1/ Several provocative papers were published on #burnout last week in @JAMA_current. I have some thoughts on both the papers and the field. I hope these might stimulate further informed discussion. I will take part in that to the extent time permits, but I do have a day job. 😀
2/ First up, the systematic review by @LisaRotenstein @srijan_sen_lab @DouglasMataMD and others. Hearty congratulations for such a visible publication based on an immense amount of work! #burnout @JAMA_current jamanetwork.com/journals/jama/…
3/ Key findings include: “marked variation in burnout definitions, assessment methods, and study quality” which “preclude definitive conclusions about the prevalence of burnout”.
4/ The authors “highlight the importance of developing a consensus definition of burnout and of standardizing measurement tools …” The positive aspects of these points bear emphasizing.
5/ Many of the instruments identified in the @LisaRotenstein paper are poorly validated as employed. Of course, validity is itself a complex construct as @CookMedEd has noted: ncbi.nlm.nih.gov/pubmed/16443422
6/ Applying inconsistent definitions or “made up” subsets of existing tools should be frowned upon, and the authors correctly throw a flag at these practices.
7/ .@NAMedicine has published a report on many of the tools in common use, and their overall validity evidence: nam.edu/valid-reliable…
8/ The authors note the complexity of #burnout as a construct defined by @CMaslach @workengagement . With 3 domains, each scored continuously, and debate about the relative importance of each domain, confusion seems inevitable. Branching out to other tools only adds to this.
9/ With this as background, I have some thoughts about the paper itself, and the questions it poses. I will not cite every key paper, but my advice is that everyone should become better educated on the decades of literature on the #burnout construct.
10/ That means both the pros and cons, and the strengths and limitations of the critiques and praises themselves. Dr. Renzo Bianchi has written consistently about overlap between #burnout and #depression, for example.
11/ #burnout and #depression are correlated but clearly (in my view) not simply variations on the same theme. Citing Bianchi without careful review of the many flaws in those arguments is quite nonscientific.
12/ Similarly, citing critiques of the MBI such as those in the development of the Copenhagen Burnout Inventory is also … interesting. Yes, Kristensen raised a number of concerns about the MBI.
13/ However, in the SAME ISSUE of Work and Stress in 2005, Schaufeli rebutted most of the critiques. You have to read both sides, not just one. wilmarschaufeli.nl/publications/S…
14/ To that end, I note that Schaufeli has himself criticized the MBI for its limitations. He and @CMaslach have written together on these topics to advance the field, and both acknowledge room for improvement. This has evolved over several decades.
15/ Having had the privilege of meeting and presenting with @CMaslach @workengagement and Dr. Schaufeli, their challenging yet cordial intellectual debates are inspiring.
16/ Let’s use that as a model rather than using language that may be easily misconstrued as undermining the entire body of work dating back decades. That’s not a critique specific to the present systematic review, but rather relevant to several works in recent years.
17/ Now, to the paper itself. Again, this is an impressive piece of work, with more than 200 citations. The effort required to produce this is immense and worthy of the greatest respect. Well done @LisaRotenstein @srijan_sen_lab @DouglasMataMD @JAMA_current !
18/ In appraising any systematic review, the first validity question is “Did the review explicitly address a sensible clinical question?” jamaevidence.mhmedical.com/book.aspx?book… This study offers two main questions it attempts to answer.
19/ The first concerns how burnout among practicing physicians has been defined across the literature. That qualitative summary seems eminently sensible to me, no concerns there at all.
20/ The second concerns the prevalence of burnout in practicing physicians. This is more problematic. Because “practicing physicians” is a very broad group, I would expect evidence across subgroups of physicians to differ, perhaps markedly.
21/ The review identified nationally representative studies along with studies of more narrow subgroups of practicing physicians. I’m not sure one would ever expect to be able to meaningfully pool these results.
22/ When variability due to instruments used, country/health care system, and year is added into the mix, the result is sure to be messy. In systematic review terms, we’ll see heterogeneity from many sources.
23/ Collecting this level of variability is interesting and informative, but given the question it should be expected, not viewed as evidence of a limitation of #burnout as a construct.
24/ A few more specific points. The authors cite Bianchi in writing of a once a week or more threshold for #burnout symptoms that “Symptoms experienced this infrequently are unlikely to reflect clinically meaningful levels of burnout”
25/ Except that a wealth of literature (not just studies my teams have been part of) has identified impactful associations of that frequency of symptoms with key outcomes relevant to nearly any health care stakeholder. These hold up whether continuous or dichotomous.
26/ Whether this reflects “symptoms of burnout” or a “clinical burnout syndrome” seems irrelevant to the importance of these feelings for physicians and their patients. By all means, let’s continue to refine our understanding of #burnout.
27/ However, suggesting that we have no idea what we’re measuring (I’ll offer thoughts on the editorial by Dr. Schwenk and @KGoldMD later) does not accurately reflect the state of the field.
28/ As of now, I see little to support a conceptualization of burnout as a clinical diagnosis treatable by psychotherapy and medication. I am open to being proven wrong about this, but the roots of burnout are primarily in our systems – on this, the authors and I seem to agree.
29/ I am not a psychiatrist or a psychologist, however. Some of the needed work would benefit from continued engagement of experts in those fields, including several of the authors.
30/ I disagree strongly with preference for the Copenhagen Burnout Inventory over the MBI. Yes, it’s free, and I sure wish the MBI was as well. However, the CBI has also been roundly criticized for its own limitations and has far less evidence in support of its validity.
31/ The MBI has limitations, freely acknowledged by its developers over several decades. To date, though, no alternative has been strong enough to replace it as the best supported instrument for #burnout as defined most commonly.
32/ Is there a better definition of #burnout to be made? Perhaps, but it hasn’t happened yet. Until it does, the Maslach construct seems to work well enough, and checks off validity criteria better than other proposals.
33/ These are my opinions as a clinician and researcher in the field who is still learning. Debate and discussion is healthy, and I applaud the authors for stimulating this debate. Let’s be sure not to lose sight of the larger issues and needs while we hammer out the details.
34/ These thoughts are by no means meant to be exhaustive. I may well also be wrong about some things. Won’t be the first or last time! Agree with me? Disagree with me? RT to spread the word, and let the discussion ensue! @dyrbye @ChristineSinsky @LarissaThomasMD @JAMA_current
