Journal Logo

Clinical Note

Reducing the use of opioids by patients with chronic pain: an effectiveness study with long-term follow-up

Nicholas, Michael K.a,b,*; Asghari, Alia,b,c; Sharpe, Louised; Beeston, Leea,b; Brooker, Charlesa,b; Glare, Paula,b; Martin, Rebeccaa,b; Molloy, Allana,b; Wrigley, Paul J.a,b

Author Information
doi: 10.1097/j.pain.0000000000001763
  • Free
  • Global Year 2020

1. Introduction

Concern about the use of opioids for chronic pain is widespread,14,28,45 and has been fueled by the growth in prescribing of these agents,28,39 alongside the lack of evidence for their effectiveness and the strong evidence for the risk of harms.3 There is also compelling evidence that the use of opioids for pain relief is associated with poorer return to work outcomes,5,22 with its major financial and health implications.43 Reducing access to these agents does not help those who are already taking them. Alternative options are required. One of the recommended alternatives is training in pain self-management.36 However, the nature and quality of self-management interventions are known to vary considerably52 and the evidence for their effectiveness among patients who were prescribed opioids is limited.56

In patients with chronic pain, opioid withdrawal and replacement by nonpharmacological, self-management approaches have been reported since the 1970s.9 Initially, this was achieved through the use of a liquid “cocktail” of an opioid and a neutral substance (whose proportions were gradually changed) while the patients participated in behavioural pain management programs where they learned to pace up their activities despite their pain and to use self-management strategies (such as goal setting and relaxation).9,40 Rather than the “pain cocktail,” subsequent studies have typically used reductions in number of tablets taken, alongside pain self-management strategies.55

The nature of the self-management training provided may also be important. In the United Kingdom, Williams et al.55 showed that an intensive, inpatient program (140 hours over 4 weeks) was more effective in reducing opioids sustainably (up to 1-year post-treatment) than the similar, but shorter, outpatient program (24 hours over 8 weeks).53,55 In the United States, Townsend et al.48 provided evidence that an intensive (120 hours over 3 weeks) outpatient interdisciplinary pain management program could reduce the use of opioids in chronic pain patients population for up to 6 months. More recently, also in the United States, Huffman et al.15 reported similar results from a retrospective analysis of data from a large sample (n = 1457) of chronic noncancer patients (with a 1-year follow-up) after another intensive outpatient interdisciplinary pain management program. This study also reported sustained improvements across a range of pain domains as well as in opioid cessation, but there was a relapse rate of 10.5% at 6 months and 30.7% at 1 year. Despite the evidence in favour of the more intensive (in terms of concentrated hours), in most countries, the more intensive programs have generally been replaced by the briefer versions, often due to cost pressures.11 Unfortunately, the effectiveness of the briefer programs in achieving opioid reduction is not well established.

In addition, the sustainability of reduced opioid use after more intensive outpatient programs is also unclear. This study sought to address this issue of effectiveness at short- and long-term follow-up using a secondary analysis of unanalyzed data from a previously published trial of a comprehensive interdisciplinary outpatient pain management program.35

2. Method

Unanalyzed data from a randomized controlled trial by Nicholas et al.35 were used to evaluate the reduction in opioid use and the degree to which these reductions were sustained over a year after the treatment program. Three main questions were addressed:

  • (1) At admission to the treatment program, is there a difference between those taking/not taking opioids in this sample on self-report measures of usual pain, depression, disability, and pain cognitions?
  • (2) Is the use of opioids reduced significantly after the treatment program, and is this reduction sustained over a 12-month follow-up?
  • (3) After the treatment program, is the cessation of opioids related to improvements in usual pain, depression, disability, and pain cognitions?

Nicholas et al.35 compared 2 approaches to reduce the threat value of pain (cognitive exposure vs cognitive distraction) within a comprehensive, 3-week outpatient interdisciplinary pain management program that was informed by cognitive-behavioural principles. The design entailed random allocation of groups to one or other version of the treatment program. Because there were no differences in outcomes either at posttreatment or 1-year follow-up between the 2 groups, the data for this study were combined to yield a total sample of 140 heterogeneous chronic pain patients (see Table 1 for characteristics of total study sample). The design, content, and conduct of the study were described fully in our earlier study35; so, we will describe only the main features of the study here.

Table 1
Table 1:
Baseline characteristics of patients who started the treatment program (means and SDs), and, where relevant, frequency (F) and percentage (%) of the total.

All potential patients were recommended for the pain program after a comprehensive multidisciplinary assessment (performed by a physician specialist in pain medicine, a clinical psychologist or psychiatrist, and a physiotherapist) at the Pain Management and Research Centre, Royal North Shore Hospital, Sydney. The patients had all been referred to the centre by their treating doctors.

2.1. Inclusion criteria

(1) Pain lasting ≥6 months and causing significant interference in normal activities despite appropriate treatments; (2) seeking help for pain; (3) no further medical or surgical treatments deemed appropriate after multidisciplinary review; (4) provided informed consent to participate in this research program; and at least one of the following: (1) reliance on medication that was minimally helpful in reducing pain severity and its impact; (2) evidence of pain-related distress or sleep disturbance; and (3) using aids to manage (sticks, hot/cold packs).

2.2. Exclusion criteria

(1) Unwilling or unable to attend the treatment program (eg, due to travel or family constraints, still seeking curative treatment, or not wanting to be part of a study); (2) unable to speak adequate English; (3) presence of active major mental disorder (eg, major depression with active suicidal ideation); or (4) a primary drug addiction problem for which formal involvement of a drug dependence team was recommended.

2.3. Participants

One hundred forty patients (aged between 18 and 65 years) were admitted to the 3-week (115 hours) outpatient pain management programme after being assessed as suitable by a multidisciplinary team and agreeing to participate.

The multidisciplinary assessment involved a day at the clinic being interviewed by 3 staff (physician specialist in pain medicine, clinical psychologist or psychiatrist, and physical therapist), and then, after a team discussion, a second meeting with the physician and discussion of our recommendations (eg, attend the group program to achieve their functional goals, including ceasing opioids). If willing to attend the group program to these ends, the patient spent 30 minutes with a nurse coordinator being briefed on the program. See Box 1 for summary of steps involved in opioid-tapering plan before admission to the program.

Box 1.

Opioid-tapering plan after multidisciplinary assessment and before attending treatment program.

  • (1) Meeting with nurse coordinator to explain program details and process to be followed before attending
  • (2) Opioid tapering explained as gradual process with goal of 40-mg oral morphine equivalent daily dose (oMMED) or less by admission to program
  • (3) Request to read pain self-management book38
  • (4) Letter with the pain centre's plan and recommendations sent to referring physician (seeking their support for plan)
  • (5) Support from pain medicine physician, nurse, and clinical psychologist on individual basis, in person or on phone, until the patient is ready to attend group program
  • (6) Preadmission attendance at the pain centre (2-3 weeks before program) with clinical psychologist or physiotherapist to check on their progress and their preparations for the program, including clarification of their treatment goals.

2.4. Treatment program

The interdisciplinary program was based directly on the British inpatient program described by Williams et al.55 but conducted in Australia on an outpatient basis. As in the U.K. program, the Australian patients were not required to pay for their attendance because this was covered either by public hospital funding or by workers compensation insurance for those claiming this cover. Patients attended in groups of 8 to 10 on weekdays for 3 consecutive weeks, from 9:00 am to 5:00 pm. The treatment team comprised a clinical psychologist, a physiotherapist, a nurse, a rehabilitation advisor, and a physician specialist in pain medicine, with between 5 and 14 years of experience on the program. The program uses an interdisciplinary style with cognitive behavioural principles informing all interactions between staff and patients.39 From a theoretical perspective, the program uses behavioural exposure whereby graduated increments in activities previously avoided due to pain or fear of pain are specifically encouraged. This is also described as activity pacing.37 Simultaneously, safety or pain avoidance behaviours (eg, medication consumption, extended rest, and use of aids) are targeted for reduction. Reductions in pain medication, including opioids, were achieved by a graduated and planned reduction in tablets after review by the physician specialist in pain medicine and monitored by the program nurse. Additional details about the program can be found in the study by Nicholas et al.34Box 2 summarises the steps undertaken during the program to complete opioid cessation.

Box 2.

Opioid-cessation steps during treatment program.

  • (1) Day 1: rationale for whole program discussed with group, including replacing use of opioids by applying self-management strategies instead.
  • Goal setting and importance of individually relevant goals discussed.
  • (2) Days 1 and 2: patients take opioids (and other pain medication) as usual for them at that time, but record each intake and dose.
  • Patients also perform and record baseline activity and exercise levels with physiotherapist. Introduction to activity pacing (continued throughout program)
  • (3) Day 2: group discussion with nurse about medication for pain and importance of ceasing use of these agents.
  • Training in use of applied relaxation started and continued daily through program
  • (4) Day 3 and throughout program: interactive discussion sessions about pain, pain mechanisms, and pain treatments (by pain medicine physician and physiotherapist).
  • Meeting with nurse, clinical psychologist, and physiotherapist to review and confirm current intake of medication, plus to confirm/revise personal goals.
  • Commence upgrading of exercises and activities (towards specified goals), coupled with training in self-management skills (for pain, sleep, stressors, anxiety, depression, etc). Skills practised and reviewed daily throughout program.
  • (5) Day 3 or 4: meet with pain medicine physician and nurse (individually) to review medication use, confirm goal of cessation for opioids (and other agents as appropriate), and then set gradual withdrawal plan for next 2 weeks to zero opioids during week 3. Usually, reducing dose every second day (monitored by nurse)
  • (6) End of week 2: families of patients (if available) attend the centre to learn about program, goals, methods, and guidance on maintenance of gains after discharge
  • (7) End of week 3: individual session with same pain medicine physician and nurse to review progress on medication withdrawal and determine whether further help needed
  • (8) End of the program: report sent to referring physician to confirm opioid use cessation and summarise other achievements by patient with details of plan (and recommended supports) for maintenance of gains achieved
  • (9) 1 month after discharge: patients return to the centre to review progress. Explore options for additional support (on individual basis) if indicated.

Those patients who had made some progress towards, but had not achieved the 40 mg (opioid drugs to a morphine equivalent dose in milligrams [oMEDD]) level before admission could still be admitted, provided they were deemed by their treating doctors and program staff (on review prior to attending) to be as ready as possible to attend. It was recognised by the centre staff that this group would probably take longer than the 3 weeks of the structured program to achieve zero opioid use and that follow-up contingency plans were likely to be needed. No one was expected or asked to rapidly cease their opioid use (in fact, it was discouraged).

Before admission, the reviewing staff member emphasized to all patients still taking opioids that further reductions in opioid use would be gradual through the program with the aim of achieving zero opioids during the third week of the program. It was also made clear to each patient coming off their opioids that simultaneously they would be learning nonpharmacological (self-management) ways of managing their pain and that these would replace the role of the opioids. These methods were outlined in the self-help book38 (written by senior staff members and based on the patient notes from the U.K. program55) that the patients were expected to have read by the time of admission.

The self-management methods taught and practised by the patients as alternatives to using opioids included graduated upgrading (despite pain) of activities related to the patients' individual goals using pacing principles (to minimise pain flare-ups while also encouraging habituation to their pain), developing a pain flare-up plan in case they had a flare-up, self-calming (relaxation) methods, cognitive coping strategies (eg, problem-solving strategies, identifying and addressing unhelpful cognitions), and education (delivered by interactive discussion) about the nature of chronic pain and the limitations of relying on medication (including opioids) alone for managing chronic pain. The education about pain was also aimed at addressing any fears of harm or threat associated with pain. During the program, all patients had regular individual time with one of the staff to discuss and problem-solve any difficulties they were having in making changes, including withdrawing from their opioids. Our group has previously published evidence that the practice of these strategies is positively correlated with better outcomes after our program, including less pain, and this finding was also discussed.34,35 Contrary to the usual expectation a patient might hold, that if they could get their pain under control with an opioid they would then upgrade their activities, in the pain program the idea was to learn to apply the self-management strategies then their pain would better controlled, and less severe.

2.5. Study design

A retrospective group design was undertaken with data on opioid use available from 6 separate occasions: the initial multidisciplinary assessment, immediately preprogram, postprogram, 1-month follow-up (in person at the pain centre), and 6- and 12-month follow-ups (by mail). The mean interval between initial assessment and admission to the pain program was 6.70 months (median = 5 months, mode: 3 months). In total, 16 programs were conducted consecutively between April 2008 and June 2009 (Fig. 1).

Figure 1.
Figure 1.:
Study flow chart.

Self-report measures for pain, depression, disability, and pain cognitions were available from pretreatment to the 12-month follow-up. Opioid data were available for the initial assessment and all assessment occasions thereafter. The self-report measures were posted to participants at the 6- and 12-month occasions. If there was no response within 10 days, the research assistant contacted them by telephone and mail for a maximum of 2 times.

The study was approved by the Northern Sydney Area Health Human Research Ethics Committee. All participants provided informed consent for their deidentified data to be used for research purposes. In total, 144 patients were originally assessed as suitable for the program but 4 refused to participate in the study, leaving the total sample at 140 (97% recruitment rate).

2.6. Measures

Data were gathered on several standardized self-report measures relating to pain, interference by pain in daily activities, mood state, and beliefs about pain. In addition, patient reports of medication use were also obtained at each collection point as part of their normal assessment process.

2.6.1. Modified Roland and Morris Disability Questionnaire

As the participants were heterogeneous for pain sites, the modified form of the RMDQ was used, which replaces the word back with pain.41 The psychometric properties of this modified version of the RMDQ have been published.2 Nicholas et al.33 reported the internal consistency (α) = 0.88. The modified RMDQ scores range between 0 and 24, with higher scores indicating more severe physical disability.

2.6.2. Pain intensity and interference scales of the Multidimensional Pain Inventory

The 3-item pain intensity scale covers pain experience in the previous week. The interference scale of the MPI with 9 items assesses interference of pain in daily activities (ie, vocational, social/recreational, and family/marital functioning) in the previous week.21 The MPI has been shown to have good reliability and validity.17 The scores range between 0 and 6, with higher scores indicating more severe pain and interference. Nicholas et al.33 reported the internal consistency of the pain intensity and interference subscales of the MPI as 0.78 and 0.92, respectively.

2.6.3. The depression scale of the Depression Anxiety Stress Scales

The depression scale of the Depression Anxiety Stress Scales24 was used to measure depression severity. The Depression Anxiety Stress Scale has been validated for use with chronic pain patients.47 Nicholas et al.33 reported the internal consistency (α) to be 0.96. The scores range between 0 and 42, with higher scores indicating more severe depressive symptoms.

In addition to the above measures, 4 cognitive process variables reflecting the threat value of pain were used.

2.6.4. The catastrophizing scale of the Pain Response Self-Statements Scale

This 9-item scale asks patients to rate on a 0 to 5 scale (0 = almost never to 5 = almost always) the frequency of catastrophic thoughts when they experience severe pain.8 Higher mean scores indicate more frequent catastrophizing. Flor et al.8 established the psychometric properties of the Pain Response Self-Statements Scale. Nicholas et al.33 reported the internal consistency (α) to be 0.86.

2.6.5. The Tampa Scale for Kinesiophobia

The Tampa Scale for Kinesiophobia23 was used to assess fear and avoidance beliefs about movement and reinjury. The 17 statements, each scored on a 4-point scale from “strongly disagree” to “strongly agree,” yield a total range from 17 to 68. Higher scores indicate more severe fear-avoidance beliefs. The Tampa Scale for Kinesiophobia has been well validated in pain populations.50 Nicholas et al.33 reported the internal consistency (α) to be 0.83. Although some have questioned the reliability and validity of the measure, it is in widespread use, and to increase comparability with other similar studies, we have included it here.

2.6.6. The Pain Self-Efficacy Questionnaire

The Pain Self-Efficacy Questionnaire31 has 10 items and measures the strength and generality of a patient's beliefs about his/her ability to accomplish various activities despite his/her pain. Scores range from 0 to 60. Higher scores indicate stronger self-efficacy beliefs. The psychometric properties of this measure are well established. Nicholas et al.33 reported the internal consistency (α = 0.93).

2.6.7. Chronic Pain Acceptance Questionnaire

The Chronic Pain Acceptance Questionnaire26 is a 20-item scale that reflects the degree to which a person engages in life activities regardless of pain and the degree to which the person feels little need to avoid or control painful experiences. In this study, a total score (range 0-120) was used, with higher scores reflecting higher pain acceptance. Nicholas and Asghari32 reported the internal consistency (α = 0.85).

2.6.8. Medication

Three methods for assessing opioid use were used.

  • (1) Taking/not taking. Because dose reports may be unreliable, the method used in the original study35 recorded total medication classes as the measure of medication used.49 For this study on opioid use, we recorded whether or not the patients were taking an opioid drug (0/1).
  • (2) Opioid dose. The opioid dose data are presented in 2 ways:
    • a. By converting reported oMEDD using a recognised conversion table developed by the Faculty of Pain Medicine, Australian and New Zealand College of Anaesthetists9 (available at For drugs reported taken as needed, unless specifically quantified, 3 times per day was used for the default frequency. Although the calculation of an oral morphine equivalent for methadone is particularly complex,29 a conservative 3:1 conversion was chosen (personal communication). Conversions were undertaken by 2 physician specialists in pain medicine (R.M. and P.J.W.) independently, who then met to resolve discrepancies.
    • b. We also report daily opioid use according to 4 categories (or bands) that are commonly used in describing opioid doses.7 The 4 categories are 0; 0.1 to <40 mg; ≥40 to <100 mg; and ≥100 mg.

2.7. Statistical analyses

All data were collected, scored, and entered into a secure database by a research assistant who was not involved in the treatments. Another researcher (also not involved in the treatments and blind to their nature) conducted all pre, post, and follow-up assessments. All analyses for this study were conducted by one researcher (A.A.) who was not involved in the treatment.

Descriptive statistics were used to summarize the baseline demographic and medical characteristics (including duration of pain and oMEDD) and psychometric measures at admission to the program. In addition, the mean scores (at admission to the program) of those who withdrew during the program and those not contactable at the 1-month and 12-month follow-up occasions were compared with those who were retained in the study throughout using independent-sample t test or χ2, as applicable. A similar analysis was conducted comparing the mean scores at posttreatment for patients who completed the program (but were lost to follow-up) and those who were followed up at 12 months. All statistical analyses for this study were performed using SPSS version 22 (SPSS, Inc).

In addition to determining the degree of statistical significance for any differences, the standardized mean difference effect size for within-subjects designs was calculated. The standardized mean difference effect size for within-subjects designs is referred to as Cohen's dz, where the Z alludes to the fact that the unit of analysis is no longer X or Y, but their difference, Z, and can be calculated directly from the t-value and the number of participants, using the following formula provided by Rosenthal42:where t represents the magnitude of observed t value in paired-sample t-tests and is the square root of sample size. Effect sizes have been classified as small above 0.20, medium above 0.50, and large above 0.80.4

The improvements in self-report measures of usual pain, depression, disability, and pain cognitions were assessed in 3 ways: statistical analysis (paired-sample t-tests and χ2 tests, as applicable); Cohen's dz (rather than Cohen's d was used to reduce the risk of inflated effect sizes); and proportion of patients who achieved a reliable change (which may be of clinical significance). Although the statistical analyses indicate the degree to which the study findings are not due to chance, reliable change estimates attempt to determine whether changes in outcome measures are substantial enough to be beyond the level of measurement error and provide an indication of the proportion of cases who achieve a change likely to be of clinical significance20,25,30 In this study, reliable change estimate cutoff points were calculated, using Jacobson's clinical significance analysis18,19—see the studies by Nicholas et al.34,35 for more details. Some researchers have also used an effect size ≥0.5 as a measure of clinically significant change,15 and for the sake of comparison of outcomes, this was examined in this study as well.

3. Results

The 140 participants and their attrition rate through the study are reported in Figure 1. The baseline data (at admission to the program) for the main outcome measures are presented in Table 1. These indicate that the cohort of patients was quite similar to patients normally seen by tertiary-level pain clinics.33 The average age was 43 years, the average length of time in pain was 6 years, sex balance was close to 50/50, a slight majority (54%) had had some education beyond secondary school (technical or university education), 67% had their main pain site in the lower back or lower back and legs, most (60%) were not working due to pain, and 67% had a compensation claim for their injury. In this context, it should be noted that at initial assessment, 103 of the 140 (73.6%) patients in this study were taking opioids. Those patients on opioids at initial multidisciplinary assessment were encouraged to start reducing those agents in preparation for the program. In some cases, this involved individual sessions with clinical psychology, physiotherapy, and medical staff with a formal medication reduction plan coordinated with the patient's general medical practitioner. The usual goal of the reduction plan was to achieve a maximum morphine equivalent dose of 40 mg per day on admission to the treatment program. Nine patients reported taking methadone with the maximum dose of 30 mg per day (mean [SD]: 21 ± 8).

On the psychological and disability measures, the average scores were generally close to the 50th percentile of the published norms for patients from this clinic collected previously.35 The mean scores (at admission to the treatment program) of those who withdrew during the program and those not contactable at the 1-month and 12-month follow-up occasions were compared with those who were retained in the study throughout. At admission to the treatment program, there were no significant differences between those who were retained in the study (at the end of the program and 1-month follow-up) and those who dropped out during the program or by the 1-month follow-up. However, when the admission scores of those retained at the 12-month follow-up (n = 84) were compared with the combined admission mean scores of those who withdrew during the program (n = 17) or were lost to follow-up (n = 39), it was found that those who withdrew or were lost to follow-up reported higher baseline usual pain severity (on the 0-6 MPI scale) (4.30 [SD = 1.02] vs 3.93 [SD = 0.99]; P < 0.04); higher depression severity (21.59 [SD = 10.97] vs 17.55 [SD = 11.76]; P < 0.04); and lower pain self-efficacy (21.82 [SD = 11.50] vs 26.51 [SD = 11.52]; P < 0.02). The differences on the other psychometric variables measured were not significant. By contrast, when end-of-program mean scores for those who completed the program but were lost to follow-up (n = 39) were compared with those retained throughout the study (n = 84), there were no differences on any measure. The latter findings suggest that those lost to follow-up after completing the program were not different (at least on the measures taken) from those retained.

In addition, when we compared the demographic (age and sex) and medical-related variables (duration of pain and oMEDD) at admission to the program between those who withdrew and those who were retained in the study over the treatment period and at the 12-month follow-up, no significant differences on these variables were found.

4. Question 1: differences between those taking vs not taking opioids at admission

As can be seen in Table 2, there were no significant differences between those taking or not taking opioids at admission to the program. This means that to the extent that the 2 groups were equivalent, the opioids did not confer any advantage in these domains.

Table 2
Table 2:
A comparison between those taking vs not taking opioids at admission to the program.

5. Question 2: opioid reduction and long-term maintenance

Examination of the data presented in Table 3 indicates that there were significant reductions in numbers of those taking opioids between initial assessment and admission to the pain program, from 103 patients to 51 patients (χ2 = 10.78, df = 1, P ≤ 0.01). However, no significant differences were found between the oMEDD (47.1 mg; SD: 56.4) at initial assessment and the oMEDD (43.4 mg; SD: 39.8) at the start of the treatment program (t = 0.337, P = 0.73). This would suggest that the patients who ceased their opioids over this period (a mean of 6.7 months) must have been taking quite low doses, but closer examination of the data revealed that during that period, some 79 of the 103 patients (76.7%) taking opioids at initial assessment ceased their opioids, but at the same time, 27 of the 37 (73%) on no opioids at initial assessment started taking them. This suggests a considerable degree of instability in medication use in this preparation period, despite the encouragement to start reducing the use of opioids among those who were taking them. It should be noted that those who started taking opioids in this period were not prescribed them by the pain centre's medical staff.

Table 3
Table 3:
Number of patients taking opioids and the mean oMEDD (SD).

After the treatment program, however, the number of patients taking the opioids declined even further, from 51 to 13. This change in number of cases was statistically significant (χ2 = 7.69, df = 1, P ≤ 0.01). In addition, there was a statistically significant reduction in the mean dose of oral morphine equivalent (43.4 mg, SD: 39.8 vs 27.1 mg, SD: 24.2; t = 6.90, P = 0.001). None of the patients on no opioids at admission to the treatment started taking them during the program, although 3 of those already on them did increase their dose to a small extent during this period (range: increased by 7-45 mg oMEDD).

At the 3 follow-up occasions, the number of patients using opioids declined, from 13 at posttreatment to 7 at 12 months for those (60%) who could be located. Interestingly, the mean dose for these 7 patients (6.8% of the original 103 at initial assessment) actually rose at the 12-month follow-up only. Close examination of Figure 2 reveals that this was due to 2 patients—and is reflected in the increase in range of dose from the 6-month follow-up (a maximum of 80 mgs) to 150 mgs at 12 month (Table 3).

Figure 2.
Figure 2.:
Proportion (%) of patients on different levels of opioids over the course of the study.

Nevertheless, as can be seen in Figure 2, the proportion of patients reporting the use of the different dose bands of opioids remained consistently low over the 12 months after the end of treatment, and conversely, the proportion taking no opioids remained consistently high over the same period. In sum, these data indicate that relapse in opioid use was minimal.

6. Question 3: relationship between reduced opioid use and self-report scales

As can be seen in Table 3, 37/51 (72.5%) patients who were still taking opioids at admission to the program had stopped them by the end. For this group, the mean oMEDD at entry was reduced from 39.8 mgs (SD = 20.9) to nil and this was statistically significant (t = 11.56, df = 36, P < 0.0001). The data in Table 4 reveal that, associated with the cessation of opioids, there were statistically significant improvements on all self-report variables from pretreatment to posttreatment for these patients. Of these improvements, large effect sizes5 (Cohen's dz ≥0.8) were found for depression severity, pain-related disability, fear-avoidance beliefs, pain self-efficacy, and pain acceptance. The effect sizes for improvements in pain interference in daily activities and catastrophizing were in the medium range (≥0.5-0.79), whereas the effect size for improvement in usual pain was in the small range (≥0.20-0.49). Overall, the mean of the treatment effect sizes for all these variables was 0.82 (range: 0.43-1.09)—a large effect size. Application of the method used by Huffman et al.,17 determination of minimal clinically significant changes (an effect size of 0.5 or greater) would have found that the improvements in all variables except usual pain would have been clinically significant in this group.

Table 4
Table 4:
Comparison of means of self-report measures at pretreatment and posttreatment for those who ceased opioids during treatment (n = 37) and for those who remained on opioids at posttreatment (n = 13).

Another way of understanding these results is to consider the proportion of patients who met the criterion for a reliable change (that is, an improvement or deterioration beyond likely measurement error). In this case, across all self-report outcome variables for those who ceased opioids, a mean of 38% (range: 24%-49%) of patients achieved this criterion for improvement, whereas only a mean of 2% (range: 0%-5%) of patients met the criterion for reliable decline in outcome (or got worse). By contrast, for those who remained on opioids at the end of the treatment, a mean of 35% (range: 23%-46%) improved, whereas a mean of 12% got worse (range: 0%-38%). Interestingly, compared with those who stopped their opioids during the treatment program, those patients who remained on their opioids (N = 13/51 or 25.4%) still reported a reduction in mean oMEDD—from 51.4 mgs (SD = 68.9) to 27.1 mgs (SD: 24.2)—but due to their small numbers (n = 13), no statistical tests were conducted.

7. Discussion

This study examined 3 questions concerning the cessation of opioids after participation in a comprehensive interdisciplinary outpatient pain management program that was aimed at enhancing pain self-management among a cohort of chronic pain patients. The results indicated that at admission to the treatment program, there were no significant differences between those taking or not taking opioids in terms of usual pain, pain interference in daily activities, pain-related disability, depression severity, as well as in pain cognitions. By the end of the program, there was a significant reduction in the use of opioids, both in numbers of patients taking opioids and in mean doses, and these gains were maintained over the 12 months after the end of the treatment program. Importantly, the sustained improvement in pain after cessation of opioids is consistent with the recent findings of a similar effect in a U.S. Veterans sample,27 which suggests that the finding is replicable.

Overall, these results are consistent with those reported in a randomized controlled trial (RCT) of the similar program but conducted in an inpatient setting55 and those reported from 2 nonrandomised studies with outpatient programs similar in length and content to the program used in this study.15,48 Together with the Townsend et al.48 and Huffman et al.15 studies, this study provides clear evidence that clinically important reductions in the use of opioids for chronic pain can be achieved after comprehensive outpatient interdisciplinary pain management programs. An important finding of this study was that this reduction can be maintained at 12-month follow-up with minimal relapse. Another important consideration16 in studies of psychologically based treatments for pain is that the adverse event rate for those who ceased opioids was minimal (a mean of 2%, range: 0%-5%, of this group deteriorated on the self-report measures). By contrast, a mean of 12% (range: 0%-38%) of those who remained on opioids at the end of the program deteriorated. Although the numbers of patients in our study who remained on opioids are too small to draw firm conclusions, the possibility that remaining on even small doses of opioids at the end of a pain rehabilitation intervention could place such patients at greater risk of worse outcomes than those who cease opioids merits future investigation.

One interesting finding was the substantial proportion of patients (79/103 or 76%) who managed to cease their opioid use in the period (mean of 6 months) after their multidisciplinary initial assessment before attending the treatment program. As indicated in the Methods section, at the initial assessment, a number of measures were used to promote reduced use of opioids, and points to the value of these steps at that time rather than waiting until the treatment program to commence the reduction. This is consistent with recent findings from brief educational interventions.6,44 However, over the same period, a similar proportion (27/37 or 73%) of those not taking opioids at initial assessment started taking them, all prescribed by medical practitioners outside the pain centre. This would seem to reflect the common problem of patients having multiple prescribing physicians—something that is discouraged in opioid guidelines generally.6 The instability of opioid use before the treatment program can be contrasted with the more stable pattern of opioid use during and after the program; no patient started opioids during the program, only 3/51 (or 5.9%) increased their doses, and 99% of cases followed up did not relapse in the 12 months after the treatment program.

Before considering the implications of these findings, we should acknowledge the strengths and limitations of the study. The strengths of the study include its prospective design, the relatively large sample size, the use of 2 specialists in pain medicine to independently calculate the morphine equivalent doses and then reach agreement on any differences, along with the evidence that the patients were generally representative of chronic pain patients attending pain management clinics, at least in Australian and in the U.K. samples.55 Compared with the sample treated in the similar U.K. inpatient program,55 our sample was slightly younger (mean age 43 vs 48 years, respectively) and included more with medicolegal involvement (67% vs 24%, respectively). In other respects, they were quite similar, especially on sex balance and proportions not working due to pain. In relation to opioid use, substantially fewer of our sample were taking opioids at admission to the program (36% vs 53%, respectively), but the mean daily morphine equivalent doses were similar (36 mg vs 30 mg, respectively). The oMEDDs in these 2 samples were generally lower than those in the U.S. samples,15,48 but other characteristics of the 4 samples were similar. Additional strengths include the modest but acceptable retention rate (60%) over the 1-year follow-up, which was superior to that reported by Huffman et al.,15 and the finding that those lost to follow-up were no different from those retained, at least in terms of opioid use and psychological profile at admission and discharge from the program. Also, the psychological measures used on each assessment occasion meant that we were able to examine pain-relevant psychological changes rather than just a description of the sample's demographics and opioid use data.

The limitations of the study include the lack of randomization to a usual care or opioid maintenance condition, which would have allowed us to evaluate the contribution of the program itself, but as indicated in the introduction, this was an effectiveness study of a program that was based on the previously reported RCT that compared 2 versions of cognitive-behavioural therapy-informed multidisciplinary treatment for patients with chronic pain relative to a wait-list control by Williams et al.55 Also, because the study was a secondary analysis of our original study,35 it was not designed or powered to investigate the use of opioids alone (as was also the case with the study by Williams et al.55).

It should also be acknowledged that opioid use was based on patient self-report and not verified by blood or urine testing. However, even these tests have limitations in determining quantity of opioids consumed. Finally, although the loss to long-term follow-up was acceptable and close to those achieved by Townsend et al.48 at 6 months and Huffman et al.15 in their similar studies (Townsend: 70% and Huffman: 55% at 6 months and 47% at 12 months vs 64% and 60%, respectively, in this study), when the limitations are considered together, there may be legitimate questions about the generalizability of these findings. It means that replication of the findings is warranted.

The instability in the number of people taking opioids between initial assessment and the treatment program may relate to the preparation undertaken in this period. It is likely that patients on opioids (particularly >40 mg oMEDD) would have engaged in a pretreatment reduction program involving exposure to the nonmedication pain management skill training used in the treatment program on an individual basis. Patients not on opioids at initial assessment would more likely have had less specific exposure to this training. The impressive reduction in those taking opioids, 79/103 (77%), supports individual therapy with opioid cessation as a goal; however, the long-term maintenance of cessation in this setting remains unclear from these results, given all patients then entered the program.

Despite these limitations, the consistency of these findings with those reported from the previous RCT in the U.K. study,55 where similar reductions in opioid use and maintenance were seen with an inpatient version of the program used in this study, and the U.S. studies15,49 provides a degree of confidence that the current findings are reliable, especially when it is considered that they were conducted in quite different health systems. It is also worth noting that in the U.K. study, the outpatient group program (24 hours over 8 weeks) achieved far less opioid cessation relative to both the inpatient program in that study and this outpatient study, with 43% of patients still taking opioids at the end of the briefer (U.K.) outpatient program, vs only 10.3% in this study. In sum, the similar findings in achieving sustained opioid cessation between the U.K. inpatient program (of about 140 hours over 4 weeks), the 2 outpatient programs conducted in the United States (3 week, 120 hours48; 3-4 weeks, 142-180 hours15), and the similar Sydney outpatient program of about 115 hours over 3 weeks suggest that this interdisciplinary, cognitive-behavioural therapy-informed approach to pain management can be effective in not only sustainably reducing reliance on opioids for patients with chronic pain, but also in enhancing functional, emotional, and cognitive outcomes at the same time.

Williams54 discussed the potential complexity of determining necessary and sufficient “doses” of pain rehabilitation treatments. But, there is broad evidence that for the more disabled, depressed, and medication-dependent patients with chronic pain, programs that are more comprehensive in content (commensurate with the complexity of the patients treated) and length (time) are likely to be more successful than briefer and less comprehensive versions of similar programs. Specific examples in support of this case can be seen in the studies by Williams et al.,53,55 Haldorsen et al.,13 and in conclusions drawn from systematic reviews on the topic.12,51

Although the necessary content of these programs may be debated, the central issue is that they offer patients an alternative to agents such as opioids or interventional procedures for managing their pain while also achieving their desired life goals. This is important because it is unlikely patients would be prepared to consider reducing their use of opioids unless they felt there was a viable alternative. Gradual withdrawal from opioids while learning pain self-management strategies are central features of programs like those described in this study. In this regard, it should be noted that Andrews et al.1 found that overdoing (or poor activity pacing) of activities in a chronic pain sample was related to higher use of opioids on an “as needed” basis. One of the strategies for overcoming the tendency of overdoing is activity pacing and this was a key feature in both the program evaluated by Williams et al.57 as well as the Sydney program described in this study. Consistent with this position, in 2 previous studies, our group has found that more consistent and regular use of identified pain self-management strategies (including pacing) was related to better outcomes (for pain, disability, depression severity, pain cognitions, and reduced medication use) after participation in our program.34,35

Conflict of interest statement

Three authors (M.K.N., L.B., and A.M.) receive royalties from the book, Manage Your Pain, used in the pain management program. The remaining authors have no conflicts of interest to declare.


The authors acknowledge the contributions to the study by other ADAPT and PMRI staff, especially Michael Cousins, Sarah Overton, Bradley Wood, Lois Tonkin, Maria de Sousa, Damien Finniss, Ann Sutherland, Mandy Corbett, Dan Costa, and Wendy O'Hanlon-Mack. The authors also thank the patients who agreed to participate in the study, and the reviewers who made a number of valuable suggestions for the study.

The original study on which this study was based was supported by a grant from the National Health & Medical Research Council of Australia to M.K.N. and L.S. (Grant No. 402524).


[1]. Andrews NE, Strong J, Meredith PJ, Fleming PJ. The relationship between overactivity and opioid use in chronic pain: a five-day observational study. PAIN 2016;157:447–55.
[2]. Asghari A, Nicholas MK. Pain self-efficacy beliefs and pain behaviour: a prospective study. PAIN 2001;94:85–100.
[3]. Chou R, Turner JA, Devine EB, Hansen RN, Sullivan SD, Blazina I, Dana T, Bougatsos C, Deyo RA. The effectiveness and risks of long-term opioid therapy for chronic pain: a systematic review for a National Institutes of Health Pathways to Prevention Workshop. Ann Intern Med 2015;162:276–86.
[4]. Cohen J. Statistical power analysis for the behavioural sciences. 2nd ed. Hillsdale: Erlbaum, 1988.
[5]. Deyo RA, Von Korff M, Duhrkoop D. Opioids for low back pain. Br Med J 2015;350:g6380.
[6]. Darnall BD, Ziadni MS, Stieg RL, Mackey IG, Kao MC, Flood P. Patient-centered prescription opioid tapering in community outpatients with chronic pain. JAMA Intern Med 2018;178:707–8.
[7]. Faculty of Pain Medicine Australian and New Zealand College of Anaesthetists. PM01: recommendations regarding the use of opioid analgesics in patients with chronic non-cancer pain. 2015. Available at: Accessed April 21, 2018.
[8]. Flor H, Behle DJ, Birbaumer N. Assessment of pain-related cognitions in chronic pain patients. Behav Res Ther 1993;31:63–73.
[9]. Fordyce W. Behavioral methods of control of chronic pain and illness. St. Louis: Mosby, 1976.
[10]. Gatchel RJ, McGeary DD, McGeary CA, Lippe B. Interdisciplinary chronic pain management: past, present, and future. Am Psychol 2014;69:119–30.
[11]. Gil KM, Ross SL, Keefe FJ. Behavioural treatments of chronic pain: four pain management protocols. In: France RD, Ranga K, Krishnan R, editors. Chronic pain. Washington: American Psychiatric Association, 1989. p. 376–413.
[12]. Guzmán J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidisciplinary rehabilitation for chronic low back pain: systematic review. Br Med J 2001;322:1511–6.
[13]. Haldorsen EMH, Grasdal AL, Skouen JS, Risa AE, Kronholm K, Ursin H. Is there a right treatment for a particular patient group? Comparison of ordinary treatment, light multidisciplinary treatment, and extensive multidisciplinary treatment for long-term sick-listed employees with musculoskeletal pain. PAIN 2002;95:49–63.
[14]. Hauser W, Schug S, Furland AD. The opioid epidemic and national guidelines for opioid therapy for chronic noncancer pain: a perspective from different continents. Pain Rep 2017;2:e599.
[15]. Huffman KL, Rush TE, Fan Y, Sweis GW, Vij B, Covington EC, Scheman J, Mathews M. Sustained improvements in pain, mood, function and opioid use post interdisciplinary pain rehabilitation in patients weaned from high and low dose chronic opioid therapy. PAIN 2017;158:1380–94.
[16]. Hunsinger M, Smith SM, Rothstein D, McKeown A, Parkhurst M, Hertz S, Katz NP, Lin AH, McDermott MP, Rappaport BA, Turk DC, Dworkin RH. Adverse event reporting in non-pharmacologic, non-interventional pain clinical trials: ACTTION systematic review. PAIN 2014;155:2253–62.
[17]. Jacob MC, Kerns RD. Assessment of the psychosocial context of the experience of chronic pain. In: Turk DC, Melzack R, editors. Handbook of pain assessment. 2nd ed. New York: The Guilford Press, 2001. p. 362–84.
[18]. Jacobson NS, Revensdorf D. Statistics for assessing clinical significance of psychotherapy techniques: issues, problems, and new developments. Behav Assess 1998;19:133–45.
[19]. Jacobson NS, Roberts LJ, Berns SB, McGlinchey JB. Methods for defining and determining the clinical significance of treatment effects: description, application, and alternatives. J Consult Clin Psychol 1999;67:300–7.
[20]. Kendal PC, Marrs-Garcia A, Nath SR, Sheldrick RC. Normative comparison for the evaluation of clinical significance. J Consult Clin Psychol 1999;67:285–99.
[21]. Kerns RD, Turk DC, Rudy TE. The West Haven–Yale multidimensional pain inventory (WHYMPI). PAIN 1985;23:345–56.
[22]. Kidner CL, Mayer TG, Gatchel RJ. Higher opioid doses predict poorer functional outcome in patients with chronic disabling occupational musculoskeletal disorders. J Bone Joint Surg Am 2009;91:919–27.
[23]. Kori SH, Miller RP, Todd DD. Kinesiophobia: a new view of chronic pain behaviour. Pain Manag 1990;3:35–43.
[24]. Lovibond PF, Lovibond SH. The structure of negative emotional states: comparison of the depression, anxiety and stress scales (DASS) with the Beck depression and Anxiety inventories. Behav Res Ther 1995;33:335–43.
[25]. McCracken LM, MacKichan F, Eccleston C. Contextual cognitive-behavioral therapy for severely disabled chronic pain sufferers: effectiveness and clinically significant change. Eur J Pain 2007;11:314–22.
[26]. McCracken LM, Vowles KE, Eccleston C. Acceptance of chronic pain: component analysis and a revised assessment method. PAIN 2004;107:159–66.
[27]. McPherson S, Smith CL, Dobscha SK, Morascod BJ, Demidenko MI, Meathd THA, Lovejoy TI. Changes in pain intensity after discontinuation of long-term opioid therapy for chronic noncancer pain. PAIN 2018;159:2097–104.
[28]. Meldrum ML. The ongoing opioid prescription epidemic: historical context. Am J Pub Health Perspect 2016;106:1365–6.
[29]. Mitchell TB, White JM, Somogyi AA, Bochner F. Slow-release oral morphine versus methadone: a crossover comparison of patient outcomes and acceptability as maintenance pharmacotherapies for opioid dependence. Addiction 2004;99:940–5.
[30]. Morley S, Williams AC. Conducting and evaluating treatment outcome studies. In: Gatchel R, Turk DC, editors. Psychosocial factors in pain. New York: Guilford Press, 2002. p. 52–68.
[31]. Nicholas MK. The pain self-efficacy questionnaire: taking pain into account. Eur J Pain 2007;11:153–63.
[32]. Nicholas MK, Asghari A. Investigating acceptance in adjustment to chronic pain: is acceptance broader than we thought? PAIN 2006;124:269–79.
[33]. Nicholas MK, Asghari A, Blyth FM. What do the numbers mean? Normative data in chronic pain measures. PAIN 2008;134:158–73.
[34]. Nicholas MK, Asghari A, Corbett M, Smeets RJ, Wood BM, Overton S, Perry C, Tonkin LE, Beeston L. Is adherence to pain self-management strategies associated with improved pain, depression and disability in those with disabling chronic pain? Eur J Pain 2012;16:93–104.
[35]. Nicholas MK, Asghari A, Sharpe L, Brnabic A, Wood BM, Overton S, Tonkin L, de Sousa M, Finniss D, Beeston L, Sutherland A, Corbett M, Brooker C. Cognitive exposure versus avoidance in patients with chronic pain: adherence matters. Eur J Pain 2014;18:424.
[36]. Nicholas MK, Blyth FM. Are self-management strategies effective in chronic pain treatment? Pain Manag 2016;6:75–88.
[37]. Nicholas MK, George S. Psychologically informed interventions for physical therapists. Phys Ther 2011;91:765–76.
[38]. Nicholas MK, Molloy A, Tonkin L, Beeston L. Manage your pain. 2nd ed. Sydney: ABC Books, 2006.
[39]. Penington Institute. Australia's annual overdose report. 2016. Available at: Accessed March 15, 2018.
[40]. Ralphs JA, Williams AC, Richardson PH, Pither CE, Nicholas MK. Opiate reduction in chronic pain patients: a comparison of patient-controlled reduction and staff controlled cocktail methods. PAIN 1994;56:279–88.
[41]. Roland M, Morris R. A study of a natural history of the back pain. Part 1. Development of a reliable and sensitive measure of disability in low-back pain. Spine (Phila Pa 1976) 1983;8:141–4.
[42]. Rosenthal R. Meta-analytic procedures for social research. Newbury Park: SAGE Publications, Incorporated, 1991.
[43]. Schofield DJ, Shrestha RN, Percival R, Callander EJ, Kelly SJ, Passey ME. Early retirement and the financial assets of individuals with back problems. Eur Spine J 2011;20:731–6.
[44]. Syed UAM, Aleem AW, Wowkanech C, Weekes D, Freedman M, Tjoumakaris F, Abboud JA, Austin LS. The effect of preoperative education on opioid consumption in patients undergoing arthroscopic rotator cuff repair: a prospective, randomized clinical trial. J Shoulder Elbow Surg 2018;27:962–7.
[45]. Sullivan M. Managing pain medications: an early text in pharmaco-psychology. In: Main CJ, Keefe FJ, Jensen MP, Vlaeyen JWS, Vowles KE, editors. Fordyce's behavioral methods for chronic pain and illness. Philadelphia, PA: IASP Press, 2015. p. 300–9.
[46]. Sullivan MD, Ballantyne J. Must we reduce pain intensity to treat chronic pain? PAIN 2015;157:65–9.
[47]. Taylor R, Lovibond PF, Nicholas MK, Cayley C, Wilson PH. The utility of somatic items in the assessment of depression in chronic pain patients: a comparison of the Zung Self-rating Depression Scale (SDS) and the Depression Anxiety Stress Scales (DASS) in chronic pain, clinical and community samples. Clin J Pain 2005;21:91–100.
[48]. Townsend CO, Kerkvliet JL, Bruce BK, Rome JD, Hooten WM, Luedtke CA, Hodgson JE. A longitudinal study of the efficacy of a comprehensive pain rehabilitation program with opioid withdrawal: comparison of treatment outcomes based on opioid use status at admission. PAIN 2008;140:177–89.
[49]. Turner JA, Calysyn DA, Fordyce WE, Ready LB. Drug utilization patterns in chronic pain patients. PAIN 1982;12:357–63.
[50]. Vlaeyen JW, de Jong J, Geilen M, Heuts PHTG, van Breukelen G. The treatment of fear of movement/(re)injury in chronic low back pain: further evidence on the effectiveness of exposure in vivo. Clin J Pain 2002;18:251–61.
[51]. Waterschoot FPC, Dijkstra PU, Hollak N, de Vries HJ, Geertzen JHB, Reneman MF. Dose or content? Effectiveness of pain rehabilitation programs for patients with chronic low back pain: a systematic review. PAIN 2014;155:179–89.
[52]. Williams ACDC, Eccleston C, Morley S. Psychological therapies for the management of chronic pain (excluding headache) in adults. Cochrane Database Syst Rev 2012;CD007407.
[53]. Williams ACdeC, Nicholas MK, Richardson PH, Pither CE, Fernandes J. Does randomisation affect the generality of findings from a controlled trial? The effects of patient preference versus randomisation on inpatient versus outpatient chronic pain management. PAIN 1999;83:57–65.
[54]. Williams ACdeC. How do we understand dose of rehabilitative treatment? Commentary. PAIN 2014;155:8–9.
[55]. Williams ACdeC, Richardson PH, Nicholas MK, Pither CE, Harding VR, Ralphs JA, Ridout KL, Richardson IH, Justins DM, Chamberlain JH. Inpatient versus outpatient pain management: results of a randomised controlled trial. PAIN 1996;66:13–22.
[56]. Windmill J, Fisher E, Eccleston C, Derry S, Stannard C, Knaggs R, Moore RA. Interventions for the reduction of prescribed opioid use in chronic non-cancer pain. Cochrane Database Syst Rev 2013;CD010323.

Opioids; Chronic pain; Interdisciplinary pain management; Pain self-management; Effectiveness study; Long-term follow-up

© 2019 International Association for the Study of Pain