As the festive season rolls around we think of time with family and friends, holidays and giving. But for anyone working in health care we also think of the increased burden of care the winter brings, rotas and busy shifts that mean missing those special moments with loved ones, and the sense of duty to ensure our patients are well provided for over the public holidays and weekends. It’s a no-brainer that over time, this can take its toll.
In a recent blog I explored the impact of the increasing clerical burden on doctors associated with e-health records, and their association with burnout. With studies showing levels of compassion fatigue and other components of burnout to affect more than half of doctors, it is perhaps timely to look at what we know about how to prevent it in the first place.
Without ways of fostering resilience and the abilities to tolerate the stressors that a career in medical practice exposes us to, we continue to damage our health, curtail our careers and ultimately do a disservice to our patients. At the very least I hope it will give pause for thought to be kind to yourself and to your colleagues through the long dark nights and short cold days of the deep mid-winter.
In a recent study published in The Lancet in September, West and colleagues review the available evidence on what can be done to prevent burnout amongst physicians and offer some pointers to individuals and organisations alike, aimed at reducing burnout rates (West et al, 2016).
They start by laying out the scale of the problem highlighted in previous studies: over half of practicing doctors and those in training are affected, with hallmark symptoms of emotional exhaustion, reduced feelings of accomplishment and depersonalisation. Further studies have revealed an association with burnout and poorer patient care, professional conduct and links with mental health problems and road traffic accidents. On the positive side it has been seen that even modest changes on self report scales of burnout can translate into a meaningful improvement for the doctor.
The authors completed a systematic review and meta-analysis of articles relating to burnout and stress (as well as other similar terms) that included measures to prevent and reduce physician burnout within the major medical databases (MEDLINE, Embase and PsychINFO) using PRISM standards.
They found 2,617 potential articles, with some 230 being reviewed in depth and having the inclusion and exclusion criteria applied. Studies had to include data on doctors (not medical students or other professionals), look at some sort of intervention to reduce burnout, and have physician specific burnout measures collected using validated rating scales. This left them with a total of 15 Randomised Controlled Trials (RCTs) and 37 Observational studies.
Appropriate statistical analysis was then used to compare the data from individual studies, measure the heterogeneity of the study populations, and to examine for bias within the published works.
The included studies covered a broad range of different interventions.
- 3 RCTs looked at organisational changes – altered working patterns, shift length, rotation length
- 12 RCTs looked at changes at an individual level – generally stress management, communication skills or self care training
- Some of these were funded to cover clinicians participating during their working day, and overall they encompassed studies in medical and surgical specialties across a range of professional grades.
- 17 cohort studies looked at an organisational level tackling practice delivery and shift length requirements.
- 20 cohort studies were individually focused and covered such interventions as small group curricula, stress management, communications skills and mindfulness training
- A smaller proportion were funded during working hours, but again they covered the full breadth of specialties.
Almost all studies used the Maslach Burnout Inventory (MBI) and only 1 RCT and 3 cohort studies used a different measure.
The authors pooled data across the studies and commented on how the interventions had affected the the following domains across both RCTs and cohort studies separately:
- Overall burnout
- Pooled data from 5 RCTs and 9 observational studies showed a reduction from 54% to 44% (10% overall difference, 95% CI 5 to 14, p<0.0001) with no difference between RCTs and cohort or for residents versus practicing clinicians
- There was a statistically significant difference showing that structural changes to the working patterns reduced overall burnout more than individualised intervention (p=0.03) especially changes to duty hours worked.
- Emotional exhaustion (EE)
- Pooled data from 12 RCTs and 28 cohort studies showed significantly reduced EE scores of 2.65 points (95% CI 1.67 to 3.64, p<0.0001) with no differences by type of study, grade of doctor or organisation vs individual interventions
- Changes to duty hours and mindfulness training had greater effects 2.88 points (95%CI 1.17 to 4.59, p=0.0001) and 4.68 points (95% CI 2.84 to 6.51, p<0.0001) respectively.
- Pooled data for 11 RCTs and 25 cohort studies showed a significant but small reduction of 0.64 points on depersonalisation scales (95% CI 0.15 to 1.14, p=0.01) with no differences between the type of study, grade of doctor or organisation vs individual interventions
- Again duty hours and mindfulness led to greater than average improvements (1.53 and 2.01 point reductions respectively).
- High emotional exhaustion
- 8 RCTs and 13 observational studies looked at high EE, showing a reduction from 38% to 24% (difference 14%, 95% CI 11 to 18, p<0.0001)
- Again type of study, or organisation vs individual interventions made no difference, but the interventions appeared slightly more effective in practicing staff compared to residents.
- High depersonalisation
- Pooled data for the 6 RCTs and 10 cohort studies looking at high depersonalisation showed a reduction from 38% to 34% following intervention (difference 4%, 95%CI 0 to 8, p=0.04)
- There were again no significant differences due to type of study, grade of doctor or organisation vs individual interventions
- Changes to duty hours and mindfulness training again showed a better response of 6% and 5% reductions respectively.
Overall there were generally few ill effects from conducting these interventions, although 4 studies reported things like dissatisfaction with new working hours, reduced ability to attend educational training, dissatisfaction with the training programme, and reduced ability to assess trainees’ clinical skills and patient care.
The researchers commented on the possibility of biases inherent in the study designs, due to the inability to blind participants and raters, nor to control for confounders, and the absence of control groups for comparison in many studies. They constructed funnel plots and did not see evidence of publication bias.
West and colleagues conclude that a range of interventions at both organisation and individual level make a significant difference to burnout levels, emotional exhaustion and depersonalisation. Even small reductions in the overall scores can mean a considerable reduction in major adverse outcomes, and, because of the narrow range for the scores, can lead to people moving from being categorised as high levels for each domain into the average level category.
They note that both organisational and individual changes can be effective, but that both are likely to be needed to get the most satisfactory outcomes; although no studies looked at this combination.
They are clear that more research should be undertaken to determine which interventions give best value for time and money to get the most effective reductions in burnout scores amongst physicians.
Strengths and limitations
The study constitutes a comprehensive evaluation of the effect of a variety of interventions on different scales within a health care system, and finds that there is considerable scope for these to make a difference to doctors’ rates of burnout.
The data are limited in that only few long term follow up measures were taken, making it difficult to know how long the effects of intervention last. Little is known from the pooled studies about the demographic make-up of the subjects in the trials, and many of the studies were inherently prone to biases through their design. There was little that could be done to control for confounding factors or to provided blinding.
The authors note that better designed and more generalisable studies are needed to determine the most effective interventions in reducing burnout amongst medical practitioners.
Implications for practice
- A range of interventions are available for health care organisations and individual doctors that can lead to a meaningful reduction in burnout rates, emotional exhaustion and depersonalisation in clinical settings.
- More work is needed to identify the individual factors and systems failures that lead to burnout developing.
- Work to help doctors become better aware of the early stages of burning out, so that (a little bit like Lord Voldemort) burnout can be recognised, named as a specific problem, and sorted out before it leads onto major difficulties.
- This will take efforts not just from clinicians, but also their leaders, managers, co-workers and ministers to build healthy systems which promote and sustain our working lives.
West CP, Dyrbye LN, Erwin PJ, Shanafelt TD. (2016) Interventions to prevent and reduce physician burnout: a systematic review and meta-analysis, The Lancet, Available online 28 September 2016, ISSN 0140-6736, http://dx.doi.org/10.1016/S0140-6736(16)31279-X.
Epstien RM, Privtera MR. (2016) Doing something about physician burnout. The Lancet, Volume 388, No. 10057, p2216–2217.