A Sequential Multiple-Assignment Randomized Trial (SMART) for Stepped Care Management of Low Back Pain in the Military Health System: A Trial Protocol

Abstract

Background

The Defense Health Agency has prioritized system-level pain management initiatives within the Military Health System (MHS), with low back pain as one of the key focus areas. A stepped care model focused on nonpharmacologic treatment to promote self-management is recommended. Implementation of stepped care is complicated by lack of information on the most effective nonpharmacologic strategies and how to sequence and tailor the various available options. The Sequential Multiple-Assignment Randomization Trial for Low Back Pain (SMART LBP) is a multisite pragmatic trial using a SMART design to assess the effectiveness of nonpharmacologic treatments for chronic low back pain.

Design

This SMART trial has two treatment phases. Participants from three military treatment facilities are randomized to 6 weeks of phase I treatment, receiving either physical therapy (PT) or Army Medicine’s holistic Move2Health (M2H) program in a package specific to low back pain. Nonresponders to treatment in phase I are again randomized to phase II treatment of combined M2H + PT or mindfulness-based treatment using the Mindfulness-Oriented Recovery Enhancement (MORE) program. The primary outcome is the Patient-Reported Outcomes Measurement Information System pain interference computer-adapted test score.

Summary

This trial is part of an initiative funded by the National Institutes of Health, Veterans Affairs, and the Department of Defense to establish a national infrastructure for effective system-level management of chronic pain with a focus on nonpharmacologic treatments. The results of this study will provide important information on nonpharmacologic care for chronic LBP in the MHS embedded within a stepped care framework.

Background and Rationale

Chronic low back pain (LBP) is a common condition whose prevalence and adverse consequences are concerns in both civilian and military health systems [1]. Operational deployments, training demands, and other stresses of active-duty service present additional challenges for chronic LBP management in the Military Health System (MHS) [2–4]. Back pain is the leading cause of medical discharge across military services [5] and the most common reason for a medical visit, with over 1 million encounters in 2015 [6] at an economic cost exceeding $1 billion [7].

Improving chronic LBP management is a priority for many health systems, including the MHS. Guidelines support focusing on nonpharmacologic management to promote self-care; reducing reliance on medication, particularly opioid therapy; and implementing stepped care [8–11]. Stepped care is a strategy to provide a continuum of evidence-based treatment across stages of pain management, emphasizing an individualized, stepwise approach as patients’ symptoms increase in complexity or as patients fail to respond with less intensive interventions [12]. Stepped care models are attractive because they begin with consistent access to evidence-based, less intensive interventions. More costly and intensive treatments are used for those who do not respond to initial care. Stepped care has been beneficial for conditions with high prevalence and multiple evidence-based interventions along a continuum of invasiveness and costs [13, 14].

There are many evidence-based, nonpharmacologic LBP treatments, including exercise, manual therapy, psychological therapies, and mindfulness interventions, yet none produce consistently large effects or are clearly superior for all patients. There is a need for research focused on tailoring treatment to individual patient factors and sequencing treatments when initial care is insufficient. The purpose of this article is to describe the Sequential Multiple-Assignment Randomization Trial for Low Back Pain (SMART LBP) study, which is designed to address these research gaps and improve chronic LBP management in the MHS.

Methods

Study Rationale and Objective

This study focuses on the initial steps of a stepped care model. Step I is primary care based. Physical therapy (PT) is often used at this step to provide exercise, education, and manual therapy [7]. Step I care should address the complex biopsychosocial nature of chronic pain [15]. The Office of the Army Surgeon General has promoted the Move to Health (M2H) program as part of a strategy to bring a biopsychosocial perspective into health care and to empower patients in self-care [16]. Operationalizing M2H to the needs of patients with chronic LBP may be an effective step I strategy. It may be that a combination of PT and M2H or evidence-based complementary health intervention, including mindfulness, are beneficial for chronic LBP. These strategies require a greater time commitment from patients and may be better conceived as step II strategies for patients who do not respond to initial efforts. The objective of this study is to compare the effectiveness of step I (PT or M2H) and step II (PT + M2H or mindfulness) care and examine strategies to sequence and individualize care. Addressing these questions can improve chronic LBP management in the MHS and provide lessons to inform strategies in other settings.

Study Design

The SMART LBP study uses a sequential multiple-assignment randomization trial (SMART) design [17]. A SMART design permits evaluation of adaptive interventions with prespecified decision rules to tailor treatment strategies to individual patients [18]. The SMART LBP study design (Figure 1) randomizes patients to 6 weeks of phase I treatment with PT or M2H. The first reassessment occurs after 8 weeks. Phase II care is based on response to initial treatment. Phase I treatment responders receive up to two additional sessions of phase I care to facilitate transition to self-care. Phase I nonresponders are randomized to phase II treatment of combined M2H + PT or mindfulness using the Mindfulness-Oriented Recovery Enhancement (MORE) program. Phase II treatment is provided over 8 weeks.

Figure 1.

Intervention and assessment flow diagram for the SMART LBP study, a sequential multiple-assignment randomized trial. SMART LBP = Sequential Multiple-Assignment Randomization Trial for Low Back Pain; R = randomization.

The SMART LBP study is designed to examine pragmatic, scalable strategies in the MHS. The study was scored across the nine domains of the PRagmatic-Explanatory Continuum Indicator Summary-2 (PRECIS-2) [19]. Based on these domains, we consider the study to be mostly pragmatic (Figure 2). The primary aims are to compare the effectiveness of phase I (aim I) and phase II treatments among nonresponders (aim II). Secondary aims compare the effectiveness of the phase I treatments when followed by MORE or by combined PT + M2H for phase I nonresponders (aims IIIa and IIIb). We also compare the four embedded two-stage treatment strategies (aim IV) and examine the comparisons of aims I–IV within prespecified patient subgroups (aim V).

Figure 2.

PRagmatic-Explanatory Continuum Indicator Summary-2 domains and scoring wheel for the SMART LBP study. Scores of 1 to 5 on each spoke of the wheel indicate how pragmatic or explanatory the trial is, with 5 representing the most pragmatic and 1 representing the most explanatory. SMART LBP = Sequential Multiple-Assignment Randomization Trial for Low Back Pain. Adapted by permission from BMJ Publishing Group Limited. [The PRECIS-2 tool: designing trials that are fit for purpose, Loudon K, Treweek S, Sullivan F, Donnan P, Thorpe KE, Zwarenstein M. BMJ 2015;350:h2147]

Study Population

The SMART LBP study is designed to recruit a representative cohort of active-duty service members and other TRICARE-eligible beneficiaries seeking primary care for chronic LBP in an MHS facility. Specific eligibility criteria are outlined in Table 1.

Table 1.

Eligibility criteria for SMART LBP trial

Inclusion Criteria	Exclusion Criteria
1. Active-duty military or member of the Reserves or National Guard on active duty, retired service members, family members, or other TRICARE-eligible beneficiaries receiving care in a participating MHS facility	1. Signs of serious or systemic pathology as a cause of LBP, including spine fracture, neoplasm, inflammatory disease (e.g., ankylosing spondylitis), vertebral osteomyelitis, and so on
2. Age, 18–65 years	2. Knowingly pregnant
3. Seen by an MHS health care provider for a chief complaint of LBP with or without symptoms in buttocks or legs in past 30 days   a. Chief complaint of LBP may be self-reported or based on ICD-10 code (M54.5, M54.9, S33.012, M51.36, M51.37, M48.06, M47.817, M54.16, M54.17, M51.26, M51.27, M54.3)	3. Interventions for LBP with providers other than primary care in the past 6 months (includes PT, pain management, counseling, specialist consultation, and so on) or any interventional pain procedure or interdisciplinary pain management in the past 6 months
	4. Any lumbar spine surgery in the past year
4. Meets NIH Task Force definition of chronic LBP [20]	5. Pending a Medical Evaluation Board, pending a discharge from the military for medical reasons, or pending or undergoing litigation for an injury
	6. Elevated acute risk for suicide based on VA or DOD guidelines [21]
	7. Planned absence of 2 weeks or more for training, vacation, or any purpose during the next 16 weeks

SMART = Sequential Multiple-Assignment Randomization Trial; LBP = low back pain; MHS = Military Health System; ICD = International Classification of Diseases; PT = physical therapy; NIH = National Institutes of Health; VA = Veterans Affairs; DOD = Department of Defense.

Screening, Recruitment, and Randomization

Recruitment will occur using complementary strategies to offer participation to a representative cohort. Primary care providers and staff may alert research staff directly about a patient seen for chronic LBP. Also, a Health Insurance Portability and Accountability Act waiver will allow approved research staff to review appointment lists in the electronic health record and contact scheduled patients with LBP to notify them about the study. If the patient is interested in more information, a research staff member will meet with the patient before or after their provider appointment to explain the study and initiate informed consent and eligibility screening. The status of all individuals with whom participation is discussed will be tracked in a screening log. Reasons for ineligibility or declining participation will be recorded in screening logs maintained in the study’s Research Electronic Data Capture (REDCap) data collection site [22].

Randomization will occur using the REDCap randomization module. A study statistician created separate randomization allocation tables for the initial randomization and for rerandomization of nonresponders. Blocked randomization with block sizes of four or six will be used. Randomization is stratified by recruitment site, gender, and active-duty status. The allocation tables were prepared using SAS software. Initial randomization will occur after consent has been given and all baseline procedures have been completed.

Participating Sites

Recruitment sites are primary care clinics at Brooke Army Medical Center and Wilford Hall Ambulatory Surgical Center at the San Antonio Military Medical Center in San Antonio, TX, and at Madigan Army Medical Center in Tacoma, WA, respectively.

Interventions

Study interventions are provided in two phases. Phase I randomly assigns patients to either PT or M2H. Phase I treatment begins within 7 days of randomization and last for 6 weeks. At the 8-week assessment, patients are determined to be treatment responders or nonresponders. Patients determined to be nonresponders are rerandomized to a phase II treatment of either MORE or combined PT + M2H. Phase II treatment lasts for 8 weeks. Consistent with a pragmatic trial, providers are allowed flexibility in determining the precise dosage and administration of all interventions. Parameters are provided within which pragmatic intervention decisions can be made by providers based on a patient’s unique needs. These parameters are described for each treatment in the following sections.

Physical Therapy

Patients randomized to PT are referred to an MHS PT clinic. The physical therapist must be licensed and have received training in study procedures. PT is focused on evidence-based treatments of education in a biopsychosocial pain model, exercise, and manual therapy [10, 23]. The number of sessions is at the discretion of the physical therapist, up to a maximum of two per week. Physical therapists are trained to guide the dosage and type of activities using evidence-based risk stratification with the STarT Back Screening Tool (SBST) [24]. The SBST categorizes patients based on risk for persistent disability (low, medium, or high), with treatment tailored to each category [25]. Low-risk patients are recommended to receive education with follow-up after 2 weeks to determine the need for additional sessions. Medium-risk patients receive education along with exercise and manual therapy to address physical factors. High-risk patients receive education, exercise, and manual therapy supplemented by principles of cognitive behavior therapy to address psychological risk factors in this category [26].

Move to Health

The MHS has been undergoing a transition from a health care system focused on disease detection, prescriptions, and procedures to a “system for health” focused on prevention and well-being through empowering patients to engage in healthier lifestyle decisions [27]. The Office of the Army Surgeon General has developed several strategies to facilitate this transition, including Move to Health, a program that is modeled on the Whole Health program in the Veterans Health Administration. The program embraces a person-centered, holistic approach emphasizing self-care and using conventional medicine and integrative approaches [28].

The M2H program for chronic LBP is provided by a health coach trained in study procedures. Health coaches delivering M2H have a health-related degree (e.g., physical therapist, exercise physiology, and so on). Patients meet with a health coach within 7 days of randomization. The initial session identifies a personal health domain that the patient wants to address to improve their well-being. Domains include sleep, physical activity, nutrition, intrinsic well-being (e.g., tobacco cessation, emotional and spiritual development), and extrinsic well-being (e.g., family or social relationships). Domain selection occurs through a shared decision-making process facilitated by a review of the patient’s health history and completion of a personal health inventory. Once a domain is selected, the coach helps the patient set a specific and measurable goal. Additional assessments may be used to determine the most appropriate strategies to assist patients in meeting their goal. An inventory of domain-specific resources is available to the health coach, including educational materials, app-based strategies, and referrals for follow-up services (e.g., sleep specialists, psychological counselor, and so on). Patients receive written instructions for a self-paced walking program, tips for sleeping with chronic LBP, and deep breathing exercises for relaxation. The number of M2H sessions is at the discretion of the health coach and patient, up to a maximum of twice weekly. Sessions may be delivered in person or with technology, including phone, video conference, and so on. Follow-up sessions are used to review progress, update goals, identify new or additional domains, and assist with setbacks and challenges.

Patients randomized to combined PT + M2H in phase II will continue their assigned phase I treatment (M2H or PT) and begin the additional component. The provider of the patient’s phase I intervention will be informed that the other component is being added to treatment.

Mindfulness Oriented Treatment

The mindfulness program used in this study is Minfulness-Oriented Recovery Enhancement (MORE). The MORE program was designed to address symptoms and underlying mechanisms of chronic pain [29, 30] using eight individual sessions with a behavioral health provider trained for the study. Behavioral health providers may be psychologists, social workers, or other licensed behavioral health providers. The core areas of MORE sessions are outlined in the following sections. At-home exercises are provided to reinforce the core areas.

Mindfulness

Patients are guided in within-session mindfulness practice to 1) become aware of when their attention is engaged by pain and aversive thoughts and feelings; 2) acknowledge and accept that attentional engagement has occurred; and 3) disengage attention from pain and aversive experience and shift attention to neutral or health-promoting stimuli via the practice of mindful breathing.

Cognitive Reappraisal and Savoring Positive Experiences

Patients receive training to use mindfulness to become aware of, decenter from, and challenge automatic thoughts while becoming open to new, more adaptive appraisals such as savoring positive experiences (e.g., an enjoyable meal, a beautiful sunset, and so on).

Treatment Adherence and Fidelity

Patients’ treatment adherence will be evaluated based on attendance at scheduled sessions. Provider fidelity to core treatment components will becaptured using checklists that are built into the electronic medical record documentation template for each session; these can be easily reviewed by research staff (who are embedded in the PT and primary care clinics). Checklists will also record off-protocol interventions. Consistent with a pragmatic study, neither patients nor providers will be removed for nonadherence or poor fidelity.

Provider Training

Investigators will train the personnel providing the study interventions. Providers are credentialed in participating MHS sites (physical and behavioral health therapists). Manuals have been developed for each intervention to facilitate training. Providers in the PT group receive a 1-day training on evidence-based LBP care, risk stratification–based decision-making, and a review of biopsychosocial education. Health coaches receive training in M2H procedures and algorithms and a minimum of 4 hours of training in motivational interviewing to help patients identify priority health domains and set goals. Behavioral health providers receive a 1-day training in MORE techniques and core components.

Schedule of Assessments

Assessments are conducted at baseline (before randomization) and at weeks 8 (before randomization of nonresponders), 18, 26, and 52 after enrollment (Table 2).

Table 2.

Schedule of assessments for SMART LBP trial

Measure	Screening	Baseline	8 Weeks	18 Weeks	26 Weeks	52 Weeks
Informed consent form	X
Eligibility criteria	X
Demographics		X
Medical history		X
Randomization		X	X*
STarT Back Screening Tool		X
PROMIS pain interference CAT		X	X	X	X	X
PROMIS physical function CAT		X	X	X	X	X
PROMIS sleep disturbance CAT		X	X	X	X	X
PROMIS depression CAT		X	X	X	X	X
PROMIS anxiety CAT		X	X	X	X	X
Patient Acceptable Symptom State		X	X	X	X	X
PEG-3		X	X	X	X	X
EQ-5D		X	X	X	X	X
Perceived readiness for duty†			X	X	X	X
Defense and Veterans Pain Rating Scale		X	X	X	X	X
Treatment side effects			X	X
Long-term opioid use						X
Limited-duty days†						X
Direct medical expenses						X

SMART = Sequential Multiple-Assignment Randomization Trial; LBP = low back pain; PROMIS = Patient-Reported Outcomes Measurement Information System; CAT = computer-adapted test; PEG-3 = Pain Intensity, Enjoyment of Life, and General Activity Three-Item Scale; EQ-5D = European Quality of Life 5-Dimension Instrument.

^*Only for nonresponders to phase I treatment.

^†Only for active-duty service members.

Outcomes

The primary outcome is the Patient-Reported Outcomes Measurement Information System (PROMIS) pain interference computer-adapted test (PI-CAT). The PI-CAT uses a 44-item bank to assess the consequences of pain on relevant aspects of life, including social, emotional, cognitive, physical, and recreational activities [31]. The PROMIS physical function CAT (PF-CAT) is the main secondary outcome. The PF-CAT uses a 121-item bank to assess current self-reported capabilities related to physical activities [32]. All PROMIS scores are reported on a T score metric, with a score of 50 aligning with the general population mean and a standard deviation of 10. Higher scores indicate more of the quantity being assessed.

Additional secondary outcomes include PROMIS CAT assessments of health domains relevant to chronic LBP, including sleep disturbance, depression, and anxiety. The Patient Acceptable Symptom State (PASS) evaluates the acceptability of the current symptom status [33]. The European Quality of Life 5-Dimension Instrument (EQ-5D) is a generic quality of life instrument covering five domains: mobility, self-care, usual activities, pain or discomfort, and anxiety or depression, combined to produce a health state ranging from 0 (death) to 1 (perfect health) [34]. The Defense and Veterans Pain Rating Scale (DVPRS) includes a numeric pain intensity rating and four pain interference questions developed with veterans and military members [35]. The Pain Intensity, Enjoyment of Life, and General Activity Three-Item Scale (PEG-3) evaluates pain severity and interference [36]. Perceived readiness for duty is evaluated for active-duty service members, with three questions assessing self-perceived ability to perform military-related tasks [37].

We will collect information about treatment side effects by asking patients about possible physical and psychological effects they believe relate to study interventions, and extent to which they were impacted (from “not at all” to “extremely”). The questionnaire was developed from assessments used in studies of physical and psychological interventions for chronic pain [38, 39].

Direct medical expenses will be extracted from the Military Health System Data Repository (MDR). Total and LBP-related costs over the 1-year follow-up will be examined. The MDR contains TRICARE claims for inpatient and outpatient encounters provided in military or civilian facilities worldwide. Opioid use will be determined from the pharmacy data transaction service file of the MDR. Unique number of prescriptions and total days’ supply will be calculated for each patient over the 1-year follow-up. We will also calculate daily opioid dose at each follow-up point based on dose of opioids at that time converted to a morphine milligram equivalent dosage. Long-term opioid use will be defined as ≥90 days of continuous opioid use in the preceding 180-day period.

Determining Responder Status

At the 8-week assessment, the PI-CAT score will be compared with the baseline score to determine if the patient is a “responder” or “nonresponder” to phase I treatment. An improvement from a baseline of ≥7 T score points is required to be considered a responder [40, 41]. All other patients will be considered nonresponders.

Sample Size Determination

Power calculations are based on an estimated 80% retention across the follow-up period and a planned enrollment of 1,200 participants. Power calculations are based on the guidelines for statistical power for SMART designs [42]. Assumptions include a SD of 8 for the PI-CAT in LBP patients [40], a minimal clinically important difference (MCID) of 2.8 [43, 44], a responder rate to phase I treatment of between 30% and 45% [45], serial correlation r=0.40 for repeated assessments, and a two-sided α=0.05 for the primary outcome. We consider intraclass correlations (ICCs) from 0 to 0.05 to account for clustering of outcomes within MORE therapists because of the small number of therapists delivering this treatment. Clustering by provider for PT and M2H was not incorporated into the power analyses due to the likelihood of patients moving between therapists in these arms.

Under these varying assumptions and with a sample size of 1,200, the detectable effect for the PI-CAT for aim I is 1.54 with 90% power and 1.33 with 80% power. Power to detect an MCID is essentially 1.0. For aim II, the detectable effects for the PI-CAT range from 1.59 to 2.37, and the power to detect an MCID value ranges from 0.97 to 0.99 (see Supplementary Data for full power analysis). The rationale for selecting a sample size larger than required to detect an MCID for the two primary aims s to support analyses of secondary objectives.

Statistical Methods

The intention-to-treat principle will be used to guide the data analyses, with patients evaluated based on randomized assignment regardless of compliance. Skewed outcomes may be transformed to approximate normality. Analyses for the project aims are outlined in the following sections.

Aim I

This co-primary aim compares the effectiveness of phase I treatments (M2H or PT) with the primary outcome after 8 weeks. Aim I analysis fits a longitudinal linear mixed model for the baseline and 8-week PI-CAT using an unstructured covariance matrix to account for serial correlation between measures and imposing the constraint that baseline PI-CAT means are equal in treatment groups. By using an unstructured covariance matrix, the model constitutes a special case of a general linear mixed model that avoids imposing specific assumptions concerning the distribution of random effects. Restricted maximum likelihood estimation will be used for the estimation of the mean adjusted difference in the 8-week PI-CAT between groups and to provide confidence intervals and hypothesis tests. In this analysis, as well as in other longitudinal analyses described in the following sections, the randomization stratification factors will be included as covariates.

Aim II

This co-primary aim compares the effectiveness of phase II treatments (MORE or M2H + PT) in phase I nonresponders by fitting a longitudinal linear mixed model to relate the randomized phase II treatments to the PI-CAT at 18, 26, and 52 weeks, with the 8-week score included as a covariate to account for the effects of phase I treatment. The model will include random effects for MORE therapists for patients randomized to this treatment. The primary treatment contrast will compare the 52-week scores.

Aims IIIa and IIIb

These aims compare two adaptive treatment regimens that extend across both phases of the SMART design. Aim IIIa compares M2H and PT when followed by mindfulness in nonresponders. Aim IIIb compares phase I treatments when followed by M2H + PT for nonresponders. The analyses willapply weighted estimating equations with robust standard errors designed for analysis of longitudinal outcomes in SMART trials [45] to compare the PI-CAT scores at weeks 8 and 18 and at weeks 26 and 52 with covariate adjustment for the baseline PI-CAT level. The 1-year assessment will serve as the primary comparison. Inverse probability of treatment weighting will account for nonresponders being rerandomized and thus split into two groups and underrepresented relative to responders. The phase I treatment in an optimal two-stage sequence may differ from the phase I treatment optimizing outcome at the end of phase I (aim I). For example, M2H may be superior to PT following phase I, but a sequence of PT followed by MORE may be the preferred two-stage sequence.

Aim IV

This secondary aim compares four embedded two-stage strategies: 1) PT with M2H + PT for nonresponders; 2) PT with MORE for nonresponders; 3) M2H with M2H + PT for nonresponders; and 4) M2H with MORE for nonresponders. The analyses will apply weighted estimating equations with robust standard errors appropriate for analysis of longitudinal outcomes in SMART designs while accounting for clustering by MORE therapists to estimate mean PI-CAT scores at weeks 8, 18, 26, and 52 [46]. The data set is augmented by including each patient either once or twice based on whether the patient’s treatment is consistent with one or two 2-stage strategies [47] (e.g., treatment of a patient who responds to PT in phase I is consistent with both regimens a and b). Inverse probability weighting will be used to account for nonresponders being rerandomized and thus split into two groups and underrepresented relative to responders.

Aim V

This secondary aim evaluates aims I–IV within subgroups of patients defined by key baseline characteristics (gender, age [≥50 years or <50 years at baseline], sleep disturbance [≥50 or <50 on the PROMIS sleep disturbance CAT at baseline], and baseline opioid use [yes or no; regular opioid use for at least 1 month]). For aims I–III, subgroup analyses will be performed by repeating longitudinal analyses within each subgroup and by extending statistical models by adding interactions between treatment and the subgroup factors. Subgroup analyses for aim IV will use Q-learning methods to investigate the more complex treatment strategies to estimate an optimal intervention in which phase I and II treatments are tailored to patient subgroup factors [48].

Procedures for Handling Missing Data

Multiple imputation will be used to ensure that statistical inferences remain valid in the presence of missing data if missingness follows a missing at random mechanism [49]. Multiple imputation will be carried out using time-ordered nested conditional imputation models to sequentially impute missing responses after phase I and subsequent treatment assignment and latter outcome variables [50]. Imputation models will include all variables in the analysis and selected additional variables considered likely to predict the outcome and likelihood of missingness. The sensitivity of the results to a possible missing not at random mechanism will be investigated within the multiple imputation approach by applying an offset term for the mean of the missing outcome variable among those patients for whom this outcome is missing and by evaluating the dependence of the primary outcome results on varying the size of this offset [51].

Implementation and Dissemination Procedures

Guidance for future implementation of effective strategies will be supported from the information collected by the Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) framework [52]. Enrollment and screening data will be used to evaluate the reach of the interventions to patients with chronic LBP in the MHS. Effectiveness will be examined through the aims outlined previously. Adoption and implementation by providers will be evaluated through fidelity checklists from treatment sessions. Finally, maintenance will be evaluated at the patient level through long-term outcome assessments. The study team has engaged leadership at participating MHS sites for implementation of the study interventions and will continue working with leadership for dissemination of the study findings.

Discussion

Stepped care models for chronic LBP are built on a premise of broad availability of less intensive and less costly interventions, and initial management strategies whose goal is facilitation of self-management [53]. The availability of numerous options for early treatment in stepped care for chronic LBP highlights the need for research examining the comparative effectiveness of different strategies, with attention to individualizing treatment and examining the outcomes of sequences of interventions. Traditional parallel-group, fixed intervention clinical trials are ill-suited to address these research questions. The SMART design employed in this study permits the evaluation of adaptive treatments that are modified based on a patient’s response in an ongoing manner, which can better reflect management of chronic conditions [54].

Our trial focuses on chronic LBP, a priority condition for system-level improvement in military, veteran, and civilian health systems [55, 56]. Improving chronic pain management with pragmatic solutions is the mission of the Pain Management Collaboratory supported by the National Institutes of Health, the Department of Defense, and the Veterans Health Administration [57]. There are unique challenges to conducting pragmatic trials in the MHS. Among these is the continuously changing operational environment of both beneficiaries (patients) and providers, as service members change duty stations every few years. Continuous changes in personnel highlight the critical need to develop scalable strategies and engage a variety of stakeholders to ensure consistent implementation across settings. This also highlights the importance of properly assessing treatment fidelity in order to understand how treatment delivery is affected by the factors in this unique setting. The SMART LBP study will contribute important information toward the overall goal of the Pain Management Collaboratory to improve chronic pain care for veterans and military personnel. Attention to implementation and engagement will also provide information that may be generalized beyond the MHS to inform care in other health systems.

Supplementary Data

Supplementary Data may be found online at http://painmedicine.oxfordjournals.org.

 

Funding sources: This research is supported by the National Institutes of Health (NIH) through cooperative agreement U24AT009769 from the National Center for Complementary and Integrative Health and cooperative agreement UG3AT009763/UH3AT009763 from the National Center for Complementary and Integrative Health at the NIH. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Conflicts of interest: The authors have no conflicts of interest to report.

Supplement sponsorship: This article appears as part of the supplement entitled “NIH-DOD-VA Pain Management Collaboratory (PMC)”. This supplement was made possible by Grant Number U24 AT009769 from the National Center for Complementary and Integrative Health (NCCIH), and the Office of Behavioral and Social Sciences Research (OBSSR). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NCCIH, OBSSR, and the National Institutes of Health.

Disclaimer: The views expressed herein are those of the authors and do not reflect the official policy or position of Brooke Army Medical Center, the US Army Medical Department, the US Army Office of the Surgeon General, the Department of the Army, the Defense Health Agency, the Department of Defense, or the US government. This article is a product of the National Institutes of Health–Department of Defense–Veterans Affairs Pain Management Collaboratory. For more information about the Collaboratory, visit painmanagementcollaboratory.org.