Speakers

Michael Ho, MD, PhD
University of Colorado

Sheana Bull, PhD
University of Colorado

Slides

Keywords

Nudge, NIH Collaboratory Trial, Cardiovascular, Adherence

Key Points

• A nudge is a strategic reminder that can help people adopt healthy behaviors, using principles from behavioral economics and cognitive psychology to improve behavior commitments, communicate social norms, and narrative stories. The Nudge study wanted to find out if nudges could help patients improve medication adherence by promoting behavioral change through positive reinforcement.

• The Year 1, UG3 Phase, aims were to develop and program a nudge message library using iterative N of 1 studies to optimize content for a range of diverse patients. The second aim was to conduct a pilot intervention to demonstrate feasibility of delivering the intervention and preliminary effects in a 3 healthcare systems. The idea was to engage the patient, providers, and healthcare system leaders to design, refine, and implement the pilot.

• In Years 2-5 (UH3 Phase) the aims were to conduct and evaluate a pragmatic patient-level randomized intervention across 3 healthcare systems to improve adherence to chronic cardiovascular (CV) medications. The second aim was to evaluate the intervention effectiveness using a mixed methods approach and apply the RE-AIM (reach, effectiveness, adoption, implementation, and maintenance) framework.

• The study recruited adult patients diagnosed with 1 or more cardiovascular condition of interest, prescribed at least 1 or more medication of interest, with a refill gap of at least 7 days. Patiens were excluded if they had neither a landline or cell phone, were enrolled in hospice or palliative care, were non-English or Spanish speaking, and residing out of the state of Colorado.

• Three healthcare delivery systems participated in the study: UC Health Clinics, Denver Health, and VA Eastern Colorado HCS Clinics.

• Nudge used an opt out approach, with IRB approval because it was a low-risk intervention and having patients consent on a large scale would be challenging and might bias the sample to more adherent patients. Nudge sent an opt-out packet with an information sheet, an opt-out sheet and a self-addressed stamped envelope. There was also an option to complete a survey to share reasons for opting out. There were 4 weeks to return the opt-out form, and there was a text opt-out option once the program started.

• More than 13,000 patients were considered eligible for the UH3 trial and more than 9,500 were randomized in the 4 arms, with about 2,300 patients in each of the arms. The 4 arms were usual care, generic text message arm, optimized text message arm that used behavioral theories that aligned with “nudge,” and a fourth arm that used a chat bot.

• All of the intervention arms had a significant influence on refills but there was no different between the intervention arms. Patients could text back with clinical questions or questions about the study. About 50% of participants responded with a question, and 9.2% responded with a question for a clinical pharmacist.

• Nudge also deployed a text message and survey to participants to get a sense for patient satisfaction with feedback on the study and text messages. The response rate was low but in general there were high levels of satisfaction with the intervention. No concerns were raised about privacy concerns or confusion about the messages.

• Lessons learned: The text message intervention improved medication refill adherence, the effectiveness was consistent regardless of the type of message delivered, and the differential impact of the intervention across some patient groups. The opt out approach is feasible and drop out rate was lower than expected. Patients were generally satisfied with text messages, and patients who were non-white and Hispanic were more likely to remain in the study. Text messages can be implemented as a low-cost intervention to improve medication refill adherence across different healthcare systems with disperate EHR systems.

Discussion Themes

-Do you think the results support a text message nudge in most clinical trials and what is the cost effectiveness of this? This is a low-cost strategy and would be a great option to use as one of multiple strategies to use in clinical trials. The nudge can be very generally defined, the timing of the message for this kind of intervention is important. It’s not clear that this messaging may not be sufficient for other types of behavior, like changing diet, but for this type of action it was sufficient.

-What was your IRB? Our IRB was the Colorado Multiple Institutional review board. Messages included PHI but the study followed the institutional security protocols that allowed the intervention to be approved.

–How did the messages account for prescription dosage changes? Patients could reply with Done and the message reminders would stop. The study would then recalculate the gaps.

Tags

#pctGR, @Collaboratory1

Speakers

Christine Goertz, DC, PhD
Professor in Musculoskeletal Research
Vice Chair for Implementation of Spine Health Innovations
Department of Orthopaedic Surgery
Duke University 

Adam Goode, DPT, PhD
Associate Professor in Orthopaedic Surgery
Vice Chair of Clinical Research and Epidemiology
Department of Orthopaedic Surgery
Duke University

 

Slides

Keywords

Pragmatic trials; research; Low Back Pain; Primary Care; Physical Therapy; Chiropractic

Key Points

• The IMPACt-LBP trial’s project goal is to implement the American College of Physicians low back pain (LBP) guideline by evaluating the impact of the Primary Spine Practitioner (PSP) model in three academic Health Care Systems (HCS). The study then evaluates the model’s effectiveness by comparing the intervention to usual medical care.
• The American College of Physicians guideline states that “physicians and patients should treat acute, sub-acute and chronic LBP with non-drug therapies.” The Primary Spine Provider (PSP) model is a multi-disciplinary collaborative care model that includes Doctors of Chiropractic (DC) and physical therapists (PT) as first line care for LBP. Their treatment approaches will focus on non-pharmacological approaches recommended by the guideline, including spinal manipulation and exercise.
• The IMPACt-LBP study design is a pragmatic, multi-site, two-arm cluster-randomized trial with randomization at the primary care provider (PCP) clinic level. It includes 22 family medicine, primary care and general internal medicine clinics affiliated with Duke University, the University of Iowa or Dartmouth-Hitchcock health systems. The study team plans to enroll 1,800 adult patients with a primary complaint of LBP who contacted a participating primary care clinic to make an appointment with a primary care provider.
• The study pathway involves a patient contacting a primary care clinic and then being transferred to an IMPACt-LBP scheduling assistant who provides a scripted introduction to the study and a baseline survey. Patients in the usual care arm clinics are then scheduled to see a PCP. Patients at the intervention care arm clinics are introduced to a DC or PT appointment option, while still being able to select a PCP appointment.
• The study’s primary endpoints are a change in PROMIS physical function from baseline to 3 months and change in PROMIS pain interference from baseline to 3 months. The study is also composed of an enrolled cohort of patients that agree to complete PRO surveys for pain interference and functional status for primary analysis. There is also a longitudinal cohort of EHR data for all patients at participating PCP clinics to assess utilization and other broad outcomes.
• The study has enrolled 187 patients since enrollment began in March 2023, and the study team has pulled together its initial lessons learned.
• In terms of administrative and regulatory lessons learned, the team reflected that the integration of PT/DC at the forefront of the patient experience is often different from many typical processes. There have also been hiring challenges related to the current job market, as well as regulatory challenges related to consent waivers and IRBs.
• Presenters also shared challenges regarding scheduling and patient insurance coverage. While not anticipated, it was also necessary to have dedicated research staff to ensure adequate recruitment. Additionally, the study team noted that patients do not all enter the health system the same way, which is an ongoing challenge.
• Differences in Cauda Equina screening by sites, which required a protocol amendment, proved to be another challenge.
• There were also lessons learned regarding data collection, as many sites are not collecting and integrating PRO outcomes to a sufficient degree for the study. The study is now using two REDCap systems to collect data, which increases the potential for data entry errors.
• From a clinical perspective, the team learned the importance of keeping providers up to date on the study progress in an efficient and streamlined way and experienced unforeseen difficulties with engaging with community PSPs.
• Patient-specific issues encountered include: patients don’t want to be approached about research during an acute pain episode; many patients have concerns about co-pays to see PT/DC; sometimes patients have multiple reasons for a PCP visit; and patients are sometimes frustrated after going through the HCS phone triage system

Discussion Themes

–How has the experience been regarding dual oversight and a single IRB? There can be many challenges related to dual oversight and a single IRB. While the single IRB streamlines some steps, there are still different and specific processes at the different sites, which can create challenges and hurdles to overcome. With IMPACt-LBP, we have been able to move forward in an efficient manner; however, we have had a few additional steps that were not initially anticipated. 

–Is working with large academic health systems a help or a hinderance? Academic health centers have the potential to drive change, so if we were able to pull it off at these sites, we felt that these interventions could be scaled to many different settings. We are learning together in real-time because we are trying to change how people think about clinical practice for low back pain. That is a real challenge. High-level administrators and clinicians are recognizing that there is a huge opportunity to provide better care for people with LBP and that we are attempting to implement evidence-based care pathways. Having this high-level support is critical and allows us to problem solve in real-time.

–Are you looking to engage with patients who have had imaging, and how does that impact the care path? We do have some sub-group analyses plans for the trial, not specific to imaging, but it is something we could consider.  Inappropriate early imaging is included in our outcome measures, but we could consider some deeper work and analyses.

–Do you think that this intervention is sustainable? We think so because we are building a playbook for this. There is continued discussion as a nation about how we treat patients for low back pain, so we think there is potential. People care about this issue, so we feel that developing this evidence is a critical next step to promote change for better care. 

Tags

#pctGR, @Collaboratory1