J O U RN A L OF GE RI A TR IC O N CO LOG Y 5 ( 20 1 4 ) 3 3 –39
Available online at www.sciencedirect.com
ScienceDirect
The association of resilience and age in individuals with colorectal cancer: An exploratory cross-sectional study Miri Cohen a,⁎, Svetlana Baziliansky b , Alex Beny c a
Department of Gerontology, University of Haifa, Haifa, Israel Social Work Department, Rambam Health Care Campus, Haifa, Israel c Oncology Institute, Rambam Health Care Campus, Haifa, Israel b
AR TIC LE I N FO
ABS TR ACT
Article history:
Background: Studies generally report lower emotional distress in older patients with cancer
Received 6 February 2013
than in younger patients with cancer. The personality construct of resilience was previously
Received in revised form 14 June 2013
found to be higher with age, but has not been assessed in relation to emotional distress in
Accepted 30 July 2013
older patients with cancer.
Available online 10 September 2013
Objective: To assess the mediating effect of resilience on the associations between age and emotional distress in patients with colorectal cancer (CRC).
Keywords:
Patients and Methods: An exploratory cross-sectional study of 92 individuals, aged 27–87 years,
Emotional distress
diagnosed with CRC stage II–III, 1–5 years prior to enrollment in the study. They completed the
Resilience
Wagnild and Young's resilience scale and Brief Symptoms Inventory-18, cancer-related
Colorectal cancer patients
problem list, and demographic and disease-related details.
Age
Results: Older age, male gender, and less cancer-related problems were associated with
Gender
higher resilience and lower emotional distress. A Structural Equation Modeling (SEM) analysis and mediation tests showed that, while controlling for cancer-related problems, resilience mediated the effects of age and gender on emotional distress. Conclusions: The study enlarges the explanation for the consistent previous findings on the better adjustment of older patients with cancer. Increased professional support should be provided for patients with low resilience levels. © 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
Older patients with cancer (>65 years) represent more than 60% of cancer cases.1 Demographic projections of developed countries anticipate that the number of older adults will expand over the coming decades and people will live to reach greater ages.2 The anticipated expansion of the > 65 age group followed by the expansion in number of older patients with cancer requires a widening of knowledge of factors that promote well-being in older patients with cancer.1 The effect of age on emotional distress in reaction to cancer diagnosis and treatment has recently enjoyed a surge of interest.
Most studies comparing the emotional distress of older and younger patients with cancer have reported that the former were less distressed (in terms of depression and anxiety) and better adjusted than the latter.3–6 Although the subjects of most of the studies were patients with breast cancer3,4 similar results were obtained for other types of cancer, such as prostate cancer5 or lung cancer.6 The few studies of patients with CRC also found that older patients (>65) reported a better quality of life and lower emotional distress.7–10 A recent study of 1800 patients with CRC, five and twelve months post-diagnosis, found higher-level emotional distress9 and a higher rate of post-traumatic symptoms in the younger patients was found in another study.10
⁎ Corresponding author at: Faculty of Social Welfare and Health Sciences, University of Haifa, Mount Carmel, Haifa 31095, Israel. Tel./fax: + 972 4 8249565. E-mail address: [email protected] (M. Cohen). 1879-4068/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jgo.2013.07.009
34
JO U RN A L OF GE RI A TR IC O N COL O G Y 5 (2 0 1 4 ) 3 3 –3 9
A very small number of studies do not support the above findings. In a study of over 700 individuals with prostate cancer, older patients reported less distress as measured by the distress thermometer and a better quality of life, but older age proved to be associated with higher levels of depression.11 However, in this study the comparison was between groups within the range of 50–93 years of age only. Another study of patients with CRC found a negative association between age and depression in married male participants, but no association was found between age and anxiety or total emotional distress score.12 However, a possible explanation of the contrasting results of these two studies may be that they examined male patients only, or that both studies presented correlational results only. Several explanations have been proposed for the lower level of emotional distress in older patients with cancer found in the majority of studies. Scholars propose that younger patients suffer more intense stressors related to their being parents of young children, being in the middle of their careers, and bearing more intense work and economic strains.4,8 Another explanation is that age groups differ in their expectations of life and of the future. Older adults may anticipate contracting various diseases with their advancing years.13 Older individuals usually experience less death anxiety than do younger ones,14 possibly due to a sense of accomplishment of life goals or personal maturation.15 However, their lower level of distress may be due to less aggressive treatments15 or to different patterns of reporting distress.16 The construct of resilience may also account for differences between age groups in coping with and adjusting to cancer. Studies showed that resilience predicted lower-level distress, better adjustment, and better quality of life among patients with cancer.17,18 Thus, the construct of resilience is used as the theoretical basis for the present study. Resilience was conceptualized as a pattern of functioning indicative of positive adaptation in the context of significant risk or adversity.18 It is expressed in an individual's capacity to recover from severe negative and traumatic experiences through the ability to adjust to different and changing situational demands, and to recover from adverse life events.18–20 It was found that resilience moderates the negative effects of stressors and promotes adaptation and is associated with mental health.21,22 However, more recent research has shown that resilience is also a construct that may change across life.23,24 It may be promoted or impaired depending on the synergy shared between individuals and their environments and experiences.23,24 The definition of resilience includes the notion that it develops through exposure to significant risks18 and is important to psychological well being in old age. While significant risks and adversities accumulate with age, research literature presents two possible trajectories of resilience in old age. In the first trajectory, resilience may weaken in old age due to physical and cognitive decline, the cumulative effects of personal and interpersonal losses, mounting negative experiences, and the loss of personal resources.25,26 The second trajectory is one of strengthening resilience with age.27,28 Over the years, with increasing encounters with challenging or negative experiences, individuals may gain experience, and learn and enrich a repertoire of efficient coping strategies.25,28 Older individuals often develop perceptions and attitudes that promote adjustment to negative life events.25 Accordingly, a high rate of
resilience was found in individuals aged 85–95.27 Also, in a Swedish study of 1719 participants, from eight different samples, aged from 19 to 103 years,29 and a Dutch study with 3265 participant aged 17 to 65+, being older was found to be associated with higher resilience.30 Lower rates of post-traumatic symptoms were found in older individuals than in younger individuals who had undergone traumatic experiences,31,32 although a few studies have found no differences between the age groups.33,34 Several studies assessed mediation models in which resilience, or components of resilience, served as a mediator between personal or environmental changes and psychological well being. For example, resilience factors were reported to mediate the relationship between social status and self-rated health,35 perceived stress, and life satisfaction.36 In addition, emotion regulation, defined as a component of resilience, was found to mediate the associations of age and affect.37 A mediating variable is one that explains possible associations between an antecedent variable and its outcome.38 This definition suggests that an antecedent variable precedes the occurrence of the mediator, and both precede the occurrence of the outcome.39 Although this definition may imply casual relations that cannot be inferred from cross sectional study designs, scholars have often suggested assessing the putative mediating role of specific variables as based on theoretical knowledge regarding the mediating role, and thus assessing the theoretical model.40,41 Accordingly, the role of personal characteristics was often assessed within theoretical models, such as models of coping with stressors (e.g., Lazarus and Folkman),42 using cross-sectional designs. For example, optimism was often assessed as a mediator between stressors and psychological well being.40,41 Only one study approached the assumption that resilience may be a mediator between age and psychological outcomes,37 while the role of resilience in explaining differences between older and younger patients with cancer in terms of emotional distress has not yet been studied. This study, although based on exploratory cross-sectional design, attempts to address this lacuna, by assessing the mediating role of resilience between age and gender and emotional distress. We chose to focus on patients with CRC in order to obtain a homogeneous sample in terms of the type of cancer and of functional disabilities, body image, or other personal and interpersonal consequences. In addition, due to its similar rate in both sexes, our sample of patients with CRC allows the examination of gender differences. According to the conceptual model, the demographic variables, age, gender and education, and the disease-related variables, stage of disease, time since diagnosis, and cancer-related problems, were hypothesized to be associated directly with resilience and indirectly via resilience with emotional distress.
2.
Methods
2.1.
Participants and Procedure
Ninety-two participants were enrolled in this study, with an age range of 27–87 year-old individuals, diagnosed with CRC 1–5 years prior to enrollment in the study. Fifty percent of the participants were younger than 65 years and the other 50%
J O U RN A L OF GE RI A TR IC O N CO LOG Y 5 ( 20 1 4 ) 3 3 –39
were 65 and older. Inclusion criteria were stages II–III CRC and currently free of disease. All individuals who fulfilled the criteria were identified from the computerized files of the CRC follow-up clinic at the main oncology center in the north of Israel and were invited to participate. All eligible individual (112 patients) were approached at their follow-up visit to the CRC clinic; twenty refused to participate, and thus the participation rate was 82.1%. The participants were asked to complete the questionnaires during their visit to the CRC clinic, after giving their informed consent. The hospital ethics board approved the study. The background characteristics of the participants are described in Table 1. The mean age was 58, age ranged between 27 and 87 years, and the mean of education was 15 years. A little more than half of the sample were men and 60% were married or living with a partner. Most participants reported an average income or higher. Participants were on average about 2.5 years post-diagnosis, and more than 60% were diagnosed with stage II CRC. Of the list of 34 problems related to cancer and its treatments, 0–16 problems were reported by the participants, with a mean of 4.8 problems per participant.
2.2.
Questionnaires
Demographic data covered information on age, gender, education, family status, and economic status. Clinical data covered stage of CRC and time since diagnosis. The cancer-related problem list6 consists of 40 items referring to practical, familial, emotional, spiritual, and physical problems. Emotional problem items were not included due to a high correlation with the Brief Symptom Inventory (BSI)
score (r =0.80, p < 0.001). The questionnaire was back-translated into Hebrew and had been used previously.43 Participants were asked to rate each problem on a yes/no scale and the total score was calculated (possible range 0–34). Wagnild and Young's resilience scale21 is a 25-item scale consisting of two factors: acceptance of self and life (e.g., ‘I feel proud that I have accomplished things in my life,’ ‘I seldom wonder what the point of it all is’), and individual competence (e.g., ‘I can depend on myself more than on anyone else,’ ‘When I'm in a difficult situation, I can usually find my way out of it’). Participants were asked to rate the items on a 7-point scale from 1 (strongly disagree) to 7 (strongly agree). The inter-correlation between the variables in the present study was r = 0.88, thus the mean score for all items was used in the subsequent analyses. Cronbach's alphas of 0.72–0.94 were reported in previous studies (reviewed in22). The scale has been translated to Hebrew and back-translated. Internal reliability (Cronbach's alpha) in the present study was 0.93. Brief Symptom Inventory-18 (BSI-18)44 is a short version of the 53-item BSI for measuring depression, anxiety, and somatization. Participants were asked to rate their feelings during the previous seven days on a 5-point scale from 0 (not at all) to 4 (extremely). A Cronbach's alpha of 0.89 was previously reported.45 The BSI has been translated into Hebrew46 and is widely used. For the present study, the mean score of anxiety and depression sub-scales was used. The sum of ratings for the two sub-scales was calculated. The possible range of scores is 0–48. The association between the means of the anxiety and depression subscales was r = 0.79 (p < 0.001). Internal reliability (Cronbach's alpha) of the two scale score for this study was 0.90.
2.3.
Table 1 – Background characteristics of the participants. Variable Age (years), M (SD) Range Gender, N (%) Male Female Education (years), M (SD) Range Familial status, N (%) Married/cohabiting Not married/cohabiting Income a, N (%) Above average Average Below average Time since diagnosis (years), M (SD) Range Stage of CRC, N (%) Stage II Stage III Cancer-related problems, M (SD) Range
N = 92 58 (11.96) 27–87 48 (52.2) 44 (47.8) 14.89 (3.49) 8–26 60 (65.2) 32 (34.8) 41 (46.1) 41 (46.1) 7 (7.8) 2.55 (1.40) 1–5 61 (66.3) 31 (33.7) 4.79 (2.97) 0–16
Abbreviations: M, mean; SD, standard deviation; CRC, colorectal cancer. a Data missing for three respondents.
35
Data Analysis
Descriptive statistics of demographic and study variables were examined and Pearson correlations between study variables were calculated. A Structural Equation Modeling (SEM) technique was chosen to test the proposed model. The sample size needed for a structural equation model (SEM) was computed.47 Given the number of observed and latent variables in the model, there is an anticipated effect size of 0.15, statistical power of 0.85, probability level lower than 0.05, and the minimum sample size for the study model structure is 89. SEM was performed using the Analysis of Moment Structures (Amos) software program, version 7.0.48 Model fit was based on generally accepted thresholds for five indices that reflect diverse criteria of fit, and are less sensitive to sample size: (1) the Normed Chi-square, which is the Chi-square divided by the degrees of freedom (χ2/df), is used to assess the model's overall fit and parsimony, and Normed Chi-square values of ≤ 2.0 indicate good fit49; (2) The Comparative Fit Index (CFI), the Tucker–Lewis Index (TLI) and the Normed Fit Index (NFI) are incremental fit indexes; (3) the Root Mean-Square Error of Approximation (RMSEA) measures the discrepancy per degree of freedom and indicates absolute fit of the model. CFI, TLI and NFI scores of > 0.90 and RMSEA values of < 0.08 indicate acceptable indices of fit,50 while CFI, TLI, NFI of > 0.95 and RMSEA of < 0.06 indicate a good model fit.44 Mediation tests were conducted using Baron and Kenny's38 guidelines and the Sobel test.52
36
JO U RN A L OF GE RI A TR IC O N COL O G Y 5 (2 0 1 4 ) 3 3 –3 9
Table 2 – Means (SDs) of the study variables. Variables
M
SD
Actual range
Possible range
Resilience BSI a
5.41 3.26
0.74 3.50
3.67–6.67 0–16.0
1–7 0–48
diagnosis, and stage of CRC. Due to the similar patterns of associations of the total BSI and of anxiety and depression scores with the study variables, the study model was assessed with the total BSI score. The resulting study model is depicted in Fig. 1. As can be seen in the figure, the paths were significant, in line with the hypotheses. The fit indices were χ2/df = 1.4; CFI = 0.953; TLI = 0.936; NFI = 0.868; RMSEA = 0.069. Thirty-four percent of the variance of resilience was explained by the background variables and cancer-related problems. Resilience was significantly predicted by age and gender: older age and male gender were associated with higher resilience. Education and cancer-related problems were not significantly associated with resilience. Sixty-one percent of the variance of BSI was explained by the variables in the model. Resilience and number of cancer-related problems were significantly associated with BSI scores. These results suggest that resilience may mediate the effects of age and gender on emotional distress. These relationships were assessed further by mediating tests. First, BSI was regressed on age (β = 0.42, p < 0.001). This association became nonsignificant when resilience was entered into the regression (β = 0.12; p > 0.05). Sobel test showed Z = 3.65 (p < 0.001). When controlling for the cancer-related problem score, the mediation effect remained significant (β = 0.07; Z = 2.09, p < .05). Thus, resilience mediated the relation of age to BSI score. Gender was positively related to BSI (r = 0.21; p < 0.001), a relation that became non-significant when resilience was entered into the regression (β = .06; p > 0.05). The Sobel test showed Z = 4.09 (p < 0.001). When controlling for the cancerrelated problems score, the mediation effect remained significant (β = 0.04; Z = 2.01, p < 0.05). The results indicate that resilience mediated the relation of gender to the BSI score.
Abbreviations: M, mean; SD, standard deviation; BSI, Brief Symptom Inventory. a Sum of depression and anxiety sub-scales.
3.
Results
Table 2 shows the means and standard deviations (SDs) of resilience and BSI (depression and anxiety sub-scales) scores. Mean score of resilience was relatively high. Also, the mean score of BSI was low, and the accepted range of scores was much below the maximum possible score. The associations among the demographic, cancer-related and study variables are given in Table 3. A higher number of cancer-related problems were associated with higher BSI scores, and lower resilience scores, while lower resilience scores were associated with higher BSI scores. Older age was associated with higher resilience scores, lower number of cancer-related problems and lower BSI scores. Female gender was associated with more cancer-related problems, higher BSI scores, and lower resilience scores. Time since diagnosis was associated with lower BSI score only, and stage of CRC was not in significant associations with the other variables. Bonferroni's correction for multiple comparisons showed that, except for the correlations of time since diagnosis with age, cancer-related problems with gender, resilience with education, BSI with gender, education and time since diagnosis, all correlations remained significant. In addition, the associations of depression and anxiety subscales with study variables were assessed. Depression (r = −0.63) and anxiety (r = −0.62) were negatively and strongly associated with resilience (p < 0.001), age (r = −0.35 and −0.39 respectively, p < 0.001), gender (r = 0.21 and 0.23 respectively, p < 0.001), and cancer related problems (r = 0.69 and 0.59 respectively, p < 0.001). Similar to the associations of study variables with BSI total scores, depression and anxiety scores were not significantly associated with education, time since
4.
Discussion
In our study, older patients with CRC reported lower emotional distress and higher resilience. Moreover, it is the first study to show that the previously described lower emotional distress in older patients with cancer is mediated by the higher resilience found in these patients, when cancer-related problems are
Table 3 – Associations among study variables.
1. 2. 3. 4. 5. 6. 7. 8.
Age Gender a b Education (years) Time since diagnosis Stage of CRC Cancer-related problems Resilience BSI c
1
2
3
– 0.10 0.01 0.23 ⁎ 0.03 −0.42 ⁎⁎ 0.54 ⁎⁎ −0.40 ⁎⁎
– −0.19 0.02 −0.04 0.27 ⁎ −0.42 ⁎⁎ 0.25 ⁎
– 0.16 0.02 −0.19 0.26 ⁎
−0.23 ⁎
Abbreviations: CRC, colorectal cancer; BSI, Brief Symptom Inventory. a Male = 0; Female = 1. b Spearman correlation coefficients. c Sum depression and anxiety sub-scales. ⁎ p < 0.05. ⁎⁎ p < 0.001.
4
5
6
7
8
– 0.01 −0.15 0.05 0.24 ⁎
– 0.05 0.05 −0.09
– −0.70 ⁎⁎ 0.68 ⁎⁎
– −0.66 ⁎⁎
–
J O U RN A L OF GE RI A TR IC O N CO LOG Y 5 ( 20 1 4 ) 3 3 –39
controlled for. The relation of gender to emotional distress was also mediated by resilience. The present results support the large number of previous studies reporting lower emotional distress with advancing age in patients with cancer3–6; there are very few exceptions.11,12 Our results are also in line with studies conducted among patients with CRC.7,8 The only study we could find of patients with CRC that did not find associations between age and distress was conducted solely with older patients (mean age 70, SD = 6, range not specified) and only included males.12 Although the construct of resilience has recently sparked growing interest in the study of adjustment to cancer,17,18 it has not yet been assessed in relation to age. Our finding that resilience is higher with age accords with previous findings in the general population,27–30 as it may increase with the accumulating experience of coping with challenging events and circumstances through life.25,27,28 The results are also in line with studies reporting a lower rate and intensity of post-traumatic symptoms in older adults in reaction to traumatic events,31,32 rather than with those that did not support this link.33,34 However, the participants in the study were relatively well educated, and mostly enjoyed an average to high income. These circumstances may point to availability of resources that support the maintenance and development of resilience.22,24 In the case of absence or decline in personal resources, a deterioration in resilience may be evident.26 The SEM and mediation tests used in this study provide evidence to support the hypothesized theoretical model. CFI and TLI indicate a good fit, and NFI and RMSEA measures indicate an acceptable fit. Although stricter thresholds for goodness of fit are also suggested,51 relying on them may result in the rejection of acceptable models when sample sizes are small and when models are not complex.50 Therefore, this study shows that resilience mediates the association between age (and gender) and emotional distress. The mediating role of resilience remained significant even when background variables and cancer-related
problems were controlled for. In order for mediating relationships to exist, the independent variable should precede the mediator.39 Although, according to studies showing the strengthening of resilience with age,27,28 it may be assumed that the age variable precedes the variable of resilience, due to the one point study design, caution is needed when interpreting the mediational relations. In previous studies, the lower levels of distress in older patients with cancer were attributed to their various situational causes (cancer and treatment characteristics),15 developmental tasks and stressors being different from those of younger patients with cancer,4,8,15 and different cognitive perceptions of the disease and its personal meaning.13,14 The present study adds to the previous results by indicating the central role of resilience, which may be a precursor of more adaptive cognitive perceptions and coping strategies, as was previously suggested.25,28 However, cancer-related problems, which include practical, familial, and physical problems, attenuated the mediating effect. In addition, they were significantly and negatively associated with age. Therefore, it may be that older persons experience fewer problems and thus report lower distress levels, as was previously suggested4,8,15 or that they perceive the problems as less disturbing.13–15 This second possibility may be explained by their greater use of positive reappraisal, a coping strategy that was also suggested to be related to resilience.25 The present findings showed that resilience was a mediator between gender and emotional distress. Previous studies have often found higher levels of distress in female than in male patients with cancer.7,9,43 In line with this result, lower levels of resilience have been found in women in previous studies (reviewed in22). The present results support these findings, but also call attention to the mediating role of resilience on distress in female patients with cancer. The present results may have practical implications for psycho-oncologists, especially given the notion that resilience
Age 0.35* Gender1
0.14
Resilience
0.80**
BSI2
Education
Time diagnosis
0.74**
Stage
Cancerrelated problems 1
male=0; female=1;
2
37
Sum of depression and anxiety subscales
*p
Available online at www.sciencedirect.com
ScienceDirect
The association of resilience and age in individuals with colorectal cancer: An exploratory cross-sectional study Miri Cohen a,⁎, Svetlana Baziliansky b , Alex Beny c a
Department of Gerontology, University of Haifa, Haifa, Israel Social Work Department, Rambam Health Care Campus, Haifa, Israel c Oncology Institute, Rambam Health Care Campus, Haifa, Israel b
AR TIC LE I N FO
ABS TR ACT
Article history:
Background: Studies generally report lower emotional distress in older patients with cancer
Received 6 February 2013
than in younger patients with cancer. The personality construct of resilience was previously
Received in revised form 14 June 2013
found to be higher with age, but has not been assessed in relation to emotional distress in
Accepted 30 July 2013
older patients with cancer.
Available online 10 September 2013
Objective: To assess the mediating effect of resilience on the associations between age and emotional distress in patients with colorectal cancer (CRC).
Keywords:
Patients and Methods: An exploratory cross-sectional study of 92 individuals, aged 27–87 years,
Emotional distress
diagnosed with CRC stage II–III, 1–5 years prior to enrollment in the study. They completed the
Resilience
Wagnild and Young's resilience scale and Brief Symptoms Inventory-18, cancer-related
Colorectal cancer patients
problem list, and demographic and disease-related details.
Age
Results: Older age, male gender, and less cancer-related problems were associated with
Gender
higher resilience and lower emotional distress. A Structural Equation Modeling (SEM) analysis and mediation tests showed that, while controlling for cancer-related problems, resilience mediated the effects of age and gender on emotional distress. Conclusions: The study enlarges the explanation for the consistent previous findings on the better adjustment of older patients with cancer. Increased professional support should be provided for patients with low resilience levels. © 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
Older patients with cancer (>65 years) represent more than 60% of cancer cases.1 Demographic projections of developed countries anticipate that the number of older adults will expand over the coming decades and people will live to reach greater ages.2 The anticipated expansion of the > 65 age group followed by the expansion in number of older patients with cancer requires a widening of knowledge of factors that promote well-being in older patients with cancer.1 The effect of age on emotional distress in reaction to cancer diagnosis and treatment has recently enjoyed a surge of interest.
Most studies comparing the emotional distress of older and younger patients with cancer have reported that the former were less distressed (in terms of depression and anxiety) and better adjusted than the latter.3–6 Although the subjects of most of the studies were patients with breast cancer3,4 similar results were obtained for other types of cancer, such as prostate cancer5 or lung cancer.6 The few studies of patients with CRC also found that older patients (>65) reported a better quality of life and lower emotional distress.7–10 A recent study of 1800 patients with CRC, five and twelve months post-diagnosis, found higher-level emotional distress9 and a higher rate of post-traumatic symptoms in the younger patients was found in another study.10
⁎ Corresponding author at: Faculty of Social Welfare and Health Sciences, University of Haifa, Mount Carmel, Haifa 31095, Israel. Tel./fax: + 972 4 8249565. E-mail address: [email protected] (M. Cohen). 1879-4068/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jgo.2013.07.009
34
JO U RN A L OF GE RI A TR IC O N COL O G Y 5 (2 0 1 4 ) 3 3 –3 9
A very small number of studies do not support the above findings. In a study of over 700 individuals with prostate cancer, older patients reported less distress as measured by the distress thermometer and a better quality of life, but older age proved to be associated with higher levels of depression.11 However, in this study the comparison was between groups within the range of 50–93 years of age only. Another study of patients with CRC found a negative association between age and depression in married male participants, but no association was found between age and anxiety or total emotional distress score.12 However, a possible explanation of the contrasting results of these two studies may be that they examined male patients only, or that both studies presented correlational results only. Several explanations have been proposed for the lower level of emotional distress in older patients with cancer found in the majority of studies. Scholars propose that younger patients suffer more intense stressors related to their being parents of young children, being in the middle of their careers, and bearing more intense work and economic strains.4,8 Another explanation is that age groups differ in their expectations of life and of the future. Older adults may anticipate contracting various diseases with their advancing years.13 Older individuals usually experience less death anxiety than do younger ones,14 possibly due to a sense of accomplishment of life goals or personal maturation.15 However, their lower level of distress may be due to less aggressive treatments15 or to different patterns of reporting distress.16 The construct of resilience may also account for differences between age groups in coping with and adjusting to cancer. Studies showed that resilience predicted lower-level distress, better adjustment, and better quality of life among patients with cancer.17,18 Thus, the construct of resilience is used as the theoretical basis for the present study. Resilience was conceptualized as a pattern of functioning indicative of positive adaptation in the context of significant risk or adversity.18 It is expressed in an individual's capacity to recover from severe negative and traumatic experiences through the ability to adjust to different and changing situational demands, and to recover from adverse life events.18–20 It was found that resilience moderates the negative effects of stressors and promotes adaptation and is associated with mental health.21,22 However, more recent research has shown that resilience is also a construct that may change across life.23,24 It may be promoted or impaired depending on the synergy shared between individuals and their environments and experiences.23,24 The definition of resilience includes the notion that it develops through exposure to significant risks18 and is important to psychological well being in old age. While significant risks and adversities accumulate with age, research literature presents two possible trajectories of resilience in old age. In the first trajectory, resilience may weaken in old age due to physical and cognitive decline, the cumulative effects of personal and interpersonal losses, mounting negative experiences, and the loss of personal resources.25,26 The second trajectory is one of strengthening resilience with age.27,28 Over the years, with increasing encounters with challenging or negative experiences, individuals may gain experience, and learn and enrich a repertoire of efficient coping strategies.25,28 Older individuals often develop perceptions and attitudes that promote adjustment to negative life events.25 Accordingly, a high rate of
resilience was found in individuals aged 85–95.27 Also, in a Swedish study of 1719 participants, from eight different samples, aged from 19 to 103 years,29 and a Dutch study with 3265 participant aged 17 to 65+, being older was found to be associated with higher resilience.30 Lower rates of post-traumatic symptoms were found in older individuals than in younger individuals who had undergone traumatic experiences,31,32 although a few studies have found no differences between the age groups.33,34 Several studies assessed mediation models in which resilience, or components of resilience, served as a mediator between personal or environmental changes and psychological well being. For example, resilience factors were reported to mediate the relationship between social status and self-rated health,35 perceived stress, and life satisfaction.36 In addition, emotion regulation, defined as a component of resilience, was found to mediate the associations of age and affect.37 A mediating variable is one that explains possible associations between an antecedent variable and its outcome.38 This definition suggests that an antecedent variable precedes the occurrence of the mediator, and both precede the occurrence of the outcome.39 Although this definition may imply casual relations that cannot be inferred from cross sectional study designs, scholars have often suggested assessing the putative mediating role of specific variables as based on theoretical knowledge regarding the mediating role, and thus assessing the theoretical model.40,41 Accordingly, the role of personal characteristics was often assessed within theoretical models, such as models of coping with stressors (e.g., Lazarus and Folkman),42 using cross-sectional designs. For example, optimism was often assessed as a mediator between stressors and psychological well being.40,41 Only one study approached the assumption that resilience may be a mediator between age and psychological outcomes,37 while the role of resilience in explaining differences between older and younger patients with cancer in terms of emotional distress has not yet been studied. This study, although based on exploratory cross-sectional design, attempts to address this lacuna, by assessing the mediating role of resilience between age and gender and emotional distress. We chose to focus on patients with CRC in order to obtain a homogeneous sample in terms of the type of cancer and of functional disabilities, body image, or other personal and interpersonal consequences. In addition, due to its similar rate in both sexes, our sample of patients with CRC allows the examination of gender differences. According to the conceptual model, the demographic variables, age, gender and education, and the disease-related variables, stage of disease, time since diagnosis, and cancer-related problems, were hypothesized to be associated directly with resilience and indirectly via resilience with emotional distress.
2.
Methods
2.1.
Participants and Procedure
Ninety-two participants were enrolled in this study, with an age range of 27–87 year-old individuals, diagnosed with CRC 1–5 years prior to enrollment in the study. Fifty percent of the participants were younger than 65 years and the other 50%
J O U RN A L OF GE RI A TR IC O N CO LOG Y 5 ( 20 1 4 ) 3 3 –39
were 65 and older. Inclusion criteria were stages II–III CRC and currently free of disease. All individuals who fulfilled the criteria were identified from the computerized files of the CRC follow-up clinic at the main oncology center in the north of Israel and were invited to participate. All eligible individual (112 patients) were approached at their follow-up visit to the CRC clinic; twenty refused to participate, and thus the participation rate was 82.1%. The participants were asked to complete the questionnaires during their visit to the CRC clinic, after giving their informed consent. The hospital ethics board approved the study. The background characteristics of the participants are described in Table 1. The mean age was 58, age ranged between 27 and 87 years, and the mean of education was 15 years. A little more than half of the sample were men and 60% were married or living with a partner. Most participants reported an average income or higher. Participants were on average about 2.5 years post-diagnosis, and more than 60% were diagnosed with stage II CRC. Of the list of 34 problems related to cancer and its treatments, 0–16 problems were reported by the participants, with a mean of 4.8 problems per participant.
2.2.
Questionnaires
Demographic data covered information on age, gender, education, family status, and economic status. Clinical data covered stage of CRC and time since diagnosis. The cancer-related problem list6 consists of 40 items referring to practical, familial, emotional, spiritual, and physical problems. Emotional problem items were not included due to a high correlation with the Brief Symptom Inventory (BSI)
score (r =0.80, p < 0.001). The questionnaire was back-translated into Hebrew and had been used previously.43 Participants were asked to rate each problem on a yes/no scale and the total score was calculated (possible range 0–34). Wagnild and Young's resilience scale21 is a 25-item scale consisting of two factors: acceptance of self and life (e.g., ‘I feel proud that I have accomplished things in my life,’ ‘I seldom wonder what the point of it all is’), and individual competence (e.g., ‘I can depend on myself more than on anyone else,’ ‘When I'm in a difficult situation, I can usually find my way out of it’). Participants were asked to rate the items on a 7-point scale from 1 (strongly disagree) to 7 (strongly agree). The inter-correlation between the variables in the present study was r = 0.88, thus the mean score for all items was used in the subsequent analyses. Cronbach's alphas of 0.72–0.94 were reported in previous studies (reviewed in22). The scale has been translated to Hebrew and back-translated. Internal reliability (Cronbach's alpha) in the present study was 0.93. Brief Symptom Inventory-18 (BSI-18)44 is a short version of the 53-item BSI for measuring depression, anxiety, and somatization. Participants were asked to rate their feelings during the previous seven days on a 5-point scale from 0 (not at all) to 4 (extremely). A Cronbach's alpha of 0.89 was previously reported.45 The BSI has been translated into Hebrew46 and is widely used. For the present study, the mean score of anxiety and depression sub-scales was used. The sum of ratings for the two sub-scales was calculated. The possible range of scores is 0–48. The association between the means of the anxiety and depression subscales was r = 0.79 (p < 0.001). Internal reliability (Cronbach's alpha) of the two scale score for this study was 0.90.
2.3.
Table 1 – Background characteristics of the participants. Variable Age (years), M (SD) Range Gender, N (%) Male Female Education (years), M (SD) Range Familial status, N (%) Married/cohabiting Not married/cohabiting Income a, N (%) Above average Average Below average Time since diagnosis (years), M (SD) Range Stage of CRC, N (%) Stage II Stage III Cancer-related problems, M (SD) Range
N = 92 58 (11.96) 27–87 48 (52.2) 44 (47.8) 14.89 (3.49) 8–26 60 (65.2) 32 (34.8) 41 (46.1) 41 (46.1) 7 (7.8) 2.55 (1.40) 1–5 61 (66.3) 31 (33.7) 4.79 (2.97) 0–16
Abbreviations: M, mean; SD, standard deviation; CRC, colorectal cancer. a Data missing for three respondents.
35
Data Analysis
Descriptive statistics of demographic and study variables were examined and Pearson correlations between study variables were calculated. A Structural Equation Modeling (SEM) technique was chosen to test the proposed model. The sample size needed for a structural equation model (SEM) was computed.47 Given the number of observed and latent variables in the model, there is an anticipated effect size of 0.15, statistical power of 0.85, probability level lower than 0.05, and the minimum sample size for the study model structure is 89. SEM was performed using the Analysis of Moment Structures (Amos) software program, version 7.0.48 Model fit was based on generally accepted thresholds for five indices that reflect diverse criteria of fit, and are less sensitive to sample size: (1) the Normed Chi-square, which is the Chi-square divided by the degrees of freedom (χ2/df), is used to assess the model's overall fit and parsimony, and Normed Chi-square values of ≤ 2.0 indicate good fit49; (2) The Comparative Fit Index (CFI), the Tucker–Lewis Index (TLI) and the Normed Fit Index (NFI) are incremental fit indexes; (3) the Root Mean-Square Error of Approximation (RMSEA) measures the discrepancy per degree of freedom and indicates absolute fit of the model. CFI, TLI and NFI scores of > 0.90 and RMSEA values of < 0.08 indicate acceptable indices of fit,50 while CFI, TLI, NFI of > 0.95 and RMSEA of < 0.06 indicate a good model fit.44 Mediation tests were conducted using Baron and Kenny's38 guidelines and the Sobel test.52
36
JO U RN A L OF GE RI A TR IC O N COL O G Y 5 (2 0 1 4 ) 3 3 –3 9
Table 2 – Means (SDs) of the study variables. Variables
M
SD
Actual range
Possible range
Resilience BSI a
5.41 3.26
0.74 3.50
3.67–6.67 0–16.0
1–7 0–48
diagnosis, and stage of CRC. Due to the similar patterns of associations of the total BSI and of anxiety and depression scores with the study variables, the study model was assessed with the total BSI score. The resulting study model is depicted in Fig. 1. As can be seen in the figure, the paths were significant, in line with the hypotheses. The fit indices were χ2/df = 1.4; CFI = 0.953; TLI = 0.936; NFI = 0.868; RMSEA = 0.069. Thirty-four percent of the variance of resilience was explained by the background variables and cancer-related problems. Resilience was significantly predicted by age and gender: older age and male gender were associated with higher resilience. Education and cancer-related problems were not significantly associated with resilience. Sixty-one percent of the variance of BSI was explained by the variables in the model. Resilience and number of cancer-related problems were significantly associated with BSI scores. These results suggest that resilience may mediate the effects of age and gender on emotional distress. These relationships were assessed further by mediating tests. First, BSI was regressed on age (β = 0.42, p < 0.001). This association became nonsignificant when resilience was entered into the regression (β = 0.12; p > 0.05). Sobel test showed Z = 3.65 (p < 0.001). When controlling for the cancer-related problem score, the mediation effect remained significant (β = 0.07; Z = 2.09, p < .05). Thus, resilience mediated the relation of age to BSI score. Gender was positively related to BSI (r = 0.21; p < 0.001), a relation that became non-significant when resilience was entered into the regression (β = .06; p > 0.05). The Sobel test showed Z = 4.09 (p < 0.001). When controlling for the cancerrelated problems score, the mediation effect remained significant (β = 0.04; Z = 2.01, p < 0.05). The results indicate that resilience mediated the relation of gender to the BSI score.
Abbreviations: M, mean; SD, standard deviation; BSI, Brief Symptom Inventory. a Sum of depression and anxiety sub-scales.
3.
Results
Table 2 shows the means and standard deviations (SDs) of resilience and BSI (depression and anxiety sub-scales) scores. Mean score of resilience was relatively high. Also, the mean score of BSI was low, and the accepted range of scores was much below the maximum possible score. The associations among the demographic, cancer-related and study variables are given in Table 3. A higher number of cancer-related problems were associated with higher BSI scores, and lower resilience scores, while lower resilience scores were associated with higher BSI scores. Older age was associated with higher resilience scores, lower number of cancer-related problems and lower BSI scores. Female gender was associated with more cancer-related problems, higher BSI scores, and lower resilience scores. Time since diagnosis was associated with lower BSI score only, and stage of CRC was not in significant associations with the other variables. Bonferroni's correction for multiple comparisons showed that, except for the correlations of time since diagnosis with age, cancer-related problems with gender, resilience with education, BSI with gender, education and time since diagnosis, all correlations remained significant. In addition, the associations of depression and anxiety subscales with study variables were assessed. Depression (r = −0.63) and anxiety (r = −0.62) were negatively and strongly associated with resilience (p < 0.001), age (r = −0.35 and −0.39 respectively, p < 0.001), gender (r = 0.21 and 0.23 respectively, p < 0.001), and cancer related problems (r = 0.69 and 0.59 respectively, p < 0.001). Similar to the associations of study variables with BSI total scores, depression and anxiety scores were not significantly associated with education, time since
4.
Discussion
In our study, older patients with CRC reported lower emotional distress and higher resilience. Moreover, it is the first study to show that the previously described lower emotional distress in older patients with cancer is mediated by the higher resilience found in these patients, when cancer-related problems are
Table 3 – Associations among study variables.
1. 2. 3. 4. 5. 6. 7. 8.
Age Gender a b Education (years) Time since diagnosis Stage of CRC Cancer-related problems Resilience BSI c
1
2
3
– 0.10 0.01 0.23 ⁎ 0.03 −0.42 ⁎⁎ 0.54 ⁎⁎ −0.40 ⁎⁎
– −0.19 0.02 −0.04 0.27 ⁎ −0.42 ⁎⁎ 0.25 ⁎
– 0.16 0.02 −0.19 0.26 ⁎
−0.23 ⁎
Abbreviations: CRC, colorectal cancer; BSI, Brief Symptom Inventory. a Male = 0; Female = 1. b Spearman correlation coefficients. c Sum depression and anxiety sub-scales. ⁎ p < 0.05. ⁎⁎ p < 0.001.
4
5
6
7
8
– 0.01 −0.15 0.05 0.24 ⁎
– 0.05 0.05 −0.09
– −0.70 ⁎⁎ 0.68 ⁎⁎
– −0.66 ⁎⁎
–
J O U RN A L OF GE RI A TR IC O N CO LOG Y 5 ( 20 1 4 ) 3 3 –39
controlled for. The relation of gender to emotional distress was also mediated by resilience. The present results support the large number of previous studies reporting lower emotional distress with advancing age in patients with cancer3–6; there are very few exceptions.11,12 Our results are also in line with studies conducted among patients with CRC.7,8 The only study we could find of patients with CRC that did not find associations between age and distress was conducted solely with older patients (mean age 70, SD = 6, range not specified) and only included males.12 Although the construct of resilience has recently sparked growing interest in the study of adjustment to cancer,17,18 it has not yet been assessed in relation to age. Our finding that resilience is higher with age accords with previous findings in the general population,27–30 as it may increase with the accumulating experience of coping with challenging events and circumstances through life.25,27,28 The results are also in line with studies reporting a lower rate and intensity of post-traumatic symptoms in older adults in reaction to traumatic events,31,32 rather than with those that did not support this link.33,34 However, the participants in the study were relatively well educated, and mostly enjoyed an average to high income. These circumstances may point to availability of resources that support the maintenance and development of resilience.22,24 In the case of absence or decline in personal resources, a deterioration in resilience may be evident.26 The SEM and mediation tests used in this study provide evidence to support the hypothesized theoretical model. CFI and TLI indicate a good fit, and NFI and RMSEA measures indicate an acceptable fit. Although stricter thresholds for goodness of fit are also suggested,51 relying on them may result in the rejection of acceptable models when sample sizes are small and when models are not complex.50 Therefore, this study shows that resilience mediates the association between age (and gender) and emotional distress. The mediating role of resilience remained significant even when background variables and cancer-related
problems were controlled for. In order for mediating relationships to exist, the independent variable should precede the mediator.39 Although, according to studies showing the strengthening of resilience with age,27,28 it may be assumed that the age variable precedes the variable of resilience, due to the one point study design, caution is needed when interpreting the mediational relations. In previous studies, the lower levels of distress in older patients with cancer were attributed to their various situational causes (cancer and treatment characteristics),15 developmental tasks and stressors being different from those of younger patients with cancer,4,8,15 and different cognitive perceptions of the disease and its personal meaning.13,14 The present study adds to the previous results by indicating the central role of resilience, which may be a precursor of more adaptive cognitive perceptions and coping strategies, as was previously suggested.25,28 However, cancer-related problems, which include practical, familial, and physical problems, attenuated the mediating effect. In addition, they were significantly and negatively associated with age. Therefore, it may be that older persons experience fewer problems and thus report lower distress levels, as was previously suggested4,8,15 or that they perceive the problems as less disturbing.13–15 This second possibility may be explained by their greater use of positive reappraisal, a coping strategy that was also suggested to be related to resilience.25 The present findings showed that resilience was a mediator between gender and emotional distress. Previous studies have often found higher levels of distress in female than in male patients with cancer.7,9,43 In line with this result, lower levels of resilience have been found in women in previous studies (reviewed in22). The present results support these findings, but also call attention to the mediating role of resilience on distress in female patients with cancer. The present results may have practical implications for psycho-oncologists, especially given the notion that resilience
Age 0.35* Gender1
0.14
Resilience
0.80**
BSI2
Education
Time diagnosis
0.74**
Stage
Cancerrelated problems 1
male=0; female=1;
2
37
Sum of depression and anxiety subscales
*p
