Lemons to lemonade: Effects of a biobehavioral intervention for cancer patients on later life changes

Abstract

Objective:

The sustainment of gains for cancer patients provided psychosocial interventions is unclear. Furthermore, it is unknown whether interventions also yield long term positive changes. The present study experimentally tests if an intervention delivered at cancer diagnosis could yield broad, long term, changes in domains such as relationships, worldview, priorities, and goals. It was hypothesized that the intervention group would report more positive and fewer negative life changes during survivorship versus the control group.

Methods:

Patients with Stage II/III breast cancer were randomized to Biobehavioral Intervention (BBI) or Assessment only. At randomization, patients completed measures of cancer stress (Impact of Events Scale, IES) and depressive symptoms (Center for Epidemiological Studies Depression scale, CES-D). At the 24 month follow-up, survivors (N=160) completed a thought listing task with 7 prompts (e.g., “relationships”). Patients listed thoughts about change since diagnosis and rated each for valence. Groups were compared on the frequency of positive/negative thoughts across prompts. Listed thoughts were content analyzed.

Results:

BBI survivors reported significantly more positive changes (p<.05), controlling for IES and CES-D. Groups did not differ on negative changes. Patients with higher IES/CES-D scores reported more negative changes at 24 months (ps<.05). Content analysis revealed a predominance (13/23) of positive thought categories.

Conclusions:

Adding support for efficacy, BBI survivors reported significantly more positive life changes since diagnosis than survivors not receiving BBI. More generally, heightened stress/depressive symptoms at diagnosis foretold survivors’ reporting of more negative changes. Thought listing is a strategy to obtain personalized accounts of life changes after breast cancer.

Introduction

When the stress of cancer goes unchecked, it covaries with concurrent poor quality of life and physical symptoms and predicts the same thereafter, with the added risk for treatment morbidities and premature cancer mortality (Chida, Hamer, Wardle, & Steptoe, 2008). Symptoms of depression at diagnosis are similarly toxic, generating stress (Wu & Andersen, 2010) and heightening risk for cancer death (RR=1.18) (Hjerl et al., 2003; Pinquart & Duberstein, 2010; Satin, Linden, & Phillips, 2009; Suppli et al., 2014). Fortunately, randomized controlled trials (RCTs) have shown psychological treatments to be differentially effective in reducing stress and depressive symptoms (e.g., Faller et al., 2013).

However, there are nagging concerns about post cancer treatment years (Hewitt, Greenfield, & Stovall, 2006; Nekhlyudov, 2017). The difficulty of this transitional period has received much clinical discussion (Kantsiper et al., 2009). The relief of treatment finally ending is juxtaposed with expected and unexpected changes, many of which may be negative. Patients may miss the regularity of contact (and support) from other patients and oncology providers, be surprised at continuance of symptoms such as fatigue or the emergence of late treatment effects, or find that resuming life activities requires a “new normal” perspective.

It is unknown if efficacious interventions delivered at diagnosis/during cancer treatment have sustained benefits or provide resilience to survivorship challenges, as follow up data are meager (Faller et al., 2013; Galway et al., 2012). Also, few interventions have been sufficiently long to help patients through the time when cancer therapy ends. Cross sectional literature suggests, not surprisingly, that survivors acknowledge positive and negative changes following cancer even when they have received no intervention (Bellizzi, Miller, Arora, & Rowland, 2007). Although there are few data from RCTs, Antoni et al. (2001) developed a measure of “benefit finding” and found intervention patients to endorse items such as “cancer helped me realize who my real friends are” at a higher frequency than control patients nine months post treatment.

Studies of life change post cancer have used measures of Posttraumatic Growth [Post-Traumatic Growth Inventory (PTGI); Tedeschi & Calhoun, 1996] or perceptions of benefits [Perceived Benefit Scales (PBS); McMillen & Fisher, 1998]. The latter have 21 and 38 items, respectively, assessing finite domains, the relevance and extent of which would be variable across people. These measures were designed for general usage, and so cancer patients were not sampled for the psychometric studies. However, data suggest type of experience or trauma impacts the process of benefit-finding (Helgeson, Reynolds, & Tomich, 2006) and perhaps by extension, the relevance of item content to different people.

Considering this, a different metric was used to study survivors’ cancer experiences since diagnosis, one which would capture cancer patients’ thoughts about changes that were personally relevant. “Thought listing” is a methodology important to attitudes and social cognition research for decades, though its use has extended to the study of psychopathology, psychotherapy, and health behaviors (e.g., Cacioppo, Hippel, & Ernst, 1997; Orbell & Kyriakaki, 2008). As conceptualized by Cacioppo and Petty (1981), thought listing is based on the assumption that the psychological significance of an individual’s thoughts and feelings, as well as underlying cognitions, can be examined via the thoughts, ideas, images, and feelings the individual describes. As used in attitudes research, for example, individuals are provided with a stimulus (picture of a Nazi swastika) or experience (watching a video of racial discrimination) and asked to report (usually in writing) their thoughts. In this manner, the thoughts are, individually, face valid and non-ambiguous, and as such, have criterion validity. Additionally, individuals are asked to rate the valence of each thought (i.e., positive, negative, neutral) which is the unit of statistical analysis. Thought listing is particularly helpful when there are few predetermined ideas and, in particular, when the aim is to determine the effects of a manipulation (Cacioppo et al., 1997).

The aim of the research was to experimentally test if a psychological intervention delivered at diagnosis and in the months thereafter could yield changes for the long term, ones that may be beneficial and/or offset survivorship difficulties. To do this, cancer patients from a well-characterized randomized clinical trial (ClinicalTrials.gov Identifier: NCT0357862), contrasting a biobehavioral intervention (BBI) and assessment versus assessment only, were studied. As delivered, BBI began in the weeks following diagnosis and extended through the next 12 months as patients began, ended, and started recovery from surgery and adjuvant therapies. BBI was subsequently found to be efficacious in reducing emotional distress, improving health behaviors, compliance, and immunity both during the intervention (Andersen et al., 2004) and when it ended at 12 months (Andersen et al., 2010). As part of the RCT, patients were regularly followed, and at 24-months the routine assessment was completed, with the addition of a thought listing task to obtain patients’ reports of life changes since diagnosis. This time point was chosen because patients would be minimally 12 months post the end of cancer treatment and further, physicians had forewarned patients that 12 months would be needed for full recovery from cancer treatment. With thought listing, all patients were anticipated to report changes, some positive, others negative, since their diagnoses. However, it was hypothesized that survivors who had received BBI would report significantly more positive changes and fewer negative changes in comparison to survivors who had been only assessed. For a more stringent test, analyses controlled for patients’ pre randomization (baseline) levels of stress and depressive symptoms. More generally, it was expected that patients reporting higher levels of stress or depressive symptoms at randomization would, as survivors, report more negative changes and fewer positive changes, consistent with prior studies (e.g., Bellizzi et al., 2007).

Methods

Design

The full trial protocol has been published (Andersen et al., 2004, 2008). An experimental design with longitudinal follow up was used. As previously described, women newly diagnosed with stage II or III breast cancer, surgically treated, and awaiting adjuvant therapy (if any) were sought. Exclusion criteria were as follows: prior cancer diagnosis, refusal of cancer treatment, age less than 20 years or more than 85 years, residence more than 90 miles from the research site, and diagnoses of mental retardation, severe or untreated psychopathology (e.g., schizophrenia), neurologic disorders, dementia, or any immunologic condition or disease. Participants were randomized within strata by the principal investigator, using White and Freedman’s (1978) minimization method; prognostic and psychosocial strata were as follows: (1) extent of disease and treatment using four levels: negative nodes but tumor more than 2 cm, one to three positive nodes, more than four positive nodes with bone marrow transplant (BMT), and more than four positive nodes without BMT; (2) hormone receptor status (positive v negative); (3) menopausal status (pre- or peri-menopausal v post-menopausal); and (4) partner status (spouse or partner v none). Patients were assessed and randomized to two groups: Biobehavioral Intervention and assessment (BBI) versus Assessment only. Patients were reassessed at 4, 8, and 12 months (spanning the 12 months of BBI delivery) in Year 1, every six months in Years 2–5, and annually thereafter. The duration of follow‐up ranged from 7 to 13 years, with a median follow up of 11 years. See Figure 1 for study flow.

Figure 1.

Study flow for a RCT accruing and assessing Stage II/III breast cancer patients postsurgery and prior to adjuvant therapy with subsequent randomization to Biobehavioral Intervention and assessment versus Assessment only arms. Non-cumulative frequency and reasons for attrition are provided.

Participants

At 24 months, 160 women remained in the trial and were disease free. The survivors were primarily Caucasian (91%), middle aged (M=51, SD=11 years), partnered (75%), and had a mean of 15 years of education (SD=3 years). Most (90%) had been diagnosed with stage II rather than stage III disease, with 86% having received chemotherapy, 59% radiation therapy, and 79% hormonal therapy.

Procedures

All procedures were reviewed and approved by the Institutional Review Board at The Ohio State University. Power analyses suggested a total number of 200 patients to test the disease endpoint (recurrence) hypothesis, and 227 patients were accrued. As previously described (Andersen et al., 2004), patients were accrued from two sources: consecutive patients at a university-affiliated National Cancer Institute–designated Comprehensive Cancer Center (n = 189) and self- and physician-referred patients from the community (n = 38). All were assessed at baseline (prior to beginning any adjuvant treatment), occurring between May 1994 and May 2000. A female research assistant blinded to group assignment aided patients’ completion of self-report measures of cancer stress, depressive symptoms, and other areas in the surgical oncology clinic. Individuals received $25.00 per assessment. At 24 months, occurring between December 1996 and September 2002, the routine assessment was completed but survivors then worked independently on the thought-listing task (see below). Thought listing was only administered at 24-months.

A brief description of the biobehavioral intervention (BBI) is provided (see Andersen et al., 2009 for details). Conceptually based on the biobehavioral model of cancer stress and disease course (Andersen, Kiecolt-Glaser, & Glaser, 1994), BBI was designed to reduce stress, enhance quality of life, increase positive health behaviors, decrease negative health behaviors, and improve compliance and health (Andersen et al., 2009). Multi-component, BBI has modules for understanding and reducing stress, disease/treatment information, problem solving, assertive communication, social support, body image/sexuality, and health behaviors. Therapists used a detailed session-by-session manual and patients received companion guidebooks. The intervention was delivered in cohorts (n = 13), ranging from 8 to 12 patients, and led by two clinical psychologists. A cohort met weekly for 1.5 hours for 18 weekly sessions (intensive) followed by 8 monthly sessions (maintenance), a total of 26 sessions in 12 months.

Previously published interim analyses showed robust, durable gains at 4- (Andersen et al., 2004) and 12-months (Andersen et al., 2007) for the BBI arm in contrast to Assessment only, both for secondary outcomes [i.e., reduced negative mood, increased social support and health behaviors, more favorable chemotherapy dose intensity, and better health] and T-cell immunity. After completion of 10-years of follow-up (median 11.5 years) endpoint data showed significantly reduced risk of breast cancer recurrence for the BBI arm versus Assessment only (Hazards Ratio = .55, p=.034; Andersen et al., 2008).

Measures

Cancer stress.

The Impact of Event Scale (IES) (Horowitz, Wilner, & Alvarez, 1979) assesses intrusive thoughts and avoidant thoughts and behaviors related to traumatic events. The 15 items were modified to assess the stress of cancer diagnosis and treatment. Items were rated on a scale from 0 (not at all) to 3 (often). Total scores range from 0 to 75, with higher scores indicating higher levels of stress. Internal consistency was 0.87.

Depressive symptoms.

The Iowa Short Form (Kohout, Berkman, Evans, & CornoniHuntley, 1993) of the Center for Epidemiological Studies Depression scale (CES-D) (Radloff, 1977) assesses depressive symptoms over the past week. Eleven items are rated on a scale from 0 (hardly ever or never) to 2 (much or most of the time). Total scores range from 0 to 22, with higher scores indicating greater depressive symptoms. Internal consistency was 0.63.

Thoughts about change.

Using the procedures of Cacioppo and Petty (1981), a survivor was given a notebook with 7 lined pages. Cover page instructions, broad and without valence, were as follows:

“Sometimes after a diagnosis of cancer, people report that their lives have changed. We are interested in the way your experience with breast cancer has affected your life. In the following pages, we ask you to reflect on how your life is different, if at all, as a result of your cancer diagnosis and treatment.”

Each lined page had a different header, i.e., relationships, view of the world, self, priorities, goals, activities, and other. On each page, a survivor was to briefly list as many thoughts about the prompt as they wished, one per line. After completing the 7 pages, a survivor then went back through and rated each change item as positive (+), negative (−), or neutral (0). The total number of positive, negative, and neutral changes reported across the 7 prompts were the predicted outcomes.

For the interest of the reader, item content was evaluated and is reported below.

Analytic Strategy

The RCT sample size was 227 with 160 participating at 24 months and included in these analyses. Preliminary analyses compare the 160 participants with the 67 non-participants on baseline sociodemographic, disease/treatment characteristics, and stress (IES) and depression symptom (CES-D) scores.

For the primary analyses, hierarchical multiple linear (HLM) regression analyses, entering IES or CES-D first followed by group (BBI v assessment only), tested for differences between groups in the number of positive, negative, or neutral thoughts listed. IES and CES-D were entered in separate models to reduce multicollinearity. Also, stress and depression have been found to contribute uniquely in studies of benefit finding and post-traumatic growth (Cadell, Regehr, & Hemsworth, 2003; Linley & Joseph, 2004; Manne et al., 2004), and thus separate regression models allowed for examination of each as a predictor of the three change outcomes. For this, six HLM analyses were conducted. The data were screened for outliers and other anomalies, and checked for violation of analysis assumptions. Significance was specified at 0.05 level and analyses were conducted using IBM SPSS 22.

For the interest of the reader, conceptual themes in the change items listed are reported. For this, the thought listings were inspected and the open coding technique of Strauss and Corbin (1998), a standard method for identifying qualitative conceptual categories of data, was used to discover consistent themes. Ten randomly selected changes from each of the seven prompts (i.e., relationships, view of the world, self, priorities, goals, activities, other) were examined by three independent coders. The three coders then noted themes commonly reported across survivors and areas, identifying specific categories of thoughts about change. All reported thoughts (positive, neutral and negative) were then sorted into categories by at least two raters. If a person’s change item had themes that corresponding to more than one of the categories, the item was categorized as two separate changes. In such instances, the item’s valence rating was maintained for both of the category assignments. For each person, the number of thoughts about change in each category was tallied. Change categories were not analyzed by group.

Results

Preliminary

Of the 227 patients, 21 (9%) had recurred or were deceased and 46 (20%) had dropped out of the assessments by 24-months (see Fig.1). Compared to the 67 persons who were not assessed at 24 months, the 160 continuing persons had, as a group, one more year of education [t= −3.43, df=225, p < 0.01; M=15 years (SD=2.72) v. M=14 years (SD=2.57)], and were more likely to have been treated with radiation therapy [t = −2.15, df=225, p = 0.03; 59% v. 43% treated] and hormonal therapy [t = −2.20, df=225, p=0.03; 79% v. 66% treated]. However, there were no differences between groups in baseline IES (p=0.41) or CES-D scores (p=0.06).

For the 24 month participants (N=160), analyses confirmed no differences between BBI (n=85) and Assessment only (n=75) groups in sociodemographic, disease, or treatment characteristics (ps>0.05), or IES (p=0.47), or CES-D scores (p=0.30).

Primary

Table 1 provides means, standard deviations, and intercorrelations of the measures and numbers of thoughts listed by positive, negative, or neutral valence.

Table 1.

Summary statistics for breast cancer survivors (N = 160) for baseline measures of stress (IES) and depressive symptoms (CES-D) and the life changes by valence reported at the 24 month follow up and the measure intercorrelations.

		Mean (SD)	Correlations
	Measure		IES	CES-D	Positive Changes	Neutral Changes	Negative Changes	Total Changes
Basline	Impact of Events Scale (IES)	25.75 (14.19)	1	0.60**	0.03	0.05*	0.26**	0.13**
	Center for Epidemiological Studies Depression Scale (CES-D)	5.78 ( 3.56)		1	0.06	0.06*	0.30**	0.20**
24 Months	Positive Changes	11.79 ( 8.85)			1	0.24*	0.45**	0.88**
	Neutral Changes	5.08 ( 7.70)				1	0.16**	0.40**
	Negative Changes	2.80 ( 2.40)					1	0.79**
	Total Changes							1

^**p<0.01 (2-tailed).

^*p<0.05 (2-tailed).

Group significantly predicted number of reported positive changes (see Table 2) above and beyond the IES (β=0.172, p=0.03) and the CES-D (β=0.170, p=0.04), with BBI survivors reporting significantly more positive thoughts about change than Assessment only survivors (see Figure 2). Group did not predict number of negative (p>0.90) or neutral (p>0.15) thoughts about change.

Table 2.

Results of hierarchical multiple regression analysis (N = 160) showing the significant effect of group (Biobehavioral Intervention versus Assessment only) as a predictor of positive changes, above and beyond baseline levels of cancer specific stress (IES) or depressive symptoms (CES-D). The Intervention arm reported significantly more positive life changes at two years post diagnosis.

Model	Block	Predictor	Outcome	Adjusted R2	Standardized β	t**
1	1	IES	Positive	−0.01	−0.016	0.203*
	2	Group	Changes	0.02	−0.172	2.145*
2	1	CES-D	Positive	−0.01	0.043	0.531*
	2	Group	Changes	0.02	0.170	−2.116*

^*p<0.05 (2-tailed).

Figure 2.

Mean number of positive, negative, and neutral changes reported on a thought listing task at the 24-month follow up for RCT arms. The Biobehavioral Intervention arm survivors reported significantly more positive changes since diagnosis in contrast to Assessment only survivors. Study arms did not differ on negative change or neutral change reports.

Regarding baseline IES and CES-D as predictors of changes, neither were predictive of positive thoughts about change, (ps>0.60). In contrast, both the IES (β=0.255, p=0.03) and CESD (β=0.302, p=0.01) significantly predicted number of negative thoughts, indicating that patients with higher levels of stress or depressive symptoms at randomization/baseline reported more negative thoughts when survivors at 24 months than those with lower levels of stress or depressive symptoms. Descriptively, an IES median split shows patients with higher IES scores reported a mean of 6.3 negative thoughts about change whereas those with lower IES scores reported a mean of 3.6 negative thoughts. Similarly, a CES-D median split shows patients with higher levels of depressive symptoms at baseline reported a mean of 5.9 negative thoughts about change whereas those with lower levels of depressive symptoms reported 4.0 negative thoughts.

Content Analyses of Thought Listings

In total, the 7 prompts yielded 998 thoughts about change from the 160 survivors. Three coders independently reviewed the thoughts listed and generated 22 distinct categories. Twentyseven thought items were unable to be categorized due to insufficient information and were placed in a “Not Enough Information” category. Inter-rater agreement across the thought categories was 89.8%. Labels for the 23 data-derived categories are provided (see Figure 3).

Figure 3.

Content of cancer survivors (N=160) reports of life changes (n=998) since diagnosis by content area. For each category, the percentage of total thoughts is provided.

Considering the 23 thought categories, the range of their mention was from 3 to 112 or alternatively, the lowest 0.3% (Alternative Treatments) to the highest, 11.2% (Close Relationships) frequency of total thoughts reported (see Figure 3). In addition to Close Relationships, other frequent thoughts were in the domains of Balance in Life (10.1%), and Perceptions of Others’ Attitudes (9.2%), all of which had a predominantly positive valence (93%, 79%, and 63%, respectively). Generally, there were more thought categories (13/23) rated positively (see Figure 4). Three categories had change thoughts that were predominantly negative: Sexuality (85%), Treatment Side Effects (78%), and Activities in Daily Living (63%). Three other frequently mentioned categories were regarded with ambivalence (i.e., roughly 50% of the thoughts for the category were positive and the remainder neutral/negative): Financial/Job/Insurance Concerns (52%), Meaning in Life (51%), and Attitude Changes (self and others; 50%).

Figure 4.

Content of cancer survivors (N=160) reports of life changes (n=998) since diagnosis by content area. The percentage of positive (white), neutral (grey), and negative (black) ratings for each topic is displayed, listed in order of percentage positive, from highest (left) to lowest (right).

Discussion

“When life gives you lemons, make lemonade” is a proverbial phrase used to encourage optimism and a positive can-do attitude in the face of adversity or misfortune. It is a phrase not uncommon in describing personal strivings and actions in the face of cancer (e.g., www.alexslemonade.org). The cancer experience comes with numerous unavoidable “lemons,” e.g., life threat, stress, time filled with all things oncology, and disruptions in normal activities, to note a few. The present study provides three perspectives on survivors’ ability (or not) to “make lemonade.” Prior research has demonstrated that the Biobehavioral Intervention (BBI) enabled breast cancer patients to address acute stressors during cancer treatment and shortly thereafter (Andersen et al., 2004). These data show an additional, unanticipated, benefit was an increase in positive life changes for intervention recipients. The data also reaffirmed the negative long-term effects of heightened stress and depressive symptoms at diagnosis (Hjerl et al., 2003; Pinquart & Duberstein, 2010; Satin et al., 2009; Suppli et al., 2014), but these findings are an extension, showing both stress and depressive symptoms to increase the likelihood of negativity in patients’ thoughts about life since diagnosis.

When individuals are randomized, it is assumed that their frames of reference are comparable, but thereafter their reference points as well as their thoughts and ideas may be impacted by the particular experimental condition (Cacioppo et al., 1997). Patients’ reports of more positive life changes long after the Biobehavioral Intervention had ended is, in our biased view, an exceptional outcome. Considered in the context of the RCT, the timing of the improved psychological, behavioral, and immune outcomes of BBI during the first 12 months (Andersen et al., 2004) and the timing of these life change reports at 24 months, early in the transition into survivorship, are plausible as correlates of the improved disease trajectory found for the BBI arm (Andersen et al., 2008).

Thought listing was a task enabling survivors to reflect on changes to their personal life, but valence ratings made the test of RCT effects possible. As to why BBI arm survivors had greater positivity in their thinking, the thought categories provide clues. The intervention included components for stress management, treatment adherence, garnering social support from friends/family, and making health behavior changes, among others. Also, a core component was problem solving, which aimed to teach patients a general strategy but which was practiced for the specific problems of fatigue and treatment-imposed time constraints. This was done by encouraging patients to shift priorities to maximize time for self/family, personally important activities, etc. and, simultaneously, to ask for support from friends and family for other important but effortful and time-consuming tasks (e.g. grocery shopping, transportation for children) and the mundane (e.g., house cleaning). These BBI components are closely related to the five most frequent and most positive thoughts about change reported: 93% positivity for close relationships, 79% positivity for balance in life, 75% positivity for perceptions of others’ attitudes, 75% positivity for appreciation/enjoyment of life, and 63% positivity for health behavior changes. Though not known for certain, the combination of content, valence, and frequency data suggests that BBI helped enhance life areas that were both common and important, and doing so resulted in their reports of overwhelmingly positive changes in these domains.

We consider the thought listing categories and their possible relevance to today’s survivors of breast cancer. In terms of absolute numbers, these survivors had many more (four times) the number of positive thoughts about the changes occurring since diagnosis than negative, a testimony to their resilience (“making lemonade”). This finding is consistent with contemporary cross sectional studies of survivors reporting a predominance of positive changes but also acknowledgement of negative (Bellizzi et al., 2007; Helgeson et al., 2006). Inspection of the few (4) predominantly negative thought categories (i.e., “lemons”) shows two to be the result of breast cancer treatment, per se (Sexuality, Side Effects). Each have some aspects which are unavoidable and without efficacious treatment (e.g., sequellae of chemotherapy induced menopause, breast changes) and others also unavoidable but which may resolve, though not readily or easily (e.g., sexual morbidity, fatigue). All of the latter difficult circumstances have remained across decades (Fallowfield & Jenkins, 2015; Mols, Vingerhoets, Coebergh, & Van De Poll-Franse, 2005). A third category, Financial Concerns, received little attention in the 1990’s and has only recently. Unfortunately, cancer patients with financial concerns will remain constant for the foreseeable future and possibly increase in numbers with, for example, restrictions to the Affordable Care Act, and despite data showing the life and death consequences of economic stressors (Kale & Carroll, 2016; Ramsey et al., 2016). Thus, the negative areas of change provided by women in the mid 1990’s remain challenges today. They are also ones requiring targeted, multi-disciplinary efforts that even the best multicomponent psychosocial interventions might not provide.

Numerous replications show that without interventions, stress or depressive symptoms at diagnosis covary with hastened recurrence or cancer death (Chida et al., 2008; Kang et al., 2009; Loberiza et al., 2002; Standish et al., 2008; Wiltschke et al., 1995). These data do show, again, that high levels of stress and depressive symptoms at diagnosis portend other negative outcomes. This example is a poignant one, as women who had reported high stress/depression wrote, with only one or two word neutral prompts, significantly more (nearly twice) the number of negative thoughts than their peers with lower stress/depressive symptoms. Stress and depression are longterm limiting factors, as early symptom declines post diagnosis may be followed with symptom rebounding (Hopko et al., 2015). The snapshots provided here underscores the importance of recommendations to screen patients at the time of diagnosis for anxiety and depression and then provide stepped care, which includes referral to receive empirically supported psychological interventions (Andersen et al., 2016; Osborn, Demoncada, & Feuerstein, 2006; Schneider, Moyer, Knapp-Oliver, Sohl, & Cannella, 2011; Sheard & Maguire, 1999).

While used for decades in other contexts, thought listing has not been previously used in psychological research in cancer and considering the content generated, it provided a new, detailed perspective of survivors’ thoughts about life changes since diagnosis. Importantly, the thought listing methodology does not involve one’s recollections of thoughts or events from times past (or even the last week), but only a listing of thoughts accessible at the time of assessment. Survivors listed many positive changes in domains not measured by the Benefit Finding Scale and/or the Post Traumatic Growth Inventory [e.g., 13 positive categories (e.g., health behaviors, side effects, altruism], in addition to listing areas of negative change (see supplementary materials for a comparison of categories/items). Once generated, coding of content can be done reliably, and doing so offers both breadth and depth to aide researchers’ and clinicians’ understanding of the experiences of breast cancer survivors. For example, information generated could be valuable to design interventions, choosing strategies to achieve change in areas patients view positively and reduce the likelihood or impact of negative change areas. Finally, the coupling of content and valence provides a nuanced understanding of individuals’ thoughts about the circumstance under study.

To consider the generalizability of the study data, a useful first comparison is that of the context of treatment of women diagnosed with stage II or III breast cancer in the 1990’s and those today as well as the psychological characteristics of this sample. While some changes in breast cancer care have occurred decades (e.g., sentinel node biopsy, shorter in hospital stay; discontinuation of bone marrow transplant) many key ones are similar if not the same. These patients received chemotherapies (primarily anthracycline-based) which are still commonly used [see Andersen et al. (2016) supplementary materials for a detailed review]. Second, the surgeries performed (46% breast-conserving surgery, 51% modified radical mastectomy) are comparable to recent rates (e.g., Morrow et al., 2009) and align with consensus treatment recommendations for loco-regional breast cancer (Kaufmann, Morrow, von Minckwitz, Harris, & The Biedenkopf Expert Panel Members, 2010).

Psychologically, use of the CES-D clinical cut-off of 10 shows 15% (n=24) of this sample with clinically significant depressive symptoms at baseline. Though studies vary, this is comparable to recent 20% point prevalence of depression among cancer patients (Mitchell et al., 2011). [For the entire (N=227) sample, the rate was 20% (Thornton, Andersen, Schuler, & Carson, 2009)]. Regarding sociodemographics, the predominantly Caucasian, educated, and above average income characteristics still, unfortunately, reflect those seeking care at comprehensive cancer centers. Thus, the thought categories generated may not represent those of samples diverse in race, ethnicity, education, or income levels or those treated in the community. For example, Tomich and Helgeson (2004), using a cross sectional design, found women with lower socioeconomic status, minorities, and those with more severe disease perceived more benefits four months after breast cancer diagnosis than women with higher socioeconomic status, Caucasian race, and early stage (Stage I) disease. But, contrary findings have also appeared (Cordova, Cunningham, Carlson, & Andrykowski, 2001; Updegraff, Taylor, Kemeny, & Wyatt, 2002). It has been suggested that race and/or socioeconomic status are confounded with psychosocial variables (e.g., cognitive appraisal, personality, coping, social support, religion, psychological distress, affect, quality of life, etc.) that have a clearer theoretical relation to post-traumatic growth, for example (Linley & Joseph, 2004). Regardless, the sociodemographic composition of this sample may have led, overall, to more positive and fewer negative thoughts about change. The latter may have been another factor contributing the lower overall frequency of negative thoughts.

The group differences reported here were achieved with an intervention that remains contemporary. The BBI is the only empirically supported, cancer specific intervention that has moved through the translational science continuum. The BBI has been disseminated to community mental health oncology providers, who have sustained BBI implementation, and shown clinical gains for patients differing from the RCT sample in sex and cancer site (Ashmore et al., in press). Based on this, it is conceivable that similar effects of positive life changes amongst other patients treated with BBI could be found. When long term intervention effects are studied, group differences are usually null, as one might expect, as across time negative affect is declining for all patients (Avis et al., 2013; Burgess et al., 2005). Other trial data (Andersen et al., 2016), trajectory analyses with the assessment only group, show a steep slope of decline in stress through 12 months, with only modest decline across the next four years. A steep slope of depressive symptom declines also occurred, though ending at 7 months and with a flat slope thereafter. These long-term trajectories would be at least similar for the BBI group. In the analyses presented here, the effect of group remained significant when controlling for baseline stress/depressive symptoms. Considering all the sources of data from the trial, the importance of the control variables, and the natural post diagnosis decline in negative affect seen in cancer patients successfully treated for their disease, accounting for 2% of the total variance in positive changes is neither surprising nor disappointing.

In conclusion, a biobehavioral intervention found efficacious in reducing negative mood and improving behavioral outcomes and health for newly diagnosed breast cancer patients appears to have facilitated more positive changes in the face of cancer treatment recovery and survivorship challenges. Conversely, the data illustrate how stress and/or depressive symptoms at diagnosis might perpetuate negative life changes in the months and years after cancer. Dissemination of efficacious psychological interventions is health psychology’s frontier and data such as these illustrate the achievement of sustained gains when patients transition to survivorship.

Appendix A. CONSORT 2010 checklist for reporting a randomized trial.

Section/Topic	Item No	Checklist item	Page No
Title and abstract
	1a	Identification as a randomized trial in the title	1
	1b	Structured summary of trial design, methods, results, and conclusions (for specific guidance see CONSORT for abstracts)	2
Introduction
Background and objectives	2a 2b	Scientific background and explanation of rationale Specific objectives or hypotheses	3–5 5
Methods
Trial design	3a	Description of trial design (such as parallel, factorial) including allocation ratio	6
	3b	Important changes to methods after trial commencement (such as eligibility criteria), with reasons	N/A
Participants	4a	Eligibility criteria for participants	6
	4b	Settings and locations where the data were collected	7
Interventions	5	The interventions for each group with sufficient details to allow replication, including how and when they were actually administered	7–8
Outcomes	6a	Completely defined pre-specified primary and secondary outcome measures, including how and when they were assessed	8–9
	6b	Any changes to trial outcomes after the trial commenced, with reasons	N/A
Sample size	7a	How sample size was determined	7
	7b	When applicable, explanation of any interim analyses and stopping guidelines	8
Randomization:
			6
Sequence generation	8a	Method used to generate the random allocation sequence	*
	8b	Type of randomization; details of any restriction
Implementation	10	Who generated the random allocation sequence, who enrolled participants, and who assigned participants to interventions	6
Blinding	11a	If done, who was blinded after assignment to	*
Allocation concealment mechanism	9	Mechanism used to implement the random allocation sequence (such as sequentially numbered containers), describing any steps taken to conceal the sequence until interventions were assigned interventions (for example, participants, care providers, those assessing outcomes) and how	*
	11b	If relevant, description of the similarity of interventions	N/A
Statistical primary and secondary outcomes	12a	Statistical methods used to compare groups for methods	9–10
	12b	Methods for additional analyses, such as subgroup analyses and adjusted analyses	N/A
Results
			Fig.1
			7
			8
			*
Participant flow (a diagram is strongly recommended)	13a	For each group, the numbers of participants who were randomly assigned, received intended treatment, and were analyzed for the primary outcome	Fig. 1
	13b	For each group, losses and exclusions after randomization, together with reasons
Recruitment		14a Dates defining the periods of recruitment and follow-up
	14b	Why the trial ended or was stopped
Baseline data		15 A table showing baseline demographic and clinical characteristics for each group
Numbers analysed	16	For each group, number of participants (denominator) included in each analysis and whether the analysis was by original assigned groups	10–11
Outcomes and estimation	17a	For each primary and secondary outcome, results for each group, and the estimated effect size and its precision (such as 95% confidence interval)	11–12
	17b	For binary outcomes, presentation of both absolute and relative effect sizes is recommended	N/A
Ancillary analyses	18	Results of any other analyses performed, including subgroup analyses and adjusted analyses, distinguishing pre-specified from exploratory	N/A
Harms	19	All important harms or unintended effects in each * group (for specific guidance see CONSORT for harms)
Discussion
			17–18
Limitations	20	Trial limitations, addressing sources of potential bias, imprecision, and, if relevant, multiplicity of analyses	17–18
Generalizability Interpretation	21	Generalizability (external validity, applicability) of the trial findings
	22	Interpretation consistent with results, balancing benefits and harms, and considering other relevant evidence	13–17
Other information
Registration	23	Registration number and name of trial registry	5
Protocol	24	Where the full trial protocol can be accessed, if available	6
Funding	25	Sources of funding and other support (such as supply of drugs), role of funders	1

^*This item has been previously described in an earlier report (see Andersen et al., 2004, 2008)