#KEYLIMEPODCAST 312: Neil Young was wrong, it’s not better to burnout than fade away

Link to episode


KeyLIME Session 312

Listen to the podcast


Naji et. al., Global prevalence of burnout among postgraduate medical trainees: a systematic review and meta-regression. CMAJ Open. 2021 Mar 8;9(1):E189-E200.


Jon Sherbino (@sherbino)


Deep in the middle of of the COVID pandemic might not be the best time to talk about burnout.  It’s akin to looking at someone with a broken femur (bone sticking through the skin, acute abnormal angle of the leg) and asking them if it hurts much.  However, my anecdote-informed opinion is that this pandemic is accelerating burnout in clinicians, especially (post)graduate trainees.  (As an aside, do we ever discuss burnout amoung undergradue health students? Is that because the stress or design of training is different?)

The second paper I wrote was about my personal experience with burnout as a senior resident during the SARS pandemic (a different scope of pandemic to our current experience, but the hot spot was Toronto in the hospital where I was training.) In that article I discussed the fear of contracting SARS, the anxiety about offering high quality care to my patients, the breakdown and pause of most teaching and learning, and the sustained social isolation.  It feels all too familiar (and amplified) nearly 20 years later as I talk with observe residents that I supervise.

The emotional and psychological weight of this pandemic comes at a time where health systems are already straining with aging populations, disparities in social determinants of health and rising resource costs requiring rationing of care. What is the cost that must be shoulder by health professional trainees? Were things better 10, 20, 30, 40 years ago?  Is burnout amplifying? 

Rather than relying on personal anecdote, let’s take a look at what the literature suggests.  Perhaps the retrospectoscope is out of focus?


“We aimed to establish the prevalence of burnout among postgraduate medical trainees, identify risk and protective factors, and assess whether burnout varied by country of training, year of study and specialty of practice.”

Key Points on the Methods

All of the PRISMA guidelines were observed; the authors are experienced in systematic reviews and meta-analysis.

MEDLINE, Embase, PsycINFO, CDSR, WoS, and Education Resources Information Centre were searched from inception to August, 2018. Hand searches of reference lists were completed. The only restriction was English language; studies must provide data to quantify burnout.

The Maslach Burnout Inventory (MBI)definition of burnout was used: emotional exhaustion, depaersonalization and lack of personal accomplishment.

Studies were evaluated, rated on methodological quality, and extracted of data using parallel independent raters.  Disagreements were resolved by consensus.

A random-effects pool prevalence model was used for analysis. Heterogenity was reported using the I2

A meta-regression analysis was conducted were burnout was binary categorized.

Secondary analyses included:

  • Associations between individual risk and protective factors (age, relationship status)
  • Assocations  with context (type of residency, region of training)

Key Outcomes

From 8505 studies, 196 were included, published from 1987-2018, representing 44 128 postgraduate medical trainees (PMT) from 47 countries. (Nearly half of the studies come from the USA).

The risk of bias was high with the vast majority of studies not controlling their study population (e.g. non-consequtive samples, high non-responder survey rates, lack of sample size validation etc.)

Overall proportion of burnout occurred in 31 210 PMTs in 114 studies with a pooled random-effects estimate of 47.3% (95% CI 43.1 to 51.5%), rising to 51.2% (95% CI 45.9-56.6%) in North America.There was VERY high heterogeneity among studies (I2 = 98.6%).  Restricting analysis to studies using the MBI markedly reduced heterogeneity in prevalence of burnout.

Meta-regression to assess for personal factors associated with burnout demonstrated:

  • No difference based on age
  • No difference based on sex
  • No difference based on relationship status
  • No difference based on stage (i.e. level) of training
Blue/first bar = no association
Orange/second bar = increases burnout
Grey/third bar = decreases burnout

Meta-regression to assess for environment factors associated with burnout demonstrated:
• No difference based on year of training (i.e. generation)
• No difference by specialty (e.g. medical v. surgical)
• North America experiences higher rates of burnout compared to other regions in the world (Data is limited to adequately represent all continents, but burnout is considerably higher in N. America compared to Europe, where there is adequate data for comparison.)
• There was mixed, inconclusive evidence on the effect of work hours on burnout

Key Conclusions

The authors conclude…

“Current wellness efforts and policies have not changed the prevalence of burnout worldwide. Future research should focus on understanding systemic factors and leveraging these findings to design interventions to combat burnout.”

Spare Keys – Other take home points for Clinician Educators

Systematic reviews and meta-analyses are excellent for taking a large literature and attempting to parse out a major narrative and some relevant sub-narratives.  But the challenge is always in design.  The more you pool data, the more you loose nuance.  It is akin to a painter mixing colours.  The addition of various palettes can provide contrast and bring vivacity to a picture.  However, if you mix and mix and mix colours of various qualities of paint, you end up with a brown -grey blob on the canvas hiding any attempt at a picture below.

Access KeyLIME podcast archives here

The views and opinions expressed in this post and podcast episode are those of the host(s) and do not necessarily reflect the official policy or position of The Royal College of Physicians and Surgeons of Canada. For more details on our site disclaimers, please see our ‘About’ page