Automated segmentation of phases, steps, and tasks in laparoscopic cholecystectomy using deep learning.

BACKGROUND: Video-based review is paramount for operative performance assessment but can be laborious when performed manually. Hierarchical Task Analysis (HTA) is a well-known method that divides any procedure into phases, steps, and tasks. HTA requires large datasets of videos with consistent definitions at each level. Our aim was to develop an AI model for automated segmentation of phases, steps, and tasks for laparoscopic cholecystectomy videos using a standardized HTA. METHODS: A total of 160 laparoscopic cholecystectomy videos were collected from a publicly available dataset known as cholec80 and from our own institution. All videos were annotated for the beginning and ending of a predefined set of phases, steps, and tasks. Deep learning models were then separately developed and trained for the three levels using a 3D Convolutional Neural Network architecture. RESULTS: Four phases, eight steps, and nineteen tasks were defined through expert consensus. The training set for our deep learning models contained 100 videos with an additional 20 videos for hyperparameter optimization and tuning. The remaining 40 videos were used for testing the performance. The overall accuracy for phases, steps, and tasks were 0.90, 0.81, and 0.65 with the average F1 score of 0.86, 0.76 and 0.48 respectively. Control of bleeding and bile spillage tasks were most variable in definition, operative management, and clinical relevance. CONCLUSION: The use of hierarchical task analysis for surgical video analysis has numerous applications in AI-based automated systems. Our results show that our tiered method of task analysis can successfully be used to train a DL model.

Patient characteristics and diagnostic tests associated with syncopal falls: A Southwestern surgical congress multicenter study.

BACKGROUND: Patients suspected of syncope frequently undergo laboratory and imaging studies to determine the etiology of the syncope. Variability exists in these workups across institutions. The purpose of this study was to evaluate the utilization and diagnostic yield of these workups and the patient characteristics associated with syncopal falls. METHODS: A multi-institutional retrospective review was performed on adult patients admitted after a fall between 1/2017-12/2018. Syncopal falls were compared to non-syncopal falls. RESULTS: 4478 patients were included. There were 795 (18%) patients with a syncopal fall. Electrocardiogram, troponin, echocardiogram, CT angiography (CTA), and carotid ultrasound were more frequently tested in syncope patients compared to non-syncope patients. Syncope patients had higher rates of positive telemetry/Holter monitoring, CTAs, and electroencephalograms. CONCLUSION: Patients who sustain syncopal falls frequently undergo diagnostic testing without a higher yield to determine the etiology of syncope.

Keeping Residents in the Dark: Do Night-Float Rotations Provide a Valuable Educational Experience?

OBJECTIVE: To qualify and characterize resident overnight activity. DESIGN: A prospective 3-phase study was conducted of surgical residents with attention to activities performed on the overnight rotation: needs assessment, direct observation of activities, and feedback. SETTING: This study was conducted at the University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma. This is both a tertiary referral center and the only American College of Surgeons (ACS) verified level 1 trauma center in the state. PARTICIPANTS: This study included current surgical residents within the residency program. RESULTS: During the study period, 270 pages were individually recorded, with 60% of these pages defined as time-sensitive activities. In addition, most of the pages involved pressing patient-care issues irrespective of postgraduate year level. Analyses revealed that residents spend most of their time performing educational activities (62%). On feedback, residents reported overall satisfaction with the learning opportunities during night-shift (6.4/7.0) and indicated their perceptions of an adequate balance of service and education on night float (6.6/7.0). This correlates with our annual rotation assessment where residents identify night-float as an overall positive experience which provides educational benefit. CONCLUSIONS: Work-hour restrictions induce residency programs to adapt to new training models. Our results report a breakdown of resident activities while on night-float and demonstrate that overnight shifts continue to provide important educational opportunities during training.

Evaluating Coding Accuracy in General Surgery Residents' Accreditation Council for Graduate Medical Education Procedural Case Logs.

INTRODUCTION: The Accreditation Council for Graduate Medical Education (ACGME) case log captures resident operative experience based on Current Procedural Terminology (CPT) codes and is used to track operative experience during residency. With increasing emphasis on resident operative experiences, coding is more important than ever. It has been shown in other surgical specialties at similar institutions that the residents' ACGME case log may not accurately reflect their operative experience. What barriers may influence this remains unclear. As the only objective measure of resident operative experience, an accurate case log is paramount in representing one's operative experience. This study aims to determine the accuracy of procedural coding by general surgical residents at a single institution. METHODS: Data were collected from 2 consecutive graduating classes of surgical residents' ACGME case logs from 2008 to 2014. A total of 5799 entries from 7 residents were collected. The CPT codes entered by residents were compared to departmental billing records submitted by the attending surgeon for each procedure. Assigned CPT codes by institutional American Academy of Professional Coders certified abstract coders were considered the "gold standard." A total of 4356 (75.12%) of 5799 entries were identified in billing records. Excel 2010 and SAS 9.3 were used for analysis. In the event of multiple codes for the same patient, any match between resident codes and billing record codes was considered a "correct" entry. A 4-question survey was distributed to all current general surgical residents at our institution for feedback on coding habits, limitations to accurate coding, and opinions on ACGME case log representation of their operative experience. RESULTS: All 7 residents had a low percentage of correctly entered CPT codes. The overall accuracy proportion for all residents was 52.82% (range: 43.32%-60.07%). Only 1 resident showed significant improvement in accuracy during his/her training (p = 0.0043). The survey response rate was 100%. Survey results indicated that inability to find the precise code within the ACGME search interface and unfamiliarity with available CPT codes were by far the most common perceived barriers to accuracy. Survey results also indicated that most residents (74%) believe that they code accurately most of the time and agree that their case log would accurately represent their operative experience (66.6%). CONCLUSION: This is the first study to evaluate correctness of residents' ACGME case logs in general surgery. The degree of inaccuracy found here necessitates further investigation into the etiology of these discrepancies. Instruction on coding practices should also benefit the residents after graduation. Optimizing communication among attendings and residents, improving ACGME coding search interface, and implementing consistent coding practices could improve accuracy giving a more realistic view of residents' operative experience.