- Email: [email protected]
Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images
Journal Pre-proof Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images Kathryn B. Sobba, MD, Adolfo Z. Fernandez, MD, FACS, Stephen S. McNatt, MD, FACS, Myron S. Powell, MD, FACS, Andrew M. Nunn, MD, FACS, Amy N. Hildreth, MD, FACS, Barbara K. Yoza, PhD, Jessica L. Gross, MD, FACS, Preston R. Miller, III, MD, FACS, Carl J. Westcott, MD, FACS PII:
S1072-7515(19)32166-0
DOI:
https://doi.org/10.1016/j.jamcollsurg.2019.09.022
Reference:
ACS 9669
To appear in:
Journal of the American College of Surgeons
Received Date: 30 July 2019 Revised Date:
30 September 2019
Accepted Date: 30 September 2019
Please cite this article as: Sobba KB, Fernandez AZ, McNatt SS, Powell MS, Nunn AM, Hildreth AN, Yoza BK, Gross JL, Miller III PR, Westcott CJ, Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images, Journal of the American College of Surgeons (2019), doi: https://doi.org/10.1016/j.jamcollsurg.2019.09.022. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc. on behalf of the American College of Surgeons.
1
Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images Kathryn B Sobba, MDa, Adolfo Z Fernandez, MD, FACSa, Stephen S McNatt, MD, FACSa, Myron S Powell, MD, FACSa, Andrew M Nunn, MD, FACSa, Amy N Hildreth, MD, FACSa, Barbara K Yoza, PhDa, Jessica L Gross, MD, FACSa, Preston R Miller III, MD, FACSa, Carl J Westcott, MD, FACSa,c Author affiliations a Department of Surgery, Wake Forest School of Medicine, Winston-Salem, NC b Department of Anesthesiology, Wake Forest School of Medicine, Winston-Salem, NC c Department of Surgery, W.G. Bill Hefner VA Medical Center, Salisbury, NC Corresponding author Carl J Westcott, MD, FACS Dept. of Surgery, Wake Forest School of Medicine Medical Center Blvd. Winston-Salem, NC, 27157 E-mail: [email protected] Phone: 1-336-716-2409 Disclosure Information: Nothing to disclose. Presented in part at the American College of Surgeons 104th Annual Clinical Congress, Boston, MA, October 2018. Brief Title Image Sharing for Surgical Quality Assurance
2
BACKGROUND: The technique for attaining photographic evidence of the critical view of safety (CVS) in laparoscopic cholecystectomy (LC) has previously been defined, however the consistency, accuracy and feasibility CVS in practice is unknown. The aim of this study was to use an already established image sharing and grading system to determine the feasibility of timely feedback after sharing intraoperative images of the CVS and to evaluate if and how cholecystitis impacts the ability to attain a CVS. STUDY DESIGN: 193 laparoscopic cholecystectomies performed by 14 surgeons between August 2017 and January 2019 were studied. Anterior and posterior intraoperative CVS images were shared using a standard multimedia messaging system (MMS). Images were graded remotely by members of the group using an established scoring system and their time to response and scores recorded. Response data was analyzed for the ability to attain timely and consistent CVS scores. RESULTS: 74 urgent laparoscopic cholecystectomies for acute cholecystitis and 119 non urgent cholecystectomies were performed during the study period. Scoring of shared images occurred in less than 5 minutes and peer review (mean 3 responses) showed agreement that was not significantly different. In patients with acute cholecystitis, a small but significant difference was observed between anterior and posterior image scoring agreement. CONCLUSION: An established image sharing and grading system for CVS can be used for realtime intraoperative feedback without increasing operative time or compromising private health information. The CVS is almost always attainable, however decreases in CVS quality and grading agreement are observed in patients with acute cholecystitis. KEYWORDS: critical view of safety, intraoperative image capture, safe cholecystectomy
3
ABBREVIATIONS CVS
critical view of safety
MMS
multimedia messaging system
PHI
private health information
LC
laparoscopic cholecystectomy
SAGES
Society of American Gastrointestinal and Endoscopic Surgeons
QA
quality assurance
4
Introduction Over 750,000 cholecystectomies are performed each year in the United States1. Despite advancements in laparoscopic techniques, the rate of bile duct injuries and leaks remains higher in laparoscopic cholecystectomies (LC) than in open1. Although reported bile duct injuries remain relatively infrequent, occurring in approximately 3 per 1,000 LCs performed, the true incidence of bile duct injury is unknown and can lead to devastating complications for patients2-5. In addition, acute cholecystitis alters normal anatomy makes adequate dissection and correct identification of cystic structures more difficult, which likely contributes to the increased frequency of bile duct injuries that is seen in these patients during LC.6 The Critical View of Safety (CVS) is a technique widely believed to decrease the risk of bile duct injuries in LC7. The technique for attaining photographic evidence of a CVS is well defined and includes the following three criteria: 1. the hepatocystic triangle is cleared of fat and fibrous tissue, 2. the lower one third of the gallbladder is separated from the liver to expose the cystic plate, 3. two and only two structures are seen entering the gallbladder8. To adopt a universal culture of safety and minimize the risk for bile duct injuries during LC, the Society of American Gastrointestinal and Endoscopic Surgeons (SAGES) safe cholecystectomy task force recommends using the CVS technique and liberal use of bile duct imaging. Attaining the CVS is recommended in all cases and they advocate an intraoperative pre-clipping timeout to confirm that the CVS has been achieved. If available, consulting a peer surgeon is advised if the anatomy is unclear or under conditions deemed “difficult” by the surgeon. In addition, a Multi-Society Bile Duct Injury Consensus Conference (2018) established consensus recommendations for the CVS itself and discussed the concept of an intraoperative time out and photo-documentation of the dissection9. However, surgeons performing LCs do not always have
5
the benefit of an available peer surgeon, time outs are not established and accepted, and the idea of photo-documentation of the dissection is understudied and generally resisted. The aim of this study was to use instant messaging images of the CVS to establish a method for photo-documentation and intraoperative time out. Image capture was by mobile phone camera and a mainstream multimedia messaging system (MMS) was used to share CVS images. A standard grading system was employed8. The study sought to establish proof of concept and to determine if pericystic inflammation affects CVS image quality and/or grading agreement. Methods Study population. LCs performed by the study surgeons at Wake Forest Baptist Medical Center between August 2017 and January 2019 were included, regardless of patient age, comorbidities, or the nature of the procedure (planned or emergent). The study was reviewed and approved by the Wake Forest School of Medicine institutional review board. The MMS group (study group) consisted of 14 attending surgeons. All were invited and able to submit and review images at their own volition. Participants were able to silence the MMS group chat on their device as they pleased. This quality assurance (QA) project was reviewed and carried out with department and institutional approval (IRB #00057007). Image acquisition, sharing and grading. During each LC, an intraoperative timeout was performed by the operating surgeon to confirm adequate CVS dissection, capture anterior and posterior CVS images on their mobile phones, and share images with the study group in real time using the standard MMS application. Care was taken to have only the operative screen in the picture devoid of any other information. The operative surgeon directed the image acquisition. Study surgeons remotely graded the adequacy of the CVS as demonstrated by the shared imag-
6
es8. The response times and scores for anterior and posterior views were recorded. No PHI was shared, and the grading surgeons were instructed to delete the photos from their devices once grading was completed. The patients identifying information was entered in a separate secure HIPPA compliant data base and the de-identified images later matched to the medical record. The grading process was discussed and practiced prior to starting the QA project. Study surgeons reviewed previously published methods to capture and grade CVS images8. Two meetings of the group and a presentation to the entire department were held to review stock images and practice grading. The grading scheme is anterior image first (3 criteria) and then the posterior image (3 criteria). The grades are listed from 0 to 2 for each aspect of the CVS and are defined in Table 1 (Table 1). The sequence of the CVS aspect grades is: two and only two structures, followed by the cystic plate for the lower 1/3 of the liver bed, and the hepatocystic triangle cleared of all fat and fibrous tissue. For example: 222/222 is a perfect score. 122/222 would indicate that the evaluator could not clearly determine if there were two and only two structures on the anterior view. In this manner the specific deficiencies are described to the submitting surgeon. Each evaluation is viewable by the study group and the surgeon requesting the evaluation. Data analysis. The primary purpose for data analysis was to determine feasibility. Average response time and % agreement between CVS scores by responding surgeons was determined. Pathology reports were reviewed to document when acute cholecystitis was present. Mean time to first response, comparison of response times during normal business hours and nights/weekends, differences in CVS scores in anterior and posterior images, and an effect of pericystic inflammation on CVS quality and grading was assessed. Final pathology reports were reviewed for all study patients and the presence or absence of acute cholecystitis was recorded; this data is reported and indicated in the figure legends. Data was presented and analyzed by one-
7
way ANOVA with Tukey’s post-test between groups (alpha 0.05, 95%Cl) or unpaired student ttest between two groups using GraphPad Prism 7.04 software (San Diego, CA). Results After a 30 day follow up, there were no bile leaks, bile duct injuries or bleeding, or return to the operating room for these patients. An average of 3 evaluations were received for all LCs performed during the study period (Fig. 1A); no difference in the number of responses was observed for surgeries performed during normal business hours compared to LCs that occurred on nights and/or weekends (Fig. 1B). The average time for a response from another surgeon was less than 5 min. (3.7+0.5 min.) and a response from a second surgeon took significantly longer (10.7+1.6 min.) (Fig.2). There was no difference in first or second feedback times during the day compared to nights/weekends (Fig.2). To determine if pericystic inflammation affects CVS quality and grading agreement, data for acute cholecystitis patients (74 patients of the total 193) were examined in a separate analysis (Table 2) . Comparable to all patients, an average of 3 evaluations was received for shared CVS images of acute cholecystitis LCs (Fig. 3) with a response from another surgeon in less than 5 min. (3.5+0.7 min.) (Fig. 4); similar to all patients, a response from a second surgeon took significantly longer (15.4+3.7 min.). There was no difference in first or second feedback times during the day compared to nights/weekends (Fig. 4). The responses were analyzed for % agreement between scores for each patient (n=173). An image was rated as demonstrating CVS satisfactorily when the mean score of >2 raters was greater than 5 on at least one image8. There was 85%+2.1% and 82%+2.2% agreement in scoring anterior and posterior images of the CVS, respectively (Fig. 5A). Further analysis of the acute cholecystitis patient image grading showed that while there was agreement between surgical
8
peers (anterior CVS, 81%+3.7%; posterior CVS 73%+4.1%), a significant difference was observed in grading agreement of the posterior CVS (Fig. 5B). These results show that adequate CVS is almost always attainable in acute cholecystitis, however pericystic inflammation decreases CVS quality and grading agreement, and the posterior view suffers and varies the most Discussion Adequate CVS dissection during LC is recommended to avoid bile duct injury.1,3,7 Studies of the quality of CVS in LC have generally questioned the understanding of the CVS among surgeons. In addition, the consistency and accuracy of the CVS in practice has been documented to be poor10-13. These studies point to the need for a mechanism to standardize and teach the CVS. Our data set is small with the intent to test feasibility and proof of concept. A multiinstitutional effort would be needed for a prospective study to test the impact on bile duct injuries and other surgery related complications. However, live QA instituted over several institutions or through a medical society project could accumulate the numbers typically discussed as needed for a prospective analysis of cholecystectomy with respect to bile duct injuries. Modifications to the approach used in this study should be considered for a multi-institutional study and are outlined later in the discussion. Our grades showed a high level of concordance. The routine practice of attaining a critical view, sharing it and having it evaluated by a group of experts, or your partners and peers should support making the CVS dissection uniform within a group. If a small group of surgeons employs this approach, they would need to agree on and then, through the sharing and grading process, migrate towards a uniform understanding of the CVS. If adopted by an institution, health care system or a specialty society, it would serve as a way to standardize dissection for
9
much larger groups. Concordance and variation in grading as well as changes with respect to grading are interesting future projects in this line of investigation. The concordance could also be explained by grader bias. An unmodified MMS system is not ideal for this application as any one reviewer’s grades and identity are visible to all other reviewers. A system where all reviewers are blinded to other reviewers is needed prior to data collection and analysis that could address this question. Our data shows using the live quality assurance (QA) system does not add a significant amount of time to a surgery, and there is potential to decrease response times if there is a larger group of evaluators. This study achieved response times <5 min. with 13 evaluators and it would be anticipated that a system with 50-100 evaluators would be faster, and become more accurate with time and experience. In addition, there is evidence that picture analysis can be accurate even if the evaluators are not experts but trained lay people14. Universal adoption of this approach would provide support for LC to surgeons in medical facilities regardless of location; cell phones are universally available and the camera resolutions are excellent, evaluators could be located anywhere and limitations of QA during typical hours could be eliminated. Artificial intelligence and computer learning applications to grade CVS is an area for future exploration and could evolve directly out of an image collection and grading system like live QA. Biliary injuries during LC are rare but consistently higher during LC than open cholecystectomy3. Previously published data shows this difference is mostly due to visual misperceptions during LC rather than errors in skill, knowledge, or judgment2,4,5. This leads us to the tenant of “getting help” or a second set of eyes when a dissection seems different or confusing. Live QA is an instant mechanism to attain that second opinion. CVS images from peer surgeons increases exposure to the CVS beyond what one would experience in their own practice, leading to a better
10
understanding of biliary anatomy and its variations that may reduce future identification errors during LC. The other mechanism of injury is the altered surgical field. This occurs with inflammation and a fusing of the gallbladder outflow tract to the cystic plate and the hepatoduodenal ligament. Although this system does not directly address the fused cystic plate, an image sharing system could be modified to help address this issue. This system also provides surgical coaching through feedback from remote peer surgeons, which provides ongoing education as well as an opportunity for identification errors to be corrected before a bile duct injury or other major error occurs. Routine intraoperative pre-clipping timeout with photo documentation is a current recurring theme in the literature and at meetings as a probable way to decrease common bile duct injury15. There are several philosophes about operative photo-documentation. On one hand, it may risk exposure to medical liability. On the other, it enables documentation of the actual conditions of the surgery and the quality of the dissection. Implementation would raise concerns with image storage, access and discoverability. CVS image sharing/grading as a part of a quality initiative and used only to promote safety may not be discoverable and may not have to be included in the medical record. Destruction of images after 30 days and/or use in QA if there are complications are reasonable and could be incorporated into an implementation plan. There are surgeon specific considerations associated with any QA intervention. Reimbursement requiring photo-documentation (eg.blephroplasties) is already established in Medicare. It is reasonable to assume these mandates from payors will expand. These complexities directly impact surgical practice. The toll on the surgeon, the addition of workflow and excessive oversight are valid and meaningful concerns. This issue is fluid as of now, but it will be surprising if photo-documentation is not a required part of medical and surgical documentation, billing
11
and quality assurance in the future. As a community, benefit could be derived from being proactive in inventing and driving any system where considerable disruption is possible. There are potential educational aspects to live QA. A grading “learners” crowd could be established separate from the QA group. Their individual performance compared to the vetted QA graders would be reported directly to them. Even the aspects of the critical view with which they are struggling would be reportable. In addition, using this system in a training program supports the desired atmosphere of cooperation and open discussion. A mobile application is a comfortable environment for sharing and feedback. This study serves as a good working model that would benefit from further modification. For example, it would be optimal if both the operating surgeon and CVS evaluators were anonymous to avoid unintended bias. The addition of a mechanism where CVS images are available for a short time before they become inaccessible to evaluators would mitigate the impression of PHI compromise. Further, the ability to assess the consistency and quality of the graders should be addressed, particularly when a system is designed to evaluate quality. These modifications could be included in software designed and developed for the purposes of CVS sharing and review. One can expect dedicated applications may become available as the need for photodocumentation, quality assurance and surgical safety and education continue to increase. Conclusions Live QA is a system that can be freely used by groups of surgeons wishing to teach and standardize cholecystectomy dissection. A similar system may be designed for other surgeries or medical determinants where image grading is useful. Pathology, joint replacement or dermatology applications come to mind. The study presented here demonstrates the feasibility of using an image sharing system during surgery to provide immediate QA. Acute cholecystitis impacted the grad-
12
ing and there were instances when a CVS was not attainable. Larger groups and crossinstitutional studies are needed to further investigate image sharing and grading and how it applies to quality in surgery.
Acknowledgements The authors would like to thank Dr Amit K Saha for statistical help, and Drs Jason Hoth, Lauren McCormack, and Clancy J Clark for assistance with the submission and evaluation of this manuscript.
13
References 1.
Strasberg SM, Brunt LM. Rationale and use of the critical view of safety in laparoscopic cholecystectomy. J Am Coll Surg. 2010;211(1):132-138.
2.
Barrett M, Asbun HJ, Chien HL, et al. Bile duct injury and morbidity following cholecystectomy: a need for improvement. Surg Endosc. 2018;32(4):1683-1688.
3.
MacFadyen BV, Jr., Vecchio R, Ricardo AE, Mathis CR. Bile duct injury after laparoscopic cholecystectomy. The United States experience. Surg Endosc. 1998;12(4):315-321.
4.
Strasberg SM, Eagon CJ, Drebin JA. The "hidden cystic duct" syndrome and the infundibular technique of laparoscopic cholecystectomy--the danger of the false infundibulum. J Am Coll Surg. 2000;191(6):661-667.
5.
Way LW, Stewart L, Gantert W, et al. Causes and prevention of laparoscopic bile duct injuries: analysis of 252 cases from a human factors and cognitive psychology perspective. Ann Surg. 2003;237(4):460-469.
6.
Strasberg SM. A three-step conceptual roadmap for avoiding bile duct injury in laparoscopic cholecystectomy: an invited perspective review. J Hepatobiliary Pancreat Sci. 2019;26(4):123-127.
7.
Strasberg SM. A teaching program for the "culture of safety in cholecystectomy" and avoidance of bile duct injury. J Am Coll Surg. 2013;217(4):751.
8.
Sanford DE, Strasberg SM. A simple effective method for generation of a permanent record of the Critical View of Safety during laparoscopic cholecystectomy by intraoperative "doublet" photography. J Am Coll Surg. 2014;218(2):170-178.
14
9.
https://www.preventbdi.org. State-of the-Art Consensus Conference of Prevention of Bile Duct Injury during Cholecystectomy: Consensus Recommendations. Final-Prevent-BDIConsensus-Recommendations-101818-web. 2018.
10.
Carr BD, Matusko N, Sandhu G, Varban OA. Cut or Do Not Cut? Assessing Perceptions of Safety During Laparoscopic Cholecystectomy Using Surgical Videos. J Surg Educ. 2018;75(6):1583-1588.
11.
Lam T, Usatoff V, Chan ST. Are we getting the critical view? A prospective study of photographic documentation during laparoscopic cholecystectomy. HPB (Oxford). 2014;16(9):859-863.
12.
Stefanidis D, Chintalapudi N, Anderson-Montoya B, et al. How often do surgeons obtain the critical view of safety during laparoscopic cholecystectomy? Surg Endosc. 2017;31(1):142-146.
13.
Wauben LS, van Grevenstein WM, Goossens RH, et al. Operative notes do not reflect reality in laparoscopic cholecystectomy. Br J Surg. 2011;98(10):1431-1436.
14.
Deal SB, Stefanidis D, Telem D, et al. Evaluation of crowd-sourced assessment of the critical view of safety in laparoscopic cholecystectomy. Surg Endosc. 2017;31(12):50945100.
15.
Singh R, Brunt LM. Critical view of safety—its feasibility and efficacy in preventing bile duct injuries. Annals of Laparoscopic and Endoscopic Surgery. 2018;3(1).
15
Table 1. Critical View of Safety Grading Scale CVS Aspect
Point Value 2
2 structures connected to the gallbladder
1
0 2 Cystic plate
1 0 2
Clearance of hepatocystic triangle
1
0 CVS, critical view of safety
Definition 2 clearly evident structures seen connecting to the gallbladder 2 structures seen connecting to the gallbladder, but there is some overlap of duct and artery or some technical artifact that interferes with clarity of determination 2 structures cannot be seen due to overlap or technical issues Cystic plate is clearly visible to approximately its bottom one third Cystic plate is visible, but overlapped with other structures and an insufficient amount of the plate is visible Cystic plate is not visible due to positioning or obstruction of view by instruments or clots Hepatocystic triangle is cleared of tissue with clear visibility of cystic structures and plate; viewer can be certain that no other structures are in the triangle Less than the whole triangle can be clearly seen with reduced confidence that there are no other structures in the triangle Tissue in the triangle obscures view of cystic structures and cystic plate preventing the determination that there are no other structures in the triangle
16
Variable Scored images, n Image score, mean ± SEM
Anterior CVS (without inflammation) 118
Posterior CVS (without inflammation) 118
Anterior CVS (with inflammation) 74
Posterior CVS (with inflammation) 74
5.4 ± 0.1
5.2 ± 0.1
5.4 ± 0.1
5.0 ± 0.1
Table 2. Mean Critical View of Safety Scores with and without Inflammation CVS, critical view of safety
17
Figure Legends Figure 1. Mean and median scores from evaluating surgeons for all cases. (A) A total of 193 laparoscopic cholecystectomies (LCs) with (n=74) and without (n=119) inflammation were analyzed for number of responses (mean ± SEM) from remote surgeons is reported. (B) No significant difference was observed between median number of responses for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD; n=37) compared with LC during normal business hours (8AM-8PM; n=156). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between groups. An unpaired Student’s t-test was used to compare the number of responses between normal and non-business hours. Figure 2. Mean response times from evaluating surgeons for all cases. Shared laparoscopic cholecystectomy (LC) images (posterior and anterior) were evaluated for critical view of safety (CVS) and first response time(s) by recorded. Real-time communication between sharing intraoperative images and CVS evaluation occurred, on average, in less than 5 minutes (n=193). Time to a second response (n=173) was significantly longer (11 minutes). No significant difference was observed between response times for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD) compared with LC during normal business hours (8AM-8PM). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between response groups. An unpaired Student’s t-test was used to compare the first and second response times. *Statistically significant. Figure 3. Mean and median scores and time of day for grading critical view of safety (CVS) images. In patients with acute cholecystitis, intraoperative images were shared between operative surgeons and other surgeons in the study located remotely. Real-time communication between
18
operative and remote surgeons and an established scoring system was used to evaluate a CVS (anterior and posterior). (A) A total of 74 laparoscopic cholecystectomies (LCs) with inflammation were analyzed for number of responses (mean ± SEM) from remote surgeons is reported. (B) No significant difference was observed between median number of responses for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD; n=29) compared with LC during normal business hours (8AM-8PM; n=45). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between groups. An unpaired Student’s t-test was used to compare the number of responses between normal and non-business hours. Figure 4. Mean and median response times for first and second grading of critical view of safety (CVS) images. In patients with acute cholecystitis, shared laparoscopic cholecystectomy (LC) images (posterior and anterior) were evaluated for CVS and first response time(s) by recorded. Real-time communication between surgeons occurred, on average, in less than 5 minutes (n=74). Time to a second response (n=64) was significantly longer (15 minutes). No significant difference was observed between response times for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD) compared with LC during normal business hours (8AM-8PM). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between response groups. An unpaired Student’s t-test was used to compare first and second response times. *Statistically significant. Figure 5. Grading consistency, biliary colic vs acute cholecystitis. Anterior and posterior intraoperative images were shared between operative surgeons and other surgeons in the study located remotely. Responses were analyzed for % agreement between scores for each patient (n=173). (A) Peer agreement (mean % ± SEM) from remote surgeons is reported. No significant
19
difference was observed between anterior and posterior % agreement. (B) In patients with cholecystitis (n=64), while there was agreement between peers, a significant difference was observed between anterior and posterior % agreement. A paired Student’s t-test was used to compare agreement between anterior and posterior scores. *Statistically significant.
20
Precis: An established image sharing and grading system for the critical view of safety (CVS) was used for real-time intraoperative feedback without increasing operative time or compromising private health information. Decreased CVS quality and grading agreement was observed in patients with cholecystitis.
Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images Prompt responses <5 min
Feasibility of medical image crowd sourcing Instant intraoperative QA
High concordance Teaching tool
Day or night Real time text string among surgeons of CVS in OR with HPI protected Sobba et al, J Am Coll Surg, February 2020
Fosters transparency and accountability Expanded uses possible
S1072-7515(19)32166-0
DOI:
https://doi.org/10.1016/j.jamcollsurg.2019.09.022
Reference:
ACS 9669
To appear in:
Journal of the American College of Surgeons
Received Date: 30 July 2019 Revised Date:
30 September 2019
Accepted Date: 30 September 2019
Please cite this article as: Sobba KB, Fernandez AZ, McNatt SS, Powell MS, Nunn AM, Hildreth AN, Yoza BK, Gross JL, Miller III PR, Westcott CJ, Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images, Journal of the American College of Surgeons (2019), doi: https://doi.org/10.1016/j.jamcollsurg.2019.09.022. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc. on behalf of the American College of Surgeons.
1
Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images Kathryn B Sobba, MDa, Adolfo Z Fernandez, MD, FACSa, Stephen S McNatt, MD, FACSa, Myron S Powell, MD, FACSa, Andrew M Nunn, MD, FACSa, Amy N Hildreth, MD, FACSa, Barbara K Yoza, PhDa, Jessica L Gross, MD, FACSa, Preston R Miller III, MD, FACSa, Carl J Westcott, MD, FACSa,c Author affiliations a Department of Surgery, Wake Forest School of Medicine, Winston-Salem, NC b Department of Anesthesiology, Wake Forest School of Medicine, Winston-Salem, NC c Department of Surgery, W.G. Bill Hefner VA Medical Center, Salisbury, NC Corresponding author Carl J Westcott, MD, FACS Dept. of Surgery, Wake Forest School of Medicine Medical Center Blvd. Winston-Salem, NC, 27157 E-mail: [email protected] Phone: 1-336-716-2409 Disclosure Information: Nothing to disclose. Presented in part at the American College of Surgeons 104th Annual Clinical Congress, Boston, MA, October 2018. Brief Title Image Sharing for Surgical Quality Assurance
2
BACKGROUND: The technique for attaining photographic evidence of the critical view of safety (CVS) in laparoscopic cholecystectomy (LC) has previously been defined, however the consistency, accuracy and feasibility CVS in practice is unknown. The aim of this study was to use an already established image sharing and grading system to determine the feasibility of timely feedback after sharing intraoperative images of the CVS and to evaluate if and how cholecystitis impacts the ability to attain a CVS. STUDY DESIGN: 193 laparoscopic cholecystectomies performed by 14 surgeons between August 2017 and January 2019 were studied. Anterior and posterior intraoperative CVS images were shared using a standard multimedia messaging system (MMS). Images were graded remotely by members of the group using an established scoring system and their time to response and scores recorded. Response data was analyzed for the ability to attain timely and consistent CVS scores. RESULTS: 74 urgent laparoscopic cholecystectomies for acute cholecystitis and 119 non urgent cholecystectomies were performed during the study period. Scoring of shared images occurred in less than 5 minutes and peer review (mean 3 responses) showed agreement that was not significantly different. In patients with acute cholecystitis, a small but significant difference was observed between anterior and posterior image scoring agreement. CONCLUSION: An established image sharing and grading system for CVS can be used for realtime intraoperative feedback without increasing operative time or compromising private health information. The CVS is almost always attainable, however decreases in CVS quality and grading agreement are observed in patients with acute cholecystitis. KEYWORDS: critical view of safety, intraoperative image capture, safe cholecystectomy
3
ABBREVIATIONS CVS
critical view of safety
MMS
multimedia messaging system
PHI
private health information
LC
laparoscopic cholecystectomy
SAGES
Society of American Gastrointestinal and Endoscopic Surgeons
QA
quality assurance
4
Introduction Over 750,000 cholecystectomies are performed each year in the United States1. Despite advancements in laparoscopic techniques, the rate of bile duct injuries and leaks remains higher in laparoscopic cholecystectomies (LC) than in open1. Although reported bile duct injuries remain relatively infrequent, occurring in approximately 3 per 1,000 LCs performed, the true incidence of bile duct injury is unknown and can lead to devastating complications for patients2-5. In addition, acute cholecystitis alters normal anatomy makes adequate dissection and correct identification of cystic structures more difficult, which likely contributes to the increased frequency of bile duct injuries that is seen in these patients during LC.6 The Critical View of Safety (CVS) is a technique widely believed to decrease the risk of bile duct injuries in LC7. The technique for attaining photographic evidence of a CVS is well defined and includes the following three criteria: 1. the hepatocystic triangle is cleared of fat and fibrous tissue, 2. the lower one third of the gallbladder is separated from the liver to expose the cystic plate, 3. two and only two structures are seen entering the gallbladder8. To adopt a universal culture of safety and minimize the risk for bile duct injuries during LC, the Society of American Gastrointestinal and Endoscopic Surgeons (SAGES) safe cholecystectomy task force recommends using the CVS technique and liberal use of bile duct imaging. Attaining the CVS is recommended in all cases and they advocate an intraoperative pre-clipping timeout to confirm that the CVS has been achieved. If available, consulting a peer surgeon is advised if the anatomy is unclear or under conditions deemed “difficult” by the surgeon. In addition, a Multi-Society Bile Duct Injury Consensus Conference (2018) established consensus recommendations for the CVS itself and discussed the concept of an intraoperative time out and photo-documentation of the dissection9. However, surgeons performing LCs do not always have
5
the benefit of an available peer surgeon, time outs are not established and accepted, and the idea of photo-documentation of the dissection is understudied and generally resisted. The aim of this study was to use instant messaging images of the CVS to establish a method for photo-documentation and intraoperative time out. Image capture was by mobile phone camera and a mainstream multimedia messaging system (MMS) was used to share CVS images. A standard grading system was employed8. The study sought to establish proof of concept and to determine if pericystic inflammation affects CVS image quality and/or grading agreement. Methods Study population. LCs performed by the study surgeons at Wake Forest Baptist Medical Center between August 2017 and January 2019 were included, regardless of patient age, comorbidities, or the nature of the procedure (planned or emergent). The study was reviewed and approved by the Wake Forest School of Medicine institutional review board. The MMS group (study group) consisted of 14 attending surgeons. All were invited and able to submit and review images at their own volition. Participants were able to silence the MMS group chat on their device as they pleased. This quality assurance (QA) project was reviewed and carried out with department and institutional approval (IRB #00057007). Image acquisition, sharing and grading. During each LC, an intraoperative timeout was performed by the operating surgeon to confirm adequate CVS dissection, capture anterior and posterior CVS images on their mobile phones, and share images with the study group in real time using the standard MMS application. Care was taken to have only the operative screen in the picture devoid of any other information. The operative surgeon directed the image acquisition. Study surgeons remotely graded the adequacy of the CVS as demonstrated by the shared imag-
6
es8. The response times and scores for anterior and posterior views were recorded. No PHI was shared, and the grading surgeons were instructed to delete the photos from their devices once grading was completed. The patients identifying information was entered in a separate secure HIPPA compliant data base and the de-identified images later matched to the medical record. The grading process was discussed and practiced prior to starting the QA project. Study surgeons reviewed previously published methods to capture and grade CVS images8. Two meetings of the group and a presentation to the entire department were held to review stock images and practice grading. The grading scheme is anterior image first (3 criteria) and then the posterior image (3 criteria). The grades are listed from 0 to 2 for each aspect of the CVS and are defined in Table 1 (Table 1). The sequence of the CVS aspect grades is: two and only two structures, followed by the cystic plate for the lower 1/3 of the liver bed, and the hepatocystic triangle cleared of all fat and fibrous tissue. For example: 222/222 is a perfect score. 122/222 would indicate that the evaluator could not clearly determine if there were two and only two structures on the anterior view. In this manner the specific deficiencies are described to the submitting surgeon. Each evaluation is viewable by the study group and the surgeon requesting the evaluation. Data analysis. The primary purpose for data analysis was to determine feasibility. Average response time and % agreement between CVS scores by responding surgeons was determined. Pathology reports were reviewed to document when acute cholecystitis was present. Mean time to first response, comparison of response times during normal business hours and nights/weekends, differences in CVS scores in anterior and posterior images, and an effect of pericystic inflammation on CVS quality and grading was assessed. Final pathology reports were reviewed for all study patients and the presence or absence of acute cholecystitis was recorded; this data is reported and indicated in the figure legends. Data was presented and analyzed by one-
7
way ANOVA with Tukey’s post-test between groups (alpha 0.05, 95%Cl) or unpaired student ttest between two groups using GraphPad Prism 7.04 software (San Diego, CA). Results After a 30 day follow up, there were no bile leaks, bile duct injuries or bleeding, or return to the operating room for these patients. An average of 3 evaluations were received for all LCs performed during the study period (Fig. 1A); no difference in the number of responses was observed for surgeries performed during normal business hours compared to LCs that occurred on nights and/or weekends (Fig. 1B). The average time for a response from another surgeon was less than 5 min. (3.7+0.5 min.) and a response from a second surgeon took significantly longer (10.7+1.6 min.) (Fig.2). There was no difference in first or second feedback times during the day compared to nights/weekends (Fig.2). To determine if pericystic inflammation affects CVS quality and grading agreement, data for acute cholecystitis patients (74 patients of the total 193) were examined in a separate analysis (Table 2) . Comparable to all patients, an average of 3 evaluations was received for shared CVS images of acute cholecystitis LCs (Fig. 3) with a response from another surgeon in less than 5 min. (3.5+0.7 min.) (Fig. 4); similar to all patients, a response from a second surgeon took significantly longer (15.4+3.7 min.). There was no difference in first or second feedback times during the day compared to nights/weekends (Fig. 4). The responses were analyzed for % agreement between scores for each patient (n=173). An image was rated as demonstrating CVS satisfactorily when the mean score of >2 raters was greater than 5 on at least one image8. There was 85%+2.1% and 82%+2.2% agreement in scoring anterior and posterior images of the CVS, respectively (Fig. 5A). Further analysis of the acute cholecystitis patient image grading showed that while there was agreement between surgical
8
peers (anterior CVS, 81%+3.7%; posterior CVS 73%+4.1%), a significant difference was observed in grading agreement of the posterior CVS (Fig. 5B). These results show that adequate CVS is almost always attainable in acute cholecystitis, however pericystic inflammation decreases CVS quality and grading agreement, and the posterior view suffers and varies the most Discussion Adequate CVS dissection during LC is recommended to avoid bile duct injury.1,3,7 Studies of the quality of CVS in LC have generally questioned the understanding of the CVS among surgeons. In addition, the consistency and accuracy of the CVS in practice has been documented to be poor10-13. These studies point to the need for a mechanism to standardize and teach the CVS. Our data set is small with the intent to test feasibility and proof of concept. A multiinstitutional effort would be needed for a prospective study to test the impact on bile duct injuries and other surgery related complications. However, live QA instituted over several institutions or through a medical society project could accumulate the numbers typically discussed as needed for a prospective analysis of cholecystectomy with respect to bile duct injuries. Modifications to the approach used in this study should be considered for a multi-institutional study and are outlined later in the discussion. Our grades showed a high level of concordance. The routine practice of attaining a critical view, sharing it and having it evaluated by a group of experts, or your partners and peers should support making the CVS dissection uniform within a group. If a small group of surgeons employs this approach, they would need to agree on and then, through the sharing and grading process, migrate towards a uniform understanding of the CVS. If adopted by an institution, health care system or a specialty society, it would serve as a way to standardize dissection for
9
much larger groups. Concordance and variation in grading as well as changes with respect to grading are interesting future projects in this line of investigation. The concordance could also be explained by grader bias. An unmodified MMS system is not ideal for this application as any one reviewer’s grades and identity are visible to all other reviewers. A system where all reviewers are blinded to other reviewers is needed prior to data collection and analysis that could address this question. Our data shows using the live quality assurance (QA) system does not add a significant amount of time to a surgery, and there is potential to decrease response times if there is a larger group of evaluators. This study achieved response times <5 min. with 13 evaluators and it would be anticipated that a system with 50-100 evaluators would be faster, and become more accurate with time and experience. In addition, there is evidence that picture analysis can be accurate even if the evaluators are not experts but trained lay people14. Universal adoption of this approach would provide support for LC to surgeons in medical facilities regardless of location; cell phones are universally available and the camera resolutions are excellent, evaluators could be located anywhere and limitations of QA during typical hours could be eliminated. Artificial intelligence and computer learning applications to grade CVS is an area for future exploration and could evolve directly out of an image collection and grading system like live QA. Biliary injuries during LC are rare but consistently higher during LC than open cholecystectomy3. Previously published data shows this difference is mostly due to visual misperceptions during LC rather than errors in skill, knowledge, or judgment2,4,5. This leads us to the tenant of “getting help” or a second set of eyes when a dissection seems different or confusing. Live QA is an instant mechanism to attain that second opinion. CVS images from peer surgeons increases exposure to the CVS beyond what one would experience in their own practice, leading to a better
10
understanding of biliary anatomy and its variations that may reduce future identification errors during LC. The other mechanism of injury is the altered surgical field. This occurs with inflammation and a fusing of the gallbladder outflow tract to the cystic plate and the hepatoduodenal ligament. Although this system does not directly address the fused cystic plate, an image sharing system could be modified to help address this issue. This system also provides surgical coaching through feedback from remote peer surgeons, which provides ongoing education as well as an opportunity for identification errors to be corrected before a bile duct injury or other major error occurs. Routine intraoperative pre-clipping timeout with photo documentation is a current recurring theme in the literature and at meetings as a probable way to decrease common bile duct injury15. There are several philosophes about operative photo-documentation. On one hand, it may risk exposure to medical liability. On the other, it enables documentation of the actual conditions of the surgery and the quality of the dissection. Implementation would raise concerns with image storage, access and discoverability. CVS image sharing/grading as a part of a quality initiative and used only to promote safety may not be discoverable and may not have to be included in the medical record. Destruction of images after 30 days and/or use in QA if there are complications are reasonable and could be incorporated into an implementation plan. There are surgeon specific considerations associated with any QA intervention. Reimbursement requiring photo-documentation (eg.blephroplasties) is already established in Medicare. It is reasonable to assume these mandates from payors will expand. These complexities directly impact surgical practice. The toll on the surgeon, the addition of workflow and excessive oversight are valid and meaningful concerns. This issue is fluid as of now, but it will be surprising if photo-documentation is not a required part of medical and surgical documentation, billing
11
and quality assurance in the future. As a community, benefit could be derived from being proactive in inventing and driving any system where considerable disruption is possible. There are potential educational aspects to live QA. A grading “learners” crowd could be established separate from the QA group. Their individual performance compared to the vetted QA graders would be reported directly to them. Even the aspects of the critical view with which they are struggling would be reportable. In addition, using this system in a training program supports the desired atmosphere of cooperation and open discussion. A mobile application is a comfortable environment for sharing and feedback. This study serves as a good working model that would benefit from further modification. For example, it would be optimal if both the operating surgeon and CVS evaluators were anonymous to avoid unintended bias. The addition of a mechanism where CVS images are available for a short time before they become inaccessible to evaluators would mitigate the impression of PHI compromise. Further, the ability to assess the consistency and quality of the graders should be addressed, particularly when a system is designed to evaluate quality. These modifications could be included in software designed and developed for the purposes of CVS sharing and review. One can expect dedicated applications may become available as the need for photodocumentation, quality assurance and surgical safety and education continue to increase. Conclusions Live QA is a system that can be freely used by groups of surgeons wishing to teach and standardize cholecystectomy dissection. A similar system may be designed for other surgeries or medical determinants where image grading is useful. Pathology, joint replacement or dermatology applications come to mind. The study presented here demonstrates the feasibility of using an image sharing system during surgery to provide immediate QA. Acute cholecystitis impacted the grad-
12
ing and there were instances when a CVS was not attainable. Larger groups and crossinstitutional studies are needed to further investigate image sharing and grading and how it applies to quality in surgery.
Acknowledgements The authors would like to thank Dr Amit K Saha for statistical help, and Drs Jason Hoth, Lauren McCormack, and Clancy J Clark for assistance with the submission and evaluation of this manuscript.
13
References 1.
Strasberg SM, Brunt LM. Rationale and use of the critical view of safety in laparoscopic cholecystectomy. J Am Coll Surg. 2010;211(1):132-138.
2.
Barrett M, Asbun HJ, Chien HL, et al. Bile duct injury and morbidity following cholecystectomy: a need for improvement. Surg Endosc. 2018;32(4):1683-1688.
3.
MacFadyen BV, Jr., Vecchio R, Ricardo AE, Mathis CR. Bile duct injury after laparoscopic cholecystectomy. The United States experience. Surg Endosc. 1998;12(4):315-321.
4.
Strasberg SM, Eagon CJ, Drebin JA. The "hidden cystic duct" syndrome and the infundibular technique of laparoscopic cholecystectomy--the danger of the false infundibulum. J Am Coll Surg. 2000;191(6):661-667.
5.
Way LW, Stewart L, Gantert W, et al. Causes and prevention of laparoscopic bile duct injuries: analysis of 252 cases from a human factors and cognitive psychology perspective. Ann Surg. 2003;237(4):460-469.
6.
Strasberg SM. A three-step conceptual roadmap for avoiding bile duct injury in laparoscopic cholecystectomy: an invited perspective review. J Hepatobiliary Pancreat Sci. 2019;26(4):123-127.
7.
Strasberg SM. A teaching program for the "culture of safety in cholecystectomy" and avoidance of bile duct injury. J Am Coll Surg. 2013;217(4):751.
8.
Sanford DE, Strasberg SM. A simple effective method for generation of a permanent record of the Critical View of Safety during laparoscopic cholecystectomy by intraoperative "doublet" photography. J Am Coll Surg. 2014;218(2):170-178.
14
9.
https://www.preventbdi.org. State-of the-Art Consensus Conference of Prevention of Bile Duct Injury during Cholecystectomy: Consensus Recommendations. Final-Prevent-BDIConsensus-Recommendations-101818-web. 2018.
10.
Carr BD, Matusko N, Sandhu G, Varban OA. Cut or Do Not Cut? Assessing Perceptions of Safety During Laparoscopic Cholecystectomy Using Surgical Videos. J Surg Educ. 2018;75(6):1583-1588.
11.
Lam T, Usatoff V, Chan ST. Are we getting the critical view? A prospective study of photographic documentation during laparoscopic cholecystectomy. HPB (Oxford). 2014;16(9):859-863.
12.
Stefanidis D, Chintalapudi N, Anderson-Montoya B, et al. How often do surgeons obtain the critical view of safety during laparoscopic cholecystectomy? Surg Endosc. 2017;31(1):142-146.
13.
Wauben LS, van Grevenstein WM, Goossens RH, et al. Operative notes do not reflect reality in laparoscopic cholecystectomy. Br J Surg. 2011;98(10):1431-1436.
14.
Deal SB, Stefanidis D, Telem D, et al. Evaluation of crowd-sourced assessment of the critical view of safety in laparoscopic cholecystectomy. Surg Endosc. 2017;31(12):50945100.
15.
Singh R, Brunt LM. Critical view of safety—its feasibility and efficacy in preventing bile duct injuries. Annals of Laparoscopic and Endoscopic Surgery. 2018;3(1).
15
Table 1. Critical View of Safety Grading Scale CVS Aspect
Point Value 2
2 structures connected to the gallbladder
1
0 2 Cystic plate
1 0 2
Clearance of hepatocystic triangle
1
0 CVS, critical view of safety
Definition 2 clearly evident structures seen connecting to the gallbladder 2 structures seen connecting to the gallbladder, but there is some overlap of duct and artery or some technical artifact that interferes with clarity of determination 2 structures cannot be seen due to overlap or technical issues Cystic plate is clearly visible to approximately its bottom one third Cystic plate is visible, but overlapped with other structures and an insufficient amount of the plate is visible Cystic plate is not visible due to positioning or obstruction of view by instruments or clots Hepatocystic triangle is cleared of tissue with clear visibility of cystic structures and plate; viewer can be certain that no other structures are in the triangle Less than the whole triangle can be clearly seen with reduced confidence that there are no other structures in the triangle Tissue in the triangle obscures view of cystic structures and cystic plate preventing the determination that there are no other structures in the triangle
16
Variable Scored images, n Image score, mean ± SEM
Anterior CVS (without inflammation) 118
Posterior CVS (without inflammation) 118
Anterior CVS (with inflammation) 74
Posterior CVS (with inflammation) 74
5.4 ± 0.1
5.2 ± 0.1
5.4 ± 0.1
5.0 ± 0.1
Table 2. Mean Critical View of Safety Scores with and without Inflammation CVS, critical view of safety
17
Figure Legends Figure 1. Mean and median scores from evaluating surgeons for all cases. (A) A total of 193 laparoscopic cholecystectomies (LCs) with (n=74) and without (n=119) inflammation were analyzed for number of responses (mean ± SEM) from remote surgeons is reported. (B) No significant difference was observed between median number of responses for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD; n=37) compared with LC during normal business hours (8AM-8PM; n=156). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between groups. An unpaired Student’s t-test was used to compare the number of responses between normal and non-business hours. Figure 2. Mean response times from evaluating surgeons for all cases. Shared laparoscopic cholecystectomy (LC) images (posterior and anterior) were evaluated for critical view of safety (CVS) and first response time(s) by recorded. Real-time communication between sharing intraoperative images and CVS evaluation occurred, on average, in less than 5 minutes (n=193). Time to a second response (n=173) was significantly longer (11 minutes). No significant difference was observed between response times for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD) compared with LC during normal business hours (8AM-8PM). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between response groups. An unpaired Student’s t-test was used to compare the first and second response times. *Statistically significant. Figure 3. Mean and median scores and time of day for grading critical view of safety (CVS) images. In patients with acute cholecystitis, intraoperative images were shared between operative surgeons and other surgeons in the study located remotely. Real-time communication between
18
operative and remote surgeons and an established scoring system was used to evaluate a CVS (anterior and posterior). (A) A total of 74 laparoscopic cholecystectomies (LCs) with inflammation were analyzed for number of responses (mean ± SEM) from remote surgeons is reported. (B) No significant difference was observed between median number of responses for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD; n=29) compared with LC during normal business hours (8AM-8PM; n=45). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between groups. An unpaired Student’s t-test was used to compare the number of responses between normal and non-business hours. Figure 4. Mean and median response times for first and second grading of critical view of safety (CVS) images. In patients with acute cholecystitis, shared laparoscopic cholecystectomy (LC) images (posterior and anterior) were evaluated for CVS and first response time(s) by recorded. Real-time communication between surgeons occurred, on average, in less than 5 minutes (n=74). Time to a second response (n=64) was significantly longer (15 minutes). No significant difference was observed between response times for LC occurring during nights or weekends (WKD), (8PM-8AM or WKD) compared with LC during normal business hours (8AM-8PM). One-way ANOVA with Tukey’s post-test (alpha 0.05, 95% CI) was used to analyze means between response groups. An unpaired Student’s t-test was used to compare first and second response times. *Statistically significant. Figure 5. Grading consistency, biliary colic vs acute cholecystitis. Anterior and posterior intraoperative images were shared between operative surgeons and other surgeons in the study located remotely. Responses were analyzed for % agreement between scores for each patient (n=173). (A) Peer agreement (mean % ± SEM) from remote surgeons is reported. No significant
19
difference was observed between anterior and posterior % agreement. (B) In patients with cholecystitis (n=64), while there was agreement between peers, a significant difference was observed between anterior and posterior % agreement. A paired Student’s t-test was used to compare agreement between anterior and posterior scores. *Statistically significant.
20
Precis: An established image sharing and grading system for the critical view of safety (CVS) was used for real-time intraoperative feedback without increasing operative time or compromising private health information. Decreased CVS quality and grading agreement was observed in patients with cholecystitis.
Live Quality Assurance: Using a Multimedia Messaging Service Group Chat to Instantly Grade Intraoperative Images Prompt responses <5 min
Feasibility of medical image crowd sourcing Instant intraoperative QA
High concordance Teaching tool
Day or night Real time text string among surgeons of CVS in OR with HPI protected Sobba et al, J Am Coll Surg, February 2020
Fosters transparency and accountability Expanded uses possible