Psychosocial Mechanisms of Cognitive-Behavioral–Based Physical Therapy Outcomes After Spine Surgery: Preliminary Findings From Mediation Analyses

Abstract

Objective

Changing Behavior through Physical Therapy (CBPT), a cognitive-behavioral–based program, has been shown to improve outcomes after lumbar spine surgery in patients with a high psychosocial risk profile; however, little is known about potential mechanisms associated with CBPT treatment effects. The purpose of this study was to explore potential mediators underlying CBPT efficacy after spine surgery.

Methods

In this secondary analysis, 86 participants were enrolled in a randomized trial comparing a postoperative CBPT (n = 43) and education program (n = 43). Participants completed validated questionnaires at 6 weeks (baseline) and 3 and 6 months following surgery for back pain (Brief Pain Inventory), disability (Oswestry Disability Index), physical health (12-Item Short-Form Health Survey), fear of movement (Tampa Scale for Kinesiophobia), pain catastrophizing (Pain Catastrophizing Scale), and pain self-efficacy (Pain Self-Efficacy Questionnaire). Parallel multiple mediation analyses using Statistical Package for the Social Sciences (SPSS) were conducted to examine whether 3- and 6-month changes in fear of movement, pain catastrophizing, and pain self-efficacy mediate treatment outcome effects at 6 months.

Results

Six-month changes, but not 3-month changes, in fear of movement and pain self-efficacy mediated postoperative outcomes at 6 months. Specifically, changes in fear of movement mediated the effects of CBPT treatment on disability (indirect effect = −2.0 [95% CI = −4.3 to 0.3]), whereas changes in pain self-efficacy mediated the effects of CBPT treatment on physical health (indirect effect = 3.5 [95% CI = 1.2 to 6.1]).

Conclusions

This study advances evidence on potential mechanisms underlying cognitive-behavioral strategies. Future work with larger samples is needed to establish whether these factors are a definitive causal mechanism.

Impact

Fear of movement and pain self-efficacy may be important mechanisms to consider when developing and testing psychologically informed physical therapy programs.

Chronic daily pain affects over 40 million individuals in the United States, with approximately 10.6 million people reporting high-impact pain that results in significant activity limitation or participation restriction.¹ Cognitive-behavioral therapy (CBT) is a well-established and effective strategy for reducing pain and disability in chronic pain populations.² Meta-analytical reviews have found small to medium effects on pain and disability when comparing CBT with usual care and small effects of CBT on disability relative to other active treatments.^2,3 CBT focuses on altering negative thoughts or feelings about the pain experience. In particular, CBT has been developed to target psychosocial risk factors for persistent pain, including elevated fear of movement and pain catastrophizing.⁴

Robust evidence supports the importance of psychosocial risk factors in the rehabilitation setting. Subsequently, studies have tested the effectiveness of combining CBT with physical therapy treatment to improve patient outcomes by addressing psychosocial risk factors of fear of movement, pain catastrophizing, and pain self-efficacy.^5–7 However, few studies have examined the mechanisms by which CBT for pain works, that is, what the mediators of treatment effects are in people with pain.⁸ Mansell et al⁹ observed reductions in fear of movement mediated the effects of a multidisciplinary intervention delivered by psychologists and physical therapists on disability in individuals with chronic low back pain (LBP). In contrast, Stevens et al¹⁰ did not find fear to be a mediator of a combined workplace intervention of CBT, ergonomics, and physical training on LBP-related outcomes. Smeets et al¹¹ found that decreased pain catastrophizing mediated the relationship between active therapy (including CBT, physical treatment, or combined treatment) and improvements in pain and disability in adults with chronic back pain, suggesting pain catastrophizing may not be a specific mediator of CBT but rather a common mechanism of active treatments. In contrast, pain self-efficacy—a positive psychosocial characteristic—may play a larger role in mediating CBT effects than factors like pain catastrophizing.^12,13 Currently, there is limited evidence on whether pain self-efficacy mediates CBT-based physical therapy interventions after spine surgery. Further examination of specific mediating factors of CBT-based rehabilitation interventions could improve theoretical models and aid the development of effective and efficient therapies.

Changing Behavior through Physical Therapy (CBPT) is a brief CBT^12,14,15 and self-management program^16,17 delivered by physical therapists over the phone. Archer et al¹⁸ demonstrated the clinical efficacy of CBPT in adults with chronic pain and high psychosocial risk undergoing spine surgery. Relative to participants assigned to an education group (eg, time and attention control), participants who participated in the CBPT program showed significantly greater reductions in pain and disability and improvements in physical health at 6 months after spine surgery.¹⁸ The overall hypothesis for the development of the CBPT program was that postoperative outcomes would improve through reductions in fear of movement and pain catastrophizing and increases in pain self-efficacy (Fig. 1). This hypothesis was informed by the fear-avoidance model of musculoskeletal pain, which posits that certain individuals exhibiting heightened fear of movement or pain catastrophizing or lower pain self-efficacy may be at risk for persistent pain and disability and worse physical health.^19,20 Archer and colleagues^21,22 and others^23–25 have found high fear of movement to be an independent predictor of poor outcomes related to pain, disability, and physical health after lumbar spine surgery, and pain catastrophizing and self-efficacy have been found to be associated with these outcomes in individuals with persistent musculoskeletal pain.^26–29

Figure 1.

Graphical representation of conceptual mediation model.

The purpose of this study was to examine whether the CBPT program improved postoperative pain, disability, and physical health through changes in fear of movement, pain catastrophizing, and pain self-efficacy. Multiple mediation analyses were conducted to test whether pre- to post-treatment changes at 3 and 6 months in pain-related cognitive factors were associated with improvement in outcomes at 6 months following surgery. We hypothesized changes in fear of movement, pain catastrophizing, and pain self-efficacy would mediate the effects of CBPT on 6-month outcomes.

Methods

Participants

Participants provided informed consent and were enrolled into a randomized trial (NCT01131611) comparing a postoperative CBPT intervention with an education control. Participants were recruited from a single academic medical center and included English-speaking adults 21 years of age or older who were undergoing lumbar spine surgery for a degenerative condition and reported high preoperative fear of movement based on a score 39 or greater on the Tampa Scale for Kinesiophobia (TSK). This score has been used as an indicator of individuals with dysfunctional pain beliefs and those at risk for poor outcomes after spine surgery.^21,30 Study inclusion criteria were having a laminectomy with or without arthrodesis for spinal stenosis, spondylosis with or without myelopathy, or degenerative spondylolisthesis; back and/or lower extremity pain for longer than 6 months; and no history of neurological movement disorder or psychotic disease. Exclusion criteria included having microsurgical techniques and surgery for spinal deformity, trauma, infection, or tumor. The Institutional Review Board at Vanderbilt University Medical Center approved the study.

Procedures

Participants were enrolled at a preoperative clinic visit, and their eligibility was confirmed after surgery. Participants completed an intake form and validated questionnaires at 6 weeks (baseline), 3 months (post-intervention), and 6 months following surgery to measure pain, disability, physical health, fear of movement, pain catastrophizing, and pain self-efficacy. Participants unable to return to the clinic for routine follow-up visits were asked to complete questionnaires at home and return in a self-addressed stamped envelope. Follow-up assessments were obtained by personnel unaware of randomized group status.

Interventions

Participants were randomized to (1) CBPT intervention, or (2) education control. The CBPT intervention and education control programs both lasted 6 weeks and consisted of weekly postoperative sessions with a trained study physical therapist (S.W.V.). In both groups, the first session was conducted in person at a standard 6-week postoperative clinic visit, and the remaining sessions were delivered over the telephone. All sessions were 30 minutes in length, except the first session, which was approximately 1 hour. Participants in both the CBPT and education groups received standard of care based on the discretion of their surgeon. Standard of care included postoperative advice to stay active, lifting restrictions, and referral to traditional physical therapy.

The CBPT intervention was developed based on brief CBT^12,14,15 and self-management programs.^16,17 CBPT included empirically supported strategies for behavioral self-management, problem solving, cognitive restructuring, and relaxation training.^14,15,31 Components of CBPT included cognitive-behavioral pain education, a graded activity plan, and weekly activity and walking goals. The study physical therapist helped participants identify distraction strategies, replace negative thoughts with positive thoughts, find balance between rest and activity, and manage setbacks by recognizing high-risk situations. Both the physical therapist and participant utilized an intervention manual that outlined the CBPT intervention (Suppl. Tab. 1).

The education control consisted of rehabilitation topics including the benefits of physical therapy, proper postoperative biomechanics, importance of daily exercise, ways to promote healing, and educational material on stress reduction, sleep, energy management, health provider communication, and injury prevention. Similar to the CBPT group, an education manual was used to guide topics. To ensure therapist fidelity and protocol adherence, all sessions were audio recorded and then reviewed. Additional details of the CBPT intervention and education control can be found in our previously published work.^18,32

Mediators

The proposed psychosocial mediators of fear of movement, pain catastrophizing, and pain self-efficacy were investigated. These mediators were chosen a priori based on our underlying theoretical model (fear avoidance model of musculoskeletal pain). All mediators were assessed at 6 weeks (baseline), 3 months, and 6 months after surgery.

Fear of movement

Fear of movement is the fear or anxiety a person experiences during physical movement or activity.³³ The 17-item TSK was used to measure fear of movement.³³ Participants rate items using a 4-point scale from “strongly disagree” to “strongly agree.” Total scores range from 17 to 68, with higher scores indicating greater fear of movement. The TSK has good psychometric properties and has been validated in the lumbar spine surgery population and individuals with various musculoskeletal conditions.^30,34,35

Pain catastrophizing

Pain catastrophizing is an exaggerated and negative outlook associated with actual or anticipated pain-related events.²⁹ The 13-item Pain Catastrophizing Scale (PCS) was used to assess catastrophic thinking associated with pain.³⁶ Participants rate items on a 5-point scale from “not at all” to “all the time.” Total scores range from 0 to 52. Scores greater than 24 differentiate “catastrophizers” and “noncatastrophizers.”³⁷ The PCS has demonstrated strong internal consistency, high test-retest reliability, and validity through associations with pain, disability, negative affect, and pain-related fear.^36,38,39

Pain self-efficacy

Pain self-efficacy is a person’s belief or judgement in his/her ability to accomplish a range of activities despite pain.⁴⁰ The Pain Self-Efficacy Questionnaire (PSEQ) was used to measure pain self-efficacy.⁴¹ The PSEQ is a 10-item questionnaire where confidence is rated on a 7-point scale from “not at all confident” to “completely confident.” Total scores range from 0 to 60, with scores greater than 40 indicating high self-efficacy.⁴² The PSEQ is a reliable measure of self-efficacy and demonstrates correlations with other psychosocial constructs such as depression, anxiety, coping strategies, pain ratings, and work-related tasks in individuals with chronic pain.⁴²

Outcomes

Outcomes included back pain intensity, disability, and physical health. All outcomes were measured at 6 weeks (baseline) and 6 months after surgery.

Back pain intensity

The Brief Pain Inventory was used to measure back pain intensity.⁴³ The back pain intensity scale includes 4 items assessing current, worst, least, and average pain. Participants rate each item on an 11-point scale, with 0 meaning “no pain at all” and 10 meaning “pain as bad as can imagine.” The Brief Pain Inventory is a reliable and valid pain measure used in surgical patients and individuals with chronic LBP.^44–46

Disability

The Oswestry Disability Index (ODI) was used to measure low back-related disability.⁴⁷ The 10-item ODI assesses 10 aspects of daily living, including pain intensity, lifting, sitting, standing, walking, sleeping, hygiene, traveling, and social and sex life. Participants rate each item using a 6-point scale ranging from 0 (high functioning) to 5 (low functioning), and a percentage score for disability is obtained. The ODI is a reliable measure of condition-specific disability and demonstrates moderate to high correlations with other disability measures.^48,49

Physical health

The 12-Item Short-Form Health Survey (SF-12) was used to measure physical health.⁵⁰ The physical component scale (PCS) of the SF-12 assesses 4 subdomains: physical functioning, role-physical, bodily pain, and general health. The PCS subscale score ranges from 0 to 100. Higher scores represent better physical health, and in the general population, the PCS scores have a mean of 50 with an SD of 10. The SF-12 PCS is a reliable, valid, and responsive measure in both generalized and various patient populations.^50–52

Data Analysis

All analyses were intent-to-treat and conducted using IBM SPSS Statistics version 26.0 (IBM Corp., Armonk, NY, USA). Pre- to post-treatment changes at 3 and 6 months were calculated for the proposed psychosocial mediators. Pearson correlations were performed to assess strength of association between mediators at each time point and between post-treatment change scores. Absolute correlations less than 0.3 were considered weak, between 0.3 and 0.5 were moderate, and greater than 0.5 were strong.⁵³

Multiple mediation analysis using ordinary least squares path analysis was performed to test whether 3-month and 6-month changes in fear of movement, pain catastrophizing, and pain self-efficacy mediate the effects of CBPT on pain, disability, and physical health at 6 months following surgery. Figure 1 illustrates the model guiding our mediation analysis. Multiple mediation provides estimates of an overall mediation effect (total indirect effect) and indirect effects for each mediator holding constant all other mediators in the model. Multiple mediation was performed using the PROCESS macro developed by Preacher and Hayes for SPSS.^54,55 A parallel mediator model was chosen as we did not have a priori hypotheses related to the serial relationship between proposed mediators and the effects of CBPT. We generated separate models for testing mediation involving 3-month and 6-month changes in psychosocial variables on each 6-month outcome. All models controlled for the baseline score of the outcome. Indirect effects (unstandardized path coefficients) and bias-corrected 95% CI for overall mediation and individual mediators were estimated based on 5000 bootstrap samples. Bootstrapping has been recommended as a superior approach for inferences in studies with small sample sizes and non-normally distributed variables.^56–58 Indirect effects with bias-corrected CIs not containing 0 were interpreted as evidence for mediation.

Role of the Funding Source

The funders played no role in the design, conduct, or reporting of this study.

Results

Participants

Participants (N = 86, mean [SD] age = 57.6 [12.2] years, N [%] females = 48 [55.8]) were randomly assigned to CBPT (n = 43) or education (n = 43). A Consolidated Standards of Reporting Trials (CONSORT) flow diagram and sample characteristics were previously published.¹⁸ Groups did not differ in demographic and clinical characteristics and baseline pain intensity, disability, and psychosocial measures (Tab. 1). A total of 93% of CBPT participants completed all treatment sessions compared with 95% of education participants. Follow-up rates at 6 months after surgery were similar across groups for participant-reported measures (CBPT: 88%; education: 97%). There were no significant differences in baseline characteristics between participants with and without complete follow-up data.

Table 1.

Baseline Characteristics of Sample^a

Characteristic	CBPT Intervention (n = 43)	Education Control (n = 43)
Demographic
Age, years	56.9 (11.1)	58.4 (13.3)
Sex, N (%) female	25 (58.1)	23 (53.5)
Race, N (%) White	36 (83.7)	33 (76.7)
Education, N (%) some college or more	30 (69.8)	32 (74.4)
Employment, N (%) working	21 (48.8)	18 (41.9)
Clinical
Fusion surgery, N (%) yes	29 (67.4)	31 (72.1)
Prior spine surgery, N (%) yes	17 (39.5)	17 (39.5)
Psychosocial scores 6 weeks after surgery
Fear of movement, mean (SD)	40.2 (7.3)	38.7 (6.5)
Pain catastrophizing, mean (SD)	14.5 (11.9)	11.5 (10.8)
Pain self-efficacy, mean (SD)	36.0 (16.8)	41.5 (15.2)
Outcome scores 6 weeks after surgery
Back pain intensity, mean (SD)	3.0 (2.2)	2.8 (2.0)
Disability, mean (SD)	38.8 (17.3)	34.0 (16.7)
Physical health, mean (SD)	29.8 (20.5)	32.1 (9.8)

^a Fear of movement measured by Tampa Scale for Kinesiophobia; pain catastrophizing measured by Pain Catastrophizing Scale; pain self-efficacy measured by Pain Self-Efficacy Questionnaire; back pain intensity measured by Brief Pain Inventory; disability measured by Oswestry Disability Index; physical health measured by physical component scale of the 12-Item Short-Form Health Survey.

At baseline (6 weeks after surgery), the mean (SD) score for fear of movement was 39.4 (6.9), for pain catastrophizing was 13.0 (11.4), and for pain self-efficacy was 38.8 (16.2) (Tab. 2). The proportion of participants at baseline reporting high fear of movement was 53.5% (n = 46), high pain catastrophizing was 14.0% (n = 12), and low self-efficacy was 50.0% (n = 43).

Table 2.

Descriptive Statistics for Mediators of Fear of Movement, Pain Catastrophizing, and Pain Self-Efficacy Assessed After Spine Surgery^a

	6 Weeks		3 Months		6 Months
	Mean (SD)	Min–Max	Mean (SD)	Min–Max	Mean (SD)	Min–Max
Fear of movement	39.4 (6.9)	24–61	36.9 (7.1)	23–53	35.1 (8.3)	18–61
Pain catastrophizing	13.0 (11.4)	0–50	10.6 (10.7)	0–40	10.2 (11.2)	0–42
Pain self-efficacy	38.8 (16.2)	6–60	43.4 (16.8)	8–60	46.0 (15.5)	1–60

^a Fear of movement measured by Tampa Scale for Kinesiophobia; pain catastrophizing measured by Pain Catastrophizing Scale; pain self-efficacy measured by Pain Self-Efficacy Questionnaire.

Association Between Proposed Mediators

Baseline (ie, 6 week) fear of movement was moderately associated with baseline pain self-efficacy (r = −0.47, P < .001). All other associations between psychosocial factors were strong (Tab. 3).

Table 3.

Associations Between Fear of Movement, Pain Catastrophizing, and Pain Self-Efficacy at 6 Weeks, 3 Months, and 6 Months After Surgery^a

	6-Week Associations		3-Month Associations		6-Month Associations
	Pain Self-Efficacy	Pain Catastrophizing	Pain Self-Efficacy	Pain Catastrophizing	Pain Self-Efficacy	Pain Catastrophizing
Fear of movement	−0.47	0.62	−0.62	0.64	−0.67	0.70
Pain catastrophizing	−0.56	—	−0.74	—	−0.73	—

^aValues are Pearson correlation coefficients and all are significant at P < .001.

Change in pain self-efficacy from baseline to 3 months was not associated with 3-month change in fear of movement (r = −0.13, P = .25) or 3-month change in pain catastrophizing (r = −0.18, P = .11). Three-month change in fear of movement and 3-month change in pain catastrophizing were moderately associated (r = 0.37, P = .001). Change in pain self-efficacy from baseline to 6 months was not associated with 6-month change in pain catastrophizing (r = −0.10, P = .39). Change in fear of movement from baseline to 6 months was associated to a weak degree with 6-month change in pain self-efficacy (r = −0.28, P = .01) and to a moderate degree with 6-month change in pain catastrophizing (r = 0.38, P < .001).

Mediation analyses

Multiple mediation analyses for CBPT effects on 6-month back pain intensity, disability, and physical health through changes in psychosocial variables at 3 months did not show evidence of mediation based on indirect effects (Tab. 4). Similar results were observed for the multiple mediation model involving 6-month psychosocial change variables and back pain intensity (Tab. 5).

Table 4.

Multiple Mediation Analyses Testing Indirect Effects of CBPT on 6-Month Back Pain Intensity, Disability, and Physical Health Through 3-Month Changes in Fear of Movement, Pain Catastrophizing, and Pain Self-Efficacy^a

	Back Pain Intensity		Disability		Physical Health
Effects of:	Coefficient	Bias-Corrected 95% CI	Coefficient	Bias-Corrected 95% CI	Coefficient	Bias-Corrected 95% CI
CBPT on outcome (total effect, c)	−0.8	−1.5 to −0.2	−9.1	−14.6 to −3.6	6.3	2.3 to 10.3
CBPT on outcome (direct effect, c’)	−0.7	−1.3 to −0.1	−7.5	−12.8 to −2.3	5.2	1.4 to 9.0
CBPT on fear of movement (a1)	−0.6	−2.8 to 1.6	−0.6	−2.8 to 1.6	−0.5	−2.8 to 1.7
CBPT on pain catastrophizing (a2)	−1.6	−5.3 to 2.0	−1.5	−5.2 to 2.1	−1.7	−5.4 to 1.9
CBPT on pain self-efficacy (a3)	4.2	−0.7 to 9.1	4.4	−0.5 to 9.3	3.7	−1.2 to 8.6
Fear of movement on outcome (b1)	0.08	0.01 to 0.15	0.6	0.1 to 1.2	−0.5	−0.9 to −0.1
Pain catastrophizing on outcome (b2)	0.003	−0.04 to 0.04	0.2	−0.1 to 0.6	0.1	−0.2 to 0.3
Pain self-efficacy on outcome (b3)	−0.01	−0.04 to 0.02	−0.2	−0.4 to 0.1	0.3	0.1 to 0.4
Effects of CBPT on outcome through:
Fear of movement (a1b1)	−0.05	−0.2 to 0.1	−0.4	−1.9 to 1.3	0.3	−1.1 to 1.4
Pain catastrophizing (a2b2)	−0.004	−0.1 to 0.2	−0.4	−1.6 to 1.1	−0.2	−1.3 to 0.2
Pain self-efficacy (a3b3)	−0.06	−0.2 to 0.1	−0.8	−3.2 to 0.7	1.0	−0.2 to 3.0
All mediators (a1b1 + a2b2 + a3b3)	−0.1	−0.4 to 0.2	−1.5	−4.5 to 1.4	1.1	−0.9 to 3.3

^a CBPT = Changing Behavior through Physical Therapy. Items in parentheses indicate mediation paths from conceptual model. Mediation models control for baseline outcome score.

Table 5.

Multiple Mediation Analyses Testing Indirect Effects of CBPT on 6-Month Back Pain Intensity, Disability, and Physical Health Through 6-Month Changes in Fear of Movement, Pain Catastrophizing, and Pain Self-Efficacy^a

	Back Pain Intensity		Disability		Physical Health
Effects of:	Coefficient	Bias-Corrected 95% CI	Coefficient	Bias-Corrected 95% CI	Coefficient	Bias-Corrected 95% CI
CBPT on outcome (total effect, c)	−0.8	−1.5 to −0.2	−9.1	−14.6 to −3.6	6.3	2.3 to 10.3
CBPT on outcome (direct effect, c’)	−0.4	−1.1 to 0.2	−4.9	−10.1 to 0.2	2.8	−1.3 to 6.9
CBPT on fear of movement (a1)	−3.7	−6.5 to −1.0	−3.9	−6.6 to −1.1	−3.8	−6.6 to −1.1
CBPT on pain catastrophizing (a2)	−1.2	−5.1 to 2.7	−1.0	−5.0 to 2.9	−1.4	−5.3 to 2.6
CBPT on pain self-efficacy (a3)	9.9	4.7 to 15.2	9.9	4.6 to 15.2	9.3	4.2 to 14.5
Fear of movement on outcome (b1)	0.1	0.002 to 0.1	0.5	0.1 to 0.9	−0.3	−0.6 to 0.05
Pain catastrophizing on outcome (b2)	0.02	−0.02 to 0.06	0.5	0.2 to 0.8	−0.05	−0.3 to 0.2
Pain self-efficacy on outcome (b3)	−0.02	−0.04 to 0.01	−0.2	−0.4 to 0.03	0.2	0.08 to 0.4
Effects of CBPT on outcome through:
Fear of movement (a1b1)	−0.2	−0.6 to 0.007	−2.0	−4.3 to −0.3	1.1	−0.2 to 2.7
Pain catastrophizing (a2b2)	−0.02	−0.2 to 0.1	−0.5	−2.6 to 1.5	0.07	−0.7 to 0.6
Pain self-efficacy (a3b3)	−0.2	−0.6 to 0.07	−1.7	−5.3 to 0.5	2.3	0.5 to 4.7
All mediators (a1b1 + a2b2 + a3b3)	−0.4	−1.1 to 0.03	−4.2	−9.6 to 0.3	3.5	1.2 to 6.1

^a CBPT = Changing Behavior through Physical Therapy. Items in parenthesis indicate mediation paths from conceptual model. Mediation models control for baseline outcome score.

For disability at 6 months, results from multiple mediation demonstrate that changes in psychosocial variables at 6 months revealed a significant indirect effect (Tab. 5; Suppl. Fig. 1). The CBPT intervention resulted in greater reductions in fear of movement (path a₁ coefficient = −3.9) and improvements in pain self-efficacy at 6 months (path a₃ coefficient = 9.9). Participants with greater reductions in fear of movement (path b₁ coefficient = 0.5) and pain catastrophizing at 6 months (path b₂ coefficient = 0.5) had lower disability at 6 months. Only changes in fear of movement were observed to be a mediator as evidenced through its indirect effect (path a₁b₁ coefficient = −2.0) and bias-corrected CI (−4.3 to −0.3).

Results from multiple mediation for CBPT effects on 6-month physical health through changes in psychosocial variables at 6 months revealed a significant indirect effect (Tab. 5; Suppl. Fig. 2). Similar to the disability model, the CBPT intervention resulted in greater reductions in fear of movement (path a₁ coefficient = −3.8) and improvements in pain self-efficacy at 6 months (path a₃ coefficient = 9.3). Participants with greater improvements in pain self-efficacy at 6 months had higher physical health at 6 months (path b₃ coefficient = 0.2). Changes in pain self-efficacy were observed to be a mediator as evidenced through its indirect effect (path a₁b₁ coefficient = 2.3) and bias-corrected CI (0.5 to 4.7).

Discussion

Beneficial effects on pain, disability, and general health were observed at 6 months after lumbar spine surgery in participants receiving a 6-week physical therapist-delivered CBPT program; this study was designed to examine the potential psychosocial mechanisms underlying these outcomes. In a multiple mediation model, 6-month changes in fear of movement mediated the effects of CBPT on 6-month disability, while 6-month changes in pain self-efficacy mediated the effects of CBPT on 6-month physical health. Pain catastrophizing was not found to be a mediator following the CBPT program.

The results of our mediation analyses do not provide definitive evidence that fear of movement and pain self-efficacy are causal mechanisms for CBPT effects on disability or physical health. This perspective is based on the finding that we did not observe a change in psychosocial variables at 3 months to have indirect effects on 6-month outcomes. For a mediator to be considered in the “causal path,” it is expected that a change in the mediator should precede a change in the outcome variable. In the current study, only 6-month changes in fear of movement and pain self-efficacy mediated the effects of CBPT on 6-month disability and physical health, respectively. Our findings may be explained by a latency effect. A greater effect over time may occur because of the time needed to adopt new skills or strategies for coping or changing behavior. DasMahapatra et al⁵⁹ reported greater mediation effects at a 6-month follow-up compared with 1 or 3 months after an online CBT self-management program in participants with chronic pain. Bergbom et al⁶⁰ and Sullivan and colleagues^61,62 also highlighted the importance of both early (eg, sequential) and late (eg, concurrent) changes in psychosocial constructs for predicting pain-related outcomes. Longer-term follow-up is needed for the CBPT intervention to determine if psychosocial changes from baseline to 6 months mediate participant-reported outcomes at 1-year following spine surgery.

Our mediation analyses focused on our theoretical psychosocial constructs of fear of movement, pain catastrophizing, and pain self-efficacy. As has been suggested by prior work,⁶³ these factors are thought to represent distinct yet overlapping constructs (supported by our correlation findings). While these psychosocial factors showed substantial correlations with each other at each time point, their longitudinal change (our mediating variable) did not. The lack of strong correlation in change may alleviate concerns for multicollinearity in our statistical modeling. This finding suggests psychosocial changes after treatment are not redundant, especially when considering the association between changes in pain self-efficacy with the other negative psychosocial factors. Fear of movement, pain catastrophizing, and pain self-efficacy have been studied in the context of chronic pain and are known determinants of persistent disability.^20,64,65 Often, studies examine these factors in isolation; however, Wideman et al⁶⁶ have shown that simultaneous consideration of these factors within statistical models may be more pertinent for risk estimation. Multidimensional screening tools have been developed to account for various contributions of psychosocial constructs on risk and for comprehensive monitoring over a course of treatment.^67–70 Moreover, studies aimed at examining multiple treatment effects may show a dominant construct(s), which has implications for how treatment is applied.^13,71

Psychosocial constructs are expected to remain stable postoperatively in the absence of targeted therapy.^21,22 Studies by Abbott et al⁷² and Monticone et al⁷³ suggest that postoperative, in-person, cognitive-behavioral interventions can improve fear immediately after spinal fusion surgery above what is expected by standard postoperative exercise alone. While we did not observe an immediate benefit following CBPT on fear of movement, we did see greater reductions at the 6-month time point that mediated the effect on 6-month disability. This was similar to the effects we observed on pain self-efficacy and 6-month physical health. Across multiple psychosocial variables, including fear, Turner et al¹³ found self-efficacy to have a unique mediating role following CBT. The results of the current study suggest potential disability and health benefits of postoperative rehabilitation strategies that not only reduce negative psychosocial characteristics such as fear of movement but also increase positive characteristics such as pain self-efficacy. For example, targeted application of treatment may need to consider strategies such as graded activity or exposure in vivo for fear of movement, and goal-setting, problem solving, or mindfulness training for pain self-efficacy.^74–78

The hypothesis that pain catastrophizing would be a mediator was not supported. Preliminary work has shown pain catastrophizing to have a negative influence on pain-related disability in patients after lumbar spine surgery.^26,79 Monticone et al⁷³ found pain catastrophizing to be a potential mediator when testing a combined cognitive-behavioral and exercise intervention after spinal fusion. In participants with chronic LBP, earlier CBT studies by Smeets et al¹¹ and Spinhoven et al⁸⁰ reported a mediating role of pain catastrophizing. Burns and colleagues⁸¹ reported a lack of treatment-specific change in pain catastrophizing following CBT for chronic pain. Our findings for pain catastrophizing may be due to the limited variability and low values at the postoperative baseline visit. Targeting CBPT to individuals with high levels of pain catastrophizing may have yielded a more robust effect. Another explanation may be the CBPT program’s greater focus on graded activity and goal setting relative to cognitive restructuring techniques, which are commonly used to address pain catastrophizing.⁷⁷ This may have resulted in CBPT being more conducive to altering fear of movement and pain self-efficacy compared with pain catastrophizing.

The current study included pre-specified a priori mediation analyses from a randomized trial. However, the study may not be adequately powered to detect mediation effects, and results need to be interpreted with caution.⁸² The results of our study should be viewed as preliminary and exploratory. Future mediation analyses should involve larger samples to validate findings from the current study and possibly disentangle more complex relationships between the proposed mediators. Our study enrolled participants with high preoperative fear of movement. However, individuals with varying levels of pain self-efficacy and lower levels of pain catastrophizing were included. The CBPT intervention may be more efficacious for individuals with higher levels across all targeted mediators. As mentioned, our results demonstrate concurrent mediation only, that is, a change in mediators did not appear to precede a change in clinical outcomes. To establish a more definitive causal path, designs involving lagged or cross-lagged analyses could be considered. Only 6-month outcome data are included in the current study. It is possible that the effects of CBPT may be reduced at longer term follow-up.⁸³ The current analyses are limited to only those cognitive mediators included within the mediation model. It is possible that other factors—unmeasured or not included in modeling—may influence our findings. Other potential psychosocial mediators of CBT include social support, acceptance, and coping as well as measures of behavioral change (ie, activity participation, adherence). Non-specific or contextual factors involving the patient-therapist relationship or expectations may also underlie the therapeutic effects of CBPT. Future trials should consider how additional constructs mediate the effects of CBT-based interventions.

The results of our exploratory mediation analyses suggest fear of movement (for disability) and pain self-efficacy (for physical health) may be important mediators of CBPT. Future work with the CBPT intervention should establish whether psychosocial changes in the first 6 months after surgery mediate long-term patient-reported outcomes following spine surgery. Fear of movement and pain self-efficacy may be important mechanisms to consider when developing and testing psychologically informed physical therapy programs.

Author Contributions

Concept/idea/research design: R.A. Coronado, D.M. Ehde, S.W. Vanston, T. Koyama, M.J. McGirt, D.M. Spengler, O.S. Aaronson, J.S. Cheng, C.J. Devin, S.T. Wegener, K.R. Archer

Writing: R.A. Coronado, D.M. Ehde, T. Koyama, D.M. Spengler, C.J. Devin, S.T. Wegener, K.R. Archer

Data collection: S.W. Vanston, S.L. Mathis, S.T. Wegener

Data analysis: J.S. Pennings, T. Koyama, S.E. Phillips

Project management: S.L. Mathis, K.R. Archer

Fund procurement: K.R. Archer

Providing institutional liaisons: S.T. Wegener

Clerical/secretarial support: R.A. Coronado

Consultation (including review of manuscript before submitting): R.A. Coronado, D.M. Ehde, J.S. Pennings, S.W. Vanston, T. Koyama, S.E. Phillips, S.L. Mathis, M.J. McGirt, D.M. Spengler, O.S. Aaronson, J.S. Cheng, C.J. Devin, S.T. Wegener, K.R. Archer

Ethics Approval

This study was approved by the Institutional Review Board at Vanderbilt University Medical Center.

Funding

This study was supported by a research grant (R21AR062880) from the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health, and a Magistro Family Foundation grant from the Foundation for Physical Therapy Research. This study used REDCap as the secure database, which was supported by a CTSA award (UL1TR000445) from the National Center for Advancing Translational Sciences.

During manuscript development and submission, Dr Coronado was supported by a Vanderbilt Faculty Research Scholars award.

Disclosures and Presentations

The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.

Data from this study were presented at the American Physical Therapy Association Combined Sections Meeting, February 16, 2017, San Antonio, Texas, and posted online at: Archer K, Coronado RA, Ehde D, et al. Fear of Movement and Pain Self-Efficacy Mediate Outcomes Following a Targeted Rehabilitation Intervention After Spine Surgery. Abstract OPL4. Orthopaedic Section Platform Presentations (Abstracts OPL1–OPL64). J Ortho Sports Phys Ther. 2016;47:A1-A29. https://www.jospt.org/doi/10.2519/jospt.2017.47.1.A1#_i129.