Key research papers on nurse burnout

How the literature is organised

Burnout entered nursing research through the Maslach Burnout Inventory, though not every study uses all of it: Dall'Ora et al. (2020) found that 39 of the 91 studies they reviewed used all three MBI subscales. Demerouti et al. (2000), working with 109 German nurses, separated job demands, which predicted exhaustion, from job resources, which predicted disengagement — an early statement of what became the job demands–resources model. Leiter and Maslach (2009), surveying 667 Canadian nurses, found worklife areas predicting burnout and burnout in turn predicting turnover intentions, with cynicism the key dimension and value conflicts and inadequate rewards the critical areas.

The second pillar is health services research. Aiken et al. (2002) connected patient-to-nurse ratios with burnout, dissatisfaction and mortality in Pennsylvania hospitals. Vahey et al. (2004) surveyed 820 nurses and 621 patients across 40 units and found that patients on units nurses described as adequately staffed and well supported were more than twice as likely to report high satisfaction, while their nurses reported significantly lower burnout. McHugh et al. (2011), with 95,499 responses, found much higher dissatisfaction and burnout among nurses in direct patient care than among nurses working elsewhere. This work treats burnout as a signal of how a hospital is run.

Dall'Ora et al. (2020), "Burnout in nursing: a theoretical review", is the orientation piece. It sorts what has been studied against Maslach's theory, finds support for the predicted predictors but not for the predicted link to turnover, and notes that 87 of its 91 studies were cross-sectional.

Main debates

Measurement and causation are the two weak points. Prevalence estimates vary widely with the instrument, the dimension and the setting: compare Galanis et al. (2021) at 34.1% emotional exhaustion with Ramírez-Elvira et al. (2021) at 46% for low personal accomplishment in intensive care. Design is the deeper problem. Dall'Ora et al. (2020) report 87 of 91 studies cross-sectional; Li et al. (2024) report 81 of 85. That leaves open whether burnout degrades care or poor conditions produce both, and it is the limitation to state plainly in a literature review. Jun et al. (2021) is the other systematic review of burnout against patient and organisational outcomes.

Where recent work is heading

The pandemic produced a wave of prevalence studies, from Hu et al. (2020), who surveyed 2,014 frontline nurses in Wuhan and found 60.5% with moderate or high emotional exhaustion and 91.2% reporting fear, to national US data — Shah et al. (2021) found that 31.5% of the US nurses who left their job in 2017 gave burnout as a reason, with a stressful work environment and inadequate staffing the most common accompanying complaints — and ICU studies (Bruyneel et al., 2021; Guttormson et al., 2022; Martin et al., 2023).

Attention is now turning to retention and to what organisations should do. Aiken et al. (2023) asked 21,050 clinicians which interventions they wanted; Li et al. (2024) meta-analysed 85 studies covering 288,581 nurses in 32 countries and found burnout associated with a worse safety climate, more missed care, more medication errors and lower patient satisfaction — but not with standardised mortality, and not with the frequency of pressure ulcers. Reporting that mixed result honestly is better practice than quoting only the significant associations.