Psychological resilience associates with pain experience in women treated for breast cancer

2.1 Participants

The study was approved by the Coordinating Ethics Committee of the Helsinki and Uusimaa Hospital District (reference number 149/13/03/00/14). All participants gave their informed consent in writing.

This is a cross-sectional study. The participants were recruited for the current NeuroPain study from the BrePainGen study, which assessed acute and persistent pain among 1,000 women treated for breast cancer at the Helsinki University Hospital during 2006–2010 [21]. Detailed description of patient selection is presented in Fig. 1. The participants with a surgical nerve lesion are described in more detail elsewhere [19].

Fig. 1:

Flow-chart of the patient selection from the original 1,000 patient cohort.

The NeuroPain cohort consisted of 402 participants who filled in multiple questionnaires during the research visit which took place 4–9 years after the initial breast-cancer surgery. Of these, 160 reported at least moderate (NRS≥4) pain severity or pain interference in any area of the body during last week and were included in the present study (Fig. 1). Information about breast-cancer treatment was collected from the patients’ records.

2.2 Questionnaires

2.3 The Cold Pressor test

The Cold Pressor test (CPT) was used to assess experimental pain sensitivity and tolerance. The participants immersed their hand in a bath with circulating cold water (+2–4 C) (JULABO USA Inc., Allentown, PA, USA), holding it in up to the wrist for the maximum time tolerated or for up to 90 s. They reported pain severity every 15 s, and at the end of the CPT when they withdrew their hand from the water on an NRS ranging from zero to 10. After the test, the participants also evaluated its unpleasantness on an NRS scale ranging from zero to 10. Relative pain unpleasantness was calculated by dividing the unpleasantness score by the pain-intensity score at the end of the test [20].

2.4 Statistical analysis

In cases of single missing items on the questionnaires (less than 20% missing values on one questionnaire), the missing values were replaced with the means of the participant’s responses from the same questionnaire. If the missing values amounted to more than 20% of the questionnaire it was removed from the analyses. When reporting descriptive statistics (Table 1), we report means and standard deviations when the variables are normally distributed and medians and inter quartile ranges when the variables are skewed. Pearson’s correlation was used to describe the strength of the association between resilience and pain severity or interference. Relative pain interference was calculated by dividing the pain-interference score by the pain-intensity score. Since linear regression is sensitive to the effect of outliers, we removed one participant with an dependent variable 6.4 standard deviations above the mean. We used the enter method in linear regression and created two models. In the first model we analyzed the association between resilience and relative pain interference. In the second model we entered resilience, anxiety and the interaction term (resilience x anxiety). The same analysis was conducted for resilience, anxiety and relative pain unpleasantness in the CPT. Resilience and anxiety were used as continuous variables. Anxiety was grouped into high and low groups for illustrative purposes. SPSS 22.0 version for Windows (SPSS INC, Chicago, IL, USA) was used for the statistical analyses.

3.1 Persistent Pain

The association between resilience and pain severity was not significant (Fig. 2A). However, pain interference and resilience had a significant negative correlation (Fig. 2B).

Fig. 2:

(A) Scatterplot of resilience score and severity of persistent pain. (B) Scatterplot of resilience and interference of persistent pain.

We created a linear regression model to predict relative pain interference by resilience, anxiety, and their interaction term as independent variables. In model 1, the resilience score was negatively associated with relative pain interference (Table 3). In model 2, the resilience score was negatively associated, the anxiety score was positively associated, and there was also a significant interaction between resilience and anxiety on relative pain interference (Table 3). The interaction is visualized in Fig. 3. The negative relationship between resilience and relative pain interference was stronger among participants with low anxiety than among those with high anxiety. Of the former, those with high resilience reported lower relative pain interference than those with low resilience. This relationship between relative pain interference and resilience was not as pronounced in participants with high anxiety.

Fig. 3:

Association of psychological resilience, anxiety and their interaction on Relative Pain Interference of persistent pain.

3.2 Experimental pain

All 160 participants participated in the CPT and of them, 67 participants (42%) held their hand in the ice-cold water in the CPT for the maximum of 90 s. There was no difference in the level of resilience between the participants who withdrew their hand before the maximum time and those who held it there for the maximum time [78.58 (12.60) vs. 76.29 (13.19), t=−1.11, p=0.27]. Neither the final NRS in the CPT nor unpleasantness at the end associated with psychological resilience (r=-0.01, p=0.94; r=−0.06, p=0.46, respectively). The relative Pain Unpleasantness of the CPT was not associated with resilience (β=−0.03, p=0.754) or anxiety (β=−0.006, p=0.955) in the linear regression model (enter method), and there was no interaction between them (β=−0.065, p=0.500).

4.1 Resilience among cancer survivors

Women treated for breast cancer react differently to potentially traumatising events related to diagnosis and treatment. Some of them have negative psychological outcomes, whereas others show remarkable resilience and even post-traumatic growth [26]. In the present study, the level of resilience was similar to previous reports of Finnish healthy sample (17–92 years) [24]. Features like optimism, hope, and a sense of coherence and positive personality-related variables e.g. personal control and low neuroticism and social support are associated with higher psychological resilience among cancer patients [26]. In some cases, experiencing and surviving cancer is a transformative process that fosters the capacity to enjoy small daily pleasures and to attach less importance to former areas of concern, which in turn is reflected in increased resilience [27].

4.2 Resilience and pain

It has been suggested that vulnerability and protective factors may work simultaneously when an individual adapts to chronic pain [28], but it is only recently that these factors have been studied simultaneously. Previous studies have suggested that anxiety may moderate the effect of resilience in cases of chronic and experimental pain [20], [29]. We did find a moderating effect of anxiety on resilience and relative pain interference in those who had persistent pain: the effect of resilience was more pronounced in those with low anxiety. This could be attributable to the nature of resilience which can only enhance adjustment among people with a substantial level of psychological wellbeing. However, another study on experimental pain reported that the effect of resilience was more pronounced among people with high anxiety [20]. It is possible that there are multiple other confounding factors. In a previous study, resilience was found to associate with pain-related dysfunction via specific pain beliefs (pain self-efficacy, pain catastrophizing and challenge appraisals) [30]. The prevalence of anxiety symptoms in women treated for breast cancer varies between 17.9 and 33.3%, which is higher than that in the average female population [31]. Anxiety is an important vulnerability factor in chronic pain. Also, anxiety, in comparison with depressive symptoms, has been reported to associate more strongly with persistent pain [2].

We included participants reporting their worst pain NRS ≥4/10 in our analysis (0=no pain, 10=worst imaginable pain). The NRS cut-off point ≥4/10 has been widely used to identify patients with moderate-to-severe or clinically significant pain [32]. Unlike in previous studies [14], [16], there was no direct association between psychological resilience and the pain-severity rating. On the other hand, we did find an association between resilience and pain interference, which is in line with a study on upper limb loss [15]. The differences in the findings could be attributable to the different patient cohorts and the type of pain. Ankawi et al. studied a sample of people reporting ongoing pain on at least 3 days a week for 6 months or more [16], whereas Ruiz-Parraga et al. focused on patients with chronic musculoskeletal pain appearing on five or more days a week for at least 6 months [14]. A pain-specific resilience scale was used in both studies, whereas we used a more general resilience scale. According to Ruiz-Parraga et al., however, the weak although statistically significant association between resilience and pain severity implies that resilience in chronic pain has a stronger relationship with low interference in daily functioning and low emotional distress than with the intensity of pain [14]. Ankawi et al. also suggest that the relationship between resilience and intensity could be mediated by the acceptance of chronic pain and pain-related self-efficacy [16]. After all, even though we found an association between resilience and pain interference in this study, the size of the effect is rather small. However, this is an important piece of information as it will help us better understand the complex phenomena of living with persistent pain. In this study, we created a theoretical model where state anxiety moderates the effect of resilience on relative pain interference. It would also be possible to create a model where resilience moderates the effect of anxiety on relative pain interference as the moderating effect is likely to be in both directions.

4.3 Experimental pain

In a previous study in an experimental pain setting, relative pain unpleasantness was associated with psychological resilience, and anxiety moderated this effect [20]. We did not find such an association in experimental pain, which may be due to the different experimental tests applied: Hemington et al. [20] used suprathreshold heat pain, whereas we used the cold pressor test. It is possible that anxiety and resilience affect pain severity differently in the different modalities. Additionally, Hemington et al. [20] studied trait anxiety, whereas we studied state anxiety. In this study, we found a protective effect of resilience and a moderating effect of anxiety on relative pain interference in clinical pain but not on relative unpleasantness in experimental pain. Experimental pain is very different from clinical pain on the psychological level in that it does not represent a threat to the person. In fact, mild anxiety has been found to improve tolerance of heat pain in healthy subjects [33].

4.4 Study limitations and strengths

Our study had only female participants. Females have reported more health-related problems than men in earlier studies, and gender differences become more pronounced with increasing age [34]. The results of previous studies on gender differences in psychological resilience are mixed [13], [24], [35], [36], [37], [38]. It has been suggested that resilience operates differently in older than in younger adults [39], [40]. At the time of the breast cancer treatments, the mean age of the present sample was 55.3 (7.9) years which is slightly younger than that of the average age (60 years) of the breast cancer patients in Finland (Finnish Cancer Registry). In this current study, the age limit was set to 75 which explains why the study group had more younger patients. This limitation is unlikely clinically significant.

Another limitation is that this study has the cross sectional design. The results are therefore correlational only. Further limitation of the study is that the pain variables are skewed to the right. That; it also affects the relative pain interference variable that, which is created from the pain severity and pain interference variables since there is not much variance in these variables. However, the relative pain interference variable itself is normally distributed. A strength of the study is the minimal amount of missing values. Another strength of this study was that in addition we asked the participant to rate possible pain in the postsurgical area and in other areas separately, which provided a better picture of their overall pain problem.

4.5 Clinical implications

According to our results, anxiety moderated the effect of resilience on pain interference. However, it seems that persons with high anxiety cannot exploit the protective role of resilience. Therefore, it may be necessary in clinical settings to target the treatment of chronic pain differently according to the patient’s level of anxiety. Appropriate treatment for anxiety may have to precede resilience-enhancing interventions. On the other hand, improving resilience among women treated for breast-cancer at an early stage could be beneficial in terms of inhibiting the development of high levels of anxiety.

Previous studies have suggested that resilience might be a viable target for intervention [41]. Improving patients’ resilience and resilient behaviour could be a significant step in the development of new treatments or rehabilitation for patients with persistent pain and for cancer survivors. The available evidence indicates a modest but consistent benefit of resilience-training programmes in improving a number of mental-health outcomes [42]. There are proposed guidelines for research on resilience intervention [9], and more studies are needed on interventions targeted specifically at enhancing resilience among cancer survivors and patients with chronic pain.

Psychosocial vulnerability has long been an important factor in pain research, whereas protective factors, especially psychological resilience, have been included only in recent years. Vulnerability and protective factors work independently to some extent [28], [43], and both should therefore be included in studies on adjustment to chronic pain. Resilient functioning (assessing both recovery and sustainability) is of interest and should be addressed in future research [8]. A pain-specific resilience scale has recently been introduced in studies on experimental and chronic pain [16], [29]. It should also be tested on cancer survivors suffering from persistent pain.