Effect of pain neuroscience education after breast cancer surgery on pain, physical, and emotional functioning: a double-blinded randomized controlled trial (EduCan trial)

Abstract

Pain is one of the most common and long-lasting side effects reported by women surgically treated for breast cancer. Educational interventions may optimize the current physical therapy modalities for pain prevention or relief in this population. Pain neuroscience education (PNE) is an educational intervention that explains the pain experience not only from a biomedical perspective but also the psychological and social factors that contribute to it. Through a double-blinded randomized controlled trial (EduCan trial) it was investigated if PNE, in addition to the standard physiotherapy program immediately after breast cancer surgery, was more effective over the course of 18 months postoperatively than providing a biomedical explanation for pain. Primary outcome was the change in pain-related disability (Pain Disability Index, 0-70) over 12 months. Secondary outcomes included change in pain intensity, upper limb function, physical activity level, and emotional functioning over 4, 6, 8, 12, and 18 months postoperatively. Multivariate linear models for repeated (longitudinal) measures were used to compare changes. Preoperative and postoperative moderators of the change in pain-related disability were also explored. Of 184 participants randomized, the mean (SD) age in the PNE and biomedical education group was 55.4 (11.5) and 55.2 (11.4) years, respectively. The change in pain-related disability from baseline to 12 months postoperatively did not differ between the 2 groups (PNE 4.22 [95% confidence interval [CI]: 1.40-7.03], biomedical 5.53 [95% CI: 2.74-8.32], difference in change −1.31 [95% CI: −5.28 to 2.65], P = 0.516). Similar results were observed for all secondary outcomes. Future research should explore whether a more patient-tailored intervention would yield better results.

In Brief

Adding 6 sessions of pain neuroscience education to standard physiotherapy after breast cancer surgery did not result in a better course of functioning compared with biomedical pain education.

1. Introduction

One of the most common side effects of breast cancer (treatment) is pain.²⁹ The literature reports average pain prevalence of 31% 1 to 2 years after breast cancer surgery.^15,64 Of particular concern is the potential impact of pain because pain can cause limitations in daily activities, participation, and interaction with the environment, with consequences for physical and emotional functioning.¹⁶ Adequate pain management in the early stage of breast cancer treatment is essential for resolving and preventing these problems, both in the short and long term.¹⁸ Educational interventions may optimize pain management by improving patient knowledge, perceived control, and attitudes toward pain.^44,45

However, clinically relevant effects of educational interventions for the management of pain in cancer populations are currently lacking.^3,46,51 Perhaps because those interventions often describe pain from a biomedical perspective (eg, explanations of structures that can cause pain and analgesic advice) and hence fail to explain other reasons why pain can persist beyond the healing process. Increased understanding of pain (neuro)physiology has resulted in a neuroscience-based educational intervention aimed to reconceptualize pain beyond the biomedical model and toward a biopsychosocial understanding.⁴³ Pain neuroscience education (PNE) explains that pain is not always a true representation of tissue damage, but rather the nervous system's interpretation of the threat of injury, which is influenced by a variety of psychosocial factors.^4,41,44 As a result of this pain reconceptualization, people may perceive pain as less threatening, and barriers to participating in (previously avoided) activities because of pain may be removed, potentially resulting in less pain-related disability and better physical and mental functioning.^39,47

To our knowledge, only 2 studies compared the effect of perioperative PNE to biomedical education in women undergoing breast cancer surgery.^10,34 The first was a pilot randomized controlled trial that examined 2 educational interventions given before breast cancer surgery.¹⁰ The intervention group watched a 90-minute pain psychoeducational video (n = 36), whereas the control group received digital health and nutrition education (n = 32). No significant effect for pain-related disability, pain intensity, and physical or emotional functioning was found up to 12 weeks after surgery. The second study was a retrospective nonrandomized case–control trial that compared perioperative PNE (n = 51) to biomedical education (n = 51) in patients with persistent postoperative pain 1 year after breast cancer surgery, excluding pain from other cancer treatments.³⁴ They found that PNE was more effective than biomedical education for pain-related disability, pain intensity, central sensitization–related symptoms, and pain-related catastrophizing, although effect sizes were small.

Given the inconclusive results and limitations of previous studies in design (non–randomized controlled trial), short follow-up, small sample size, and study population (no generalization to the general breast cancer population), we conducted a double-blinded randomized controlled trial with 18-month follow-up to determine whether breast cancer patients who received postoperative PNE reported more favorable changes in functioning than control group who received biomedical education, both in addition to standard physiotherapy. The primary outcome was the change in pain-related disability over 12 months. Secondary outcomes included changes in pain intensity and physical and emotional functioning 4, 6, 8, 12, and 18 months postoperatively.

2. Methods

2.1. Study design

The EduCan trial was a parallel, two-arm randomized controlled trial approved by the Ethical Committee of the University Hospitals Leuven (s60702) and registered at ClinicalTrials.gov (NCT03351075). A detailed description of the protocol has been published.¹²

2.2. Participants

Recruitment took place at the Multidisciplinary Breast Center of the University Hospitals Leuven campus Gasthuisberg (Belgium) between November 2017 and March 2020. Potential participants signed informed consent before inclusion. Inclusion criteria were as follows: being diagnosed with histologically confirmed invasive or noninvasive primary breast cancer; scheduled for one of the following surgeries: mastectomy with sentinel node biopsy or axillary lymph node dissection (with or without breast reconstruction) or breast conserving surgery with axillary lymph node dissection; no distant metastasis; female; aged 18 years or older; could comply with the study protocol; and comprehension of the Dutch language (reading, listening, writing, and speaking).

2.3. Randomization

After enrollment, participants were randomly assigned (1:1) to the intervention group (PNE) or control group (biomedical education). This computer-generated randomization was performed by an independent coworker (T.D.V.) using permuted blocks (size = 4).

2.4. Blinding

Participants, assessors, and physical therapists performing the standard physical therapy program were all blinded to group allocation. Before consenting to participate in the study, participants were informed that they would be randomized to either a “traditional biomedical education” or a “modern educational intervention.” The difference between these interventions was neither explained during recruitment nor in the written consent document. To prevent contamination between the 2 groups, both the educational sessions and physiotherapy were one-on-one, minimizing possible interaction between participants. An informative session on prevention and treatment of lymphedema did take place in a group of approximately 10 participants. A communication sheet was drawn up to ensure that standardized answers were given to patient's questions during the standard physical therapy program. If the participants asked pain-related questions, they were referred to the physiotherapist delivering the educational interventions. An independent statistician (S.F.) of the Leuven Biostatistics and statistical Bioinformatics Center analyzed the data to ensure additional blinding of the research team.

2.5. Interventions

2.5.1. Standard physical therapy program

All participants attended a one-on-one 30-minute standard physical therapy program once or twice weekly (intensive phase) starting the first week postsurgery, independent of group allocation. These sessions took place at the Department of Physical Medicine and Rehabilitation of the University Hospitals Leuven campus Gasthuisberg (Belgium) and were delivered by 4 physiotherapists (L.D., V.H., and E.V.d.G.) with a master's degree in Rehabilitation Sciences and Physiotherapy. The program included 3 modalities tailored to the individual needs of the participant: (1) manual techniques (passive mobilizations to restore shoulder range of motion, myofascial techniques to improve muscle flexibility, and scar tissue massage to improve flexibility of the scar(s)), (2) specific exercises to improve shoulder range of motion and upper limb strength, and (3) advice on general exercises to increase physical activity level.

After 4 months, this intensive program was replaced by 3 individual follow-up sessions (maintenance phase) at 6, 8, and 12 months after surgery. At these time points, participants received a physiotherapy session (by L.D. or E.V.d.G.) and were referred to a physiotherapy practice in primary care for further intensive follow-up if needed.

Additionally, participants were asked to attend one informative group session regarding prevention and treatment of lymphedema, given by a physical therapist with a master's degree in Rehabilitation Sciences and Physiotherapy and specialized in treatment of breast cancer–related lymphedema (T.D.V. or L.V.). If the participants reported symptoms of lymphedema, they were referred for thorough evaluation and treatment at the Center for Lymphedema of the University Hospitals Leuven.

2.5.2. Educational sessions

Throughout the whole study period, participants attended 6 one-on-one, 30-minute educational sessions on pain after breast cancer treatment. These sessions took place at the Department of Physical Medicine and Rehabilitation of the University Hospitals Leuven campus Gasthuisberg (Belgium) and were delivered by a physical therapist proficient in pain management (A.D.G. or K.B.). Three sessions were scheduled in the intensive phase (starting 1-3 weeks after surgery) and 3 sessions in the maintenance phase at 6, 8, and 12 months after surgery. In case a face-to-face session in the hospital was not possible, a digital session with live therapist interaction covering the same content was provided. During the educational sessions, information was presented both verbally and with a PowerPoint presentation. Additionally, participants received a booklet and a web-link to an online summarizing presentation to read at home. Knowledge regarding the principles that were covered during the educational sessions was tested before the first and after the third session (intensive phase) and before the start of the fourth session (maintenance phase) by means of a questionnaire based on the Neurophysiology of Pain Test,³⁵ adapted to the educational content of both groups.

Women who did not experience pain at the time of the educational sessions were taught how to cope with possible future pain. During the 3 maintenance sessions, the physiotherapist went through the information provided in the intensive phase and discussed the implementation in future stages of the recovery process. For these sessions, the participants received a second booklet with specific information for this phase.

Participants in the control group and intervention group had the same schedule and format of educational sessions, only the content of the education differed between the groups.

2.5.2.1. Control group: biomedical education

The learning goal consisted of gaining biomedically oriented knowledge about pain after breast cancer treatment. The participants were explained that pain is related to tissue injury caused by the different treatment procedures for breast cancer. The physical therapist providing the education talked about the side effects of these different treatment modalities, the role of different structures, and injured vs healthy tissue in acute and persistent pain. Additionally, patients received guidance on activity management based on the load—loadability principle (physical activity—rest). The physical therapist advised them to listen to their body and adjust their physical activity level accordingly. He or she also went over the current recommendations for general exercises after cancer treatment, based on the American Cancer Society Guidelines for Physical Activity: at least 150 minutes of moderate-intensity activity or 75 minutes of vigorous-intensity activity each week (or a combination of these), preferably spread throughout the week.³¹ Finally, advice on work resumption in the context of the different (persistent) side effects of the treatments was provided, with a focus on ergonomic factors. The different factors influencing return to work were discussed, and patients were informed on whom to contact to address those factors (this information was the same in both groups).

2.5.2.2. Intervention group: pain neuroscience education

The learning goal consisted of gaining biopsychosocial oriented knowledge about pain after breast cancer treatment. Participants were explained the physiological and psychological processes involved in the pain experience to help reconceptualizing pain. The authors did so by adapting the content and images from the books “Explain Pain,”⁴ “Pijneducatie een praktische handleiding voor (para)medici,”⁶² and “The Pain Toolkit”⁴⁰ for use in a breast cancer population. The sessions included the following topics: characteristics of acute vs persistent pain, specific side effects of breast cancer treatment modalities in relation to pain, how pain is a product of the brain, how pain becomes persistent (plasticity of the nervous system, modulation, modification, and central sensitization), and potential sustaining factors of pain (such as emotions, stress, pain cognitions, and pain behavior). Additionally, the experimental intervention included advice on activity management while experiencing pain and other symptoms, considering the intertwinement of influencing biopsychosocial factors. Participants learned about increasing general exercises and activities according to the principles of graded activity and pacing reported by the International Association for the Study of Pain.³⁶ This includes general exercise activities according to pacing strategies for “persisters” (ie, restructuring the activity pattern to avoid peaks of overactivity and exacerbations of their pain) and graded activity for “avoiders” (ie, time-contingent increase of physical activity). Finally, it was explained that work resumption could break the vicious cycle of biopsychosocial components and persistent pain. In addition, the principles described above for activity management were applied to the working situation. The different factors influencing return to work were discussed and patients were informed on whom to contact to address those factors (this information was the same in both groups).

2.6. Outcomes

All participants were evaluated preoperatively and postoperatively (within 1 week before and after surgery) and at 4, 6, 8, 12, and 18 months postoperatively.

2.6.1. Primary outcome measure

The primary outcome was the change in pain-related disability from before surgery to 12 months after surgery. Pain-related disability was evaluated using the Dutch language version of the Pain Disability Index (PDI-DLV).^50,61 The primary endpoint of the study was set at 12 months postoperatively because the majority of recovery from breast cancer surgery occurs within 12 months of surgery, and the studied intervention was designed to operate primarily within this time frame. The PDI assesses the degree of pain interference with normal role functioning in 7 different life domains (family/home responsibilities, recreation, social activity, occupation, sexual behavior, self-care, and life-support activity) on a 10-point Likert scale ranging from 0 (no disability) to 10 (total disability) (total score range 0-70).^50,61 One of the 2 researchers (L.D. or E.V.d.G.) administered the PDI-DLV during the evaluation consultations at the Department of Physical Medicine and Rehabilitation of the University Hospitals Leuven campus Gasthuisberg (Belgium). In addition to written instructions, the researchers provided a standard verbal instruction, stating that mean pain-related disability from any cause should be indicated.

2.6.2. Secondary outcome measures

Secondary outcomes were 3-fold: (1) pain symptoms and characteristics; (2) physical functioning; and (3) emotional functioning. Assessments of secondary outcomes were completed by the participants at home, either electronically via the digital patient record or on paper, within 1 week before or after the evaluation consultation.

2.6.2.1. Pain symptoms and characteristics

2.6.2.1.1 Pain intensity

The visual analog scale (VAS) is a horizontal 100-mm line with 2 endpoints representing the extreme states: “no pain” and “worst pain possible.”²⁸ Participants were asked to rate the global mean pain intensity experienced in the past week. In addition to written instructions, the researchers provided a standard verbal instruction, stating that mean pain-related disability from any cause should be indicated. The VAS was found to have good psychometric properties to evaluate pain in women diagnosed with breast cancer.¹⁹

2.6.2.2. Physical functioning

2.6.2.2.1. Physical activity level

A waist-worn tri-axial accelerometer (ActiGraph wGT3X-BT1+, Pensacola, FL, USA) was used to evaluate physical activity level. More precisely, the parameters derived from this device were physical activity energy expenditure (kcals/day), sedentary time (minute/day), moderate-to-vigorous physical activity (minute/day), (very) vigorous physical activity (minute/day), and step counts (steps/day). Physical activity outcomes were only evaluated postoperatively and at 4 and 12 months postoperatively. Participants were instructed to wear the device on the right hip during 7 consecutive days for at least 12 hours during waking hours (except for showering or swimming).³⁷ Data collection was considered valid when at least 4 days with a recording period of ≥600 minutes were available.³⁷ ActiLife software (version 6.13.4 Full Edition) was used to process the data. A sample rate of 90 Hz, 60-second epoch setting and modified version of the Choi algorithm (60-0-1 using vector magnitude) was applied. The Freedson VM3 combination cut points were used to evaluate categories of activity intensity (moderate: 2691-6166, vigorous: 6167-9642, and very vigorous physical activity: 9643-∞ counts per minute).^37,55 A minimum bout length of 10 minutes, with maximum 150 counts per minute using the vertical axis from hip accelerations was used to categorize sedentary time.^30,37 The Freedson VM3 combination algorithm was used to measure physical activity energy expenditure.¹ The ActiGraph GT3X+ has demonstrated excellent relative reliability (2-week interval) for sedentary behavior and good relative reliability for moderate-to-vigorous physical activity in patients 12 months after breast cancer surgery.⁴⁹

2.6.2.2.2. Upper limb function

The Disabilities of the Arm, Shoulder and Hand questionnaire (DASH) is a 30-item questionnaire that assesses symptoms and functional status, with a focus on physical function, in populations with upper extremity musculoskeletal conditions.^23,48 The items cover upper extremity-related symptoms and measure functional status at the level of disability. Patients score each item on a 5-point Likert scale (1-5), with higher scores reflecting higher disability/worse symptoms. The total score of the DASH ranges between 0 and 100. Because of its consistently large effect sizes for construct validity and responsiveness, the DASH is recommended for assessing upper extremity function in breast cancer survivors.²⁰

2.6.2.3. Emotional functioning

2.6.2.3.1. Pain-related catastrophizing

The Pain Catastrophizing Scale (PCS) is a 13-item questionnaire that reflects on previous painful experiences and asks to indicate the degree to which each of the 13 described thoughts or feelings were experienced while in pain. Each question is scored on a 5-point Likert scale from 0 (not at all) to 4 (all the time). Total PCS scores can be evaluated (ranging from 0 to 52), with higher scores corresponding to more pain-related catastrophizing, as well as 3 subscales scores assessing rumination about pain (rumination, score range 0-16), magnification of negative consequences in the context of pain (magnification, score range 0-12), and experienced helplessness in the context of pain (helplessness, score range 0-24).^8,59,60

2.6.2.3.2. Depression, anxiety, and stress

The Depression, Anxiety and Stress Scale (DASS-21) evaluates the presence of negative emotional states of severity of depression, anxiety, and stress over the past week. Each subscale consists of 7 questions, with each question scored on 4-point Likert scale from 0 (did not apply to me at all) to 3 (applied to me very much or most of the time).^11,33

2.6.2.3.3. Psychological symptoms, existential well-being, and support

The McGill Quality of Life Questionnaire (MQOL) is multidimensional tool that evaluates the overall quality of life over the past 2 days. The following MQOL subscales are related to psychosocial functioning and included in the present study: psychological symptoms (4 items), existential well-being (6 items), and support (2 items). Each item is scored on an 11-point Likert scale from 0 to 10 with opposite anchors at the end, with higher scores reflecting a better psychosocial functioning.^7,13

2.7. Sample size

The sample size was based on a comparison of the changes for the primary outcome measure (PDI-DLV) at 12 months after surgery. Because no information was available on the SD of the changes, the calculation was based on a comparison of the values at 12 months (note this corresponds to assuming a correlation between baseline and 12 months equal to 0.5, in which case the SD of the change equals the SD at 12 months). Assuming a coefficient of variation (CV) equal to 0.5, 87 participants per group were needed based on a 2-sample pooled t test of a mean ratio with lognormal data and α = 0.05 to detect with 80% power a difference of 20% in PDI.^6,57 To anticipate a drop-out rate of 5%, a total of 184 subjects were needed to be recruited.

2.8. Statistical methods

Descriptive statistics for continuous values are presented as mean (SD) and median (interquartile range). Categorical variables are presented as frequency and proportion (%).

A multivariate linear model for longitudinal measures with an unstructured covariance matrix (fitted on the measurements preoperatively and 4, 6, 8, 12, and 18 months postoperatively) was applied for each continuous (primary and secondary) outcome, correcting for the postoperative assessment (allowing the relation between the postoperative value and the value at the other timepoints to be timepoint specific). From this model, changes vs the preoperative value were reported (with 95% confidence interval [CI]) and compared between both groups. Using likelihood-based estimation, subjects with a missing value at one or more timepoints were still included in the analyses. For the outcomes with missing postoperative values, the model was fitted applying a multiple imputation approach (MCMC method within each group and using 20 imputed datasets). For the physical activity outcomes that were measured postop, at 4 months and at 12 months, changes vs the postoperative value were evaluated.

If model residuals showed a right-skewed distribution and zero values were present, an inverse hyperbolic transformation was applied. Because this transformation is a log-transform, a change refers to a ratio after back-transformation.

Furthermore, moderator analyses were performed for the effect on PDI-DLV at 12 months. First, a multiple imputation was performed (MCMC method). To ensure that imputed values were in the 0 to 70 range, the imputation was performed on logit transformed values (PDI-DLV values equal to 0 were replaced by value 0.5 before applying the transformation). Second, Spearman correlations were reported between preoperative and postoperative variables and the change in PDI-DLV after 12 months, as well as the slopes from a linear regression model. In the latter model, the interaction was verified, referring to a difference in slope between both groups. Rubin's rule was used to combine the results from the imputed datasets. Following preoperative and postoperative variables were explored as potential moderators: pain-related disability, pain intensity, upper limb function, pain-related catastrophizing, depression, anxiety, stress, psychological symptoms, existential well-being, and support.

Data were analyzed using SAS software, version 9.4. P < 0.05 was considered significant. No corrections for multiple testing were performed.

3. Results

In total, 493 women were eligible, of which 184 were included in the study (Fig. 1). Both intervention and control groups had similar demographic and clinical characteristics at baseline (Table 1). Attendance at educational interventions was 98% in the intervention group and 99% in the control group during the intensive phase. During the maintenance phase, 97% of participants in the intervention group and 98% of participants in the control group attended the educational interventions.

Figure 1.: Flowchart of the EduCan trial according to the Consort 2010 flow diagram.^{56, 38} A0: baseline assessment, A1: postoperative assessment, A4: 4 months postsurgery assessment (end of intensive phase), A6: 6 months postsurgery assessment, A8: 8 months postsurgery assessment, A12: 12 months postsurgery assessment, and A18: 18 months postsurgery assessment (end of maintenance phase). Missing data and reason are shown for the primary outcome measure (pain-related disability evaluated with Pain Disability Index).

Table 1 - Characteristics of participants according to treatment allocation.

	Intervention group, n=92	Control group, n=92
Age (y), mean (SD); median (IQR)	55.4 (11.5); 54.0 (14.1)	55.2 (11.4); 54.0 (15.4)
BMI (kg/m2), mean (SD); median (IQR)	25.4 (4.3); 24.2 (6.4)	25.9 (5.9); 24.7 (6.8)
Educational level*
Primary education or no diploma	3 (3.5%)	3 (3.5%)
Lower secondary education	5 (6%)	5 (6%)
Upper secondary education	23 (26%)	28 (33%)
Higher education: professional bachelor	30 (34.5%)	32 (37%)
Higher education: academic bachelor or master	25 (29%)	19 (22%)
Surgery at dominant side	44 (48%)	41 (45%)
Type of surgery
BCS + ALND	4 (4%)	9 (10%)
ME + SLNB	41 (45%)	43 (47%)
ME + ALND	47 (51%)	40 (43%)
Tumor size
pTis	5 (5%)	8 (9%)
pT0	8 (9%)	9 (10%)
pT1	30 (33%)	26 (28%)
pT2	30 (33%)	38 (41%)
pT3	17 (18%)	11 (12%)
pT4	2 (2%)	0 (0%)
Lymph node stage
pNx	1 (1%)	0 (0%)
pN0	43 (47%)	50 (54%)
pN1	36 (39%)	29 (31.5%)
pN2	9 (10%)	7 (8%)
pN3	3 (3%)	6 (6.5%)
Radiotherapy	74 (80%)	66 (72%)
Breast	3 (3%)	9 (10%)
Thorax	63 (69%, N = 91)	53 (58%)
MSP	68 (75%, N = 91)	59 (64%)
Axilla	5 (5%, N = 91)	7 (8%)
Hormone therapy (ongoing)	69 (75%)	68 (74%)
Tamoxifen	18 (20%)	11 (12%)
Aromatase inhibitors	51 (55%)	57 (62%)
Chemotherapy	63 (68.5%)	55 (60%)
Neo-adjuvant	25 (27%)	25 (27%)
Adjuvant	38 (41%)	30 (33%)
Anthracyclines	39 (42%)	38 (41%)
Taxane-based	63 (68.5%)	38 (41%)
Xeloda	2 (2%)	5 (5%)
Target therapy (ongoing)	23 (25%)	20 (22%)
No. of physical therapy sessions, mean (SD); median (IQR)	20.1 (6.6); 19.5 (8.0)	20.3 (7.7); 20.0 (10)

Numbers (%) are given unless specified otherwise.

^*Assessed retrospectively at 4 months after surgery, so only calculated from data available at 4 months postoperatively.

ALND, axillary lymph node dissection; BCS, breast conserving surgery; BMI, body mass index; IQR, interquartile range; ME, mastectomy; MSP, median subclavian and parasternal lymph node areas; SLNB, sentinel lymph node biopsy.

The primary analysis (Table 2) revealed no significant difference in pain-related disability change from baseline to 12 months after surgery between the intervention and control groups (intervention 4.2, 95% CI: 1.4-7.0; control 5.5, 95% CI: 2.7-8.3; difference in change −1.31, 95% CI: −5.3 to 2.7; P = 0.516). Similar results were found at 4, 6, 8, and 18 months after surgery. Figures 2 and 3 graphically show the scores and changes over time for pain-related disability, respectively.

Table 2 - Observed results and estimates for within- and between-group changes in primary outcome pain-related disability and secondary outcome pain intensity at different time points after surgery vs before surgery from a multivariate linear model for longitudinal measures.

	Observed information						Change within groups		Difference in change between groups
	Intervention group			Control group			Intervention group	Control group	Estimate (CI)	P
	Mean	Median (IQR)	n	Mean	Median (IQR)	n	Estimate (CI)	Estimate (CI)
Pain-related disability (PDI-DLV 0-70) (primary outcome)
Preoperatively	4.9	0.0 (4.0)	91	4.6	0.0 (8.0)	92
1 wk postoperatively	20.5	17.5 (27.0)	92	21.6	22.0 (27.0)	89
At 4 mo	8.5	3.25 (11.9)	88	9.3	5.0 (12.0)	89	3.70 (1.37; 6.04)	4.67 (2.35; 6.99)	−0.97 (−4.26; 2.33)	0.5655
At 6 mo	8.8	3.5 (14.0)	86	9.7	5.0 (13.5)	88	4.06 (1.71; 6.42)	5.07 (2.74; 7.40)	−1.01 (−4.32; 2.31)	0.5521
At 8 mo	7.9	2 (10.0)	86	10.0	4.0 (15.0)	87	3.22 (0.67; 5.77)	5.48 (2.94; 8.01)	−2.26 (−5.86; 1.34)	0.2180
At 12 mo (primary endpoint)	9.3	4.0 (12.0)	82	9.9	3.5 (15.5)	84	4.22 (1.40; 7.03)	5.53 (2.74; 8.32)	−1.31 (−5.28; 2.65)	0.5163
At 18 mo	7.9	3.0 (11.0)	83	8.7	5.0 (14.0)	86	3.13 (0.59; 5.67)	4.19 (1.68; 6.70)	−1.07 (−4.64; 2.51)	0.5592
Pain intensity (VAS 0-100)
Preoperatively	15.5	10.0 (21.0)	91	15.1	7.0 (23.0)	92
1 wk postoperatively	31.8	28.0 (31.0)	92	28.9	25.0 (32.5)	92
At 4 mo	22.7	16.0 (34.5)	88	24.6	20.0 (46.0)	89	7.26 (1.87; 12.65)	9.49 (4.14; 14.85)	−2.23 (−9.84; 5.37)	0.5630
At 6 mo	24.1	19.0 (34.0)	85	25.7	20.5 (39.0)	88	8.72 (2.80; 14.63)	10.42 (4.58; 16.27)	−1.70 (−10.03; 6.62)	0.6869
At 8 mo	21.2	14.5 (36.0)	86	23.4	18.0 (30.0)	87	5.95 (0.41; 11.49)	8.67 (3.17; 14.17)	−2.72 (−10.54; 5.09)	0.4928
At 12 mo	21.6	16.0 (25.0)	81	25.5	24.5 (41.0)	84	6.31 (1.06; 11.57)	10.61 (5.43; 15.79)	−4.30 (−11.68; 3.09)	0.2524
At 18 mo	19.0	15.0 (28.0)	83	23.7	14.5 (45.0)	86	3.91 (−1.70; 9.52)	8.71 (3.17; 14.24)	−4.80 (−12.69; 3.09)	0.2315

Within-group changes vs preoperatively moment and comparison of these changes between both the groups are derived from the multivariate linear model for longitudinal measures.

CI, confidence interval; IQR, interquartile range; PDI-DLV, Dutch language version of the Pain Disability Index; VAS, visual analog scale.

Figure 2.: Pain Disability Index (PDI, total score) over time. Pre: baseline assessment before surgery, post: postoperative assessment, 4 m: 4 months postsurgery assessment (end of intensive phase), 6 m: 6 months postsurgery assessment, 8 m: 8 months postsurgery assessment, 12 m: 12 months postsurgery assessment, and 18 m: 18 months postsurgery assessment (end of maintenance phase).

Figure 3.: Change in Pain Disability Index (PDI, total score) vs preoperative assessment. Pre: baseline assessment before surgery, 4 m: 4 months postsurgery assessment (end of intensive phase), 6 m: 6 months postsurgery assessment, 8 m: 8 months postsurgery assessment, 12 m: 12 months postsurgery assessment, and 18 m: 18 months postsurgery assessment (end of maintenance phase).

Analysis of secondary outcomes showed that there were no statistically significant differences in change between the intervention and control groups (Tables 2–4). Over the 18-month follow-up, both groups experienced increases in pain intensity (intervention 3.91, 95% CI: −1.70 to 9.52; control 8.71, 95% CI: 3.17-14.24) (Table 2) and a decline in upper limb function (intervention 6.56, 95% CI: 3.15-9.97; control 9.72, 95% CI: 3.17-14.24) (Table 3) relative to the preoperative level. Regarding emotional functioning (Table 4), both groups experienced decreases in psychological symptoms (intervention 2.07, 95% CI: 1.57-2.57; control 2.09, 95% CI: 1.60-2.57) and social support (intervention −0.54, 95% CI: −0.96 to −0.11; control −0.83, 95% CI: −1.24 to −0.42) and an increase in existential well-being (intervention 1.07, 95% CI: 0.64-1.49; control 0.61, 95% CI: 0.20-1.03) compared with the preoperative level, over the 18-month follow-up.

Table 3 - Observed results and estimates for within- and between-group changes in physical functioning at different time points after surgery vs before surgery from a multivariate linear model for longitudinal measures.

	Observed information						Change within groups		Difference in change between groups
	Intervention group			Control group			Intervention group	Control group	Estimate (CI)	P
	Mean	Median (IQR)	n	Mean	Median (IQR)	n	Estimate (CI)	Estimate (CI)
Sedentary time (accelerometry, min/d)
1 wk postoperatively	400.4	396.2 (97.9)	64	398.5	418.3 (110.5)	66
At 4 mo	372.3	380.8 (138.7)	72	377.3	391.7 (109.7)	71	−19.9 (−39.4; −0.3)	−20.8 (−40.6; −1.0)	0.9 (−26.9; 28.7)	0.9479
At 12 mo	356.5	371.1 (113.3)	48	372.2	364.3 (113.6)	48	−34.9 (−59.6; −10.1)	−23.9 (−48.6; 0.8)	−10.9 (−45.9; 24.0)	0.5376
Time in moderate-to-vigorous physical activity (accelerometry, min/d)
1 wk postoperatively	29.5	30.2 (29.01)	64	31.4	22.4 (34.3)	66
At 4 mo	33.6	28.5 (37.5)	72	32.8	26.3 (29.0)	71	3.45 (−1.43; 8.33)	2.54 (−2.41; 7.50)	0.90 (−6.05; 7.86)	0.7980
At 12 mo	45.4	44.2 (38.4)	48	38.1	35.1 (25.5)	48	13.41 (6.55; 20.26)	6.75 (−0.11; 13.61)	6.66 (−3.04; 16.35)	0.1770
Time in (very) vigorous activity (accelerometry, min/d)
1 wk postoperatively	0.47	0.0 (0.14)	66	0.48	0.0 (0.17)	66
At 4 mo	1.31	0.0 (0.29)	72	0.67	0.0 (0.25)	71	0.18 (−0.01; 0.36)	0.12 (−0.07; 0.31)	0.06 (−0.21; 0.32)	0.6620
At 12 mo	1.91	0.14 (0.95)	48	1.24	0.14 (0.54)	48	0.39 (0.12; 0.66)	0.22 (−0.05; 0.49)	0.17 (−0.21; 0.55)	0.3731
Step count average (accelerometry, steps/day)
1 wk postoperatively	6499	5909 (4232)	64	6201	5909 (3798)	66
At 4 mo	6673	7241 (4560)	72	6609	5943 (3079)	71	50 (−600; 700)	342 (−316; 1000)	−293 (−1218; 632)	0.5332
At 12 mo	7905	8221 (3593)	48	6642	6390 (3365)	48	1100 (349; 1850)	261 (−489; 1010)	839 (−221; 1900)	0.1201
Upper limb function (DASH 0-100)
Preoperatively	13.2	7.1 (20.8)	86	12.0	7.5 (17.7)	88
1 wk postoperatively	41.8	41.4 (24.2)	85	39.2	40.0 (22.5)	85	28.39 (24.76; 32.03)	27.10 (23.47; 30.73)	1.29 (−3.84; 6.43)	0.6197
At 4 mo	22.0	17.5 (27.5)	79	20.5	17.5 (24.4)	86	10.05 (7.05; 13.05)	7.81 (4.89; 10.72)	2.24 (−1.94; 6.43)	0.2915
At 6 mo	23.1	20.0 (21.3)	83	22.5	19.6 (24.2)	82	10.02 (7.01; 13.02)	10.40 (7.44; 13.37)	−0.39 (−4.60; 3.83)	0.8563
At 8 mo	22.1	14.6 (24.4)	84	23.3	20.3 (26.7)	86	9.16 (5.73; 12.58)	10.60 (7.22; 13.97)	−1.44 (−6.25; 3.37)	0.5558
At 12 mo	21.6	15.8 (20.0)	81	22.8	19.2 (26.7)	83	8.51 (5.39; 11.64)	10.15 (7.07; 13.22)	−1.63 (−6.02; 2.75)	0.4632
At 18 mo	19.1	12.5 (21.7)	79	22.1	19.2 (27.1)	82	6.56 (3.15; 9.97)	9.72 (6.37; 13.07)	−3.16 (−7.94; 1.63)	0.1944

Within-group changes vs preoperatively moment and comparison of these changes between both groups are derived from the multivariate linear model for longitudinal measures.

CI, confidence interval; DASH, Disabilities of the Arm, Shoulder and Hand questionnaire; IQR, interquartile range.

Table 4 - Observed results and estimates for within- and between-group changes in emotional functioning at different time points after surgery vs before surgery from a multivariate linear model for longitudinal measures.

	Observed information						Change within groups		Difference in change between groups
	Intervention group			Control group			Intervention group	Control group	Estimate (CI)	P
	Mean	Median (IQR)	n	Mean	Median (IQR)	n	Estimate (CI)	Estimate (CI)
Pain-related catastrophizing (PCS 0-52)*
Preoperatively	10.3	8.5 (13.0)	90	8.6	7.0 (11.0)	89
1 wk postoperatively	9.2	8.0 (11.0)	90	8.5	6.5 (9.0)	90
At 4 mo	9.6	7.0 (15.0)	79	9.2	7.0 (12.0)	86	0.896 (0.676; 1.188)	1.077 (0.818; 1.419)	0.831 (0.561; 1.233)	0.3590
At 6 mo	9.6	9.0 (12.0)	84	8.8	6.5 (13.0)	82	0.870 (0.661; 1.145)	0.958 (0.727; 1.261)	0.908 (0.615; 1.340)	0.6277
At 8 mo	9.9	8.0 (12.5)	84	8.9	7.0 (13.0)	87	0.877 (0.662; 1.162)	1.005 (0.761; 1.328)	0.873 (0.587; 1.297)	0.5001
At 12 mo	9.1	7.0 (11.0)	81	9.5	8.0 (14.0)	83	0.894 (0.679; 1.176)	1.065 (0.811; 1.399)	0.839 (0.570; 1.236)	0.3744
At 18 mo	8.3	5.0 (9.5)	80	9.4	6.0 (12.0)	81	0.762 (0.562; 1.033)	0.898 (0.663; 1.216)	0.848 (0.552; 1.304)	0.4533
Depression (DASS-21 0-42)*
Preoperatively	7.5	4.0 (8.0)	90	6.3	6.0 (8.0)	91
1 wk postoperatively	6.8	4.0 (8.0)	90	5.6	4.0 (6.0)	91
At 4 mo	6.3	4.0 (10.0)	79	5.9	4.0 (6.0)	84	0.802 (0.588; 1.092)	0.930 (0.686; 1.261)	0.862 (0.558; 1.330)	0.5018
At 6 mo	7.1	4.0 (11.0)	84	6.2	4.0 (8.0)	83	0.819 (0.611; 1.100)	0.956 (0.713; 1.282)	0.857 (0.565; 1.299)	0.4674
At 8 mo	6.7	2.0 (8.0)	85	5.7	3.0 (8.0)	86	0.822 (0.608; 1.112)	0.748 (0.554; 1.010)	1.099 (0.717; 1.683)	0.6653
At 12 mo	5.3	2.0 (8.0)	82	5.3	4.0 (8.0)	83	0.625 (0.459; 0.851)	0.774 (0.570; 1.052)	0.808 (0.522; 1.248)	0.3362
At 18 mo	6.2	4.0 (8.0)	80	5.5	2.0 (8.0)	82	0.718 (0.535; 0.963)	0.749 (0.560; 1.003)	0.958 (0.633; 1.449)	0.8375
Anxiety (DASS-21 0-42)*
Preoperatively	5.4	4.0 (6.0)	90	5.1	4.0 (6.0)	91
1 wk postoperatively	4.9	4.0 (6.0)	90	4.8	2.0 (8.0)	91
At 4 mo	5.5	4.0 (6.0)	79	5.7	4.0 (8.0)	84	1.045 (0.772; 1.415)	0.987 (0.733; 1.328)	1.059 (0.693; 1.620)	0.7902
At 6 mo	5.2	4.0 (5.0)	84	6.1	4.0 (6.0)	84	0.963 (0.731; 1.269)	1.213 (0.922; 1.596)	0.794 (0.538; 1.171)	0.2445
At 8 mo	4.1	2.0 (6.0)	85	5.5	4.0 (8.0)	86	0.746 (0.562; 0.992)	0.924 (0.697; 1.224)	0.808 (0.541; 1.206)	0.2973
At 12 mo	4.6	2.0 (6.0)	82	4.9	4.0 (8.0)	83	0.783 (0.575; 1.067)	0.839 (0.617; 1.139)	0.934 (0.604; 1.443)	0.7578
At 18 mo	4.7	3.0 (8.0)	80	5.0	4.0 (8.0)	82	0.829 (0.625; 1.100)	0.821 (0.621; 1.085)	1.010 (0.679; 1.504)	0.9600
Stress (DASS-21 0-42)*
Preoperatively	10.6	10.0 (12.0)	90	9.6	8.0 (10.0)	91
1 wk postoperatively	7.5	7.0 (10.0)	90	7.5	6.0 (12.0)	91
At 4 mo	8.5	8.0 (10.0)	80	8.3	8.0 (11.0)	84	0.812 (0.598; 1.101)	0.760 (0.563; 1.026)	1.068 (0.696; 1.639)	0.7619
At 6 mo	9.4	10.0 (12.0)	84	9.1	7.0 (12.0)	84	0.909 (0.675; 1.225)	0.866 (0.644; 1.165)	1.049 (0.689; 1.597)	0.8225
At 8 mo	9.8	8.0 (8.0)	85	9.5	10.0 (12.0)	86	0.984 (0.736; 1.314)	0.920 (0.690; 1.226)	1.069 (0.711; 1.608)	0.7482
At 12 mo	9.0	8.0 (12.0)	82	8.0	6.0 (11.0)	84	0.782 (0.574; 1.064)	0.721 (0.532; 0.979)	1.084 (0.702; 1.673)	0.7169
At 18 mo	9.1	10.0 (12.0)	80	8.8	8.0 (10.0)	82	0.857 (0.629; 1.167)	0.868 (0.639; 1.178)	0.987 (0.639; 1.525)	0.9535
Psychological symptoms (MQOL 0-10)
Preoperatively	5.4	5.5 (3.7)	89	5.5	5.7 (4.0)	92
1 wk postoperatively	7.0	7.2 (3.2)	90	7.0	7.6 (4.2)	90
At 4 mo	7.4	7.7 (2.7)	78	7.2	7.7 (3.5)	86	1.93 (1.37; 2.48)	1.68 (1.13; 2.22)	0.25 (−0.53; 1.03)	0.5306
At 6 mo	7.3	8.0 (3.5)	84	7.0	7.5 (3.7)	83	1.88 (1.35; 2.40)	1.41 (0.89; 1.93)	0.47 (−0.27; 1.21)	0.2118
At 8 mo	7.0	8.0 (4.5)	83	7.4	7.7 (3.2)	87	1.57 (0.97; 2.16)	1.92 (1.33; 2.50)	−0.35 (−1.19; 0.49)	0.4118
At 12 mo	7.6	8.1 (3.0)	82	7.3	8.1 (3.5)	84	2.07 (1.49; 2.65)	1.82 (1.25; 2.39)	0.25 (−0.56; 1.06)	0.5441
At 18 mo	7.6	8.5 (3.7)	79	7.6	8.0 (2.5)	83	2.07 (1.57; 2.57)	2.09 (1.60; 2.57)	−0.02 (−0.71; 0.68)	0.9646
Existential well-being (MQOL 0-10)
Preoperatively	6.5	6.5 (1.8)	89	6.5	6.8 (2.7)	90
1 wk postoperatively	6.5	6.5 (2.5)	90	6.6	6.7 (2.2)	90
At 4 mo	6.7	7.0 (2.5)	80	7.0	7.2 (2.2)	86	0.19 (−0.14; 0.52)	0.51 (0.19; 0.84)	−0.32 (−0.78; 0.14)	0.1745
At 6 mo	7.0	7.0 (2.7)	84	7.2	7.7 (2.6)	83	0.48 (0.11; 0.85)	0.66 (0.29; 1.03)	−0.18 (−0.70; 0.34)	0.5014
At 8 mo	7.1	7.5 (2.7)	83	7.0	7.3 (2.3)	87	0.61 (0.22; 1.00)	0.58 (0.20; 0.96)	0.03 (−0.52; 0.58)	0.9192
At 12 mo	7.4	7.5 (2.3)	82	7.2	7.3 (1.9)	84	0.90 (0.52; 1.27)	0.71 (0.34; 1.09)	0.18 (−0.35; 0.71)	0.5025
At 18 mo	7.5	7.7 (2.2)	79	7.1	7.7 (3.3)	83	1.07 (0.64; 1.49)	0.61 (0.20; 1.03)	0.45 (−0.14; 1.05)	0.1363
Support (MQOL 0-10)
Preoperatively	8.3	8.5 (2.0)	88	8.3	8.5 (2.0)	91
1 wk postoperatively	8.2	8.5 (2.0)	89	8.2	8.5 (2.0)	90
At 4 mo	7.6	8.0 (2.0)	80	7.7	8.0 (2.0)	86	−0.59 (−0.97; −0.22)	−0.58 (−0.94; −0.22)	−0.02 (−0.54; 0.50)	0.9479
At 6 mo	7.6	8.0 (2.0)	84	7.8	8.0 (2.0)	81	−0.60 (−0.95; −0.25)	−0.48 (−0.83; −0.13)	−0.12 (−0.61; 0.38)	0.6482
At 8 mo	7.5	8.0 (2.0)	83	7.6	8.0 (2.0)	87	−0.75 (−1.14; −0.36)	−0.67 (−1.05; −0.28)	−0.09 (−0.64; 0.46)	0.7552
At 12 mo	7.9	8.0 (2.0)	82	7.8	8.0 (2.2)	84	−0.34 (−0.70; 0.01)	−0.45 (−0.80; −0.10)	0.11 (−0.39; 0.61)	0.6705
At 18 mo	7.7	8.0 (2.0)	79	7.4	8.0 (2.5)	83	−0.54 (−0.96; −0.11)	−0.83 (−1.24; −0.42)	0.29 (−0.30; 0.89)	0.3289

Within-group changes vs preoperatively moment and comparison of these changes between both groups are derived from the multivariate linear model for longitudinal measures.

^*Because the analysis has been performed on transformed values (inverse hyperbolic sign), the estimates for the within-group changes and between-group differences in these changes are back transformed and thus refer to ratios.

CI, confidence interval; DASS, Depression, Anxiety and Stress Scale; IQR, interquartile range; MQOL, McGill Quality of Life Questionnaire.

Considering the results of the moderator analysis, there was no evidence of a relation with the effect of the intervention on pain-related disability 1 year after surgery for any of the preoperative or postoperative variables (pain-related disability, pain intensity, upper limb function, pain-related catastrophizing, depression, anxiety, stress, psychological symptoms, existential well-being, and support) (Supplement 1, available at https://links.lww.com/PAIN/B759).

4. Discussion

This study shows that women who received PNE in addition to standard physiotherapy immediately after breast cancer surgery had no different change in pain-related disability from before surgery to 12 months postoperatively when compared with a group that received biomedical pain education. Similar results were observed for secondary outcomes including pain intensity, physical functioning, and emotional functioning up to 18 months after surgery.

Two previous studies investigated the effectiveness of perioperative PNE in women undergoing breast cancer surgery.^10,34 The first was a pilot randomized controlled trial that examined 2 different educational interventions given before breast cancer surgery.¹⁰ The intervention group watched a 90-minute pain psychoeducational video that included information on cognitions, emotions, and physiologic hyperarousal related to pain (n = 36), whereas the control group received digital education about health and nutrition (n = 32). Our study differed in that our educational interventions were provided postoperatively, and we had more and face-to-face educational sessions, additional physiotherapy, and a longer follow-up period. Similar to our findings, they found no significant effect for pain-related disability, pain intensity, and physical or emotional functioning up to 12 weeks after surgery. One possible explanation is that those with low pain scores were unable to recognize the pain information at the time provided, potentially impeding the learning process, which is important to the success of pain psychoeducation or PNE. This may also have influenced our findings, given that 54% of participants in our study rated their pain intensity lower than 30/100 at their first postoperative PNE session. In a study that found significant effects of perioperative PNE in a musculoskeletal pain population, perioperative pain scores were indeed higher.³²

The second study was a retrospective nonrandomized case–control trial that compared perioperative PNE (n = 51) to biomedical education (n = 51) in patients with persistent postoperative pain 1 year after breast cancer surgery.³⁴ Similar to our study, educational interventions were face-to-face, delivered by a physiotherapist with use of a pamphlet and combined with physiotherapy. In contrast to our findings, they found that PNE was more effective than biomedical education in reducing pain-related disability, pain intensity, symptoms related to altered central somatosensory functioning, and pain-related catastrophizing.³⁴ The effect sizes were all small (r = 0.20-0.29), with the exception of the effect size for pain intensity, which was found to be moderate (r = 0.31). One explanation for the discrepancy in findings could be the difference in study population. Although they excluded participants whose main pain was cancer treatment related other than persistent postsurgical pain (eg, chemotherapy-induced peripheral neuropathy and aromatase inhibitor–induced arthralgia), these patients were included in our study. This is because many breast cancer patients receive (neo)adjuvant therapies, and one-fifth of those who receive taxane-based chemotherapy will develop peripheral neuropathy,^54,58 and about half of women treated with aromatase inhibitors will experience hormonal therapy–induced joint pain.² If we excluded these patients, we would be limiting our ability to apply our study results to the general population after breast cancer. Aside from the study population, there was also a difference in the way educational interventions were integrated into the perioperative rehabilitation process. The other study provided an educational session before surgery and at each physiotherapy session (ie, once every 1-2 weeks for 3 months), whereas in our study, a fixed number of 6 educational sessions were provided alongside the physiotherapy sessions at predetermined timepoints after breast cancer surgery. The fact that educational sessions were not integrated into the physiotherapy sessions in our study (to specifically assess the effect of the educational interventions) may have also contributed to the difference in the additional effect of PNE.

Aside from the low prevalence of clinically relevant postoperative pain and the implementation of PNE alongside physiotherapy, a number of other factors could have influenced our study results. First, increased psychological distress in the early postoperative stage may have hampered the process of conceptual change.^21,24,52 The first 3 educational sessions were given within the first month after surgery. Increased psychological distress is common at this stage, which is known to negatively impact cognitive functions (eg, attention and memory).^21,24,52 Consequently, participants may have been less receptive/engaged in the educational sessions or not ready to reconceptualize pain. Second, conceptual change learning is shaped around challenging existing knowledge rather than simply learning new information.^42,63 Given that existing knowledge is often limited to a biomedical understanding of pain and the biomedical point of view is more widely accepted, it is possible that the intensity of PNE was insufficient to allow a paradigm shift to a biopsychosocial explanation of pain.^5,9 Third, the attention given to the project participants as a result of their participation may have influenced the results. Because both groups engaged with the same rigorous physiotherapy program at a specialized institution, the possibility of obtaining an additional effect from PNE may have been diminished (ceiling effect). A strong therapeutic alliance has been shown to increase satisfaction and outcomes in pain patients.^17,27 Despite a comprehensive physical therapy program at a specialized institution and a potentially strong therapeutic alliance, we were unable to completely prevent an increase in pain-related impairment compared with preoperative levels in either group at any follow-up. The objective of the current study was to achieve a greater change toward preoperative level in the PNE group than in the biomedical education group. We cannot, however, conclude that educational interventions are ineffective in breast cancer patients because we did not include a control group that received physiotherapy without any type of education.

Some limitations of the current study need to be acknowledged. First, sample size was calculated based on PDI scores in non–breast cancer populations, which could have led to an approximation that differed if PDI scores from a population with breast cancer were used instead. Second, both educational interventions were given by the same physiotherapist convinced of the importance of a biopsychosocial approach to pain, which might have influenced treatment fidelity. Treatment fidelity was not assessed in the study (eg, by recording educational sessions and inspecting recordings for forbidden elements). However, having the same therapist teach both educational interventions may have minimized the impact of nonspecific therapy factors. Third, no validated outcome measures assessing pain knowledge, attitudes, and beliefs were included. We also had to limit ourselves to a small number of moderators for the moderator analysis. We concentrated on preoperative and postoperative moderators that could be used in clinical practice to identify patients before or immediately after surgery who might benefit more or less from PNE after surgery. Fourth, the study design of a randomized controlled trial does not mimic a real-life situation and might have undermined the external validity. However, in addition to being a limitation, the research design could be viewed as a strength. Randomized controlled trials ensure internal validity and provide a rigorous tool for investigating cause–effect relationships between intervention and outcome. Other strengths of this study were its large sample size, long follow-up, double blinding, consistent assessment of the primary outcome parameter, and incorporation of maintenance sessions of education at 6, 8, and 12 months after surgery.

Minimizing symptom burden after treatment is paramount to restore quality of life after breast cancer. Pain neuroscience education is a convenient technique for improving pain-related functioning in persistent pain populations.⁶⁵ Despite the fact that we found no additional effect of PNE in patients immediately after breast cancer surgery, our findings add to the body of knowledge about PNE in this population and provide a basis for future research to fine-tune the optimal delivery format. Perhaps PNE should be integrated, administered only to patients experiencing postoperative pain or at risk of persistent pain, and only when patients are willing to receive information. If patients and healthcare professionals agree on treatment decisions, applying PNE would be an inherent choice based on the patient's individual needs and readiness, instead of a one-size-fits-all imposed formula.^14,22,66 This (more pragmatic) approach has the potential to enhance the process of conceptual change learning that PNE aims to accomplish.^25,26,53

5. Conclusion

Adding 6 sessions of PNE to physiotherapy after breast cancer surgery did not result in a better course of pain-related disability, pain intensity, and physical or emotional functioning up to 18 months postoperatively as compared with biomedical pain education. Future research on PNE should look into the effects of a more patient-tailored approach, depending on a patient's specific needs and readiness.

Conflict of interest statement

B. Morlion reports personal fees from Pfizer, Gruenenthal, Kyowa-Kirin, GSK, Reckit and Benckiser, and Shionogi, outside the submitted work. The other authors have no conflicts of interest to declare.

Appendix A. Supplemental digital content

Supplemental digital content associated with this article can be found online at https://links.lww.com/PAIN/B759.