Intervention-related changes in coping ability drives improvements in mood and quality of life for patients taking adjuvant endocrine therapy

Abstract

Objective:

A recent randomized trial of a group psychosocial telehealth intervention (STRIDE) improved anxiety, depression, quality of life (QOL), symptom distress, coping, and self-efficacy to manage symptoms related to taking adjuvant endocrine therapy (AET) in women with non-metastatic hormone receptor-positive breast cancer. This study examined whether changes in coping and self-efficacy mediated intervention effects on anxiety, depression, QOL, and symptom distress.

Method:

Women (N=100) were recruited between 10/2019–06/2021 from Massachusetts General Hospital and were randomized to STRIDE or to the medication monitoring control group. Participants completed self-report measures of anxiety and depression (Hospital Anxiety and Depression Scale), QOL (Functional Assessment of Cancer Therapy-Breast Cancer scale), symptom distress (Breast Cancer Prevention Trial Symptom Scale), coping (Measure of Current Status–Part A), and self-efficacy (Self Efficacy for Managing AET Symptoms) at baseline and 24-week follow-up. Mediated regression models tested whether changes in coping ability and self-efficacy mediated the intervention effects on anxiety, depression, QOL, and symptom distress, controlling for key variables.

Results:

Improvements in coping across the 24-week study period mediated the effect of STRIDE on anxiety symptoms (indirect effect, B=−0.61, SE=0.28, 95% CI: −1.28, −0.17), depressive symptoms (indirect effect, B=−0.50, SE=0.21, 95% CI: −0.97, −0.15), and QOL (indirect effect, B=3.80, SE=1.25, 95% CI: 1.54, 6.49), but not symptom distress. Changes in self-efficacy did not mediate improvements in any of the proposed outcomes.

Conclusion:

Coping is an essential component of a brief group psychosocial intervention that drives improvements in mood and QOL for women with non-metastatic hormone receptor-positive breast cancer taking adjuvant endocrine therapy.

Introduction

Adjuvant endocrine therapy (AET) is a daily, orally administered treatment provided to women with hormone receptor-positive breast cancer, which encompasses approximately 75% of all breast cancer diagnoses.¹ AET medications, such as tamoxifen or aromatase inhibitors, are recommended for up to ten years¹ and are associated with a 29% relative risk reduction in mortality.² Despite the benefits of AET, rates of adherence are low, with research suggesting that up to 59% of patients do not adhere to AET as prescribed.³ Non-adherence to AET is most commonly a result of burdensome medication side effects (i.e., hot flashes, pain, fatigue), which also increase anxiety and depressive symptoms, interfere with patients’ quality of life (QOL), and exacerbate distress.⁴ These negative effects can simultaneously affect patients’ attitudes towards and adherence to AET.⁵ For example, patients’ self-efficacy for taking medication and their ability to cope with AET symptoms are constructs that are associated with AET adherence.⁶ Therefore, equipping patients with behavioral skills to manage AET symptoms may buffer against QOL decrements and mood symptoms, and, in turn, improve adherence to this critical therapy.⁶

We developed an evidence-based behavioral intervention (‘STRIDE’) to improve symptom management and adherence for patients with non-metastatic hormone receptor-positive breast cancer taking AET.⁷ STRIDE is a small, group-based, six-session (+2 booster sessions) telehealth intervention that comprises assessment and psychoeducation to overcome AET adherence barriers as well as cognitive and behavioral skills⁸ to manage physical and emotional AET symptoms and cope with distress related to breast cancer and fear of recurrence.⁷ In our pilot randomized controlled trial (RCT; N=100), patients assigned to STRIDE reported fewer anxiety and depressive symptoms, improved QOL, and lower AET symptom distress across 24 weeks, compared to a Medication Monitoring (MM) control group.⁹ Additionally, participants in STRIDE reported better coping ability (i.e., better ability to use cognitive, behavioral, and relaxation skills to cope with their distress). Participants in STRIDE also reported better self-efficacy for managing AET symptoms (i.e., more confidence in their ability to manage specific symptoms of AET). Additionally, those who reported having difficulty adhering to their AET at study enrollment exhibited objective improvements in adherence to AET, compared to those in MM.¹⁰

With these promising findings, an important next step is to identify the intervention’s underlying mechanisms, which is instrumental for increasing efficacy, improving patient outcomes, and maximizing scalability.¹¹ Given the intervention’s theoretical framework, grounded in the Cognitive Model for Menopausal Symptoms¹² and Murray’s Framework for Medication Adherence,¹³ we examined whether processes such as improving coping ability and strengthening self-efficacy for managing symptoms mediated the effects of STRIDE on anxiety and depressive symptoms, breast cancer QOL, and breast cancer symptom distress.

Materials and Method

Study Design

The current study is a pre-specified exploratory analysis of an RCT comparing the STRIDE intervention to a MM control (ClinicalTrials.gov identifier: NCT03837496). Participants were recruited between October 2019 and June 2021 from outpatient clinics at the Massachusetts General Hospital Cancer Center and community affiliates. Prior to initiating the trial, the Dana-Farber/Harvard Cancer Center Institutional Review Board approved the study protocol (IRB #18–603), which also has been published.⁷

Participants

Eligible patients were English-speaking females at least 21 years old, diagnosed with non-metastatic hormone receptor-positive breast cancer (stage 0-IIIb), who completed primary treatment, were within 1 week to 36 months of starting AET, and were functioning independently with an Eastern Cooperative Oncology Group¹⁴ performance status ≤2. Patients also reported distress about AET adherence and symptoms with an adapted National Comprehensive Cancer Network distress thermometer¹⁵ and met eligibility criteria if they scored ≥4 (range: 0–10) on any one of three questions.⁷ Patients were ineligible if they were enrolled in a clinical trial related to AET, psychotherapy intervention study, or other group psychotherapy; were undergoing primary treatment for another cancer diagnosis; or had a condition that would impact study participation (e.g., uncontrolled psychosis, cognitive impairment).

Procedures

Study staff reviewed electronic health records (EHR) and approached potentially eligible patients by phone or in person after obtaining permission from the treating oncology clinician. We obtained informed consent electronically via REDCap survey. All consented patients completed baseline patient-reported outcome measures (PROMs) and stored their AET in a study-provided Medication Event Monitoring System pill bottle and cap (MEMS Caps).¹⁶ We then randomized patients to either the STRIDE intervention or the MM control group using a computer-generated randomization scheme. Randomization was stratified by baseline distress level on the Hospital Anxiety and Depression Scale,¹⁷ using a cut-off of ≥8 (high versus low). Patients repeated PROMs at 12 and 24 weeks after baseline and were compensated $20 per assessment.

STRIDE Intervention

Patients randomized to STRIDE received follow-up oncology care as usual and participated in six weekly one-hour small group (two to three participants per group) telehealth sessions plus two individual booster sessions four- and five-months after their respective baseline assessment with a trained psychologist or psychology fellow. A comprehensive description of the intervention is available in a prior publication.⁸ This theoretically guided and evidence-based intervention teaches cognitive and behavioral skills, coping strategies, and psychoeducation for symptom management and adherence optimization. Specifically, STRIDE targets coping behaviors and self-efficacy via skills such as coping effectiveness training per stress, appraisal, and coping theory;¹⁸ mindfulness skills¹⁹ to cultivate present awareness and acceptance in light of cancer uncertainty; cognitive-behavioral tools for managing specific AET symptoms such as pain,²⁰ fatigue,²¹ hot flashes,²² and sleep disturbance;²³ and relaxation training for recognition and management of the stress response.²⁴

Medication Monitoring Control

Participants randomized to the MM control group received follow-up oncology care as usual and continued to store their AET in the MEMS Caps. After completing their final study assessment, participants in this group were offered the STRIDE workbook.

Measures

Participants self-reported sociodemographic characteristics (e.g., age, sex, race, ethnicity) as part of the baseline survey. Study staff extracted cancer and treatment-related information (e.g., stage, type of AET) from the EHR.

Potential Mediators

Changes in coping ability.

We computed a change score to indicate difference in patients’ perceived coping ability from baseline to 24 weeks using the total score of the Measure of Current Status–Part A (MOCS-A).²⁵ The MOCS-A, developed in a sample of women with nonmetastatic breast cancer, measures one’s ability to cope with distress using cognitive and relaxation skills. The MOCS-A is a broad measure of coping, and items are not specific to breast cancer or AET. It is a 13-item measure scored from zero (I cannot do this at all) to four (I can do this extremely well), with greater total scores suggesting greater coping ability. In the current sample, the MOCS-A demonstrated strong internal consistency (Cronbach’s α=.91).

Changes in self-efficacy.

We computed a change score representing differences in self-efficacy to manage AET symptoms from baseline to 24 weeks, measured by the Self Efficacy for Managing AET Symptoms Questionnaire (SESM-AET).²⁶ The SESM-AET is a validated 8-item questionnaire that measures a person’s confidence in their ability to manage specific AET symptoms, though the measure is not exhaustive of all AET-related symptoms women may experience. Items are rated on a one (not at all confident) to ten (very confident) scale, with higher scores indicating greater self-efficacy. In the current sample, the SESM-AET demonstrated good internal consistency (Cronbach’s α=.86).

Outcome Measures

Anxiety and depressive symptoms.

We assessed mood symptoms with the Hospital Anxiety and Depression Scale (HADS),¹⁷ a 14-item measure of anxiety and depressive symptoms in the past week. The HADS uses a zero to three Likert-type rating scale and sums to a total score, with higher scores indicating greater anxiety or depressive symptoms. This measure was chosen to best capture anxiety and depression in the context of individuals experiencing significant treatment-induced symptoms that may overlap with mood symptoms, such as fatigue. In the current sample, the HADS demonstrated good internal consistency (Cronbach’s α=.87).

QOL.

To assess QOL, we administered the Functional Assessment of Cancer Therapy-Breast Cancer scale (FACT-B),²⁷ which measures well-being across physical, social, emotional, functional, and breast cancer specific subscales. Participants recall aspects of QOL over the past seven days, with response options ranging from zero to four, and greater total scores indicating better QOL. In the current sample, the FACT-B demonstrated strong internal consistency (Cronbach’s α=.90).

Breast cancer symptom distress.

We assessed symptom distress using the Breast Cancer Prevention Trial Symptom Scale (BCPT),²⁸ a 26-item measure that asks participants to rate how bothered they have been over the past week about eight symptoms of breast cancer treatment (e.g., hot flashes, sexual problems) on a zero to four scale, with higher scores indicating greater distress. The BCPT demonstrated good internal consistency in the current sample (Cronbach’s α=.88).

Statistical Analyses

We used the Statistical Package for the Social Sciences (version 28.0.1.0) first to calculate descriptive statistics for participant sociodemographic and clinical characteristics. We previously reported that patients in STRIDE showed significantly improved anxiety symptoms, QOL, and symptom distress, and improvements in depressive symptoms that approached significance, compared to patients in the MM control (see Table 4, Jacobs et al.).⁹ Therefore, we used mediated regression models via the PROCESS macro (version 4.3.1)²⁹ to test whether the effects of STRIDE on anxiety, depression, breast cancer specific QOL, and symptom distress at 24 weeks were mediated through changes either in coping ability or self-efficacy for managing AET symptoms. For each dependent variable, two mediators were tested: 1) 24-week change in coping ability and 2) 24-week change in symptom management self-efficacy. Each model was adjusted for whether the patient was also receiving ovarian suppression (due to baseline group differences), baseline distress (due to randomization stratification), baseline level of the respective mediator, and baseline level of the respective dependent variable. Therefore, each regression testing the effect of STRIDE on the mediating variable produced unique results depending on the covariates included. We modeled four regressions: the independent variable to the dependent variable (total effect), the independent variable to the mediator, the mediator to the dependent variable, and the independent variable to the dependent variable including the mediator (direct effect). We examined the indirect effect with bootstrapped 95% confidence intervals to determine whether the intervention effect on the dependent variable was mediated through either of the proposed mediators. Finally, we conducted a sensitivity analysis that controlled for how many months the patient had been taking AET to account for differences in these mediation models based on the time elapsed since AET initiation.

Results

Sociodemographic characteristics of the sample are presented in Table 1. Participants were women who were on average 56.1 years old (SD=10.94, range: 31, 81) and primarily identified as White (n=91, 91%) and non-Hispanic (n=94, 94%). Over three-quarters of participants were diagnosed with stage I breast cancer (n=77, 77%) and had been taking AET for an average of 17.91 months (SD=8.64). Most participants were employed full time/part time or a student (n=62, 62%), were married (n=73, 73%), and received a college (n=34, 34%) or master’s-level education (n=32, 32%).

Table 1.

Baseline Characteristics

Characteristic (N (%) or M (SD))	N = 100
Age, years; M (SD, range)	56.1 (10.94, 31–81)
Months since AET initiation (AET start to enrollment)	17.91 (8.64)
Race
Asian	4 (4)
Black or African American	1 (1)
White	91 (91)
Other	3 (3)
Not Reported	1 (1)
Ethnicity
Hispanic or Latino/a	3 (3)
Not Hispanic or Latino/a	94 (94)
Not Reported	3 (3)
Education
11th grade or less	1 (1)
High school graduate/GED	8 (8)
Some college/Technical School	16 (16)
College Graduate	34 (34)
Master’s Degree	32 (32)
Advanced Professional Degree	9 (9)
Household Income
$25,000–$49,999	7 (7)
$50,000–$99,999	22 (22)
$100,000–$149,999	18 (18)
> $150,000	50 (50)
Declined to respond/unknown	3 (3)
Employment Status
Full-time/part-time work or student	62 (62)
Caring for home or family	10 (10)
Unemployed	4 (4)
Not working due to illness/disability	1 (1)
Retired	18 (18)
Other or missing	5 (5)
Relationship Status
Married/Cohabitating	73 (73)
Non-cohabitating relationship	3 (3)
Single, never married	9 (9)
Divorced/separated	12 (12)
Loss of long-term partner/widowed	3 (3)
Breast Cancer Stage
Stage 0	8 (8)
Stage I	77 (77)
Stage II	11 (11)
Stage III	4 (4)
Type of AET
Aromatase Inhibitor	60 (60)
Tamoxifen	40 (40)
Primary Treatment Type
Surgery only	23 (23)
Radiation only	1 (1)
Surgery & Radiation	47 (47)
Surgery & Chemotherapy	8 (8)
Surgery, Chemotherapy, & Radiation	21 (21)
Ovarian Suppression
Receiving Ovarian Suppression	28 (28)
Not Receiving Ovarian Suppression	72 (72)

Note. AET = Adjuvant Endocrine Therapy, GED = General Educational Development Test. Advanced Professional Degree include MD, PhD, or JD degrees.

Changes in Coping as a Mediator

We tested whether changes in coping from baseline to 24 weeks mediated the effects of STRIDE on anxiety symptoms at the 24-week follow-up. As previously reported, participation in STRIDE was associated with improved coping ability (mediator) from baseline to 24 weeks, compared to MM patients. Next, changes in coping (mediator) from baseline to 24 weeks were significantly associated with reductions in anxiety at 24 weeks. While STRIDE was associated with significant reductions in anxiety, this relationship became less significant when change in coping was included as a mediator in the model; see Figure 1. In fact, a significant indirect effect through coping showed that the STRIDE intervention effect on anxiety was driven by improvements in patients’ coping ability (B=−0.61, SE=0.28, 95% CI: −1.28, −0.17; see Table 2). Of the total effect of STRIDE on anxiety, 40.3% can be attributed to improvements in coping ability.

Figure 1.

STRIDE Intervention Effects on Anxiety Symptoms Indirectly Through Coping Skills

Note: Controlled for baseline distress, ovarian suppression, anxiety, coping. HADS-A = Hospital Anxiety and Depression Scale – Anxiety, MOCS-A = Measurement of Current Status–Part A.

Table 2.

Mediation Effects of Change in Coping and Self-Efficacy

Δ Coping as Mediator
	B	SE	95% CI	p-value
Effect of Δ Coping on Anxiety Symptoms at 24 weeks (HADS-A)†
STRIDE on Δ coping	4.33	1.41	1.52, 7.14	.003
Δ coping on anxiety symptoms	−0.14	0.05	−0.23, −0.05	.003
Total effect	−1.52	0.62	−2.76, −0.28	.02
Direct effect	−0.91	0.63	−2.16, 0.34	.15
Indirect effect	−0.61*	0.28	−1.28, −0.17	--
Effect of Δ Coping on Depressive Symptoms at 24 weeks (HADS-D)†
STRIDE on Δ coping	4.36	1.39	1.60, 7.11	.002
Δ coping on depressive symptoms	−0.11	0.03	−0.18, −0.05	.001
Total effect	−0.90	0.45	−1.79, −0.01	.048
Direct effect	−0.40	0.45	−1.28, 0.48	.37
Indirect effect	−0.50*	0.21	−0.97, −0.15	--
Effect of Δ Coping on Breast Cancer QOL at 24 weeks (FACT-B)†
STRIDE on Δ coping	4.48	1.40	1.71, 7.25	.002
Δ coping on breast cancer QOL	0.85	0.19	0.47, 1.23	.001
Total effect	8.91	2.71	3.53, 14.29	.002
Direct effect	5.11	2.59	−0.05, 10.27	.052
Indirect effect	3.80*	1.25	1.54, 6.49	--
Effect of Δ Coping on Symptom Distress at 24 weeks (BCPT)†
STRIDE on Δ coping	4.33	1.41	1.53, 7.13	.003
Δ coping on symptom distress	0.03	0.12	−0.20, 0.26	.81
Total effect	−4.79	1.49	−7.74, −1.83	.002
Direct effect	−4.91	1.57	−8.04, −1.77	.003
Indirect effect	0.12	0.54	−1.08, 1.16	--
Δ Self-Efficacy as Mediator
	B	SE	95% CI	p-value
Effect of Δ Self-Efficacy on Anxiety Symptoms at 24 weeks (HADS-A)††
STRIDE on Δ self-efficacy	0.97	0.35	0.27, 1.67	.007
Δ self-efficacy on anxiety symptoms	−0.02	0.19	−0.41, 0.37	.92
Total effect	−1.47	0.63	−2.72, −0.22	.02
Direct effect	−1.45	0.66	−2.77, −0.14	.03
Indirect effect	−0.02	0.20	−0.46, 0.36	--
Effect of Δ Self-Efficacy on Depressive Symptoms at 24 weeks (HADS-D)††
STRIDE on Δ self-efficacy	0.98	0.35	0.29, 1.68	.006
Δ self-efficacy on depressive symptoms	−0.20	0.14	−0.48, 0.09	.17
Total effect	−0.85	0.46	−1.77, 0.07	.07
Direct effect	−0.66	0.48	−1.61, 0.30	.18
Indirect effect	−0.19	0.19	−0.60, 0.16	--
Effect of Δ Self-Efficacy on Breast Cancer QOL at 24 weeks (FACT-B)††
STRIDE on Δ self-efficacy	1.01	0.35	0.31, 1.71	005
Δ self-efficacy on breast cancer QOL	1.36	0.83	−0.30, 3.02	.11
Total effect	8.77	2.72	3.36, 14.19	.002
Direct effect	7.40	2.82	1.78, 13.01	.01
Effect of Δ Self-Efficacy on Breast Cancer QOL at 24 weeks (FACT-B)††
Indirect effect	1.38	1.07	−0.61, 3.71	--
Effect of Δ Self-Efficacy on Symptom Distress at 24 weeks (BCPT)††
STRIDE on Δ self-efficacy	0.96	0.35	0.26, 1.65	.008
Δ self-efficacy on symptom distress	−0.80	0.46	−1.71, 0.11,	.09
Total effect	−4.74	1.50	−7.71, −1.77	.002
Direct effect	−3.97	1.54	−7.04, −0.91	.01
Indirect effect	−0.77	0.56	−2.10, 0.06	--

Note: Δ = change;

^*p <.05;

^†= controlled for baseline distress, ovarian suppression, dependent variable, and coping;

^††= controlled for baseline distress, ovarian suppression, dependent variable, and self-efficacy. BCPT = Breast Cancer Symptom Distress, FACT-B = Functional Assessment of Cancer Therapy – Breast, HADS-A = Hospital Anxiety and Depression Scale – Anxiety, HADS-D = Hospital Anxiety and Depression Scale – Depression, QOL = quality of life.

Since intervention-related improvements in depressive symptoms approached significance in our RCT,⁹ we next tested whether changes in coping from baseline to 24 weeks mediated the effect of the STRIDE intervention on depressive symptoms. As previously reported, participation in STRIDE was associated with improved coping ability from baseline to 24 weeks, compared to MM patients, with the respective covariates. Next, changes in coping were significantly associated with reductions in depressive symptoms at 24 weeks. STRIDE was associated with reductions in depressive symptoms that approached significance, and this effect became less significant when coping was included in the model; see Figure 2. A significant indirect effect showed that the intervention effect on depressive symptoms was driven by improvements in coping (B=−0.50, SE=0.21, 95% CI: −0.97, −0.15; see Table 2). Of the total effect of STRIDE on depressive symptoms, 55.5% can be attributed to improvements in coping ability.

Figure 2.

STRIDE Intervention Effects on Depressive Symptoms Indirectly Through Coping Skills

Note: Controlled for baseline distress, ovarian suppression, depressive symptoms, coping. HADS-D = Hospital Anxiety and Depression Scale – Depression, MOCS-A = Measurement of Current Status–Part A.

We also tested whether changes in coping from baseline to 24 weeks mediated the effect of the STRIDE intervention on breast cancer specific QOL on the FACT-B. Again, patients assigned to STRIDE reported improved coping ability from baseline to 24 weeks compared to MM control, with the respective covariates. Next, changes in coping were significantly associated with improvements in QOL. As was previously established, STRIDE was also associated with significant improvements in QOL, and this effect became less significant when coping was included as a mediator in the model; see Figure 3. There was a significant indirect effect through coping, such that the intervention effect on QOL was driven by improvements in patients’ coping ability (B=3.80, SE=1.25, 95% CI: 1.54, 6.49; see Table 2). Of the total effect of STRIDE on breast cancer QOL, 42.6% can be attributed to improvements in coping ability.

Figure 3.

STRIDE Intervention Effects on Breast Cancer QOL Indirectly Through Change in Coping Skills

Note: Controlled for baseline distress, ovarian suppression, breast cancer QOL, coping. FACT-B = Functional Assessment of Cancer Therapy – Breast, MOCS-A = Measurement of Current Status–Part A.

Finally, we tested whether changes in coping from baseline to 24 weeks mediated the effect of the STRIDE intervention on breast cancer symptom distress on the BCPT. Patients assigned to STRIDE reported improved coping from baseline to 24 weeks, compared to MM control, with the respective covariates. However, the changes in coping were not associated with breast cancer symptom distress at 24 weeks, and the indirect effect was not significant (B=0.12, SE=0.54, 95% CI: −1.08, 1.16; see Table 2).

Changes in Self-Efficacy as a Mediator

We followed this same process to test whether the intervention effect on anxiety was mediated by improvements in self-efficacy for managing AET symptoms. We first re-established that those patients assigned to STRIDE reported improved self-efficacy (mediator) from baseline to 24 weeks, compared to MM, with the respective covariates. However, change in self-efficacy from baseline to 24 weeks was not significantly associated with reductions in anxiety at 24 weeks, and the indirect effect was not significant (B=−0.02, SE=0.20, 95% CI: −0.46, 0.36; see Table 2).

We similarly tested whether the intervention effect on depressive symptoms was mediated by improvements in self-efficacy. As reported previously, patients assigned to STRIDE demonstrated improved self-efficacy (mediator) from baseline to 24 weeks, compared to MM. However, changes in self-efficacy were not associated with reductions in depressive symptoms at 24 weeks. Further, the indirect effect through changes in self-efficacy was not significant (B=−0.19, SE=0.19, 95% CI: −0.60, 0.16; see Table 2).

We next tested whether the intervention effect on breast cancer specific QOL was mediated by improvements in self-efficacy. We again re-established that those patients assigned to STRIDE reported improved self-efficacy from baseline to 24 weeks, compared to MM control. However, changes in self-efficacy were not associated with improvements in breast cancer specific QOL at 24 weeks, and the indirect effect through change in self-efficacy was not significant (B=1.38, SE=1.07, 95% CI: −0.61, 3.71; see Table 2).

Similarly, we tested whether changes in self-efficacy from baseline to 24 weeks mediated the effect of the STRIDE intervention on breast cancer symptom distress on the BCPT, which was observed. However, changes in self-efficacy were not associated with breast cancer symptom distress at 24 weeks, and the indirect effect through change in self-efficacy was not significant (B=−0.77, SE=0.56, 95% CI: −2.10, 0.06; see Table 2).

In our sensitivity analysis adjusting for time taking AET (months since AET initiation), we observed no differences in any of the mediation models, such that the findings were the same regardless of how long a patient had been taking AET. To retain the most parsimonious model, we did not include this variable in our final results.

Discussion

This pre-specified exploratory analysis demonstrated that STRIDE-related improvements in patients’ coping abilities were the primary drivers of the intervention effect on mood and QOL outcomes for patients with non-metastatic hormone receptor-positive breast cancer taking AET. Specifically, we found that changes in coping across the 24-week study period mediated the effect of STRIDE on anxiety and depressive symptoms as well as breast cancer-related QOL. Changes in coping did not drive STRIDE-related improvements in symptom distress. We also found that 24-week changes in self-efficacy did not mediate the relationships between STRIDE and any of the four outcome measures, despite patients in STRIDE reporting improved self-efficacy, compared to the MM control.⁹

This study identified coping ability as a significant mediator of STRIDE’s effect on anxiety symptoms. Approximately half of women with breast cancer experience significant fear of cancer recurrence,³⁰ which can increase at different periods across survivorship.³¹ The relationship between relaxation-focused coping and anxiety has been corroborated in other studies, such as in survivors of breast cancer experiencing menopausal symptoms²² and more broadly in patients with varying cancer types and stages.³² STRIDE modules focus on relaxation training, including diaphragmatic breathing and progressive muscle relaxation, as well as strategies to schedule worry time to decrease rumination about recurrence. These evidence-based, relaxation-focused coping skills,³³ as measured by the MOCS-A,²⁵ equipped participants with tangible tools to reduce their experience of anxiety and minimize the intensity of physical anxiety symptoms, such as tightness and tension.⁹

Improved coping ability in STRIDE participants also drove reductions in depressive symptoms. Interestingly, a recent study showed that patients with breast cancer who reported depressive symptoms were likely to employ maladaptive coping strategies, such as avoidance.³⁴ Content in STRIDE offers more adaptive coping alternatives to avoidance or withdrawal, such as cognitive reframing to challenge overgeneralized thoughts and worries, and acceptance.⁹ In fact, a systematic review on psychological adjustment after breast cancer showed that patients who used evidence-based coping skills, such as cognitive restructuring and problem solving, reported fewer depressive symptoms.³⁰ STRIDE also teaches behavioral activation, a first-line treatment for moderate depression in cancer according to the American Society of Clinical Oncology.³⁵ It is likely that patients in STRIDE learned to employ both cognitive and behavioral strategies to cope with difficult thoughts and situations, leading to a reduction in depressive symptoms, such as sadness, anhedonia, and feelings of guilt and/or worthlessness.

Equipping patients with the tools to cope with their cancer and treatment-related experiences drove improvements in QOL as well. Other emotion-focused coping strategies have been positively associated with QOL in patients with breast cancer,³⁶ and those taking AET have described the importance of knowing how to cope with symptoms to improve QOL.⁴ STRIDE delivers problem-focused and emotion-focused coping strategies to improve breast cancer related QOL through management of disruptive symptoms such as hot flashes, fatigue, pain, sleep disturbance, and uncertainty related to taking AET.⁹ The intervention also teaches several non-pharmacological approaches to managing these symptoms such as stimulus control, time-based activity pacing, and relaxation training.

While STRIDE participants did report lower levels of breast cancer symptom distress in the overall trial,⁹ changes in coping did not mediate this finding. This observation may be due to differences in measurement. While the MOCS-A measures broad coping strategies taught by the STRIDE intervention, the BCPT measures distress related to very specific AET symptoms. It may be that the STRIDE-related improvements in breast cancer symptom distress are not fully captured by the MOCS-A coping measure and a different intervention component may underly this effect. For example, STRIDE teaches strategies to enhance interactions and communication with healthcare providers, and patient-provider relationship satisfaction may be a driver of intervention-related improvements in symptoms measured on the BCPT.

While we did not find that self-efficacy for managing symptoms was a mediator of the effect of STRIDE on the proposed outcomes, other studies have documented such associations. For example, lower self-efficacy to manage the sexual symptoms of AET was associated with poorer QOL in postmenopausal women.³⁷ The self-efficacy measure in this study (SESM-AET) asks about participants’ management of eight specific AET-related symptoms, which is not exhaustive of all symptoms that patients may be experiencing and does not refer to general self-efficacy as a broader construct. Therefore, the finding that self-efficacy was not a mediator may be unique to this self-efficacy measure and the eight symptoms it refers to. It seems that in the current study, coping, rather than self-efficacy, was a significant driver of these positive changes in anxiety symptoms, depressive symptoms, and QOL in our sample. In contrast, the MOCS-A used in this study is broad, not symptom-specific, and focused more on overall psychological distress, which may help to explain why changes in coping ability mediated improvements in psychological symptoms, as compared to self-efficacy for managing eight specific symptoms. Importantly, despite the improvements in self-efficacy seen in STRIDE, symptoms persisted, suggesting that a patient’s mood might still be affected by symptoms even when feeling more efficacious to manage them.

Research Implications

The current mediation analysis is supported by strong theoretical frameworks, the Cognitive Model for Menopausal Symptoms¹² and Murray’s Framework for Medication Adherence,¹³ that justify coping and self-efficacy as intervention targets and mediators of the effects on patient-reported outcomes. Mediation analyses are vital for understanding how the intervention produces the changes observed and what key factors might sustain long-term intervention effects. However, many mediation analyses of psychosocial interventions are never conducted, and proposed mediators are rarely identified by way of a theoretical framework. In fact, coping is a common mechanism of action in psychosocial interventions in cancer but is not always formally tested as such.¹¹ Importantly, mediation analyses allow the testing of intervention models across contexts and diagnoses to understand how we can best provide clinical care.³⁸ For example, the current intervention efficacy may be enhanced by refining the intervention to include even more coping strategies. A briefer intervention could also be tested in minimally resourced clinical care settings and could be equally efficacious, assuming it retains these primary coping strategies.

Clinical Implications

Women with non-metastatic hormone receptor-positive breast cancer taking AET experience clinically significant impairments to their well-being due to symptoms related to taking AET.⁴ As non-adherence to AET is well-documented in this population,³ identifying clinical mechanisms that can improve the experience of taking AET are vital to improving patient care. The cognitive, behavioral, and relaxation skills taught in STRIDE⁸ likely enhanced patients’ ability to cope with AET symptoms and improved their mood and QOL, carving a path towards enhancing the experience of taking AET.

Study Limitations

Limitations of the current study include the homogenous sample comprised mainly of non-Hispanic White patients with non-metastatic disease at an academic medical center in the northeast region of the United States. Further inquiry using a more racially, ethnically, and geographically diverse sample would strengthen the generalizability of results. Given that patients of a racial or ethnic minority group are at greater risk for AET symptoms⁴ and non-adherence,³⁹ it is imperative to test the mediating role of coping to improve mood and QOL in a sample that represents patients most at risk of developing this symptom profile. Coping ability was measured using the total score of the MOCS-A, and future research may examine the subscales of this measure to identify which specific coping strategies (e.g., relaxation, cognitive reframing) may drive improvements in anxiety, depression, and breast cancer QOL. Further, participants self-report of their own coping and self-efficacy may introduce bias into the results. Given that STRIDE is a coping skills-based intervention, the MOCS-A may reflect a more direct assessment of the skills participants acquired, as opposed to their perceptions for managing specific symptoms assessed by the SESM-AET, which could additionally help to explain the current findings. Future studies using ecological momentary assessments may help to capture actual use of the skills acquired and practiced through participation in STRIDE. This type of assessment may offer increased accuracy about actual strategies used in response to symptoms as they arise, as compared to recall at the time of the self-report assessment which may take place days or weeks later.

This exploratory analysis, in the context of a pilot randomized controlled trial, is hypothesis-generating. Though this pilot trial evidenced adequate effect sizes (e.g., d=.48 for the effect of STRIDE on coping ability),⁹ these results are preliminary and would benefit from testing in a larger, more representative sample.

Conclusion

Guidelines from the American Society for Clinical Oncology recommend AET for up to 10 years⁴⁰ which highlights the need to identify and leverage modifiable factors that improve patients’ psychological and emotional experience during AET. Importantly, anxiety and depressive symptoms, QOL, symptom distress, coping with symptoms and self-efficacy are all well known to contribute to AET non-adherence.^5,6 The current study demonstrates that enhancing patients’ ability to cope with stress and symptoms by way of a brief intervention can improve mood and QOL. These mediation findings provide insights for how to sustain long-term effects and optimize adherence interventions for patients with breast cancer taking AET.⁹

Data availability statement:

Data is available upon reasonable request from the authors.