Neck pain is common in adults aged >50 years and is the fourth leading cause of disability in the United States [1]. Chronic neck pain (CNP) reduces functional status and quality of life and is associated with deleterious psychological outcomes in affected individuals. CNP accounts for more than 10 million ambulatory medical visits per year [2].

Medications are most commonly used to treat CNP in clinical practice; however, systematic reviews find limited evidence for effective treatments relative to low back pain, and current therapies show only modest effect sizes [3]. Because medications and other conventional treatments often fail to substantially relieve pain, patients frequently seek other treatments.

In all, 58% of the older adults surveyed used some type of alternative treatment [4], and pain is the primary reason individuals use complementary treatments [5]. After low back pain, CNP is the most common pain condition for complementary health use [6]. Massage is the second (after chiropractic) most commonly used complementary treatment for CNP [7]. Studies have shown that massage is safe, with few risks and rare serious adverse effects [8,9].

Despite its safety and potential benefits, the expense associated with massage therapy limits its accessibility. Teaching informal care allies to provide massages has the potential to improve accessibility. Kozak et al [10] demonstrated the feasibility of caregiver-delivered massage, which led to significant decreases in pain, stress or anxiety, and fatigue in Veterans with cancer. Collinge et al [11] recruited 97 patient or caregiver dyads to practice massage for individuals experiencing cancer. Caregiver-applied interventions led to decreases in patients’ pain, depression, and other cancer-related symptoms [11]. Although the feasibility and benefit of the care ally–applied massage approach have been established in cancer populations, it has not been examined specifically in non–cancer-related musculoskeletal pain populations.

The Trial Outcomes for Massage: Care Ally–Assisted Versus Therapist-Treated (TOMCATT) study began as a 3-arm, randomized controlled trial of 2 massage therapy delivery approaches for CNP. The primary aim of this study was to compare care ally–assisted massage (CA-M) and therapist-treated massage (TT-M) with a wait-list control group (WL-C) for pain-related disability. The secondary outcomes included pain severity, health-related quality of life, depression, anxiety, and stress. Before pausing the TOMCATT activities due to the COVID-19 pandemic, the study design was modified to a 2-arm study. The CA-M arm was discontinued owing to recruitment and adherence challenges.

The primary purpose of this manuscript is to describe the original TOMCATT study design, the challenges faced during the study’s first 2 years of recruitment, and the resultant design modification to a 2-arm study. Specifically, we describe the original TOMCATT methodology, materials, and procedures in the Methods section and report the initial enrollment and intervention initiation outcomes for the first 30 recruitment months, rationale for the modified study design, and subsequent modified TOMCATT methodology in the Results section. TOMCATT is an active study that has restarted from the COVID-19–related pause and is projected to be concluded by early 2023.

TOMCATT’s study population consisted of Veterans with CNP randomized to one of 3 study arms: CA-M, TT-M, or WL-C. All eligible patients had an identified care ally, but only those randomized to the CA-M arm ultimately participated in TOMCATT with their care ally. Those randomized to the CA-M arm attended a group training workshop that taught them a massage routine. After the training, CA-M dyads were asked to complete the learned 30-minute massage routine 3 times per week for 12 weeks and document the massages that were delivered in a log. The TT-M intervention consists of twice weekly, hour-long massage sessions provided by massage therapists trained to deliver semistandardized, individualized treatments.

TOMCATT outcome assessments were conducted via interviews at baseline and 1, 3, and 6 months in all study arms. The primary outcomes were neck pain and related disabilities. The secondary outcomes were other pain measures, health-related quality of life, depression, anxiety, and pain cognition.

TOMCATT was reviewed and approved by the Indiana University Institutional Review Board (#1604689005_9034) and Veterans Administration Research Review Committee (VA project ID:IIR 15-333) and registered with ClinicalTrials.gov (NCT03100539). All participants engaged in a one-on-one informed consent process with the study personnel and provided written informed consent.

Veterans are eligible if they meet criteria outlined in Textbox 1.

Eligible Veterans who provided written informed consent were enrolled and underwent a baseline assessment. Randomized care allies also provided written consent to participate at the beginning of the intervention workshop training. Care allies were eligible if their partnered Veteran was randomized to the care ally arm, they attended the CA-M training workshop with their Veteran partner and did not have medical concerns that might interfere with giving a massage.

Primary care providers at the Roudebush Veterans Affairs (VA) Medical Center in Indianapolis, Indiana, and surrounding community-based outpatient clinics were informed of the TOMCATT study details and asked to provide signed approval for the research team to contact their patients for possible study participation. Potential participants were identified by querying the VA’s electronic medical record to create a master list of Veterans meeting the following criteria: (1) neck pain diagnoses per International Classification of Diseases, Ninth Revision codes 721.0 to 723.9 and (2) primary care clinic visits in the past year (a proxy measure of VA care engagement).

A recruitment letter signed by their provider is mailed to qualifying Veterans to describe the study. Letters contain an initial screening for neck pain–related disability that interested Veterans may return if they would like to be contacted to assess their eligibility and possible participation. An appointment is scheduled for eligible Veterans to sign an informed consent statement and the Health Insurance Portability and Accountability Act authorization for those indicating a desire to participate. Baseline interviews and assessments are conducted by a research assistant (MK, EE, or BL) before randomization to minimize ascertainment bias. Veterans also self-identified as being interested in study participation after learning of TOMCATT through acquaintances, word of mouth, or study pamphlets.

Patients were initially assigned to one of the three study arms (TT-M, CA-M, and WL-C) using randomization lists created by the study statistician. Stratified block randomization, with random block sizes of 3 and 6, was used for the original plan to enroll 468 participants (excluding care allies enrolled within the CA-M arm). Sex (male or female) was the only randomization strata used.

Table 1 and Table 2 outline the data collection protocol for the Veteran and care ally participants, respectively. Veteran and care ally participants in the CA-M arm also completed a brief learning objectives survey following the training workshop.

Veteran participants were reimbursed US $25 per completed outcome assessment, with 4 scheduled assessments (baseline and 1, 3, and 6 months). Participants in the TT-M and WL-C arms were invited to receive all CA-M training and materials after their 6-month interview. Veterans randomized to WL-C are invited to receive a massage from one of the TOMCATT massage therapists. Care allies received a US $50 gift card at the completion of the care ally training and a complimentary massage session.

Participants in the WL-C arm received check-in calls at months 2 and 4 from the study staff and were administered outcome assessments on the same schedule (baseline and 1, 3, and 6 months) as the treatment groups. Participants in the WL-C were instructed to continue their medical care as normal and to not begin any massage treatment during the 6 months of the study.

The TOMCATT pretrial activities were launched in July 2016 and included massage therapist recruitment, training, and onboarding; the finalization of training materials; data collection documents and databases created; and research personnel hired and trained.

Participant recruitment began in May 2017. Figure 3 depicts the results of screening, eligibility determination, enrollment, randomization, and treatment initiation for the first 30 months of TOMCATT. Potential participants were identified through the electronic medical record as having CNP, of whom 7032 were sent a study invitation letter that included a screening NDI form for interested individuals to return to learn more and be further screened for enrollment eligibility. Various additional recruitment methods have been used, but their yield is unclear. The NDI screening forms yielded 33.2% (305/919) of the enrolled participants. In all, 14.8% (136/919) of the received NDI form scores did not meet the neck pain with disability thresholds required to participate, whereas another 26.1% (240/919) of potentially eligible participants returned NDI forms but could not be contacted.

Randomization resulted in 102, 102, and 101 participants in the CA-M, TT-M, and WL-C arms, respectively, in the first 30 months. All WL-C participants progressed to the 12-week nonintervention phase of the study, with 93.1% (94/101) completing an outcome assessment (data not shown). In all, 98% (100/102) of participants in the TT-M initiated the intervention. Of the 102 participants, only 2 (2%) failed to schedule and attend their first TT-M treatment and withdrew from the study owing to schedule conflicts.

For CA-M, only 54.9% (56/102) of the randomized patients attended training and initiated treatment. Disinterest was the primary reason (22/46, 48%) indicated for noninitiation, followed by scheduling issues (12/46, 26%). Among those who attended training, an average of 42 days passed between baseline data collection and training or treatment initiation. Participant engagement was further challenged by inconsistent treatment log-return compliance, despite most participants (52/56, 93%) completing one or more follow-up outcome assessments (data not shown). Nearly 45% (25/56) of participant dyads returned no compliance logs, whereas nearly one-third of the participants (17/56) returned logs for each of the 12 intervention weeks. A majority of CA-M dyads (31/56, 55%) returned at least one week of compliance logs.

Several mitigatory steps were taken to improve treatment initiation in the CA-M arm over approximately 14 months. To accommodate individual schedules, training workshops were scheduled 3 times per month, which included at least one Saturday, a morning training session during the week, and an afternoon training session during the week per month; any of which was held for as long as at least one dyad attended. Several additional training workshops were scheduled per individual dyad scheduling needs.

In an effort to retain participants in the CA-M arm and facilitate treatment initiation for those with care ally barriers, a matching approach was launched whereby former Veteran participants who expressed willingness were matched with enrollees who ultimately did not have a care ally willing or able to attend training and provide care ally–assisted intervention for 12 weeks. An enrollee with care ally participation barriers agreed to be matched but did not attend the scheduled training, despite the matched, stand-in care ally attending.

Although not initiated, a modified training approach was developed for CA-M arm participants as an alternative to the 3.5-hour training seminar. The modified approach was composed of a combination of at-home learning and supportive applied laboratory experiences. The developed approach included participants accessing digitally recorded didactic information and materials from a supportive DVD on their own time in a structured and prompted format. Once complete, the dyad would schedule a 1-hour hands-on application session with NM to reinforce content and ensure safe and appropriate application of the intervention techniques.

Figure 4 displays the flow diagram for recruitment, enrollment, and treatment initiation from November 2019 to March 2020, when the COVID-19 pause began. Recruitment efforts resulted in 50% of the recalculated needed enrollment achieved in just over 4 months. Of those enrolled, all but one TT-M participant initiated treatment (TT-M).

The TOMCATT trial was initially designed to test the effectiveness of 2 different massage interventions versus WL-C. Treatment engagement and retention challenges have emerged and raised questions about the feasibility of CA-M. Consequently, the study design was modified, resulting in a simpler, 2-arm study. This decision, although difficult, allowed the TOMCATT study team to focus and redirect recruitment efforts on the 2 remaining study arms. Promising recruitment and retention success for the postmodification efforts point to promising expectations for TOMCATT’s resumption efforts from the COVID-19 pause.

Caregiver-delivered massage has been shown to be feasible and effective in 2 previous Veteran-focused studies [10,11], but such an approach has not been studied for a chronic musculoskeletal pain population. The treatment engagement and retention challenges that disproportionately affected the CA-M arm were greater than expected, based on previously published studies [10,11]. It is speculated that these challenges have emerged for several reasons. First, participants randomized to the CA-M group may have been disappointed that they were not randomized to the TT-M group given the popularity of massage. It is well known that participants have strong preferences for treatment allocation within trials [34], and these preferences can impact follow-up rates, attrition, and treatment outcomes [35]. Second, dyadic research poses unique challenges in recruitment, retention, attrition, data collection, and analysis [36] relative to nondyadic research. Studies of chronic pain may heighten the challenges of dyadic research, because chronic pain can have a significant impact on a person’s social relationships by triggering symptoms of anxiety, depression, and anger [37]. Third, CA-M was an intensive intervention that involved attending a one-time, multihour workshop; a DVD for regular home use; a treatment session log to complete weekly; and delivery of 30-minute massage sessions by the care ally 3 times a week for 12 weeks. The substantial time investment and significant requirement for active dyadic participation may have been too great for some participants who never engaged in the intervention or withdrew from the study.

The qualitative interview data collected in a subset of TOMCATT participants are anticipated to elucidate some challenges that emerged within the original design and may provide potential insights into ways to improve treatment engagement, retention, and interpret eventual trial results. In these interviews, the participants are asked about their prerandomization arm preferences, treatment perceptions, and outcome expectations. Qualitative data collection and analysis may highlight a desire to incorporate treatment preferences into the design of future massage trials. Furthermore, assessing the sociodemographic and clinical correlates of treatment engagement and retention, especially among CA-M participants, are planned.

This manuscript reports only preliminary findings related to enrollment and intervention uptake to explain the modification rationale from a 3-arm to a 2-arm study design and to describe the modification methodology. Furthermore, the COVID-19 pandemic has caused significant disruptions to TOMCATT, especially the delivery of TT-M. All in-person noncritical research activities were suspended in March 2020. Thus, all in-person TOMCATT activities, including delivery of massages, were forced to halt owing to safety concerns and institutional mandates. Because of intermittent surges in COVID-19 infection rates throughout 2020 and early 2021, TOMCATT continued to suspend in-person research activities owing to safety concerns for participants, massage therapists, and study staff. As a result, recruitment and massage treatment delivery for TOMCATT did not restart until June 2021, after infection rates had declined and safety mitigation factors were in place and believed to be effective. Study recruitment and enrollment began gradually with study capacity reached and a smooth process implemented by December 2021.

Although the use of informal caregivers to provide massage is an innovative care delivery model, it was not feasible in our study of Veterans with CNP, which was hampered by low engagement and high attrition. TOMCATT’s continued efforts with the modified study design will provide a better understanding of the feasibility and effectiveness of an on-site therapist-applied massage treatment model for Veterans with CNP, an approach that demonstrates feasibility and acceptability. The incorporation of patient preferences into the study design is planned for future trials to improve engagement and retention in all care approaches.