In this post, we’re offering summaries and comments on articles #1-5 from BMJ Quality & Safety’s Top 10 Articles of 2021. To check out the full list of our 21 finalists, click here. To read more about articles #6-10 in our Top 10 list, 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.

5. Improving surgical quality in low-income and middle-income countries: why do some health facilities perform better than others? by S Alidina et al. BMJ Quality & Safety 2021;30:937-949. Accompanying editorial available here.

In interviewing more than 100 providers across 10 hospitals participating in the multi-phase Safe Surgery 2020 initiative in Tanzania, Alidina and colleagues sought to identify factors leading to greater improvement on a composite of 14 surgical safety and communication indicators. Examples of these 14 indicators included administration of prophylactic antibiotics within 60 minutes before the procedure, completion of instrument and supply counts post-procedure, and discussion of the anticipated procedure length and difficulty. Although all 10 hospitals improved substantially across the composite indicators, the four highest performers improved at least twice as much as the lowest three performers. Using a positive deviance framework, the authors compared themes distinguishing the highest from lowest performers. Six themes with multiple subthemes eventually emerged, with many identifying higher performing facilities as having more engaged leadership and focusing more on team improvement and collective knowledge transfer for organisational learning, rather than primarily targeting short-term improvement in individual surgical safety practices. Specifically, these higher-performing teams had freer communication with and diffusion of responsibilities to non-surgical providers than lower-performing peers during the intervention and likely had stronger cultures of teamwork prior to it. Higher performing teams were also better at adapting the learning methods to set clear targets and identify gaps in performance and addressed resistance to these changes earlier than lower performing hospitals. Often, higher-performing teams tried to engage those least enthusiastic about potential changes to take ownership over parts of the transformation. While both higher and lower-performing institutions valued the increases in knowledge, capacity, and mentorship from the initiative, maximizing the benefits of such structured programs may therefore require focus on engaging leaders and the non-technical aspects of surgery such as teamwork, communication, and collective learning.

4. Socioeconomic deprivation and ethnicity inequalities in disruption to NHS hospital admissions during the COVID-19 pandemic: a national observational study by M Warner et al. Published Online First: 25 November 2021. Accompanying editorial available here.

Comparing hospitalizations across the English National Health Service in March through December 2020 to the same months in 2019, Warner and colleagues examined the disruptions to non-COVID hospital admissions during the pandemic. The authors specifically reviewed differences in non-COVID, non-maternity emergency and elective admissions to NHS hospitals in socioeconomic and ethnicity groups known to have had greater rates of hospitalizations, serious illness, and deaths related to the pandemic. Across England, the disruption was staggering – there were 3 million fewer elective and 1.2 million fewer emergency admissions during the ten-month period in 2020 relative to 2019, reductions of 36% and 22% respectively. There also were disparities by socioeconomic status in the number of non-COVID-19 admissions during 2020, with a 2.1% larger fall in elective admissions in areas in the highest quintile for socioeconomic deprivation than in the lowest quintile. Emergency admissions, however, had an absolute 2.3% smaller reduction (10.1% relative reduction) in the highest quintile areas for socioeconomic deprivation than the lowest quintile. Areas in the highest quintile for share of ethnic minority patients had different patterns of changes in admissions, with a U-shaped reduction in elective admissions and the 3^rd quintile having the largest reduction (1.5% relative to lowest/highest quintile). For emergency admissions, however, the differences relative to the lowest quintile increased by quintile of higher non-white populations, with the highest quintile for ethnic minority population having an absolute 6.8% greater reduction than the lowest quintile, or a 37% relative reduction. Extrapolated across the English population, these reductions meant the most deprived areas had 3.0 fewer elective admissions and 2.8 more emergency admissions per 1,000 residents compared with the least deprived, and areas with the largest ethnic minority populations had 6.1 fewer emergency admissions per 1,000 residents compared to those with the smallest ethnic minority populations. Given these differences cannot be explained by admissions for COVID-19, the differences may have been driven, at least in part, by avoidance of necessary care during the pandemic, risking exacerbating existing health inequities.

3. Determining the skills needed by frontline NHS staff to deliver quality improvement: findings from six case studies by D Wright et al. Published Online First: 27 August 2021. Accompanying editorial available here.

Rather than focus on managerial or leadership personnel and skills for quality improvement (QI) work, Wright and colleagues studied competencies for frontline professionals trying to improve care. The authors chose three different NHS hospital trusts with varying Care Quality Commission ratings and followed two multidisciplinary QI projects at each trust. Across these six projects, the team performed 122 interviews with 87 staff. Through their thematic analysis, the authors identified six ‘Socio-Organisational Functional and Facilitative Tasks’ (SOFFTs) with 20 subcomponent skills and tasks across technical, learning, and soft skill domains. These SOFFTs were not uniformly required for implementation across the six projects, but projects where the frontline staff either lacked the necessary SOFFTs or were somehow impeded in carrying them out were noted to struggle. The six SOFFT domains were as follows: 1) Adopting and promulgating the appropriate organisational environment; 2) Managing the QI rollercoaster; 3) Getting the problem right; 4) Getting the right message to the right people; 5) Enabling learning to occur; and 6) Contextualising experience. Ultimately, the researchers found that, while technical skills related to specific methodologies are important to QI work, knowing how to choose and adapt different methodologies when barriers present may be more important for frontline team members. Similarly, although understanding data is important for QI, complex data analysis capabilities may be less essential for frontline staff. Finally, many of the soft skills traditionally de-emphasized in QI training were found to be key for problem solving as projects progressed. Thus, teams may be best positioned for success in QI by ensuring a wide diversity of participation to maximize the range of skills and contributions, while maintaining flexibility and employing different SOFFTs at different times throughout the duration of a project.

2. Concordance with urgent referral guidelines in patients presenting with any of six ‘alarm’ features of possible cancer: a retrospective cohort study using linked primary care records by B Wiering et al. Published Online First: 04 October 2021. Accompanying editorial available here.

In certain cancers, shorter diagnostic intervals from the time of first symptomatic presentation may lead to earlier stage at diagnosis. To this end, the English Two Week Wait National Institute for Health and Care Excellence (NICE) guidelines created in 2005 and updated in 2015 highlight ‘red-flag’ features that should lead to specialist assessment within two weeks – breast lumps, haematuria, and iron-deficiency anaemia. These guidelines were designed such that the risk of cancer threshold for referral should have a positive predictive value of 3%. Reviewing more than 48,000 records of patients presenting in 2014-15 with one of six symptoms featured in the Two Week Wait guidelines, Wiering and colleagues sought to determine rates of non-referral within the recommended timeframe, whether patient factors were associated with referral/non-referral, and the proportion of patients diagnosed with cancer within one year of a guideline discordant non-referral. During the study period, 40% of patients received an urgent referral, although the frequency varied by symptom (from 17% for dysphagia to 68% for breast lumps) and by age, number of comorbidities, and, in the case of referrals for breast lumps, social deprivation. For comparison, 9.9% of those referred urgently were diagnosed with cancer within one year vs. 3.6% who did not receive an urgent referral. Thus, 35% of those diagnosed with cancer at one year had guideline discordant non-referral at the time of initial presentation. This was particularly pronounced in those presenting with anaemia but who were not referred, where 9.5% ultimately were diagnosed with cancer within one year. Many general practitioners may have had understandable rationales for not referring some patients, including allowing an opportunity for follow-up and later referral if the symptoms persisted in patients deemed to be at lower risk. In some cases, such non-referral may be valid; however, better adherence to these guidelines may increase cancer detection rates, even for patients with alarm features that already have high urgent referral rates.

1. Timeout procedure in paediatric surgery: effective tool or lip service? A randomised prospective observational study by OJ Muensterer et al. BMJ Quality & Safety 2021;30:622-.Accompanying editorial available here.

Although surgical timeout checklists are an established tool for reducing harm, it has been unclear how adherent surgical teams are to such checklists in practice and, thus, the overall performance of such processes. By purposefully introducing errors such as the wrong name, intervention, age, or site into timeout routines for 120 randomly selected surgical cases, Muensterer and colleagues were able to evaluate how well these errors were recognized and reported during the timeout process. All timeout personnel were blinded to the study aside from the surgeon investigators, with error detection defined as verbal interjection after the surgeon completed the timeout. If the error was not detected, the surgeon investigator would then correct the timeout before proceeding. Of the 120 introduced errors, only 65 (54%) were verbally noted and corrected by team members other than the surgeon investigators. Of the 65 recognized errors, wrong gender was most often identified (77%), while wrong age was least often (30%). Anaesthesiologists were most likely to identify to an error, accounting for two-thirds of those identified, while nurses reported one-quarter and surgical trainees and medical students reported far fewer. This study did not query team members after either the timeout or the case to evaluate whether errors were noted but not verbally reported, and so these response rates may not reflect rates of actual recognition of errors in the timeout process. Possibilities for not identifying errors include that junior members of surgical teams are less likely to notice errors or may be less likely to speak up when errors are noticed given the hierarchy of their teams. While some technological advances and human-centered design approaches may reduce some types of timeout errors in the future (such as by automatically verifying allergies and patient demographics/medications audially before a case), changing cultural hierarchy and improving safety systems will require long-term strategies and are arguably most important.

–Joel Boggan