Individual and dyadic development of personal ...

Support Care Cancer (2014) 22:53–62 DOI 10.1007/s00520-013-1949-6

ORIGINAL ARTICLE

Individual and dyadic development of personal growth in couples coping with cancer Alfred Künzler & Fridtjof W. Nussbeck & Michael T. Moser & Guy Bodenmann & Karen Kayser

Received: 28 September 2012 / Accepted: 13 August 2013 / Published online: 29 August 2013 # Springer-Verlag Berlin Heidelberg 2013

Abstract Rationale Couples share distress as well as potential personal growth (PG) after a cancer diagnosis. It is essential for professionals to learn more about the ways couples cope together with adversity. Dyadic results may help to understand controversial results in the PG literature and inform clinicians in optimizing psychological support for couples. Objective We examine the temporal and dyadic development of PG among patients and their intimate partners. In addition, life threat is examined as a potential factor influencing PG after cancer diagnosis. Methods We assessed PG using the Personal Growth Inventory in a clinically representative, mixed-type and mixed-stage cancer

A. Künzler Division of Oncology/Haematology, Cantonal Hospital Aarau, Aarau 5001, Switzerland e-mail: [email protected] A. Künzler Division of Onco-Psychology, Psychiatric Services Aargau, Aarau, Switzerland F. W. Nussbeck Department of Psychology, University of Bielefeld, Universitätsstraße 25, Bielefeld 33615, Germany e-mail: [email protected] M. T. Moser Clinical Research Aargau, Aarau, Switzerland G. Bodenmann Department of Psychology, University of Zurich, Binzmühlestrasse 14/Box 23, Zurich CH-8050, Switzerland e-mail: [email protected] K. Kayser (*) Kent School of Social Work, Oppenheimer Hall, University of Louisville, 2217 S. 3rd Street, Louisville, KY 40292, USA e-mail: [email protected]

cohort (N =154 couples) 6 and 12 months after cancer diagnosis. Medical data on cancer diagnoses and treatments were collected from physicians. Actor–partner interdependence models were applied. Results PG was reported by patients and their partners. Women (either as patients or partners) reported more PG than male patients or partners. PG remained relatively stable over 6 months and was related to whether the patient was receiving curative or palliative treatment. Female patients experienced less PG 6 months after the cancer diagnosis and if treatment was curative. Male patients experienced less PG if their partners experienced PG, and the treatment was curative. Conclusions Dyadic growth is a phenomenon not limited to breast cancer or female patient couples and may represent a form of dyadic coping. Patients and partners seem to develop individual and dyadic growth, depending on a combination of gender and life threat. Psycho-oncology services may want to promote both couple level coping and support in order to optimize cancer care. Keywords Cancer . Oncology . Couples . Posttraumatic growth . APIM

Introduction When faced with a life-threatening disease, patients’ reactions vary widely, depending on the different appraisals of the situation [1]. Cancer patients often report distress, with prevalence rates varying from 30 to 90 % depending on time of assessment, tumor location, treatment, and sample characteristics [2–7]. A growing body of literature suggests that some individuals coping with cancer also derive benefits from their experiences. Various labels such as posttraumatic growth or personal growth have been used to identify these positive changes, with personal growth (PG) being the most commonly used [8, 9]. In the past few years, studies in the field of PG have increased significantly

54

Support Care Cancer (2014) 22:53–62

and several reviews are now available [10–16]. However, studies have resulted in mixed findings regarding the relationships between PG and other variables. While PG has been found to be associated with less depression and more positive well-being, it has also been associated with more intrusive thoughts and unrelated to anxiety and global distress [10]. In another study, PG during the first year after a cancer diagnosis predicted better adjustment in the long term [17]. Demographic and personality variables have been found to predict PG [11]. The present study aims to expand our understanding of how PG develops by examining it over time and within the context of the patient and intimate partner’s relationship. It also explores how the potential life threat of cancer influences the experience of PG.

adverse events [10, 11, 32], and female partners tend to provide more support than male partners [23]. Couple communication and spouse behavior can significantly affect adjustment to cancer for couples with a female patient [34] while less of a correlation is found with male patient couples [35]. It is unclear as to how gender and role (patient vs. partner) influence PG of couples coping with cancer [34–41]. Many studies on couples and cancer have used homogeneous breast cancer samples, which lead to a confounding of gender and role. Discrepant findings may also be the result of narrowbased cross-sectional investigations [24, 42].

Cancer as a traumatic event

Knowledge about cancer-related stress, the dyadic system, and the ways couples cope may help to more fully understand how PG can be facilitated and support strategies enhanced within couples. Specific research questions are the following: (1) Do cancer patients relatively soon after diagnosis (6 and 12 months) experience personal growth? (2) Do their intimate partners experience personal growth at the same time? (3) Are there gender differences in personal growth? (4) Do patients’ and partners’ perceived personal growth affect each other over time (dyadic actor and partner effects)? (5) Is there a differential effect of curative vs. palliative purpose of cancer treatment on personal growth?

It has been hypothesized that stressors must be perceived as sufficiently threatening and therefore psychologically distressing in order to produce PG [13]. PG has been examined after various types of traumatic events and differences between type of trauma and amount of growth have been found. The objective severity of disease and subjective perception of distress seem to be related to PG. However, there is no clear evidence whether the traumatic event itself or the subjective experience (e.g., life threat) promotes PG. A consistent relationship between PG and disease variables has not been established [18]. Some researchers propose a curvilinear association, with very low and very high stress negatively correlated with PG [18–21]. The finding of more perceived benefit among breast cancer patients with an advanced (but not the most severe) stages of disease supports this theory [22]. The social context Support from an intimate partner often plays a critical role in a cancer patient’s adjustment. However, the role of partners should not be limited to only support provision [23]. Partners are also a part of a distressed system with both partners perceiving high levels of distress [3–6, 24], as well as posttraumatic growth [25]. Coping with a life-threatening stressor may be viewed as a genuine systemic task involving dyadic coping [26–28]. The way both partners deal independently but also jointly with the stress will impact their adjustment to the threat. In general, couples intensify their dyadic interaction when stressors are present [26, 29]. Dyadic processes thus merit high attention in couples affected by cancer [30–32]. Gender and role differences It has been claimed that women and men cope differently with their own illnesses [33] as well as illnesses in the family [24]. For example, women seem to derive more benefits from

Research questions

Method Sample and recruitment procedure Our sample is derived from a large coping study assessing patients who had recently been diagnosed with any type and stage of cancer [5]. The study was conducted with approval by the ethics committees of involved regions. We followed the Total Design Method [43] in order to maximize response rates. The primary recruitment site was a cancer center and outpatient clinics treating cancer patients as part of an interdisciplinary rural hospital in Switzerland. Adult patients (1) receiving any type of primary cancer diagnosis within 8 weeks and (2) living with an intimate partner for the previous 6 months were approached by phone with their partners to participate in the study. Return rate for mailed questionnaires was 74 %. Participants were not compensated. Medical data about their cancer disease were provided by the patients’ primary doctor. Further details regarding recruitment procedures have been reported previously [5]. In the current study, we analyzed data from 154 couples with both partners participating in a longitudinal study. Six months (t1) and 1 year after cancer diagnosis (t2), patients and partners provided ratings on personal growth they perceived considering the preceding 6 months. Since posttraumatic

Support Care Cancer (2014) 22:53–62

personal growth (PG) takes time to emerge [44] and “growth” reported soon after the event may rather reflect a cognitive strategy to reduce stress [45]), we began assessing PG only 6 months after the time of the diagnosis. The second point of assessment, 1 year after diagnosis, was chosen because at that time in many cases, primary cancer treatment is complete. It thus reflects an important step in the coping process. In this study, 94 male and 60 female patients participated with their respective partners. Basic demographics are given in Table 1. Patients and partners were well educated (88 % attended school and other education programs for more than 10 years). Couples had been living together for 29 years on average. Patients had a variety of cancer types with tumor stages ranging from early to late (see Table 2). Most of the patients received an initial curative-intent treatment that included a combination of surgery, chemotherapy, and radiation. Participating patients did not differ from nonparticipating patients on all variables that were initially available (age, tumor site, stage of disease, physician-rated physical functioning). Measures Personal growth Personal growth was assessed using the Posttraumatic Growth Inventory (PGI) [18] in its German translation [38]. The PGI is a retrospective self-evaluative measure suitable for cancer patients [46] and their partners. Personal growth of patients will be named self personal growth (SPG) in this study in contrast to partners’ personal growth (PPG). The scores at t1 (6 months after diagnosis; SPG1 and PPG1) describe the development of personal growth during the time between the moment of diagnosis and 6 months later. The scores at t2 (1 year after diagnosis; SPG2 and PPG2) describe the period between t1 and t2. For our analyses, we computed the mean score of the 21 items of the PGI as total score for each time of measurement. Based on consistent advice from R. Tedeschi (personal communication, 22 April 2004) and A. Maercker (personal communication, 21 April 2004), we applied the original 6-point scale instead of the German 3-point scale. The total score of the German questionnaire has a very high internal consistency (for our sample Cronbach’s alpha=0.93 for SPG1, 0.94 for SPG2, 0.93 for PPG1, 0.94 for PPG2). Examining the course of personal growth we control for the influence of sex of the patient and the purpose of cancer therapy at the moment of diagnosis. The purpose of cancer therapy can be either (1) curative (eliminating the disease) or (2) palliative (promoting quality of life while living with the disease). A palliative situation does not imply that death is imminent but that the disease cannot be eliminated; it may be controlled for a time ranging from months up to decades. Still, with a palliative purpose of cancer treatment, the patient continues to experience cancer as life-threatening.

55

Statistical analyses In order to investigate the course of personal growth over time and to simultaneously estimate the influences partners exert on each other, we estimated a series of multigroup actor– partner interdependence Models (APIM) [47]. APIM allowed for a simultaneous analysis of four effects (see Fig. 1): (1) how a patient’s personal growth at t1 (SPG1) influences her or his personal growth at t2 (SPG2; annotated arrow a1), (2) how a partner’s personal growth at t1 (PPG1) influences her or his personal growth at t2 (PPG2; a2), (3) how a patient’s personal growth at t1 (SPG1) influences her or his partner’s personal growth at t2 (PPG2; p1), and (4) how a partner’s personal growth at t1 influences the patient’s personal growth at t2 (PPG2; p2). In addition to the regression weights, APIM allowed for an analysis of the partial correlation between the two variables representing personal growth at t2 (SPG2 and PPG2) identifying these parts of common variance that cannot be predicted by personal growth at t1. The quality of the prediction was determined by inspection of the determination coefficients and the sign and height of the regression weights as well as their associated p values. In order to determine whether couples react differentially to the diagnoses if the patient is female or male, we simultaneously estimated one APIM for each of the constellations. Figure 1 presents a multigroup APIM (model 1): one for male patients and one for female patients. Multigroup analysis in the framework of path models allows for a flexible testing if different model parameters (i.e., regression coefficients) are identical across groups. We were able to examine whether actor or partner effects have an identical impact on the dependent variables for both men and women. These assumptions can be tested restricting specific paths (e.g., a1) to be equal across groups. The information criteria as Akaike information criterion (AIC) and Bayesian information criterion (BIC) of the models can be compared to decide if the model with equality restrictions fits equally well as the model with path coefficients that are allowed to differ between the submodels [48]. Additionally, a substantial decline in the determination coefficient for the restricted model would indicate a less good predictive power of the model. Attention was paid to the amount of explained variance since we were interested in the genesis of PG at t2. In addition to the measures of personal growth, we included the purpose of the initial cancer treatment (curative vs. palliative) as predictor in a second model (model 2a). In a first step (Fig. 2a), we only included the main effect of purpose of treatment on personal growth at t2. In a second step (Fig. 2b; model 2b), we included the interaction effects between purpose of treatment and personal growth at t1.

56

Support Care Cancer (2014) 22:53–62

Table 1 Participants’ basic demographics

Variable

Patients Male

N Age Education

M (SD) Elementary school Secondary school Vocational school High school College/University

Statistical program and procedure All analyses were carried out using Mplus 5.2 [49], applying the FIML estimator which allows for flexible handling of missing data points. Mplus provides information about the sample statistics (mean, variances, and covariances/correlations) and allows for a flexible specification of the multigroup SEM. All models were estimated as described above. Model evaluations were executed based on the amount of explained variance (determination coefficients) and for nonsaturated models based on the chisquare values. We examined the significance of the regression coefficients. Model 2b (Fig. 2b) was simplified by the introduction of equality constraints which were incorporated based on an inspection of the model results. The models with and without simplification were compared to each other. For all analyses, we centered the metric variables in order to avoid multicolinearity in the models with interactions [50]. We centered the variables representing personal growth for the patients at the grand mean (male and female patients) at t1. This allowed for an identification of mean differences between the two sexes at t1 but also for an identification of an overall increase Table 2 Disease and treatment variables Variable

Cancer type

Initial treatment purpose Disease state after 1 year

Male patients

Female patients

Lung Breast Colorectal Hematologic/lymphatic Genitourinary Stomach

21 % 0 19 % 18 % 13 % 13 %

13 39 11 18 7 4

% % % % % %

Central nervous system Ear–nose–throat Others Curative Palliative Full remission Partial remission Progression

7 7 2 53 47 42 26 32

2 2 4 69 31 69 21 10

% % % % % % % %

% % % % % % % %

Partners Female

Male

Female

94

60

60

94

61 (11.3) 11 % 63 %

52 (13.5)

55 (14.2) 13 % 62 %

58 (12.5)

20 % 6%

18 % 7%

or decrease in SPG over time. Additionally, we centered all personal growth variables for partners around the grand mean for partners at t1 according to the same rationale.

Results Descriptives Descriptive statistics for the complete sample can be found in Table 3. The mean values for all scores of PG fall into the range of “experienced a bit” of growth and “experienced some growth” (ranging from 2.28 to 3.18). Thus, patients and their partners experience some PG shortly after the diagnosis (research questions 1 and 2). Mean values do not change over time; all differences between t1 and t2 were nonsignificant (p >.20). Female patients score higher than male patients on PG at t1 (mean difference=0.67, df =1, p