Setting and Participants

Peer coaches and patients were recruited from the 5 primary care clinics at Roudebush VA Medical Center (RVAMC) in Indianapolis, IN. We first obtained permission from primary care providers (PCPs) to recruit from their patient panel. Because this was a pilot study, we recruited from two PCPs’ panels to meet our recruitment goal. Patients had been diagnosed with musculoskeletal pain (ICD-9 codes 715, 719, 721, 722, 723, 724, 726, 729.0, 729.1, 729.3, 729.5, 738.4, 738.5) that had persisted for at least 6 months, and had at least moderate pain severity, defined by pain ≥ 5 on a 0 (no pain) to 10 (worst pain imaginable) scale. Patients were excluded if they had been hospitalized for psychiatric or substance abuse reasons in the last 6 months, had active suicidal ideation, prior or pending back surgery, severe medical conditions (e.g., New York Heart Association Class III or IV heart failure) that precluded participation, or severe hearing or speech impairment.

Peer Coaches

Peer coaches had participated in one of two prior studies at RVAMC involving pain self-management and had consented to be contacted for future studies.

Intervention

Peer coaches (n=10) attended a 3-hour training session co-led by the study psychologist and nurse. Training consisted of a didactic session, which explained and reviewed chronic pain basics and pain self-management strategies; goal setting, including teaching coaches to guide others in this activity; and motivational interviewing strategies. Demonstrations and role-playing were used.

After training, each peer coach was assigned 2 patients to “coach” and support for 4 months. To the extent possible, assignments were based on pain location. When this was not achievable, pairs were matched as closely as possible according to age. We assigned 2 patients per coach in an effort not to over-burden any individual coach. Peer coach-patient pairs were instructed either to meet in person, through phone calls, or a combination of both, a minimum of twice per month for the 4-month period. All participants were given a study manual with the following 6 sections: 1) chronic pain basics; 2) relaxation skills; 3) activity pacing; 4) cognitive behavioral skills, 5) self-care skills, and 6) interpersonal skills. In addition, the following sections were unique to the peer coach manual: 1) what is a peer; 2) cultural competence; 3) communication skills; 4) managing crisis and emergency situation; and 5) motivational strategies.

Peer coaches were asked to draw on the manual as they saw appropriate, while being flexible and responsive to each patient's needs. Coaches were encouraged to share their own experiences with pain management, including successes and failures, to share strategies that worked for them, help find appropriate strategies for assigned patients, and help set pain self-management goals. In addition, coaches were encouraged to engage patients in social conversation, as appropriate, and offer support and motivation.

Intervention Fidelity

We used several facilitation strategies during the intervention to optimize fidelity. First, peer coaches participated in a 3-hour training session at the beginning of the study. Three separate training sessions were held; all were audio recorded to ensure quality and consistency. Second, peer coaches participated in supervision calls twice per month. During these “booster” sessions, the study psychologist (MK) reviewed, emphasized, and, if necessary, re-educated coaches on expectations for the intervention, such as setting and reviewing goals with patients. Third, a detailed study manual, described above, provided content for peer coaches and their patients to reference and use as needed. In addition to these facilitation strategies, peer coaches and patients were asked in an interview at the end of the study about their experiences with the trial, including the content of their meetings. These included open-ended questions (“What did you talk about in your meetings?”) as well as closed-ended questions (“Did you set goals with your peer coach?”). Responses to these questions will be used to facilitate development of a systematic fidelity checklist for use in the follow-up study.

Measures

All patient outcomes were assessed at baseline and 4-month post-intervention follow-up. Pain was the primary outcome and was assessed with the PEG, a validated 3-item version of the Brief Pain Inventory (12, 13), and the PROMIS Pain Interference measures. The PROMIS symptom measures have had extensive development and population validation by NIH and their use in research is being encouraged across multiple studies, facilitating intra- and inter-disease comparisons (14).

We also assessed several secondary measures. Depression was measured with the PHQ-9. Several studies have validated the PHQ-9 as a diagnostic measure with excellent psychometric properties. Internal consistency has consistently been shown to be high (Cronbach's α > 0.80) and test-retest assessment shows the PHQ-9 to be a responsive and reliable measure of depression treatment outcomes (15).

Anxiety was measured with the GAD-7, which has demonstrated reliability (α = 0.89) and validity (criterion, construct) as a measure of anxiety in the general population and primary care (16).

Self-efficacy was measured with the Arthritis Self-Efficacy Scale (17), a 6-item measure that has been used in prior studies of patients with chronic pain (5, 18).

Perceived social support was measured with the Multi-Dimensional Scale of Perceived Social Support (MPSS). The MPSS includes 12, 7-point Likert scale items. The test-retest reliability and internal consistency is high, ranging from α=.84-.95 across a variety of studies (19, 20).

Patient activation refers to a patient's knowledge, skill, and confidence to self-manage one's chronic health condition (21). Activation was measured with the Patient Activation Measure (PAM) 13-item Short Form. The PAM has been demonstrated reliable and valid in a variety of studies, with reliability ranging from α =.87-.88 (21-24).

Negative pain cognitions were assessed with two measures: the Pain Catastrophizing Scale and the Centrality of Pain Scale. The Pain Catastrophizing Scale is a 13-item scale that assesses catastrophizing—a pain belief that has been found to be a strong predictor of poor treatment response. Validation studies have found strong evidence of criterion-related, concurrent, and discriminant validity (25). Centrality of pain refers to the degree to which a person views pain as a dominant feature of one's life and identity (26). The Centrality of Pain Scale is a 10-item instrument, with responses measured on a 5-point Likert scale that range from “strongly disagree” to “strongly agree.” In its original validation study, the scale demonstrated high internal consistency (Cronbach's α=.90) and construct validity (26). Questions include “Pain controls my life,” and “My pain consumes all of my energy.”

Participants were not compensated directly for participation in the study (i.e., for meeting with their assigned partners). However, peer coaches were paid $30 to attend the initial training, and peer coaches and patients were paid $30 for outcome assessments.

Meetings

Based on peer coach report, number of meetings ranged from 3 to 16, with a median of 6 meetings. The majority of meetings took place over the phone. Some peer-patient pairs chose to meet at the VA, either in the cafeteria or coffee shop, at least once during the intervention, but met via telephone for the remainder of meetings.

Fidelity

To enhance fidelity for this pilot study we used several facilitation strategies, described in the Methods section. Data from post-intervention interviews will be used to create a systematic fidelity checklist for use in a large follow-up study.

Baseline characteristics

Peer coaches’ ages ranged from 50-71 (Mean=60, SD=7) years and all were male veterans. Eight were White, 1 Black, 1 Hispanic. See Table 1 for peer coach and patient demographics. Baseline characteristics of peer coaches are in Table 2.

Patient demographics and baseline scales did not differ significantly between completers (n = 17) and non-completers (n = 3), with the exception of employment status (Fisher's exact test p-value = .046). All non-completers were employed or retired, whereas 65% of completers were unable to work. For these reasons and the small number of non-completers, all non-completers were dropped from analysis.

Patients’ ages ranged from 35-66 (Mean=58, SD=8) years; 9 were White and 8 were Black. All were male veterans. Patients’ pain locations were as follows: low back (n=8), neck (6), knees (1), shoulders (1), “everywhere” (1).

The project reported here was supported by the Department of Veterans Affairs, Veterans Health Administration, Quality Enhancement Research Initiative (QUERI) and Health Services Research and Development (RRP 12-438 and CDA 10-034). The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs or the United States Government. The authors would like to thank Christy Sargent and Erica Evans for their assistance with data collection.