The effect of peer group management intervention on chronic pain intensity, number of areas of pain, and pain self-efficacy

2.1 Study settings

Our study was a stepped wedge cluster randomized controlled trial (Figures 1 and 2) conducted at OH, where the OH-using patients had requested a peer group intervention. Occupational health professionals (doctors, psychologists, physiotherapists, and nurses) recruited 60 patients with chronic pain to pain peer groups for 1 month in 2016, from mid-January to mid-February. Subjects who met the inclusion criteria and were interested in participating in the study were interviewed individually (H.G. or M.R.) and informed about the study, such as that participation is voluntary, and the study will not interfere with their use of occupational health care. After written informed consent, the interviewer randomized the blinded participants into three active and three waiting groups, 4–10 people per group (Figure 1). At the end of the interview, participants selected one, randomly selected, sealed envelope from all mixed, sealed, white envelopes, containing information from the randomization into one of six groups (A, B, C, D, E, and F).

Figure 1

Flow chart of the study. Scr = screening questionnaire, Bef1 = before intervention questionnaire for groups ABC and DEF (first before intervention questionnaire), DEF had a 5-month waiting period, Bef2 = before the intervention questionnaire for group DEF (second before the intervention questionnaire), Aft = immediately after the intervention questionnaire for groups ABC and DEF, Aft6m = 6 months after the intervention questionnaire for groups ABC and DEF, and Aft 12m = 12 months after the intervention questionnaire, only for half of the dusters, for groups ABC.

Figure 2

Study design of a stepped-wedge cluster randomized controlled trial. Shaded cells represent after intervention periods, and blank cells represent before intervention/control periods. This trial has six data-collection points. Data-collection point questionnaires are in Figure 1.

Three groups, A, B, and C, started the intervention immediately and had six meetings from March 15, 2016, to April 21, 2016. Three groups, D, E, and F, started the intervention after 5 months of waiting, from August 16, 2016, to September 30, 2016. The 6th-month follow-up meeting for groups A, B, and C was in October, and the 12th-month follow-up meeting was in March. The 6th-month follow-up meeting for groups D, E, and F was also in March (Figure 1).

2.2 A stepped wedge cluster randomized controlled trial

In stepped wedge trials, the randomized clusters start intervention at different times; each participant is finally in the intervention. Randomization, as in this study, increases the validity of the study [31]. Many stepped wedge trials are analyzed using mixed model statistics including a random effect for cluster and fixed effects for period to account for secular trends, combining both vertical and horizontal effects of the intervention and control cluster comparisons. Randomization in stepped wedge trials is crucial due to the potential influences of time on the data, allowing for improved control over biases arising from secular trends [31].

2.3 Sample size calculation

In the presence of several outcomes, the sample size should be calculated for the outcome with the smallest difference in effect size before and after [28]. Salaffi et al. [29] stated that the minimal cut-off point measure of the Numeric Rating Scale of pain is −1.0 cm. When pain is 7.5 before intervention, and after intervention 6.5, the effect size is 13%. The difference in pain acceptance and fear-avoidance beliefs may even be larger than 15% pre- and post-intervention [29]. By using stepped wedge analysis on the study design matrix (six groups with four observations of each group) and with a 15% detectable minimum difference of the outcome (chronic pain, mean 7.14 on a scale from 0 to 10), we calculated the minimum group sample size to be 10 (when a-error is 0.05 and b-error 0.80) and total number of participants to be 60 [30].

2.4 Intervention

The intervention consisted of a pain management group that included peer support, pain education about the biopsychosocial nature of pain, and chronic pain. Before the intervention, there was a coffee break and a mindfulness exercise or relaxation. During the intervention, the homework included chapters from Helena Miranda’s book “Rethinking Pain” [34]. In the first meeting, the discussion subject was the pain management tools already in use; participants had used a wide range of different drug-free chronic pain treatment methods. In the second meeting, the discussion subject was pain mechanisms. The theme of the third meeting was (1) to do enjoyable things that would help improve pain management and (2) to discuss how pain affects social life. In the fourth meeting, the group leaders led discussions on the meaning of work, the future in life, work goals and professional development, and how to improve workability. At the fifth meeting, the topics were feelings; love, touch, sexuality, maintaining a positive attitude, and having a pet. In the last sixth meeting, participants discussed new tools to improve pain management, and they made a pain management plan. The themes in the 6th-month meeting were psychological flexibility, the importance of education and awareness skills, working on rehabilitation, and updating the pain management plan [26].

Interventions included CBT, mindfulness meditation, relaxation, ACT, homework, and the perspective of return to work [26].

CBT, which we used in this study, is based on talk therapy by psychotherapist Aaron Beck in 1964. It focuses on changing thoughts (cognitions) and actions (behavior), which is why it is called cognitive behavioral therapy [33].

In mindfulness practice, we are consciously present in the present moment, we welcome everything that comes to mind, and we immediately let it go [39].

ACT is more comprehensive than CBT including mindfulness practices, the pursuit of psychological flexibility by “unpacking” cognitions, using metaphor, and engaging in value-related activities [6].

2.5 Outcomes, variables, and covariates

2.6 Questionnaires

We used a modified Örebro Musculoskeletal Pain Screening Short Questionnaire (OMPSQ) [35]. In OMPSQ, we changed the first question “How long have you had your current pain problem?” to “During the past week, where have you felt pain?” We asked for pain in ten areas: head, upper back, neck or shoulder area, lower back, shoulder or upper arm, hip or thigh, elbow or arm, knee or lower leg, wrist or hand, and ankle or foot. The intensity of pain during the past week was assessed by a VAS for pain [36] of OMPS [35]. The Pain Self-Efficacy Questionnaire included ten items about coping with pain [37].

The number of observations (n) is greater than the number of participants due to the stepped-wedge design in which each participant receives the intervention. The questionnaires were filled out many times by the same participants (Table 2 and Figures 1 and 2).

2.7 Study population and recruitment

The inclusion criteria were as follows:

• Helsinki city employee,

• Age 18–65 years, and

• Any chronic pain, i.e., pain lasting for >3 months.

• The study participant is willing to share their thoughts in a peer group intervention.

• The risk of work disability was high in this population, i.e., ≥50 points (0–100) by the modified Örebro questionnaire [35].

The exclusion criteria were as follows:

• Malignant diseases such as cancer or severe mental illness,

• Participation in another pain-management group,

• A major current psychological or physical life crisis, and

• Risk of work disability low, i.e., <50 points by the modified Örebro questionnaire.

2.8 Statistical analyses

We compared the post-intervention period with the pre-intervention (control) period and adjusted for clustering effect and calendar time. We used linear mixed-effects models to analyze data from repeated measures, and we adjusted the differences in the outcomes of interest between intervention and control periods for the clustering effect and calendar time. We used random effects to model the correlation between individuals in the same cluster and a fixed effect for the time interval.

4.1 Results

In this stepped wedge randomized study conducted in occupational health care, a long-term effect on multisite pain and its tolerability was achieved with the group intervention. The number of areas of pain decreased, and pain-related self-efficacy increased for 6–12 months.

Most (80%) of our study participants were satisfied with all nonpharmacological treatments, and only a few (20%) of open answers were critical. Some felt that the study intervention was too short, and some were a bit tired from the length and homework [38]. At the doctor’s office, strengthening the pain patient’s self-efficacy increases pain tolerance and reduces disability, workability, and functional ability [37].

4.2 Finnish OH

Finnish occupational health care is part of basic health care. Since its goal is to prevent and detect work-related health problems and support work ability, occupational health care personnel know the working conditions [32]. We conducted this study in an OH unit, where the staff had expertise in promoting group therapy, especially in supporting pain management measures.

4.3 Comparing the results with other intervention studies

In our stepped wedge study, we accepted all types of pain except cancer pain. However, an RCT by Turner et al. compared Mindfulness-Based Stress Reduction (MBSR), CBT, and usual care in back pain patients, and CBT and MBSR increased self-efficacy [39]. Research results from Turner et al. are in line with our research results.

Also, the Lamb et al. study included 701 patients with subacute or chronic LBP. They had six CBT intervention meetings as we did, and pain self-efficacy improved significantly in the intervention group after a 12-month follow-up [40]. Research results from Lamb et al. are in line with our research results.

In addition, Carpenter et al. found that an online CBT intervention improved self-efficacy in chronic LBP patients post-intervention but not at the 6-week follow-up. The results parallel our research results; in our study, self-efficacy also improved after intervention and at the 6–12-month follow-up [41].

An Australian RCT by Nicholas et al. compared pain self-management (PSM), including CBT, exercises, and completing a home exercise chart with praise and encouragement. Furthermore, they compared exercise-attention control (EAC) including CBT, exercise, and clinical psychologists who offered discussions about living with pain. They included also a waiting list (WL) control like our study. EAC group participants were not encouraged to exercise or practice at home and completed no home exercise charts. WL controls started intervention (EAC/PSM group) 12 weeks later [42]. In the Nicholas et al. study, the PSM group increased pain self-efficacy after and at 1-month follow-up compared to the EAC group [42].

Also, in our study, the intervention included, e.g., CBT and homework, which participants discussed in every meeting, and they also discussed “exercise with joy” in the fifth meeting. In our study, the waiting list started intervention 5 months after the first three groups [26]. In addition, in our study, each participant completed their pain management plan with the encouragement and support of the group instructors, which probably strengthened their self-efficacy and praised and encouraged the group participants [26]. Research results from Nicholas et al. [42] parallel our research results, but in our study, self-efficacy increased up to the 6–12-month follow-up.

A German meta-analysis and systematic review by Bernardy et al. found that CBT in fibromyalgia patients had no effect on pain but an effect on pain self-efficacy at the median intervention duration of 9 weeks [43]. Bernardy et al. systematic review results parallel our research results, but our results lasted longer, 6–12 months.

4.4 Multisite pain

Common comorbidities with recurrent and widespread pain are depression and anxiety [9]. When pain, depression, and anxiety are simultaneously treated, pain may decrease, and high self-efficacy prevents depression and multisite pain [47]. Prolonged or frequent sick leave and physical incapacity weaken the financial situation of a pain patient, and a poor economic situation increases the risk of multisite pain [48]. Furthermore, opioids increase depressive symptoms, risk of disability, disability increases poverty, and the use of health services for patients with chronic pain [49]. Our participants had a high number of pain sites at baseline, 5.96, i.e., almost six different areas of pain on average. Our intervention increased self-efficacy, which may have influenced the number of areas of pain.

Comparing the results of areas of pain with other intervention studies is challenging, as our study is the only study in which psychological treatments reduced the number of areas of pain. One of the common causes of chronic pain, single and multisite pain, is perceived stress and stress reaction systems [22]. However, psychological treatments are effective for stress [50]. Because our study included stress-reducing psychological treatments, using CBT, mindfulness, and learning relaxation skills, they may have helped to reduce pain areas [26]. Patient selection, working-aged people, and OH settings with good therapeutical resources may have also influenced the results. Also, versatile intervention methods may have affected the results.

Only a few previous prospective studies have evaluated the effectiveness of psychosocial interventions in treating multisite pain, and this is the first study in occupational health that has attempted to reduce multisite pain. Rundell et al. found multisite pain in back pain patients to predict disability, which is why reducing multisite pain is necessary. However, pain research has focused on individual pains [44]. The more extensive the pain, the more sick leave days, and work restrictions, so decreasing multisite pain should increase working capacity [45]. Physical work is a risk for multisite pain, especially in female-dominated fields and middle- or low-income earners [46]. Also, the most common occupation of the pain patients who participated in our study was physical, such as the social and health sector or nursing home nurse, home nurse, early childhood educator, and as well lighter office work, because office work mainly involves working in a static sitting position, a factor influencing chronic pain.

In other studies, the number of areas of pain has been reduced mainly by exercise, but in our intervention, we did not have structured exercise. However, pain management tools already in use; like exercise were discussed in pain management groups. In the Cantarero-Villanueva et al. study, 8 weeks of water exercise therapy reduced neck, shoulder, or arm pain in breast cancer survivors. However, widespread pressure pain hyperalgesia did not change significantly [51]. Reducing multisite pain is necessary because, in older age (50–64 years old), multisite pain increases the risk of poor workability and earlier retirement [52]. In addition to being associated with disability, multisite pain increases healthcare costs [53].

4.5 Pain

Measuring pain intensity in VAS daily is challenging because changes in psychological or functional parameters may invalidate or compensate for changes in pain intensity [54]. Patients with chronic pain have often suffered from disabling chronic pain for years or even their entire life [55], which is accompanied by many pain-related symptoms, such as depression and insomnia [56]. For example, the OMPSQ is a broader questionnaire with a VAS and questions about depression and insomnia [35].

A Cochrane review by Williams et al. studied chronic pain (excluding headache) in adults and found that CBT has minor to moderate effects on pain immediately after treatment but disappeared at a 6-month follow-up [23]. The results of a Cochrane review by Williams et al. are parallel to our research results.

Macea et al. conducted a systemic review and meta-analysis and found that web-based CBT interventions lead to modest reductions in chronic pain [25]. According to a systematic review by Skelly et al., CBT slightly improved fibromyalgia pain [20]. In addition, CBT improved the functional ability of the fibromyalgia patient, and the effects persisted to 1–6-month follow-up [20]. The results of a systematic review by Skelly et al. and Macea et al. are parallel to our study results.

RCT by Basler et al. studied chronic low back patients, for over 7 weeks. The intervention lasted 12 weeks, longer than in our study [57]. Basler et al. compared a combination of usual care and CBT to usual care and had a 6-month follow-up [57]. Both the Basler et al. study and our study had a waiting list group. In Basler’s study, those who received CBT and usual care reported less pain than those who received only usual care. Pain may have decreased more in Basler’s study than in our study because the duration of the intervention was more prolonged, 12 weeks [57].