Burnout in doctors and the quality of patient care: Our systematic review

There is increasing evidence that healthcare staff burnout is linked to a range of negative outcomes, including increased staff absences, higher rate of staff turnover, and poorer quality of patient care (see my previous blog on this here). In a systematic review I co-authored, we found that that 21 out of 30 (70%) studies looking at the link between higher staff burnout and poorer patient safety reported a significant association between the two (Hall et al., 2016). The review found hard evidence for what many clinicians could see happening in their wards and surgeries: when staff are hard pressed, patient care suffers. It was well received on social media and has since been cited dozens of times.

However, while this previous review found clear evidence for a link between staff burnout and patient safety, it seemed to me that two questions remained unanswered. The first was whether there is also a link between burnout and other aspects of patient care, such as patient satisfaction. The second was what the strength of this relationship is: that is, just how much do increases in burnout impact patient care?

So, when I was invited to contribute to a systematic review on the links between burnout and patient care in doctors by Dr Maria Panagioti, I jumped at the chance. The review led by Dr Panagioti aimed to answer both these questions. It gathered studies which investigated burnout in doctors in relation to a broader range of outcomes, including:

  1. Patient safety incidents, (e.g., adverse events, medication errors, diagnostic incidents)
  2. Low professionalism (e.g., adherence to treatment guidelines, quality of communication, malpractice claims, empathy)
  3. Low patient satisfaction

It also quantified the strength of these relationships using meta-analysis, which was not employed in the previous review.

Is burnout linked with patient safety incidents?

The review identified 21 studies which reported on the association between burnout and patient safety incidents. The results of the meta-analyses suggested that burnt-out doctors were twice as likely to be involved in a patient safety incident as those not suffering from burnout. All aspects of burnout (exhaustion, disengagement and low accomplishment) were associated with a significantly higher risk of being involved in a patient safety incident.

Is burnout linked with low professionalism?

28 studies were found which reported on the link between burnout and low professionalism (e.g., showing low empathy, having received a malpractice claim). The results of the meta-analyses suggested that burnt-out doctors were twice as likely to show low professionalism. When the different aspects of burnout were examined separately, disengagement was the aspect most linked with low professionalism. Doctors who were disengaged from their patients were 3-times as likely to exhibit low professionalism. Doctors high in emotional exhaustion or low in personal accomplishment were over 2.5-times as likely to exhibit low professionalism.

Importantly, the review found that the link between burnout and low professionalism was twice as high in trainee and early career doctors compared with more experienced doctors. This is particularly concerning when the recent GMC survey results showing that a quarter of trainee doctors are burnt-out are considered.

Is burnout linked with low patient satisfaction?

7 studies reported measures of patient satisfaction. It was found that burnt-out doctors were at twice the risk of having dissatisfied patients. Again, disengagement was the aspect of burnout most closely linked with low patient satisfaction, with disengaged doctors showing a 4.5-fold increased risk. Low personal accomplishment was also linked with twice the risk of low patient satisfaction. No link was found with emotional exhaustion.

Where now?

This review finds strong evidence that burnt-out doctors are at significantly higher risk of being involved in patient safety incidents, showing low professionalism and having dissatisfied patients. Having clarified the presence and size of the problem of burnout for patient care, the next step for us as researchers is to identify evidence-based solutions to this problem. While a number of interventions to reduce burnout have been proposed (see Panagioti et al., 2017), there is a need to identify 1) which interventions are most feasible and most effective, and 2) whether reducing burnout can improve patient care.

For my previous blog on tackling burnout, please see here.

The review described in this article was published in JAMA: Internal Medicine. To read it, please see here.

Can training improve doctors’ skills in breaking bad or difficult news? Our review

Breaking bad or difficult news in healthcare settings is challenging. What is the best way to tell someone they have cancer? How do you find words to say their treatment hasn’t worked? When I first became interested in this topic, I found a large body of literature had explored these questions. Studies suggested that the way difficult news is delivered has a lasting impact on patients’ subsequent symptoms of depression (Mager & Andrykowski, 2002), and that these events can have a negative impact on healthcare professionals themselves, increasing their stress levels (Shaw et al., 2013).

Despite this, one thing was still unclear: can this situation be improved? Can we train healthcare professionals to be better at delivering difficult news? This question seemed so big, so obvious, that my co-author Dr Maria Panagioti and I were sure it would have been answered before. We were surprised; while several individual studies had tested an intervention to break bad news, no one had yet systematically reviewed these studies and analysed the data together using meta-analysis. Meta-analysis combines results from multiple studies, providing better evidence about whether an intervention works and how effective it is than individual studies can provide. At first I was frustrated, why had no one done this yet? My frustration was quickly replaced with a sense of anticipation: if no one else had done it, there was an opening to do this ourselves! We got to work.

The search process
We decided to focus our review on doctors. We searched databases for potentially relevant studies and identified 2,270 records altogether. We filtered these by reading their abstracts, leaving 71 studies for more careful inspection. After reading the full texts of these, we excluded a further 53, leaving 17 studies for inclusion. The studies we included were from a range of countries, including the United States, Hong Kong and Israel. They were also from a range of healthcare settings, including palliative care, obstetrics and primary care.

Does training improve doctors’ skills in breaking bad or difficult news?
Altogether, the studies provided data from 1322 participants to answer our main question. The results of the meta-analyses were clear: interventions improved doctors’ abilities in breaking bad news with a large effect size. This effect wasn’t explained by sources that can bias results (such as the included studies being low in quality). Interestingly, interventions were most effective when they used an established framework for news delivery, known as ‘SPIKES’ (Baile et al., 2000). SPIKES outlines clear steps that doctors should follow when delivering bad or difficult news, such as delivering a ‘warning shot’ before breaking the news itself.

Could training help doctors too?
We asked a second question of the data, which was whether interventions can improve doctors’ confidence in delivering difficult news. Again, the results were clear: interventions improved doctors’ confidence in news delivery with a medium effect size. This suggests that difficult news delivery training may be one way to help reduce the stress levels of healthcare professionals who are regularly involved in these events, which is important considering current high rates of burnout in the healthcare workforce (see my previous blog on healthcare staff burnout here).

What now?
It is now clear that physicians can indeed be trained to deliver difficult news, and this training both improves their skills in difficult news delivery and also their confidence. What is less clear is whether such interventions can improve the experience of patients receiving this news. We only identified one study which looked at this. This study was based in Japanese cancer services (Fujimori et al., 2014). It found that when physicians were trained in breaking bad news, their patients reported significantly lower depression and anxiety symptoms later on, although the effect size was small. Further research into patient experience is warranted.

A need for further research in obstetric settings
One area future research on news delivery should focus on is obstetrics. The challenges in this setting are particularly complicated, as expectant parents may have no reason to think there is anything out of the ordinary with their pregnancy. Furthermore, if a healthcare professional has identified that a baby has a disability, it can be hard to know the extent of this before birth or how it will affect their lives. Our own research suggests that some parents still have a terrible experience at this time, which leaves a strong emotional imprint (Johnson et al., 2018; for my blog on this see here). Future research is needed to understand how healthcare practitioners working in obstetrics can be better trained and supported to deliver difficult news.

To read the review discussed in this blog, see here.