Abstract
Objective:
We compared pain prevalence and intensity between caregivers and non-caregivers, as well as between different types of caregivers.
Method:
Using two rounds of the National Social Life, Health, and Aging Project, we used regression models to analyze differences in pain prevalence and intensity between caregivers and non-caregivers at baseline (2010–2011) and follow-up (2015–2016). The sample consisted of 2,332 participants aged 62 and older (352 caregivers and 1,980 non-caregivers).
Results:
Cross-sectionally, caregivers were more likely than non-caregivers to report pain. Longitudinally, caregivers with pain at baseline were more likely to report the presence of pain and greater pain intensity five years later compared to non-caregivers who also reported pain at baseline. More hours of care per week was associated with greater pain intensity at baseline and follow-up among caregivers with pain.
Discussion:
Caregivers’ pain is a public health concern, with efforts to mitigate the effects of caregiving on pain needed.
Keywords: caregiving, physical pain, National Social Life, Health, Aging Project
Introduction
In the United States, family caregivers serve as the primary source of support for adult family members who need help with daily functioning due to disability and/or old age. Currently, one in five adults in the United States are caregivers (AARP, 2020), with 8 million providing care for an older relative with dementia (Freedman & Wolff, 2024). Family caregiving can challenge caregivers’ physical, mental, and emotional health (Sörensen & Conwell, 2011). Importantly, caregivers’ health is an independent predictor of poor outcomes among their care recipients such as unmet needs (Beach & Schulz, 2017), hospitalization (Kuzuya et al., 2011) and mortality (Pristavec & Luth, 2020; Schulz et al., 2021). Therefore, understanding and improving caregivers’ health constitutes a priority, especially among professionals engaged in supporting older adults (Haley & Elayoubi, 2024; Talley & Crews, 2007). Indeed, the National Alliance for Caregiving (2022) established a framework for family caregiving and public health, which emphasizes the need for greater surveillance of caregivers’ health issues and heightened awareness of caregivers’ challenges among public health officials and practitioners.
The Challenge of Caregivers’ Physical Pain
Physical pain is a highly prevalent and highly consequential, yet understudied, health issue among family caregivers. One in three caregivers to adults in the United States has arthritis, one of the most common causes of pain (Jackson et al., 2022). Among caregivers to older adults, the prevalence of arthritis increases to 40% (Turner, Robinson et al., 2024), which is nearly twice as high as the prevalence in the general population of adults ages 18 and older (22.7%) and over 10% higher than the general population aged 45–64 (29.3%; Barbour et al., 2017). Moreover, over half of all family caregivers to older adults report bothersome physical pain. Among family caregivers who report bothersome pain, 25% indicate that pain limits their activities on at least some days of every month (Turner, Robinson et al., 2024).
Notably, these reports assessed the prevalence of caregivers’ pain among samples of adults of all ages rather than focusing specifically on older adult samples. Yet, over half of all caregivers to older adults are over the age of 65, with nearly 22% over the age of 75 (Wolff et al., 2018). Given strong associations between pain and age, pain may be a particularly salient issue among older caregivers. Specifically, the prevalence of pain may be higher, and more strongly associated with caregiving outcomes in this age group.
Navigating the burden of pain alongside caregiving responsibilities can make both pain management and caregiving more difficult. Indeed, caregivers’ pain is associated with greater psychological distress and caregiver burden, and pain makes caregiving more challenging often by compromising caregivers’ ability to perform care tasks (Ahn & Logan, 2022; Blyth et al., 2008; Perrin et al., 2014). Pain is associated with both diminished quality of life and mental health in caregivers (Fauziah et al., 2022). Given the high prevalence and consequences of caregivers’ pain, it is important to better understand the intersection between caregiving and pain, including how caregiving may contribute to or possibly exacerbate pain.
Pain Comparisons Between Caregivers to Non-Caregivers
An initial step in understanding how caregiving may be associated with pain would be to compare pain experiences between caregivers and non-caregivers. In the current study, we hypothesized that caregivers would have greater pain prevalence and higher pain intensity than non-caregivers. Our hypothesis is informed by literature suggesting that caregivers often report greater emotional and psychological stress that non-caregivers (Janson et al., 2022), and mental health is strongly associated with greater pain perception (Abdallah & Geha, 2017; Gatchel et al., 2004). Relatedly, theories of stress and coping in caregivers point to the unique stressors of caregiving that may contribute to physical challenges such as pain. Namely, Monin and Schulz (2010) posit that there are interpersonal effects of suffering whereby caregivers who witness a. care recipient’s decline in health status are more likely to experience caregiver distress. We hypothesize that these interpersonal effects of suffering may make caregivers especially prone to pain in ways that are not witnessed in non-caregivers. Lastly, caregiving-related injuries may put caregivers at higher risk of pain compared to non-caregivers. Indeed, injuries and associated musculoskeletal pain are common among nursing staff and other paid caregivers within clinical settings who perform tasks such as lifts and transfers (Davis & Kotowski, 2015).
Studies conducted outside of the United States have compared pain between caregivers and non-caregivers. In a study across 17 European countries, Barbosa et al. (2021) found that spousal caregivers were more likely to report being troubled by pain than non-caregivers (OR = 1.30, 95% CI = 1.13 -- 1.48, p < .001). In a Swedish study, Fagerstrom et al. (2020) found that caregivers and non-caregivers were equally as likely to report pain, but on a 10-point visual analog scale (VAS) measuring pain intensity, non-caregivers reported higher pain intensity pain. However, other than a lone report comparing rates of arthritis between caregivers and non-caregivers (Jackson et al., 2022), this type of robust analysis comparing the pain experience of both caregivers and non-caregivers utilizing large, population-based datasets has not yet been conducted in the United States. Because of cross-cultural differences in caregiving experiences (Afram et al., 2014; Chan et al., 2021) and pain experiences (Zimmer et al., 2022), the existing international studies cannot be confidently generalized to caregivers in the United States. Thus, a study comparing pain between caregivers to non-caregivers in the United States is needed.
Indeed, in the United States, much of the research on caregivers’ pain is conducted among samples of caregivers, rather than comparing caregivers to non-caregivers (Turner, Pillemer et al., 2024). These studies offer insight into the experience of caregiving with pain, and have helped identify which caregivers are at risk for pain. But, because they are unable to compare caregivers to non-caregivers, they are limited in their ability to capture whether and how caregiving itself might be associated with pain experiences. Additionally, because these studies are typically cross-sectional, they are unable to offer evidence on whether caregiving might contribute to the development of new pain or worsening of existing pain over time.
Pain Comparisons Between Different Types of Caregivers
A recent study compared the relative prevalence of pain between different types of caregivers (Turner, Robinson et al., 2024). The study found that certain caregiving characteristics, such as whether the caregiver had physical difficulty providing care, were associated with a greater likelihood of pain. However, as with studies comparing pain experiences between caregivers and non-caregivers, studies comparing the pain experience between different types of caregivers are scarce (Turner, Pillemer et al., 2024) and no studies to our knowledge analyze how different types of caregiving (e.g., more intense caregiving) may contribute to the development of new pain or worsening of existing pain over time. Because caregiving is not a monolithic experience and there is great diversity in caregiving arrangements comparing pain between different types of caregivers is necessary to develop a fuller picture of the intersection between caregiving and the worsening of pain over time, including offering insights into which caregivers are at a greater risk for long term pain implications of caregiving.
The Present Study
In this study, we analyzed longitudinal data to compare pain experiences over a 5-year period between caregivers and non-caregivers in a sample of middle-aged and older adults. As a related objective we sought to determine if various caregiving dynamics (e.g., hours of care per week, dementia versus non-dementia caregiving) were associated with pain presence and changes in pain intensity over time.
Hypothesis 1:
We hypothesized that caregivers would be more likely to report having any pain in the previous four weeks both at baseline and follow-up.
(Hypothesis 1a)
Among participants reporting any pain in the past four weeks, we hypothesized that caregivers would report higher pain intensity at baseline and at follow-up compared to non-caregivers.
Hypothesis 2:
Given that existing research suggests dementia caregivers, older caregivers, and caregivers with greater caregiving intensity have greater physical demands from caregiving and more negative health consequences from caregiving (Bom et al., 2019; Parker et al., 2022; Zehner Ourada & Walker, 2014), we hypothesized that dementia caregivers (versus non-dementia caregivers), primary caregivers (versus non-primary caregivers), spousal caregivers (versus adult-child caregivers), and caregivers with a lot of mental strain (versus some strain or no strain) would have a greater likelihood of reporting pain at baseline and follow-up and have greater pain intensity at baseline and follow-up. We further hypothesized that the more hours of care provided per week would be associated with greater levels of pain intensity among caregivers at both baseline and follow-up.
Method
Study Procedure and Participants
We analyzed data from the United States’ National Social Life, Health, and Aging Project (NSHAP). Launched in 2005, NSHAP includes four rounds of data collected every five years plus a fifth round exclusively devoted to studying the COVID-19 pandemic. NSHAP is nationally representative of community-dwelling older adults in the United States. Participating older adults complete an in-home interview with field interviewers from NORC at the University of Chicago. Topics included in the questionnaire span social networks, physical and mental health and well-being, healthcare utilization and medications, and illness. Further details about NSHAP, including sample recruitment and data collection for each round, are available on NSHAP’s webpage (NORC at the University of Chicago, n.d.) and the National Archive of Computerized Data on Aging (Waite et al., 2023).
In this study, we analyzed data from middle-aged and older adults who participated in both Round 2 (2010–2011) and Round 3 (2015–2016) of NSHAP. Of note NSHAP’s Round 2 survey includes more extensive caregiving questions than the surveys in Round 1 and Round 3. Thus, we selected Round 2 as the baseline for our analyses (rather than Round 1).
Measures
Pain
Participants responded “yes” or “no” to the following question: In the past four weeks, have you had any pain?” If they indicated that they did have pain, they then reported the intensity of that pain with the following question: Please check the box next to the phrase that best describes the level of pain in the past four weeks. Participants responded via a Likert-scale ranging from 0 = no pain to 6 = the most intense pain imaginable. A small number of participants (15 in wave 2 and 4 in wave 3) responded that they did have pain in the past four weeks, but when asked to describe the level of pain, denoted “no pain.” We recategorized these participants as having no pain in the past four weeks and re-coded their response to the pain intensity variable as missing/not applicable.
Caregiver Characteristics
Participants indicated whether they were currently assisting an adult who needed help with day-to-day activities because of age or disability (1 = yes, 0 = no), whether the person required care because of Alzheimer’s Disease or another form of dementia (1 = yes, 0 = no), whether they were the primary caregiver (1 = yes, 0 = no), their relationship to the care recipient (1 = spouse, 2 = adult-child, 3 = other), and how much mental strain they had from caregiving (0 = no strain, 1 = some strain, 2 = a lot of strain).
Caregivers also reported how many days per week they spent caring for the person (numeric response) and how many hours per day they spent caring for the person (less than two hours, two hours or more but less than four hours, four to eight hours, more than eight hours, all of the time). We coded the number of hours per day spent caregiving numerically (1 = less than two hours, 2 = two hours or more but less than four hours, 3 = four to eight hours, 4 = more than eight hours, 5 = all of the time) and multiplied the recoded values by the number of days spent caring per week to create a pseudo-continuous variable representing, as closely as possible with this dataset, the total number of hours per week spent providing caregiving. Six participants responded that they were a caregiver, but when asked to report the number of days per week they provide care, wrote in “0.” We excluded these participants from analysis.
Covariates
We included the following demographic covariates in our analyses: age, gender, race and ethnicity, education status, and working status. NSHAP participants provided their date of birth, which was then used to calculate age. NSHAP operationalized gender as male and female. Participants denoted their race (White/Caucasian, Black/African American, Asian/Pacific Islander/American Indian or Alaskan Native) and ethnicity (Hispanic or Latino versus not Hispanic or Latino). Regarding education, participants reported the highest degree or certification they had earned, which NSHAP recorded into a four-category variable: less than high school, high school or equivalent, vocational certificate/some college/associates, bachelor’s degree or more. For working status, participating indicated “no” or “yes” when asked whether they were currently working.
Analytic Strategy
We conducted all analyses in SAS Studio version 9.04. We used logistic regression models to calculate the odds that caregivers (versus non-caregivers) at baseline would report pain in the previous four weeks at baseline and at follow-up (Hypothesis 1). Among participants who did report experiencing pain in the previous four weeks, we used regression models to estimate and compare mean differences (caregivers versus non-caregivers) in self-reported pain intensity at both baseline and at follow-up (Hypothesis 1a) while accounting for the complex survey design. Specifically, we tested for group differences using least squares means comparisons. For Hypothesis 1a, we utilized regression with contrast statements rather than an Analysis of Variance (ANOVA) model because the procedure to conduct ANOVA models in our statistical software (i.e., SAS PROC GLM) does not have a survey option to account for the complex sampling design of this survey data (Cassell, 2006). In all models, we controlled for participants’ age, race, gender, education, and employment status. In models where follow-up pain was the dependent variable, we also controlled for baseline pain. Across all models, we weighted the data using baseline weight, cluster, and strata variables to account for the NSHAP’s complex survey design.
For our analysis of Hypothesis 1 and 1a, we recognized that participants in our sample who endorsed caregiving at baseline may not have continued caregiving at follow-up and, conversely, non-caregivers at baseline could have become caregivers by follow-up. In these cases, participants’ caregiving status at follow-up, as well as their transitions into or out of caregiving roles, may have resulted in a misclassification bias. To address this issue, we also conducted an additional longitudinal analysis in which we restricted the comparisons for Hypothesis 1 and Hypothesis 1a to only individuals whose caregiving status did not change (that is, they were caregivers or non-caregivers at both waves). This method offered allowed assessment of whether changes in caregiver status between the two waves influenced study results. However, the sample size of caregivers was much smaller in this restricted sample. Because the restricted analysis may have been underpowered to detect differences, these findings should be taken as reinforcement of, but not replacement of, the main results.
We then compared pain intensity at follow-up between various subgroups of caregivers who reported having pain in the previous four weeks (Hypothesis 2) using regression models with estimate and contrast statements as with Hypothesis 1a. In all models, we controlled for participants’ age, race, gender, education, and employment status. In models where follow-up pain was the dependent variable, we also controlled for baseline pain.
Missing Data and Attrition
There were 3,190 participants in Round 2 (i.e., baseline). Our statistical procedure utilized case (i.e., listwise) deletion to handle missing data, thus observations were removed from analysis if there were any missing values in any of the variables included in the model. Eight-hundred-fifty-eight participants were excluded because of missing data, yielding an analytic sample of 2,332 for the main cross-sectional analysis.
Additionally, eight-hundred-forty participants did not participate in Round 3 and were necessarily not a part of the longitudinal analysis. The following demographic characteristics were associated with the likelihood of a participant dropping out between baseline and follow-up: age (higher age more likely to drop out), gender (men more likely to drop out), employment status (participants currently working less likely to drop out), and pain level (higher pain more likely to drop out). Finally, 329 participants either stopped or started caregiving between rounds and were thus excluded from the restricted longitudinal analysis (Figure 1).
Figure 1.
Flow chart depicting attrition and missing data
Results
Sample Characteristics
At baseline, participants had a mean age of 71.91 years and over half (52.19%) were female. The sample was predominantly White/Caucasian (88.48%), with 6.53% Black/African American 4.99% Asian/Pacific Islander/American Indian/Alaskan Native, and non-Hispanic (93.92%) participants. Participants’ education levels were as follows: 13.50% had less than a high school diploma, 24.62% had a high school diploma or equivalent, 33.16% had a vocational certification/some college/associates degree, and 28.72% had a bachelor’s degree or higher. 13.74% (n = 352) were caring for an adult at baseline (mean age of care recipient at baseline = 77.22, range = 19–100; SD = 15.03). 70 reported caregiving at both waves, and 1,093 reported no caregiving at either wave; these 1,163 participants were included in the restricted analysis (Figure 1). Demographic characteristics by caregiver status at baseline are shown in Table 1.
Table 1.
Demographic Statistics
| Caregiver at Baseline (n=352) |
Non-Caregiver at Baseline (n=1,980) |
|
|---|---|---|
|
Sample Size (%) | ||
| Sex | ||
| Male | 153 (45.22%) | 963 (48.22%) |
| Female | 199 (54.78%) | 1017 (51.78%) |
| Race and ethnicity | ||
| White/Caucasian | 266 (84.69%) | 1646 (89.08%) |
| Black/African American | 57 (10.21%) | 208 (5.95%) |
| Asian, Pacific Islander, American Indian, or Alaskan Native | 29 (5.10%) | 126 (4.97%) |
| Ethnicity | ||
| Not Hispanic/Latino | 310 (93.06%) | 1795 (94.06%) |
| Hispanic/Latino | 42 (6.94%) | 185 (5.94%) |
| Educational level | ||
| Less than high school diploma | 68 (15.94%) | 305 (13.11%) |
| High school diploma or equivalent | 83 (21.03%) | 493(25.20%) |
| Vocational certificate/some college/associate’s degree | 118 (37.35%) | 632 (32.49%) |
| Bachelor’s degree or more | 83 (25.69%) | 550 (29.20%) |
| Employment | ||
| Currently working | 278 (75.95%) | 445 (24.71%) |
| Not currently working | 74 (24.05%) | 1535 (75.29%) |
| Pain at baseline | ||
| Reported pain | 255 (71.78%) | 1291 (65.94%) |
| Did not report pain | 97 (28.22%) | 689 (34.06%) |
| Mean (SE) |
||
| Age | 72.11 (0.44) | 71.88 (0.23) |
| Caregiver Subgroups | Sample Size* (%) | |
|
|
||
| Dementia caregiving | ||
| Non-dementia caregiver | 265 (83.09%) | |
| Dementia caregiver | 64 (16.91%) | |
| Primary caregiver | ||
| Non-primary | 159 (49.65%) | |
| Primary | 182 (50.35%) | |
| Relationship to care recipient | ||
| Spouse | 124 (32.55%) | |
| Adult-child | 53 (16.79%) | |
| Other | 167 (50.66%) | |
| Mental strain | ||
| No strain | 140 (42.72%) | |
| Some strain | 162 (48.31%) | |
| A lot of strain | 34 (8.97%) | |
| Mean (SE) | ||
|
|
||
| Hours of care per week | 14.03 (0.91) | |
|
|
||
Sample sizes in caregiver subgroups may not equal total number of caregivers (352) because of missing data.
At baseline, 71.78% of caregivers and 65.94% of non-caregivers reported pain in the previous four weeks. At follow-up 76.47% of caregivers and 70.69% of non-caregivers reported pain in the previous four weeks. Among participants reporting pain in the previous four weeks at baseline, mean pain intensity was 2.97 for caregivers and 2.98 for non-caregivers. Among participants reporting pain in the previous four weeks at both baseline and follow-up, mean pain intensity at follow-up was 3.15 for caregivers and 2.86 for non-caregivers.
Caregivers Compared to Non-Caregivers
Cross-Sectional Analysis
Differences in Pain Presence as a Function of Caregiving Status.
Caregivers were more likely than non-caregivers to report pain in the previous four weeks at baseline (OR = 1.31; CI = 1.01–1.70; Table 2).
Table 2.
Odds of reporting pain in the last four weeks at baseline and follow-up
| Pain in the Past 4 Weeks at Baseline | Pain in the Past 4 Weeks at Follow-Up | |
|---|---|---|
| Odds Ratio (95% CI) | Odds Ratio (95% CI) | |
|
|
||
| Age | 0.98 (0.97 – 1.00) | 0.99 (0.97 – 1.02) |
| Gender | ||
| Male | ref | ref |
| Female | 1.33 (1.06 – 1.67) | 1.16 (0.90 – 1.51) |
| Race | ||
| White/Caucasian | ref | ref |
| Black/African American | 0.91 (0.66 – 1.24) | 0.80 (0.45 – 1.40) |
| Asian, Pacific Islander, American Indian, or Alaskan Native | 0.87 (0.51 – 1.51) | 0.52 (0.21 – 1.30) |
| Ethnicity | ||
| Not Hispanic/Latino | ref | ref |
| Hispanic/Latino | 0.71 (0.49 – 1.03) | 0.45 (0.23 – 0.90) |
| Education | ||
| Less than high school diploma | ref | ref |
| High school diploma or equivalent | 0.81 (0.59 – 1.11) | 0.67 (0.30 – 1.48) |
| Vocational certificate/some college/associate’s degree | 0.95 (0.62 – 1.47) | 0.78 (0.38 – 1.60) |
| Bachelor’s degree or more | 0.96 (0.65 – 1.41) | 0.68 (0.31 – 1.52) |
| Employment Status | ||
| Not currently working | ref | ref |
| Currently working | 0.80 (0.59 – 1.09) | 1.05 (0.77 – 1.44) |
| Baseline Pain | ||
| No pain | -- | ref |
| Pain | -- | 5.07 (3.92 – 6.57) |
| Caregiving Status | ||
| Non-Caregiver | ref | ref |
| Caregiver | 1.31 (1.01 – 1.70) | 1.37 (0.90 – 2.10) |
Note. Bolded text signifies significant difference from the reference group.
Differences in Pain Intensity as a Function of Caregiving Status.
Among participants who reported pain at baseline, there was no difference in the level of pain intensity as a function of caregiver status (estimate = −0.03; p = 0.75; Figure 2).
Figure 2.
Differences in Pain Intensity at Baseline and Follow-Up as a Function of Baseline Caregiving Status
Longitudinal Analysis
Differences in Pain Presence as a Function of Caregiving Status.
Caregivers (as reported at baseline) did not differ from non-caregivers (as reported at baseline) in the likelihood of reporting pain at follow-up (OR = 1.37; CI = 0.90–2.10; Table 2).
Because many of these participants changed caregiving status between the two waves (e.g., changing from caregiver to non-caregiver and vice-versa), we also ran this analysis with a restricted sample that only included participants who remained caregivers (N=70) or non-caregivers (N=1,093) at both waves. Among this restricted sample, caregivers were more likely than non-caregivers to report pain at follow-up (OR = 2.30; CI = 1.05–5.03).
Differences in Pain Intensity as a Function of Caregiving Status.
Among participants who reported pain at both baseline and follow-up, caregivers (as reported at baseline) reported greater follow-up pain intensity compared to non-caregivers (as reported at baseline) (estimate = 0.23; p = 0.02; Figure 2).
Because many of these participants changed caregiving status between the two waves (e.g., changing from caregiver to non-caregiver and vice-versa), we also ran this analysis with a restricted sample that only included participants who remained caregivers or non-caregivers at both waves. Among this restricted sample, we found a similar pattern: caregivers reported higher pain intensity scores at follow-up than non-caregivers (estimate = 0.43; p = 0.04).
Comparisons Between Caregiver Subgroups
Differences in Pain Presence among Pre-Specified Caregiver Subgroups
At both baseline and follow-up, there were no differences in the likelihood of reporting pain in the previous four weeks between dementia caregivers (versus non-dementia), primary (versus non-primary) caregivers, spousal caregivers (versus adult-child caregivers or caregivers related to the care recipient in another way), or caregivers with a lot of mental strain (versus caregivers with no strain or some strain).
Differences in Pain Intensity among Pre-Specified Caregiver Subgroups
Cross-sectional Analyses.
Adult-child caregivers reported less baseline pain intensity compared to spousal caregivers (estimate = −0.37; p < .0001). Compared to caregivers with no mental strain, caregivers with some mental strain (estimate = 0.20; p < .0001) and a lot of mental strain (estimate = 0.29; p < .0001) reported higher baseline pain intensity (Table 4).
Table 4.
Odds of reporting pain in the last four weeks and between-group differences in pain intensity at baseline and follow-up
| Pain in the Past 4 Weeks at Baseline | Pain in the Past 4 Weeks at Follow-Up | |||
|---|---|---|---|---|
| Odds Ratio (95% CI) | Odds Ratio (95% CI) | |||
|
|
||||
| Dementia Caregiving | ||||
| Non-dementia caregiver | ref | ref | ||
| Dementia caregiver | 0.88 (0.40 – 1.90) | 1.04 (0.44 – 2.47) | ||
| Primary Caregiver | ||||
| Non-primary | ref | ref | ||
| Primary | 1.02 (0.53 – 1.95) | 1.04 (0.45 – 2.37) | ||
| Relationship Type | ||||
| Spouse | ref | ref | ||
| Adult-child | 1.18 (0.47 – 2.93) | 0.56 (0.15 – 2.07) | ||
| Other | 1.35 (0.67 – 2.68) | 1.11 (0.45 – 2.69) | ||
| Mental Strain | ||||
| No strain | ref | ref | ||
| Some strain | 1.17 (0.62 – 2.21) | 1.58 (0.71 – 3.53) | ||
| A lot of strain | 2.14 (0.43 – 10.56) | 1.73 (0.49 – 6.13) |
||
| Pain Intensity at Baseline | Pain Intensity at Follow-Up | |||
| Mean (SE) | Comparison Estimate, p value | Mean (SE) | Comparison Estimate, p value | |
|
|
||||
| Dementia Caregiving | ||||
| Non-dementia caregiver | 3.25 (0.09) | ref | 3.43 (0.04) | ref |
| Dementia caregiver | 3.18 (0.05) | −0.08 (0.05), p = 0.14 | 3.08 (0.04) | −0.35 (0.07), p < .0001 |
| Primary Caregiver | ||||
| Non-primary | 3.32 (0.06) | ref | 3.32 (0.05) | ref |
| Primary | 3.33 (0.08) | 0.01 (0.02), p = 0.63 | 3.21 (0.09) | −0.11 (0.11), p = 0.32 |
| Relationship Type | ||||
| Spouse | 3.38 (0.07) | ref | 3.39 (0.15) | ref |
| Adult-child | 3.02 (0.07) | −0.37 (0.05), p < .0001 | 2.94 (0.07) | −0.45 (0.21), p = 0.04 |
| Other | 3.32 (0.09) | −0.06 (0.07), p = .38 | 3.32 (0.58) | −0.07 (0.16), p = 0.65 |
| Mental Strain | ||||
| No strain | 3.21 (0.09) | ref | 3.17 (0.05) | ref |
| Some strain | 3.40 (0.08) | 0.20 (0.03), p < .0001 | 3.32 (0.08) | 0.14 (0.10), p = 0.16 |
| A lot of strain | 3.50 (0.07) | 0.29 (0.04), p < .0001 | 3.23 (0.05) | 0.06 (0.05), p = 0.26 |
| Hours of care per week | 0.01 (0.002), p = .004 | -- | 0.02 (0.004), p = .004 | -- |
Note. Bolded text signifies significant difference from the reference group. Each model ran separately. Covariates for each model are as follows: gender, age, race, ethnicity, education, and employment status. Pain prevalence/intensity at baseline was also included in follow-up models.
Longitudinal Analyses.
Dementia caregivers (as reported at baseline) reported less follow-up pain intensity than non-dementia caregivers (estimate = −0.35; p < .0001). Adult-child caregivers (as reported at baseline) reported less follow-up pain intensity compared to spousal caregivers (as reported at baseline) (estimate = −0.45; p = 0.04; Table 4).
Differences in Pain Intensity as a Function of Number of Hours of Caregiving per Week
Hours of care per week was associated with higher pain intensity both at baseline (estimate = 0.01; p = .004) and at follow-up (estimate = 0.02; p = .004).
Discussion
Family caregivers are relied on as the primary supporters of adults with health and disability challenges, including the increasingly aging population. The evidence that caregiving can negatively impact caregivers’ physical and mental health is strong (Bom et al., 2019). This study adds to this literature by identifying pain as a possible health consequence of caregiving. Indeed, results suggest that family caregiving contributes to the worsening of pain over time for caregivers with existing pain. Further, among caregivers, certain caregiving characteristics (relationship to care recipient, amount of care-related mental strain, hours of care provided each week) are associated with greater pain intensity.
Key Findings
Pain Prevalence
In cross-sectional analyses, caregivers were slightly more likely to report pain in the previous four weeks compared to non-caregivers. This result offers evidence that being a caregiver is associated with a greater risk of reporting pain. In longitudinal analyses, caregivers were more likely than non-caregivers to report pain at follow-up in the restricted sample but not in the main sample. Taken together, these contrasting longitudinal results from the main and restricted samples suggest that duration of caregiving impacts the likelihood of reporting pain. Indeed, given it only included participants who were either a caregiver at both waves or a non-caregiver at both waves, the restricted sample analysis offers a way to capture duration of caregiving. That the restricted sample resulted in caregivers being more likely to report pain, but that main sample (which included participants who moved in and out of caregiving statuses) suggests that the longer a person provides care, the more likely they appear to be to report pain.
Notably, in this specific sample, the prevalence of pain was high among both caregivers and non-caregivers. For example, though the Barbosa et al. (2021) study comparing caregivers to non-caregivers across Europe utilized a similarly-aged sample of participants aged 65 and older, only 50% of non-caregivers reported having pain (compared to 66% in our sample). It is possible that the high prevalence of pain in our non-exposed group (i.e., non-caregivers) influenced the magnitude of between-group differences, which would have been even larger in another sample. limited our ability to detect between-group differences in pain (Hsieh, 1989).
Gender as a Significant Predictor of Baseline Pain Prevalence
Gender was the only significant covariate in the model of caregiving at baseline predicting baseline pain, with women being more likely to report pain (Table 3). Nationally, women are disproportionally impacted by pain, and are typically responsible for family caregiving. As such, one way in which caregiving might create gender-based health disparities is through pain. Post-hoc analysis of this sample revealed that associations between caregiving and pain were not dependent on (i.e., moderated by) gender. Nonetheless, given how highly gendered both pain and caregiving are, we encourage future research on the intersection between caregiving, pain, and women’s health.
Table 3.
Between-group differences in pain intensity at baseline and follow-up
| Pain Intensity at Baseline | Pain Intensity at Follow-Up | |||
|---|---|---|---|---|
| Mean (SE) | Comparison Estimate, p value | Mean (SE) | Comparison Estimate, p value | |
|
|
||||
| Age | −0.01 (0.005) | p = .30 | −0.003 (0.004) | p = .51 |
| Gender | ||||
| Male | 3.06 (0.35) | ref | 3.15 (0.12) | ref |
| Female | 3.09 (0.33) | 0.03 (0.05), p = .54 | 3.11 (0.11) | −0.04 (0.06), p = .52 |
| Race | ||||
| White/Caucasian | 3.05 (0.07) | ref | 3.20 (0.16) | ref |
| Black/African American | 2.93 (0.13) | −0.12 (0.11), p = .27 | 3.19 (0.16) | −0.01 (0.10), p = .90 |
| Asian, Pacific Islander, American Indian, or Alaskan Native | 3.25 (0.14) | 0.20 (0.15), p = .42 | 3.01 (0.13) | −0.19 (0.20), p = .34 |
| Ethnicity | ||||
| Not Hispanic/Latino | 2.99 (0.07) | ref | 3.01 (0.08) | ref |
| Hispanic/Latino | 3.16 (0.13) | 0.17 (0.14), p = 0.21 | 3.26 (0.25) | 0.25 (0.28), p = 0.38 |
| Education | ||||
| Less than high school diploma | 3.33 (0.11) | ref | 3.26 (0.14) | ref |
| High school diploma or equivalent | 3.08 (0.10) | −0.25 (0.11), p = .03 | 3.17 (0.14) | −0.09 (0.14), p = .55 |
| Vocational certificate/some college/associate’s degree | 3.04 (0.09) | −0.29 (0.12), p = .02 | 3.12 (0.13) | −0.14 (0.13), p = .28 |
| Bachelor’s degree or more | 2.86 (0.08) | −0.47 (0.10), p < .001 | 2.98 (0.14) | −0.28 (0.14), p = .05 |
| Employment Status | ||||
| Not currently working | 3.14 (0.07) | ref | 3.14 (0.11) | ref |
| Currently working | 3.01 (0.10) | −0.13 (0.08), p = .09 | 3.13 (0.13) | 0.01 (0.08), p = .87 |
| Baseline Pain | -- | -- | 0.29 (0.03) | p < .0001 |
| Caregiving Status | ||||
| Non-Caregiver | 3.09 (0.07) | ref | 3.02 (0.11) | ref |
| Caregiver | 3.06 (0.10) | −0.03 (0.09), p = .75 | 3.25 (0.31) | 0.23 (0.10), p = .02 |
Note. Bolded text signifies significant difference from the reference group. Age and baseline pain intensity treated as continuous variables.
Pain Intensity
Regarding pain intensity among participants who reported having pain in the previous four weeks, results were both expected and unexpected. We found that caregivers and non-caregivers reported similar pain intensity at baseline, which contrasted our hypothesis. However, in line with our hypothesis, we found that caregivers reported greater pain intensity at follow-up than non-caregivers even after controlling for baseline pain intensity. Our analysis with the restricted sample of participants who were either caregivers or non-caregivers at both waves yielded similar results, reinforcing our results observed in the main longitudinal sample. As such, our results suggest that in addition to being a risk factor for pain in general, caregiving is also a risk factor for the worsening pain over time.
Further, among caregivers reporting pain, the finding that greater caregiving intensity was associated with higher levels of pain intensity offers additional evidence that caregiving itself may intensify pain. Indeed, such a finding is indicative of a dose-response effect, whereby more intense caregiving (higher dose) is associated with more intense pain (higher response). Paired with the finding that caregivers had greater pain intensity at follow-up than non-caregivers, we conclude that caregiving – especially high-intensity caregiving – is likely a risk factor for worsening pain. This finding expands existing evidence regarding how caregiving may negatively impact caregivers’ health. In so doing, it reinforces caregivers’ pain as a necessary public health concern, with efforts to mitigate the effects of caregiving on caregivers’ pain intensity urgently needed especially as the number of family caregivers grows as the population ages.
Finally, there were several interesting findings related to differences in pain intensity between caregiver subgroups. Namely, the finding that dementia caregivers (versus non-dementia caregivers had less pain intensity is interesting and contrasts with our hypothesis. Of note, the number of dementia caregivers in this sample was small, and thus these analyses may have been underpowered. Yet, it offers some evidence of a counter-intuitive finding regarding caregiving and pain. It is possible, for example, that those who are healthier are more likely to become caregivers and that the high demands of caring for a person with dementia and/or a very old care recipient help to keep the caregiver healthy, a nod to the healthy caregiver effect (Bertrand et al., 2011; Leggett et al., 2019). To offer some context to this possibility, we compared self-rated health between dementia caregivers and non-dementia caregivers. We found that dementia caregivers had better self-rated health than non-dementia caregivers (estimate = 0.50, p < .001), reinforcing the possibility of a healthy caregiver effect for dementia caregivers.
It is also possible that dementia caregivers engage in fewer or less demanding physical tasks compared to other caregivers who may be more frequently support care recipients’ with tasks such as mobility. Indeed, the extent to which caregivers report physical difficulty with care is more dependent on the care recipients’ level of disability than whether they have dementia (Wolff et al., 2016). Also, as care recipients get older and/or if they have dementia, other resources might be available to caregivers which helps to mitigate pain. For example, additional caregivers, including paid caregivers, may be available to assist with care tasks, though we were unable to analyze this possibility in the current study. Ultimately, differences in pain experiences between dementia and non-dementia caregivers warrant future investigations.
Limitations
Several limitations of this study warrant discussion. First as with much of the literature on caregiving, much of this sample was White. Given existing racial and ethnic disparities in pain, we encourage research specifically focused on pain experiences among racial and ethnic minority caregivers. Previous studies suggest Black caregivers are less likely to report pain than White caregivers (Turner et al., 2024), which is counter-intuitive to national research on pain within Black communities and reflects a need for strong research to better understand pain experiences among historically marginalized caregivers.
Additional limitations pertain to the longitudinal analyses. First, the NSHAP dataset does not include the same robust caregiver questions in Round 3 (i.e., follow-up) as it does in Round 2 (i.e., baseline). It is possible that dynamics of the caregiving situation changed between waves (e.g., hours of care may have increased; care recipient might have developed dementia), which could have impacted results in ways that were not considered in our models. Relatedly, we did not capture how long participants served as caregivers prior to the baseline assessment. Participants who had been providing care for a longer time may have had time to adjust and cope with caregiving such that caregiving had less of an impact on their health (i.e., pain). Finally, although we conducted an analysis with a restricted sample of participants who either remained caregivers or remained non-caregivers between the two waves, we did not capture caregiving status at follow-up in the primary analysis. Ultimately, a future more robust consideration of how transitions into and out of caregiving impact pain is necessary to fully capture associations between caregiving and pain.
Conclusion
In this population-based study, we compared the presence of pain and the extent of pain intensity between caregivers and non-caregivers for the first time in the United States. In so doing, we offer evidence that caregiving may contribute to the worsening of existing pain over time, especially for caregivers providing more hours of care each week. Further research on caregivers’ pain is needed, and the avenues to conduct such work are numerous. For example, researchers should identify the incidence of caregivers’ new pain conditions after the onset of caregiving, consider how caregivers’ pain impacts their ability to complete care tasks and meet care recipient needs, analyze how caregivers’ cope with their pain, and given population aging is a global phenomenon, should identify the scope of the problem of caregiver pain and its consequences for caregiving internationally.
Funding:
This work was supported by the National Institutes of Health/National Institute on Aging [T32 AG049666 to SGT, 5K24AG053462–07 to MCR, and P30AG022845 to MCR and KAP].
Footnotes
Conflicting Interests: We have no conflicts of interest to disclose.
Ethical approval and informed consent: Our analysis of this data did not require ethical approval from our university because the dataset we used is publicly available and de-identified.
Data availability:
Researchers can apply for access to the data used in this study via the following webpage: https://www.norc.org/research/projects/national-social-life-health-and-aging-project.html. Statistical code for this study is made available by email request to the corresponding author. This study was not pre-registered.
References
- AARP. (2020). Caregiving in the U.S. https://www.aarp.org/content/dam/aarp/ppi/2020/05/full-report-caregiving-in-the-united-states.doi.10.26419-2Fppi.00103.001.pdf
- Abdallah CG, & Geha P (2017). Chronic pain and chronic stress: Two sides of the same coin?. Chronic Stress, 1, 10.1177/2470547017704763 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Afram B, Stephan A, Verbeek H, Bleijlevens MHC, Suhonen R, Sutcliffe C, Raamat K, Cabrera E, Soto ME, Hallberg IR, Meyer G, & Hamers JPH (2014). Reasons for institutionalization of people with dementia: Informal caregiver reports from 8 European countries. Journal of the American Medical Directors Association, 15(2), 108–116. 10.1016/j.jamda.2013.09.012 [DOI] [PubMed] [Google Scholar]
- Ahn S, & Logan JG (2022). Perceived role overload and physical symptom experience among caregivers of older adults: The moderating effect of social support. Geriatric Nursing, 43, 197–205. 10.1016/j.gerinurse.2021.11.021 [DOI] [PubMed] [Google Scholar]
- Barbosa F, Delerue Matos A, Voss G, & Costa P (2021). Spousal care and pain among the population aged 65 years and older: A European analysis. Frontiers in Medicine, 8, 602276. 10.3389/fmed.2021.602276 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barbour KE, Helmick CG, Boring M, & Brady TJ (2017). Vital signs: prevalence of doctor-diagnosed arthritis and arthritis-attributable activity limitation—United States, 2013–2015. Morbidity and Mortality Weekly Report, 66(9), 246–253. 10.15585/mmwr.mm6609e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Beach SR, & Schulz R (2017). Family caregiver factors associated with unmet needs for care of older adults. Journal of the American Geriatrics Society, 65(3), 560–566. 10.1111/jgs.14547 [DOI] [PubMed] [Google Scholar]
- Bertrand RM, Saczynski JS, Mezzacappa C, Hulse M, Ensrud K, & Fredman L (2012). Caregiving and cognitive function in older women: Evidence for the healthy caregiver hypothesis. Journal of Aging and Health, 24(1), 48–66. 10.1177/0898264311421367 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bom J, Bakx P, Schut F, & van Doorslaer E (2019). The impact of informal caregiving for older adults on the health of various types of caregivers: A systematic review. The Gerontologist, 59(5), e629–e642. 10.1093/geront/gny137 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chan CY, Cheung G, Martinez-Ruiz A, Chau PY, Wang K, Yeoh E, & Wong EL (2021). Caregiving burnout of community-dwelling people with dementia in Hong Kong and New Zealand: A cross-sectional study. BMC Geriatrics, 21, 1–15. 10.1186/s12877-021-02153-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Davis KG, & Kotowski SE (2015). Prevalence of musculoskeletal disorders for nurses in hospitals, long-term care facilities, and home health care: a comprehensive review. Human factors, 57(5), 754–792. 10.1177/0018720815581933 [DOI] [PubMed] [Google Scholar]
- Fagerstrom C, Elmstahl S, & Wranker LS (2020). Analyzing the situation of older family caregivers with a focus on health-related quality of life and pain: a cross-sectional cohort study. Health and Quality of Life Outcomes, 18(79), 1–10. 10.1186/s12955-020-01321-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fauziah W, Kato M, Shogenji M, Tsujiguchi H, & Taniguchi Y (2022). Factors associated with depression among family caregivers of patients with stroke in Indonesia: A cross-sectional study. The Journal of Nursing Research, 30(5), e231. 10.1097/jnr.0000000000000515 [DOI] [PubMed] [Google Scholar]
- Freedman VA, & Wolff JL (2024). Dementia and caregiving trends and the emerging policy environment. Presented at the 2024 NHATS/NSOC Research Conference on Caregiving to Older Adults with Dementia June 20–21, 2024, Ann Arbor, Michigan. [Google Scholar]
- Gatchel RJ (2004). Comorbidity of chronic pain and mental health disorders: the biopsychosocial perspective. American Psychologist, 59(8), 795–805. 10.1037/0003-066X.59.8.795 [DOI] [PubMed] [Google Scholar]
- Haley WE, & Elayoubi J (2024). Family caregiving as a global and lifespan public health issue. The Lancet Public Health, 9(1), e2–e3. 10.1016/S2468-2667(23)00227-X [DOI] [PubMed] [Google Scholar]
- Jackson EM, Omura JD, Boring MA, Odom EL, Foster AL, Olivari BS, McGuire LC, & Croft JB (2022). Prevalence and characteristics of arthritis among caregivers—17 States, 2017 and 2019. Morbidity and Mortality Weekly Report, 71(44), 1389–1395. 10.15585/mmwr.mm7144a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Janson P, Willeke K, Zaibert L, Budnick A, Berghöfer A, Kittel-Schneider S, ... & Keil T. (2022). Mortality, morbidity and health-related outcomes in informal caregivers compared to non-caregivers: A systematic review. International Journal of Environmental Research and Public Health, 19(10), 5864. 10.3390/ijerph1910586 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kuzuya M, Enoki H, Hasegawa J, Izawa S, Hirakawa Y, Shimokata H, & Akihisa I (2011). Impact of caregiver burden on adverse health outcomes in community-dwelling dependent older care recipients. The American Journal of Geriatric Psychiatry, 19(4), 382–391. 10.1097/JGP.0b013e3181e9b98d [DOI] [PubMed] [Google Scholar]
- Leggett AN, Sonnega AJ, & Lohman MC (2020). Till death do us part: Intersecting health and spousal dementia caregiving on caregiver mortality. Journal of Aging and Health, 32(7–8), 871–879. 10.1177/0898264319860975 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Monin JK, & Schulz R (2010). The effects of suffering in chronically ill older adults on the health and well-being of family members involved in their care. GeroPsych, 23 (4), 207–213. 10.1024/1662-9647/a000024 [DOI] [PMC free article] [PubMed] [Google Scholar]
- National Alliance for Caregiving (2022). Chronic disease family caregiving thorugh a public health lens: The framework for family caregiving and public health. https://www.caregiving.org/wp-content/uploads/2022/09/NAC_Chronic-Disease-Caregiving-Report_090722.pdf
- NORC at the University of Chicago. (n.d.). National Social Life, Health, and Aging Project. https://www.norc.org/research/projects/national-social-life-health-and-aging-project.html
- Parker LJ, Fabius C, Rivers E, & Taylor JL (2022). Is dementia-specific caregiving compared with non-dementia caregiving associated with physical difficulty among caregivers for community-dwelling adults? Journal of Applied Gerontology, 41(4), 1074–1080. 10.1177/07334648211014352 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perrin PB, Morgan M, Aretouli E, Sutter M, Snipes DJ, Hoyos GR, Buraye JA, & Arango-Lasprilla JC (2014). Connecting health-related quality of life and mental health in dementia caregivers from Colombia, South America. Journal of Alzheimer’s Disease, 39(3), 499–509. 10.3233/JAD-130764 [DOI] [PubMed] [Google Scholar]
- Pristavec T, & Luth EA (2020). Informal caregiver burden, benefits, and older adult mortality: A survival analysis. The Journals of Gerontology, Series B: Psychological Sciences and Social Science, 75(10), 2193–2206. 10.1093/geronb/gbaa001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schulz R, Beach SR, & Friedman EM (2021). Caregiving factors as predictors of care recipient mortality. The American Journal of Geriatric Psychiatry, 29(3), 295–303. 10.1016/j.jagp.2020.06.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sörensen S, & Conwell Y (2011). Issues in dementia caregiving: Effects on mental and physical health, intervention strategies, and research needs. American Journal of Geriatric Psychiatry, 19(6), 491–496. 10.1097/JGP.0b013e31821c0e6e [DOI] [PMC free article] [PubMed] [Google Scholar]
- Talley RC, & Crews JE (2007). Framing the public health of caregiving. American Journal of Public Health, 97(2), 224–228. 10.2105/ajph.2004.059337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Turner SG, Pillemer K, Demetres M, Heaney K, Joshi S, Luebke M, Messay G, & Reid MC (2024). Physical pain among family caregivers to older adults: A scoping review of the literature. Journal of the American Geriatrics Society. [online ahead of print] 10.1111/jgs.19037 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Turner SG, Robinson JRM, Pillemer KA, & Reid MC (2024). Prevalence estimates of arthritis and activity-limiting pain among family caregivers to older adults. The Gerontologist, 64(5), gnad124. 10.1093/geront/gnad124 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Waite LJ, Cagney KA, Dale W, Huang ES, Laumann EO, McClintock MK, O’Muircheartaigh CA, Schumm LP, & Cornwell B (2023). National Social Life, Health, and Aging Project (NSHAP): Round 2 and Partner Data Collection, [United States], 2010–2011 Inter-university Consortium for Political and Social Research [distributor]. 10.3886/ICPSR34921.v5 [DOI] [Google Scholar]
- Wolff JL, Mulcahy J, Huang J, Roth DL, Covinsky K, & Kasper JD (2018). Family caregivers of older adults, 1999–2015: Trends in characteristics, circumstances, and role-related appraisal. The Gerontologist, 58(6), 1021–1032. 10.1093/geront/gnx093 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zehner Ourada VE, & Walker AJ (2014). A Comparison of physical health outcomes for caregiving parents and caregiving adult children. Family Relations, 63(1), 163–177. 10.1111/fare.12046 [DOI] [Google Scholar]
- Zimmer Z, Fraser K, Grol-Prokopczyk H, & Zajacova A (2022). A global study of pain prevalence across 52 countries: examining the role of country-level contextual factors. Pain, 163(9), 1740–1750. 10.1097/j.pain.0000000000002557 [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.
Data Availability Statement
Researchers can apply for access to the data used in this study via the following webpage: https://www.norc.org/research/projects/national-social-life-health-and-aging-project.html. Statistical code for this study is made available by email request to the corresponding author. This study was not pre-registered.