Lessons from the implementation of a trauma center-based program to support primary care providers in managing opioids and pain after trauma hospitalization

Collaborative Opioid Taper After Trauma study

The Collaborative Opioid Taper After Trauma (COTAT) study is a clinical trial that randomized patients discharged from a level I trauma center after trauma to usual care or to a program in which their PCPs were supported by a PA pain specialist in managing the patients’ opioid tapers for their pain. Patients recruited to the study were 18 or older, spoke English or Spanish, were admitted to the level I trauma center with an Injury Severity Score of 4 or greater,11 and were discharged to their home outside the trauma center’s county. The study intended to recruit equal numbers of patients who were opioid naïve and opioid tolerant (ie, using prescribed opioids in the month before trauma), but the study was unable to do so given challenges that COVID-19 posed on study recruitment. Patients with OUD or evidence of illicit drug use in the past month, active cancer, palliative or hospice care, or serious psychiatric disorders (psychotic symptoms, or a psychiatric hospitalization or suicide attempt in the past year) were excluded. Of the 73 patients providing consent for the trial, 37 were randomized to the PA support intervention for their PCPs; three of these were opioid tolerant.

Post-trauma hospitalization opioid taper support program

The study intervention or program consisted of an advanced practice provider, namely a PA (LAK), with expertise in pain management to serve as a supportive resource over the phone to a patient’s PCP about post-trauma pain management and opioid tapering for up to 5 months after hospital discharge. The PA had access to a psychiatrist with chronic pain expertise (MS), a family physician (L-MB), and a trauma surgeon (SA) for guidance when providing individual case advice to the PCP. The PA did not directly manage the patient’s care or prescriptions.

The following summarizes the support services planned by the PA:

• Contacting the patient within a few days of hospital discharge to review their discharge instructions and pain management plan, to confirm their PCP’s identity, to assist in identifying a PCP if the patient did not have one, and to encourage follow-up with their PCP.

• Faxing the patient’s discharge summary, discharge instructions, and a detailed study instruction sheet to the PCP within a few days of the patient’s discharge.

• Reaching out to the PCP by phone within the first week after the patient’s discharge to review the discharge instructions, to ensure that the PCP is aware of the support program and its offerings as well as how to reach the PA, and to determine the PCP’s preferences on frequency and preferred method of communication.

• Contacting the PCP’s office weekly for the first 2 weeks after trauma hospitalization discharge, then monthly for up to 5 months to see if the PCP had questions or wanted support in managing their patient’s pain or in tapering their opioid medication. After the first contact, the PA conducted additional consultations only when appropriate (eg, no further consultations were conducted for patients already tapered off opioids or not following up with their PCP).

The PA offered support such as:

• Contacting the hospital trauma team for questions about trauma recovery.

• Advising on the opioid taper plan if not proceeding as planned.

• Problem solving if the PCP had concerns about their patient’s pain management.

• Arranging a case presentation to a telehealth multidisciplinary pain specialist panel about the patient if the PCP desired additional advice.

Early in the support program, the PA discovered that the PCPs often had delegates (eg, medical assistants, nurses) who could receive information and support on their behalf, and these PCP delegates were contacted as alternatives to the PCP, depending on provider availability.

Data sources

We used five sources of data in this study:

1. Field notes. The PA recorded all interactions with patients, PCPs, and their delegates on forms that were managed using REDCap electronic data capture tools hosted by the Institute of Translational Health Sciences at the University of Washington.12 13 Each encounter with the PA was entered on a separate form. These forms included the ability to record answers to both closed-ended and open-ended questions. Forms completed after talking with the patient after discharge included questions related to their follow-up care plan, pain levels, and opioid taper, as well as the opportunity to record other issues that the patient reported in their health since trauma hospitalization. The PA also asked the patient to confirm or identify their PCP. After speaking with the PCP office, the PA recorded: whether the PCP identified themself as the patient’s provider, whether the patient had completed their opioid taper, and whether the PCP needed any support in caring for the patient, particularly related to opioid management.

2. PCP surveys. The study team sent a survey to identified PCPs or their delegates at the end of the intervention, asking six questions concerning the acceptability (eg, ‘I like the COTAT guidance’), appropriateness (eg, ‘The COTAT guidance seems applicable to my practice’), and feasibility (eg, ‘The COTAT guidance seems doable for my practice’) of the intervention (Supplemental Digital Content, online supplemental file 1). These questions were adapted from Weiner et al’s Acceptability of Intervention Measure, Intervention Appropriateness Measure, and Feasibility of Intervention Measure.14

3. Structured debrief with the PA. The PA was interviewed at the end of the pilot study using semistructured open-ended questions (Supplemental Digital Content, online supplemental file 2). The interview was audiotaped and transcribed, then reviewed by three study team members (L-MB, DBG, MS) to summarize modifications made to the intervention throughout the study, what went well and what the greatest challenges were in implementing the intervention, the responses of patients and providers/delegates to being contacted, the perceived benefits and challenges to the providers about the intervention, and the PA’s perspective on modifying the intervention or on other effective ways to improve opioid management after trauma hospitalization.

4. National Provider Identifier Standard (NPI) data. We used the NPI data to identify provider gender and specialty (family medicine, internal medicine, pediatrics). We used the hospital discharge summary, conversations between the PA and the patient, and online primary care clinic information to identify provider types (Doctor of Medicine [MD], Doctor of Osteopathic Medicine [DO], Advanced Registered Nurse Practitioner [ARNP], PA), practice ZIP code and practice type (eg, federally qualified health center [FQHC], hospital-affiliated clinic). We used rural-urban commuting area codes linked to the primary care clinic ZIP code to categorize practice location into three groups—urban, large rural city/town, and small rural town.

5. Electronic health record data. The electronic health record provided demographic information (ie, age, gender, language preference, residence location, insurance status) on the patients randomized to the intervention, whether or not the patients were admitted to the intensive care unit, hospital length of stay, and information used to calculate the Injury Severity Score.

Implementation outcomes

We included measures of five implementation outcomes defined by Proctor and colleagues15 16 and listed in table 1.

The implementation process was measured by identifying barriers, facilitators, and key lessons from the PA’s perspective.

Analysis

We conducted simple descriptive frequencies of patient and PCP characteristics and the discrete study measures obtained from the PA in the study database. We conducted a content analysis of the open-ended data recorded by the PA, as well as of the PA interview transcription to identify facilitators, challenges, and key learnings related to implementation of the intervention.

Patient characteristics

The 37 patients had a mean age of 45, were nearly three-quarters male, 95% white, 3% Hispanic, and almost all had an English language preference (table 2). Nearly half were publicly insured, 38% privately insured. They had a mean Injury Severity Score of 13, and almost a quarter spent time in the intensive care unit. The patients’ mean hospital length of stay was 4.5 days. Just over 40% of the patients lived in rural areas.

PCP characteristics

The 18 PCPs were largely family physicians, and nearly half were female (table 3). Nearly a quarter of the PCPs were advanced practice providers (three ARNPs, one PA) with the remainder MDs and DOs. Almost 40% of the PCPs’ practice locations were in rural locations. The largest proportion of PCPs practiced in FQHCs (33%), followed by independent practices and hospital/health system-affiliated clinics. The minority practiced in integrated health systems.

Adoption of and fidelity to the support program by the patients

The PA reached 32 (86%) of the 37 participating patients by phone after discharge from the trauma center (adoption), and when reached, the PA documented that they consistently discussed the planned content for the call (pain management, opioid and other medication management, trauma hospitalization follow-up, and making an appointment with a PCP; measure of fidelity). The PA documented that the vast majority of patients (26 of the 32 patients, 81%) discussed topics beyond the planned content, the most common being social context or financial issues (72%), more pain than expected (47%), opioid side effects (28%), aftercare with other providers (25%), and issues with mood or well-being (25%). Half of the patients were not planning to see a PCP, citing various reasons: not wanting to travel to the PCP office, competing priorities (eg, family member with more pressing medical concerns), feeling that follow-up at the trauma center was adequate and logistical difficulties (eg, insurance, time off work).

Adoption of and fidelity to the support program by the PCPs

The PA support program was provided to all of the PCPs caring for the 19 patients whose PCP was reached and confirmed (100% adoption, table 4). The PA completed at least one consultation with a PCP or PCP delegate on behalf of each of these patients (fidelity). The number of consultations ranged from one to four, with an average of 1.9 (fidelity). A total of 36 consultations were conducted over the course of the study. The minority of consultations were conducted with PCPs (22%); most were conducted with PCP delegates (56% with medical assistants, 22% with nurses). The PA documented that the three topics planned for the consultations (pain management, opioid management, and trauma hospitalization follow-up) were discussed for the majority of the patients (fidelity). Other topics that the PA documented as discussed during the consultations with PCPs or their delegates included other medication management, aftercare with other providers, opioid side effects, and social context or financial issues.

Acceptability, appropriateness, and feasibility of the support program

The survey of PCPs or their delegates conducted after completion of the support program (data not shown) had a low response rate (18%, 3 of the 17 surveyed). Two of the three responding PCPs agreed that the program was acceptable, appropriate, and feasible, and the third was neutral about the program.

Facilitators and barriers to implementing the support program

The PA’s work in the support program was facilitated by a high level of receptivity by the patients, and by the shift to providing the support to the PCPs directly, as they were hard to reach, and to the PCP delegates (table 5). The PA cited many challenges to providing support. First, the study took place exclusively during the COVID-19 pandemic, which decreased the ability to meet patients in the hospital prior to discharge, the willingness of patients to see their PCPs due to COVID-19 concerns, and the availability of the PCPs. Second, PCPs were difficult to reach, and the PA was not available at all times for call backs from the providers. Third, many patients did not plan to follow-up with their PCP, limiting the utility of a PCP-focused intervention. Finally, the PA expressed concern that the PA support could be construed as the trauma center trying to provide oversight of the PCP’s patient care, and that it was challenging to recommend a care plan that differed from that already offered by the PCP.

Key lessons for trauma centers

Several key lessons emanated from the interview with the PA (table 6). First, patients were easier to reach than PCPs, and were appreciative of post-trauma hospitalization follow-up. They discussed many post-trauma issues beyond pain and opioid management. Patients discharged home after trauma often did well after hospitalization, tapered off their prescribed opioids early, and their providers may not have needed a post-trauma hospitalization intervention. Notably, it was not routinely clear to patients or PCPs who was responsible for post-trauma care and the opioid taper, and taper instructions were also not clear to many patients. Finally, patients were receptive to and routinely used non-opioid options for pain management after trauma.

Implications for level I trauma centers

Hospital discharge services routinely include development of a discharge summary that is sent to the patients’ follow-up care providers.28 Ensuring that these discharge summaries clearly and specifically designate the provider responsible for post-trauma care and opioid management after discharge and a clear opioid taper plan would be a first step toward supporting patients and their PCPs. Discharge care coordinators who facilitate the discharge planning for inpatients could take responsibility for ensuring that this tailored information is provided consistently in all discharge summaries. Second, Zatzick et al’s care management program for patients both prior to and after trauma discharge resulted in fewer serious concerns in the 6 months after trauma discharge as well as fewer emergency department visits 3–6 months after discharge.27 Trauma hospitalizations are frequently complex and require multiple services (eg, multiple surgeons, pain management specialists, physical and occupational therapists) to optimize outcomes. This complex care from multiple services can lead to complex discharge management needs, making a navigator like a care coordinator critical for patients after discharge. For post-trauma patients who live distant from the trauma center and are at high risk of adverse outcomes, this can be even more important. Testing a multilevel intervention that includes both a patient-level intervention with a navigator or care manager and a PCP-level intervention with an opioid management component would shed light on whether Zatzick’s care management intervention findings could be replicated and perhaps amplified for this selected group. Preliminary investigation supports the feasibility of a patient-level care management intervention for patients with OUD.29

Implications for policy makers seeking to optimize post-trauma patient care

The American College of Surgeons Committee on Trauma publishes clinical best practice guidelines that inform operation of US trauma centers and link trauma center designation to quality indicators.30 The American College of Surgeons’ Resources for Optimal Care of the Injured Patient (2022 Standards)31 includes a section on discharge planning, which requires that all trauma centers have a process to determine the level of care and the rehabilitation services required after trauma center discharge. These standards recommend, but do not require, that level I and II trauma centers adopt patient-centered strategies for facilitating patient transition into the community. They include ongoing care management as one of those strategies, citing Zatzick et al’s research.27 Shifting this recommendation to a requirement that includes care coordination, especially for high-risk patients prescribed opioids after discharge, and then monitoring the requirement’s implementation and outcomes would help ensure that level I trauma centers include a post-trauma discharge program for supporting their patients as they transition from inpatient to their home settings. Additionally, high-quality post-trauma discharge care coordination requires funding. Current payment structures for global surgical care, including the surgical hospitalization and postoperative outpatient follow-up care, do not support the cost of this type of service. Development of new billable codes or value-based allocation of existing bundled payments for perioperative care management services to support high-risk populations would be critical to implementing a comprehensive inpatient and care transition program.32

Limitations

This study is limited by the relatively small number of trauma patients recruited to the support program’s intervention. Recruitment was conducted during the COVID-19 pandemic, which impacted recruitment as well as the primary care practice landscape throughout the study. Despite this disruption, the study had 100% adoption of the intervention by the PCPs or their delegates who were confirmed and reached. The trial initially had planned to recruit a stratified sample of patients who were opioid naïve versus patients using opioids chronically for pain. However, few patients using opioids chronically for pain qualified for the study, so we were unable to examine the implementation of the support program in this higher risk population. Although 2 of the 3 PCPs responding to a postintervention study survey thought that the support program was acceptable, appropriate, and feasible, the survey response rate was low—18%. Finally, implementation of the program was complicated by a new set of state rules concerning opioid prescribing that went into effect around the same time as the study. These rules refer to guidelines that in opioid-naïve patients, any opioids prescribed during the first 6 weeks postoperatively should be managed solely by the surgeon.33 This may have led some patients to contact one of several surgical services involved in their inpatient care for pain management issues rather than their PCPs, even if they lived more distant from the trauma center, although we did not measure this.