In this post, we’re offering summaries and comments on some articles from the BMJQS’s Top 10 of 2019. To check out a full list of our 20 finalists, click here. The editors and the Editorial Board used data such as citation rates and social media engagement in addition to their own judgement to create these lists. Overall, they demonstrate the wonderful breadth of valuable articles published in BMJ Quality & Safety.
- Rate of avoidable deaths in a Norwegian hospital trust as judged by retrospective chart review – January 2019
In reviewing 1000 consecutive nonpsychiatric hospital deaths in a trust in Norway, Rogne et al. classified 4.2% of deaths as avoidable (defined as > 50% chance of avoidability). Ultimately, invasive procedures accounted for one-third of avoidable deaths. Interestingly, deaths classified as avoidable were much more likely to occur in patients with only mild or no functional impairment. Only 3 of these 42 deaths were identified simultaneously from the hospital’s mandatory reporting system, suggesting that strategies such as periodically reviewing an otherwise unselected sample of deaths may provide a more accurate window into avoidable hospital mortality and related quality of care. The study’s findings are consistent with Hogan et al.’s classic study in England from 2012 and further adds to the critique of claims of high rates of deaths from medical error.
- Effectiveness of double checking to reduce medication administration errors: a systematic review – August 2019
Medication administration is a process repeated hundreds of times daily in hospitals, but it is prone to error. Double-checking meds using two nurses rather than one is a frequently used safety practice with high face validity, particularly for high-risk meds. However, this takes time from both nurses, potentially impacting overall staffing and workflow on units and generating other unintended consequences. Using systematic reviewing methods, Koyama et al. identified 13 articles measuring differences between processes of double-checking meds for verification and at least one outcome of medication adverse events (MAEs) or adherence to hospital policy. Only three studies were considered to be of ‘good’ quality. One showed an association between double-checking meds and a reduction of MAEs in simulated patient scenarios, while another showed a reduction in MAEs when utilising this practice across 122 patients over two months in a large academic hospital (OR for MAE 0.44, 95% CI 0.27-0.72). The third good quality study showed no association with a reduction in MAEs. Studies reporting adherence showed rates ranging from 52%-97%. Given this relative paucity of higher quality studies, the authors conclude that no best practices for effective double-checking of meds have been identified. This is in many ways a frustrating finding: a very common practice without much of an evidence-base. The need for high quality studies is pressing.
- Exposure to incivility hinders clinical performance in a simulated operative crisis – September 2019
Incivility towards colleagues is unpleasant – but does it matter for performance? Here, Katz et al. studied completion of critical checklist elements and overall performance relative to that expected by training level in anaesthesia trainees exposed to either a courteous or uncivil surgeon during a simulated intraoperative haemorrhage. Across the 67 simulated encounters reviewed, those encountering an uncivil surgeon felt the simulated behaviours negatively affected their performance (65% affected in experimental group vs. 25% in controls, p=0.009). Although these individuals did not report any difference in their perceptions of their overall personal performance (p=0.112), five of nine aspects of the critical checklist were performed significantly less frequently in those working with uncivil surgeons. Overall, significantly fewer trainees exposed to uncivil behaviour performed at the level expected of their training year when reviewed by multiple raters (63.6% met standards in the experimental group vs. 91.2% of controls, p=0.009). When adjusting for multiple variables within a regression model for overall performance, only exposure to incivility was significantly correlated with worse performance (OR 0.110, 95% CI 0.022-0.544). The finding that communication, performance of technical skills, and decision-making are all impaired with exposure to incivility is deeply troubling. Good evidence on how to address incivility effectively is now needed.
- Sicker patients account for the weekend mortality effect among adult emergency admissions to a large hospital trust – March 2019
The apparent ‘weekend effect’ of increased mortality in patients admitted on Saturdays and Sundays has been noted for several years and across multiple health systems. Recent data suggest this effect may be secondary to fewer relatively low-risk patients being admitted on weekends, with the remaining necessary admissions therefore being more severely ill. In this study, Sun et al. reviewed >163,000 admissions over four years at a large university hospital and extracted multiple demographic factors and patient data elements. This enabled the authors to calculate National Early Warning Score (NEWS) scores within the first 24 hours of admission. They found the number of admissions was 25% lower on weekends (91 vs. 120 per day), mostly driven by a mean 17 fewer referrals from GP clinics and five fewer from other outpatient clinics on weekends. Additionally, more patients either were admitted directly or transferred to the ICU within 24 hours on weekends (4.2% vs. 3.0%). Of the 90% of admissions with a calculated NEWS score, those admitted on weekends had a significantly higher mean score (1.8 vs. 1.7, p=0.008) and a significantly higher proportion fell into a more severe category. Overall, 5.6% of weekend admissions died within 30 days vs. 5.0% of weekday admissions (OR 1.13, 95% CI 1.08-1.19). However, adjustment with NEWS scores alone attenuated this effect (OR 1.02, 95% 0.96-1.08). The weekend mortality effect therefore seems to be an effect of sicker patients. This finding does not obviate the need to investigate quality of care at weekends, but it does suggest that attention to patterns of admission is needed as a basis of improvement practice and policy.
In this study, Jones et al. facilitated an intervention for senior hospital leaders to develop and implement organisation-wide quality improvement (QI) work at six provider organisations within the National Health Service (NHS). At the outset, three of the six were rated as ‘requires improvement’ by the NHS regulator. Each site nominated three board members for a series of events with the goal of developing specific organisation-wide QI strategies and implementing at least one broad QI project per site. In their qualitative analysis, the research team found that three of six (including one rated as ‘requires improvement’) had a positive outcome. The defining features for successful organisational responses related to the availability of ‘slack’ institutionally and the functioning of the board. Organisations required to follow external improvement plans had limited local autonomy and restricted funding for other initiatives. In contrast, ‘slack’ was available in organisations not experiencing external pressure, allowing them time and energy to focus on additional priorities and possibilities. The three successful organisations had stable leadership and were able to develop a shared vision for QI, while the other three had significant personnel turnover or only superficial governance by the board of directors. Ultimately, sustainable and continuous QI may require institutional slack to be created and maintained within organisations through stable leadership and functioning governance. Paradoxically, this may mean the organisations with the biggest need for broad improvement may be the least likely to achieve it.