9

A Research Agenda to Advance Clinician Professional Well-Being

As the committee reviewed the research and evidence relevant to its charge, it found multiple gaps. This chapter discusses these gaps and charts a course forward for the research community and other stakeholders.

LIMITATIONS OF THE EXISTING LITERATURE

In its discussions about the approach to the study charge, the committee deliberated over various manifestations of workplace stress, such as burnout, compassion fatigue, and posttraumatic stress disorder, and concluded that the report should focus on burnout as a barrier to professional well-being. The scope of the literature review and the committee’s approach are discussed in Chapters 1 and 2.

Nearly all of the research that the committee found in its extensive review of the literature focused on physicians, residents, medical students, and, to a lesser degree, nursing populations. The committee identified few publications related to burnout and professional well-being in the dentistry and pharmacy professions, suggesting the evidence base in those fields is largely non-existent. Similarly, few studies were identified on burnout and professional well-being among advance practice providers (e.g., advanced practice nurses, nurse anesthetists), other members of the health care team (e.g., physician assistants, physical therapists), fellows in graduate medical education, and learners in nursing, pharmacy, and dentistry. Even the foundational information in these fields, such as the national prevalence of burnout among dentists and pharmacists in practice as well as in training, is not available. There is a strong sense, however, that these clinicians and learners are not immune to the workplace stressors outlined in this report

and that they likely have a substantial prevalence of burnout. It also became apparent to the committee that little is known about the contributing factors and the consequences of burnout among these clinician types and that a deeper understanding of the drivers and consequences of burnout across career and life stages for all clinicians is needed.

Overall, the committee found limited evidence for causal relationships between clinician or learner burnout and many of the possible contributing and outcome variables. Most studies reviewed by the committee were cross-sectional correlational and explored a limited number of personal and system factors. The few published longitudinal studies that the committee identified generally had smaller sample sizes and were still correlational in nature. The committee found few randomized trials evaluating the causes of, or solutions to, clinician burnout.

Many studies that the committee reviewed had small sample sizes. In the larger studies it did review, the response rates were typically low and usually the authors did not include a non-response analysis. In many studies the key independent variables, such as clinical workload, were subjectively reported by the same individuals who also provided the key dependent variables, such as burnout or job satisfaction, thereby introducing inherent biases. Most of the available intervention studies focused on individual strategies and often lacked rigorous study design, data collection with validated instruments, appropriate control groups, and long-term follow-up (Panagioti et al., 2017; West et al., 2016). Overall, the evidence for systems solutions for combating burnout is scant across all clinician types.

Questions about generalizability also apply to much of the published research. For many of the contributing factors discussed in the report, there are multiple studies in different settings confirming that the factors do indeed contribute to clinician burnout. Many of the studies examining interventions, however, are limited to a single site. Study sites may have unique features that are difficult to quantify (or even identify or describe) and that may affect the outcome of the intervention. While there is much to learn from the evidence established from single-site studies, their scalability and generalizability are, in many cases, untested. This is important because interventions with positive outcomes in one location may not work in the same way elsewhere without local adjustments.

NEEDED RESEARCH

Moving forward, the research on clinician burnout and professional well-being should employ robust quantitative or qualitative methodologies and study design principles, should include objectively obtained measures of key independent variables, and should use measures of burnout and professional well-being with strong psychometric properties. Validated

instruments to measure burnout should be used, as discussed in Chapter 3. A review on this topic has recently been published (NAM, 2018). Similar to the case with the assessment of mental health issues, such as depression, more than one measure can be used to measure the same construct. Although it is not necessary for every study to use the same instrument, each study should use a validated instrument with defined performance characteristics. Although the Maslach Burnout Inventory (MBI)–Human Services Survey is the most widely used instrument and has the most evidence for its validity, other validated tools are available. Other available instruments to measure burnout include the Oldenburg Burnout Inventory, the Copenhagen Burnout Inventory, and the Stanford Professional Fulfillment Index. Instruments to measure dimensions of professional well-being, such as professional fulfillment and engagement, also exist but would benefit from additional validity work, particularly with respect to predictive validity (i.e., whether the scores correlate with outcomes of interest for health care). Instruments are available to measure other professional well-being outcomes of interest, such as compassion fatigue, posttraumatic stress disorder, and engagement. In developing instruments to measure new constructs of professional well-being, it will be essential to use established rigorous scientific processes of instrument development and validation, paying particular attention to concurrent and predictive validity.

Another high-priority research need identified by the committee is the carrying out of longitudinal study designs that enable the exploration of causation as well as the trajectory of burnout over time. Additionally, intervention studies should have randomized controlled or cohort study designs with crossover or appropriate comparison groups and include follow-up at least 6 to 12 months after the end of the intervention. Consideration should also be given to exploring the impact of an intervention in one clinician group on the work-lives and professional well-being of other team members. Select intervention studies should be designed at the interprofessional team level. Interventions should report on cost and be scalable. Pragmatic research that identifies best practices for the implementation of intervention strategies shown to have a positive effect on professional well-being is also needed.

Several investigators have proposed research agendas (Dyrbye et al., 2017a,b; Linzer, 2018). To move the field forward, the committee believes methodologically rigorous research should be conducted within the following five major areas:

1. Foundational epidemiologic research is needed to better define the prevalence of burnout among select groups of clinicians and learners within select health profession education programs.

Studies that include large samples of clinicians other than nurses and physicians (e.g., dentists, pharmacists, advance practice providers, and other clinicians) and learners other than medical students and residents (e.g., those in nursing, pharmacy, dentistry, and other clinical health profession education programs and fellows in graduate medical education training programs), preferably from the United States and across practice settings, institutions, and demographic groups, are needed to better define the extent of the problem of burnout in these professional groups. Special attention should be given to understanding how burnout affects underrepresented groups, such as women and racial and ethnic minorities. Studies exploring the prevalence of professional well-being at the interdisciplinary team level are also needed. These studies should employ validated instruments to measure burnout and other dimensions of professional well-being. If studied in combination with samples of other workers from the general population, a more in-depth understanding of the problem could be acquired. There also continues to be a need for well-conducted, longitudinal studies of burnout among physicians, nurses, and other clinicians across practice settings (e.g., private practices, hospitals, nursing care facilities, and community-based organizations in both urban and rural areas) and across modalities of care delivery, including in-person and virtual interactions. Such longitudinal studies in health care profession learners than span the educational continuum would provide useful information about the course of burnout among learners and about whether the experience of burnout as a learner affects the risk of experiencing burnout subsequently once in practice.

2. Hypothesis-generating research is needed to define optimal professional fulfillment and well-being.

More than the absence of burnout is required for clinicians and learners to thrive professionally and personally. A number of frameworks of professional fulfillment, engagement, and professional well-being have been developed in an effort to define this concept, and both qualitative and quantitative studies are needed to advance this work. How these constructs relate to enhanced personal, professional, and societal outcomes is also needed, as are intervention studies evaluating how system design can best cultivate and support these qualities for clinicians and learners.

3. Research is needed to identify work system factors, learning environment factors, and individual mediating factors that increase the risk for burnout or that promote professional well-being among clinicians and health profession learners.

Research is needed that identifies additional job demands and job resources and individual mediating factors that relate to clinician professional well-being and risk for burnout, recognizing that these issues may need to be examined separately for different types of clinicians (role-specific factors) and different practice settings (e.g., physician versus nurse versus pharmacist, etc.; inpatient versus outpatient; large practice versus solo or small group practice; metropolitan versus rural practice). Research should also be carried out to identify the elements of the learner environment, including online education settings, that contribute to learners developing burnout and the elements that are critical to promoting learner professional well-being. Specifically, longitudinal studies are needed to better identify external environment, health care organization (HCO), learning environment (as applicable), frontline care delivery, and individual factors that contribute to burnout and enhance professional well-being among clinicians and health profession learners. A better understanding of the effects of health information technology on nurses and other non-physician clinicians is also needed, as is a better understanding of the relationship between work and learning environments and the impact that changes in one have on the other and on learners. Efforts should incorporate measures of workload, work complexity, teamwork, professionalism, the learning environment, and other relevant factors (DiAngi et al., 2017; Josiah Macy Jr. Foundation, 2018). Select research priorities within this domain are listed in Box 9-1. Investigators should approach their work with the understanding that identified factors and their relative contributions may vary across different types of clinicians, demographics, career stages, and learners.

4. Research is needed to gain further understanding of the implications of clinician and learner burnout and professional well-being on patients, clinicians, learners, health care organizations, and society.

Prospective study designs that measure a variety of independent outcomes (rather than self-report), such as quality, safety, and costs of care, are needed. Studies are needed to advance our understanding of the personal and professional consequences of professional burnout (and its levels of severity), high engagement, and professional well-being. Economic models that estimate the costs of clinician burnout across disciplines, including medicine, nursing, pharmacy, dentistry, and other clinical disciplines, are also needed. Research priorities within this domain are listed in Box 9-2.

5. Research is needed to evaluate systems-based interventions to prevent and mitigate the risk of burnout and optimize professional well-being across the career span as well as help clinicians and learners with burnout recovery.

Methodologically robust intervention studies are needed to identify effective systems approaches to preventing and mitigating the risk of clinician and learner burnout and to identifying evidence-based practices to help individuals recover from burnout. This research should identify system approaches to optimizing clinician and learners’ professional well-being, including strategies that assist individuals in optimizing the personal factors that mediate stress response. How best to engage clinicians, learners,

and patients in the design of interventions (and system redesign) needs to be explored, as does how best to facilitate these individuals’ engagement in research studies. Potential targets for intervention research are listed in Box 9-3 and will be further informed by the research proposed in Boxes 9-1 and 9-2. Interventions that include diverse groups of clinicians and learners as well as studies that target the unique needs of each type of clinician and learner (e.g., physicians, nurses, nurse practitioners, physician assistants, pharmacists, and other clinical disciplines) are both needed. Studies should explore if improving the work environment also improves the learning environment and examine the similarities and differences between effective interventions. Strategies that are effective under research design settings should subsequently be tested further using principles of implementation science to

determine the best way to scale and spread such advances. Studies should also explore the unintended consequences of interventions (e.g., impact on other members of the health care team, impact on the work or learning environment). Finally, intervention research that assesses not only the impact on clinician burnout and professional well-being but also the downstream impact on quality, safety, cost of care, and access to care is critically needed. Research priorities within this domain are listed in Box 9-3.

KEY FINDINGS

As documented throughout this report, much is known about the prevalence, causes, and consequences of clinician burnout in physicians and nurses. Less is known about the issue in other clinicians and learners. There is also little known about systems-based approaches to mitigating burnout and promoting professional well-being in all clinicians, including physicians and nurses. The proposed research agenda is robust and has the potential to be transformative. Success in moving the research agenda forward and, ultimately, reducing suffering and improving patient care outcomes will require methodologically strong studies, substantial funding, and collaboration.

A multi-pronged approach involving all stakeholders to addressing research barriers is critical to realizing viable and sustainable solutions. Such effort must be coupled with HCOs making clinician professional wellbeing a priority, surfacing and testing new ideas, and sharing them with one another and the field to accelerate improvements within their individual organizations and the field at large.

REFERENCES

NAM (National Academy of Medicine). 2018. Valid and reliable survey instruments to measure burnout, well-being, and other work-related dimensions. https://nam.edu/validreliable-survey-instruments-measure-burnout-well-work-related-dimensions/#purpose (accessed April 17, 2019).

Panagioti, M., E. Panagopoulou, P. Bower, G. Lewith, E. Kontopantelis, C. Chew-Graham, S. Dawson, H. van Marwijk, K. Geraghty, and A. Esmail. 2017. Controlled interventions to reduce burnout in physicians: A systematic review and meta-analysis. JAMA Internal Medicine 177(2):195–205.

West, C. P., L. N. Dyrbye, P. J. Erwin, and T. D. Shanafelt. 2016. Interventions to prevent and reduce physician burnout: A systematic review and meta-analysis. The Lancet 388(10057):2272–2281.

This page intentionally left blank.