The Relationship Between the Pain Experience and Emotion Regulation in Older Adults

Correspondence to: Thomas Hadjistavropoulos, PhD, Department of Psychology, University of Regina, 3737 Wascana Parkway, Regina, SK S4S 0A2, Canada. Tel: 306-585-4457; Fax: 306-337-2321; E-mail: thomas.hadjistavropoulos@uregina.ca.

Pain Medicine, Volume 21, Issue 12, December 2020, Pages 3366–3376, https://doi.org/10.1093/pm/pnaa135

Cite

Omeed Ghandehari, Natasha L Gallant, Thomas Hadjistavropoulos, Jaime Williams, David A Clark, The Relationship Between the Pain Experience and Emotion Regulation in Older Adults, Pain Medicine, Volume 21, Issue 12, December 2020, Pages 3366–3376, https://doi.org/10.1093/pm/pnaa135

Abstract

To investigate the relationship of emotion regulation strategies (i.e., emotional suppression and reappraisal) with pain catastrophizing, fear of pain, pain intensity, worry, and depression as function of age in samples of older and younger adults.

Cross-sectional design using validated questionnaires.

Participants resided in the community. They completed validated measures using online questionnaires.

Two-hundred fifty-seven older adults and 254 younger adults with chronic pain participated.

Participants completed validated questionnaires of emotion regulation strategies, pain-related functioning and mental health.

Emotion regulation varied as a function of age and gender. Among our chronic pain sample, older adult males reported lower use of reappraisal and suppression than younger adult males, while older adult females reported higher use of reappraisal than younger adult females. Emotional suppression was positively related to pain catastrophizing, pain intensity, worry, and depression. Reappraisal was negatively related to depression and worry. Interestingly, age showed a positive relationship with fear of pain, pain catastrophizing, worry, depression, and pain intensity, while gender was related to fear of pain and worry. Finally, emotional reappraisal partially mediated the relationship between the affective dimensions of pain intensity and pain catastrophizing among older adults.

Our results indicate that reappraisal strategies are important for older and younger adults with chronic pain, pointing to the necessity of considering these strategies when working clinically with such populations. However, given our findings as well as those in the literature, gender should also be considered.

Introduction

Emotion regulation (ER) is the process of monitoring and evaluating emotional reactions and using strategies for altering when and how emotions occur [ 1, 2]. People differ in their use of ER strategies. They may rely on either adaptive (e.g., reappraisal, acceptance, mindfulness, problem solving) or less adaptive strategies, such as avoidance of situations that may elicit an emotion or normal internal experiences. Commonly studied in the ER literature are the strategies of expressive suppression and reappraisal (e.g., [ 3]). Reappraisal involves changing the way one thinks about an emotional event (e.g., reframing a negative state, such as pain, in order to reduce its negative impact), especially a stressful situation, while expressive suppression can, for example, involve hiding emotional states by inhibiting nonverbal behaviors indicative of such states (e.g., 4, 5]). Emotional suppression has been shown to increase physiological symptoms of anxiety and to decrease positive mood compared with reappraisal [ 6].

Emotional states, such as depression and anxiety, as well as ER, play an important role in the pain experience (e.g., [ 7–9]). Koechlin et al. [ 1] systematically reviewed the literature on ER and chronic pain. Overall, most of the studies they reviewed showed that the relationship between ER and pain was mediated by a psychological response (e.g., anxiety, negative mood) [ 1, 10]. Although cognitive reappraisal was not found to have an impact on the level of pain reported in these studies, expressive suppression negatively impacted anxiety, depression, catastrophic thinking, and daily functioning.

Fear of pain and catastrophic thinking about pain also have a negative impact on the pain experience and contribute to disability. High levels of catastrophizing about pain and fear of pain are inversely related to successful recovery from an injury and positive adaptation to chronic pain conditions [ 11]. Pain catastrophizing is also a prospective predictor of chronic pain after knee surgery [ 12] and of temporal summation of pain in chronic low back pain patients [ 13]. Similarly, two meta-analyses have led to the conclusion that fear of pain is related to increased disability and greater pain experienced among chronic pain patients, although further longitudinal studies are required to tease apart the relationships [ 14, 15].

The relationships among pain, negative emotional states (e.g., depression, anxiety), pain catastrophizing, fear of pain, and ER are not entirely consistent across studies. For instance, Wong and Fielding [ 16] investigated the associations among pain catastrophizing, negative affect, ER (cognitive reappraisal and expressive suppression), and level of pain in chronic pain patients. Although they did not find a direct relationship between either of the ER strategies and pain intensity or pain disability, they found that expressive suppression mediated the relationship between negative affect and catastrophizing. In contrast to suppression, cognitive reappraisal was not a mediator in this relationship. Guimond et al. [ 3] found that among cancer patients, experiential avoidance and expressive suppression were related to negative psychological symptoms including anxiety, fatigue, depression, fear of cancer recurrence, insomnia, and cognitive impairments, but not to level of pain. In the experimentally induced pain literature, findings are more varied with regard to the impact of ER strategies (e.g., [ 17–20]). For example, in one study, participants were randomized to one of two ER conditions (i.e., reappraisal or distraction) or a control condition during a cold pressor task (i.e., placing the hand in a bath of cold water to induce pain) [ 21]. These researchers found that neither reappraisal nor distraction was related to pain level experienced during the tasks or pain level remembered. In contrast, in another laboratory experimental investigation, Hampton, Hadjistavropoulos, Gagnon, Williams, and Clark [ 22] found that using either suppression or reappraisal during a pain task was related to significantly lower pain intensity. Use of reappraisal was also related to lessened pain unpleasantness, anxiety, and tension ratings during the pain task.

The relationships among ER, pain level, catastrophizing, fear of pain, and negative emotions such as depression and anxiety have not been investigated among older adults, especially given some contradictory findings in the literature involving younger individuals. It is important to note that pain-related psychological processes often vary as a function of age (e.g., [ 23–25]) and gender (e.g., [ 26–28]). For example, Ruscheweyh et al. [ 25] determined that, while catastrophizing did not generally differ between age groups (i.e., dispositional catastrophizing), catastrophizing about pain directly following noxious stimulation (i.e., situation-specific catastrophizing) was associated with emotional responses to pain (i.e., reappraisal, suppression) in younger adults, whereas it was associated with greater pain intensity in older adults.

Regarding gender, Sullivan et al. [ 27] found that males reported less pain than females, with catastrophizing accounting for this gender difference in reported pain (women had higher catastrophizing scores than men). A theoretical model of ER and pain has been described for women, specifically in relation to genito-pelvic pain/penetration disorder [ 29]. However, this model has not been evaluated empirically, and the difference between men and women with regards to ER and pain has yet to be investigated. Knowledge on how gender and age influence the relationship between pain-related processes and ER is limited. Suppression of negative emotion (e.g., anger) has been linked to greater self-reported pain symptoms in a sample of middle-aged chronic pain patients (e.g., [ 9]), but it is unknown whether these findings generalize to older participants. Investigations of emotions in relation to older adults with pain have generally been focused on those who experience acute pain (e.g., [ 30]), although some investigations examine ER in older populations without pain (e.g., [ 31]). Results from those investigations examining acute pain are limited in that homogeneous convenience samples are often only recruited from hospital or acute care settings (e.g., [ 30]), which hinders the generalizability of the results [ 32].

Our goal was to examine ER strategies (i.e., expressive suppression, reappraisal) in relation to fear of pain, pain catastrophizing, and pain intensity among older and younger adults with chronic pain. In examining the relationship of ER strategies with pain, the roles of age and gender were considered. We formulated the following hypotheses:

1. We expected that older participants, in general, would show greater levels of emotion reappraisal and lower levels of suppression than younger participants (Hypothesis I). However, we were not able to make specific hypotheses about gender (and age), given contrary information in the research and different measures of ER used [ 33]. Although research on age and gender differences in ER has not been conducted among patients with chronic pain, there have been studies examining these differences among a general community sample (e.g., [ 33, 34]). We deemed it important to extend this research in chronic pain samples that differ in age.
2. Based on previous research (e.g., [ 6, 7, 35–38]), we expected that fear of pain and pain intensity would be positively related to pain catastrophizing (Hypothesis II).
3. We expected that reappraisal would be negatively associated and use of suppression would be positively associated with fear of pain, pain intensity, and pain catastrophizing, as well as worry and depression (Hypothesis III).
4. Based on literature that managing negative pain-related emotional experiences through use of appropriate ER strategies can mitigate unpleasant pain-related outcomes [ 9, 22] we predicted that reappraisal and suppression would mediate the influence of fear of pain and pain intensity on pain catastrophizing (Hypothesis IV) ( Figure 1). More specifically, as pain and fear of pain lead to coping strategy usage, the selected coping strategy will, in turn, determine the extent of catastrophizing ( Figure 1). Where pain leads to reappraisal, we expected to see less catastrophizing. Suppression was expected to result in more catastrophizing.

Hypothesized mediational relationships. ERQ = Emotion Regulation Questionnaire; PCS = Pain Catastrophizing Scale; SFMPQ = Short Form McGill Pain Questionnaire; SVFPQ-III = Short Version Fear of Pain Questionnaire-III.

Methods

Participants

Participants (N = 511) were both younger (i.e., between 18 and 64 years, N = 254) and older (i.e., 65 years or older, N = 257) adults with chronic pain (i.e., participants were invited to participate in this research if they had experienced pain for at least three months). The three-month time frame was chosen because this corresponds to the operationalization of chronic pain offered by the International Association for the Study of Pain [ 39]. Participants were recruited through Qualtrics Panels (www.qualtrics.com). This recruitment approach has been successfully used in an increasing number of health-related studies (e.g., [ 40–42]). Qualtrics Panels are a useful method of participant recruitment, as online questionnaires can be created consistent with recommendations made by King et al. [ 43], which provide guidance for reaching populations for research participation that are difficult to recruit, such as older adults with chronic pain. This method allows researchers to maximize sample representativeness by recruiting participants from across North America based on specific inclusion criteria. Before participants take part in a study, demographic information is verified by Qualtrics to protect against misrepresentation of participants. Verification of data accuracy is also incorporated throughout the survey by Qualtrics Panels. Information is checked against a database to ensure that participants are accurately representing themselves. Before participating in this study, participants provided informed consent. All participants were compensated for their time by Qualtrics. Approval for this study was granted by our institutional ethics review board.

Measures

Demographic Questionnaire

Our Demographic Questionnaire consisted of questions about age, gender (participants were asked to tell us what gender they were; it was assumed that participants would respond according to the gender they identified as, rather than their biological assignment at birth), relationship status, and whether they were currently living with pain (answered “yes” or “no”). If they answered “yes” to this question, we further queried the length of time living with chronic pain (measured via self-report regarding when pain was first noticed) and current pain medications.

Emotion Regulation Questionnaire

The Emotion Regulation Questionnaire (ERQ) [ 4] is a 10-item scale that measures the degree to which respondents control their emotions through cognitive reappraisal and expressive suppression. Items are rated on a seven-point Likert scale ranging from 1 (i.e., strongly disagree) to 7 (i.e., strongly agree). The ERQ is comprised of a six-item cognitive reappraisal subscale and a four-item expressive suppression subscale. Elevated scores on each subscale reflect a higher likelihood of using that respective ER strategy. Confirmatory factor analysis supports a two-factor structure of the ERQ [ 44]. Adequate internal consistencies have also been reported [ 4, 45]. In administering this measure to older adults, Brady et al. [ 46] reported satisfactory internal consistency of α = 0.89 for the reappraisal subscale and α = 0.80 for the suppression subscale. Our internal consistencies were consistent with these (α = 0.86 for the reappraisal subscale and α = 0.80 for the suppression subscale).

Center for Epidemiologic Studies–Depression Scale

The Center for Epidemiologic Studies–Depression Scale (CES-D) [ 47] is a 20-item scale that assesses symptoms of depression over the past week. Items are rated on a four-point Likert-type scale ranging from 0 (i.e., rarely or none of the time) to 4 (i.e., most or all of the time). The initial development of the CES-D was tailored for the general population; however, support for the use of the CES-D in older adult populations exists (e.g., [ 48, 49]). Among samples of older adults, internal consistency is adequate [ 49]. In this study, the internal consistency of the CES-D was excellent (α = 0.93).

Penn State Worry Questionnaire–Abbreviated

The Penn State Worry Questionnaire–Abbreviated (PSWQ-A) [ 50] is an eight-item scale, shortened from the original 16-item PSWQ [ 51], which measures generalized worry. Items are rated on a five-point Likert-type scale ranging from 1 (i.e., not at all typical of me) to 5 (i.e., very typical of me). The original PSWQ has demonstrated adequate validity and reliability [ 51], but concerns about its generalizability to older adults have been raised (e.g., [ 50, 52]). The PSQW-A, however, has demonstrated adequate internal reliability, test–retest reliability, and construct validity among younger and older adults [ 50, 52]. In our sample, the PSWQ-A had excellent internal consistency (α = 0.96).

Shortened Version of the Fear of Pain Questionnaire–III

The Shortened Version of the Fear of Pain Questionnaire–III (SVFPQ-III) [ 53] is a 15-item shortened version of the original 30-item measure designed to assess fear of severe pain, minor pain, and medical pain. Each pain-related situation constitutes its own five-item subscale. Items are rated on a five-point Likert-type scale ranging from 1 (i.e., not at all) to 5 (i.e., extreme). McNeil and Rainwater [ 54] reported adequate internal consistency and test–retest reliability for the SVFPQ-III. Albaret et al. [ 55] determined that the SVFPQ-III yielded adequate internal consistency for all three subscales in samples of younger and older adults. For the purposes of this study, we used the total score, which has excellent internal consistency (α = 0.89).

McGill Pain Questionnaire Short Form

The McGill Pain Questionnaire Short Form (SFMPQ) [ 56] measures pain qualities through 11 sensory (e.g., stabbing) and four affective (e.g., fearful) words. Each word is rated on a four-point Likert-type scale ranging from 0 (i.e., none) to 3 (i.e., severe). The measure includes a visual analog scale assessing pain intensity and a verbal descriptor scale assessing current pain. Wright, Asmundson, and McCreary [ 57] showed that the measure has adequate intraclass correlation coefficients and test–retest reliability with younger and older adults. The SFMPQ has been used successfully with older adults who have chronic pain (e.g., [ 58–60]). In this study, internal consistencies were adequate for the affective (α = 0.82) and sensory (α = 0.83) subscales and excellent for the total scale (α = 0.89). The total score was used to test Hypotheses II and III, whereas the subscale scores were used to test Hypothesis IV.

Pain Catastrophizing Scale

The Pain Catastrophizing Scale (PCS) [ 37] measures catastrophic cognitions associated with pain. The 13-item measure asks respondents to reflect on past painful experiences and to indicate the frequency with which they experienced each thought or feeling associated with pain. Items are rated on a five-point Likert scale ranging from 0 (i.e., not at all) to 4 (i.e., all the time). The PCS has demonstrated adequate internal consistency and test–retest reliability in younger and older adults [ 37]. For the purposes of this study, the total scale score was shown to be excellent (α = 0.94).

Procedure

Participants completed the Demographic Questionnaire, ERQ, PSWQ-A, CES-D, SVFPQ-III, SFMPQ, and PCS online through Qualtrics Panels.

Analyses

Frequencies and descriptive statistics were calculated for demographic characteristics as well as questionnaire scores on the ERQ, PSWQ-A, CES-D, SVFPQ-III, SFMPQ, and PCS for all participants. If significant differences other than with respect to age were found in demographic characteristics between older and younger adults, those demographic characteristics were considered as possible covariates in subsequent analyses.

To examine age and gender differences (Hypothesis I), two 2 × 2 analyses of variance (ANOVAs) or analyses of covariance (ANCOVAs) were completed. In the first ANOVA, ERQ reappraisal was the dependent variable, with age group and gender as between-subject factors. In the second ANCOVA, age group and gender were the between-subject factors, relationship status was a covariate (because there were age differences with respect to that variable), and the ERQ suppression subscale was the dependent variable.

The second and third hypotheses involved multiple regression analyses for which a conservative regression approach was utilized (e.g., [ 61]), wherein the full model was tested first. If the full model was significant, each variable’s independent and unique contribution to the regression was examined after all other variables were entered into the equation. With respect to Hypothesis II (i.e., that fear of pain and pain intensity would relate to pain catastrophizing scores), a multiple regression analysis was conducted with PCS as the dependent variable and gender, total SFMPQ score, and total SVFPQ-III score as the independent variables. To test Hypothesis III (i.e., that reappraisal would be negatively associated and suppression positively associated with fear of pain, pain intensity, pain catastrophizing, depression, and worry), five multiple regression analyses were conducted with the ERQ subscales as independent variables (both reappraisal and suppression were entered into the same model). A separate analysis was conducted for each of the dependent variables (i.e., PSWQ-A, CES-D, SVFPQ-III–total score, SFMPQ–total score, PCS).

To test Hypothesis IV (i.e., inclusion of reappraisal will decrease the effect of fear of pain and pain intensity on catastrophizing, and inclusion of suppression will increase the effect of fear of pain and pain intensity on catastrophizing), mediation analyses were conducted to examine the mediating effect of ER strategies (i.e., reappraisal, suppression) on the relationship between fear of pain (SVFPQ-III total score) and pain catastrophizing, as well as pain intensity (SVMPQ affective subscale scores and SVMPQ sensory subscale scores) and pain catastrophizing ( Figure 1). These analyses were conducted separately across younger and older adults. Consistent with previous research [ 62–64], each analysis was conducted using 95% confidence intervals and 5,000 bootstrapped samples. All mediation analyses were completed through Hayes’ [ 65] process macro (http://www.processmacro.org/index.html) for IBM SPSS Statistics 21. This macro used path analyses to estimate the direct and indirect effects of an independent variable on an outcome without requiring data to adhere to assumptions of normality. In these analyses, intervals not including zero were considered significant.

Results

Demographic Characteristics and Questionnaire Scores

The sample included 511 participants. Additional demographic characteristics and questionnaire scores are presented in Table 1. All participants reported chronic pain. To confirm that the sample was taken from a chronic pain population, the means and standard deviations for the SFMPQ and PCS were compared with those of other chronic pain samples and were found to be similar [ 37, 56, 66, 67]. A chi-square analysis revealed significant differences between older and younger adults with respect to self-reported use of pharmacological interventions to manage chronic pain (χ 2 (1, N = 511) = 19.24, P 0.001). Participants’ use of pharmacological interventions to manage their pain was therefore considered as a potential covariate in exploratory analyses.

Demographic characteristics and questionnaire scores

CES-D = Center for Epidemiological Studies–Depression Scale; ERQ = Emotion Regulation Questionnaire; PCS = Pain Catastrophizing Scale; PSWQ-A = Penn State Worry Questionnaire–Abbreviated; SFMPQ = McGill Pain Questionnaire Short Form; SVFPQ-III = Shortened Version of the Fear of Pain Questionnaire–III.

Demographic characteristics and questionnaire scores

CES-D = Center for Epidemiological Studies–Depression Scale; ERQ = Emotion Regulation Questionnaire; PCS = Pain Catastrophizing Scale; PSWQ-A = Penn State Worry Questionnaire–Abbreviated; SFMPQ = McGill Pain Questionnaire Short Form; SVFPQ-III = Shortened Version of the Fear of Pain Questionnaire–III.

Hypothesis I

Older and younger adults were compared on demographic variables of interest to determine the use of covariates in subsequent ANOVAs. Significant differences were found with respect to duration of chronic pain (t(504) = 8.55, P 0.001), relationship status (χ 2 (5, N = 511) = 195.31, P 0.001), and medication use (χ 2 (1, N = 511) = 19.24, P 0.001). Older and younger adults did not differ with respect to gender. Significant variables were considered as covariates in follow-up ANOVAs only if the variables were also significantly correlated with the ERQ reappraisal or suppression subscales, which were the dependent variables of interest. No covariates were used for reappraisal. For suppression, a 2 × 2 ANCOVA was run with relationship status as the covariate because there were age differences for that variable. No other covariates were used for suppression.

For reappraisal, results indicated a significant interaction effect between age group and gender (F(1, 507) = 11.78, P = 0.001, η 2 = 0.02). Follow-up t tests revealed that younger adult males (M [SD] = 30.42 [6.19]) scored significantly higher on the ERQ reappraisal subscale than older adult males (M [SD] = 28.45 [5.51], t(250) = −2.64, P = 0.009). Also, younger adult females (M [SD] = 28.91 [6.70]) scored lower than older adult females (M [SD] = 30.72 [5.84], t(257) = 2.24, P = 0.026). For suppression, a main effect for gender (F(1,506) = 47.08, P 0.001, η 2 = 0.08) and relationship status (F(1,506) = 4.28, P = 0.039, η 2 = 0.008) was identified. Males (M [SD] = 17.75 [4.55]) scored significantly higher than females (M [SD] = 14.59 [5.27]) on the suppression subscale of the ERQ (t(509) = 7.24, P 0.001). Younger adult males (M [SD] = 18.43 [4.56]) scored significantly higher on the ERQ suppression subscale than older adult males (M [SD] = 17.09 [4.47], t(250) = −2.39, P = 0.018). Younger females (M [SD] = 14.55 [5.40]) scored almost the same as older females (M [SD] = 14.47 [5.05]).

Hypothesis II

Regression equation examining the unique variance in pain catastrophizing accounted for by demographic characteristics and questionnaire scores

CES-D = Center for Epidemiological Studies–Depression Scale; PSWQ-A = Penn State Worry Questionnaire–Abbreviated; SFMPQ = McGill Pain Questionnaire Short Form; SVFPQ-III = Shortened Version of the Fear of Pain Questionnaire–III.

Regression equation examining the unique variance in pain catastrophizing accounted for by demographic characteristics and questionnaire scores

CES-D = Center for Epidemiological Studies–Depression Scale; PSWQ-A = Penn State Worry Questionnaire–Abbreviated; SFMPQ = McGill Pain Questionnaire Short Form; SVFPQ-III = Shortened Version of the Fear of Pain Questionnaire–III.

Hypothesis III

The regression results are presented in Table 3. All five full models that we tested were significant (i.e., for pain intensity, fear of pain, pain catastrophizing, depression, and worry).

Regression equation examining the unique variance in dependent variables accounted for by demographic characteristics and emotion regulation strategies

CES-D = Center for Epidemiological Studies–Depression Scale; ERQ Reappraisal = Emotion Regulation Questionnaire–Reappraisal; ERQ Suppression = Emotion Regulation Questionnaire–Suppression; PSWQ-A = Penn State Worry Questionnaire–Abbreviated; SFMPQ = McGill Pain Questionnaire Short Form; SVFPQ-III = Shortened Version of the Fear of Pain Questionnaire–III.

Full model was significant.

Regression equation examining the unique variance in dependent variables accounted for by demographic characteristics and emotion regulation strategies

CES-D = Center for Epidemiological Studies–Depression Scale; ERQ Reappraisal = Emotion Regulation Questionnaire–Reappraisal; ERQ Suppression = Emotion Regulation Questionnaire–Suppression; PSWQ-A = Penn State Worry Questionnaire–Abbreviated; SFMPQ = McGill Pain Questionnaire Short Form; SVFPQ-III = Shortened Version of the Fear of Pain Questionnaire–III.

Full model was significant.

The full model involving SFMPQ total scores (i.e., pain intensity) as the dependent variable was significant (F(6,499) = 3.70, P 0.05, R 2 = 0.04). An examination of each variable’s unique contribution to the equation suggested that age and ERQ suppression scores made independent and unique contributions to the model. Age made the greatest contribution (younger adults had higher scores on the SFMPQ), followed by suppression (positively associated with SFMPQ) ( Table 3).

The full model with SVFPQ-III total scores (i.e., fear of pain) as the dependent variable was also significant (F(6,499) = 4.67, P 0.001, R 2 = 0.05). An examination of each variable’s unique contribution to the equation suggested that gender (females had higher total scores on the SVFPQ-III) and age (younger adults had higher total scores on the SVFPQ-III) made independent and unique contributions to the model. Gender contributed to a greater degree than age.

The full model with PCS total scores (pain catastrophizing) as the dependent variable was significant (F(6, 499) = 10.87, P 0.05, R 2 = 0.12). An examination of each variable’s unique contribution to the equation suggested that age (younger adults scored higher on the PCS) and ERQ suppression scores (positively associated with the PCS) made independent contributions to the model. Age contributed more than suppression, but both were highly significant (P < 0.001).

The full model was significant for PSWQ-A total scores (i.e., worry; F(6,499) = 19.56, P < 0.001, R 2 = 0.19). Examination of each variable’s unique contribution to the equation suggested that, consistent with the study’s hypotheses, the ERQ reappraisal subscale score (negatively associated with the PSWQ-A) and ERQ suppression subscale score (positively associated with the PSWQ-A) made independent and unique contributions to the model. Participant gender (women scored higher on the PSWQ-A) and age group (younger adults scored higher on the PSWQ-A) also made independent and unique contributions to the model. Age contributed the most, followed by gender, reappraisal, and suppression.

The full model with CES-D total scores (depression) as the dependent variable was significant (F(6,499) = 22.95, P 0.001, R 2 = 0.22). An examination of each variable’s unique contribution to the equation suggested that age, ERQ suppression scores, ERQ reappraisal scores, and length of time living with pain made independent contributions to the model. Age contributed the most (younger adults scored higher on the CES-D), followed by suppression (positively associated with the CES-D), reappraisal (negatively associated with the CES-D), and duration of chronic pain (positively associated with the CES-D).

To summarize the independent and unique contributions, 1) concerning the model with pain intensity as the dependent variable: age and suppression both contributed to the model; 2) for fear of pain as the dependent variable: gender and age contributed to the model; 3) for pain catastrophizing: age and suppression contributed to the model; 4) for worry: age, gender, reappraisal, and suppression contributed to the model; and 5) for depression: age, suppression, reappraisal, and duration of chronic pain contributed to the model. Specifically summarizing the results for the ER variables, suppression was related to pain intensity, pain catastrophizing, worry, and depression, while reappraisal was related to worry and depression. As expected, suppression was positively associated with all the variables, and reappraisal was negatively associated with depression and worry.

Hypothesis IV

With regards to older adults, ERQ reappraisal subscale scores did not fully mediate any of the hypothesized models; however, partial mediation of a single model was found. That is, the direct effect of affective dimensions of pain intensity (SFMPQ affective subscale) on pain catastrophizing (b = 2.00, 95% confidence interval [CI] = 1.69 to 2.31) and the indirect effect of affective dimensions of pain intensity (SFMPQ affective subscale) on pain catastrophizing were significant (b = –0.06, 95% CI = –0.17 to –0.01). The direct effects of fear of pain (SVFPQ-III total score) on pain catastrophizing (b = 0.45, 95% CI = 0.35 to 0.55) and the direct effects of sensory dimensions of pain intensity (SFMPQ sensory subscale) on pain catastrophizing (b = 0.97, 95% CI = 0.81 to 1.12) were both significant, but the indirect effects for these two models were nonsignificant.

Among older adults, ERQ suppression subscale scores did not mediate any of the hypothesized models. However, the direct effects of fear of pain (b = 0.44, 95% CI = 0.35 to 0.54), sensory dimensions of pain intensity (SFMPQ sensory subscale; b = 0.94, 95% CI = 0.79 to 1.09), and affective dimensions of pain intensity (SFMPQ affective subscale; b = 1.90, 95% CI = 1.61 to 2.20) on pain catastrophizing were significant (although the indirect effects were nonsignificant).

With regards to younger adults, ERQ reappraisal subscale scores did not mediate any of the hypothesized relationships. The direct effect of fear of pain (SVFPQ-III total score) on pain catastrophizing was significant (b = 0.48, 95% CI = 0.37 to 0.60), but the indirect effect was not. The direct effect of sensory dimensions of pain intensity (SFMPQ sensory subscale) on pain catastrophizing was significant (b = 1.14, 95% CI = 0.95 to 1.32), but the indirect effect was not. The direct effect of affective dimensions of pain intensity (SFMPQ affective subscale) on pain catastrophizing was significant (b = 2.32, 95% CI = 1.97 to 2.66), but the indirect effect was not. Among younger adults, ERQ suppression subscale scores also did not mediate any of the hypothesized models. Although the direct effects of fear of pain (SVFPQ-III total score; b = 0.47, 95% CI = 0.35 to 0.59), sensory dimension of pain intensity (SFMPQ sensory subscale; b = 1.14, 95% CI = 0.95 to 1.32), and affective dimension of pain intensity (SFMPQ affective subscale; b = 2.32, 95% CI = 1.98 to 2.66) on pain catastrophizing were significant, the indirect effects were not.

To summarize, for older adults, although we did not find that either reappraisal or suppression fully mediated the hypothesized relationships, reappraisal was found to partially mediate the relationship between the affective pain subscale score and pain catastrophizing. None of our hypothesized relationships were confirmed for the younger adults.

Discussion

We examined use of ER strategies in separate samples of older and younger adults with chronic pain. We also investigated the relationship of two ER strategies (emotional suppression and reappraisal) with pain, fear of pain, pain catastrophizing, and negative emotions (worry and depression) while taking age and gender into consideration. In the first instance, we found that use of ER varied as a function of age and gender. Older men reported lower use of both re-appraisal and expressive suppression compared with younger men, whereas older women reported higher use of reappraisal than younger women. It is not clear why women seemed to increase their use of re-appraisal over time in our sample and men did not. There are inconsistencies across ER studies, possibly related to measurement differences, but our findings are more similar to those of John and Gross [ 34], who used the ERQ. Although emotional regulation has not been investigated among older adults with chronic pain, among older community-residing women who were not selected for chronic pain and who were asked to retrospectively report their emotion regulation strategies at age 20, reappraisal was found to increase with age, whereas suppression decreased [ 34]. Nolen-Hoeksema and Aldao [ 33], who did not use the ERQ, found that reappraisal decreased in later years (age 65–75) for men and women, whereas suppression increased for older women but not for men. It is possible that these differences across studies are related to different measurement methodologies. More research on the relationship of ER with age and gender is needed to clarify existing discrepancies in study results.

We also found that, among older adults, reappraisal partially mediated the relationship between the affective dimensions of pain intensity and catastrophizing. This was partially consistent with our hypothesis (we expected full mediation). Contrary to our expectations, neither ER strategy fully mediated the relationship between pain and pain catastrophizing or partially mediated this relationship among younger adults (although, as stated above, one model was significant among older adults). In previous research, it has been suggested that various psychological variables, including catastrophizing, mediate the relationship between ER and pain [ 1], and it has been shown that suppression mediates the relationship between negative affect and catastrophizing [ 16]. It may be that the affective subscale of the SFMPQ (the measure we used to capture negative affect) did not in actuality capture the negative affect component measured in the study by Wong and Fielding [ 16] and instead relates more to pain intensity. It could be that negative affect stands as an intermediate step between pain intensity and catastrophizing and that ER mediates the relationship between negative affect and catastrophizing. This would explain the discrepancy between our results and those of Wong and Fielding [ 16] but requires empirical investigation. Future models investigating the complex relationship between pain level, ER, and catastrophizing should consider negative emotional variables aside from the affective subscale of the SF-MPQ.

When examining all participants, both ER strategies demonstrated important effects in our results. We found that emotional suppression significantly related to pain catastrophizing, pain intensity, worry, and depression. The negative relationship of reappraisal with worry and depression was also established. There also may be benefit in further elucidating why suppression was related to pain catastrophizing and pain intensity as well as depression and worry, whereas reappraisal was only related to depression and worry in this study. We had anticipated that both suppression and reappraisal would relate to all the variables examined, albeit in opposite directions. This latter finding contrary to our hypothesis, but consistent with Wong and Fielding’s [ 16] finding in their mediation study; suppression displayed a positive relationship with pain intensity and pain catastrophizing. However, it is uncertain why the use of the ER suppression strategy did not relate to fear of pain.

Overall, the results of this study support age and gender differences with respect to ER strategy use. These findings regarding age and gender are consistent with the previous investigation by John and Gross [ 34]. For example, in this study, males reported more use of suppression than females, consistent with the belief that males tend to avoid emotional reactivity [ 66]. Taken together, our findings on ER strategies point to the value of focusing on ER strategies in clinical work with pain patients, especially older adults, and focusing these interventions differently among men and women. Offering an intervention to increase older men’s use of reappraisal strategies and lessening their use of emotional suppression may lead to improved pain-related outcomes, whereas older women may require less focus on increasing their reappraisal strategies.

Limitations and Directions for Future Research

Despite our identification of age differences in ER strategies and of other relationships between ER strategies and pain outcomes, our findings were cross-sectional in nature, limiting our confidence about the direction of causality in the identified relationships. As a future direction, longitudinal studies would allow for the examination of causality in these relationships. Also, laboratory examinations could provide more insight into the effect of the use of suppression and reappraisal on variables of interest. Although our sample sizes were sufficient to detect medium effect sizes in the mediation models, having larger sample sizes may have enabled us to detect additional small effect sizes [ 67]. Finally, although the ERQ has been used in pain contexts [ 16, 22, 68], the use of self-report has limitations; it is not known how well self-report would compare to more objective indices of emotional regulation (e.g., thinking-aloud procedures describing responses in real time) in the pain context, as, to the best of our knowledge, this has not previously been investigated with regards to ER. Nonetheless, this issue has been investigated in the context of emotional intelligence. Specifically, people who were the least skilled in emotional intelligence had limited insight into their deficits and were more reluctant to pursue improvement than people who were more skilled in emotional intelligence (e.g., [ 69]).

As part of this study, participants living with chronic pain were recruited independent of pain location or type of pain experienced (e.g., headache pain vs arthritis pain). Rustøen et al. [ 70] determined that participants in an older age group reported longer pain durations and more comorbidities while receiving more pain interventions. Despite these findings, older adults had higher quality of life scores and reported better mood. Similar findings were replicated in this study. Specifically, despite living with chronic pain for longer periods of time, older adults report better mood and lower overall anxiety compared with younger adults. Even so, future studies should account for the type of chronic pain being experienced with regards to the type of ER strategy relied upon by younger and older adults.

Finally, results from the present study demonstrate a positive correlation between the use of suppression and both depression and anxiety. We also showed that use of reappraisal partially mediated the relationship between the affective qualities of pain and pain catastrophizing among older adults. These findings suggest that future investigations of clinical ER strategies aimed at improving chronic pain management are warranted. This may be helpful for practitioners of brief therapies targeting chronic pain among older adults, including Brief Acceptance and Commitment Therapy (ACT) (e.g., [ 71]) and Brief Cognitive Behavioral Therapy (CBT) [ 72]. Although ER training is a component of both therapies [ 73, 74], further focus on these aspects, especially reappraisal, may be of value for chronic pain management among older adults.

Conclusions

We documented age and gender differences in the use of ER strategies. Moreover, we replicated findings of the positive relationship between ER suppression strategy and depression and anxiety. The negative relationship between ER appraisal and depression and generalized worry was also established. Despite the finding in an older adult sample that reappraisal partially mediated the relationship between the affective dimensions of pain intensity and catastrophizing, evidence that suppression and reappraisal played a fully mediating role between pain intensity or fear of pain and pain catastrophizing in older or younger adults with chronic pain was not found. It is likely that coping strategies (e.g., [ 75]) other than suppression and reappraisal are just as effective, and possibly more influential, in chronic pain management.

Funding sources: This work was supported by the Saskatchewan Health Research Foundation under Grant 2908.

Conflicts of interest: The authors have no conflicts of interest to declare.

Trial registration: orcid.org/0000-0002-8586-0450.