Perceived Patient Pain and Spousal Caregivers’ Negative Affect: The Moderating Role of Spouse Confidence in Patients’ Pain Management

Abstract

Objectives:

This study examined whether older patients’ greater daily pain perceived by their spouses was associated with spouses’ higher daily negative affect. We further investigated whether spouses’ lower confidence in patients’ ability to manage pain exacerbated the daily association between perceived patient pain and spouses’ negative affect.

Method:

We used baseline interviews and a 22-day diary of knee osteoarthritis patients and their spouses (N = 144 couples). Multilevel models were estimated to test hypotheses.

Results:

Daily perceived patient pain was not associated with spouses’ daily negative affect. However, spouse confidence significantly moderated the association. Only spouses with lower confidence in patients’ pain management experienced higher negative affect on days when they perceived that patients’ level of pain was higher than usual.

Discussion:

Findings suggest that spousal caregivers’ lack of confidence in patients’ pain management may be a risk factor for spouses’ affective distress in daily life.

Spousal caregivers of patients with chronic pain are exposed to their loved one’s suffering in everyday life. According to the caregiver stress-health model proposed by Monin and Schulz (2009), witnessing a loved one in pain may lead to affective distress in the caregiver. That is, observing a loved one’s suffering may influence caregivers’ emotions because close relationship partners empathize with each other. The current study tests this theoretical proposition by examining whether perceived patient pain is associated with spousal caregivers’ negative affect.

It is important to examine the extent to which perceived patient pain is associated with spousal caregivers’ negative affect, because increasing levels of negative affect are associated with caregivers’ poorer health and quality of life (Chung et al., 2009; Mausbach et al., 2007). Empirical studies on the association between perceived patient pain and caregivers’ psychological distress have produced mixed findings. Although some studies suggested that being exposed to a loved one’s pain is distressing for caregivers (Monin et al., 2017; Redinbaugh et al., 2002), Geisser and colleagues (2005) did not found such an association. Most of these previous studies on perceived patient pain and spouses’ distress used data from a single time point (Geisser et al., 2005; Redinbaugh et al., 2002). An exception is a 7-day daily diary study among older adults with a painful musculoskeletal condition, which found that spousal caregivers of these patients reported higher distress on days when they perceived their partner’s suffering to be greater than usual (Monin et al., 2017).

Given that affect is a short-term state that varies from day to day (Ekman & Davidson, 1994; Watson et al., 1988), the current study examines the relation between daily perceived patient pain and spousal caregivers’ daily negative affect, using daily diary data. This study focuses on patient pain perceived by their spouses rather than patients’ ratings of their pain, given that observers are most likely to respond affectively to pain that they notice (Goubert et al., 2005). Based on the caregiver stress-health model (Monin & Schulz, 2009), we hypothesize that spousal caregivers of chronic pain patients experience higher negative affect on days when they perceive that patients’ pain is greater than usual (Hypothesis 1).

Prior studies have paid less attention to conditions under which perceived patient pain has the strongest influence on spousal caregivers’ affective distress. However, it is important to investigate these conditions, since it allows us to identify spousal caregivers of chronic pain patients who are most vulnerable to affective distress and thus may benefit from a psychosocial intervention. One potentially important contextual factor is spouses’ confidence in patients’ ability to manage their pain. Social cognitive theory posits that a person’s self-efficacy, which is the belief about his or her capabilities, may modulate how much negative affect is experienced in response to stressful situations (Bandura, 1997). Extending social cognitive theory, Lent and Lopez (2002) suggested that partners in close relationships develop not only self-efficacy related to their own capabilities but also other-efficacy, which is an individual’s beliefs about his or her partner’s ability to perform particular behaviors. This other-efficacy, which may or may not correspond to the other’s self-efficacy or actual abilities, stems from perceptions of the other’s accomplishment in past situations or certain aspects of the perceiver (Lent & Lopez, 2002).

In the context of chronic illness management, the associations between spousal caregivers’ beliefs about patients’ control over their illness and caregivers’ psychological outcomes have been investigated (Cano et al., 2009; Karademas & Giannousi, 2013; Sterba & DeVellis, 2009). Some studies reported that spouses’ perceptions about patients’ control of their cancer or chronic pain are not linked to their depressive symptoms or anxiety (Cano et al., 2009; Karademas & Giannousi, 2013). In contrast, one study found that husbands’ lack of confidence in wives’ control over their rheumatoid arthritis (RA) predicted husbands’ worse psychological adjustment at a 4-month follow-up (Sterba & DeVellis, 2009). Yet, little is known about whether spousal caregivers’ beliefs about patients’ ability to manage their pain moderates the association between perceived patient pain and caregivers’ negative affect.

We expect that spousal caregivers’ lower confidence in patients’ ability to manage their pain strengthens the association between daily perceived pain and spousal caregivers’ daily negative affect (Hypothesis 2). According to social cognitive theory, it is not stressful life conditions per se but the perceived inability to manage stress that is distressing (Bandura, 1997). Spousal caregivers with lower confidence in their partners’ ability to manage their pain may be more worried about their partners’ future health and feel increasing hopelessness when perceiving partners’ greater pain than those with higher confidence. Indeed, spouses who catastrophized more about their partners’ chronic pain had significantly greater depressive symptoms than those with lower levels of catastrophizing (Cano et al., 2005).

To summarize, the current study tests the following hypotheses, using a 22-day daily diary study of spousal caregivers of patients coping with knee osteoarthritis (OA).

Hypothesis 1: Spousal caregivers of patients with chronic pain will experience more negative affect on days when they perceive that patients’ pain is greater than usual.

Hypothesis 2: Daily perceived patient pain will be more strongly associated with spousal caregivers’ daily negative affect for caregivers with lower confidence in patients’ ability to manage pain than for those with higher confidence.

OA is a degenerative disease prevalent in older adults in which joint cartilage breaks down over time, resulting in persistent joint pain, varying degrees of inflammation, and joint stiffness (Keefe et al., 2002). Considering that persistent pain is the primary symptom of OA and is the major reason why individuals seek medical attention (Creamer, 2000), knee OA provides a good model in which to study whether patients’ daily pain perceived by their spouses relates to spouses’ higher daily negative affect.

Method

Study Design

The current study used data from a larger study of patients diagnosed with knee OA and their spouses. The parent study combined in-person interviews conducted over an 18-month period (i.e., Time 1, Time 2 at a 6-month follow-up, and Time 3 at an 18-month follow-up) with a 22-day assessment of daily experiences immediately after the Time 1 interview. During the daily assessment protocol, patients and their spousal caregivers answered questions about their health and affect 3 times per day (i.e., morning, afternoon, and end-of-day). This study used data from Time 1 interviews and 22-day daily assessments.

Participants

To be eligible for the study, patients had to be diagnosed with knee OA by a physician, rate their usual knee pain as moderate or severe, be at least 50 years of age, and be married or in a long-term relationship in which they shared a residence with their partner (Martire et al., 2013). Exclusion criteria were a comorbid diagnosis of fibromyalgia or RA, use of a wheelchair to get around, and a plan to have hip or knee surgery within the following 6 months. Couples were excluded from the study if the spouse also had usual knee pain of moderate or greater intensity, used a wheelchair to get around, or required assistance with personal care activities. Both partners had to be cognitively functional as indicated by the accuracy of their answers to questions regarding the current date, day of the week, their age, and birth date. Both partners also had to be free of any major hearing, speech, or language problems that would interfere with the comprehension and completion of data collection conducted in English.

Primary sources of recruitment were research registries for rheumatology clinic patients and older adults interested in research; flyers distributed to University of Pittsburgh staff and faculty, and word of mouth. A total of 606 couples were screened for eligibility. Of these, 221 couples declined to participate, and the most frequent reasons were lack of interest (N = 87) or illness in the family (N = 55). A total of 233 couples were not eligible, and the most frequent reasons were lack of OA in the knee (N = 55) or knee OA pain that was mild (N = 47). The total enrolled sample comprised 152 couples (i.e., 304 individuals) which included three same-sex couples. A total of 145 couples completed the diary assessment component of the study, and 144 of these couples provided sufficient data for our primary analyses. Given that spouses in this study provided daily instrumental (M = 1.47, SD = 0.49, range = 1–3) and emotional support (M = 1.74, SD = 0.57, range = 1–3) to patients, we referred to spouses as spousal caregivers.

Data Collection Procedures

All procedures were approved by the University of Pittsburgh Institutional Review Board (IRB # 07030113) and written informed consent was obtained prior to baseline data collection. Trained staff interviewed patients and spouses independently in their home. Following these interviews, couples were trained in use of the handheld computer (i.e., the Palm TX) as well as the format and content of the diary questions. Participants were trained to complete their diary assessments independently of their spouse, using their own handheld computer. The goal was to capture the participants’ experiences within the general time frames of morning, afternoon, and end-of-day. Therefore, participants were instructed to answer questions: (a) within 60 minutes of rising in the morning (i.e., morning), (b) between 2:00 and 4:00 p.m. (i.e., afternoon), and (c) upon retiring at night (i.e., end-of-day). The current study utilized daily morning, afternoon, and end-of-day assessments.

Completion and compliance rates were examined for the diary data. Of 3,168 potential morning, afternoon, or end-of-day assessments (144 Participants × 22 Days), spousal caregivers completed a total of 2,911 morning assessments (92%), 2,933 afternoon assessments (92%), and 2,914 end-of-day assessments (93%).

Measures

Daily perceived patient pain.

Each evening, spousal caregivers rated patients’ pain severity with one item using a 0 (none) to 3 (severe) scale: “How would you describe your spouse’s overall arthritis pain today?” The average level of daily perceived patient pain was mild to moderate (M = 1.34, SD = 0.82, Skewness = 0.06).

Daily negative affect.

Spousal caregivers reported their negative affect over the past 30 minutes at the beginning of day, afternoon, and end of day. This measure is the average of five items (depressed or blue, frustrated, angry or hostile, unhappy, and worried or anxious) that were rated from 0 (not at all) to 6 (extremely) (Thomas & Diener, 1990). The three assessments within a day were averaged to create a daily negative affect score, with higher scores indicating greater daily negative affect (between-person Cronbach’s α > .80; M = 0.47, SD = 0.75, Skewness = 2.56).

Spouse confidence in patients’ pain management.

During in-person interviews, spousal caregivers responded to five items from the Arthritis Self-Efficacy Scale (Lorig et al., 1989), which were developed to measure patients’ confidence in managing pain, but were modified in this study to assess spousal caregivers’ confidence in patients’ ability to manage their pain (e.g., “How confident are you that s/he can decrease his or her pain quite a bit?”). Items were rated from 1 (not at all confident) to 10 (totally confident) and summed for analysis (Cronbach’s α = .75; M = 31.94, SD = 8.84, Skewness = −0.24).

Covariates

Chosen on a conceptual basis, covariates in this study are spousal caregivers’ sex, age, and baseline depressive symptoms, as well as patients’ daily ratings of their own pain and patients’ daily negative affect. We controlled for spousal caregivers’ sex to account for potential gender differences in daily negative affect (Almeida & Kessler, 1998). We also controlled for spousal caregivers’ age, as older people tend to report lower daily negative affect (Sliwinski et al., 2009). We included spousal caregivers’ baseline depressive symptoms as a covariate, because depressed individuals report greater daily negative affect than do healthy individuals (Bylsma et al., 2011). Spousal caregivers’ depressive symptoms were assessed using the Center for Epidemiologic Studies Depression Scale-10 (CESD-10; Andresen et al., 1994). Total summed CESD scores ranged from 0 to 30, with higher scores indicating greater depressive symptoms (Cronbach’s α =.77; M = 5.92, SD = 4.50). We controlled for patients’ daily negative affect, the average of the three assessments within a day (between-Person Cronbach’s α > .80; M = 0.52, SD = 0.77), due to potential covariance between patients’ and their spousal caregivers’ affect (Saxbe & Repetti, 2010). Finally, we included patients’ daily ratings of their own pain (M = 1.51, SD = 0.66) to adjust for covariation between patients’ and caregivers’ ratings of patients’ pain (Redinbaugh et al., 2002).

Data Analysis

We tested our hypotheses in two multilevel models (Snijders & Bosker, 2012). To test Hypothesis 1, we first estimated a main effect model to examine the within-person association between daily perceived patient pain and spousal caregivers’ daily negative affect. To test Hypothesis 2, we added a cross-level interaction term between daily perceived patient pain and spouse confidence in patients’ pain management to the main effect model. Both models included all covariates and were estimated using SAS 9.4 PROC MIXED (SAS Institute Inc, 2013). Restricted maximum likelihood (REML) was used to handle missing data. All Level 1 predictors and covariates were centered relative to each person’s mean score to remove between-person variance in these scores (i.e., group-mean centered), and all continuous Level 2 covariates were grand-mean centered. The interaction model for testing Hypothesis 2 was specified as

$${\text{Spouse’s Negative Affect}}_{it}={\gamma }_{00}+{\gamma }_{10}{\left(\text{Perceived Patient Pain}\right)}_{it}+{\gamma }_{20}{\left(\text{Patient Pain}\right)}_{it}+{\gamma }_{30}{\left(\text{Patient Negative Affect}\right)}_{it}+{\gamma }_{01}{\left(\text{Spouse Confidence}\right)}_i+{\gamma }_{02}{\left(\text{Spouse’s Sex}\right)}_i+{\gamma }_{03}{\left(\text{Spouse's Age}\right)}_i+{\gamma }_{04}{\left(\text{Spouse's Depressive Symptoms}\right)}_i+{\gamma }_{11}{\left(\text{Perceived Patient Pain}\right)}_{it}\times {\left(\text{Spouse Confidence}\right)}_i+{\mu }_{1i}{\left(\text{Perceived Patient Pain}\right)}_{it}+{\mu }_{0i}+e_{it}$$

where subscript i indicates a person i, and subscript t indicates time t. Therefore, parameters with subscript i denote time-invariant person-level predictors, whereas parameters with subscript it denote time-varying predictors. γ₁₀ to γ₄₀ represent the influences of daily perceived patient pain reported by spouses, daily patients’ ratings of their own pain, and patients’ daily negative affect on spouses’ daily negative affect, respectively. γ₀₁ to γ₀₄ represent the influences of spousal caregivers’ confidence in patients’ pain management, sex, age, and depressive symptoms on spousal caregivers’ daily negative affect, respectively. γ₁₁ represents the influence of spouse confidence on the association between daily perceived patient pain and spouses’ daily negative affect, which is our particular interest for Hypothesis 2. μ_0i is the intercept variance that is uncorrelated with the residual variance e_it. μ_1i is the variance of the regression slopes for daily perceived patient pain on spouses’ daily negative affect.

To estimate the local effect size of the interaction term (γ₁₁), we computed Cohen’s f² which is an appropriate statistic for comparing the strength of a variable’s influence while controlling for other variables (Selya et al., 2012). Cohen’s f² was interpreted by convention in terms of small (.02), medium (.15), or large (.35) effects (Cohen, 1988).

Results

Demographic and medical characteristics of the sample are displayed in Table 1. Spousal caregivers were 65.31 years of age on average (SD = 11.46), primarily White (85%), and the majority were male (57%). Spousal caregivers had been educated for 15.88 years (SD = 2.03) on average, and 46% of them were employed. The couples have been married for 34.24 years (SD = 16.61) on average, and most of them reported their annual household income greater than US$40,000. Patients have been suffering from knee OA for 12.79 years (SD = 11.34) on average. Cortisone injections into joints (62%) and physical therapy (61%) were frequently used by the patients to manage their pain. About 22% of patients were engaged in strength or endurance exercise.

Table 1.

Demographic and Medical Characteristics of Sample (N = 144 Couples).

	Patients	Spouses
Variables	M (SD) or %	M (SD) or %
Age (years)	65.58 (9.78)	65.31 (11.46)
Sex (male)	43%	57%
Race (White)	87%	85%
Years of education	16.08 (2.01)	15.88 (2.03)
Employed (yes)	42%	46%
Years married/in relationship	34.24 (16.61)	—
Household income
US$10,000–US$39,999	26%	—
US$40,000–US$79,999	38%	—
US$80,000 or more	31%	—
Duration of knee osteoarthritis (years)	12.79 (11.34)	—
Pain management
Cortisone injections into joints	62%	—
Physical therapy	61%	—
Strength/endurance exercise	22%	—

Between-person correlations among study variables are presented in Table 2. Perceived patient pain reported by spouses was moderately correlated with patients’ rating of their own pain (r = .450, p< .001). As expected, perceived patient pain was positively correlated with spouses’ negative affect (r = .281, p < .001). Spousal caregivers’ confidence in patients’ pain management was correlated with perceived patient pain (r = −.321, p< .001) but not with their negative affect (r = −.145, p = .083).

Table 2.

Between-Person Correlations among Study Variables and Covariates.

Variable	1.	2.	3.	4.	5.	6.	7.	8.
1. Perceived patient paina	—
2. Spouse confidence	−.321***	—
3. Spouses’ negative affecta	.281***	−.145	—
4. Patients’ negative affecta	.088	−.218**	.114	—
5. Patients’ rating of their paina	.450***	−.296***	.192*	.290**	—
6. Spouses’ age	−.036	.090	−.066	−.179*	−.131	—
7. Spouses’ sex	.123	.022	.127	−.136	−.013	−.204*	—
8. Spouses’ depressive symptoms	.119	−.125	.556***	.161	.187*	−.028	.112	—

Note. N_spouses = 144; N_patients = 144. Spouses’ sex was coded as male = 0, female = 1.

^aTo compute between-person correlations, daily variables were averaged across the 22-day study period.

^*p < .05.

^**p < .01.

^***p < .001.

Results of multi-level models are exhibited in Table 3. The main effect model indicated that, inconsistent with Hypothesis 1, daily perceived patient pain was not significantly associated with their own negative affect (γ₁₀; b = .017, p = .330). The interaction model revealed that Hypothesis 2 was supported in that a significant cross-level interaction emerged between daily perceived patient pain and spouse confidence in patients’ pain management (γ_11;b= −.005, p = .013). The local effect size of the interaction term (γ₁₁) was small (Cohen’s f² = .001). As depicted in Figure 1, a simple slope analysis indicated that greater daily perceived patient pain was associated with spouses’ higher daily negative affect only for spouses with lower (−1SD) confidence in patients’ ability to manage pain (b = .038, p = .040). In contrast, daily perceived patient pain was not significantly associated with spouses’ daily negative affect among those with average (b = .022, p = .211) or higher (+1SD; b = .007, p = .717) confidence in patients’ pain management.

Table 3.

Associations between Spouses’ Perception of Patients’ Pain and Spouses’ Daily Negative Affect, Moderated by Spouse Confidence in Patients’ Pain Management.

	Main Effect	Interaction
	b (SE)	b (SE)
Fixed effects
Intercept, γ00	0.443*** (0.056)	0.443*** (0.056)
Patients’ rating of their pain, γ20	0.001 (0.016)	0.002 (0.017)
Patients’ negative affect, γ30	0.100** (0.028)	0.093** (0.028)
Spouses’ sex, γ02	0.071 (0.099)	0.068 (0.099)
Spouses’ age, γ03	−0.001 (0.005)	−0.002 (0.005)
Spouses’ depressive symptoms, γ04	0.072*** (0.014)	0.071*** (0.014)
Spouses’ perception of patients’ pain, γ10	0.017 (0.017)	0.022 (0.018)
Spouse confidence, γ01	−0.006 (0.004)	−0.006 (0.004)
Spouses’ perception of patients’ pain × Spouse confidence, γ11	—	−0.005* (0.002)
Random effects
Variance intercept, ${\sigma }_{\mathrm{\mu }0\text{i}}^2$	0.256*** (0.032)	0.256*** (0.032)
Variance spouses’ perception of patients’ pain, ${\sigma }_{\mathrm{\mu }1\text{i}}^2$	—	0.012* (0.006)
Cor intercept, spouses’ perception of patients’ pain, σμ0i, μ1i	—	0.015 (0.012)
Residual variance, ${\sigma }_{\in \text{it}}^2$	0.183*** (0.005)	0.180*** (0.005)

Note. N_spouses = 144; N_observations = 2,682; Spouses’ sex was coded as male = 0, female = 1;

^*p < .05;

^**p < .01;

^***p < .001.

Figure 1.

Cross-Level Interaction between Daily Perceived Patient Pain and Spouse Confidence in Patients’ Pain Management Predicting Spouses’ Daily Negative Affect.

Note. The person-mean centered daily perceived patient pain was plotted along the x-axis, and spousal caregivers’ daily negative affect was plotted along the y-axis. The values of spouse confidence were −8.84 (−1SD), 0 (average), and 8.84 (+1SD), respectively. Each person-mean centered covariate and spouses’ sex equaled zero.

Estimated random effects of the interaction model showed that the slope variance (${\sigma }_{{\mu }_{1i}}^2=.012$, p = .030) was significant, suggesting that the within-person associations between daily perceived patient pain (γ₁₀) and spouses’ daily negative affect varied significantly across individuals after adjusting for the effect of spouse confidence in patients’ pain management on the associations. The covariance between the slope and the intercept was not significant (${\sigma }_{u_{0i},{\mu }_{1i}}=.015$, p = .194), implying that there was no significant association between the regression slopes for daily perceived patient pain (γ₁₀) and the intercept of spouses’ daily negative affect.

Discussion

Using data from a diary study of spousal caregivers of chronic pain patients, the current study examined whether spouses experience more negative affect on days when they perceive that patients’ pain is greater than usual, and whether spouse confidence in patients’ ability to manage pain modulates this association. Contrary to our prediction, there was no significant association between daily perceived patient pain and spouses’ daily negative affect. Yet, consistent with our second hypothesis, spousal caregivers’ lower confidence in their partners’ ability to manage pain significantly moderated the daily association between perceived patient pain and spouses’ negative affect. That is, only spousal caregivers with a relatively low level of confidence in patients’ pain management reported higher negative affect on days when they perceived their partners’ pain to be greater than usual.

This finding is in line with prior research which showed that husbands’ lack of confidence in their wives’ control over RA was related to husbands’ worse psychological adjustment (Sterba & DeVellis, 2009). Social cognitive theory posits that people who do not believe that they can exercise control over threats have disturbing thoughts that produce stress and depression (Bandura, 1997). Similarly, spousal caregivers with lower confidence in patients’ pain management may find it difficult to control their worries and hopelessness, especially when they perceive that patients are facing more pain than usual, and thereby experience greater daily negative affect. Furthermore, according to social cognitive theory (Bandura, 1997), efficacy beliefs may regulate affective distress through their impact on coping behavior. Hemphill and colleagues (2016) reported that spousal caregivers who are less confident in patients’ arthritis management respond to patients’ pain with greater instrumental support, which is often related to higher caregiver distress (Pinquart & Sörensen, 2003; Poulin et al., 2010). Spouses who have a relatively low level of confidence in patients’ pain management may experience more negative affect when perceiving that patients are having pain that is more severe than usual, because they provide more instrumental support to patients than those with higher spouse confidence.

Our finding that the full sample of spousal caregivers did not experience higher negative affect on days when their partners’ pain was greater than usual is consistent with that of Geisser et al.’s (2005) study which found no association between perceived patient pain and spousal caregivers’ affective distress. However, our result appears to contradict the results from some prior research (Monin et al., 2017; Redinbaugh et al., 2002). For instance, Monin and colleagues (2017) reported that spouses of chronic pain patients felt greater daily distress in response to their partners’ physical suffering. Inconsistency in findings across the two studies may be due to different predictor and outcomes examined in each study. We focused on perceived patient pain, while Monin et al. (2017) addressed perceived patients’ physical suffering because of any physical symptoms such as pain or fatigue. Given that patients’ fatigue is associated with their spousal caregivers’ greater depressive symptoms (Passik & Kirsh, 2005), perceived patients’ fatigue might contribute to the significant association between perceived patient suffering and spouses’ distress found in Monin et al.’s (2017) study. Moreover, we asked spousal caregivers about their general negative affect, whereas Monin et al. (2017) asked about caregivers’ distress specific to partners’ physical suffering, which is more likely to be associated with their perceptions of partners’ suffering.

In addition, the different analytic approaches used in each study might contribute to the mixed findings about the association between perceived patient pain and spousal caregivers’ distress. Redinbaugh et al. (2002) reported the significant association, based on between-person analysis. The current study also found the significant between-person correlation between perceived patient pain and spousal caregivers’ negative affect. However, between-person approach is poorly suited for evaluating within-person processes (Molenaar, 2004). Given that an affective reaction to perceived partners’ pain is a within-person process, our finding from within-person analysis may more accurately capture the association than that of prior research. More future research needs to investigate whether perceived patient pain is related to spousal caregivers’ affective distress, using the within-person analysis.

The findings from our study have important theoretical implications. First, our findings support the caregiver stress-health model which proposes that numerous contextual factors may influence the extent to which perceived patient pain relates to their caregivers’ affect (Monin & Schulz, 2009). The current study demonstrated that spousal caregivers’ confidence in patients’ pain management significantly modulated the association between perceived patient pain and spouses’ negative affect. This finding suggests that being exposed to a loved one’s pain is not necessarily distressing for all spouse caregivers, but distressing for particular subgroups of caregivers. Second, the findings imply that spousal caregivers’ confidence in patients’ pain management may play an important role in determining not only patients’ health and well-being (Gere et al., 2014; Hemphill et al., 2016; Rohrbaugh et al., 2004) but also caregivers’ own psychological outcomes. Future research could benefit from examining caregivers’ confidence in patients’ ability to manage symptoms as a contextual factor that influences the associations between stressors and well-being of family members caring for older individuals with other types of chronic illness.

Our findings also inform clinical practice with spousal caregivers of chronic pain patients. We found that the effect size of the interaction between perceived patient pain and spouse confidence was small. However, the estimated effect size is usually small when random regression coefficients are included in the model (Snijders & Bosker, 2012). Furthermore, the effect size found in this study is comparable with those reported in the existing ecological momentary assessment or diary studies on affect (Liao et al., 2017; Zhaoyang & Martire, 2019). Given that spousal caregivers of chronic pain patients witness their loved one in pain in daily life and increased levels of negative affect are associated with caregivers’ poorer health and quality of life (Chung et al., 2009; Mausbach et al., 2007), the small effect of spouse confidence in patients’ pain management can accumulate over time and have a clinically important influence on spousal caregivers’ health and well-being. Therefore, it is important to implement an intervention that elevates spousal caregivers’ confidence in their partners’ pain management. Keefe and colleagues (2004) found that spouses included in OA patients’ pain coping skills and exercise training rated their partners’ pain control and rational thinking higher than spouses who were not included. Similarly, another experimental study showed that wives’ confidence in their husbands’ physical and cardiac capability increased after the experiment among wives who also performed their husbands’ exercise test, but not among those who only watched the test (Taylor et al., 1985). Taken together, couple-oriented interventions in which spousal caregivers actively participate in cognitive-behavioral training with patients may be effective in enhancing caregivers’ beliefs about patients’ ability to manage chronic pain.

There are limitations of the present study that indicate promising directions for future research. First, we did not address why spousal caregivers with a low level of confidence in patients’ pain management experienced higher negative affect on days when they perceived that their partners’ pain was greater than usual. An important next step in this area of research is to examine underlying mechanisms (e.g., spouses’ catastrophizing regarding patients’ pain, spousal support) through which a low level of spouse confidence strengthens the association between perceived patient pain and spousal caregivers’ negative affect. Second, this study measured spousal caregivers’ confidence in patients’ pain management once at the Time 1 interview, based on the assumption that spouse confidence is relatively stable across days. Some studies (Keefe et al., 2001; Zhaoyang et al., 2017) reported that patients’ self-efficacy varied more at the between-person level (74%–79%) than at the within-person level (21%–26%). Future research can examine the extent to which spouse confidence in patients’ pain management varies daily and whether daily spouse confidence moderates the daily association between perceived patient pain and spouses’ negative affect. Third, our sample consisted of spousal caregivers who were primarily middle-class Whites and who did not have usual knee pain of moderate or greater intensity. Future research ought to examine the generalizability of our findings to spousal caregivers with different demographic characteristics or who are also dealing with their own pain.

Despite these limitations, the daily diary methodology used here represents an ecologically valid means of capturing spousal caregivers’ affective responses to their partner’s pain on a daily basis. Our results suggest that spousal caregivers’ daily perceptions of their partners’ pain are significantly associated with greater daily negative affect only for those with lower confidence in their partners’ ability to manage their pain. Thus, these findings suggest that, if replicated using other study populations, spousal caregivers’ lack of confidence in patients’ pain management may be a risk factor for spouses’ higher negative affect, especially when perceiving patients’ greater chronic pain in daily life.