Abstract
The purpose of this study was to describe the referrals and use of a hybrid care model for low back pain that includes on-site care by physical therapists, physical activity training, and psychologically informed practice (PiP) delivered by telehealth in the Improving Veteran Access to Integrated Management of Low Back Pain (AIM-Back) trial.
Data were collected from November 2020 through February 2023 from 5 Veteran Health Administration clinics participating in AIM-Back, a multisite, cluster-randomized embedded pragmatic trial. The authors extracted data from the Veteran Health Administration Corporate Data Warehouse to describe referral and enrollment metrics, telehealth use (eg, distribution of physical activity and PiP calls), and treatments used by physical therapists and telehealth providers.
Seven hundred one veterans were referred to the AIM-Back trial with 422 enrolling in the program (consult-to-enrollment rate = 60.2%). After travel restrictions were lifted, site visits resulted in a significant increase in referrals and a number of new referring providers. At initial evaluation by on-site physical therapists, 92.2% of veterans received pain modulation (eg, transcutaneous electrical nerve stimulation, manual therapy). Over 81% of enrollees completed at least 1 telehealth physical activity call, with a mean of 2.8 (SD = 2.0) calls out of 6. Of the 167 veterans who screened as medium to high risk of persistent disability, 74.9% completed at least 1 PiP call, with a mean of 2.5 (SD = 2.0) calls out of 6. Of those who completed at least 1 PiP call (n = 125), 100% received communication strategies, 97.6% received pain coping skills training, 89.6% received activity-based treatments, and 99.2% received education in a home program.
In implementing a hybrid care pathway for low back pain, the authors observed consistency in the delivery of core components (ie, pain modulation, use of physical activity training, and risk stratification to PiP), notable variability in telehealth calls, high use of PiP components, and increased referrals with tailored provider engagement.
These findings describe variability occurring within a hybrid care pathway and can inform future implementation efforts.
Background
Embedded pragmatic clinical trials (ePCTs) present a unique opportunity to study the effectiveness and implementation of health care treatments in real-world clinical settings.1,2 These trial designs use existing clinical staff and workflows to deliver core components of a trial intervention while allowing sites flexibility to organize care based on local resources, staff arrangements, and institutional policies.3 In doing so, ePCTs have the potential to answer important questions about outcomes, scalability, and treatment uptake that explanatory randomized controlled trials cannot answer.4,5 ePCTs are particularly well suited to study novel ways of organizing and delivering care, such as hybrid models that integrate telehealth with traditional in-person care.6–11
Telehealth has gained significant attention in health care over the last decade for musculoskeletal rehabilitation.12–14 Telehealth offers numerous potential benefits, including the ability to complement in-person interventions with access to treatments like cognitive behavioral therapies, mindfulness training, and physical activity coaching.15–18 Despite the many potential benefits of combining in-person care with telehealth, hybrid models also have numerous challenges. Two key challenges include the need to integrate in-person treatment with virtual care, especially when delivered by different providers, and to gain buy-in from patients and health care professionals on the efficiency and effectiveness of hybrid care compared to in-person care.12,19–22
Hybrid care models are endorsed by people seeking care for musculoskeletal conditions21 and are likely to have an important role in the growth and evolution of physical therapy. Emerging value-based care initiatives incentivize low cost, high-value care options, as well as treatments that take a biopsychosocial approach to improving long-term health and self-management,23–25 both of which may be facilitated by hybrid models. For this reason, physical therapists, and other providers involved with hybrid care pathways, must develop a better understanding of how these models function in real-world settings. To date, most hybrid care models described in the literature are conceptual,16,26,27 delivered in the context of tightly controlled randomized controlled trials,28–32 or “blended models”33 that involve only a single provider type delivering care through telehealth or alongside a digital resource (eg, an app).6,11,32,34–37 Therefore, existing examples provide little guidance on how to integrate interdisciplinary hybrid care models into “real world” health systems.
Improving Veteran Access to Integrated Management of Back Pain (AIM-Back; trial identification: NCT04411420) is a multisite, cluster-randomized ePCT comparing 2 novel clinical care pathways that provide early access to nonpharmacological care for veterans with low back pain (LBP) in the Veterans Health Administration (VHA) Health Care System.38 One of the AIM-Back trial pathways is a hybrid model where physical therapists deliver care alongside a specially trained central provider who conducts physical activity training and psychologically informed practice (PiP) through telehealth. Briefly, this pathway, known as the Sequenced Care Pathway (SCP), is initiated by a primary care referral, followed by on-site physical therapy interspersed with the aforementioned telehealth options.
The purpose of this original research article is to describe pathway uptake of the AIM-Back SCP as an indicator of its scalability potential and real-world implementation. Given the paucity of data reported on hybrid care pathways involving physical therapists, we were specifically interested in: reporting on the number of referrals and program enrollments; describing the effects of tailored site visits on program referral rates; describing variability in use of on-site and telehealth care, including core intervention components; and determining how this information could be used to identify components of hybrid care vulnerable to implementation (ie, on-site vs telehealth) and inform additional implementation efforts.
Methods
Setting
This article reports on data from 5 VHA primary care clinics around the USA that were participating in the AIM-Back study. Nineteen clinics implemented an AIM-Back pathway, with randomization occurring in 2 blocks: 10 in the first block (5 randomized to SCP in August 2020) and 9 in the second block (4 randomized to SCP in December 2021). Within blocks, we took a staggered approach to implementing the SCP, with the first site launching the program in November 2020, the second and third sites launching in February 2021, and the fourth and fifth sites launching in April 2021. We focus on SCP clinics in the first randomization block as recruitment is completed or near completion for these clinics. Clinics in the second randomization block were not included in this study due to being in the middle of enrollment. Additional details on AIM-Back were published in protocol38 and pathway development39 articles, with key pathway features highlighted below.
Participants
We provided AIM-Back primary care providers (PCPs) with the following criteria to identify veterans eligible for participation in AIM-Back: diagnosis of LBP, veteran enrolled in the participating Department of Veterans Affairs primary care clinic, and veteran was not receiving or referred for palliative or hospice care. AIM-Back was delivered at each site as a clinical program with veterans receiving evidenced-based treatments as standard practice for LBP.
Overview of the SCP
The SCP has both on-site and telehealth components (Fig. 1). The program begins when a veteran with LBP receives a referral from a PCP to an AIM-Back physical therapist in the local Department of Veterans Affairs health care system (ie, “on-site” care). The physical therapist initial evaluation visit is designed to provide a guideline-concordant examination, evaluation, and treatment.40 AIM-Back physical therapists are encouraged to use treatments that include but are not limited to pain education, physical activity promotion, and evidence-based pain modulatory interventions (eg, manual therapy, electrotherapy). Veterans also complete the STarT Back Screening Tool (SBST) to identify risk for prolonged disability due to LBP.41 The SBST does not inform care decisions at the initial on-site physical therapist visit, but captured to document the baseline risk profile of veterans entering the program. After the initial on-site physical therapist visit, the program includes 6 weekly telehealth visits (via phone or video call) focused on physical activity coaching from a specially trained “central provider.”
Sequenced Care Pathway (SCP) for Improving Veteran Access to Integrated Management of Low Back Pain (AIM-Back). PT = physical therapist; ROM = range of motion; SBST = STarT Back Screening Tool.
Physical activity coaching sessions include physical activity and home exercise program promotion, facilitation, and progression.42 Furthermore, central providers engage veterans in goal setting, pacing skills, and establishing weekly action plans to optimize home program success. After 6 weeks of physical activity coaching via telehealth, veterans are scheduled to return to their local physical therapist for reexamination, treatment, and readministration of the SBST. Veterans who are “low risk” on the SBST at 6-week follow-up are discharged from the AIM-Back program. Veterans who are “medium to high risk” are eligible to receive 6 additional weekly PiP interventions delivered via telehealth by the central provider. PiP content builds across the 6 calls and includes strategies to improve communication, pain coping skills training (eg, relaxation techniques, pain education), activity-based treatments (eg, value-based goal setting, pacing, and problem solving), and education in home program. Upon completing the 6-week PiP intervention, veterans are either discharged, or they follow-up with their PCP, thus completing the AIM-Back pathway.
Central providers and physical therapists document their encounters within the same local Computerized Patient Reporting System, which facilitates communication and continuity of care. They can also communicate over a secure chat system, Microsoft Teams, to discuss essential patient needs and care plans to ensure a seamless handoff between providers. AIM-Back included 4 central providers, all of whom were located in North Carolina.
Pathway Modifications Due to COVID-19
Due to national and local polices that affected health care access during the COVID-19 pandemic, some VHA clinics only offered initial evaluation and follow-up visits by on-site physical therapists via phone or video call. In these situations, local physical therapists completed the same evaluation and documentation as in-person care, except that interventions were delivered virtually (eg, education in self-mobilization techniques, instruction on transcutaneous electrical nerve stimulation [TENS] at home). For the purposes of this descriptive study, we did not distinguish between on-site physical therapist visits completed in person versus via phone or video.
Overview of Health Care Provider Training
As this was a new program in the VHA, we focused our initial implementation efforts on training on-site physical therapists, PCPs, and central providers on how to deliver the pathway and document care. All training occurred prior to the program launching at each site, although we offered ad hoc training as new providers joined each clinic. We originally planned for in-person site visits as part of AIM-Back physical therapist and PCP training. However, the onset of COVID-19 and its associated travel restrictions required us to conduct all initial training virtually, including the training of central providers. Licensed physical therapists with board-certified orthopedic specializations serve as central providers, although we designed the role to be agnostic to clinical training backgrounds. The Supplementary Appendix provides detailed information on physical therapist, PCP, and central provider training.
Ongoing Site Engagement and Monitoring
Following program launch, we scheduled regular communications with PCPs and physical therapists through emails, direct Microsoft Teams messages, and virtual meetings. Weekly emails to PCPs and physical therapists consisted of a brief report on recruitment progress and an acknowledgement of PCPs who submitted consults in the previous week. Monthly briefings sent via email to PCPs, physical therapists, and health care system points of contact included a more comprehensive report on recruitment progress and addressed common barriers to enrollment at each site. The AIM-Back research team joined clinic-wide monthly virtual PCP staff meetings to remind providers about the program, give enrollment updates, and answer questions. Additionally, we routinely reached out via Microsoft Teams messenger or email to PCPs to remind them of AIM-Back and answer questions about the program or referral process. Our research team met quarterly with each AIM-Back physical therapist to discuss progress and troubleshoot challenges.
Once COVID-19 travel restrictions were lifted in early 2022, we conducted in-person site visits to 3 of the 5 SCP sites that had not yet met their recruitment goals. These sites had launched AIM-Back approximately 1 year prior to our visit. Site visits generally lasted 1 to 2 days. We tailored site visits to the needs of each clinic and included meetings with PCPs, physical therapists, and executive leadership teams to build relationships, discuss progress, troubleshoot challenges, train new providers, deliver education and marketing materials, and generate enthusiasm for the program.
Data Collection and Management
We conducted descriptive analyses of AIM-Back referral and used data to evaluate patient flow through the pathway. Data sources included VHA’s Corporate Data Warehouse, which is a repository of Department of Veterans Affairs electronic health record data. Specifically, we extracted data on AIM-Back referrals, enrollment, and treatment use in the SCP from health factors and structured text fields generated from templated notes designed for the pathway. Data from program initiation through February 28, 2023 are reported for each Block 1 clinic randomized to the SCP pathway, coinciding with the timeline necessary for data extraction and analysis for this manuscript.
We decided a priori to report on specific metrics that would best describe real-world uptake of the program. These metrics assessed referral and enrollment (number of veterans seen for LBP in participating AIM-Back clinics; number of veterans issued a referral to AIM-Back; number of veterans enrolling in the program by attending the physical therapist initial evaluation visit; referral-to-enrollment rate), telehealth use (distribution of physical activity and PiP calls among those eligible to receive services), and specific intervention use (distribution of skills used during the initial physical therapist evaluation, distribution of skills used during PiP calls). These metrics would allow us to determine whether core components of the pathway were being delivered as intended in the design of the SCP. We established the following core components during the development phase of AIM-Back based on recommended pragmatic clinical trial design features5,43: pain modulation at initial evaluation (eg, TENS, manual therapy), guided physical activity delivered by the central provider, and risk stratification using SBST at 6 weeks followed by PiP if indicated. Descriptive statistics are presented with proportions calculated for categorical variables and means/medians for continuous variables.
Role of the Funding Source
The funders played no role in the conduct or reporting of this study.
Results
Referral and Enrollment
From study initiation at each AIM-Back clinic through February 2023, there were 6994 clinic visits for veterans with an AIM-Back trained provider that included a documented LBP diagnosis (range = 927–2331 visits across clinics) (Tab. 1). Across all 5 clinics, PCPs issued AIM-Back referrals to 701 unique veterans. Of the 701 referrals, 241 were discontinued or cancelled (eg, veteran did not attend the initial evaluation, veteran decided not to pursue care, referral was administratively discontinued) and 422 attended the physical therapist initial evaluation visit, indicating enrollment in the AIM-Back program (mean referral-to-enrollment rate = 60.2%). As of the end of February 2023, 38 referrals were still in process, meaning veterans had not yet had their initial physical therapist visit and/or the referral had not been discontinued or cancelled. Across clinics, the number of referrals ranged from 37 to 179, while enrollment ranged from 14 to 129 veterans (range of rates of referral to enrollment = 37.8%–72.1%).
Distribution of Use Metrics Across the 5 AIM-Back Sequenced Care Pathway Clinical Sitesa
| Measure | Site 1 | Site 2 | Site 3 | Site 4 | Site 5 |
|---|---|---|---|---|---|
| Time in trial (mo) | 25 | 25 | 28 | 23 | 23 |
| No. of clinic visits with documented LBP diagnosis | 1446 | 1278 | 1012 | 2331 | 927 |
| No. of AIM-Back referrals issued | 179 | 179 | 158 | 148 | 37 |
| No. of enrolled veterans | 115 | 129 | 95 | 69 | 14 |
| Rate of referral to enrollment (%) | 64.25 | 72.07 | 60.13 | 46.62 | 37.84 |
| Proportion of those enrolled with ≥1 PA call (%) | 79.13 | 87.60 | 73.08 | 79.03 | 100.00 |
| Mean no. of PA calls | 2.81 | 3.13 | 2.46 | 2.47 | 3.00 |
| Proportion of enrolled veterans with SBST medium to high risk at baseline (%) | 89.57 | 86.05 | 76.92 | 90.32 | 100.00 |
| Proportion of those with SBST medium to high risk at 6 wk and ≥1 PiP call (%) | 73.91 | 78.18 | 66.67 | 79.31 | 70.00 |
| Mean no. of PiP calls | 3.50 | 33.85 | 3.52 | 3.69 | 2.80 |
| Proportion of veterans who received a pain modulatory intervention at baseline (%) | 93.91 | 93.02 | 75.79 | 95.65 | 100.00 |
| Measure | Site 1 | Site 2 | Site 3 | Site 4 | Site 5 |
|---|---|---|---|---|---|
| Time in trial (mo) | 25 | 25 | 28 | 23 | 23 |
| No. of clinic visits with documented LBP diagnosis | 1446 | 1278 | 1012 | 2331 | 927 |
| No. of AIM-Back referrals issued | 179 | 179 | 158 | 148 | 37 |
| No. of enrolled veterans | 115 | 129 | 95 | 69 | 14 |
| Rate of referral to enrollment (%) | 64.25 | 72.07 | 60.13 | 46.62 | 37.84 |
| Proportion of those enrolled with ≥1 PA call (%) | 79.13 | 87.60 | 73.08 | 79.03 | 100.00 |
| Mean no. of PA calls | 2.81 | 3.13 | 2.46 | 2.47 | 3.00 |
| Proportion of enrolled veterans with SBST medium to high risk at baseline (%) | 89.57 | 86.05 | 76.92 | 90.32 | 100.00 |
| Proportion of those with SBST medium to high risk at 6 wk and ≥1 PiP call (%) | 73.91 | 78.18 | 66.67 | 79.31 | 70.00 |
| Mean no. of PiP calls | 3.50 | 33.85 | 3.52 | 3.69 | 2.80 |
| Proportion of veterans who received a pain modulatory intervention at baseline (%) | 93.91 | 93.02 | 75.79 | 95.65 | 100.00 |
aAIM-Back = Improving Veteran Access to Integrated Management of Back Pain; LBP = low back pain; PA = physical activity; PiP = psychologically informed practice; SBST = STarT Back Screening Tool.
Distribution of Use Metrics Across the 5 AIM-Back Sequenced Care Pathway Clinical Sitesa
| Measure | Site 1 | Site 2 | Site 3 | Site 4 | Site 5 |
|---|---|---|---|---|---|
| Time in trial (mo) | 25 | 25 | 28 | 23 | 23 |
| No. of clinic visits with documented LBP diagnosis | 1446 | 1278 | 1012 | 2331 | 927 |
| No. of AIM-Back referrals issued | 179 | 179 | 158 | 148 | 37 |
| No. of enrolled veterans | 115 | 129 | 95 | 69 | 14 |
| Rate of referral to enrollment (%) | 64.25 | 72.07 | 60.13 | 46.62 | 37.84 |
| Proportion of those enrolled with ≥1 PA call (%) | 79.13 | 87.60 | 73.08 | 79.03 | 100.00 |
| Mean no. of PA calls | 2.81 | 3.13 | 2.46 | 2.47 | 3.00 |
| Proportion of enrolled veterans with SBST medium to high risk at baseline (%) | 89.57 | 86.05 | 76.92 | 90.32 | 100.00 |
| Proportion of those with SBST medium to high risk at 6 wk and ≥1 PiP call (%) | 73.91 | 78.18 | 66.67 | 79.31 | 70.00 |
| Mean no. of PiP calls | 3.50 | 33.85 | 3.52 | 3.69 | 2.80 |
| Proportion of veterans who received a pain modulatory intervention at baseline (%) | 93.91 | 93.02 | 75.79 | 95.65 | 100.00 |
| Measure | Site 1 | Site 2 | Site 3 | Site 4 | Site 5 |
|---|---|---|---|---|---|
| Time in trial (mo) | 25 | 25 | 28 | 23 | 23 |
| No. of clinic visits with documented LBP diagnosis | 1446 | 1278 | 1012 | 2331 | 927 |
| No. of AIM-Back referrals issued | 179 | 179 | 158 | 148 | 37 |
| No. of enrolled veterans | 115 | 129 | 95 | 69 | 14 |
| Rate of referral to enrollment (%) | 64.25 | 72.07 | 60.13 | 46.62 | 37.84 |
| Proportion of those enrolled with ≥1 PA call (%) | 79.13 | 87.60 | 73.08 | 79.03 | 100.00 |
| Mean no. of PA calls | 2.81 | 3.13 | 2.46 | 2.47 | 3.00 |
| Proportion of enrolled veterans with SBST medium to high risk at baseline (%) | 89.57 | 86.05 | 76.92 | 90.32 | 100.00 |
| Proportion of those with SBST medium to high risk at 6 wk and ≥1 PiP call (%) | 73.91 | 78.18 | 66.67 | 79.31 | 70.00 |
| Mean no. of PiP calls | 3.50 | 33.85 | 3.52 | 3.69 | 2.80 |
| Proportion of veterans who received a pain modulatory intervention at baseline (%) | 93.91 | 93.02 | 75.79 | 95.65 | 100.00 |
aAIM-Back = Improving Veteran Access to Integrated Management of Back Pain; LBP = low back pain; PA = physical activity; PiP = psychologically informed practice; SBST = STarT Back Screening Tool.
Referrals Before and After Site Visits
The primary goal of in-person site visits was to increase referral rates after COVID-19-related travel restrictions were lifted. We observed a marked increase in AIM-Back referrals following on-site visits for the 3 clinics. Increases were due to a boost in referral rate from providers that had already been referring into the program, as well as an increase in the number of new providers referring to the program (Fig. 2). Specifically, we observed a 440% increase in referrals from 3 months prior to site visits to 3 months following site visits, with a total of 8 new referring providers, when considering recruitment from all 3 sites that received a visit.
Cumulative provider referrals into Improving Veteran Access to Integrated Management of Low Back Pain (AIM-Back) in the 3 mo before and after site visits across the 3 sites that received a visit.
Telehealth Use
As of the time of this analysis, 396 of the 422 enrolled veterans were 3 or more months past their initial on-site physical therapist visit and therefore had the opportunity to complete the pathway. We focus on this “completed” subset for the remainder of our analyses. Table 2 includes demographic characteristics of this sample. Of these, 322 (81.3%) completed at least 1 physical activity call (range = 73.1%–100% across sites). The mean number of completed physical activity calls per veteran was 2.8 (SD = 2.0; median = 3) out of a maximum of 6 possible calls. Figure 3 provides the distribution of physical activity calls. There was no statistical difference in the mean number of physical activity calls completed by veterans at low versus medium to high risk on the baseline SBST (2.65 [SD = 1.92] vs 2.85 [SD = 1.87], respectively; P = .47). Only 23 veterans (5.8%) completed all 6 physical activity calls (4 in the low-risk group, 19 in the medium- to high-risk group). Although veterans were given the option of telephone or videoconference call, nearly all telehealth visits occurred by telephone.
Demographic Information for Veterans Enrolled in AIM-Back at Least 3 Months After Initial Physical Therapist Evaluation On-Site (n = 396)a
| Characteristicb | Value |
|---|---|
| Age, y | |
| Mean (SD) | 52.3 (15.5) |
| Range | 21–90 |
| Sex (reported as man/woman) | |
| Men | 344 (86.9) |
| Women | 52 (13.1) |
| Ethnicityb | |
| Not Hispanic or Latino | 337 (96.8) |
| Hispanic or Latino | 11 (3.2) |
| Raceb | |
| White | 232 (62.5) |
| Black or African American | 134 (36.1) |
| American Indian, Alaska Native, Native Hawaiian, or other Pacific Islander | 5 (1.4) |
| High-impact chronic pain (HICP)c status at baselineb | |
| Yes | 329 (83.5) |
| No | 65 (16.5) |
| Characteristicb | Value |
|---|---|
| Age, y | |
| Mean (SD) | 52.3 (15.5) |
| Range | 21–90 |
| Sex (reported as man/woman) | |
| Men | 344 (86.9) |
| Women | 52 (13.1) |
| Ethnicityb | |
| Not Hispanic or Latino | 337 (96.8) |
| Hispanic or Latino | 11 (3.2) |
| Raceb | |
| White | 232 (62.5) |
| Black or African American | 134 (36.1) |
| American Indian, Alaska Native, Native Hawaiian, or other Pacific Islander | 5 (1.4) |
| High-impact chronic pain (HICP)c status at baselineb | |
| Yes | 329 (83.5) |
| No | 65 (16.5) |
aData are reported as numbers (percentages) of enrollees unless otherwise indicated. AIM-Back = Improving Veteran Access to Integrated Management of Back Pain.
bMissing data: ethnicity, n = 48; race, n = 25; HICP status, n = 2.
cHICP was defined as reports of pain every day or most days in the past 3 mo and reports of pain limiting life and work every day or most days in the past 3 mo.
Demographic Information for Veterans Enrolled in AIM-Back at Least 3 Months After Initial Physical Therapist Evaluation On-Site (n = 396)a
| Characteristicb | Value |
|---|---|
| Age, y | |
| Mean (SD) | 52.3 (15.5) |
| Range | 21–90 |
| Sex (reported as man/woman) | |
| Men | 344 (86.9) |
| Women | 52 (13.1) |
| Ethnicityb | |
| Not Hispanic or Latino | 337 (96.8) |
| Hispanic or Latino | 11 (3.2) |
| Raceb | |
| White | 232 (62.5) |
| Black or African American | 134 (36.1) |
| American Indian, Alaska Native, Native Hawaiian, or other Pacific Islander | 5 (1.4) |
| High-impact chronic pain (HICP)c status at baselineb | |
| Yes | 329 (83.5) |
| No | 65 (16.5) |
| Characteristicb | Value |
|---|---|
| Age, y | |
| Mean (SD) | 52.3 (15.5) |
| Range | 21–90 |
| Sex (reported as man/woman) | |
| Men | 344 (86.9) |
| Women | 52 (13.1) |
| Ethnicityb | |
| Not Hispanic or Latino | 337 (96.8) |
| Hispanic or Latino | 11 (3.2) |
| Raceb | |
| White | 232 (62.5) |
| Black or African American | 134 (36.1) |
| American Indian, Alaska Native, Native Hawaiian, or other Pacific Islander | 5 (1.4) |
| High-impact chronic pain (HICP)c status at baselineb | |
| Yes | 329 (83.5) |
| No | 65 (16.5) |
aData are reported as numbers (percentages) of enrollees unless otherwise indicated. AIM-Back = Improving Veteran Access to Integrated Management of Back Pain.
bMissing data: ethnicity, n = 48; race, n = 25; HICP status, n = 2.
cHICP was defined as reports of pain every day or most days in the past 3 mo and reports of pain limiting life and work every day or most days in the past 3 mo.
Distribution of physical activity calls completed between the veteran and the central provider among veterans participating in Improving Veteran Access to Integrated Management of Low Back Pain (AIM-Back) after the 3-mo follow-up time point (n = 396).
At initial on-site physical therapist visit, 342 of 396 veterans (86.4%) screened as medium to high risk on the SBST (range across sites = 76.9%–100%). Of the 396 veterans, 266 (67.2%) returned for an on-site follow-up physical therapist visit and completed the SBST, with a mean time between initial evaluation and on-site follow-up of 50.0 (SD = 15.4) days (range = 13–180 days). Of those, 167 veterans (62.8%) screened as medium to high risk on the SBST and were subsequently offered PiP interventions by the central provider. Figure 4 provides the distribution of PiP calls. The mean number of completed PiP calls per veteran was 2.5 (SD = 2.0; median = 3) out of a maximum of 6 possible calls, with 74.9% completing at least 1 PiP call (range = 66.7%–78.2% across sites). Fifteen veterans completed at least 5 physical activity and 5 PiP calls, while only 1 veteran completed all 6 physical activity and 6 PiP calls (Suppl. Tab. 1).
Distribution of psychologically informed practice (PiP) calls completed between the veteran and the central provider among veterans with medium to high risk on the STarT Back Screening Tool (SBST) at the 6-wk follow-up time point (n = 167).
Interventions Provided by Physical Therapists
At the initial on-site physical therapist evaluation and consistent with the core component of pain modulation, TENS was used with 5.2% of veterans in the clinic, and 40.6% were instructed in TENS use at home. Manual therapy to soft tissue was used with 5.4% of veterans, and joint mobilization/manipulation was used with 15.7% of veterans. Other interventions at initial evaluation were reported in text fields within the notes and included delivery of therapeutic exercise, dry needling, postural education, thermal modalities, and self-mobilization techniques. Altogether, 92.2% of veterans received some form of pain modulation during the initial on-site physical therapist visit. Physical therapists used pain neuroscience education with 155 veterans (39.1%).
At the 6-week follow-up, physical therapists used these interventions less frequently: TENS in the clinic (3.4%), joint mobilization/manipulation (12.4%), and soft-tissue manual therapy (6.8%). No specific intervention was provided for 95 veterans (35.7%) at follow-up. Pain neuroscience education was delivered to 79 veterans (29.7%) at the 6-week follow-up.
PiP Interventions Delivered by the Central Provider
For the 74.9% (n = 125) of veterans who were eligible for PiP and received at least 1 PiP call, 100% received communication strategies at some point in their program, 97.6% received pain coping skills training, 89.6% received activity-based treatments, and 99.2% received education in a home program. Of these, 88.8% (n = 111) received all 4 PiP strategies at some point during their treatment. Like the physical activity calls, nearly all PiP telehealth calls occurred by telephone, rather than video call, at the veterans’ requests.
Discussion
Real-world use of hybrid telehealth care models initiated in primary care and involving other disciplines is not well documented in the literature. These models, whether they feature physical therapists or not, are likely to have unique use profiles and implementation challenges. An important finding is the wide variation in completion rates of telehealth calls and broad range of rates of referral to enrollment across clinics (37.8%–72.1%). The degree to which variation was a result of the program’s hybrid design is difficult to assess, especially due to the concurrent impact of COVID-19 on health care use and access. However, we would expect patterns that reflect the natural variability of real-world care given the embedded, pragmatic nature of the trial. Lower participation rates for individual components of hybrid care pathways in pragmatic clinical trials, when compared to randomized controlled trials, are consistent with prior studies of a pragmatic design emphasizing telehealth44 and was observed for on-site treatment in the TARGET trial in LBP.45
The findings from our study further inform the field because there are limited data describing real-world uptake of hybrid clinical pathways—not just specific treatments—for LBP or other common chronic conditions. To date, it is unclear whether the variability we observed needs correction or simply reflects what we should expect from a hybrid pathway implemented in a highly pragmatic way. Of note, this variability should not impact the primary analysis for pathway effectiveness because our design incorporated elements that increase the chance that clinic characteristics are similar for both treatment pathways (eg, covariate constrained randomization). Secondary analyses (eg, heterogeneity of treatment effects) on the completed trial data will provide further insight on whether variability influences outcomes and therefore should be addressed. In particular, we would want to know if there is a dose–response relationship between the number of completed calls and better outcomes for physical function and pain interference. If the number of calls is associated with trial outcomes, it raises the question of whether future efforts should focus on reducing the intensity of the program (ie, fewer touchpoints) to make participation easier, or on efforts to improve implementation of the current program.
Although our study best positions us to describe variability observed in implementing this pathway for an ePCT, it can also inform future clinical implementation of hybrid care pathways for LBP and other chronic noncommunicable conditions. Currently, there are very little data by which to set benchmarks for care use in hybrid models, and there is no consensus on minimum content, frequency, or duration parameters for effective psychologically informed treatment.46 By highlighting the variability inherent in delivering hybrid care, our data indicate what may be considered acceptable by providers and veterans in the VHA, with an average of approximately 3 completed calls for the physical activity and PiP content each. Moreover, we observed a far greater preference by veterans for telephone calls versus video calls. Anecdotally, many veterans experienced challenges setting up the videoconferencing offered by the VHA and preferred the easier telephone option. These findings may be useful for designing hybrid care models that will evaluate their effectiveness in future ePCTs. These data may also help set standards for referral rates and variation in care that are indicative of acceptable implementation of hybrid care models involving multiple providers.
We established a priori core components of the pathway based on recommended pragmatic clinical trial design features.5,43 Adherence to delivering core components was generally good. Use of pain modulation interventions was high, with over 90% of veterans receiving some form of treatment. Over 80% completed at least 1 physical activity training call. The follow-up rate to on-site care was 67%, with a follow-up time frame of approximately 7 weeks (SD = 15 days) (compared to a prescribed 6 weeks). Among those who returned to clinic and were identified as medium to high risk, 3 of 4 veterans completed at least 1 PiP call and almost 90% of those completing at least 1 call received all 4 PiP strategies at some point during their treatment.
We did not track specific reasons for failure to complete all planned calls or missed on-site visits as specified by the program, therefore can only speculate on these reasons. Some appointments were administratively discontinued (eg, referrals were inappropriate or administered in error). In other cases, central providers were unable to reach veterans by phone. Central providers attempted at least 1 or 2 calls per week for both the physical activity and the PiP portions of the pathway. When possible, veterans would receive voicemails and secure messages with instructions on how to contact the central provider. Other reasons for missed visits may have included veterans experiencing improvements or worsening of their symptoms that led them to no longer need care, or seek escalation of care, respectively. In some cases, veterans may have decided to pursue a different option for their care. Veterans may have also decided that care delivered by phone or video was not preferable or did not meet their needs. Central providers gave veterans the choice of telehealth by phone or video, with the majority requesting care by phone. Although suitable for physical activity guidance and psychological interventions, telehealth by phone may provide more limited opportunities than in-person or video-based care to build strong patient–provider relationships. Though speculative, this could have resulted in suboptimal participation.
Less than 10% of clinic visits for veterans with a documented LBP diagnosis seeing an AIM-Back trained provider resulted in an AIM-Back referral. However, when a referral was issued, nearly 2 out of every 3 resulted in enrollment into the program. The optimal referral rate is presently unclear, and the lack of similarly designed pragmatic trials in this population makes it challenging to contextualize this finding. Low primary care referral rate was a challenge in the aforementioned TARGET trial, with 36% of high-risk veterans referred for PiP delivered by physical therapists.45 ePCTs in other settings and health conditions show highly variable referral rates among potentially eligible participants, from 7.9% to 81.0%.44,45,47,48 The observed referral rate in our study and others underscores the challenges associated with implementing new clinical pathways, even when designed with pragmatic features intended to ease provider burden and streamline access to guideline concordant care. These challenges persisted in AIM-Back despite significant and purposeful PCP engagement efforts, and despite the program having been designed with input from VHA clinicians, veterans, and caregivers.
Low referral rates could also be a function of our site training and engagement strategy, which was entirely virtual through the first year of the program due to COVID-19 travel restrictions. On-site visits help researchers develop rapport with clinical delivery teams, understand site-specific nuances of care delivery, identify barriers, troubleshoot problems, and build enthusiasm for the program. These are difficult to replicate in a virtual format. Although we designed our virtual trainings with these goals in mind, we found the quality of virtual engagement inferior to in-person interactions for these purposes. To this point, referral rates improved significantly following our 3 site visits. Ongoing qualitative interviews with AIM-Back PCPs will help us uncover barriers to program implementation. Specifically, we are interviewing AIM-Back PCPs with high and low referral rates to identify ways to improve our training and engagement processes, understand PCP perspectives on the program, identify workflow constraints that limit referrals, and gain insight on how PCPs describe the program to patients. We are collecting this information to inform future implementation studies of AIM-Back to ensure the program is deemed a feasible and acceptable treatment option for pain.
A notable finding was the large increase in AIM-Back referrals following site visits. We believe the biggest reason for improvements in referrals post-site visit was due to the higher quality of engagement we had with providers and delivery teams in person versus through telephone or video meetings. We often had limited opportunities to promote the program or answer questions when joining monthly clinic meetings virtually. On site visits, we had a more captive audience of providers and PCPs, which allowed us to develop rapport with PCPs, answer questions about the consult process, clear up misconceptions about the program, troubleshoot problematic workflows, and promote AIM-Back. One benefit of on-site visits after AIM-Back was underway at a site was that we could see where clinics were struggling with implementation, and then design the site visit around engagement of the most relevant parties to efficiently troubleshoot strategies. Future trials may benefit from initial standard site visits to develop rapport and train providers, with more tailored follow-up site visits occurring later to address site-specific challenges.
Limitations
Readers should consider our findings in the context of a few limitations. First, we are conducting this trial in VHA, which includes a population impacted more by musculoskeletal pain than nonveteran populations.49 This impact was supported in our trial by the high percentage of veterans with medium to high SBST risk classification and HICP at baseline. Health care use patterns reported with this hybrid care pathway may differ if delivered outside of the VHA, or among patients not enrolled in primary care. Second, although AIM-Back did not cover expenses associated with treatment, the VHA has different payment structures, policies, and priorities that impact pathway implementation and health care use compared to non-VHA settings. Similarly, the VHA employs an electronic health record that permits tracking of health care use and outcome measures and allows remote write access for central providers. These capabilities significantly enhanced our ability to make this program truly embedded, which may not be the case in other health systems. Third, this program was initiated from primary care, thus we cannot generalize these findings to care initiated from specialty providers (eg, orthopedics) or in other settings (eg, emergency departments). Finally, we did not assess the acceptability of PiP by veterans in this study. If veterans felt stigmatized by receiving psychologically focused treatment, or felt the treatment would not be helpful for their pain condition, it could have contributed to low completion rates. Future work should study methods to optimize acceptability of PiP interventions in veteran populations.
Conclusion
Experiences with the AIM-Back trial highlight the variation inherent in implementing new hybrid models for LBP that are initiated by primary care referral and include on-site care by physical therapists along with telehealth delivered by specially trained central providers. Pathway metrics observed in this study may help determine if this clinical program for LBP can be scaled to other VHA clinics not involved with AIM-Back and inform implementation efforts of future hybrid care models for other common chronic conditions.
Author Contributions
Trevor A. Lentz (Conceptualization, Data curation, Investigation, Methodology, Project administration, Supervision, Writing—original draft, Writing—review & editing), Cynthia J. Coffman (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing), Tyler Cope (Investigation, Project administration, Writing—original draft, Writing—review & editing), Zachary Stearns (Investigation, Project administration, Writing—original draft, Writing—review & editing), Corey B. Simon (Conceptualization, Investigation, Project administration, Supervision, Writing—original draft, Writing—review & editing), Ashley Choate (Conceptualization, Investigation, Project administration, Supervision, Writing—original draft, Writing—review & editing), Micaela Gladney (Project administration, Writing—original draft, Writing—review & editing), Courtni France (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Validation, Visualization, Writing—original draft, Writing—review & editing), S. Nichole Hastings (Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing—original draft, Writing—review & editing), and Steven Z. George (Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing—original draft, Writing—review & editing)
Ethics Approval
Activities to support AIM-Back implementation were considered to be nonresearch operations activities as defined in VHA Handbook 1058.05. Individual-level outcome measurement was approved as research by the Institutional Review Board of the Durham Veterans Affairs.
Funding
This work is supported through cooperative agreement UH3 AT009790 from the National Center for Complementary and Integrative Health (NCCIH). This study was also supported in part by the Center of Innovation to Accelerate Discovery and Practice Transformation (CIN 13–410) within the Durham Veterans Affairs Health Care System and by the Duke Claude D. Pepper Older Americans Independence Center (Hastings, NIA P30AG028716). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or U.S. Department of Veterans Affairs.
Clinical Trial Registration
This study was registered in the ClinicalTrials.gov registry (NCT04411420).
Data Availability
Data are available from the author upon reasonable request.
Disclosures
The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.
For more information about the AIM-Back study, visit https://sites.duke.edu/aimback/learn-about-aim-back/. The AIM-Back study is part of the NIH-DOD-VA Pain Management Collaboratory. For more information about the Collaboratory, visit https://painmanagementcollaboratory.org/.
Comments