In this post, we’re offering the second half of our summaries and comments on articles from BMJ Quality & Safety’s Top 10 of 2020. To check out the 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.

5. Identifying and encouraging high-quality healthcare: an analysis of the content and aims of patient letters of compliment by A Gillespie & TW Reader. Published Online First: 08 July 2020. Accompanying blog post available here.

Patients and families may be able to identify important safety issues and opportunities for improvement within healthcare systems. Rather than focus on shared patient concerns or complaints, Gillespie and Reader instead reviewed a sample of more than 1,200 letters of compliment sent to 54 English hospitals (~2% of the total over the study period) to learn more about what patients or their families value enough to acknowledge by sending an unsolicited message. They found that 80% of compliment letters included a theme of gratitude, either for an individual, a care team or a unit. In 44% of cases, these sentiments were sent or copied to a unit, departmental, or senior manager, presumably to ‘reward’ the person(s) named. In almost half of compliment letters, specific behaviors were commended, again presumably to encourage such behaviors in the future. Often, these behaviors involved those named going ‘above and beyond’ what was perceived as their normal duties. From a positive perspective, managers sharing these compliments with front-line staff may build better professional relationships, improve work culture, and reinforce the underlying motivations many front-line staff have in choosing a clinical career. On the other hand, managerial tracking of such compliments may lead to front-line staff perceptions of an obligation to solicit such compliments, or the compliment itself may actually indicate a staff member going ‘above and beyond’ to compensate for inadequate resources being provided. Ideally, managers and health systems would use trends and themes in such compliment letters to identify opportunities for redirecting or adding resources to improve the overall patient experience.

4. Assessment of a quality improvement intervention to decrease opioid prescribing in a regional health system by CS Brown (@CraigBrownMD) et al. Published Online First: 16 September 2020. Accompanying editorial available here.

Prescriptions for opioid medications after surgical procedures have been implicated as a significant contributor to the ongoing opioid public health crisis in the US, prompting the adoption of prescribing guidelines and even legislation limiting prescription sizes after many procedures. In this study, Brown et al. examined postoperative prescription size after nine different surgical procedures following issue of three sets of guidelines for these procedures created by the Michigan Opioid Prescribing Engagement Network partnership, as well as patient reports of opioid consumption, satisfaction and postoperative pain. In addition to publishing guidelines specifying the maximum number of oxycodone 5 milligram tablets that should be provided on a discharge prescription, each of the 69 hospitals in the analysis had access to an institution-specific dashboard; thousands of clinicians and laypersons also received communication regarding the practice and effects of overprescribing via platforms such as Twitter and newsletters across the entire state. Although prescription size was already declining for these procedures prior to release of the first guideline, this rate of decrease accelerated for several months thereafter before flattening after release of the second and third guidelines. Overall, the mean opioid prescription size decreased 50% and significant declines in prescription size were noted across all nine procedure types. Similarly, in the third of patients who returned surveys, patient-report opioid consumption and the percentage of patients receiving a refill of opioid medications also decreased by 50%. Importantly, the balancing measures of patient satisfaction and patient-reported postoperative pain remained unchanged throughout the study period. While this multistakeholder partnership group may be unique to the state of Michigan, this work illustrates both that variation in opioid prescription patterns after surgeries can be reduced, hopefully ultimately reducing downstream risks to patients, and that surgeons can affect postoperative opioid use by altering prescription practices. Furthermore, these changes can occur without affecting patients’ perceptions of postoperative pain care.

3. Association of open communication and the emotional and behavioural impact of medical error on patients and families: state-wide cross-sectional survey by JC Prentice et al. BMJ Quality & Safety 2020;29:883-894. Accompanying editorial available here.

Up to one in four American adults have been involved in or know someone who has suffered a medical error. Using a sample from a wider multiyear survey in the state of Massachusetts, Prentice et al. report an important population-based assessment of the impact of medical errors on patients. The authors surveyed 253 patients who had perceived personal, family or other contact involvement in a medical error. They evaluated the physical and emotional impacts of these errors (including sadness, anger, depression, and feelings of abandonment/betrayal), as well as subsequent avoidance of certain providers and loss of trust in healthcare. Furthermore, their survey addressed whether there was open communication about the error from anyone broadly involved in the care event, and analysed the effects of the error based on this communication. Initially after the error, more than 80% of respondents reported at least one significant emotional impact, and 43% reported substantial physical health effects or a resultant death. Many studied impacts had the potential to linger for several years – for example, 42% of those whose error occurred in the previous year had ongoing physical effects, as did 27% of those whose error was 3-6 years earlier. Additionally, more than half the group perceiving an error 3-6 years ago still experienced emotional effects, with all but the feeling of anger persisting at relatively stable levels no matter the time from the event. Similarly, patterns of healthcare avoidance and loss of trust in healthcare persisted over several years. Overall, 34% of respondents received no communication from providers or the healthcare system about the error, while 29% had the error acknowledged and only 19% received an apology. As increasing numbers of the six studied elements of open communication (acknowledgment, speaking openly and truthfully, speaking in easily understood language, conveying information about health consequences of the error, welcoming questions, and providing opportunities to express feelings) were experienced by the respondents, the numbers reporting sadness, depression, and abandonment/betrayal or avoidance decreased. Apologies were also independently associated with reductions of some persistent effects. These results suggest that provider training and system investments should prioritize formally structured communication after an error that incorporates several such elements of open communication, in addition to a clear apology. Unfortunately, even inclusion of all these elements and apologies do not seem to protect patients from ongoing perceptions related to lost faith in the efficacy and safety of care, highlighting the need for additional efforts to further improve reconciliation and reduce long-lasting harm.

2. Associations between double-checking and medication administration errors: a direct observational study of paediatric inpatients by JI Westbrook (@JWestbrook91) et al. Published Online First: 07 August 2020.

Double checking medication administration is a frequently followed and time-consuming nursing practice and is particularly emphasized in paediatric hospitals. These checks are intended to be performed independently; however, there are opportunities to share information (termed ‘priming’) between individuals during this process, which may reduce the theoretical protective effect of the practice. After performing a systematic review of double checking published in 2019, finding few studies of high quality and mixed results in reducing medication administration errors (MAEs), Westbrook et al. performed the first study to explore the effectiveness of double-checking procedures for preventing MAEs in a paediatric setting. In all, the study team observed 5,140 med administrations by 298 nurses over a 22-week period, almost 70% of which required double checking, per hospital policy. Extrapolating from the noted 7.9 min to double check each IV dose and the 5.5 min for non-IV doses, the team estimated the practice to require 133 nurse hours/day and cost $A2.7 million/year at the facility. Given this time and expense, it is unfortunate that the authors found no association between mandated double checking and the odds of an occurrence of MAEs (OR 0.89, 95% CI 0.65-1.21) or of the MAE being of higher severity (OR 0.86, 95% CI 0.65-1.15). The one instance in which double checking was noted to reduce MAEs was when applied to doses for which double checking was optional, with lower overall odds of an MAE (OR 0.71, 95% 0.54-0.95) or the MAE being of higher severity (OR 0.75, 95% 0.57-0.99). For 92.5% of doses for which a double check was mandated, the check was completed but was ‘primed’, perhaps because the additional time required for a truly independent verification was not possible. The authors conclude that an overall review of these checking processes should be pursued more widely given, particularly given changes in medication administration technology, packaging and practices over the past two decades.

1. Seven features of safety in maternity units: a framework based on multisite ethnography and stakeholder consultation by EG Liberati (@ElisagLib) et al. Published Online First: 25 September 2020. Accompanying editorial available here.

Improving maternity care is an important goal worldwide, having implications for everything from local hospital system cost structures and practitioner liability to universal human rights and the international community’s women’s and children’s health agenda. Building on a previous single example of a high-performing maternity unit, Liberati et al. expanded their ethnography to five additional maternity units that had participated in a training programme by the original high-performing unit and refined their previous framework to share ‘what good looks like’ in the maternity ward context. The resulting plain-language ‘For Us’ (For Unit safety) framework lists seven characteristics, believed to act synergistically, to create a robust safety maternity environment. It is expected most maternity units would exhibit some combination of these factors that could be analysed locally to highlight areas of current success and opportunity. The seven For Us features are as follows:

• Commitment to safety and improvement at all levels, with everyone involved:  units exhibiting this authentic commitment had trusted risk management processes, systems, and roles and all staff constantly were engaged in seeking to improve care.
• Technical competence, supported by formal training and informal learning:  in areas with the highest levels of staff confidence and readiness to respond to safety events, structured training was mandatory and supported with protected time for attendance to develop and maintain proficiency, role modelling and peer mentoring were emphasized across staff relationships and common areas were available for debriefing after complex clinical situations to allow informal learning.
• Teamwork, cooperation, and positive working relationships:  for the most effective unit teams, an emphasis on shared goals and flexible hierarchies helped manage any potential conflicts and fostered local support to help others and maintain well-being and morale.
• Constant reinforcing of safe, ethical, and respectful behaviors:  in respectful units, the expected behaviors were explicitly shared and modeled by team members, creating clear expectations for all staff roles and helping integrate newcomers into the local work culture.
• Multiple problem-sensing systems used as a basis of action:  by creating psychologically safe environments, units where frontline staff and families could share ideas and concerns, and where management also incorporated other clinically-relevant data, were better able to continuously learn and improve.
• Systems and processes designed for safety and regularly reviewed and optimised:  areas where the physical environment, equipment, and information technology have been tested and standardised to reduce frontline staff cognitive load allowed increased risk anticipation and problem sensing.
• Effective coordination and ability to mobilise quickly:  strategies such as handovers, huddles, rounds and structured emergency protocols reinforced by simulation allow updated information sharing to increase coordination and situational awareness throughout successful units and, therefore, emergency responses to proceed quickly.

–Joel Boggan