CHAPTER 3
LITERATURE REVIEW (STUDY 1)

OBJECTIVES AND METHODS

Search Design and Strategy

First, the Workgroup and National Academies’ Research Center (the Research Center) developed a search term matrix that was used for the literature review. The Research Center then searched electronic databases (Embase, MEDLINE, PubMed, and Scopus) to identify peer-reviewed articles. In addition, a search of the internet was conducted to identify reports in the grey literature (government, consensus, and white papers) that could contribute to the overall understanding of PPGs. The search was limited to articles (peer-reviewed or government reports) that were published in English in the United States between 2009 and 2019. Search terms reflected the five identified professions—medicine (MD and DO), nursing (RN and APRN), physician assistant, dentistry (DDS and DMD), and pharmacy (pharmacists and pharmacy technicians)—as well as relevant treatment and conditions, health care professional competencies, collaboration with patients and families, and patient outcomes (see Appendix A).

Of the 822 articles initially identified using the above criteria, perspective and editorial articles (213) and articles not available in full text (62) were excluded from the sample, as the Workgroup decided to focus analysis on research articles and articles that described quality improvement projects. The remaining 547 articles underwent abstract screening, of which only US-based research studies that focused on professional practice gaps among practicing clinicians were included, or 310 of the 547 original articles. The decision to review US-based research studies was made because the Action Collaborative is focused on the opioid crisis In the United States.

Coding and Analysis

Members of the Workgroup then developed inclusion criteria, including a working definition of what constitutes a professional practice gap, to select articles from the initial search for analysis (see coding

inclusion criteria matrix in Appendix C). The working definition of a PPG was based on the cited Accreditation Council for Continuing Medical Education (ACCME) definition, which states that PPGs are the difference between health care processes or outcomes observed in practice and those potentially achievable on the basis of current professional knowledge. The Workgroup conducted a reliability study to evaluate interrater reliability among coders for inclusion in the review. Seven members of the Workgroup, whose professional backgrounds reflected all of the professions included in the search terminology, independently reviewed 10 randomly selected articles. Determining whether the article described a PPG was the area of greatest variation among coders and was addressed through group discussion and consensus for rationale among the team members. For example, an article that described differences in how physicians and nurse practitioners prescribed opioids was classified by four of six reviewers as meeting inclusion criteria for describing a PPG (difference in practices between two professions), while two reviewers were unsure. By reviewing and reinforcing the definition of a PPG, Workgroup members were able to resolve discrepancies.

An independent research team with expertise in coding and analysis was subcontracted to complete the article coding, using the matrix developed by the Workgroup. The research team was led by a doctor-ally prepared, tenured university professor with extensive expertise in this type of analysis.

Results

Quantitative

A total of 310 articles (310/547; 57%) met the inclusion criteria for this review. The predominant reason articles were excluded was that they failed to describe a professional practice gap (n = 86; 36%).

Table 1 summarizes research article composition. The research articles reflected research or quality improvement studies and were classified as quantitative, qualitative, or mixed methods.

TABLE 1 | Literature Review Research Article Composition

Table 2 describes health care professionals by type represented in the literature review. Physicians were the most common health care profession, followed by nurses, pharmacists, physician assistants, and dentists.

TABLE 2 | Health Care Professionals by Type Represented in the Literature Review

Overall, areas of specialty were indicated by 43% of respondents (see Table 3). The majority of articles reflected practice of physicians only, but 20% of articles included analysis of two professions. A small number of articles included more than two professions.

TABLE 3 | Specialties Represented in Literature Review

As seen in Table 4, the articles in this review described practices in a variety of care settings, including primary care/outpatient, acute care/inpatient, clinic/outpatient-unspecified, community/outpatient, clinic/inpatient, and other. The most common care setting was primary care/outpatient.

TABLE 4 | Practice Environments Represented in Literature Review

Table 5 describes the domains of practice included in the literature review. Chronic pain management was the most common domain of practice. Additional domains of practice included acute pain management, substance use disorders, and other practice domains.

TABLE 5 | Domains of Practice in Literature Review

Table 6 summarizes the types of data sources included in the literature review. Data sources used to identify or describe PPGs were predominantly descriptive self-reports. Other data sources included medical record and other. Categories were not mutually exclusive; therefore, one article could include multiple types of data sources.

TABLE 6 | Data Sources Used to Identify or Describe PPGs in the Literature

As seen in Table 7, the predominant patient population referred to in the articles reviewed was adult. Other patient populations included “across the lifespan” and pediatric.

TABLE 7 | Patient Populations Referred to in Literature Review Articles

Table 8 summarizes PPGs by type or stage in the care process. The articles in this review most commonly reflected PPGs associated with prescribing or tapering opioids. Additional types or stages included monitoring, screening/assessment, non-pharmacological treatment, identification/diagnosis, prescribing non-opioids, referral, and other. Categories were not mutually exclusive; therefore, one article could include multiple types or stages in the care processes.

TABLE 8 | PPGs by Type or Stage in the Care Process Included in the Literature Review

The majority of articles cited gaps in clinical knowledge attitudes and biases, and/or the use of (failure to use/lack of available) evidence-informed tools and resources as the root causes for the identified PPGs (see Table 9). Communication with patients/families, constraints in the practice setting, and/or communication with other members of the health care team were also cited as PPGs. Categories were not mutually exclusive; therefore, one article could include multiple types of PPGs. Coders also captured qualitative data into two additional categories that reflected (1) health care professionals and patients/families; and (2) the environment where care is delivered.

TABLE 9 | Professional Practice Gap by Category in Literature Review

A summary of the qualitative data is presented below.

Qualitative

Qualitative data captured from the articles in this review contribute to a deeper understanding of PPGs among the five represented health professions. The Workgroup co-leads extrapolated qualitative data from the articles reviewed and entered those data into the coding matrix under two major categories: (1) PPGs related to health care professionals and patients/families, and (2) PPGs related to the environment where care is delivered. The Workgroup co-leads then, by consensus, grouped the data based on the description of the PPG and organized the data thematically in order to facilitate a logical presentation for the reader. The qualitative data are presented in the two major categories below, and each category contains several themes (see Table 10).

Category 1: Professional practice gaps related to health care professionals and patients/families

Health Care Professionals Theme 1: Education and Training

PPGs attributed to the education and training of health care professionals can be further categorized into gaps related to competency, perceived lack of evidence, and lack of access to existing evidence. Gaps related to the competency of health care professionals included descriptions such as: lack confidence and/or training to prescribe opioids and/or to use a multi-modal approach to control pain (Andrilla, Coulthard, and Patterson, 2018; Carey et al., 2018; Choo et al., 2016; Jamison et al., 2016; Jamison et al., 2014; Keller et al., 2012; Lum et al., 2011; Phelan et al., 2009; Regunath et al., 2016; Samuels et al., 2016); problems converting between different opioids (McCalmont et al., 2018); inconsistent medical education related to addiction assessment and ability to identify drug diversion (McCauley et al., 2019); and lack of experience in prescribing opioids (Barry et al., 2010; Khalid et al., 2015; Samuels et al., 2016).

Health Care Professionals Theme 2: Guidelines

Guidelines were another area of PPGs that reflected a lack of competence in the clinical setting. Articles cited unawareness of an evidence-based guideline or lack of application of guidelines by health care professionals as key gaps (Goesling et al., 2018; Mafi et al., 2015; McCalmont et al., 2018; McCann et al., 2018; Mehta et al., 2010; Morse et al., 2011; Starrels et al., 2011). Specifically, articles described health care professionals who reported a willingness to perform opioid harm reduction interventions, but did not provide these services to their patients (Samuels et al., 2016); who reported that they had implemented evidence-based guidelines, but rates of drug screening and specialty referral remained low (Chen et al., 2016); and those who chose to use a clinical impression or personal preference for prescribing opioids despite the available evidence-based guideline (Irvine et al., 2014; Park et al., 2019).

Health Care Professionals Theme 3: Lack of Evidence, Tools, or Resources

Health care professionals reported a lack of high-quality evidence for prescribing opioids or co-prescribing sedatives and opioids, and tools that were not user-friendly (Franklin et al., 2013; Gaither et al., 2016; Huang and Kuelbs, 2018; Kircher et al., 2014; Kraus et al., 2015; Larochelle et al., 2015; Leverence et al., 2011; Linnaus et al., 2019; Morse et al., 2011). Health care professionals also reported not knowing risk mitigation strategies for prescribing opioids, including how to screen patients for SUDs, how to provide patient education, and types of prescription drug diversion programs that were available as resources (McCarthy et al., 2016; Reid et al., 2010).

Health Care Professionals Theme 4: Attitudes or Biases

A number of articles described negative attitudes or biases held by health care professionals toward patients. Findings indicated health care professionals may exhibit negative attitudes and biases toward patients who have chronic pain and depression, who have illicit benzodiazepine use, who use Medicaid insurance to pay for an office visit, and who have an opioid-using spouse. (Hirsh et al., 2014; Knudsen et al., 2018).

Health care professionals also expressed concern about prescribing opioids due to the potential for addiction and side effects (Leong et al., 2010; Lum et al., 2011); fear of causing harm to the patient (Jamison et al., 2016; Leong et al., 2010; Linnaus et al., 2019; Lum et al., 2011; Macerollo et al., 2014; Schuman-Olivier et al., 2013); concern of opioid misuse by family members or caregivers (Spitz et al., 2011); and acknowledging patients’ concerns with the stigma of medications for OUD (i.e., methadone) (Shah and Diwan, 2010). Some health care professionals reported that the patient or family was reluctant to try an opioid to control pain (Spitz et al., 2011).

Health Care Professionals Theme 5: Lack of Interprofessional Collaboration, Interest, and Trust

Health care professionals reported a lack of interprofessional collaboration in the care of patients with SUDs or chronic pain (Mehta et al., 2010). There was also a reported lack of interest from some health care professionals for prescribing opioids (Barry et al., 2010). Finally, lack of trust was a theme

in some articles with health care professionals describing challenges in trusting the patient’s description of pain and the subjectivity of pain scales, sometimes manifested as the health care professional not documenting the pain score in the medical record (Brown et al., 2015; Calcaterra et al., 2016; Mehta et al., 2010; Regunath et al., 2016).

Health Care Professionals Theme 6: Differences in Prescribing Practices

Differences in prescribing practices between groups was also a common theme in the literature reviewed. Different practices can be categorized into two general areas: provider type and type of pain. In the literature review, differences in prescribing practices were found between physicians and APRNs (Franklin et al., 2013; McCalmont et al., 2018; Muench et al., 2019); physicians and physician assistants (Ganem et al., 2015); primary care physicians and pain specialists (McCarberg et al., 2013); resident physicians and attending physicians (Khalid et al., 2015); and junior and senior resident physicians (Linnaus et al., 2019). The root cause of the differences in prescribing patterns was not well understood.

There were differences noted in prescribing practices for patients who had different types of pain. Specifically, there were differences in prescribing practices between patients who had acute versus chronic pain (Larochelle et al., 2015), and between patients who had unclassified pain versus a known pain source (e.g., fibromyalgia vs. broken bone) (Romanelli et al., 2017). There were also differences between patients who experienced breakthrough pain (BTP). For example, patients reported lower BTP in the community setting as compared to the pain clinic setting, and patients reported more episodes of BTP for non-cancer pain as compared to cancer-related pain (Portenoy et al., 2010).

Patients and Families Theme 1: Patient Demographics

There were significant differences in prescribing practices that reflected patient demographic variables, including age (Grasso et al., 2017; Monitto et al., 2017; Okunseri et al., 2015; Reid et al., 2010; Shah, Hayes, and Martin, 2017); gender (Chen et al., 2011; Grasso et al., 2016; Manchikanti et al., 2013; Monitto et al., 2017; Murphy, Phillips, and Rafie, 2016; Oliva et al., 2015; Ringwalt et al., 2014; Romanelli et al., 2017; Shah, Hayes, and Martin, 2017); race (Chen et al., 2011; Grasso et al., 2016; Moskowitz et al., 2011; Okunseri et al., 2015; Rasu and Knell, 2018; Ringwalt et al., 2014; Ringwalt et al., 2015; Romanelli et al., 2017); socioeconomic status (Platts-Mills et al., 2012; Ray et al., 2017); geographic location (area of United States; rural vs non-rural) (McDonald, Carlson, and Izrael, 2012; Rasu and Knell, 2018; Shah, Hayes, and Martin, 2017); and payor type (private insurance, public insurance or no insurance) (Okunseri et al., 2015; Rasu and Knell, 2018; Romanelli et al., 2017; Shah, Hayes, and Martin, 2017). There were also differences in prescribing practices between patient populations (e.g. geriatric versus orthopedic; surgery versus medical) and between civilian and active duty military patients (Ganem et al., 2015; Mehta et al., 2010). There were differences in prescribing practices for patients with past medical histories that included chronic pain and/or SUDs as compared to patients who did not have similar past medical histories, and between patients who had comorbidities of mental illness as compared to those who did not (Grasso et al., 2016; Grasso et al., 2017; Nugent et al., 2017; Rasu and Knell, 2018).

Patients and Families Theme 2: Patient-Reported Differences

Patient-reported differences in how their pain was treated were also prevalent in the literature. Patients reported gaps that included undertreatment of their pain (McCauley et al., 2014) and using the emergency department for pain management (McCauley et al., 2014). Patients reported not being provided treatment options for substance dependence (McCauley et al., 2014); general lack of access to therapy services for pain and SUDs (Nugent et al., 2017; Penney et al., 2017); and lack of a structured process for tapering dosages (Penney et al., 2017). Patients reported that physicians spent insufficient time educating them on pain management (McCauley et al., 2014); presenting them with alternative pain management options (Phelan et al., 2009); or discussing the risks of opioid dependence (Phelan et al., 2009). Patients also reported not being included in decision-making processes around their own pain management (Penm et al., 2019).

Category 2: Professional practice gaps related to the environment where care is delivered

Care Environment Theme 1: Challenges for Prescribers due to Organizational Policies

Articles also described challenges in relation to patient or organizational goals. For example, a hospital may implement a strict policy on prescribing opioids. An overly strict policy might conflict with the needs or goals of the prescribing health care provider, depending on the population of patients cared for by the provider. Finally, providers may have concerns that their income and employment would be impacted by low patient satisfaction scores if they did not prescribe opioids requested by patients or if patients’ pain was not well controlled (Henry et al., 2018; Penm et al., 2019).

Care Environment Theme 2: Burdens on Health Care Professionals

Insufficient time and resources in the practice setting were commonly cited as contributing to PPGs. Authors described providers struggling with competing tasks (Barry et al., 2010; Behar et al., 2017; Bergman et al., 2013; Harle et al., 2015; Hawkins et al., 2017; Huhn and Dunn, 2017; Kohlbeck et al., 2018; Kraus et al., 2015; McCann et al., 2018); high administrative burden (Kahler et al., 2017; Shugarman et al., 2010); excessive cognitive load (Burgess et al., 2014); and institutional pressure to reduce hospital re-admissions and discharges (Calcaterra et al., 2016) as negatively impacting their ability to manage patients with SUDs or OUDs. Providers also cited fear of litigation, in particular fear of the medico-legal consequences related to opioid diversion and fraud as significant concerns (Andraka-Christou and Capone, 2018; Calcaterra et al., 2016).

System-related constraints that contributed to PPGs were also common. These constraints can be categorized into five overarching areas: insurance coverage, mandatory continuing education, lack of referral resources, lack of institutional guidelines, and issues with data interoperability.

Care Environment Theme 3: Insurance Coverage

Descriptions of constraints related to health insurance coverage included whether the patient was covered by an insurance policy or not, and if covered, what specific treatment was covered under the policy. Articles cited low reimbursement rates and limited or no insurance coverage for mental health services and addiction counselors as constraints (Andraka-Christou and Capone, 2018; Barry et al., 2010; Behar et al., 2017; Cheng et al., 2019; Huhn and Dunn, 2017).

Care Environment Theme 4: Mandatory Continuing Education

Regulatory restrictions were cited as a constraint that contributed to PPGs. Descriptions of these types of constraints included concern that physicians would not be willing to comply with the mandatory continuing education requirements for prescribers of extended-release and long-acting opioid medication under the Food and Drug Administration’s (FDA’s) Risk Mitigation and Evaluation Strategies (REMS) requirements, which would decrease the number of physicians eligible to prescribe opioids controlled by REMS requirements (Slevin and Ashburn, 2011). Another constraint cited was requirements related to buprenorphine waivers (Rosenblatt et al., 2015; Stein et al., 2015). However, the U.S. Department of Health and Human Services (HHS) loosened buprenorphine waiver requirements in April 2021, allowing eligible medical professionals to treat up to 30 patients with buprenorphine without completing the federal certification process (HHS, 2021). Articles also described lack of planning at the state level to address adequate numbers of providers who could prescribe controlled substances to meet population health needs as a constraint in the practice setting (Sera et al., 2017).

Care Environment Theme 5: Lack of Referral Resources

Lack of available referral resources across multiple health care settings was cited as contributing to PPGs. Articles described insufficient numbers of mental health services practitioners, addiction counselors, and pain management specialists as constraining health care practitioners ability to care for patients with OUD or other SUDs (Andrews et al., 2013; Andrilla, Coulthard, and Patterson, 2018; Barry et al., 2010; Leverence et al., 2011; Morse et al., 2011; Wiznia et al., 2017). Articles also recognized that lack of available referral resources were particularly challenging for rural providers of care (McCalmont et al., 2018; McCann et al., 2018).

Care Environment Theme 6: Lack of Institutional Guidelines

Lack of institutional guidelines or resources were described by a number of articles as contributing to PPGs. They described lack of standardization in opioid prescribing within organizations (Huang and Kuelbs, 2018; Raneses et al., 2019; Regunath et al., 2016; Ringwalt et al., 2014; Schwartz et al., 2018); how lack of institutional standardization manifested in practice variations, such as more liberal opioid prescribing practices in the emergency department as compared to other departments in the institution; prescribing practices that were medical or surgical specialty dependent; and institutions that had a “prescribing culture” (Gernant, Bastien, and Lai, 2015; Gugelmann et al., 2013; Irvine et al., 2014; Myers et al., 2017; Raneses et al., 2019).

Care Environment Theme 7: Data Interoperability

Data interoperability, or lack of, was identified as a health care system constraint that contributed to PPGs. Articles cited the large volume of clinical notes that were not accessed by providers due to time constraints and questioning the efficacy of electronic health record order sets as challenges related to data use (Kahler et al., 2017; Luk et al., 2016). In addition, lack of integration between prescription drug monitoring programs and the electronic health record, especially across state lines, was identified as being particularly problematic (Perrone, DeRoos, and Nelson, 2012).

Additionally, patient-related constraints in the practice setting was another source that contributed to PPGs.

Care Environment Theme 8: Patients

Constraints in the practice setting that were patient-related included logistics issues (e.g., patient cannot make the medical appointment), and lack of opportunity for a provider-patient relationship in some settings (e.g., the emergency department) (Chambers et al., 2016; McCauley et al., 2014; Nugent et al., 2017).

Table 10 provides a summary of the PPG categories and corresponding themes identified above.

TABLE 10 | Summary of PPG Themes by Category

Validation Results

Validation surveys were also conducted for three weeks, from mid-September to early October 2020. The objective was to confirm the findings of the literature review and to identify any potential areas that were not captured in the published, peer-reviewed literature. The target audiences included practicing clinicians and organizations with national data on clinical practices and outcomes.

The surveys were distributed through the Collaborative’s email lists, using a convenience sampling method. A total of 44 respondents completed the surveys, of which 24 were clinicians and 20 were individuals responding on behalf of their organizations. The surveys included statements about clinical practice that pulled from the quantitative and qualitative data from the literature review. Both clinicians and organizations indicated how reflective the statements were of their experience and findings, respectively. The clinician survey, in particular, asked respondents to indicate if the statements were reflective of their own practice and/or the practices of their colleagues. In addition, the clinician survey included an opportunity to enter professional practice gaps that were missing from the literature review. Respondents from the clinician survey reported PPGs across the following categories: systems (86.00%), treatments and resources (54.00%), attitudes and biases (49.00%), practice variation of undetermined origin (48.40%), and those pertaining to health care professionals (41.60%). These results varied across each category when individual clinicians described statements reflective of their peers. Respondents from the organization survey reported PPGs across the following categories: systems (88.20%), practice variation of undetermined origin (83.50%), health care professionals (70.30%), attitudes and biases (62.90%), and treatment and resources (62.60%). Notably, results from the clinician and organization surveys both described systems as the greatest contributor to PPGs. Additionally, results from the organization survey were higher values across each category of PPGs when compared to the results from the clinician survey.

Given the discrepancy in responses between statements that were reflective of individual clinician practice compared to the practices of colleagues and organizational findings, the survey data may be indicative of response bias. In addition, the small number of respondents and the non-random sampling method are significant limitations of the validation survey results. While the number of respondents to the validation survey was very low, the data are included for transparency and for the opportunity for researchers to consider replicating in the future.

Concurrently with the literature review and validation survey, the Workgroup conducted an analysis of health care professional requirements related to opioid use and SUD across accrediting, certifying, licensing, and regulatory bodies of the five identified health professions. Chapter 4 describes the process and outcomes used to collect, analyze, and aggregate those results.