A Single-Session Acceptance and Commitment Therapy Intervention among Women Undergoing Surgery for Breast Cancer: A Randomized Pilot Trial to Reduce Persistent Postsurgical Pain

Abstract

Objective.

Oncologic breast surgeries carry a risk for persistent postsurgical pain. This study was a randomized pilot and feasibility study of a single-session Acceptance and Commitment Therapy (ACT) intervention compared to treatment as usual among women undergoing surgery for breast cancer or ductal carcinoma in situ (DCIS).

Methods.

Participants were recruited via letter of invitation and follow-up phone call from a single site in the United States from 2015–2017. Participants were at risk for persistent post-surgical pain, based on: young age (<50), a pre-existing chronic pain condition, or elevated anxiety, depression, or pain catastrophizing.

Results.

The 54 participants were female with a mean age of 52.91 years (SD=11.80). At 3-months post-surgery, 11% of the sample reported moderate-severe pain (>3 on a 0–10 numeric rating scale) in the operative breast or with arm movement. Written qualitative responses indicated that the majority of participants who received the intervention understood the concepts presented and reported continued practice of exercises learned in the session. The between group effect sizes for moderate-severe pain and elevated anxiety at 3-months post-surgery were small (Phi=0.08 and 0.16, respectively). The between group effect sizes for depression, pain acceptance, and pain catastrophizing at 3-months post-surgery were minimal.

Conclusions.

This study found small positive effects on post-surgical pain and anxiety for a single-session ACT intervention among women with breast cancer. This study supports the use of ACT with this population.

Introduction

Breast cancer is the second most common cancer in American women, affecting about 1 in 8 (12%) in their lifetime [1]. Nearly 30% of women have persistent pain following breast cancer surgeries [2], with 14% reporting moderate to severe pain [3]. Persistent pain following breast cancer surgeries is associated with significant disability and distress [4] and emotional distress in cancer patients is associated with poorer compliance, higher risk of mortality, and poorer quality of life [6].

Several factors confer risk for persistent postsurgical pain in this population. Preoperative psychological distress (depressive and anxious symptoms as well as pain catastrophizing) has emerged as an important risk factor [7–9]. Of note, the relation between pain and depression can be bi-directional, such that increased pain is also associated with increased depression [10]. Co-morbid chronic pain also conveys risk for persistent postsurgical pain in multiple surgical populations [11] and younger age is a risk factor for both persistent post-surgical pain and depression among women undergoing breast cancer surgery [3, 7, 8][5, 12]. Other factors associated with depression and anxiety among women undergoing treatment for breast cancer include: not being married, children living in the home, having to travel long distances for treatment, lower income, stage of cancer, religion, and high symptom-burden [5, 12].

Psychological interventions addressing depression, anxiety, and psychological approaches to pain management in surgical patients may serve to prevent the development of persistent post-surgical pain, depression, and anxiety in at-risk individuals. Cognitive behavioral therapy and relaxation therapy have demonstrated positive effects in pain reduction following orthopedic, cardiac, and colorectal surgeries. However, the difference in pain between intervention and non-intervention groups was about 1 (on average) on a 0–10 scale [13]. A meta-analysis of psychological studies during the perioperative period found that single-session educational interventions demonstrated the greatest patient adherence; however, education only sessions did not improve post-surgical pain outcomes [13]. The current study proposed to pilot a single session intervention, to increase adherence, while offering an active psychotherapy: Acceptance and Commitment Therapy (ACT).

ACT is a contextual behavioral therapy based on functional contextualism [14] and the science of Relational Frame Theory [15]. ACT aims to increase psychological flexibility by cultivating psychological acceptance and committed action in the direction of one’s personally identified values [16]. ACT includes six core processes: Acceptance, cognitive defusion, being present, self-as-context, values, and committed action [16]. Acceptance in this context refers to an active willingness to come fully into contact with private events (thoughts, emotions, physical sensations, etc), without attempting to change their frequency or form, and acceptance is in service of increasing values-based actions [16, 18]. Increased pain acceptance is associated with improved functional outcomes among patients with chronic pain [17].

ACT is an empirically supported treatment for minimizing the impairing impact of chronic pain [19], depression and anxiety [20]. Among breast cancer treatment completers, a pilot 8 session ACT group intervention decreased depressive and anxious symptoms [21]. Another feasibility trial of a 6-session telephone-based ACT intervention for symptom interference among patients with metastatic breast cancer found reductions in fatigue and sleep interference [22]. These findings indicate the potential utility of ACT among breast cancer patients.

The current study sought to trial a single-session ACT intervention that could feasibly be completed in conjunction with patients’ post-surgical follow-up appointment. Single session ACT interventions are gaining empirical support [23–26], including a recent case series of a one-hour individual ACT session to promote health behavior change [27]. The current study designed an intervention that is more intensive than the one-hour health behavior change intervention (tailored to a breast cancer population), but still brief enough to add-on to an extant clinic appointment.

This study aims to investigate the feasibility and acceptability of a single-session ACT intervention following surgeries for breast cancer or ductal carcinoma in situ (DCIS). This study also aims to examine preliminary effects, compared to a non-intervention group, on persistent post-surgical pain, anxiety, depression, pain catastrophizing, and pain acceptance at 3-months post-surgery.

Methods

Design.

This study was a randomized pilot feasibility trial of a single session ACT intervention among women undergoing surgery for breast cancer. The study is registered at clinicaltrials.gov: NCT03430765. Consistent with recommended pilot sample sizes derived from commonly-used approaches for pilot and feasibility trials [28–30], the target sample size was N=60 (30 per group). An attrition rate of 20% was anticipated so the targeted enrollment was 72 participants.

Participants.

Following local institutional review board approval (IRB # 201502709), patients were recruited from the Holden Comprehensive Cancer Center breast clinic in Iowa, United States, between 2015–2017. All participants completed the consenting process. Inclusion criteria were: 1) ≥18 years of age, 2) scheduled for mastectomy or lumpectomy for breast cancer or DCIS, 3) identified to be at risk for persistent post-surgical pain. Participants were determined to be “at risk” for persistent postsurgical pain if they met any of the following criteria: under the age of 50, having a pre-existing chronic pain condition, or elevated anxiety (≥10 on the Generalized Anxiety Disorders 7-item scale [31]), depression (≥10 on the Personal Health Questionnaire Depression 8-item Scale [32]), or pain catastrophizing (≥30 on the Pain Catastrophizing Scale [33]). Exclusion criteria included: 1) language or cognitive barriers preventing completion of questionnaires, 2) active severe psychiatric disorder: bipolar or psychotic disorders, 3) significant surgical complications. Language and cognitive barriers were self-report or as indicated in the medical record. Severity of psychiatric disorder was determined based on active bipolar or psychotic disorder, per medical record review or patient report.

Data Collection Protocol.

Potential participants were sent a letter about the study and contacted by phone to discuss interest in participating. Interested participants were screened for eligibility. Those who were eligible were consented and then randomized to either treatment as usual (standard breast cancer care) or the ACT intervention plus standard care, by a research team member. The randomization sequence was in permuted blocks of 4 and 6. Randomization was stratified by surgery type (mastectomy and lumpectomy). Following randomization and prior to surgery, participants completed baseline survey measures (described below) and the same surveys were completed again 3-months following surgery. One-week post-surgery, participants were also asked to complete a pain assessment. Additionally, those randomized to ACT were asked to complete written qualitative responses regarding the ACT intervention. All data were collected via Research Electronic Data Capture (REDCap)[34].

Intervention Arm.

Participants randomized to the intervention arm received a 2-hour individual therapy session two weeks after surgery (on the day of their surgical follow-up) based on ACT. The session was delivered by either a clinical psychologist or an advanced counseling psychology doctoral student, under the supervision of the psychologist. The therapist followed a manualized format for the session (outlined in supplemental file 1) and each participant received a workbook of handouts. The session covered: a) check-in about experiences in treatment, b) values clarification; c) brief mindfulness exercise; d) differentiating between private events (thoughts, emotions, physical sensations) and behaviors, e) acceptance and willingness exercises, f) cognitive defusion exercises; and g) committed action (including goal-setting) [16]. After each session, the therapist completed a fidelity checklist to ensure each of these domains was covered (which they were with each participant).

Treatment as Usual.

All participants received medical care as indicated based on a combination of their pathology results, staging, hormone receptor status, genetic risk, discussion at tumor board, and stated patient preferences. Treatment involved surgery for all participants, radiation for most who underwent lumpectomy, chemotherapy as indicated, and frequently hormone therapy.

Measures.

Usability Questions.

Participants who completed the ACT intervention were asked to highlight aspects of the intervention that they found helpful and unhelpful and were asked whether they continued to use anything that they learned in the ACT session. Participants provided written responses to these survey questions (Supplemental file 2).

Pain.

Pain intensity was measured using a numeric rating scale (0–10 NRS), including half points, where 0 represents “no pain” and 10 represented “pain as bad as you can imagine.” Participants were asked to rate pain in their breast and pain associated with holding their arms in front of themselves and lifting them above their head. A cut-off of ≤3 was used to determine no worse than mild pain as a clinically meaningful end point [35]. General pain assessment (not specific to surgery) was measured with the Brief Pain Inventory (BPI) Short Form. The BPI is a widely used 15-item measure with established validity, which includes 4 items assessing pain severity [36, 37]. Internal consistency for the four pain severity items on this measure was good (preoperative α=0.91; 3-months post-surgery α=0.93).

Depression.

The Personal Health Questionnaire Depression Scale (PHQ-8) was used to assess for depressive symptoms. The PHQ-8 is based on diagnostic criteria for Major Depressive Disorder in the Diagnostic and Statistical Manual of Mental Disorders–Fourth Edition (DSM-IV). This measure has demonstrated validity and reliability for assessing severity of depressive symptoms [32]. Internal consistency in the current study was good (preoperative α=0.80; 3-months post-surgery α=0.85). A cut-off of 10 was used on this measure to indicate moderate to severe depressive symptoms; this cut-point has demonstrated optimized sensitivity and specificity in identifying clinical cases of major depressive disorder [32].

Anxiety.

The Generalized Anxiety Disorder Scale (GAD-7) was used to assess for symptoms of anxiety. The GAD-7 is a 7-item self-report measure designed to assess Generalized Anxiety Disorder. Scores can range between 0–30. This widely used measure has demonstrated reliability and validity as a measure of anxiety in the general population [31]. Internal consistency in the current study was good (preoperative α=0.85; 3-months post-surgery α=0.86). A cut-off of 10 was used on this measure to indicate elevated levels of anxiety, as this cut-off has demonstrated optimized sensitivity and specificity in identifying clinical cases of Generalized Anxiety Disorder [31].

Pain Catastrophizing.

The Pain Catastrophizing Scale (PCS) is a 13-item self-report measure that produces a total score and three subscale scores: Rumination, Magnification, and Helplessness. Internal consistency in the current study was good (preoperative α=0.91; 3-months post-surgery α=0.93). A cut-off of 20 was used to indicate elevated levels of pain catastrophizing (i.e., above the 50^th percentile compared to patients with chronic pain) [33].

Pain Acceptance.

Pain acceptance was measured with the Chronic Pain Acceptance Questionnaire 20-item short-form (CPAQ-20). This is a 20-item self-report measure assessing participants’ attitudes toward pain. Internal consistency in the current study was good (preoperative α=0.87; 3-months post-surgery α=0.87). This measure has demonstrated reliability and validity [38].

Data Analyses.

Qualitative data were analyzed for thematic content using deductive content analysis [39]. Analyses were completed by the principle investigator who organized responses into thematic categories (broadly positive and negative responses to the intervention) and sub-categories (responses to specific exercises and aspects of the treatment structure).

Quantitative data were checked for normality and descriptive statistics were run. On multi-item measures, surveys with at least 85% complete data were used and mean imputations were calculated for missing items on these measures. The pain, depression, anxiety, and pain catastrophizing instruments offered clear cut-points for clinical significance, so these measures were divided into categorical variables. Pain acceptance was kept as a continuous variable. Means or medians for demographic and clinical measures are presented at baseline, one-week post-surgery (for pain) and 3-months post-treatment for the full group and by treatment arm (ACT vs Treatment as Usual: TAU). Differences by treatment arm at all time-points were examined via Chi-Squared, t-test, or Mann-Whitney U tests as appropriate. Effect size differences on the outcome variables between the ACT and TAU arms were calculated using Phi (where 0.1 is considered a small effect size, 0.3 medium, and 0.5 large) for all categorical variables (pain, depression, anxiety, and pain catastrophizing), and Cohen’s d (where 0.2 is considered small, 0.5 medium, and 0.8 large) for the continuous variable (pain acceptance).

Data Availability Statement.

The data that support the findings of this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy restrictions.

Results

Between June 2015 and August 2017, a total of 264 women were scheduled for mastectomy or lumpectomy to treat breast cancer or DCIS. We were unable to contact 59 (22%) patients prior to surgery. Of the remaining 205, a further 12 (5%) were excluded (4 due to self-reported cognitive limitations; 1 cancelled surgery; 3 had significant comorbid psychiatric disorders including 2 with bipolar disorder and 1 with Schizophrenia; and 4 were getting treatment for additional cancers). Of the 193 potentially eligible participants (73%), a further 80 (30%) declined participation. Of the patients who declined participation we were able to solicit specific reasons beyond “I do not want to participate” from 58/80. Reasons for non-participation fell into a few broad categories: A sense of general overwhelm/not having time/”too much on my plate” (26 patients); not wanting to provide personal information (7 patients); concerns about travel or having drivers wait for them during the intervention (6 patients); not wanting to spend two hours in an intervention (8 patients); and being specifically disinterested in the intervention itself (not wanting to talk about stressors, not being stressed, aversion to anything that includes mindfulness, perceiving that the intervention would not be helpful to them: 11 patients). The remaining 113 (43%) potential participants expressed interest in participating, but a further 51 (19%) did not meet the “at risk” screening criteria. The remaining 62 (23%) individuals were consented and randomized (31 to TAU and 31 to ACT). Of those randomized, 30 in the TAU arm completed the study (1 withdrew). Of those randomized to ACT, 24 completed the ACT intervention and the study (2 withdrew, 3 were excluded due to delayed surgery or surgical complications, and 2 did not receive the intervention). The 2 participants who withdrew from the study in the ACT arm both did so prior to the ACT intervention. See Figure 1 for the CONSORT diagram.

Figure 1.

CONSORT diagram

The 54 participants were female with a mean age of 52.91 years (SD = 11.80). Over half (56%) underwent lumpectomy and 44% mastectomy. Thirty-seven percent of the sample had simultaneous reconstruction at the time of mastectomy and 19% underwent auxiliary lymph node dissection. Participants met “at risk” inclusion criteria based on: 20% screened positive for depression, 30% for anxiety, 17% for pain catastrophizing, 57% based on comorbid chronic pain, and 48% based on age <50 years (participants could screen positive on more than one criterion). There were no significant differences between treatment and control arFmarriedms on any baseline demographic, surgical, pain, or psychological variables (Table 1).

Table 1.

Baseline Demographics and Clinical Characteristics.

	Total (N = 54)	ACT (n = 24)	TAU (n = 30)	Difference
Age (Mean, SD)	52.91 (11.80)	55.63 (11.22)	50.66 (11.99)	p=0.13
Income (%)
< 40,000	17%	23%	14%	p=0.20
40,000 – 79,999	24%	23%	24%
≥ 80,000	59%	55%	62%
Ethnicity
Hispanic or Latino	0%	0%	0%	p=0.38
Not Hispanic or Latino	93%	92%	93%
Unknown	7%	8%	6%
Racial Identification
Asian	4%	4%	3%
African American	2%	4%	0%	p=0.47
Caucasian	87%	83%	90%
Unknown	7%	9%	6%
Surgery Type
Mastectomy	44%	37.5%	50%	p=0.36
Lumpectomy	56%	62.5%	50%
Preoperative
Breast Pain (Median, IQR)	0.00 (0.00–0.50)	0.00 (0.00–0.50)	0.00 (0.00–0.00)	p=0.08
Depression (Median, IQR)	4.50 (2.82–8.00)	4.00 (2.07–7.75)	5.00 (3.00–8.00)	p=0.27
Anxiety (Median, IQR)	5.00 (2.00–8.00)	4.00 (2.00–6.75)	5.50 (2.75–9.00)	p=0.22
Pain Catastrophizing (Median, IQR)	10.00 (3.00–14.04)	11.00 (1.08–14.00)	9.50 (3.75–14.81)	p=0.62
Pain Acceptance (Mean, SD)	75.44 (15.94)	73.80 (16.75)	76.54 (15.59)	p=0.58

Preoperative Pain and One-Week Postoperative Pain.

Moderate–severe pre-operative breast pain (>3 on a 0–10 scale) was reported by only 2% of the sample prior to surgery. The average generalized pain score (on the BPI item ‘average pain’) was 2.19 (SD = 2.09). Twenty-seven percent of the sample reported moderate-severe generalized pain pre-operatively.

At one-week post-surgery, 48% of the sample reported moderate-severe post-operative pain. This variable was not significantly different between the ACT (55%) and TAU (42%) treatment arms. Median one-week post-operative breast pain was 2.00 (IQR: 1.00–3.00) and median postoperative movement pain was 3.00 (IQR: 1.00–5.00), neither were significantly different between ACT and TAU groups (Supplemental File 3).

Qualitative Feedback.

Written qualitative responses indicated that the majority of participants who received the intervention understood the concepts presented and reported continued practice of exercises learned in the session. Two participants did not indicate finding the session helpful: one noted concerns about keeping family waiting and another commented that she was “doing fine” and “didn’t need the session”. The brief mindfulness exercise was noted by many participants as being helpful (e.g., “I have found the meditation to be really helpful”), although 2 participants specifically noted that they did not like it. Values work was also frequently identified as being useful, although not specifically by name (e.g., “It made me realize how much of my life is about family and friends”). Many participants noted continued use of mindfulness after the session, a couple reported continued use of goal setting, and a couple reported using willingness in relation to emotions and cognitive defusion. Multiple participants also noted finding it useful to talk about their experiences and emotions related to cancer treatment. The majority of the respondents also reported that they were fine with the length and pace of the sessions, with multiple participants noting that the time went by more quickly than expected. A couple participants stated preference for longer, a couple for shorter, and one for multiple sessions. Most responded that the pace was good with a couple participants wanting a faster, and a couple wanting a slower, pace.

Intervention Effect at Three Months

Pain.

A total of 11% of the sample reported moderate-severe pain (>3 on a 0–10 scale) in the operative breast or with arm movement at 3-months follow-up. In the ACT group, 8.3% of the sample reported moderate to severe arm or movement pain (2/24) compared to 13.3% in the TAU group (4/30). See Figure 2. The effect size was small: Phi = 0.08.

Figure 2.

Breast and movement pain severity at 3-months post-surgery.

Anxiety and Depression.

A total of 9.26% of participants screened positive for anxiety (≥10 on the GAD-7) and 14.81% screened positive for depression (≥10 on the PHQ-8) at 3-months follow-up. In the ACT arm, 4.2% of the sample reported elevated anxiety (1/24) compared to13.3% (4/30) in TAU. The effect size for difference between groups in anxiety was small Phi = 0.16. In the ACT group, 12.5% reported elevated depression (3/24) compared to 13.3% (4/30) in TAU. The effect size for difference between groups in depression was minimal Phi = 0.01.

Pain Catastrophizing.

At 3-months follow-up, 3.7% of participants screened positive (above the 50^th percentile) for pain catastrophizing (>20 on the PCS). In the ACT arm 4.2% (1/24) reported elevated pain catastrophizing and 3.3% (1/30) in TAU. The effect size was minimal Phi=0.02.

Pain Acceptance.

Mean pain acceptance score was 82.39 (SD=17.0) for the full sample (ACT M=83.35, SD=16.89; TAU M=81.69, SD=17.34). The effect size for difference between groups on pain catastrophizing was minimal: Cohen’s d=0.10.

Discussion

This pilot study found that the majority of women undergoing surgery for breast cancer or DCIS found a single session ACT intervention useful and continued to use content they had learned in the session. Participants particularly valued the mindfulness practice and values orientation work. By 3-months follow-up, small positive effects for treatment were found on post-operative pain and anxiety.

The overall levels of persistent pain in the current sample (11% moderate-severe pain) were smaller than reported in previous literature, including a recent systematic review of 3,746 patients reporting a prevalence of 29.8% [2]. This suggests that the current study may not have captured the most “at risk” patients. The current sample was also not highly distressed on average, despite inclusion criteria necessitating risk for persistent post-surgical pain, which included distress-based criteria. Average baseline depression, anxiety, and pain catastrophizing scores were well below clinical cut-offs indicating elevated symptoms [31–33], suggesting some restriction of range in this sample. In addition, baseline pain acceptance found in this study was higher than previously reported in both an orthopedic surgery sample [23] and in samples of patients with chronic pain [17, 40], suggesting good pain coping at baseline. Further, one-third of potential participants who declined participation noted a general sense of overwhelm as a reason for non-participation, suggesting that some participants who may have benefited most from the intervention declined to participate. Creative approaches to increase ease of participation, such as offering a telehealth intervention may increase participation among the most highly distressed patients.

Study Limitations.

The sample in the current study was comprised of breast cancer patients, which limits generalizability to other disease groups. In addition, to increase intervention effects, a more intensive intervention may be needed in this population. Another limitation is the use of mean imputation on multi-item measures with at least 85% complete, due to potentially reducing variance in these variables. Finally, a significant limitation to this study was that the sample was largely Caucasian and higher income, so findings may not generalize to other patient populations.

Research and Clinical Implications.

A randomized control trial of an ACT intervention to prevent persistent post-surgical pain is warranted in a larger more diverse sample. The single-session intervention trialed in this pilot feasibility study could be expanded to multiple sessions, including a pre-operative session to deepen patient understanding and to support continued practice of the skills learned, which may translate to larger treatment effects. Additionally, switching the format to telehealth could increase accessibility and feasibility and may even increase participation among the most distressed patients. Clinically, this study aligns with the broader literature supporting the use of ACT with this population.

Conclusion

This study found small positive effects on post-surgical pain and anxiety for a single-session ACT intervention among women undergoing surgery for breast cancer or DCIS. Qualitative feedback suggested applicability of ACT in this population. The subsequent larger trial of this pilot could include a more intensive, but still accessible version of the ACT intervention.