A New Quality Improvement Toolkit to Improve Opioid Prescribing in Primary Care

Constance van Eeghen, DrPH, MHSA, MBA; Amanda G. Kennedy, PharmD, BCPS; Mark E. Pasanen, MD; Charles D. MacLean, MD


J Am Board Fam Med. 2020;33(1):17-26. 

In This Article


The study design used in this project was an iterative toolkit-based QI approach with 3 cohorts of clinics over a 7-year period. We designed the toolkits for primary and specialty ambulatory care clinics in our region and described strategies to improve opioid prescribing based on best available evidence and accrued expertise. The third of these trials is still underway.

Trial 1

In 2012, as the potential for negative health outcomes of opioid prescribing became more apparent, the complexity of providing opioid therapy grew as well. Primary care providers found it challenging to care for patients with ongoing opioid prescriptions[14] and some began avoiding pain management care altogether.[15] Based on the clinical experience of the authors and discussion with colleagues, we predicted that clinicians and office staff would find a structured process improvement approach, laying out a team-based problem-solving method in steps, would be a helpful method of influencing opioid prescribing by providing best prescribing practices and increasing provider satisfaction with opioid prescribing. We assembled a list of best practices available at that time and offered an in-office, team-based project based on Lean process improvement, an easy-to-use structured QI approach. Lean is a systematic method of redesigning workflows to improve both care and provider satisfaction.[16–18] Each project was facilitated by one of the authors (CvE) to Vermont providers through the Office of Primary Care and Area Health Education Centers Program at the University of Vermont Larner College of Medicine. In accordance with the policy of the University of Vermont Committees on Human Research, this effort was determined to be a QI project and, therefore, exempt from review.

From 2012 to 2013, 9 Vermont primary care clinics and 1 orthopedic office specializing in back pain were recruited by convenience sample. Clinics were given an overview of the project goal and asked to commit 1 provider, 1 nurse, and 1 front desk staff to regular meetings for approximately 8 hours over 4 months. Recruitment closed after 10 clinics volunteered to participate. These meetings were facilitated by a Lean expert to select from a menu of strategies taken from opioid-prescribing best practices. Team members selected strategies based on analysis of their patients' needs and clinic's current workflow. Each clinic determined its own schedule of team meetings and duration of the project. In addition, we offered each clinic the opportunity to attend an individualized medical education conference to review the epidemiology, medication prescribing issues, and strategies for opioid prescribing. When requested, we also provided consultative visits and phone calls with local experts in the management of chronic pain. Participating clinics received a stipend of $1000 at the completion of their projects, determined to be 3 months after their last team meeting with their facilitator.

Our primary outcome measure was change in provider satisfaction with the clinic's management of opioid prescribing, measured by pre-and postproject surveys. We conducted a paired response analysis using Wilcoxon sign-rank to test changes in preproject results to postproject results for significance. Providers and staff also completed postproject surveys on the degree of completion of their chosen prescribing strategies to implement. We used STATA 15 (Stata Corporation, College Station, TX) for data management and analysis. In addition, the facilitator collected team process measures (number of team meetings, number of hours in meetings, number of strategies selected by type, and completion of project) and field notes on contextual factors.

Trial 2

Based on the outcomes of this initial trial, we developed the toolkit (first ed) as a manual for QI facilitators—in our experience many clinics have access to QI facilitation through a parent organization (eg, hospital, Federally Qualified Health Center [FQHC]) or through other avenues (eg, state-sponsored QI initiatives, Department of Health). The toolkit included 3 stages: (1) preparing the team and baseline measures, (2) designing workflows using Lean with selected opioid-prescribing strategies (see Table 1), and (3) implementing the resulting plan with appropriate follow-up measures. An extensive appendix included applicable state laws, QI assessment tools, and sample patient assessment and care protocols. Trial 2 recruited 7 clinics new to the study, 4 primary care and 3 specialty (dentist, orthopedic rehabilitation, and rheumatology) from 2014–2016, 2 of which were affiliates of the study institution. Recruitment closed in 2015. Teams underwent the same recruitment, introductions, support, and stipends as in trial 1. Clinics were allowed to customize their pre- and postproject surveys and could choose not to use them.

Our primary measure of success for trial 2 was the percent of clinics that completed the toolkit's activities. Second, we evaluated process measures collected by facilitators reflecting the uptake of the toolkit for each team: number of meetings, number of team hours, duration in months, and number of strategies selected by category (regulatory compliance; streamlining workflow; providing peer support; and responding to patients at risk). Facilitators also collected field notes on contextual factors.

During trial 2, we revised the toolkit (second ed) based on the new 2016 CDC guidelines,[10] feedback from trained facilitators who utilized the toolkit with clinics in our region, and the passage of opioid-prescribing laws in Vermont.[19]

Trial 3

During our recruitment process, we consistently heard that clinics had a high interest in improving prescribing practices but were unable to engage formally trained QI facilitators to support them. Learning from other Lean-based toolkits we had developed,[18] we redesigned the toolkit from 2018 to 2019 by using a workbook format (third ed). This format uses team-based problem-solving rubrics and checklists for use by clinic staff not formally trained in facilitation, thus expanding the audience for the toolkit, and added new strategies on dosing and alternative therapies (see Table 1). We recruited 3 health care organizations representing 10 Vermont primary care clinics, 1 of which was a study affiliate (3 sites). Teams underwent the same recruitment, introductions, support, and stipends as in trials 1 and 2, with the addition of coaching support for internal facilitators. Teams were free to choose their own strategies and QI data. In this study, we report only on the improvement strategies trial 3 clinics chose, as their projects were not finished at the time of publication. For an overview of each trial's goals, toolkit versions, requirements, outcome measures, and results, see Figure 1.

Figure 1.

Evolution of a Quality Improvement (QI) toolkit for opioid prescribing practices.