Article Text
PDF
Abstract
Background We describe a novel process using positive deviance (PD) with the Canadian Association of Thoracic Surgeons members, to identify perioperative best practice to minimise anastomotic leak (AL) and length of stay (LOS) following oesophagectomy. To our knowledge, this is the first National combination of level 1 evidence with expert opinion (ie, PD seminar) aimed at reducing AL and LOS in oesophageal surgery. Our primary hypothesis is that a multicentre National PD seminar is feasible, and could lead to the generation of best practices recommendations aimed at reducing AL and LOS in patients with oesophageal cancer.
Methods Adverse events, LOS and AL incidence/severity following oesophagectomy were prospectively collected from seven Canadian thoracic institutions using Thoracic Morbidity and Mortality classification system (2017–2020). Anonymised display of centre’s data were presented, with identification of centres demonstrating PD. Surgeons from PD sites discussed principles of care, culminating in the consensus recommendations, anonymously rated by all (5-point Likert scale).
Results Data from 795 esophagectomies were included, with 25 surgeons participating. Two centres were identified as having the lowest AL rates 44/395 (11.1%) (vs five centres 71/400 (17.8%) (p<0.01)) and shortest LOS 8 days 45 (IQR: 6–14) (vs 10 days (IQR: 8–18) (p<0.001)). Recommendations included preoperative (prehabilitation, smoking cessation, chemotherapy for patients with dysphagia, minimise stents/feeding tubes), intraoperative (narrow gastric conduit, intrathoracic anastomosis, avoid routine jejunostomy, use small diameter closed-suction drains), postoperative day (POD) (early (POD 2–3) enteral feeding initiation, avoid routine barium swallow studies, early removal of tubes/drains (POD 2–3)). All ranked above 80% (4/5) in agreement to implement recommendations into their practice.
Conclusion We report the feasibility of a National multicentre PD seminar with the generation of best practice recommendations aimed at reducing AL and LOS following oesophagectomy. Further research is required to demonstrate whether National PD seminars can be an effective quality improvement tool.
- Quality improvement
- Surgery
- Patient-centred care
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Anastomotic leak (AL) following oesophagectomy has profound negative implications for patients with oesophageal cancer, and despite major quality improvement (QI) efforts, AL continues to plague both surgeons and patient alike.
WHAT THIS STUDY ADDS
We describe a novel QI approach to minimise AL and length of stay following oesophagectomy by using a positive deviance (PD) seminar. Using PD highlights innovative strategies to prevent and manage AL, by distinguishing exemplary practice and generating actionable, patient centred consensus recommendations.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
PD enables an innovative and supportive environment for surgeons, by identifying best practices while not demeaning less successful surgical groups. The PD approach can champion a culture of professionalism and respect for patient centred care, and at its most granular level, brings forth limitations to current practices, allowing for well-designed clinical trials.
Introduction
There is an urgent need for a learning healthcare system that continuously and systematically uses data to inform improvements in care. Ample evidence demonstrates how surgical adverse events (AEs, ie, deviations from expected uncomplicated recovery) are a critical determinant of surgical quality, as AEs contribute to increased length of stay (LOS), costs of care and impaired patient experience.1 2 Traditional quality improvement (QI) interventions such as Morbidity and mortality rounds are limited by a myopic focus on errors and harm and have failed to reduce preventable medical errors.3 Surgery is individualistic and surgeons often purport that technical expertise is the essential requirement for optimum patient outcomes. However, surgical excellence and improved outcomes have been overwhelmingly shown to be superior when an integrated team is associated with large numbers of cases and experience.4 In particular when there is a personal and sensitive accountability, and given the critical importance of AEs to both patients and providers, there is a vital need to identify data-driven methods to reduce their occurrence.
In an attempt to overcome some of these difficulties, the concept of positive deviance (PD) has emerged in healthcare literature. This is a process designed to distinguish exemplary practice performed by individuals, teams or organisations, whose actions lead to improved safety and quality of given care.5 The PD approach has its roots within the public health domain, where it was observed that within communities there are individuals and teams whose uncommon practices lead to improved health outcomes.6 PD has since been applied to other settings, including infectious disease,7 orthopaedic surgery,8 cardiology,9 as well as at the health system level.10 The use of PD requires several important group characteristics such that recommendations gleaned are both clinically relevant and individually acceptable. This includes (1) a set of standard performance measures which are both accessible and endorsed by the participants, (2) identification of variation in participant behaviour to enable identification of ‘PD’, (3) generation of best practice recommendations to be tested in other participants practices and (4) participants who are both willing and agreeable to adopt recommendations.10 This positive feedback serves to strengthen the community of care providers, ultimately improving patient outcomes. As such, PD is a promising technique enhancing sensitive surgical AE data.
A multicentre application of PD to the reduction of surgical complications has been pioneered by the Canadian Association of Thoracic Surgeons (CATS).11 The process aims to provide best supporting level-1 evidence, married to consensus expert opinion, leading to the generation of practice recommendations for clinically relevant topics. This goal is in part achieved by way of systematic recording of both the incidence and severity of AEs within all centres, using a standardised taxonomy (ie, the Thoracic Morbidity & Mortality (TM&M) based on the Clavien-Dindo schema).12 This enables tracking of site- and operative procedure-specific complications. TM&M data are then used to inform individual PD seminars, where one AE and one procedure are chosen by a group of participating surgeons for analysis and discussion. PD seminars combine best evidence with best experience. A literature review is initially presented to the group by a staff member, providing a summary of current studies and guidelines. Individual-data are then presented anonymously, and the group identifies the best performer(s) (demonstrating the practice displaying PD). The chosen surgeon(s) or centre(s) describe their current practice and why they believe that their methods lead to superior patient outcomes. Discussion among all surgeons leads to the generation of consensus recommendations that agree to implement into practice. Since 2014, CATS has facilitated 14 local and 4 multicentre PD seminars on a variety of topics, resulting in the generation of 144 consensus recommendations.
Here, we describe a National multicentre PD seminar focusing on oesophagectomy complications and LOS. Our primary focus was on anastomotic leak (AL), representing one of the most dreaded complications following oesophagectomy.13 14 Given the devastating consequences of an AL, our primary hypothesis was that a multicentre National PD seminar is feasible, and could lead to the generation of best practices recommendations aimed at reducing AL and LOS in patients with oesophageal cancer.
Methods
We conducted a virtual multicentre PD seminar on AL and LOS following oesophagectomy. Contemporary, prospectively collected operative and AE data of seven Canadian thoracic surgery centres were compared anonymously. Aggregate data prepared for the seminar included patient age, LOS (with and without outliers removed; outliers defined by being greater than 2 SDs from the mean), AL incidence and severity (all AEs were all graded based on the TM&M system based on Clavien-Dindo classification) and re-admission rates.
A flow chart describing the PD process is shown in figure 1. The PD seminar commenced with a didactic session with a systematic literature review of peer-reviewed evidence focused on best-evidence around oesophageal leaks and their management (level-1 evidence). This was followed by anonymous display of institutional self-assessment data (figure 2). Two sites (identified by group consensus) displayed practice demonstrating PD in terms of lowest rates of AL and shortest LOS. These two sites were unmasked, and surgeons from each site discussed the particular institutional practices that were felt to contribute most improved patient outcome. A discussion with all the participants ensued, allowing for collegial in-depth assessment and hypothesis generation regarding the identified perioperative strategies. Ultimately, consensus recommendations were created and grouped according to their preoperative, intraoperative and postoperative timing. At the conclusion of the PD seminar, to ensure a consensus agreement on all recommendations, surgeons were asked to anonymously indicate their level of agreement (1—strongly disagree; 5—strongly agree) with each recommendation using the Mentimeter platform.
Positive deviance seminar flow chart
Anonymised multicentre data display with top two sites demonstrating positive deviance. LOS, length of stay.
A postseminar survey was then distributed to the participants, allowing formative evaluation of the PD seminar process, value and impact. Each site had local ethics and privacy approval for their local patient-level data collection. CATS has ethics approval through Clinical Trials Ontario for the aggregate average and rate data comparisons.
Results
Data from 795 oesophagectomies, performed by 31 thoracic surgeons at 7 centres between 2017 and 2020, were included in the data presented. Twenty-five surgeons from 15 centres, representing seven Canadian provinces, participated in the 2-hour PD seminar discussion. The number of oesophagectomies from each centre is presented in table 1.
Number of oesophagectomies performed by academic Canadian centre (2017–2020)
Two centres (McGill, Toronto UHN) were identified as demonstrating PD, with rates of major or minor AL of 11.1% and median LOS of 8 days, compared with 17.8% and 9.8 days, respectively, for the other sites (figure 2). Discussion of the data led to the generation of 13 practice recommendation divided into the preoperative, intraoperative and postoperative settings (figure 3). Preoperative recommendations included: early patient assessment by a thoracic surgeon, a dedicated focus on prehabilitation (including daily exercise/walking, smoking cessation), expedited chemotherapy for patients with significant dysphagia and minimising interventions (stents and surgically placed jejunostomy feeding tubes) unless complete dysphagia or severe cachexia. Intraoperative recommendations included: constructing a narrow, non-redundant conduit, creating an intrathoracic anastomosis when appropriate based on patient/tumour factors, placing the anastomosis as low on the greater curvature (ie, as close to the gastroepiploic arcade as possible), avoiding routine surgically placed feeding jejunostomy and using small diameter closed suction drains. Postoperative recommendations included: early removal of nasogastric tube (postoperative day (POD) 2), early initiation of enteral feeding (POD 2–3) without routine barium swallow studies, and early removal of chest tubes/drains. Surgeons indicated a strong level of agreement with the preoperative (4.75/5), intraoperative (4.5/5) and postoperative (4.4/5) recommendations, verifying the consensus on all 13 recommendations to be put into practice by all participating surgeons (figure 4).
Top ranked consensus recommendations. POD, postoperative day.
Level of agreement of proposed recommendations among participating surgeons. POD, postoperative day.
A post-PD seminar survey was distributed to all surgeons who attended. 10 of the 25 surgeons (40%) responded. The majority (90%) agreed that the PD seminar offered important and relevant self-reflection and self-assessment practices, which could offer direct improvements in patient care. The data presented were thought to be valuable and to identify practice changing perioperative management strategies (90%). Importantly, most surgeons (80%) indicated they felt motivated and intended to modify their clinical practice considering the recommendations and would be willing to participate in further seminars (90%).
We investigated the impact on patient outcomes at 12 months by analysing the effect of attending the PD seminar on AL rate and LOS (table 2).
Attending surgeons versus non-attendee surgeons to the PD seminar and impact on AL rate and LOS
Here, the outcomes of interest (AL, LOS) of surgeons who attended versus non-attendees was compared. In the surgeons who attended and whose surgeon-specific data were available for analysis in our database (n=9), an average AL rate of 18.2% (pre-PD seminar) was reduced to 12.8% (post-PD seminar), although this was not statistically significant (p=0.2). In the group of non-attendee surgeons with available data (n=14), an average AL rate of 14.3% (pre-PD seminar) remained similar at 14.7% (post-PD seminar) (p=0.9). In both groups (surgeons who attended and non-attending surgeons), the LOS was reduced, although this too was non-statistically significant. Specifically, average LOS following oesophagectomy for patients who attended the PD seminar was reduced from 10 days (pre-PD seminar) to 8 days (postseminar) (p=0.09), while patients under the care of surgeons who did not attend remained on average 11 days (pre-PD seminar), which decreased to 9 days in the postseminar period (p=0.08).
Discussion
Learning from sensitive provider-specific data in a manner that standardizes and improves care remains a vital challenge. For the first time, we report the feasibility of a National PD seminar to reduce AL and LOS following oesophagectomy. Two sites were identified as demonstrating best performance, not surprisingly as they were also the sites with greatest volume (see table 1). We identified consensus recommendations through collaborative discussion with multiple thoracic surgeons from different centres across Canada. The process enabled the generation of 13 consensus driven, clinically relevant perioperative practice recommendations for patients undergoing oesophagectomy, and the participating surgeons involved collectively agreed to implement these consensus recommendations in their practice. Here, the desired outcome (reduction in AEs) is derived from practices which are based in a community of surgeons whose principal interest is paramount to their own success (optimal patient outcome). Identifying either an individual surgeon, team-practices, or patient factors that produce better outcomes compared with surgical peers is a powerful way of engendering change. This represents the essence of PD. Furthermore, when we performed an analysis of the outcomes of interest (AL, LOS) at 12 months, comparing surgeons who attended the PD seminar to those who did not, we observed an encouraging decrease in the AL rate in participants, although this was not our primary outcome of interest. Further research is required to demonstrate whether National PD seminars can be an effective tool, capable of QI. Here, an interrupted time series analysis and mixed-methods evaluation will provide a more rigorous analysis of actual patient-level impact.
Standardisation of care from a national perspective remains an objective surgeons, patients and healthcare administrators aspire to. Previous research has identified unacceptably high practice variation, resulting in inconsistent patient outcomes.15–19 National coordination and standardisation of data collection is the first step required to enable feedback to surgeons, followed by collegial discussion such as the PD approach. As practice standardisation has been associated with decreased postsurgical AEs,20 the PD approach has the potential to not only reduce practice variation, but also improve patient outcomes across all participating centres. The PD seminar gathers data on site-specific practices, where best performers are identified and all centres are aware of their own data and that of the best performers. Assessing accurate causal information of events that have multiple possible causations has proven elusive (ie, leak following oesophagectomy). Here, causations range from underlying patient factors, surgical team factors and cultural factors such as population nutritional state, fitness and gut micro-flora. Teasing out individual positive and negative contributors and their possible interaction has proven difficult, this resulting from not only the complexity of patient-disease-surgeon interaction, but also due of the sensitivity of patient complications which involves surgical pride, patient blame and economics. Consequently, the PD approach to surgical complications has proven particularly valuable in the context of surgery, overcoming some of these drawbacks. Ranking can be done reliably based on valid performance measures that vary both between individual surgeons and across centres. Furthermore, the PD process is aligned with and complementary to existing hospital QI objectives.
Surgeons have been exceptionally collegial and open to the PD approach, previously reporting their favourable impressions of both the team-building and positivistic nature of the PD process within their centre.21 Surgeons participating voluntarily in the PD approach record the data, in the belief that it will improve care. Indeed, this was reflected in the post-PD seminar questionnaires, where the majority (90%) of surgeons reported the PD process allowed for self-reflection and assessment of their own clinical practices, and where agreed on recommendations could improve the quality of their patients’ care. Furthermore, analysis from other PD seminars has shown that implementation of this process at the local level in a single centre resulted in a 34% reduction in atrial fibrillation, a 38% reduction in air leak, and a 25% reduction in AL after all major thoracic operations.22
Though PD has most often been used in non-surgical specialties, this approach is particularly valuable in the context of surgery in that ranking can be done reliably based on valid performance measures that vary both between individual surgeons and across centres. PD could be applied to any interventional discipline that performs procedures with potential costly harmful AEs. Using the PD approach as a repeatable process, the seminars can be performed in the same manner for various combinations of AEs and surgical procedures. Furthermore, the PD process is aligned with and complementary to existing hospital QI objectives. Rare and unusual surgical techniques may permit participation in large QI programmes that would otherwise not be feasible. The PD approach offers a potential innovative approach to both standardise and improve patient care and surgical quality.
Limitations
An analysis of the impact of the PD seminar recommendations, and the effectiveness of PD seminar as a QI tool, is beyond the scope of this paper. Indeed, the recommendations are not all-encompassing, and it remains unclear to what extent individual surgeons are willing to adopt the practice recommendations. As an example, despite the widespread adoption of minimally invasive oesophageal surgery, data from a high-volume centre used in our current study routinely perform oesophagectomies as a hybrid procedure (combining laparoscopy and thoracotomy) or, at times, as an open approach (laparotomy/thoracotomy). This was not considered a best practice recommendation given the current existing evidence supporting minimally invasive surgical approaches. In addition, adherence to the consensus recommendations is the subject of further study; thus, we cannot report whether PD seminars are an effective tool capable of QI.
Despite limitations in completeness of data from participating surgeons, all major centres that routinely perform oesophagectomies contributed to the consensus recommendations (7/7 centres). As such, these sites are able to provide insights into the discussion, which is the ultimate goal of the PD seminar (ie, a collaborate discussion among surgeons).
We normally perform confidential surgeon self-assessment as part of the PD seminar (ie, provide each surgeon with their own data); however, we did not have individual surgeon level data available at the time of the seminar due to limited local research support during the COVID-19 pandemic. Our future intent is to make surgeon level data confidentially available to each participating surgeon during PD seminars, following identification of sites demonstrating PD.
Finally, this study is limited by the low response rate from the surgeons (40%) and the possibility that non-respondents would have provided different answers. Given the format of the postseminar survey (email), this may have contributed to the observed low response rate. It is possible that surgeons who completed the post-PD seminar survey were in favour of the recommendations, while those that did not had differing views of the recommendations. Although the individual response rate was low, good institutional coverage was observed. Additionally, in meeting, live-polling response rate was excellent as this was an in-person opinion poll. Overall, our analysis exposes limitations to current practices, and at its most granular level, allows for hypothesis generation, potentially helping future research into specific areas.
Conclusion
We report the feasibility of a National multicentre PD seminar with the generation of best practice recommendations aimed at reducing AL and LOS in patients with oesophageal cancer. The goal of the PD approach is to improve healthcare delivery by identifying best practices while not demeaning less successful surgical groups. This demands courage of all participants, requiring openness, transparency, honesty and humility. The PD approach can champion a culture of professionalism and respect for patient-centred care. Further research is required to establish whether PD seminars can represent an effective tool capable of QI in patients with oesophageal cancer.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
References
Footnotes
Contributors Conception and design: DJ, MG, CA and AS. Administrative support: DJ, MG and CA. Provision of study materials or patients; manuscript writing; final approval of manuscript: All authors. Collection and assembly of data: DJ, MG, CA, SN and AS. Data analysis and interpretation: DJ, MG, CA, GD, LF and AS. AS, DJ and LF were responsible for the overall content and acted as guarantors.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.