Cohort Profile: Baseline Characteristics of Veterans from Improving Veteran Access to Integrated Management of Back Pain (AIM-Back) – an Embedded Pragmatic, Cluster Randomized Trial in the United States

Study Design

AIM-Back (registered at clinicaltrials.gov: NCT04411420) is a multisite, ePCT that compares the effectiveness of two clinical care pathways for Veterans with LBP within the Veterans Health Administration (VHA) Health Care System. The AIM-Back trial has been described previously in our protocol paper, including a description of its pragmatic components.[11] Both pathways were initiated by a referral from a trained primary care provider and are briefly summarized below. The Sequenced Care Pathway (SCP) started with an evaluation and treatment by a local VA physical therapist. Veterans were then offered six sessions of weekly telehealth physical activity calls through a remote centralized care provider. Veterans then returned for a follow-up assessment with their local physical therapist, which included risk screening for persistent disability using the STarT Back Screening Tool (SBST).[14] Those identified as medium or high risk on the SBST received an additional six weeks of remote care focused on pain coping skills training delivered by physical therapists trained in psychologically informed care. Additional details for the SCP and metrics for its usage have been previously described.[15]

The Pain Navigator Pathway (PNP) involved referral of the Veteran to a Pain Navigator, i.e., a local VA provider that worked with Veterans primarily via telehealth to identify appropriate non-pharmacologic treatment options, decide on the preferred option(s), and coordinate access to selected service(s). Veterans in the PNP had follow-up appointments with their Pain Navigator at 6 and 12 weeks post-initial encounter. Additional details for the PNP and metrics for its usage have been previously described. [16]

The 4-item short form PROMIS® Physical Function and Pain Interference were the primary outcome measures for the AIM-Back study. A selected set of additional outcomes were collected to further inform the effectiveness of the program.[11]

Patient and Public Involvement

The SCP and PNP were reviewed and modified in the planning stages of this trial based on input from our Veteran Research Engagement Panel (VetREP) as well as other partners including administrators, and clinical providers. Details on this process have been described in a prior publication.[17]

Participating Clinics Eligibility, Recruitment, and Randomization

Veterans Affairs clinics were recruited to participate in AIM-Back in two waves from February 2020-September 2020 and from March 2021-December 2021. Primary Care clinics were eligible to participate in AIM-Back if they could provide clinical personnel to deliver either of the treatment interventions and had seen between 800 and 5000 unique patients diagnosed with LBP in the previous year, with two exceptions decided by study team as allowable during the site recruitment phase. The first exception was for a rural serving clinic with LBP visits slightly below 800 (n=733) to help diversify the location of our participating sites. The other exception was made for a clinic that had >5000 visits because we identified specific patient aligned care teams (PACT) within their primary care clinic to participate and maintain an appropriate eligible participant volume. The number of primary care personnel trained to refer to AIM-Back during the trial ranged from 4-69 per enrolled clinic.

The AIM-Back trial received site participation agreements from and randomized 19 VA Primary Care clinics to one of the two pathways. Two clinics withdrew in the early stages of rolling out the AIM-Back program and data from these clinics are excluded from all analyses. The 17 clinics resided in ten different states: West Virginia, South Carolina, Nevada, Missouri, Texas, Pennsylvania, North Carolina, Kansas, Ohio, and Kentucky. Participating clinics were randomized into two blocks using a blocked covariate constrained randomization [11] and, after randomization, each clinic underwent training on how to implement their assigned pathway. Block 1 consisted of ten clinics (five in each pathway) that were randomized in September 2020, with the first clinic actively referring Veterans in February 2021. Block 2 randomization was completed in December 2021 and included five PNP clinics and four SCP clinics, with the first clinic referring Veterans in March 2022. The 2 clinics that withdrew (described above), one from each pathway, were both from Block 2.

Cohort and Sample Definition

To be eligible for AIM-Back, Veterans were referred to the randomized pathway at their clinic by trained primary care providers across 17 clinics (9 PNP; 8 SCP) between February 1, 2021 and January 31, 2024 (Referral Cohort). Veterans from the Referral cohort (n=2767; 1481 PNP; 1286 SCP) that participated in the study included: 1) an Electronic Health Record (EHR) sample of Veterans (n=1817; 1006 PNP, 811 SCP) enrolled in the program (i.e., attended initial AIM-Back visit with no consent required), approximately 66% of the Referral cohort, and 2) a Survey sample of Veterans (n=996; 480 PNP; 516 SCP) that agreed at the AIM-Back referral to be contacted about the research survey study were consented and completed baseline surveys (see Figure 1). Approximately 80% of the Referral cohort agreed to be contacted for the survey study (n=2183), 45.6% of whom were included in the survey sample There were n=799 Veterans in the EHR sample (44%) that were also in the Survey sample (see Figure 1); we planned for approximately 50% of the EHR sample to be in the Survey sample in the design of the study.

• Download figure
• Open in new tab

Figure 1.

AIM-Back Baseline CONSORT

Each clinic had a targeted enrollment of 105 Veterans for the EHR sample with a minimum enrollment target of 65 and a maximum of 130 Veterans to minimize cluster imbalance; for the Survey sample the target enrollment was n=54 per clinic.[18] Trial enrollment concluded in January 2024.

Providers at participating clinics were trained to refer Veterans to AIM-Back who had low back pain, received primary care at a randomized clinic, had a valid telephone number, were not receiving or had not been referred to hospice or palliative care, and were informed that AIM-Back was best suited for Veterans for whom conservative care was appropriate. Individual consent was not obtained for Veterans in the EHR sample to participate because each clinic had agreed to deliver the guideline-concordant AIM-Back program as part of their usual clinical care; therefore, the care delivery was not considered research. To be eligible to be contacted about participating in the Survey sample from research staff, Veterans referred to AIM-Back had to agree, at the referral, to be contacted for the research study that collected more granular data in surveys with additional longer follow-up time points. Further eligibility criteria for the survey sample included not being in institutional care (nursing home or hospital), no cognitive impairment, dementia, or lack of decision-making capacity, no serious mental illness (diagnosis of schizophrenia, bipolar disorder, psychiatric hospitalization) in the previous year or no current high-risk suicide flag in their medical record, and available access to, with ability to communicate on, a telephone. After confirming some of the additional eligibility criteria via an EHR chart review, the study team attempted to call to further screen for any remaining eligibility and, if eligible, obtained consent and administered the baseline survey. Further details on survey enrollment are detailed in the protocol paper.[11]

Methods of Data Collection

The AIM-Back trial had two data sources: 1) data entered through the Computerized Patient Record System (CPRS), the Veterans Health Administration Health Care System (VHAHCS) EHR, and extracted from the Corporate Data Warehouse (CDW) and 2) data collected via telephone by AIM-Back research staff.

Primary care clinicians referred patients to the pathway randomized to their clinic through the EHR. Once an AIM-Back referral was issued, clinics’ usual care protocols were followed to schedule the baseline intervention visit by an AIM-Back trained physical therapist or pain navigator (depending on the pathway to which the clinic was randomized). Demographic information (age, race, ethnicity, and sex), opioid use based on a prescription fill in the year prior to referral, PTSD diagnoses in year prior to referral, and Care Assessment Needs (CAN) scores prior to referral date (closest)[19] were extracted from CDW for the Referral cohort (see list of drugs for opioid use and ICD10 codes for PTSD diagnosis in Supplemental Materials). CAN score is a comorbidity measure and is the estimated risk for hospital admission or death within a year calculated weekly for all eligible Veterans. The score is the risk percentile and ranges from 0 (lowest risk) to 99 (highest risk). During the initial evaluation, the provider used a clinical note template specific to the clinic’s assigned AIM-Back pathway to collect data that included the 2-item CDC high impact pain questionnaire [15], the 2-item NIH Task Force on Research Standards for Chronic Low Back Pain [20], the 4-item PROMIS Pain Interference [21], the 4-item PROMIS Physical Function [22], the 4-item PROMIS Sleep Disturbance[23], and other relevant care notes. The data from these EHR templates are stored as structured text fields and uploaded nightly to CDW.

For the survey sample, patient demographics (baseline only), PROMIS pain interference, PROMIS physical function, PROMIS sleep disturbance, CDC and NIH definitions of HICP, Catastrophizing, PEG pain intensity, Self-efficacy, Depression (PHQ-2, AUDIT-C, Quality of life (EUROQOL), and Nonpharmacological and Self-Care Approaches (NSCAP)[24,25] measures were collected. Data were collected electronically into REDCap [26] at baseline, 3, 6, and 12 months.

In addition, to evaluate representativeness of the Referral Cohort for each clinic, we identified all primary care LBP visits based on ICD-10 codes (see list in Supplemental Materials) during the study active periods at each clinic and extracted demographics (age, race, ethnicity, and sex) for the Veterans with the LBP visits.

Strengths and Limitations

The findings reported here indicate that the AIM-Back trial has several strengths, including a) having data from the Referral Cohort to understand provider and patient engagement with these pathways; b) the enrollment goal was met, which allows for adequate power of all planned analyses; c) the use of both EHR and survey samples provide for a comprehensive assessment, allowing for future robust evaluation and sensitivity analyses of trial findings; and d) the integration of AIM-Back features within existing clinical settings demonstrates the potential for these care pathways to be implemented more broadly in Veteran Health Administration clinics and in other healthcare environments outside the Veteran Health Administration. Additionally, since cluster randomization can result in imbalances, it is encouraging to see balance in our primary and secondary outcome measures at baseline across care pathways.[18]

Some of the limitations of the trial include recruitment during the COVID-19 pandemic that affected practice patterns for clinics participating during shutdowns compared to care delivered pre- and post-shutdown. Leveraging the EHR as a data collection tool presented challenges related to consistent documentation which may impact our analysis, such as the PPR, that is reliant on ICD-10 code documentation for LBP visits which is known to be inconsistently used across healthcare systems and personnel. [28] Lack of appropriate diagnostic documentation may miss detection of Veterans eligible to receive AIM-Back. An important consideration for future research and implementation efforts is to further explore participation among older Veterans to overcome the current limitations of the trial’s generalizability due to the lower referral of this subgroup to the AIM-Back program.

Collaboration

Deidentified data will be made available upon request and permission for use granted consistent with VA policy via inquiry to vhaduraimback{at}va.gov.