative care recipient time spent across all settings, because such an approach would not cancel out caregiver spillover effects. Future work with larger samples and caregiver interviews needs to be done to confirm these findings and establish whether the same settings affect care recipient and caregiver QoL in the same ways. Interventions in these settings could then target preventing any negative effects to the members of a dyad individually or jointly.

Supplementary Material

Appendix

Key points

  • It is unclear if patient measures of home time, or days alive and not in acute and post-acute healthcare settings, impact family caregiver quality of life.

  • Based on the substantial variation in the relationship with caregiver quality of life by setting, we conclude that a single home time measure is not reflective of the quality of life impacts to caregivers.

  • Future work with larger samples and caregiver interviews could confirm these findings and establish whether the same settings affect care recipient and caregiver quality of life in the same ways.

Why does this paper matter?

In the United States, approximately 75% of older adults with disability have a family caregiver. This research matters because health payment models are increasingly considering caregiver burden and caregiver quality of life and measures like home time, with low reporting burden, are needed to track these phenomena. Understanding whether less care recipient time in healthcare settings increases caregiver quality of life is important to quantify positive spillovers of patient interventions that support more home time.

ACKNOWLEDGMENTS

We thank the ASPIRE advisory panel for their valuable input and as well as VetREP panel members who advised the larger ASPIRE study. We also thank the family caregiver and veteran participants in the HI-FIVES and iHI-FIVES trials.

SPONSOR’S ROLE

The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The contents do not represent the views of the Department of Veterans Affairs or the United States Government.

Footnotes

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts.

FINANCIAL DISCLOSURE

This research was funded by a Research to Impact for Veterans (RIVR) Award (Grants Log Number: RVR 19-472; PI: Van Houtven). This research was also supported, in part, by the Durham VA Center of Innovation to Accelerate Discovery and Practice Transformation (Grants Log Number: CIN 13-410). Courtney Van Houtven is supported by the U.S. Department of Veterans Affairs, Veterans Health Administration, Office of Research Development, Research Career Scientist Program (RCS-21-137). Megan Shepherd-Banigan is supported by a U.S. Department of Veterans Affairs, Veterans Health Administration, office of Research and Development, Career Development Award (CDA-17-006). Support for VA/CMS Data provided by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development Service, VA Information Resource Center (Project Numbers SDR 02-237 and 98-004). The funding organization was not involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

REFERENCES

  • 1.Dennis PA, Stechuchak KM, Van Houtven CH, et al. Informing a home time measure reflective of quality of life: a data driven investigation of time frames and settings of health care utilization. Health Serv Res. 2023; 58(6): 1233–1244. doi: 10.1111/1475-6773.14196 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Lee H, Shi SM, Kim DH. Home time as a patient-centered outcome in administrative claims data. J Am Geriatr Soc. 2019; 67(2): 347–351. doi: 10.1111/jgs.15705 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Shen E, Rozema EJ, Haupt EC, et al. Assessing the concurrent validity of days alive and at home metric. J Am Geriatr Soc. 2022; 70(9): 2630–2637. doi: 10.1111/jgs.17506 [DOI] [PubMed] [Google Scholar]
  • 4.Services CfMM. Guiding an Improved Dementia Experience (GUIDE) Model. Accessed December 20, 2023. https://www.cms.gov/priorities/innovation/innovation-models/guide
  • 5.Van Houtven CH, Konetzka RT, Taggert E, Coe NB. Informal and formal home care for older adults with disabilities increased, 2004–16. Health Aff (Millwood). 2020; 39(8): 1297–1301. doi: 10.1377/hlthaff.2019.01800 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Ariti CA, Cleland JG, Pocock SJ, et al. Days alive and out of hospital and the patient journey in patients with heart failure: insights from the candesartan in heart failure: assessment of reduction in mortality and morbidity (CHARM) program. Am Heart J. 2011; 162(5): 900–906. doi: 10.1016/j.ahj.2011.08.003 [DOI] [PubMed] [Google Scholar]
  • 7.Edwards ST, O’Neill A, Niederhausen M, et al. Trajectories of care and outcomes of veterans receiving home-based primary care. J Am Geriatr Soc. 2023; 72: 80–90. doi: 10.1111/jgs.18607 [DOI] [PubMed] [Google Scholar]
  • 8.O’Brien EC, Greiner MA, Xian Y, et al. Clinical effectiveness of statin therapy after ischemic stroke: primary results from the statin therapeutic area of the patient-centered research into outcomes stroke patients prefer and effectiveness research (PROSPER) study. Circulation. 2015; 132(15): 1404–1413. doi: 10.1161/circulationaha.115.016183 [DOI] [PubMed] [Google Scholar]
  • 9.Van Houtven CH, Smith VA, Sperber NR, Coffman CJ, Hastings SN. Advancing the science of population-based measures of home-time. Healthcare (Amst). 2020; 8(4):100463. doi: 10.1016/j.hjdsi.2020.100463 [DOI] [PubMed] [Google Scholar]
  • 10.Arya S, Langston AH, Chen R, et al. Perspectives on home time and its association with quality of life after inpatient surgery among US veterans. JAMA Netw Open. 2022; 5(1):e2140196. doi: 10.1001/jamanetworkopen.2021.40196 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Sperber NR, Shapiro A, Boucher NA, et al. Developing a person-centered, population based measure of “home time”: perspectives of older patients and unpaid caregivers. Healthcare (Amst). 2021; 9(4):100591. doi: 10.1016/j.hjdsi.2021.100591 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Van Houtven CH, Smith VA, Lindquist JH, et al. Family caregiver skills training to improve experiences of care: a randomized clinical trial. J Gen Intern Med. 2019; 34(10): 2114–2122. doi: 10.1007/s11606-019-05209-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Wang V, Allen K, Van Houtven CH, et al. Supporting teams to optimize function and independence in veterans: a multi-study program and mixed methods protocol. Implement Sci. 2018; 13(1): 58. doi: 10.1186/s13012-018-0748-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Van Houtven CH, Voils CI, Weinberger M. An organizing framework for informal caregiver interventions: detailing caregiving activities and caregiver and care recipient outcomes to optimize evaluation efforts. BMC Geriatr. 2011; 11: 77. doi: 10.1186/1471-2318-11-77 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Ma JE, Grubber J, Coffman CJ, et al. Identifying family and unpaid caregivers in electronic health records: descriptive analysis. JMIR Form Res. 2022; 6(7):e35623. doi: 10.2196/35623 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Le QA, Doctor JN. Probabilistic mapping of descriptive health status responses onto health state utilities using Bayesian networks: an empirical analysis converting SF-12 into EQ-5D utility index in a national US sample. Med Care. 2011; 49(5): 451–460. doi: 10.1097/MLR.0b013e318207e9a8 [DOI] [PubMed] [Google Scholar]
  • 17.Devlin NJ, Brooks R. EQ-5D and the EuroQol group: past, present and future. Appl Health Econ Health Policy. 2017; 15(2): 127–137. doi: 10.1007/s40258-017-0310-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Law EH, Pickard AS, Xie F, Walton SM, Lee TA, Schwartz A. Parallel valuation: a direct comparison of EQ-5D-3L and EQ-5D-5L societal value sets. Med Decis Making. 2018; 38(8): 968–982. doi: 10.1177/0272989x18802797 [DOI] [PubMed] [Google Scholar]
  • 19.Lee EC, Wright J, Walters SJ, Cooper CL, Mountain GA. Estimating the minimum important difference in the DEMQOL instrument in people with dementia. Qual Life Res. 2021; 30(10): 2995–3005. doi: 10.1007/s11136-021-02900-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Martin MP, McEntee ML, Suri Y. Caregiver quality of life: how to measure it and why. Am J Health Promot. 2021; 35(7): 1042–1045. doi: 10.1177/08901171211030142f [DOI] [PubMed] [Google Scholar]
  • 21.Bowling A Measuring Health. A Review of Quality of Life Measurement Scales. Open University Press; 1997. [Google Scholar]
  • 22.Sivertsen H, Bjørkløf GH, Engedal K, Selbæk G, Helvik AS. Depression and quality of life in older persons: a review. Dement Geriatr Cogn Disord. 2015; 40(5–6): 311–339. doi: 10.1159/000437299 [DOI] [PubMed] [Google Scholar]
  • 23.Wilkinson P, Izmeth Z. Continuation and maintenance treatments for depression in older people. Cochrane Database Syst Rev. 2012; 11:Cd006727. doi: 10.1002/14651858.CD006727.pub2 [DOI] [PubMed] [Google Scholar]
  • 24.Wilson KC, Mottram PG, Vassilas CA. Psychotherapeutic treatments for older depressed people. Cochrane Database Syst Rev. 2008; 23(1):Cd004853. doi: 10.1002/14651858.CD004853.pub2 [DOI] [PubMed] [Google Scholar]
  • 25.Brouwer WB, van Exel NJ, van de Berg B, Dinant HJ, Koopmanschap MA, van den Bos GA. Burden of caregiving: evidence of objective burden, subjective burden, and quality of life impacts on informal caregivers of patients with rheumatoid arthritis. Arthritis Rheum. 2004; 51(4): 570–577. doi: 10.1002/art.20528 [DOI] [PubMed] [Google Scholar]
  • 26.Contreras ML, Mioshi E, Kishita N. Factors related to the quality of life in family carers of people with dementia: a meta-analysis. J Geriatr Psychiatry Neurol. 2021; 34(5): 482–500. doi: 10.1177/0891988720924713 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Sittironnarit G, Emprasertsuk W, Wannasewok K. Quality of life and subjective burden of primary dementia caregivers in Bangkok, Thailand. Asian J Psychiatr. 2020; 48:101913. doi: 10.1016/j.ajp.2019.101913 [DOI] [PubMed] [Google Scholar]
  • 28.Wicks MN, Milstead EJ, Hathaway DK, Cetingok M. Subjective burden and quality of life in family caregivers of patients with end stage renal disease. ANNA J. 1997; 24(5): 527–528; 531–8; discussion 539–40. [PubMed] [Google Scholar]
  • 29.Ann Arbor MUoMNUA. Health and Retirement Study. Accessed July 1, 2024. https://hrs.isr.umich.edu/
  • 30.Andresen EM, Malmgren JA, Carter WB, Patrick DL. Screening for depression in well older adults: evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression Scale). Am J Prev Med. 1994; 10(2): 77–84. [PubMed] [Google Scholar]
  • 31.Kroenke K, Spitzer RL, Williams JB. The Patient Health Questionnaire-2: validity of a two-item depression screener. Med Care. 2003; 41(11): 1284–1292. doi: 10.1097/01.Mlr.0000093487.78664.3 [DOI] [PubMed] [Google Scholar]
  • 32.Bédard M, Molloy DW, Squire L, Dubois S, Lever JA, O’Donnell M. The Zarit burden interview: a new short version and screening version. Gerontologist. 2001; 41(5): 652–657. doi: 10.1093/geront/41.5.652 [DOI] [PubMed] [Google Scholar]
  • 33.Fillenbaum GG, Smyer MA. The development, validity, and reliability of the OARS multidimensional functional assessment questionnaire. J Gerontol. 1981; 36(4): 428–434. doi: 10.1093/geronj/36.4.428 [DOI] [PubMed] [Google Scholar]
  • 34.Zou H, Hastie T. Regularization and variable selection via the elastic net. J R Stat Soc Series B Stat Methodology. 2005; 67(2): 301–320. doi: 10.1111/j.1467-9868.2005.00503.x [DOI] [Google Scholar]
  • 35.Clogg CC, Petkova E, Haritou A. Statistical methods for comparing regression coefficients between models. Am J Sociol. 1995; 100(5): 1261–1293. [Google Scholar]
  • 36.Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992; 30(6): 473–483. [PubMed] [Google Scholar]
  • 37.Van Houtven CH, Coffman CJ, Decosimo K, et al. A stepped wedge cluster randomized trial to evaluate effectiveness of a multi-site family caregiver skills program. Health Serv Res. Forthcoming 2024. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix
NIHMS2035634-supplement-Appendix.pdf (121.5KB, pdf)

RESOURCES