Article Text

Download PDFPDF
The physician burnout conundrum: where do we go from here?
  1. Kyle M Fargen
  1. Neurological Surgery and Radiology, Wake Forest University, Winston-Salem, North Carolina, USA
  1. Correspondence to Dr Kyle M Fargen, Neurological Surgery and Radiology, Wake Forest University, Winston-Salem, North Carolina, USA; kylefargen{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Over the last 10 years, physician burnout has become a hot button topic within medicine. To put the recent focus on burnout into perspective, a PubMed query using the keywords “physician burnout” reveals over 5000 individual articles, with over 2000 of these published in the last 3 years alone, encompassing nearly every field in medicine. Our specialty is no different: we recently surveyed neurointerventionalists both prior to1 and during the coronavirus pandemic,2 as well as interventional radiology staff3 regarding burnout prevalence and predictive factors. These studies found that roughly half of both physicians and non-physician staff met criteria for burnout based on self-reported inventory scores. These studies mimic the prevalence of burnout across all medical specialties; an often cited survey from 2015 demonstrated that roughly 50%–60% of doctors, regardless of practice, exhibited burnout.4 Consequently, professional burnout is now widely accepted among the medical community and has even recently been included in the International Classification of Diseases (ICD-11).

These observations generate obvious questions. Is nearly half of our workforce truly composed of disaffected, exhausted, walking zombies? Are our patients at grave risk of medical errors or staff complacency from impaired physicians and nurses? The answer to these questions is almost certainly ‘No.’ But why then is there such a disconnect between the data and real life?

Burnout and how it is measured

The ICD-11 describes professional burnout as a triad of (1) feelings of energy depletion or exhaustion; (2) feelings of negativism or cynicism related to one’s job; and (3) sense of ineffectiveness and lack of accomplishment. This definition is drawn from the original description of burnout and its components by Maslach and Jackson in 1981.5 These authors developed the Maslach Burnout Inventory (MBI) and its derivations to score three major domains (emotional exhaustion, depersonalization, and personal accomplishment). Each score was designed to lie on a spectrum, with higher exhaustion and depersonalization scores and lower accomplishment scores correlating to greater degrees of experienced burnout. Alternative survey instruments have been developed to study burnout (Stanford Professional Fulfillment Index, Mini-Z Burnout Survey, Copenhagen Burnout Inventory, etc) but these are not as frequently used.

It has become routine in published research to dichotomize scores from burnout inventories into ‘burnout’ versus ‘no burnout’ categories. However, the majority of burnout instruments, including the MBI, were designed to measure experienced burnout on a spectrum using Likert scales. In fact, the MBI was never intentioned to stratify respondents into distinct categories.5 Furthermore, the thresholds chosen in previous studies to define burnout versus no burnout are mostly arbitrary: the thresholds have not been validated statistically in terms of risk of burnout adverse outcomes (such as suicidal ideation, medical errors, etc). Even the most widely cited authors on burnout routinely identify respondents with a symptom of burnout, which is often mistaken by readers to mean that those individuals are ‘burned out.’ One could argue that this strategy of dichotomization is reasonable in that it, at the very least, segregates respondents with more burdensome experienced burnout versus those with less, but this approach lacks validity.

Causality between burnout and patient care metrics

The overwhelming majority of published physician burnout studies are cross-sectional in nature and have identified associations between burnout scores and medical errors, malpractice lawsuits, poor patient care practices, depression, alcohol abuse, and suicidal ideation.6–10 Importantly, these studies are regularly cited in the burnout literature as demonstrating the high toll of burnout on patient care, yet these studies do not assess causality. Are higher burnout scores the cause of medical errors, or is professional dissatisfaction a natural consequence of making mistakes and hurting patients in a career that you are devoted to? Similarly, are higher scores due to dealing with the added stress of a malpractice lawsuit, or was burnout causative in the malpractice claim in the first place? These questions are exemplified by a recent study by Menon and colleagues of over 1300 physicians comparing burnout scores from three separate inventories, depression, suicidal ideation, and medical errors. Interestingly, when controlling for depression, burnout scores had no association with suicidal ideation, while depression did.11 This study highlights the lack of clarity between quantified burnout and the risk to physicians regarding outcomes of interest. There is essentially no high-quality evidence on this topic to date, but burnout has become an accepted and widespread scientific entity plaguing healthcare workers (an ‘epidemic’).

Meaning is lost: if everyone is burned out, no one is burned out

What does it mean to be ‘burned out’ if 50%–60% of physicians meet these criteria? In the Menon study, suicidal ideation over the prior year was reported by 5.5% of respondents.11 While this number is woefully high and concerning, this percentage is 10 times lower than the number of physicians4 identified as having burnout using conventional reporting. This disconnect between the way we identify burnout and what actually matters poses serious challenges.

Based on the way the current burnout instruments are designed, any physician could easily answer subjective questions in a manner that would suggest they are experiencing burnout, particularly if they are feeling tired from call or had a few recent bad outcomes. After a particularly rough week of stroke call, most could say they are fatigued, feel like they’re not helping anyone, or are disconnected with their patients emotionally. However, there is a big difference between these occasional stressful periods, common to all physicians, and pervasive, inescapable, and dangerous thoughts and actions. The current instruments and conventions for measuring and reporting burnout do not adequately differentiate these two.

Failing to differentiate between the minority of individuals with true burnout and the larger majority who are dealing with manageable periodic stress invalidates the term and its associations. The definition of burnout is therefore of little academic value in identifying and mitigating risk factors for developing burnout, which limits the value of research in this realm. Second, the colloquial use of the word ‘burnout’ has become a synonym for ‘tired’ or ‘bored’ and is now recognized by many as synonymous with ‘whining.’ This is particularly damaging if administrative bodies merely accept a “most doctors are burned out, it’s normal” mentality, failing to act when one of their physicians is truly at breaking point.

Are we all burned out or are we living the natural history of human satisfaction?

Metrics of well-being and satisfaction in the general population have been well-documented to follow age-related curves. figure 1 showcases overlapping estimated curves for overall happiness,12 health,13 marital satisfaction,14 and worrying15 based on age. Note that overall unhappiness, marital dissatisfaction, and worrying peak during the ages of independent physician practice (35–60 years), while physical health continues to deteriorate. In fact, a large physician burnout study demonstrated very similar trends as the overall happiness U-shaped curve demonstrated in figure 1: 50.5% of early-stage physicians (median age 38 years), 53.9% of mid-career doctors (median 49 years), and 40.4% of late-career physicians (median 61 years) reported burnout.16 These studies collectively suggest that it is only natural and expected for individuals to feel more dissatisfied with their overall lives and careers during these at-risk (‘mid-life crisis’) years.

Figure 1

Overlapping curves estimating the natural history of satisfaction and well-being in the general population.

But physicians are different than the general population, right? Aren’t our jobs naturally more stressful, putting us at high risk of burnout? While survey data suggests that roughly two-thirds of American full-time employees report job-related burnout,17 studies do indeed suggest that doctors and trainees work longer hours, have higher emotional exhaustion, depersonalization and burnout, and report less satisfaction with work–life balance than the general population.18 However physicians harbor protective factors as well: we have higher levels of resilience19 and do not have to contend with financial, housing, or healthcare insecurities. Further, doctors are one of the most trusted and respected occupations in society.20 These reports suggest that physicians may have factors that both promote and protect against burnout compared with the general population, but our overall occupational risk does not seem excessive compared with other types of work.

Future of physician burnout

Burnout is a complex problem that we need to better understand. How do we make burnout a more useful scientific entity in the future? We should separate ‘experienced burnout’ from ‘burnout.’ New nomenclature should be developed to describe physicians that are experiencing true burnout and unable to practice safely or are at high risk of suicide, such as ‘occupational stress impairment.’ It is important that physicians that are truly impaired are differentiated from those that are experiencing normal or temporary work-related or life stress. Categorization of burnout scores into risk groups needs to be validated in terms of outcomes. Using conventional arbitrary composite score thresholds, or the mere presence of a symptom of burnout, as the means of stratifying respondents is scientifically misguided. Continuing to use this methodology as ‘convention’ (of which I am personally guilty!) only perpetuates the problem. Identifying risk of suicidal ideation or errors at different score thresholds will allow for more scientifically rigorous evaluation of the utility of the instrument itself as well as study of risk factors and benefit of interventions.

Ethics statements

Patient consent for publication

Ethics approval

This study does not involve human participants.



  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Commissioned; internally peer reviewed.