Use of implementation mapping in the planning of a hybrid type 1 pragmatic clinical trial: the BeatPain Utah study

Background Considerable disparities in chronic pain management have been identified. Persons in rural, lower income and minoritized communities are less likely to receive evidence-based, nonpharmacologic care. Telehealth delivery of nonpharmacologic, evidence-based interventions for persons with chronic pain is a promising strategy to lessen disparities, but implementation comes with many challenges. The BeatPain Utah study is a hybrid type I effectiveness-implementation pragmatic clinical trial investigating telehealth strategies to provide nonpharmacologic care from physical therapists to persons with chronic back pain receiving care in Community Health Centers (CHCs). CHCs provide primary care to all persons regardless of ability to pay. This paper outlines the use of implementation mapping to develop a multifaceted implementation plan for the BeatPain study. Methods During a planning year for the BeatPain trial we developed a comprehensive logic model including the 5-step implementation mapping process informed by additional frameworks and theories. The five iterative implementation mapping steps were addressed in the planning year; 1) conduct needs assessments for involved groups; 2) identify implementation outcomes, performance objectives and determinants; 3) select implementation strategies; 4) produce implementation protocols and materials; and 5) evaluate implementation outcomes. Results CHC leadership/providers, patients and physical therapists were identified as involved groups. Barriers and assets were identified across groups which informed identification of performance objectives necessary to implement two key processes; 1) electronic referral of patients with back pain in CHC clinics to the BeatPain team; and 2) connecting patients with physical therapists providing telehealth. Determinants of the performance objectives for each group informed our choice of implementation strategies which focused on training, education, clinician support and tailoring physical therapy interventions for telehealth delivery and cultural competency. We selected implementation outcomes for the BeatPain trial to evaluate the success of our implementation strategies. Conclusions Implementation mapping provided a comprehensive and systematic approach to develop an implementation plan during the planning phase for our ongoing hybrid effectiveness-implementation trial. We will be able to evaluate the implementation strategies used in the BeatPain Utah study to inform future efforts to implement telehealth delivery of evidence-based pain care in CHCs and other settings. Trial registration Clinicaltrials.gov Identifier: NCT04923334. Registered June 11, 2021 (https://clinicaltrials.gov/study/NCT04923334


INTRODUCTION
One in 5 Americans live with chronic pain. 1 Back pain is the most prevalent form of chronic pain. 2 Evidence-based interventions (EBIs) for chronic low back pain (LBP) include nonpharmacologic treatments provided by physical therapists (PTs) and others. 3,4Guidelines advise against opioids, [5][6][7] yet evidence-practice gaps persist.10][11][12][13][14] Considerable disparities exist in chronic pain prevalence and management.Prevalence is higher among persons with less income or education, and in rural communities. 15,16These individuals are more likely to receive opioids [17][18][19] and less likely to receive nonpharmacologic care. 20,21Many persons with these characteristics receive primary care in Community Health Centers (CHCs), 22;23,24 yet geographic and other barriers limit access to nonpharmacologic care. 23About a third of CHC patients are Latino/a, introducing additional barriers related to language and sociocultural t. 25,26 Collectively these barriers contribute to pain disparities. 27lehealth delivery of nonpharmacologic care could lessen disparities, 28 but application has been limited. 29,30Experiences during COVID support telehealth's potential to increase access, 31 but issues speci c to implementation in underserved communities must be considered. 32Implementation mapping (IM) is a systematic approach to iteratively develop scalable and sustainable EBI implementation strategies. 33,34This paper describes application of IM in the BeatPain Utah study examining telehealth EBIs for patients with chronic LBP in CHCs.

BeatPain Utah Study
BeatPain Utah 35 (Clinicaltrials.govIdenti er: NCT04923334) is a pragmatic clinical trial examining two PT-led telehealth EBIs; a brief consult and an extended PT program; delivered across two treatment phases.Further details are published. 35BeatPain is a hybrid type 1 trial, primarily focusing on effectiveness and secondarily on implementation outcomes. 36Implementation mapping occurred during a planning year before enrollment.Planning activities, including community-member interviews, were approved by the University of Utah Institutional Review Board.
BeatPain Utah is being conducted in 9 CHC organizations in Utah serving urban and rural communities.Among persons served by these clinics approximately 49% identify as Hispanic/Latino/a, 37% are non-English speakers, 45% are uninsured, and 59% are below the federal poverty level. 37Patients in CHC clinics are referred to BeatPain through standards-based, HIPAA-compliant electronic referral (e-referral) from a CHC EHR using phiMail® (EMR Direct, Inc., San Diego, CA).Once an e-referral is received a BeatPain team member contacts the patient.Patients opting to enroll provide oral consent.Those choosing not to participate or ineligible are offered care without study data collection.

Implementation Logic Model
Implementation mapping was informed by additional frameworks to develop a logic model (Fig. 2).The Consolidated Framework for Implementation Research (CFIR) helped identify contextual factors across 5 domains that could in uence e-referral and telehealth implementation; the innovation being implemented, outer setting, inner setting, individuals involved, and the implementation process.Social-Cognitive Theory (SCT) emphasizes factors in uencing behavior including; 1) environmental factors (e.g., social support, cultural context); 2) cognitive factors (e.g., knowledge, self-e cacy); and 3) behavioral factors (e.g., coping strategies, outcome expectancies). 38Proctor's taxonomy of Outcomes for Implementation Research, 39 informed IM step 5.

Implementation Mapping
Implementation mapping was based on a 5-step model 33 detailed below and depicted in Fig. 2.
Step 1 -Needs Assessment Step 1 identi ed barriers and facilitators using needs assessment for groups involved in implementation including patients, CHC leadership/providers, and PTs. 40,41For CHC leadership/providers, needs assessments focused on the existing and desired clinic work ows for identifying and treating individuals with LBP; and the type of feedback desired about referred patients.The goal was collection of data to inform e-referral implementation. 42,43For patients, we conducted individual, remote interviews with individuals with LBP in communities served by CHC clinics.Interviews used an ethnographic approach focused on persons' lived experience including their understanding of LBP and healthcare experiences. 44eeds assessments for PTs about telehealth implementation were conducted in group meetings.COVID restrictions necessitated remote assessments for all stakeholder groups.
Step 2 -Identify Implementation Outcomes, Performance Objectives and Determinants Step 2 used step 1 ndings to identify implementation outcomes and performance objectives for each group.Implementation outcomes identi ed behavior goals.Performance objectives identi ed tasks required to achieve an implementation outcome.Next, determinants were speci ed, informed by step 1 ndings and the SCT framework. 45Determinants are modi able, internal factors helping explain why persons would achieve a performance objective. 46For example, self -e cacy is an important determinant of behavior change and relates to one's beliefs in their capacity to perform an action and persist despite barriers. 38Another behavior change determinant is outcomes expectancy, i.e., is the belief that performing an action will lead to a particular outcome, and is another behavior change determinant. 47ep 3-Select Implementation Strategies In step 3, methods, or processes that can change a determinant based on evidence or theory (e.g., increase knowledge, change attitudes) 48 were selected.Methods informed the choice of implementation strategies.Strategies 49 were then operationalized as practical applications that t within the context described by CFIR and the needs assessments. 46ep 4-Create Implementation Protocols Step 4 operationalized implementation strategies by producing protocols and materials for training and delivery.
Step 5-Evaluate Implementation Outcomes Step 5 developed a plan to evaluate outcomes of the implementation processes during the BeatPain study.Outcomes were based on Proctor's taxonomy 39 and included acceptability, adoption, feasibility, and delity.

RESULTS
Step 1 -Needs Assessment Needs assessment outcomes are described in Table 1.Assessments were conducted with CHC leaders/providers from 7 organizations.Participants expressed receptivity to PT-led telehealth.
Facilitators included the ability to accommodate patients in Spanish or English, without cost, and receiving feedback about referred patients.Potential barriers centered on EHR capabilities for making ereferrals, time constraints and competing demands, and integrating a new work ow, particularly in light of COVID impacts on clinic operations.Prior experience providing care in low income and Spanishspeaking communities.
Availability of team members with Spanish language skills and cultural background.
Limited experience or training to provide treatments using telehealth.
Inexperienced providing care using audio-only telehealth delivery with no video access.
Concerns about the ability to engage and motivate patients and provide effective exercise instruction via telehealth.
We interviewed ve female patients with LBP, each of whom identi ed as Hispanic/Latina.Three preferred communication in Spanish, two in English.Facilitators identi ed included positive experiences and trust in CHCs and availability of cell phones for telehealth.Barriers included lack of reliable internet and technology for video telehealth sessions.There was a general lack of awareness that PT could be provided by telehealth and that it may be bene cial.Some interviewees expressed preferences for passive pain coping, including medication or rest, that are not EBIs.
Facilitators for PTs included commitment to providing care to persons in historically marginalized communities.Most were bilingual and some had experience providing PT with Spanish-speaking patients.Barriers included lack of telehealth experience, the need to adapt treatments for phone-only and video telehealth delivery, and ability to engage and motivate patients using telehealth.
Step 2 -Identify Implementation Outcomes, Performance Objectives and Determinants We identi ed implementation outcomes as participating in BeatPain and referring persons with LBP for CHC leadership/providers; engaging in telehealth for patients; and providing telehealth with delity to intervention core components for PTs.Performance objectives derived from these outcomes, and associated determinants are outlined in Table 2.

Regular team meetings for problemsolving and skills practice
Determinants informed our choice of implementation strategies to help each group achieve their performance objectives.For CHC leadership/providers we identi ed knowledge of nonpharmacologic pain care and expectations for telehealth PT as determinants of agreeing to participate and place ereferrals.The ability to provide technical support for the implementation of e-referrals within clinic EHRs and avoid work ow disruptions were determinants of sustainment of e-referrals.For patients, we considered knowledge and outcomes expectancy for telehealth PT as determinants for engaging in BeatPain.Determinants for PTs included knowledge of how to deliver the BeatPain intervention core components using telehealth, and self-e cacy to engage patients who may have different cultural backgrounds.
Step 3-Select Implementation Strategies Implementation methods relevant across groups included increasing knowledge, changing awareness, changing attitudes and beliefs, developing skills, capabilities and self-e cacy, and outcome expectations. 48We matched these methods to implementation strategies (Table 2).
Implementation strategies for CHC leadership/providers included education about nonpharmacologic LBP care and the BeatPain program; and training on how to explain BeatPain to patients to increase selfe cacy for referring patients.Additional implementation strategies included hands-on technical support to implement e-referrals within each clinic's EHR (three different EHR products are used across organizations) and provide ongoing technical support.We used a secure, EHR-based process designed to be minimally disruptive to existing work ows.We chose secure messaging based on the Direct standard protocol using phiMail because this approach was HIPAA-compliant, standards-based, inexpensive, and bidirectional.Since the Direct protocol is required for EHR certi cation, it was supported by all EHR systems used in Utah CHC clinics. 50This strategy also provided the infrastructure for PTs to return feedback to providers on patients' status, helping to build positive expectations about BeatPain.
Although in-clinic e-referrals are minimally disruptive, they require clinician recall during a visit.We therefore included a second implementation strategy using text messaging to capture individuals for whom a referral may have been unaddressed during the visit.We used a population health management system (Azara Healthcare, Burlington, Massachusetts, USA) that interfaces with clinic EHRs and identi es eligible patients (recent appointment for LBP), then automatically sends a bidirectional text message introducing the BeatPain project and offering a connection to telehealth.Patients who respond positively are noted on an electronic dashboard and clinic staff can place an e-referral.
Implementation strategies for patients focused on engagement to build positive expectancies for telehealth PT.Explanations used by BeatPain personnel to describe the program to patients and patientfacing materials such as the project's webpage were tailored to provide information on what telehealth PT involves and its potential bene ts.Additionally, a strategy of adapting and tailoring telehealth PT interventions addressed the determinant of building patient self-e cacy for active pain coping.
Telehealth PT interventions were adapted to include a motivation-and-problem-solving (MAPS) approach found effective for chronic care management and substance use treatment. 51The MAPS approach is appropriate for persons irrespective of their readiness to change and explicitly targets motivation and self-e cacy as behavior change mechanisms. 52,53ysical therapist implementation strategies included training and education on providing care using telehealth, especially when communication is audio-only.Integration of mHealth resources including appbased exercise and education platforms helped support patients' self-management.Physical therapists were trained in MAPS using didactic and interactive strategies to build self-e cacy for delivering the BeatPain core components.MAPS includes motivational interviewing and cognitive behavioral techniques that help patients set personalized goals and manage barriers towards achieving these goals. 51To further build self-e cacy, the MAPS expert on the BeatPain team provided coaching through role-playing a PT session and providing feedback.Physical therapists were also trained on culturally competent care to meet patients' sociocultural and linguistic needs. 54ep 4-Create Implementation Protocols Step 4 operationalized implementation strategies.For CHC leadership/providers we developed brief (10-15 minute) presentations, for in-person or remote delivery, focused on evidence supporting nonpharmacologic care and the BeatPain program.Detailed instructions for placing e-referrals in the clinic's EHR were provided along with suggested language providers could use to describe BeatPain to patients.Ongoing updates were provided during staff meetings, including reminders on e-referral procedures, trouble-shooting barriers, and anecdotal patient experiences.
We operationalized patient implementation strategies by developing a study webpage, in English and Spanish, accessible through a QR code on recruitment materials, to describe BeatPain and build positive expectations (https://health.utah.edu/physical-therapy-athletic-training/research/clinical-outcomesresearh/beatpain-utah/eng).The webpage described the partnership between BeatPain and CHC clinics, gave biographies of BeatPain personnel, and described telehealth PT treatment.We used the MedBridge phone app (MedBridge, Inc., Bellevue, WA) to provide exercise and education videos.
Physical therapist training focused on the intervention core components and integration of MAPS for English and Spanish-speaking patients. 35Training in culturally competent care used Hays ADDRESSING framework as a structured self-exploration method of how the PT's cultural background may interact with their patient's background and in uence care. 55We also used Betancourt's framework for cross-cultural communication to help PTs consider major cultural issues they may encounter and provide personcentered care characterized by empathy and respect for patients' values and preferences. 56 Step 5-Evaluate Implementation Outcomes Implementation outcomes for the BeatPain trial 35 were selected to evaluate the success of the IM process across groups.Because BeatPain adapted nonpharmacologic pain EBIs for underrepresented populations using novel delivery strategies, we identi ed important implementation outcomes as adoption, acceptability, feasibility, and delity as de ned in Table 3. barriers to e-referral implementation and sustainment. 57These determinants were in uenced by COVID impacts on sta ng and the varied EHR systems used within clinics.We used brief, intermittent trainings on the e-referral process, ongoing EHR support, and a secondary text message recruitment approach as strategies to address these determinants.Lack of knowledge and uncertain expectations for telehealth PT were not surprising given limited exposure to nonpharmacologic EBIs or telehealth PT in CHCs; 20,23,58 but could adversely impact providers' self-e cacy for advising patients with LBP about BeatPain.
Educating providers and sending feedback on referred patients addressed these concerns.
Patient needs assessments re ected uncertainty that telehealth PT is equivalent to in-person care, consistent with other reports. 59,60Also, studies during COVID reported persons of Hispanic ethnicity expressed less willingness to use telehealth. 61,62We therefore considered positive expectations about telehealth PT a determinant of patients' attendance, which we addressed by emphasizing the personalized nature of telehealth and ability to individualize care in patient-facing materials.Some patients expressed preferences for passive coping strategies that are associated with lower self-e cacy for engaging in active EBIs such as physical activity. 635][66] We addressed this through training PTs in cross-cultural communication emphasizing the patient as teacher; and strategies to negotiate differences in a patient-centered manner. 56,67ke most PTs, BeatPain therapists had limited telehealth experience. 68Thus, building PTs' self-e cacy to deliver telehealth care and build effective patient-therapist relationships, particularly with phone-only communication, was a key determinant.We trained PTs in MAPS which combines motivational interviewing and cognitive behavioral techniques to help patients set and achieve personally meaningful goals. 51Motivational interviewing is a person-centered communication strategy well-suited to phone delivery [69][70][71] and circumstances where the patient and PT have different cultural backgrounds, 72,73 possibly due to motivational interviewing's collaborative, non-judgmental nature which may reduce risks for implicit biases by providers. 74

CONCLUSION
BeatPain Utah is a hybrid type I trial evaluating effectiveness and implementation outcomes. 36IM provided a systematic, theory-driven process to develop and evaluate implementation strategies.
Assessment of implementation outcomes will allow us to evaluate the success of our implementation strategies for future trials and clinical applications.

Table 1
Results of the needs assessments conducted for implementation mapping step 1 across groups (CHC = Community Health Center; EHR = Electronic Health Record).

Table 2
53 )ings from implementation mapping steps 2 and 3 across groups (CHC = Community Health Center; EHR = Electronic Health Record; ‡ implementation strategies based on the taxonomy of Waltz et al53 )

Table 3
43plementation outcomes evaluated in the BeatPain study (CHC = Community Health Center; ‡ outcome domains and de nitions based on the framework of Proctor et al43).Of note, the performance objectives in Table2were identi ed as the critical steps for achieving the implementation outcomes listed here.Implementing telehealth pain care in CHCs creates an opportunity to increase the reach of EBIs and reduce pain management disparities.BeatPain Utah uses an e-referral process of persons with LBP from CHC clinics to a centralized telehealth PT team; requiring behavior changes and new work processes for clinical teams, patients and PTs.The 5-step IM process informed by additional models helped us to;understand needs and assets for CHC leadership/providers, patients and PTs; identify actions necessary to achieve implementation outcomes; identify determinants of those actions; and operationalize implementation strategies to address key determinants.Through this process we developed a multifaceted implementation plan to connect patients with telehealth EBIs.The nal IM step identi ed implementation outcomes for the hybrid effectiveness-implementation study.We identi ed knowledge and positive expectations around nonpharmacologic pain care, EHR support for placing e-referrals, and minimizing work ow disruptions as important determinants of e-referrals from CHC clinics; consistent with other ndings that technology challenges and work ow disruptions are This study was approved by the University of Utah institutional review board.Consent for participation in the activities reported in this study was obtained following guidelines from the University of Utah institutional review board.