Chronic Pain, Opioid Misuse
Conditions
Keywords
Long-Term Opioid Therapy, Algorithms, Opioid Misuse, Chronic Pain
Brief summary
The NIH Helping to End Addiction Long-term (HEAL) initiative has identified a critical next step to addressing the opioid crisis: improving treatments for opioid misuse behaviors (e.g., using more opioids than prescribed, illicit substance use) in patients prescribed long-term opioid therapy for chronic pain. In previous work, the investigators have developed innovative consensus-based algorithms to manage these behaviors. By developing implementation strategies for these algorithms, this project is directly responsive to the HEAL initiative and promises to reduce opioid misuse-related harms.
Detailed description
Despite a growing understanding of the risks of long-term opioid therapy (LTOT), it continues to be frequently prescribed and remains a mainstay of treatment for chronic pain. The Centers for Disease and Control (CDC) Guideline for Prescribing Opioids for Chronic Pain is geared toward primary care providers and has been adopted as the standard of care by many healthcare organizations and insurers. Importantly, it encourages monitoring of patients on LTOT for opioid-related harms. By implementing monitoring, primary care providers may uncover various concerning behaviors, sometimes called aberrant drug-related behaviors or opioid misuse behaviors, that arise among individuals prescribed LTOT for chronic pain. These behaviors (e.g., missed appointments, using more opioid medication than prescribed, asking for an increase in opioid dose, aggressive behavior, and alcohol and other substance use) are common, concerning, and may represent unsafe use of LTOT or a developing opioid use disorder (OUD). However, the CDC Guideline and other existing evidence do not provide specific, detailed guidance about how to address concerning behaviors when they occur. Therefore, there is a critical need to understand how to best respond to these behaviors. The long-term goal of our program of research is to reduce LTOT-related harms, particularly from opioid misuse, and diminish their impact on the U.S. opioid epidemic. As a first step toward accomplishing this goal, the investigators conducted a Delphi study to rigorously establish consensus-based approaches to managing common and challenging concerning behaviors, from which algorithms were created. Identifying and operationalizing implementation strategies using an evidence-based framework are the critical next steps that must occur before any testing of the algorithms. The investigators successfully uncovered optimal implementation strategies through primary care provider experiences with Standardized Patients (SPs) followed by Consolidated Framework for Implementation Research (CFIR)- and Expert Recommendations for Implementing Change (ERIC)-guided individual interviews. Using our prior expertise developing clinic-wide opioid risk reduction strategies and a Patient-Provider advisory board, the investigators developed a comprehensive "implementation package" that can be delivered to primary care practices. The investigators now aim to conduct a pilot trial to test the algorithm implementation package. Guided by the CFIR-based implementation plan and using the implementation package that the investigators developed, pilot trial will be conducted to investigate feasibility, acceptability, and preliminary effectiveness of the algorithm implementation package.
Interventions
BEHAVIORALPilot study of algorithms implementation package
The algorithm implementation package includes a link to the algorithms in the Electronic Health Record, Smartphrases, audited feedback, and instructions.
Sponsors
University of Pittsburgh
National Institute on Drug Abuse (NIDA)
Study design
Allocation
NA
Intervention model
SINGLE_GROUP
Primary purpose
HEALTH_SERVICES_RESEARCH
Masking
NONE
Intervention model description
Investigators will conduct a pilot trial to assess the feasibility, acceptability, and preliminary effectiveness of the Algorithm Implementation Package across three University of Pittsburgh Medical Center (UPMC) primary care practices. The intervention will be implemented using a staggered rollout across the three clinics - General Internal Medicine Oakland will receive the intervention first, followed by Northern Medical Associates Hampton and Wexford. All clinics will receive the same intervention. Feasibility and acceptability will be assessed once only after the 6- or 9-month implementation period through clinician surveys and qualitative interviews. Preliminary effectiveness will be evaluated using electronic health record (EHR) data assessing reductions in concerning opioid-related behaviors and increases in the diagnosis and treatment of Opioid Use Disorder.
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
Yes
Inclusion criteria
* Clinicians practicing at UPMC community primary care clinics.
Exclusion criteria
* Clinicians not practicing at UPMC community primary care clinics.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Feasibility of Algorithms | At the end of the 6- or 9-month implementation period | The number of algorithms used by physicians was assessed via a survey administered at the end of the 6- or 9-month implementation period, measuring self-reported toolkit utilization during the study. Our primary feasibility benchmark will be that 80% of physicians report using at least one algorithm during the study period. |
| Acceptability of Algorithms | At the end of the 6- or 9-month implementation period | Acceptability of the algorithms was assessed via a self-report survey administered at the end of the 6- or 9-month implementation period, measuring physicians' awareness of the algorithms and self-reported toolkit use within six months of implementation. Our primary acceptability benchmark is that at least 80% of physicians report awareness of the algorithm implementation and at least 50% report using the algorithms during the study period. Additionally, qualitative interviews with physicians and staff provided further insights, which were analyzed using thematic analysis. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Preliminary Effectiveness of Algorithms - MME Reduction ≥10% | Pre-implementation (12 months), implementation (6 to 9 months), post-implementation (12 months) | Number of long-term opioid therapy (LTOT) patients whose 90-day average Morphine Milligram Equivalents (MME) decreased at or above a margin of 10% from the start of the reporting period to the end of the reporting period. |
| Preliminary Effectiveness - Average MME Within Last 90 Days | Pre-implementation (12 months), implementation (6 or 9 months), post-implementation (12 months) | Average morphine milligram equivalents (MME) among long-term opioid therapy (LTOT) patients during the last 90 days of each period. |
| Preliminary Effectiveness of Algorithms - Opioid Discontinuation | Pre-implementation (12 months), implementation (6 to 9 months), post-implementation (12 months) | Number of long-term opioid therapy (LTOT) patients whose 90-day average Morphine Milligram Equivalents (MME) at the start of this reporting period was 0. |
| Preliminary Effectiveness of the Algorithms - New OUD Diagnoses in LTOT Patients | Pre-implementation (12 months), implementation (6 to 9 months), post-implementation (12 months) | New opioid use disorder (OUD) diagnoses documented in the electronic health record (EHR) among all LTOT patients seen by participating physicians, by period. No new OUD diagnoses were documented in any period in LTOT patients. |
Countries
United States
Contacts
PRINCIPAL_INVESTIGATORJessica Merlin
University of Pittsburgh
Participant flow
Recruitment details
Three University of Pittsburgh Medical Center (UPMC) clinics were recruited for the study, with clinicians at these clinics participating in the implementation. The toolkit was sequentially integrated into the practices between September 2022 and September 2023.
Pre-assignment details
We reduced our sample to 3 clinics: two community with the same director, and one larger academic practice. We started the intervention at the academic clinic 2 months before the two community clinics for logistical reasons. Because not all physicians reported information about clinic affiliation, data are presented into a single arm. The academic clinic implementation lasted 6 months and the two community clinics 9 months. All 3 clinics were assessed 12 months pre- and post-implementation.
Participants by arm
| Arm | Count |
|---|---|
| Implementation Bundle The 'Implementation Bundle' was integrated into participating clinics over six to nine months. This algorithm implementation package included a link to the algorithms in the Electronic Health Record, Smartphrases, audited feedback, and instructions. | 49 |
| Total | 49 |
Baseline characteristics
| Characteristic | Implementation Bundle |
|---|---|
| Age, Categorical <=18 years | 0 Participants |
| Age, Categorical >=65 years | 0 Participants |
| Age, Categorical Between 18 and 65 years | 49 Participants |
| Ethnicity (NIH/OMB) Hispanic or Latino | 0 Participants |
| Ethnicity (NIH/OMB) Not Hispanic or Latino | 0 Participants |
| Ethnicity (NIH/OMB) Unknown or Not Reported | 49 Participants |
| Provider at UPMC | 49 Participants |
| Race (NIH/OMB) American Indian or Alaska Native | 0 Participants |
| Race (NIH/OMB) Asian | 0 Participants |
| Race (NIH/OMB) Black or African American | 0 Participants |
| Race (NIH/OMB) More than one race | 0 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 0 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 49 Participants |
| Race (NIH/OMB) White | 0 Participants |
| Region of Enrollment United States | 49 Participants |
| Sex: Female, Male Female | 20 Participants |
| Sex: Female, Male Male | 29 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk |
|---|---|---|
| deaths Total, all-cause mortality | 0 / 49 | 0 / 49 |
| other Total, other adverse events | 0 / 49 | 0 / 49 |
| serious Total, serious adverse events | 0 / 49 | 0 / 49 |
Outcome results
Primary
Acceptability of Algorithms
Acceptability of the algorithms was assessed via a self-report survey administered at the end of the 6- or 9-month implementation period, measuring physicians' awareness of the algorithms and self-reported toolkit use within six months of implementation. Our primary acceptability benchmark is that at least 80% of physicians report awareness of the algorithm implementation and at least 50% report using the algorithms during the study period. Additionally, qualitative interviews with physicians and staff provided further insights, which were analyzed using thematic analysis.
Time frame: At the end of the 6- or 9-month implementation period
Population: Physicians from the three participating clinics who agreed to participate in post-implementation surveys.
| Arm | Measure | Group | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|---|
| Implementation Bundle | Acceptability of Algorithms | Physicians aware of algorithm implementation | 19 Participants |
| Implementation Bundle | Acceptability of Algorithms | Physicians who used at least one algorithm | 7 Participants |
Primary
Feasibility of Algorithms
The number of algorithms used by physicians was assessed via a survey administered at the end of the 6- or 9-month implementation period, measuring self-reported toolkit utilization during the study. Our primary feasibility benchmark will be that 80% of physicians report using at least one algorithm during the study period.
Time frame: At the end of the 6- or 9-month implementation period
Population: Physicians from the three participating clinics who agreed to participate in the post-implementation survey.
| Arm | Measure | Category | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|---|
| Implementation Bundle | Feasibility of Algorithms | Physicians who used at least one algorithm | 7 Participants |
| Implementation Bundle | Feasibility of Algorithms | Physicians who did not use any algorithm or did not respond to this question | 13 Participants |
Secondary
Preliminary Effectiveness - Average MME Within Last 90 Days
Average morphine milligram equivalents (MME) among long-term opioid therapy (LTOT) patients during the last 90 days of each period.
Time frame: Pre-implementation (12 months), implementation (6 or 9 months), post-implementation (12 months)
Population: Outcome values are aggregated from electronic health record (EHR) data for LTOT patients attributed to participating physicians (n=49). Patient EHR data contributed to the outcomes, but patients were not enrolled as study participants. Row values reflect patient-level means (SD) by period. Means aggregate monthly 90-day MME within each period.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Implementation Bundle | Preliminary Effectiveness - Average MME Within Last 90 Days | Pre-implementation Period | 105.25 90-day average MME (mg/day) | Standard Deviation 16.27 |
| Implementation Bundle | Preliminary Effectiveness - Average MME Within Last 90 Days | Implementation Period | 96.76 90-day average MME (mg/day) | Standard Deviation 21.06 |
| Implementation Bundle | Preliminary Effectiveness - Average MME Within Last 90 Days | Post-implementation Period | 96.48 90-day average MME (mg/day) | Standard Deviation 16.34 |
Secondary
Preliminary Effectiveness of Algorithms - MME Reduction ≥10%
Number of long-term opioid therapy (LTOT) patients whose 90-day average Morphine Milligram Equivalents (MME) decreased at or above a margin of 10% from the start of the reporting period to the end of the reporting period.
Time frame: Pre-implementation (12 months), implementation (6 to 9 months), post-implementation (12 months)
Population: Outcome values are aggregated from electronic health records (EHR) for LTOT patients attributed to participating physicians (n=49). Patients EHRs contributed outcome data but were not enrolled study participants. Number of LTOT patients whose 90-day average MMEs decreased at or above a margin of 10% from the start of the reporting period to the end of the reporting period.
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Implementation Bundle | Preliminary Effectiveness of Algorithms - MME Reduction ≥10% | Pre-implementation Period | 297 Clinic patients |
| Implementation Bundle | Preliminary Effectiveness of Algorithms - MME Reduction ≥10% | Implementation Period | 122 Clinic patients |
| Implementation Bundle | Preliminary Effectiveness of Algorithms - MME Reduction ≥10% | Post-implementation Period | 270 Clinic patients |
Secondary
Preliminary Effectiveness of Algorithms - Opioid Discontinuation
Number of long-term opioid therapy (LTOT) patients whose 90-day average Morphine Milligram Equivalents (MME) at the start of this reporting period was 0.
Time frame: Pre-implementation (12 months), implementation (6 to 9 months), post-implementation (12 months)
Population: Outcome values are aggregated from electronic health records (EHR) for LTOT patients attributed to participating physicians (n=49). Patients EHRs contributed outcome data but were not enrolled study participants. Row values reflect patient-level totals by period.
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Implementation Bundle | Preliminary Effectiveness of Algorithms - Opioid Discontinuation | Pre-implementation Period | 95 Clinic patients |
| Implementation Bundle | Preliminary Effectiveness of Algorithms - Opioid Discontinuation | Implementation Period | 109 Clinic patients |
| Implementation Bundle | Preliminary Effectiveness of Algorithms - Opioid Discontinuation | Post-implementation Period | 166 Clinic patients |
Secondary
Preliminary Effectiveness of the Algorithms - New OUD Diagnoses
New opioid use disorder (OUD) diagnoses documented in the electronic health record (EHR) among all patients seen by participating physicians, by period. No new OUD diagnoses were documented in any period.
Time frame: Pre-implementation (12 months), implementation (6 to 9 months), post-implementation (12 months)
Population: Outcome values are aggregated from electronic health records (EHR) for LTOT patients attributed to participating physicians (n=49). Patients EHRs contributed outcome data but were not enrolled study participants. Row values reflect patient-level totals by period.
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Implementation Bundle | Preliminary Effectiveness of the Algorithms - New OUD Diagnoses | Pre-implementation period | 0 Diagnoses |
| Implementation Bundle | Preliminary Effectiveness of the Algorithms - New OUD Diagnoses | Implementation period | 0 Diagnoses |
| Implementation Bundle | Preliminary Effectiveness of the Algorithms - New OUD Diagnoses | Post-implementation period | 0 Diagnoses |