Abstract
Perspective
This pilot study compared the effects of a 1-day preventive behavioral intervention (ACT) to TAU in at-risk veterans undergoing orthopedic surgery. Three months following the intervention, veterans receiving ACT exhibited quicker cessation of pain and opioid use. Focusing on preoperative pain management may help prevent chronic postsurgical pain.Free full text
Acceptance and Commitment Therapy for Prevention of Chronic Post-surgical Pain and Opioid Use in At-Risk Veterans: A Pilot Randomized Controlled Study
Abstract
High levels of pain, significant anxiety, and/or depressive symptoms prior to surgery put patients at elevated risk forchronic pain and prolonged opioid use following surgery. The purpose of this preliminary study was to assess the efficacy of a one-day Acceptance and Commitment Therapy (ACT) workshop in “at-risk” Veterans for the prevention of chronic pain and opioid use following orthopedic surgery. In a randomized controlled trial, 88 at-risk Veterans undergoing orthopedic surgery were assigned to Treatment as Usual (TAU; N =44) or TAU plus a 1-day ACT workshop (N = 44). Pain levels and opioid use were assessed up to 3 months following surgery. Pain acceptance and values-based behavior were assessed at baseline and 3-month follow-up. Participants who completed the ACT workshop reached pain and opioid cessation sooner than those in TAU. Post-operative complications exhibited a moderating effect on these outcomes such that the effects of ACT were greater in patients without complications. Increases in pain acceptance and values-based behavior, processes targeted in ACT, were related to better outcomes. These promising results merit further investigation in a larger clinical trial. Providing an intervention prior to surgery for at-risk veterans has the potential to change clinical practice from a focus on management of postoperative pain to prevention of chronic pain in at-risk individuals. This study is registered at clinicaltrials.gov no.01364870.
Introduction
Chronic post-surgical pain (CPSP) is a common and debilitating problem that occurs in 5-85% of patients following a variety of surgeries22. Orthopedic surgeries are consistently associated with high postoperative pain, with the prevalence of chronic pain at 3–24 months to be 20% or more after Total Knee Arthroplasty (TKA) and 8% or more after Total Hip Arthroplasty (THA)2,4,43,33,11. Disabling pain has a dramatic impact on functioning, quality of life, mood, sleep, cognitive functions and social life6,38.
Chronic pain is also highly correlated with prolonged opioid use following surgery. For example, an examination of the Total Joint Replacement Registry between 2008 and 2011 found that 41% of patients who underwent TKA were still using opioids for their knee pain after 90 days.30 Growing evidence shows that repeated opioid use causes “opioid-induced hyperalgesia” (i.e. increased pain sensitivity caused by opioid exposure), making pain control even more difficult20,21. When pain and addiction are present, the patient’s level of desperation rises16, medical and psychiatric illness severity increases, and there is a greater utilization of health care3.
In recent years, attention has increasingly focused on identifying predictors of postoperative pain, in hopes that this would facilitate early intervention and better pain management. In several reviews, preoperative pain consistently emerges as an important predictor for the development of chronic postoperative pain15,17. Additionally, patients with significant psychological distress (anxiety and/or depression) prior to surgery are at high risk of having their pain persist after surgery10,37,31. Thousands of orthopedic surgeries are performed yearly within the Veterans Affairs health care service. In this population, distress-based conditions, including post-traumatic stress disorder, other anxiety disorders, and depression, are common, particularly for those who have experienced atrocities or massacres during deployment34,35. These Veterans are at high risk for developing CPSP and are the patients targeted in the present study.
Acceptance and Commitment Therapy (ACT) is a trans-diagnostic (applicable to a variety of conditions) and evidence-based treatment for depression, chronic pain, and mixed anxiety7. However, the use of ACT as a preventive intervention for chronic post-surgical pain and prolonged opioid use is still novel. To date, there has only been one non-randomized study1 and a detailed case report42 describing the effects of ACT on CPSP and opioid use. Importantly, ACT has been effectively implemented in various treatment-delivery formats, including one-day workshops8. Flexibility in delivery format and duration allows focus to be placed on how best to package and deliver the intervention to meet the needs of the patient population, ensure treatment adherence, and increase successful dissemination into everyday clinical settings.
In summary, high levels of preoperative pain, depression, and/or anxiety put Veterans at high risk for chronic pain and prolonged opioid use following surgery. Providing a preventive intervention for the substantial number of Veterans who experience significant distress or pain prior to surgery may prevent deleterious outcomes, including CPSP and prolonged opioid use. ACT is an evidence-based treatment for chronic pain and distress. However, its use as a preventive intervention for CPSP and prolonged opioid use has not studied.
The aim of this pilot study was to determine if a 1-day preoperative ACT workshop with an individualized “booster” session postoperatively reduces the length of pain and opioid use following surgery when compared to treatment as usual (TAU). We hypothesized that the 1-day ACT workshop would lead to quicker cessation of pain and opioid use compared to TAU.
Study Design
This was a single-blinded, prospective, randomized controlled study with assignment to one of two groups: treatment as usual (TAU) or ACT plus TAU (ACT + TAU).
Setting/Sample
With approval of the local Institutional Review Boards, veterans scheduled for an orthopedic surgery were recruited from the Iowa City Veterans Affairs Medical Center (VAMC). The Iowa City VAMC serves Veterans living in 50 counties in Eastern Iowa, Western Illinois and Northern Missouri, has a surgical complexity designation of “complex,” and serves as a tri-level care system including primary, secondary and tertiary care to Veterans requiring orthopedic surgery.
Inclusion criteria for patients in this study were: 1) Scheduled for orthopedic surgery that typically requires at least six weeks of postoperative rehabilitation; and 2) identified to be at-risk for chronic post-surgical pain and prolonged opioid use. At-risk criteria included severe preoperative pain (> 7 on a 0-10 scale) or moderate preoperative pain (3.5 - 7 on a 0-10 scale) with anxiety (≥15 on the clinician administered Hamilton Rating Scale for Anxiety [HAM-A]) or depression (≥17 on the Hamilton Rating Scale for Depression [HAM-D]). The cuts-offs for anxiety and depression indicate significant levels of anxiety and depression36,29. Exclusion criteria included bipolar or psychotic disorder, history of brain injury, inability to complete study forms due to mental incapacity or language barrier, imminent risk of suicide, and participation in ACT or mindfulness-based therapy during the previous year.Consistent with CONSORT guidelines, de-identified information were tracked regarding the number of eligible patients who declined to enroll, who were excluded after enrollment, or who enrolled but withdrew prior to completing the entire protocol.
Randomization and Blinding
Prior to enrollment of participants, a randomization sequence was generated by the study statistician (author M.B.Z.) using SAS v9.2 PROC PLAN with randomization occurring in permuted blocks of 4 and 6, stratified by type of orthopedic surgery (TKA, THA, shoulder, spine, other). The study statistician was not involved in patient recruitment or enrollment. Allocation to treatment group remained concealed via background programming until consent and inclusion/exclusion criteria were determined. The principal investigator and research assistants assessing outcomes were blinded to group assignment. The project director and the recruiter were not blinded so they could be available throughout the study to ask and answer questions regarding the participants’ treatment.
Intervention Groups
Treatment-as-Usual (TAU)
All participants received TAU. Pre-surgery treatment included a nurse-led patient education class covering the post-operative course and what to expect for pain control and recovery. Participants were prescribed analgesics, sedatives and/ or anxiolytics immediately prior to surgery. Intraoperatively, regional (i.e., spinal and femoral) anesthesia and analgesia was given. In the immediate postoperative period, all participants received opioids; some patients also received non-opioids, anticonvulsants and/or anxiolytics. These were administered either around the clock or as needed until discharge. Participants were sent home with analgesia (often a combination medication of an opioid and acetaminophen) for breakthrough pain.
Acceptance and Commitment Therapy (ACT) plus TAU
Participants randomized to receive ACT were scheduled to attend a 1-day workshop before their preoperative clinic visit. Food and money for transportation were provided to facilitate attendance. The 5-hour ACT workshops were co-facilitated by two clinical psychologists. Topics covered in the workshop included: 1) Acceptance and Mindfulness Training emphasizing new ways of managing troubling thoughts, feelings, and physical sensations (e.g., learning how to recognize, and develop cognitive distance from, unhelpful thoughts such as “I can’t take this pain anymore” or “This is too much to bear”) and learning how to willingly face experiences that cannot be changed; and 2) Behavioral Change Training involving a) teaching patients how to recognize ineffective patterns of behavior and habits, b) exploring and setting life goals and goals related to mental and physical health, and c) promoting effective and committed actions to achieve these goals despite the urge to do otherwise. The workshop was structured around the ACT Matrix32.
Overall, the goal of the ACT intervention was to increase patients’ awareness of thoughts, feelings, and behaviors that exacerbate pain (such as struggling with it) and recognition of avoidance strategies that interfere with life engagement. For example, pain involves both the noxious physical experiences (e.g., physical sensations of stabbing, throbbing, etc.) and the reactions to them, which include thoughts (e.g., this is awful, I can’t bear this, I can’t do anything), feelings (e.g., hopelessness), and behaviors (e.g., avoidance of activities). The ACT intervention does not focus directly on the noxious experiences themselves but, instead, on the reactions to them and the effect that these reactions have on mental health and functioning. This is done by teaching new ways to respond to thoughts and feelings related to pain – acceptance and mindfulness – and encouraging behavioral (re)engagement in meaningful life activities.
Participants who attended an ACT workshop were provided a manual reinforcing the training exercises to use at home. They were also scheduled to receive an individualized “booster” session over-the-phone with one of the workshop facilitators 2-4 weeks after workshop attendance.
Primary Outcome Measures
Pain and Opioid Cessation
After surgery, participants used the Daily Log of Pain and Pain Medication (DLPM) to record their maximum pain intensity each day and the total amount of each pain medication taken each day until pain and opioid cessation was achieved (pain < 3 for 5 consecutive days; zero doses of opioids recorded for 5 consecutive days). The medical record was reviewed for pain medications prescribed at discharge and the name and dose of each was included on the DLPM. The DLPM was either delivered to the participant in the hospital shortly before discharge or mailed to the participant’s home after discharge. The medical record was reviewed again at follow-up visits to confirm that opioid cessation by non-renewal of prescription was consistent with self-report.
Process Measures
Pain Acceptance
Pain acceptance was measured using the Chronic Pain Acceptance Questionnaire (CPAQ)27. This is a 20-item questionnaire to assess Pain Willingness (i.e. the willingness to have pain without trying to avoid or reduce it); and Activity Engagement (i.e. the degree to which one engages in life activities regardless of pain). Each item is scored on a 7-point Likert scale ranging from 0 (never true) to 6 (always true). The CPAQ Total Score is the sum of the two subscales. The CPAQ has been used in treatments of chronic pain patients, is related to improvements in emotional functioning, and has demonstrated good reliability and validity. It yields valid inferences regarding acceptance of pain, based on significant negative correlations with measures of avoidance and emotional distress and positive correlations with measures of patient functioning27. In this study, reliability values ranged from 0.89 at enrollment to 0.87 at follow-up.
Engagement in Values-Based Behaviors
Engagement in values-based behavior was measured using the Chronic Pain Values Inventory (CPVI)28. Respondents are asked to rate the importance of six domains of values: family, intimate relations, friends, work, health, and growth or learning. Respondents are also asked to rate their success living according to their values in each domain over the past two weeks. Each item was rated on a scale from 0 (not at all important/successful) to 5 (extremely important/successful). The mean discrepancy (i.e. difference between importance & success) ratings were used as a measure of Values-based Action: the extent to which patients see their behavior as guided by their values. Lower scores on the discrepancy measure indicate stronger values-based engagement and activation. Studies support the internal consistency and construct validity of this measure as a reflection of values-based engagement and activation24,23. In this study, Cronbach’s alpha were greater than .80 for both scales.
Other Variables
Other variables were assessed at baseline for the purpose of characterizing the sample and as potential control variables, based on the success of randomization. They included demographics, preoperative opioid use, pain intensity and interference using the Brief Pain Inventory (BPI)19, opioid risk using the Opioid Risk Tool (ORT)40, substance use disorder using the CAGE-AID5, and post-traumatic stress disorder (PTSD) symptoms using the PTSD checklist for DSM-V (PCL-5)39. These are all standardized questionnaires with established reliability and validity in Veteran and pain populations. The Hamilton Rating Scales for Anxiety and Depression were only used to identify participants who met “at-risk” criteria.
Procedures
Following consent, participants completed baseline anxiety and depression assessments by a clinical psychologist. Those who met at-risk criteria were then asked to additionally complete the demographics survey, the Brief Pain Inventory, the Chronic Pain Acceptance Questionnaire, the Chronic Pain Values Inventory, the Opioid Risk Tool, the CAGE-AID and the PCL-5.
Eligible participants were then randomized to either Treatment as Usual (TAU) or ACT+TAU. Participants randomized to ACT+TAU were scheduled for a 1-day ACT workshop prior to surgery. Those who were unable to attend a workshop prior to surgery were scheduled for a workshop soon after surgery. All participants were provided 14 weeks of daily logs (DLPM). Participants were instructed to return the logs on a weekly basis following surgery and were called each week to be reminded or thanked, as appropriate. Three months after surgery, participants completed surveys on pain acceptance (CPAQ) and engagement in values-based behavior (CPVI) during an in-person visit or by phone. All post-operative assessments were conducted by a blinded assessor and participants were instructed not to discuss whether they attended a workshop or not so the assessor would remain blinded.
Data analysis
Study data were collected and managed using REDCap (Research Electronic Data Capture)13. Data are described using percentages for categorical variables and mean ± standard deviation (SD) or median and 25th to 75th percentiles (interquartile range [IQR]) for continuous variables. Demographic and baseline clinical variables were compared between ACT+TAU and TAU, using two-sample t-test or Wilcoxon rank-sum test for continuous variables and Pearson chi-square test of Fisher’s exact test for categorical variables. Baseline variables that were found to differ between groups were used as covariates in assessing the efficacy of ACT+TAU compared to TAU.
Intent-to-treat (ITT) analyses were performed including all participants randomized to treatment groups that had surgery, regardless of adherence to treatment protocols. Since randomization had to occur prior to surgery to provide the ACT workshop preoperatively, and because end-point outcomes of pain and opioid cessation are based on post-operative pain and post-operative opioid intake, randomized participants who did not end up having surgery were not included in the ITT analyses. The efficacy of ACT+TAU compared to TAU was also examined in a per-protocol analyses that excluded 9 participants randomized to ACT+TAU; these 9 participants did not complete the treatment protocol (8 did not attend a workshop and 1 did not meet the study protocol requirements).
Efficacy of ACT+TAU compared to TAU for post-operative pain cessation and postoperative opioid cessation was examined using survival analysis methods. Time to cessation from date of surgery was used as the endpoint. Those without the endpoint were considered as censored observations at the time of last known follow-up. Kaplan-Meier curves were constructed showing the product-limit estimate of the cumulative probability of survival (non-cessation) at the follow-up times for the observed endpoints. Median time to cessation was also computed for each endpoint. ACT+TAU treatment effect size relative to TAU, expressed as hazard ratio for these endpoints, was estimated by fitting a Cox proportional hazard regression model that included treatment group, and covariates to account for baseline group differences. The hazard ratio compares the hazard rate in ACT+TAU versus TAU groups, with the hazard rate defined as the probability that at a given time t, at an individual who has remained event-free up to that time will experience the event. For this study, the hazard ratio is the relative likelihood of cessation in the ACT versus TAU subjects at any given point in time.
In addition, the Cox regression model was expanded to examine presence of postoperative complication as an effect moderator. There were 4 participants in each group who experienced postoperative complications that influenced their postoperative recovery (including surgical site cellulitis requiring re-hospitalization and IV antibiotics; gallbladder infection requiring re-hospitalization; deep vein thrombosis; urethral damage due to catheter insertion and subsequent prolonged pain and incontinence; kidney stones requiring placement of bilateral stents and weekly lithotripsy; staph infection at surgical site requiring re-hospitalization and surgery for irrigation and debridement with polyethylene exchange; and two cases of severely limited range-of-motion requiring manipulation under anesthesia).
The effect of ACT+TAU on mean change in CPAQ and CPVI was examined using linear mixed model analysis for repeated measures, with treatment group, time, and treatment by time interaction as fixed effects. Efficacy of ACT+TAU versus TAU was expressed as the difference in mean change from baseline at 3 months (with 95% CI).
All statistical analyses were performed using SAS (version 9.4). SAS procedures LIFETEST and PHREG were used for the cessation analyses, and the MIXED procedure was used for the CPAQ and CPVI analyses.
Results
Between Jan 2015 and August 2016, 451 Veterans were scheduled for orthopedic surgery. Of these, 32 (7%) met exclusion criteria based on review of medical records. The remaining 419 were approached by letter and then by phone, and, if interested, consented and screened. Of the 419 approached, 187 (45%) declined to participate1, 104 (25%) could not be reached or screened before surgery, 32 (7%) did not meet inclusion criteria (i.e. surgery cancelled or did not have sufficient pain, anxiety and/or depression to meet the at-risk criteria); and 8 (2%) were excluded following chart review and prior to randomization (N=7 due to exclusionary medical conditions and N=1 non-English speaking). The remaining 88 individuals were randomized to treatment group (44 to TAU and 44 to ACT+TAU). Of the 44 randomized to each group, 36 in the TAU group and 40 in the ACT+TAU group received surgery (8 in the TAU group and 4 in the ACT+TAU group had their surgery cancelled or postponed beyond the study period). (See Figure 1 for CONSORT Diagram).
Consort Diagram.
Consort diagram depicting participant enrollment, randomization to groups, and follow-up visits. Note. TAU=Treatment as usual; ACT=Acceptance and Commitment Therapy; AE = Adverse event.
These 76 Veterans ranged in age from 25 to 83 years with a mean (± SD) of 62.6 (± 10.4) years. Participants were primarily male (94%), married (54.5%), non-Hispanic/Latino (95.5%), Caucasian (82.5%), had completed a vocational degree or some college (52.5%), had been in a combat zone (57%), were scheduled for total knee arthroplasty (68%), had at least one other chronic pain condition (71.5%) and were not taking opioids for their pain (68%). Significantly more participants randomized to TAU had been in a combat zone (p=0.04) and this group also had significantly better baseline CPVI scores (p=0.004-0.005). Surprisingly, all 5 female Veteran participants were randomized to the ACT+TAU group (p=0.06). There were no other significant differences in baseline characteristics between treatment groups (see Table 1). Veterans who had surgery (and were, therefore, included in the Intent-To-Treat analyses) were not significantly different from Veterans who did not have surgery (and were, therefore, excluded from the ITT analyses), except that Veterans who had surgery were significantly more likely to have been in a combat zone (p=0.03). Of note, all Veterans who attended the ACT workshop completed it. Twenty Veterans attended the workshop before surgery (median: 44 days preoperatively) and 11 attended it after surgery (median: 35 days postoperatively). Of those who had surgery and completed the workshop, 91% completed the booster session. Furthermore, the follow-up rates were excellent with 90% of the participants in ACT+TAU and 91% of those in TAU completing the 3-month follow-up visit.
Table 1
Participant characteristics by treatment
| Variable | ACT (n=40) | TAU (n=35) | p-value |
|---|---|---|---|
|
| |||
| Age, mean (SD) | 62.2 (10.1) | 63.0 (10.7) | 0.737 |
|
| |||
| Sex - Male | 35 (88%) | 35 (100%) | 0.057 |
|
| |||
| Marital Status | (n=39) | (n=34) | |
| 22 (56%) | 18 (53%) | 0.766 | |
|
| |||
| Ethnicity | (n=35) | (n=32) | |
| 35 (100%) | 29 (91%) | 0.104 | |
|
| |||
| Race | |||
| 34 (85%) | 28 (80%) | 0.568 | |
|
| |||
| Education | (n=39) | (n=34) | |
| 11 (31%) | 11 (33%) | ||
| 22 (56%) | 19 (56%) | 0.399 | |
| 6 (15%0 | 4 (12%) | ||
|
| |||
| Combat Zone (Yes) | 18 (45%) | 24 (69%) | 0.040 |
| 10 (26%) | 17 (49%) | 0.041 | |
|
| |||
| Surgery | |||
| 26 (65%) | 25 (71%) | 0.524 | |
| 7 (18%) | 7 (20%) | ||
| 7 (18%) | 3 (9%) | ||
|
| |||
| Other Pain Conditions (Yes) | 28 (72%) | 24 (71%) | 0.910 |
|
| |||
| Opioid use Preoperatively (Yes) | 12 (30%) | 12 (34%) | 0.691 |
|
| |||
| ORT, median [IQR] | 0.5 [0-4] | 1 [0-6] | 0.419 |
|
| |||
| CAGE, median [IQR] | 1 [0-3] | 1 [0-2] | 0.343 |
|
| |||
| PCL-5 | 12.5 [5.5-22.5] | 8 [3-18] | 0.236 |
|
| |||
| BPI severity, mean (SD) | 4.8 (1.9) | 5.0 (2.5) | 0.653 |
|
| |||
| BPI interference | 5.3 (2.2) | 5.0 (2.0) | 0.609 |
|
| |||
| CPAQ Total, mean (SD) | 66.8 (19.7) | 68.2 (20.4) | 0.761 |
|
| |||
| CPVI Mean Success | 2.3 (1.1) | 3 (1) | .005 |
|
| |||
| CPVI Discrepancy score | 1.4 (1.0) | 0.8 (0.9) | 0.004 |
Note: BPI=Brief Pain Inventory; CPAQ=Chronic Pain Acceptance Questionnaire; CPVI=Chronic Pain Values Inventory; ORT = Opioid Risk Tool; CAGE: Substance Use Risk.
Post-operative Pain and Opioid Cessation
From Kaplan-Meier curves for ITT analyses, median days to pain cessation was 66 (IQR: 28 days to > 3 months) for the ACT+TAU group and 74 (IQR: 32 to > 3 months) for the TAU group (see Figure 2a). The pain cessation hazard ratio for ACT+TAU (vs. TAU) was 1.42 (95% CI: 0.68, 2.95) after controlling for baseline differences in gender, combat zone, and baseline CPVI discrepancy score and mean success. Similar, but stronger results were obtained in the per protocol analysis, with median days to pain cessation for the ACT+TAU group of 61 compared to 74 for the TAU group (IQR: 28 days to > 3 months) and a slightly increased hazard ratio of 1.50 (95% CI: 0.71, 3.14) (see Figure 2b; Table 2). Of note, in both the ITT and per protocol analyses, postoperative complications exhibited a moderating effect on treatment outcome such that ACT+TAU exhibited a greater treatment effect in those with no complications, with a hazard ratio of 1.58 (95% CI: 0.74, 3.37) for ITT, and 1.64 (95% CI: 0.76, 3.52) for per protocol (see Figure 2c and Table 2). Again, days to pain cessation in the per protocol group was lower for the ACT+TAU (58 days) compared to the TAU group (87 days; IQR: 32 - >3 Months)
Kaplan-Meier curve for the outcome of pain cessation showing the product-limit estimate of the cumulative probability with pain>3 at post-surgery follow-up for a) intent to treat, b) per protocol, and c) per protocol without complication. This shows the effect of adherence to protocol and complication status in assessing the efficacy of ACT on pain cessation.
Table 2
Hazard Ratio (with 95% CI) of opioid cessation, and pain cessation, and Mean difference (with 95% CI) between ACT vs. TAU for outcome measures
| Variable | Statistic | Intent to Treat | Per Protocol | |||
|---|---|---|---|---|---|---|
| ALL | No Complication | ALL | No Complication | |||
| Time to pain cessation | Hazard ratio | Unadjusted | 0.99 (0.54, 1.81) | 1.19 (0.63, 2.28) | 1.09 (0.59, 2.04) | 1.36 (0.70, 2.65) |
| Adjusted* | 1.42 (0.68, 2.95) | 1.58 (0.74, 3.37) | 1.50 (0.71, 3.14) | 1.64 (0.76, 3.52) | ||
| Time to opioid cessation | Hazard ratio | Unadjusted | 1.19 (0.70, 2.04) | 1.41 (0.80, 2.52) | 1.30 (0.74, 2.29) | 1.62 (0.89, 2.95) |
| Adjusted* | 1.44 (0.74, 2.78) | 1.62 (0.82, 3.21) | 1.56 (0.81, 3.01) | 1.69 (0.85, 3.34) | ||
| CPAQ Total | Mean diff (change) | Unadjusted | 1.46 (−7.07, 10.00) | 2.70 (−6.44, 11.85) | 2.83 (−6.02, 11.67) | 4.23 (−5.30, 13.76) |
| Adjusted** | 1.42 (−7.11, 9.57) | 2.68 (−6.47, 11.82) | 2.82 (−6.02, 11.66) | 4.23 (−5.30, 13.75) | ||
| CPVI Mean success | Mean diff (change) | Unadjusted | 0.13 (−0.33, 0.59) | 0.28 (−0.21, 0.78) | 0.11 (−0.36, 0.57) | 0.27 (−0.24, 0.77) |
| Adjusted** | 0.13 (−0.33, 0.59) | 0.28 (−0.22, 0.78) | 0.11 (−0.36, 0.57) | 0.27 (−0.24, 0.77) | ||
| CPVI Discrepancy score | Mean diff (change) | Unadjusted | −0.42 (−0.85, 0.01) | −0.50 (−0.98, −0.03) | −0.32 (−0.76, 0.12) | −0.39 (−0.88, 0.10) |
| Adjusted** | −0.42 (−0.85, 0.01) | −0.51 (−0.98, −0.04) | −0.31 (−0.75, 0.13) | −0.38 (−0.87, 0.10) | ||
For ITT analyses of opioid cessation, median days to opioid cessation from Kaplan-Meier curve was 42.5 (IQR: 30 to 96) for the ACT+TAU group and 51 (IQR: 34 to 94) for the TAU group (see Figure 3a). Among Veterans in the ACT+TAU group, 29% were still taking opioids at 7 weeks follow-up vs. 52% in the TAU group (Odds ratio 0.39; 95% CI 0.14-1.08). The Hazard ratio controlling for baseline differences in gender, combat zone, and CPVI discrepancy score and mean success was 1.44 (95% CI: 0.74, 2.78). Slightly increased hazard ratios were observed in the per protocol analysis (see Figure 3b; Table 2). Again, postoperative complications was a treatment effect moderator, and the opioid cessation hazard ratio of ACT+TAU relative to TAU increased to 1.62 (95% CI: 0.82, 3.21).
Kaplan-Meier curve for the outcome of opioid cessation showing the product-limit estimate of the cumulative probability of opioid use at post-surgery follow-up for a) intent to treat, b) per protocol, and c) per protocol without complication. This shows the effect of adherence to protocol and complication status in assessing the efficacy of ACT on opioid cessation.
As noted above, the ACT group comprised of Veteran patients who attended the workshop before and patients who attended the workshop after surgery. Thus, the possible impact of timing of the intervention on the effect of ACT relative to TAU was assessed. This was done by performing a per protocol analysis using Cox proportional hazard regression with 3 groups (ACT pre-surgery, ACT post-surgery, and TAU), adjusted for covariates. The hazard ratio estimate of pain cessation from this analysis for ACT relative to TAU was 1.42 (95% CI: 0.66, 3.08) for ACT pre-surgery, and 1.91 (95% CI: 0.62, 5.94) for ACT post-surgery. The opioid cessation hazard ratio was 1.45 (95% CI: 0.72, 2.94) for ACT pre-surgery, and 1.96 (95% CI: 0.74, 5.24) for ACT post-surgery. For both endpoints, although hazard ratio estimate and CI upper limit of CI was larger for post-surgery, the CI lower limits were similar, resulting in overlapping Cis.
Pain Acceptance and Values-Based Behaviors
Estimates of group differences in mean change in pain acceptance (CPAQ) and engagement in values-based behaviors (CPVI) are shown in Table 2. Importantly in relation to pain cessation, Veterans reporting increased pain acceptance and CPVI mean success met pain cessation criteria at a significantly higher rate than Veterans reporting less change in the per protocol analyses (for change in CPAQ-Active Engagement [per +5] hazard ratio, HR=1.20, 95% CI: 1.04, 1.38; change in CPAQ Total [per +5] HR=1.10, 95% CI: 1.00, 1.20; change in CPVI Mean Success [per +0.5] HR=1.28, 95% CI: 0.82, 1.21) (see Figure 4). Additionally, opioid cessation rate was significantly higher in Veterans reporting increased engagement in values-based behaviors than Veterans reporting less change (for change in CPVI-Mean Success [per +0.5] HR=1.30, 95% CI: 1.11, 1.53; change in CPVI-Discrepancy [per −0.5] = HR 1.19, 95% CI: 1.01, 1.41) (see Figure 5).
Predicted probability distribution for post-operative pain>3 from Cox proportional hazard regression for specific values of a) change in CPAQ active engagement, b) change in CPAQ Total, and c) change in CPVI mean success. The figure illustrates the effect of 5 points difference in CPAQ active engagement change or CPAQ Total change, or a 0.5 difference in CPVI mean success change on the probability distribution for post-operative pain>3.
Predicted probability distribution for post-operative opioid use from Cox proportional hazard regression for specific values of a) change in CPVI mean success and b) change in CPVI discrepancy score. The figure illustrates the effect of a 0.5 difference in CPVI mean success change, or CPVI discrepancy score change on the probability distribution for postoperative opioid use.
Discussion
This pilot study supports the preliminary efficacy of an intensive 1-day treatment in this difficult-to-treat patient population. All Veterans who attended the ACT+TAU intervention completed the ACT workshop and 90 percent completed the individualized booster session. Patients in the 1-day ACT intervention reached pain cessation (66 versus 74 days) and opioid cessation (42.5 versus 51 days) sooner than patients receiving TAU. Seven weeks post-surgery, 29% and 52% of the patients in ACT+TAU and TAU, respectively, were taking opioids. Importantly, these results were strengthened when the analyses were restricted to patients who attended the workshop and particularly to those who did not experience complications following surgery. Thus this pilot study provides meaningful information about the moderating effect of surgical complications on treatment outcome at 3 months and suggests that this patient subgroup may have differing needs from patients who do not have complications following surgery. This pilot trial also showed that increases in pain acceptance and values-based behavior, the processes that are targeted in ACT, were related to quicker pain cessation. Increases in values-based behavior was also associated with quicker cessation of opioids. These findings are consistent with ACT treatment studies for patients with chronic pain whereby increases in acceptance of pain and values-based behavior are correlated with improvements in various outcomes, such as depression, anxiety, and disability26. The results of our study are thus also noteworthy because the ACT+TAU group exhibited lower levels of values-based behaviors at baseline compared to TAU.
There are several innovative features of this pilot work, including the targeted patient population, the selection of the risk factors, the timing of the intervention, and the treatment model and delivery strategy. In this study, we identified 1) a subset of surgical procedures, orthopedic surgery, that are associated with high rates of chronic post-surgical pain22; 2) particular risk factors (high pre-surgical pain, depression, and anxiety) that put patients at risk for chronic post-surgical pain and prolonged opioid use31; and 3) a Veteran population because they suffer from high rates of the psychological risk factors (depressive and anxiety disorders)34. We then offered a more comprehensive preventive intervention including a psychological treatment to target the risk factors. The timing of the psychological intervention, i.e. prior to the surgery, occurred at an important junction in the patient’s medical care. The treatment delivery strategy utilizing a 1-day workshop took into account a growing body of literature that highlights problems with treatment adherence and completion of the traditional weekly, in-person, psychotherapy. The brief format is also more accessible to surgical patients who have conflicting demands on their time9. Finally, we utilized the ACT treatment model because it is transdiagnostic14 and applicable to the range of difficulties experienced by the patient population recruited, including depression, anxiety, and/or chronic pain7. Over seventy percent of the Veteran patients recruited for this study reported having another chronic pain condition. Providing a transdiagnostic intervention that targets common coping skills across separate pain conditions may thus be particularly useful for patients undergoing surgery.
The long-term aim of this work is to translate our growing understanding of the psychological risk factors associated with CPSP into effective psychological interventions that may curb the growing problem of CPSP. Behavioral interventions have a long history of success for the treatment of chronic pain26. Less work, however, has focused on the prevention of chronic pain. Chronic pain is not only associated with significantly decreased quality of life and productivity, but also with long term opioid use25. Prolonged opioid use can result in significant tolerance and can paradoxically increase pain sensitivity (i.e., hyperalgesia)20, and, unfortunately, the opioid crisis in the United States is related to the use of opioids among surgical patients. Surgery serves as a “critical event” where patients are exposed to opioids for the management of acute pain. However, an increasing number of patients undergoing surgery are remaining on prescription opioids for longer periods than medically indicated12. Increased recognition of the risk factors that put patients at risk for chronic pain and prolonged opioid use following surgery allows opportunity for preventive and comprehensive care41. This is the first study utilizing an ACT approach to prevent the development of chronic post-surgical pain and prolonged opioid use in at-risk Veterans undergoing orthopedic surgery. Broadly, this approach focuses on preventing the transition from acute to chronic pain rather than treating pain only after it has become chronic41.
The findings of this pilot trial have important implications, pending replication in a larger-scale randomized controlled trial. First, offering a preventive psychological intervention for surgical patients at risk for the development of chronic pain has the potential to move the field from a focus on treating chronic pain to preventing it17. In recent years, several pre-surgical risk factors have been consistently associated with the development of chronic pain post-surgery. Greater attention to, and treatment of, these factors may reduce the incidence of chronic pain41. Second, it is critical to identify a treatment delivery strategy that can be completed within a limited time frame by Veterans who are undergoing surgery. These Veterans have limited functional capacity and flexibility to engage in weekly treatments. I. The availability of an effective intervention that more comprehensively addresses the needs of distressed Veterans and that can be completed in 1-day should be considered as an alternative to the regularly prescribed weekly treatments because it ensures treatment adherence and completion and is more feasible for patients who are undergoing surgery9, especially for those patients who live several hours away from the VAMC. Third, enhancing pain acceptance and values-based behaviors, key processes of ACT, may be particularly important for this patient population.
At the same time, more work is needed to better understand how psychological interventions can be maximally effective in addressing the risk factors associated with the development of chronic pain and prolonged opioid use. For example, when can a psychological intervention help most? Does it need to be completed prior to surgery or can it also effectively be completed soon after surgery? What “dose” is maximally effective? How may the results obtained in this study be strengthened while still being feasible for surgical patients? When, and how often, should booster sessions be provided? A novel clinical initiative, the “Transitional Pain Service” has been developed at the Toronto General Hospital to help patients as they transition through the various stages of surgical recovery, including pre- and post-operatively18. Patients who may be at risk for developing chronic pain are identified and offered comprehensive care by a multidisciplinary team of pain physicians, psychologists, physiotherapists, and nurses within the surgical setting. Thus, the interventions are embedded in the surgical service, offered across the hospital-to-home trajectory, and comprehensive. Such programs may lead to substantial benefits including decreased incidence of chronic pain, opioid misuse and hospital savings.
It is also worth addressing other possible risk factors or coping strategies. For example, in this pilot study patients who experienced complications following the surgery reported descriptively worse outcomes, both in pain severity and increased opioid use. Thus, this subset of patients may represent a particular group of patients who could benefit from increased attention and therapeutic guidance postoperatively. Furthermore, research has shown that a history of substance abuse or chronic opioid use puts patients at increased risk for a range of negative outcomes17, 41, 44. Do interventions need to be tailored specifically to meet the needs of subgroups of patients? In this pilot study, substance use was assessed with a screening measure and the rates were generally low. Future studies should more systematically and comprehensively assess the role of substance abuse history on outcomes of interest.
These and many other questions still remain in a field that is increasingly appreciating the importance of addressing acute pain prior to it becoming chronic. It will be important for researchers and clinicians to continue to find ways to prevent and manage these difficulties among surgical patients. In this pilot study, we obtained promising preliminary data to suggest that targeting patients with specific risk factors for chronic pain and prolonged opioid use prior to their surgery leads to a shortened time on opioids and with pain. At the same time, these results await replication with a more rigorous design. For example, the 3-month follow-up period in this study is relatively short. Future studies should examine whether the positive treatment effects obtained here persist over longer periods of time or if they begin to wear off at certain times. This could provide important information about timing for additional booster sessions. In this study, timing of opioid cessation was obtained by Veteran self-report followed by a review of the medical record to confirm consistency between the two. Future studies may assess opioid use through pill counts to enhance accuracy. Future studies should also enroll a larger number of participants to ensure the ability to detect treatment effects and to examine a larger number of mediators and moderators of treatment response. Finally, it will be important to study additional important outcomes such as the impact of the treatment on functioning and quality of life.
Acknowledgments
This work was supported by the National Institutes of Health NCCIH 5R34AT008349-02 to PI Barbara Rakel and the National Institute of Neurological Disorders and Stroke (T32 NS045549). This work was also partially supported by the use and resources of the Houston VA HSR&D Center for Innovations in Quality, Effectiveness and Safety (CIN13-413). Data was collected with REDCap electronic data capture tools hosted at University of Iowa (supported by NIH 54TR001013).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The authors do not have any conflicts of interest.
1The decline rate dropped significantly over the course of the study with improvements in recruitment strategies.
Contributor Information
Lilian Dindo, Baylor College of Medicine, 2002 Holcombe Blvd (152), Houston, Texas, USA, Center for Innovations, Quality, and Effectiveness, Houston, Texas, USA.
M. Bridget Zimmerman, University of Iowa, College of Public Health, N373 CPHB, Iowa City, IA, 52242, USA.
Katherine Hadlandsmyth, University of Iowa, Department of Anesthesia, 200 Hawkins Drive, Iowa City, 52242, USA.
Barbara StMarie, University of Iowa, College of Nursing, 440 CNB, Iowa City, Iowa, 52242, USA.
Jennie Embree, University of Iowa, College of Nursing, 318 CNB, Iowa City, Iowa, 52242, USA.
James Marchman, University of Iowa, Department of Psychology, 720 S Dubuque St, Iowa City, IA 52240, USA.
Toni Tripp-Reimer, University of Iowa, College of Nursing, 399 CNB, Iowa City, Iowa, 52242, USA.
Barbara Rakel, University of Iowa, College of Nursing, 318 CNB, Iowa City, Iowa, 52242, USA.
References
Citations & impact
Impact metrics
Citations of article over time
Alternative metrics
Article citations
Demystifying the virtualization process: A call for standard reporting of virtual modifications to evidence-based psychotherapies, using the FRAME model.
Transl Behav Med, 14(8):452-460, 01 Jul 2024
Cited by: 0 articles | PMID: 38777341
An Online Single-Session Cognitive Behavioral Therapy for Depression and Anxiety Associated with Multiple Sclerosis-Pilot Study.
Behav Sci (Basel), 14(7):620, 21 Jul 2024
Cited by: 0 articles | PMID: 39062443 | PMCID: PMC11274225
Acceptance and commitment therapy for patients with chronic pain: A systematic review and meta-analysis on psychological outcomes and quality of life.
PLoS One, 19(6):e0301226, 14 Jun 2024
Cited by: 0 articles | PMID: 38875241 | PMCID: PMC11178235
Review
Free full text in Europe PMC
Evaluation of Pain Catastrophizing Scale for surgical referral to pain psychology in patients undergoing spinal surgery.
N Am Spine Soc J, 19:100330, 30 May 2024
Cited by: 0 articles | PMID: 39021894 | PMCID: PMC11254176
A Single-Session Process-Based Cognitive-Behavioral Intervention Combined with Multimodal Rehabilitation Treatment for Chronic Pain Associated with Emotional Disorders.
Behav Sci (Basel), 14(4):327, 15 Apr 2024
Cited by: 2 articles | PMID: 38667123 | PMCID: PMC11047417
Go to all (46) article citations
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Funding
Funders who supported this work.
NCATS NIH HHS (1)
Grant ID: U54 TR001013
NCCIH NIH HHS (1)
Grant ID: R34 AT008349
NINDS NIH HHS (1)
Grant ID: T32 NS045549