Tele-collaborative outreach to rural patients with chronic pain: pragmatic effectiveness trial protocol for the CORPs study

Abstract

Background

Despite the increased availability of evidence-based treatments for chronic pain, many patients in rural areas experience poor access to services. Patients receiving care through the VA may also need to navigate multiple systems of care.

Objective

To examine the effectiveness of a remotely delivered collaborative care intervention for improving pain interference among veterans with high-impact chronic pain living in rural areas.

Design

We will conduct a four-site pragmatic effectiveness trial of remotely delivered collaborative care for high-impact chronic pain. Participants (n = 608) will be randomized to the Tele-Collaborative Outreach to Rural Patients (CORPs) intervention or to minimally enhanced usual care (MEUC). Participants randomized to CORPs will complete a biopsychosocial assessment and five follow-up sessions with a nurse care manager (NCM), who will collaborate with a consulting clinician to provide personalized recommendations and care management. CORP participants will also be invited to a virtual 6-session pain education group class. Participants randomized to MEUC will receive a one-time education session with the NCM to review available pain services. All participants will complete quarterly research assessments for one year. The primary study outcome is pain interference. This trial will oversample veterans of female birth sex and minoritized race or ethnicity to test heterogeneity of treatment effects across these patient characteristics. We will conduct an implementation process evaluation and incremental cost-effectiveness analysis.

Discussion

This pragmatic trial will test the real-world effectiveness of a remotely delivered collaborative care intervention for chronic pain. Study findings will inform future implementation efforts to support potential uptake of the intervention.

Introduction

In 2009, the Veterans Health Administration (VA) began implementation of a stepped-care model for chronic pain management, which builds on a foundation of patient education for pain self-management approaches. Many patients adequately manage pain using self-management approaches; however, patients with high-impact chronic pain often require more intensive care. Within the VA, stepped-up treatment engages primary and specialty care, including physical therapy, pharmacy, complementary and integrative health (CIH) approaches, mental health, substance use, and other services. When indicated, treatment may be elevated to interdisciplinary pain teams or tertiary pain centers.¹

Collaborative care approaches to pain management have demonstrated effectiveness and align with the VA stepped model of pain care. Collaborative care is a team-based approach to the management of chronic illnesses, often involving nurse care managers (NCMs) and consulting clinicians collaborating to improve management of targeted conditions such as depression, anxiety, chronic pain, and their comorbidities.² Several randomized trials have demonstrated the efficacy of collaborative care for chronic pain,^3,4 with subsequent studies extending the intervention to predominantly telephone-delivered care management.^5,6

Additional strategies are needed to address pain care for veterans who reside in rural areas. Recent decades have seen the decentralization of VA primary care from large, urban VA medical centers to VA community-based clinics, with the goal of providing care closer to where veterans reside.⁷ Veterans may also receive VA-paid care from a non-VA community provider,⁸ or access care from community providers through other means, such as Medicare, Medicaid, private insurance, or out-of-pocket payments. Although veterans in rural areas may have access to care through multiple systems, there are limitations in care coordination between providers and across healthcare systems.⁹ Team-based collaborative pain care has demonstrated efficacy in addressing pain outcomes in several clinical trials, but has not been systematically tested in veterans located in rural areas, who comprise nearly one-third of all veterans seeking care in the VA.

This clinical trial will test the tele-Collaborative Outreach to Rural Patients with chronic pain (CORPs) intervention, which aims to improve pain and quality of life outcomes by assisting rural-residing veterans with navigation to treatment services and coordination of care across VA and community health systems. To inform potential future implementation efforts, this trial includes an implementation process evaluation and incremental cost-effectiveness analysis (CEA). This study is part of the Pain Management Collaboratory (PMC), a coordinating center and consortium of pragmatic effectiveness trials funded by the National Institutes of Health, Department of Defense, and Department of Veterans Affairs.¹⁰

Methods

Study objective

The goal of this multi-site pragmatic effectiveness trial is to examine the utility of a remotely delivered collaborative care intervention for high-impact chronic pain among veterans who live in rural areas. Aim 1 will test the hypothesis that participants who receive the CORPs intervention, relative to minimally enhanced usual care (MEUC), will experience improvements in pain-related interference. Secondary outcomes will include measures of pain intensity, functioning, quality of life, mental health symptom severity and suicidal ideation, sleep, and utilization of nonpharmacologic pain treatments. This study will oversample veterans of female birth sex and minoritized race and ethnicity. Aim 2 hypothesizes that the observed treatment effects will be evident across birth sex, race, and ethnicity. Aim 3 will involve an implementation process evaluation¹¹ and incremental CEA¹² to inform the development of an implementation toolkit to support a future rollout of CORPs across VA sites nationally.

Study population

The study population comprises veterans with high-impact chronic pain who reside in rural areas. High-impact chronic pain is defined as moderate to severe pain intensity on most days or every day in the past three months, with at least one major activity restriction (eg, being unable to work outside the home, go to school, or complete household chores).¹³

The VA uses Rural Urban Commuting Area (RUCA) codes to define rurality.¹⁴ Each US zip code is assigned a RUCA number from 1 to 10, with the VA allocating urban, rural, and highly rural status to each value—Urban: RUCA = 1; Rural: RUCA = 2-9; Highly Rural: RUCA = 10. For the purposes of this study, a RUCA of 2-10 is used to define rurality.

Eligible participants will be a US veteran, age 18 years or older, English speaking, with reliable phone access, confirmed rural residence based on RUCA, who meet the Centers for Disease Control and Prevention’s definition of high-impact chronic pain.¹³ Study exclusion criteria include: Cognitive impairment that would preclude participation in study activities, plans to move outside the VA catchment area in the next 3 months, residing in long-term or hospice care, surgery in the past 3 months, terminal illness, and current enrollment in another pain-related clinical trial. Consistent with a pragmatic trial design, we will not exclude patients with comorbid mental health or substance use disorder diagnoses.

Participating sites

Participants (N = 608) will be recruited from the VA Portland Health Care System (Coordinating Center), Portland, Oregon; VA Minneapolis Health Care System, Minneapolis, Minnesota; VA North Texas Health Care System, Dallas, Texas; and VA Tennessee Valley Health Care System, Nashville, Tennessee. Although these VA health centers are in urban areas, each system has affiliated Community-Based Outpatient Clinics (CBOCs) that provide primary care to patients residing in rural areas.

Recruitment, screening, and randomization procedures

With Institutional Review Board (IRB) approval, this study will use centralized and local site-based recruitment strategies. The centralized recruitment approach will leverage VA electronic health record (EHR) data to identify patients at participating sites who meet preliminary eligibility criteria, defined as rural residence based on RUCA codes, a pain-related diagnosis (musculoskeletal, neuropathic, or headache pain), and EHR-derived pain numeric rating scores indicative of moderate-to-severe pain intensity.¹⁵ We will oversample veterans of female birth sex and/or minoritized race and ethnicity (50% female and 50% minoritized race/ethnicity). Veterans who pass prescreening will be mailed a recruitment letter and opt-out form. Individuals who do not contact the research team within 10 business days will be phoned to ascertain interest.

Local site-based recruitment will rely on VA clinicians distributing recruitment letters to potentially eligible patients. Veterans who learn about the study during clinical encounters, or by other means, may self-refer to undergo screening assessment by the research team.

Regardless of the recruitment mechanism, upon contact with research staff, individuals will be provided with a study overview and phone eligibility screening will be completed if the patient is interested in consenting to participate. Eligible veterans will complete informed consent by phone using paper forms that are mailed, or electronically using a VA-approved service. During the consent call, research staff will verbalize the information from the consent materials and ensure comprehension. Veterans who wish to enroll will sign the consent documents, which will be reviewed for accuracy prior to scheduling the baseline enrollment visit.

The baseline enrollment assessment will be completed by phone and compensated with $50. At the end of the baseline assessment, participants will be randomized to CORPs or MEUC. The study will utilize permuted block randomization with blocks of varying length to randomize patients 1:1 to intervention arm. Randomization will be conducted using REDCap; study coordinators who perform randomization will be masked to the randomization scheme.

Interventions

CORPs intervention

The CORPs intervention includes an initial biopsychosocial pain assessment with a nurse care manager (NCM). Participants will collaboratively engage in treatment planning with the NCM and receive five scheduled follow-up visits (2-, 4-, 8-, 12-, and 16-week postintake). Additional appointments may be scheduled as clinically indicated throughout the study. Follow-up appointments will serve to reinforce the treatment plan and address barriers to treatment engagement. All sessions with the NCM will be conducted by phone or via VA Video Connect (VVC), the VA’s internal videoconferencing system, based on participant preference.

All veterans randomized to the CORPs intervention will be encouraged to participate in a 6-session pain education group that is held virtually and led by the NCM. Topics include: (1) neurobiology of pain, (2) psychosocial approaches to pain management, (3) nutrition and physical activity, (4) CIH pain treatments, (5) sleep, and (6) managing negative emotions. The primary purpose of the pain education group is to provide an overview of approaches included in a multimodal pain management strategy. Participants who desire treatment in a particular domain may be referred to the appropriate service during their follow-up appointments. Patients who are unable to attend one or more group education sessions may receive the session content during individual follow-up encounters with the NCM.

Consistent with collaborative pain care models, the NCM is supported with consultation from a clinician who, with the NCM, establishes pain treatment recommendations for the primary care team.² The NCM functions as a patient navigator and care coordinator, communicating decision support recommendations. The NCM will engage participants in a review of nonpharmacologic treatment options for chronic pain, refer to services (in the VA or community as clinically indicated and available), and provide care management to ensure engagement with treatment. The NCM also communicates recommendations to the primary care team concerning analgesic pharmacotherapy options. The NCM and consulting clinician discuss patient treatment plan development and progress weekly. See Figure 1 for a schematic of the CORPs intervention.

MEUC comparator

Consistent with pragmatic trial methodology, this study utilizes usual care for the comparator arm. For this study, the MEUC comparator involves a one-time educational session with the NCM. The educational session will take place by phone or VVC and may include a brief review of the veteran’s prior pain treatments, along with some general recommendations about pain-related services and activities in their area or available online. The NCM may provide collated resources created by the VA and/or the study team via mail or through the VA’s secure messaging system. Ultimately, the veteran must continue to independently navigate and coordinate receipt of services they are interested in, including between providers and the healthcare systems.

Assessment procedures

Participants will complete interviewer-administered surveys by telephone at baseline, 3-, 6-, 9-, and 12-month timepoints; participants will be compensated $50 for completing each of the five research assessments. Follow-up assessments will be administered by masked assessors and will last 45-60 minutes. Additional clinical data will be collected from the EHR through administrative data extraction.

For both the CORPs and MEUC arms, fidelity will be measured through EHR manual chart abstraction. Consistent with pragmatic trial design, fidelity ratings will not be used to alter the course of care. Rather, fidelity will be determined after study completion.

Primary and secondary outcomes

The primary study outcome is a change in Brief Pain Inventory (BPI)¹⁶ pain interference. Secondary outcomes include pain intensity, functioning, quality of life, mental health symptom severity and suicidal ideation, sleep, and utilization of nonpharmacologic pain treatments. See Table 1 for a list of study outcomes and measures.

Other measures

Demographic data to be collected include age, gender, birth sex, race, ethnicity, marital status, employment status, and disability/VA service connection.

The study cost-effectiveness analyses will require additional measures. The clinical outcome used in the CEA is quality-adjusted life years (QALYs). We will calculate QALYs using data from the Veterans Rand 12-item Health Screen (VR-12)¹⁹ using a previously developed algorithm to convert the VR-12 responses to QALYS.^30,31 In addition, pain intensity and pain interference from the Brief Pain Inventory¹⁶ will be used for secondary cost-effectiveness analyses in Aim 3, to calculate pain-free days. The CEA will also include the cost of usual healthcare services used by study participants. We will use the Participant Health Care Utilization Survey (PHCUS)²⁹ to measure healthcare utilization received from outside the VA, in addition to self-reported engagement in nonpharmacological interventions for chronic pain²⁵ and EHR-derived clinical encounters to quantify healthcare utilization received within the VA. The healthcare utilization data will be converted into costs using published unit costs (eg, VA schedules for the cost of an emergency department visit).

Analytic methods

The primary study outcome is a change in pain interference, assessed with the BPI. An intent-to-treat random effect regression model will compare changes in the pain interference outcome variable measured continuously over the 12 months of patients’ participation, comparing patients who receive CORPs vs MEUC as a fixed effect. The study will account for the fixed effect of clustering across sites. A between-effect of treatment (CORPs vs MEUC) will be estimated to test the primary hypothesis. We will also test a within-effect of time and an interaction between treatment and time, and if significant, will include these effects in the final model selection. Random effects will be specified for participants at baseline and over time to model the change over time in outcome variables using the identity link, given their approximately continuous distribution. Analyses will include prespecified, time-invariant covariates of age, gender, race, and ethnicity collected at baseline. Analyses will adjust for baseline levels of alcohol use, other illicit substance use, prescribed opioid use, and prescribed nonopioid analgesic pharmacotherapy use. This covariate-adjusted analytic approach will use data from all participants.

The analysis of the secondary outcomes will follow an almost identical pattern. The primary difference will be the link function used to analyze any noncontinuous outcomes such as count dependent variables. The secondary outcomes that evaluate nonpharmacologic pain treatment utilization and suicide behavior follow count distributions and will be evaluated using a series of tools to help to determine which type of count model (ie, Poisson, negative binomial, zero-inflated Poisson, zero-inflated negative binomial) provides the best fit to the data.³²

For Aim 2, to assess the moderating role of birth sex (female vs. male) and race/ethnicity (White non-Hispanic vs. minoritized race or ethnicity) on primary and secondary outcomes, a series of analyses comparing effect sizes between subgroups of rural veterans who receive the CORPs intervention will be conducted. We acknowledge that dichotomizing race/ethnicity does not account for the heterogeneity within this construct. However, we consider minoritized status as an acknowledgement of power differentials between the dominant White non-Hispanic majority and Black, Indigenous, and people of color (BIPOC) minority. Our operationalization of this construct thus speaks to this inherent power differential. As described, the random effects framework will be utilized to capitalize on the longitudinal nature of the data, statistical power, and flexibility of this modeling strategy. However, this aim will examine a between effect for birth sex (Model 1) and race/ethnicity (Model 2).

Handling of missing data

Missing data will be handled consistent with current expert recommendations.^33,34 This approach emphasizes (1) either maximum likelihood or multiple imputations, and (2) extensive sensitivity analyses, including missing not at random approaches in order to examine the robustness of the treatment effects across varying assumptions.³⁵ This strategy involves the inclusion of multiple predictors of missing data including treatment success, attendance, baseline values of the outcome(s), age, birth sex, and other covariates that can likely account for much of the missing data. The approach also includes a variety of missing not at random³⁶ strategies including Diggle-Kenward or Wu-Carroll selection modeling, or various types of pattern mixture modeling³⁷ if necessary in order to test the sensitivity of an observed treatment effect across a variety of different missing data assumptions.

Sample size determination

The target sample size of N = 608 for this study is based on Aim 2 analyses of heterogeneity of treatment effects across birth sex, race, and ethnicity. The study assumed equal enrollment of rural veterans with chronic pain who report birth sex of female vs. male and race/ethnicity of White non-Hispanic vs. minoritized race or ethnicity based on the sampling scheme and recruitment strategy. A clinically meaningful effect on primary outcomes was specified to be d = 0.50. For pain interference, this effect size equates to an approximately 30% reduction.³⁸ A 15% reduction, or d = 0.25, is not believed to be of clinical significance.³⁸ Using birth sex for illustrative purposes, the CORPs sample size of n = 243 is needed to detect a statistically significant effect size difference between females and males of d = 0.25 with 80% power. If the observed effect size difference is smaller than 0.25, the treatment effects of CORPs will be deemed to be clinically equivalent between females and males. If the observed effect size difference is larger than 0.25, and approaching d = 0.50, it can be concluded that treatment effects of CORPs differ between female and male veterans. The same assumptions are followed for rural veterans who receive the CORPs intervention and who identify as White non-Hispanic vs. minoritized race or ethnicity. A CORPs sample of 243 results in an overall study sample (both CORPs and MEUC participants) of N = 486. Accounting for 20% attrition, the study must enroll 608 rural veterans with chronic pain.

For Aim 1 primary and secondary outcome analyses, power is reported for an effective sample size of N = 486. Study preliminary data support a minimal effect size of d = 0.50 across primary study outcomes. The study conservatively assumes a two-tailed test, an α = 0.05, an intraclass correlation of 0.4 over time between assessments, which would provide 99% power to detect a minimal effect size of d = 0.50. Similar treatment effects are assumed for secondary outcomes that follow approximately normal distributions. In addition, for the count outcome of nonpharmacologic pain treatment utilization, assuming a base rate of engagement among 20% of the sample in the MEUC comparator group, and the anticipated increase in engagement among CORPs participants of an additional 20% (totaling 40% engagement among the CORPs group), this results in 95% power to detect such a treatment utilization increase in each class of nonpharmacologic pain treatment.

Incremental CEA

The primary clinical outcome for the analysis will be quality-adjusted life years (QALYs) and will follow recommended procedures for the estimation of the costs of the intervention and usual care services costs used by study subjects.¹² All costs will be adjusted to a year in the middle of the study period to account for inflation. The timeframe of the CEA is from enrollment through 12 months.

The primary analysis will calculate an Incremental Cost Effectiveness Ratio (ICER) with cost per QALY as the outcome. The use of QALYs as an outcome metric reflects a utility-based approach³⁹ and is recommended by the Public Health Task Force.¹² In addition, the evaluation of QALYs provides information on possible life years saved as well as changes in health-related quality of life. The VR-12 was chosen for assessing QALYs because it includes important quality-of-life domains specific to veterans and includes items related to pain.¹⁹ We will convert data from the VR-12 to QALYS using established algorithms.^30,31

Patient-level intent-to-treat CEA analyses will be conducted. The primary analyses will be total costs at 6 and 12 months after participant enrollment. Costs will be estimated using healthcare utilization data from the PHCUS, VA’s EHR data, and published unit costs. Cost data will likely not be normally distributed. Thus, the net benefit regression method with ordinary least squares regression analyses will be used to examine cost-effectiveness.^39,40 The robustness of the parametric tests will be confirmed using nonparametric bootstrapping with a single model with 1000 replications using the bias-corrected and accelerated method.^41,42 Adjusted differences between the conditions will be estimated using ordinary least squares regression models with bootstrap interval estimates. All analyses will be adjusted for baseline characteristics including age, gender, race/ethnicity, recruitment site, baseline costs, and baseline pain symptom severity. Hypothesis tests for the clinical and cost outcomes will be based on the significance of the group variable in the bootstrapped multiple regression equations.⁴² Missing data will be imputed using multiple imputations with chained equations. Five imputation datasets will be created and combine estimates such that the standard errors reflect the variability introduced by the imputation process. To represent uncertainty around the ICER estimates, we will use scatter plots of bootstrapped cost-and-effect pairs, and will construct cost-effectiveness acceptability planes to present the probability that the intervention is cost-effective across a range of willingness to pay values.^39,41,42

Implementation process evaluation

The implementation process evaluation for the CORPs study is guided by the Consolidated Framework for Implementation Research (CFIR).⁴³ CFIR describes five constructs important to the implementation of interventions into clinical settings. These include:

Outer Setting—policies and systemic pressures that support, or sometimes mandate, implementation; Inner Setting—local contextual factors that require consideration to support implementation; Characteristics of Individuals—attitudes of implementers, including clinicians and administrators; Intervention Characteristics—attributes of the intervention that may facilitate or impede implementation; and Process—strategies through which implementation is executed.

The CORPs study will collect multiple forms of qualitative data, including: online diaries that collect real-time data from NCMs and consulting clinicians; virtual learning community observations; and semistructured interviews with patients, researchers, clinicians, and clinic managers, which will allow participants to recount their experience and reflect on the intervention.

Rapid qualitative analysis methods⁴⁴ will be used to analyze data collected during the implementation process evaluation, following guidance for best practices in qualitative methods for rapid turn-around in health services research.^45,46 First, a summary of each interview will be generated and condensed to a 1-2 page document (ie, data reduction). Second, team members will carefully review the summaries, examining their usability and relevance. The format of summaries will be iteratively modified until the team identifies appropriate (most useful) domains to categorize the data into a templated matrix. Once the template is established, each interview will be summarized and data added to the matrix for rapid analysis of interview content. Concurrently, data will be extracted from the online diaries and added to a templated matrix. Matrix domains will be updated as sites progress through implementation of CORPs to capture process data. Finally, direct observation of the virtual learning community will be documented in the form of handwritten or typed fieldnotes. Fieldnotes will be reviewed monthly by members of the qualitative analysis team and data extracted for incorporation into the same templated matrix used for interview and diary data.

Discussion

This is the first pragmatic trial to test the effectiveness of remotely delivered collaborative care to rural veterans living with chronic pain. Throughout the development of the study protocol, we have worked closely with key partners, including (1) leadership from national VA pain and rural health program offices, (2) veterans who reside in rural locations and those with lived pain experience, (3) rural VA and community clinicians, and (4) VA clinical administrators at the facilities where the pragmatic trial will be executed. As a member of the Pain Management Collaboratory (PMC),¹⁰ we also receive expert consultation on all aspects of the trial design through the PMC’s many work groups, including measure selection, statistical analysis plan, ethical and regulatory elements of the trial, implementation and dissemination planning, and stakeholder engagement.

To ensure the study’s pragmatism, the trial design was further guided by the PRagmatic Explanatory Continuum Indicator Summary (PRECIS-2) tool.⁴⁷ As shown in Figure 2, the CORPs trial is predominantly “very pragmatic” across the nine PRECIS-2 domains, indicated by larger numbers that approach “5” on the PRECIS-2 wheel: (1) eligibility: broad inclusion of veterans with chronic pain who reside in rural areas, including those with co-occurring mental health or substance use disorders; (2) recruitment and (3) setting: Targeted outreach and care delivery through the VA, emulating patients’ typical care settings; (4) organization and (5) flexibility-delivery: study interventionists include VA clinicians who will receive training in core elements of the intervention using methods the VA offers when implementing new interventions; (6) flexibility-adherence: NCMs will ask patients about adherence with pain treatment plans, which is part of routine clinical care and may increase engagement; (7) follow-up: quarterly assessments over 12 months that leverage data available from the EHR; (8) primary outcome: study primary outcome is pain interference which is highly relevant to patients; (9) primary analysis: all available data with intent-to-treat analyses.

This pragmatic trial purposely includes detailed implementation process data and a CEA. Should CORPs demonstrate clinical effectiveness, these supplementary data could help to support wider-scale dissemination of the intervention to VA healthcare systems nationally, thus improving the quality of life for rural veterans living with chronic pain.

Funding

This work is supported through cooperative agreement UH3/UG3 AT012257 from the National Center for Complementary and Integrative Health (NCCIH) and VHA Office of Rural Health Veterans Rural Health Resource Center-Portland (NOMAD No. PROJ-04142). The study was also supported by resources from the Center to Improve Veteran Involvement in Care at the VA Portland Health Care System. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, Department of Veterans Affairs, or U.S. Government. This manuscript is a product of the Pain Management Collaboratory. For more information about the Collaboratory, visit https://painmanagementcollaboratory.org/.

Conflicts of interest: None declared.

Supplement statement

This article appears as part of the supplement entitled “Pain Management Collaboratory: Updates, Lessons Learned, and Future Directions.”

This manuscript is a product of the Pain Management Collaboratory. For more information about the Collaboratory, visit https://painmanagementcollaboratory.org/.

References