Coping in Patients with Incurable Lung and Gastrointestinal Cancers: A Validation Study of the Brief COPE

Abstract

Context

Patients with incurable cancer engage in several coping styles to manage the impact of cancer and its treatment. The Brief COPE is a widely used measure intended to capture multiple, distinct types of coping. The Brief COPE has not been validated among patients with incurable cancer.

Objectives

We sought to validate seven subscales of the Brief COPE in a large sample of patients newly diagnosed with incurable lung and non-colorectal gastrointestinal cancers (N=350).

Methods

Participants completed the Brief COPE and measures assessing quality of life (Functional Assessment of Cancer Therapy – General) and psychological distress (Hospital Anxiety and Depression Scale) within eight weeks of diagnosis of incurable cancer. We evaluated the psychometric properties of the Brief COPE using a confirmatory factor analysis and tests of correlation with the quality of life and distress scales.

Results

The Brief COPE factors were consistent with the original subscales, though the behavioral disengagement scale had low internal consistency. Factors showed anticipated relationships with quality of life and distress measures, except emotional support coping, which was correlated with increased depression and anxiety. We also conducted an exploratory higher-order factor analysis to determine if subscales’ score variances grouped together. The higher-order factor analysis resulted in two factors, with active, emotional support, positive reframing, and acceptance loading onto one factor and denial and self-blame loading onto the second.

Conclusion

The selected subscales of the Brief COPE are appropriate measures of coping among individuals newly diagnosed with incurable lung and gastrointestinal cancers.

Introduction

Individuals diagnosed with incurable cancer face a life-threatening stressor that elicits various coping responses. Lazarus and Folkman (1) define coping as an individual’s “constantly changing cognitive and behavioral efforts to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person.” Before implementing coping strategies, individuals cognitively appraise the nature of the stressor and their abilities and/or resources to cope with the threat (2).

Research on the assessment of coping styles among patients with incurable cancer is sparse. Nonetheless, understanding coping in this population is clinically meaningful given that these patients endure considerable stress, experience high symptom burden, and encounter difficult decisions regarding treatment at the end of life (3–5). In a large study of coping in patients with cancer, investigators used factor analysis to identify five coping patterns: seeking or using social support, focusing on the positive, distancing, cognitive escape-avoidance, and behavioral escape-avoidance (6). However, because this study included patients with mixed cancer types and severity, the results may not be applicable to patients with incurable cancer. Patients with incurable cancer may cope differently than those with early-stage cancer or other non-terminal illnesses given the life-threatening and often rapidly progressive nature of the disease. Advancing the assessment and measurement of coping in patients with cancer may improve our ability to identify those most vulnerable and enhance adaptive coping strategies in this population.

The Brief COPE is an abbreviated version of the COPE measure, an inventory of various types of coping styles (7). Based on Lazarus and Folkman’s model of coping along with behavioral self-regulation, the scale is intended to assess individuals’ various coping strategies in an efficient manner. As one of the most frequently employed measures to assess coping, the Brief COPE was included in the Coping with Cancer study, a large, multi-site investigation of patients with advanced cancer (8) as well as in other studies of cancer populations (9–11), providing empirical support for use of the measure in this population. However, to the best of our knowledge, no investigators to date have assessed the psychometric properties of the Brief COPE in a cohort of patients with incurable cancer, especially lung and gastrointestinal cancers which have a poor prognosis and high symptom profile (12, 13). Because of the broad use of the Brief COPE, we wanted to ensure the measure accurately and consistently captures the construct of coping within this unique population and decided to test its psychometric properties.

The purpose of this psychometric study was to evaluate the validity of the Brief COPE in a sample of patients with incurable lung and gastrointestinal cancers. We tested two primary hypotheses to evaluate the validity of the scale: (a) a confirmatory factor analysis of the Brief COPE would result in subscales congruent with the original subscales of the Brief COPE and (b) associations between the Brief COPE and established measures of quality of life (QOL) and psychological distress in our sample would reflect the relationship of these measures previously established in other cancer studies.

Methods

Research Sample

As part of a randomized clinical trial of early palliative care integrated with standard oncology care, we enrolled patients within eight weeks of diagnosis of incurable lung or non-colorectal gastrointestinal (GI) cancers who received care at a large academic medical center. Patients were considered to have incurable cancer based upon having advanced disease for which no curative treatment options existed. Patients were eligible if they had been previously diagnosed with an earlier stage of disease. We selected this patient population because their baseline data would reflect their coping needs during this particularly stressful period (14) along with their psychological needs which are known to be largely unmet (15). Inclusion criteria included: (a) age ≥ 18 years; (b) no prior therapy for metastatic disease; (c) an Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 2; (d) the ability to read and respond to questions in English or be able to complete patient-reported measures with minimal assistance; and (e) plan on receiving care from Massachusetts General Hospital Cancer Center. We excluded patients who had already received palliative care services, required an immediate referral to palliative care or hospice, or had significant co-morbidities or psychiatric condition that prohibited consent or participation. The Dana-Farber/Harvard Cancer Center Institutional Review Board approved this study prior to initiation.

Study Procedures

Study staff screened patients who presented to participating oncology clinics for eligibility by examining clinical documentation in the electronic health record and consent forms for chemotherapy or clinical trials. If eligible, the study staff member sent an email to the treating clinician notifying him or her that the patient was eligible to participate and also placed a reminder about the study on the patient’s paper clinic chart at the time of visit. After being introduced to the study by their medical oncologist, all participants provided written informed consent prior to randomization. Cancer Center protocol staff, who were separate from the study staff, administered baseline questionnaires to participants immediately following informed consent and prior to randomization. Participants completed questionnaires either on paper or via a web-based, HIPAA-compliant survey tool using a tablet PC.

Measurements

Coping Styles

Participants originally completed all 14 Brief COPE coping style subscales, with two survey items assessing each coping strategy. We selected this measure because it was one of the most commonly used measures of coping within this patient population. After repeated feedback from patients regarding the length and redundancy of the measure, we decided to administer only seven coping strategies that we considered highly relevant to patients with incurable cancer: emotional support, positive reframing, active, acceptance, self-blame, denial, and behavioral disengagement. The developer of the instrument encouraged such selective and flexible use of subscales (7). Participants selected the degree to which they had used each type of coping (1 = “not at all,” 2 = “a little bit,” 3 = “a medium amount,” and 4 = “a lot”).

Quality of Life

We measured QOL with the Functional Assessment of Cancer Therapy – General (FACT-G) (16). This valid, reliable scale consists of 28 items across four QOL domains including physical, functional, emotional, and social well-being. We utilized total FACT-G scores for this analysis. The response format consists of a 5-point Likert scale in which patients rate their QOL in the past week. Scores on the FACT-G range from 0 to 112, with higher scores indicating a better QOL.

Psychological Distress

We measured psychological distress (anxiety and depression) in the Hospital Anxiety and Depression Scale (HADS) (17). The HADS is widely used among patients with cancer. It consists of two 7-item subscales (14 items total) with 4-point Likert-type response options assessing depression and anxiety symptoms during the past week. Scores on each subscale range from 0 to 21, with higher total and subscale scores indicating worse psychological distress.

Analytic Approach

We first examined assumptions for conducting a factor analysis including multivariate normality, collinearity between variables as evidenced by the correlation matrix and anti-image correlation matrix, and homoscedasticity by examining bivariate scatterplots. After identifying high levels of skewness (± 4) on several Brief COPE items related to floor and ceiling effects of item responses, we used logarithmic transformations to correct these variables in order to proceed with the factor analysis.

To test our first hypothesis that a confirmatory factor analysis of the Brief COPE in our patient sample would result in subscales similar to the developer’s (i.e., Carver) original measure, we performed an exploratory factor analysis with an unforced number of factors. We used Varimax rotation and Maximum Likelihood extraction to uncover the underlying dimensions of the Brief COPE. We examined scree plots, eigenvalues, factor loadings on the pattern matrix, Bartlett’s test of sphericity for an identity matrix, the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy, and Cronbach’s alpha tests of internal consistency to arrive at factor structures. Our goal was to find factor structures that corresponded theoretically to Carver’s conceptualization of the Brief COPE and maintained clear factor loadings, high percentage of variance explained, and high communalities. We examined multiple iterations of the exploratory analysis to compare different modes of factor extraction, rotation, and factor loadings to ensure our results remained strong across varying types of analysis. Based on the final factor structure, we then examined higher order factors among the Brief COPE subscales to determine if underlying latent factors explained the variance in participant scores across these subscales.

We examined our second hypothesis by evaluating supposed relationships between participants’ scores on the Brief COPE and their scores on the FACT-G and HADS. We chose these measures because (a) strong evidence exists in the literature describing the relationships between each construct and coping from which we could develop hypotheses and (b) we collected both measures as a part of the parent study. When comparing the Brief COPE to the FACT-G, we hypothesized that the positive reframing subscale (e.g., adjusting one’s appraisal of a stressor to be less catastrophic or threatening) would be positively associated with the FACT-G total score. We based this hypothesis on research suggesting that individuals who appraise their cancer as more threatening have lower well-being and greater psychological distress (9, 18). Because of the extensive literature regarding positive reframing, we selected this analysis to provide the strongest psychometric evidence for the Brief COPE. For descriptive purposes, we reported additional associations between other Brief COPE subscales and QOL.

When comparing the Brief COPE to the HADS Depression and Anxiety subscales, we hypothesized that (a) active and engaged types of coping would be negatively associated with depression and anxiety and (b) passive and disengaged types of coping would be positively associated with depression and anxiety. We considered active and engaged types of coping to include the eight subscale items pertaining to active, positive reframing, acceptance, and emotional support coping whereas passive and disengaged types of coping included the six subscale items pertaining to denial, self-blame, and behavioral disengagement. While Carver cautioned against creating false dichotomies of the Brief COPE subscales, we examined these classifications for the purpose of confirming the Brief COPE’s psychometric properties in this population, not to redefine the original measure’s conceptual underpinnings. We tested these hypotheses using Pearson correlations between scale scores with a two-tailed significance level set at α = 0.05.

Results

Sample

Between May 2011 and July 2015, we enrolled in the trial 350 patients with incurable cancer who completed baseline measures, which are included in this analysis. The rate of eligible participants who enrolled in the study was 72.9%. Table 1 shows the sample demographics and clinical characteristics. Approximately half the sample was male (54.0%), and the mean age was 64.86 years old (SD = 10.86). Participants were primarily White (92.3%) and had an ECOG performance score of a 0 or 1 (91.1%). The most common diagnoses were non-small cell lung cancer (NSCLC) (44.0%) and pancreatic cancer (24.9%). The mean number of days between being informed of advanced disease and enrolling in the study was 25.0 (SD = 14.1).

Table 1.

Sample Characteristics

	N (%) N = 350
Age (M ± SD)	64.86 ± 10.86
Male gender	189 (54.0)
White race	323 (92.3)
Marital status
Married or living with significant other	244 (69.7)
Divorced or separated	35 (10.0)
Widowed or loss of long-term partner	35 (10.0)
Single or never married	27 (7.7)
Non-cohabitating relationship	6 (1.7)
Other	3 (0.9)
Education
High school degree or less	131 (37.5)
Some college or technical school	79 (22.6)
College graduate	66 (18.9)
Master’s degree or more	74 (21.2)
Annual household income
< $25,000	55 (17.1)
$25–50,000	78 (24.2)
$51–100,000	92 (28.6)
$101 – 150,000	56 (17.4)
> $150,000	41 (12.7)
Cancer type
Non-small cell lung cancer	154 (44.0)
Pancreatic	87 (24.9)
Small cell lung cancer	30 (8.6)
Esophageal	24 (6.9)
Other	55 (15.7)
Days between notice of advanced disease and study enrollment (M ± SD)	25.0 (14.1)
Performance status at baseline
0 (fully active)	95 (27.1)
1 (ambulatory but no strenuous   functioning)	224 (64.0)
2 (ambulatory but no work activities)	31 (8.9)
Measurements	M ± SD
Brief COPE – Acceptance (range 2–8)	6.59 ± 1.64
Brief COPE – Active (range 2–8)	6.79 ± 1.48
Brief COPE – Behavioral Disengagement   (range 2–8)	2.33 ± 0.81
Brief COPE – Denial (range 2–8)	3.12 ± 1.51
Brief COPE – Emotional Support (range   2–8)	7.57 ± 0.94
Brief COPE – Positive Reframing (range   2–8)	5.47 ± 1.85
Brief COPE – Self-blame (range 2–8)	2.80 ± 1.29
FACT-G (range 0–112)	78.31 ± 15.17
HADS – Anxiety (range 0–21)	5.31 ± 3.92
HADS – Depression (range 0–21)	4.65 ± 4.01

Construct Validity - Factor Analysis

The KMO statistic of 0.72 represented a moderate sampling adequacy, and significant Bartlett’s test (p ≤ 0.001) indicated that factor correlations do not reflect an identity matrix. Therefore, factor analysis testing can be conducted. The scree plot demonstrated that two or three factors explained the majority of the item variance with several additional factors present. As we hypothesized in evaluating the psychometric validity of the scale, the factor analysis resulted in rotated factor loadings consistent with the original Brief COPE factors (Table 2). These seven factors explained 60.9% of the variance in item responses, and many subscales were significantly correlated with each other (Table 3). When we compared this 7-factor model to alternate models using Promax rotation, principal axis factoring, and fewer factors, we decided to retain the original 7-factor model given its superior conceptual congruency and strong statistical results. Notably, the emotional support subscale was negatively associated with more approach-oriented types of coping (e.g., acceptance, positive reframing, and active) and positively associated with more avoidance types of coping (e.g. self-blame and behavioral disengagement).

Table 2.

Factor Analysis Results for Brief COPE

Subscale/Item	Factora
	I	II	III	IV	V	VI	VII
Self-blame (α = 0.733) (I)
I’ve been criticizing     myself.	.984	−.043	.125	.027	−.035	.040	.095
I’ve been blaming     myself for things that     happened.	.572	−.052	.149	.123	−.007	−.016	.073
Acceptance (α = 0.814) (II)
I’ve been learning to live     with it.	−.092	.949	−.126	−.126	.233	.058	−.035
I’ve been accepting the     reality of the fact that     it has happened.	−.084	.586	−.252	−.172	.168	.192	−.118
Denial (α = 0.724) (III)
I’ve been saying to     myself “this isn’t     real.”	.151	−.124	.978	.019	−.039	.038	−.013
I’ve been refusing to     believe that it has     happened.	.203	−.198	.553	.065	.041	−.069	.238
Emotional Support (α = 0.721) (IV)
I’ve been getting     emotional support     from others.	.072	−.116	−.002	.889	−.112	−.074	.080
I’ve been getting     comfort and     understanding from     someone.	.108	−.098	.070	.601	−.092	−.176	.010
Positive Reframing (α = 0.573) (V)
I’ve been looking for     something good in     what is happening.	−.074	.222	−.042	−.072	.658	.062	.014
I’ve been trying to see it     in a different light, to     make it seem more     positive.	−.009	.052	.038	−.083	.609	.133	−.127
Active (α = 0.667) (VI)
I’ve been concentrating     my efforts on doing     something about the     situation I’m in.	−.021	.109	.001	−.164	.154	.791	−.076
I’ve been taking action     to try to make the     situation better.	.015	.152	−.051	−.292	.401	.442	−.097
Behavioral   Disengagement (α =   0.223) (VII)
I’ve been giving up     trying to deal with it.	.121	−.038	.117	.058	−.081	−.054	.801
I’ve been giving up the     attempt to cope.	.242	−.104	.027	.032	−.108	−.101	.241

^aFactor loadings are bolded according to the factor on which they most strongly loaded.

Table 3.

Brief COPE Factor Correlations

	Factor
	I	II	III	IV	V	VI	VII
Self-blame (I)	1.00
Acceptance (II)	−0.17a	1.00
Denial (III)	0.31a	−0.36a	1.00
Emotional Support (IV)	0.19a	−0.28a	0.10	1.00
Positive Reframing (V)	−0.07	0.33a	−0.04	−0.19a	1.00
Active (VI)	−0.05	0.37a	−0.06	−0.39a	0.38a	1.00
Behavioral Disengagement (VII)	0.29a	−0.20a	0.19	0.15a	−0.14a	−0.20a	1.00

^ap <0.01

The internal consistency was acceptable at α = 0.563 across the fourteen items with subscales’ alphas ranging from 0.223 (behavioral disengagement) to 0.814 (acceptance). The behavioral disengagement subscale performed poorly when considering its factor loadings: one of the two items (“I’ve been giving up the attempt to cope.”) had a very low loading (0.241) onto the subscale with a cross-loading (0.242) on the self-blame subscale. Despite these shortcomings, the subscale was maintained to retain conceptual clarity with the original measure. Removal of the items did not significantly improve the overall factor structure or the scale’s explanation of score variance. In addition, moving the item to the self-blame subscale did not significantly impact the subscale’s internal constancy or scale’s explained variance. Besides this subscale, the subscales’ internal consistencies were similar to Carver’s original measure (7).

We entered the seven factors into a second factor analysis using the same criteria as the first analysis to determine if higher order factors existed, as indicated on the scree plot. The behavioral disengagement subscale did not load onto any factor, so we eliminated it. We analyzed the remaining 6 factors. Our findings resulted in two factors that explained 39.0% of score variance in the subscales. Table 4 lists the factor loadings for each subscale and the internal consistency of each factor, α = 0.757 and 0.701, respectively. The first higher-order factor included active, positive reframing, emotional support (negatively loaded), and acceptance. The second higher-order factor included denial and self-blame. While the acceptance subscale cross-loaded onto both factors, we retained it in the first higher-order factor given its stronger loading and theoretical similarity with the other subscales in this factor.

Table 4.

High Level Factor Analysis

Subscale	Factorsa
	I	II
Factor One (α = 0.757)
Active	.718	−.051
Positive Reframing	.523	−.068
Emotional Support	−.498	.108
Acceptance	.517	−.405
Factor Two (α = 0.701)
Denial	−.023	.844
Self-blame	−.103	.360

^aFactor loadings are bolded according to the factor on which they most strongly loaded.

Hypothesized Relationships with Other Constructs

The Brief COPE positive reframing subscale was significantly positively related to the FACT-G total scale as hypothesized, r = 0.23, p ≤ 0.001. All Brief COPE subscales were significantly correlated with the FACT-G with more engaged types of coping demonstrating a positive association and less engaged types having a negative association. In addition, the HADS anxiety and depression subscales correlated with the Brief COPE subscales as anticipated with the exception of the emotional support subscale being positively correlated with the anxiety (r = 0.16, p ≤ 0.01) and depression (r = 0.29, p ≤ 0.001) subscales rather than negatively correlated as originally hypothesized. Also, the positive reframing subscale was not significantly correlated with the anxiety subscale (r = −0.06, p = 0.28), though it was with the depression subscale (r = 0.23, p ≤ 0.001). Table 5 reports the correlations between coping, anxiety and depression, and quality of life.

Table 5.

Correlations between Brief COPE, HADS Subscales, and FACT-G

Brief COPE Subscale	HADS Subscale				FACT-G
	Anxiety		Depression
	r	p-value	r	p-value	r	p-value
Factor One
Active	−.173	≤ .001	−.304	≤ .001	.270	≤ .001
Positive Reframing	−.058	.281	−.258	≤ .001	.226	≤ .001
Emotional Support	.159	≤ .01	.288	≤ .001	.319	≤ .001
Acceptance	−.289	≤ .001	−.368	≤ .001	.383	≤ .001
Factor Two
Denial	.388	≤ .001	.261	≤ .001	−.348	≤ .001
Self-blame	.378	≤ .001	.279	≤ .001	−.339	≤ .001
Behavioral Disengagement	.225	≤ .001	.232	≤ .001	−.216	≤ .001

Discussion

This study provides evidence for good psychometric properties of the Brief COPE in patients newly diagnosed with incurable cancer. Our results largely supported a seven factor solution matching the original subscales, with the exception of the behavioral disengagement subscale that had an extremely low internal consistency and a weak factor loading. This pattern may be related to patients’ hesitancy to endorse items that suggest they are “giving up” trying to improve their lives with cancer (a term used within both behavioral disengagement items) when newly confronted with an incurable cancer diagnosis. The Brief COPE subscales showed generally expected inter-relationships and associations with QOL and psychological distress symptoms with some exceptions.

The emotional support subscale demonstrated unexpected results throughout our analysis. It was negatively associated with engaged types of coping, had a moderately strong but negative factor loading in the higher-order factor analysis, and was positively associated with anxiety and depression. Indeed, our results reflect previous findings of the emotional support being related to distress in cancer patients (19). Nonetheless, using emotional support is generally considered an active, engaged form of coping, yet our results indicate that it is distinct from other forms of active coping given its association with behavioral disengagement and self-blaming coping. These unexpected findings may reflect the functional role emotional support serves in helping patients cope with their cancer (e.g., patients who find it difficult to cope with their cancer on their own may rely on others to assist them in managing their illness).

Our exploratory higher-order factor analysis revealed a two-factor solution with active, positive reframing, emotional support (negatively loaded), and acceptance loading onto the first higher-order factor, and denial and self-blame loading onto the second higher-order factor. Similarly, Carver’s (7) original higher-order factor analysis of all 14 subscales found active, positive reframing, and one acceptance item loading onto one factor, with denial and self-blame loading onto another factor. These similarities are notable as Carver’s study population was hurricane survivors and ours was patients newly diagnosed with incurable cancer. These groups were experiencing qualitatively different types of stress and expectations for recovery.

Coping researchers have debated how best to label higher order categories of coping and at least 14 categorization systems have been identified in the literature (20). However, we refrain from assigning such labels to our higher-order factors. Given the conflicting ways of naming these subscale groupings, we defer to Carver’s intention to use the Brief COPE primarily as a flexible model of various coping methods while recognizing that some subscales tend to group together and reflect broader themes in the coping literature.

Research Implications

Our findings are consistent with the literature suggesting that greater use of active, acceptance and positive reframing coping is associated with less emotional distress (21–23). We also found that greater use of self-blame, denial, and behavioral disengagement coping is associated with greater emotional distress (9, 22, 24). Since previous studies were conducted mostly among patients with early-stage breast cancer, our data corroborate and reinforce these findings in a new sample of patients with incurable lung and GI cancers.

These coping patterns are important because cancer elevates risk for psychological distress (25) which in turn has been linked with poorer QOL (26–28), greater barriers to quality cancer care (29–33) and worse morbidity and mortality (32–36). Evidence also suggests that the stress of a cancer diagnosis can increase use of more defensive coping, such as denial (37), and that poor perceived skills for coping with stress have predicted worse well-being among patients (38). Adding the Brief COPE to clinical studies may help elucidate the role of coping style in facilitating or impeding adjustment to stressors relevant to patients with incurable cancer.

Clinical Implications

The Brief COPE is a practical, simple measure that quickly assesses an individual’s coping strategies. Such information may help clinicians assist patients to cope more adaptively by using effective strategies to improve QOL and mood. Patients utilize different coping strategies based on their understanding of their illness, available supports, personality, past experience with adversity, and level of psychological distress, among other factors. Seeking emotional support (which was associated with increased depression and anxiety and lower QOL in our sample) is not always effective; it may be a proxy for high distress and influenced by the quality of support available. While the current psychometric analysis used a cross-sectional approach to evaluate the Brief COPE, prospective studies could elucidate how patients’ coping evolves over time. For example, denial may be adaptive for some patients at certain moments in their disease trajectory to maintain functioning in the face of future uncertainty. However, this style may later become maladaptive if it leads to inaccurate illness understanding and poor treatment decisions.

To advance the field, researchers may want to examine the performance of the Brief COPE over time. This will help identify how patient coping styles remain consistent, fluctuate, or evolve throughout the course of terminal cancer and at critical points in cancer care. Certain behaviors may have complex or unique associations with patient outcomes. In this study, greater emotional support unexpectedly was associated with less use of active coping, positive reframing, and acceptance, as well as greater psychological distress. Future studies should address the temporal relationships among these factors, with attention to potential dynamic interactions between cancer-related stressors, coping behaviors, and psychological distress over time.

This study has several limitations. Participants completed the Brief COPE within eight weeks of their cancer diagnosis and may not have had sufficient time to engage in the potential range of coping strategies. Also, the majority of participants were diagnosed with incurable non-small cell lung and pancreatic cancer and may have different stressors and coping styles than patients with other types of incurable cancer. Finally, future testing should directly address the influence of violations of normality such as skewness and consider adjustments to item response options.

To our knowledge, this study is the first to psychometrically validate the Brief COPE within a sample of patients newly diagnosed with incurable lung and GI cancers. We tested specific hypotheses to evaluate the accuracy of the Brief COPE for assessing patients’ use of coping strategies. While the results support many of our hypotheses, noted variations included poor performance of the behavioral disengagement subscale and unexpected positive correlations of the emotional support subscale with psychological distress and more avoidant types of coping. Future research should confirm our findings and seek to improve our understanding of the coping strategies employed by patients with incurable cancer.