Publications

Importance: 

Sacral neuromodulation (SNM) requires removal for infectious complications in 3-11%.

Objective: Does preoperative antibiotic choice impact device removal rates?

Study Design: 

This was a retrospective cohort analysis, using the Health Facts® Database, representing >750 hospitals. We included female patients undergoing SNM
implantation 2010-2018. Univariate and multivariate logistic regression identified factors associated with removal. 35 comorbidities were evaluated. Those with p<0.2 on univariate analysis were included in the multivariate analysis. We decided a priori to include prophylactic antibiotic choice in the final model.

Results: 

Of 1,433 patients, 170 (11.9%) had device removal. Subjects were 70.0 ± 14.9 years old, predominantly Caucasian (90.0%), treated in urban hospitals (94.1%),
and married (54.2%). 11.8% were obese, and 18.0% smoked. Those in the removal cohort were more likely from the Northeastern U.S. 52.3% received 1st gen cephalosporins (CPSN), 7.4% 2nd or 3rd gen CPSNs, 9.1% vancomycin, 13.4% aminoglycosides, 4.6% clindamycin, and 13.3% fluoroquinolones. Compared to vancomycin, more removals were associated with first generation CPSNs (OR=3.1, 95% confidence interval [1.4, 6.8]); clindamycin (OR=3.2, [1.2, 8.4]); second/third generation CPSNs (OR=3.1, [1.3, 7.6]); and aminoglycosides (OR = 3.1, [1.3, 7.4]). Additionally, patients treated in the Northeast were more likely to undergo removal (OR = 1.9, [1.0, 3.7]).

Conclusions: Vancomycin as a prophylactic antibiotic was associated with fewer device removals compared to most antibiotics in this retrospective cohort analysis. While prospective trials could confirm this benefit, low removal rates may make this impractical.

Keywords: Sacral Neuromodulation, antibiotics, urinary incontinence, postoperative complications, device removal, infection

BACKGROUND: Urinary tract infections (UTI) are among the most common infections encountered in both community and healthcare settings. Differentiating between community-associated UTI (CA-UTI) and healthcare-associated UTI (HA-UTI) is crucial for understanding their epidemiology, identifying risk factors, and developing appropriate treatment strategies. Machine learning (ML) techniques have shown significant potential in improving the accuracy of predicting these infections, enabling more effective interventions and better patient outcomes. While previous studies have demonstrated the utility of ML models in various healthcare settings, there is still a need for a comparative analysis of different ML approaches, particularly in distinguishing between CA-UTI and HA-UTI and assessing the risk of UTI among hospitalized patients.

OBJECTIVE: Using 2019-2023 patient demographics, hospital, and socioeconomic data, this study aims to build, validate, and compare machine learning models-Decision Tree (DT), Neural Network (NN), Logistic Regression (LR), Random Forest (RF), and Extreme Gradient Boosting (XGBoost) to differentiate between the incidences of HA-UTI and CA-UTI. Additionally, it seeks to identify key predictors of UTI using demographic, hospital, and socioeconomic variables.

RESULTS: The DT model demonstrated the highest sensitivity, particularly in handling the highly imbalanced data of HAI, with a sensitivity of 87%. LR achieved the best overall accuracy, at 95.9% for HA-UTI and 93.2% for HA-UTI vs. CA-UTI. RF performed best in cross-validation, reaching 99.1% for HA-UTI and 96.2% for HA-UTI vs. CA-UTI. NN showed the highest specificity, at 93.4%, for HA-UTI vs. CA-UTI. The AUC values further supported these findings, ranging from 71.9% for NN to 96% for RF, reflecting the robustness of these models across different annual datasets. Among patient demographics, hospital, and socioeconomic variables, all models consistently identified the nurse units (e.g., inpatient units and mental health units) as the most significant predictors of UTI. In addition to nurse units, LR and DT identified location (e.g., various clinics and medical centres) as a key predictor. For HA-UTI versus CA-UTI, variations were observed across the years, with patient age, median household income, and gender intermittently emerging as key predictors.

CONCLUSION: The predictive accuracy of the machine learning models is relatively the same, with some differences in sensitivity and specificity for identifying both HA-UTI vs. CA-UTI and HA-UTI. Nurse units consistently emerge as the most significant predictors across all years. The importance of all predictors, such as socioeconomic factors and location, varies from year to year, highlighting the need for incorporating those variables in the surveillance systems to optimize the accuracy of predictions.

Miscommunication in the OR is a threat to patient safety and surgical efficiency. Our objective was to measure the frequency and causes of communication interference between robotic team members. We observed 78 robotic surgeries over 215 h. 65.4% were General Surgery, most commonly cholecystectomy, identifying Speech Communication Interference (SCI) events, defined as “surgery-related group discourse that is disrupted according to the goals of the communication or the physical and situational context of the exchange”. We noted the causes and strategies to correct the miscommunication, near misses, and case delays associated with each SCI event. Post-surgery interviews supported observations and were analyzed thematically. Overall, we observed 687 SCI events (mean 8.8 ± 6.5 per case, 3.2 per hour), ranging from one to 28 per case. 48 (7.0%) occurred during docking and 136 (19.8%) occurred during a critical moment. The most common causes were concurrent tasks (66.1%); loud noises (10.8%) from patient cart, lightbox fan, and suction machine; and overlapping conversations (4.2%). 94.8% resulted in a case delay. These events distracted from monitoring patient safety and resulted in near misses. Mitigating strategies included leaning out of the surgeon console to repeat the message and employing a messenger. These findings help characterize miscommunication in robotic surgery. Possible interventions include microphones and headsets, positioning the surgeon console closer to the bedside, moving loud equipment further away, and upgrading the patient cart speaker.

Keywords: Anesthesia; Communication; Interprofessional teamwork; OR nursing staff; Robotic surgery; Surgical error.

PURPOSE:

The surgical clerkship provides medical students with valuable hands-on experience. This study examined why medical students speak or remain silent in the OR to improve progression from novice to engaged surgical team member.

METHODS:

Using Constructivist Grounded Theory 37 interviews were conducted concerning expectations and behaviors that encourage or discourage students from speaking during their clerkship. Transcripts were coded, analyzed, and triangulated to develop a conceptual model.

RESULTS:

Students' decision to speak or remain silent was based on their perception of the OR as a safe learning space. Our findings suggest that better preparation, awareness of critical moments, and informal communication with team members encouraged student speech.

CONCLUSIONS:

Medical students remain conflicted about their speaking in the OR and their evaluation. Key to improving students' psychological safety is establishment of interpersonal relationships, awareness of OR mood, and assignment of case-related tasks to assist with OR assimilation and improved learning.

Introduction and hypothesis:

Surgeon kinematics play a significant role in the prevention of patient injury. We hypothesized that elbow extension and ulnar wrist deviation are associated with bladder injury during simulated midurethral sling (MUS) procedures.

Methods:

We used motion capture technology to measure surgeons’ flexion/extension, abduction/adduction, and internal/external rotation angular time series for shoulder, elbow, and wrist joints. Starting and ending angles, minimum and maximum angles, and range of motion (ROM) were extracted from each time series. We created anatomical multibody models and applied linear mixed modeling to compare kinematics between trials with versus without bladder penetration and attending versus resident surgeons. A total of 32 trials would provide 90% power to detect a difference.

Results:

Out of 85 passes, 62 were posterior to the suprapubic bone and 20 penetrated the bladder. Trials with versus without bladder penetration were associated with more initial wrist dorsiflexion (−27.32 vs −9.03°, p = 0.01), less final elbow flexion (39.49 vs 60.81, p = 0.03), and greater ROM in both the wrist (27.48 vs 14.01, p = 0.02), and elbow (20.45 vs 12.87, p = 0.04). Wrist deviation and arm pronation were not associated with bladder penetration. Compared with attendings, residents had more ROM in elbow flexion (14.61 vs 8.35°, p < 0.01), but less ROM in wrist dorsiflexion (13.31 vs 20.33, p = 0.02) and arm pronation (4.75 vs 38.46, p < 0.01).

Conclusions:

Bladder penetration during MUS is associated with wrist dorsiflexion and elbow flexion but not internal wrist deviation and arm supination. Attending surgeons exerted control with the wrist and forearm, surgical trainees with the elbow. Our findings have direct implications for MUS teaching.

Keywords: Biomedical engineering; Kinematics; Midurethral sling; Motion capture; Patient safety; Surgical education; Surgical simulation.

Objective: 

In our previous work, teaching surgeons used potentially ambiguous language in the OR 12.3 times per minute. Our objectives were to examine ambiguous examples featuring Directional Frame of Reference (DFoR), which involves instructions containing directional terms like "up" or "left," and to uncover what contributes to understanding or misunderstanding of such instruction.

Design: 

We videorecorded the critical moments of 6 surgeries, as chosen by the surgeons. With a semanticist, we applied constructs from formal semantics to choose potentially ambiguous DFoR terms commonly flagged in our previous work. We separately interviewed attending and resident surgeons, asking each to describe the meaning of those DFoR terms while they viewed case recordings alongside transcripts. We compared their responses, analyzing them for agreement in direction. We performed thematic analysis on case and interview transcripts for themes related to DFoR.

Setting: 

Midwestern academic university teaching hospital.

Participants: 

Six attending and 6 resident surgeons.

Results: 

Attending and resident surgeons disagreed on direction in 9 of the 26 (34.6%) DFoR examples. Misunderstanding arose from using linear direction to describe three-dimensional space, e.g., "up" for anterior/cephalad/right. It also arose when combining degree modifiers with DfoR, e.g., "we're far enough back" combines the ambiguities of "back" (DfoR) and "far enough" (degree modifier). Use of axial parts (noun-like directional terms), e.g., "bottom," and confusing "left" for "right" also provoked misunderstanding. Misunderstanding was associated with lack of experience and mitigated by pointing with a finger or instrument, concurrent with speech.

Conclusions: 

Use of ambiguous language with DFoR incurs a high potential for misunderstanding, especially with novice surgeons. We recommend avoiding linear directions and axial parts, and instead physically pointing to represent complex 3D directions. Degree modifiers can be replaced with exact distances e.g., replace "little more anterior" with "1 centimeter anterior," and semaphores can be used to clarify direction.

Keywords: communication; linguistics; semantics; surgery; surgical education.

Highlights

• This is the first study applying linguistic inquiry and word count to intraoperative spoken language.

• Attending surgeons spoke more words and used more language of clout.

• Residents used more informal speech with an emotional tone and assent.

• There were no meaningful differences in language use according to gender.

Objective

Previous work has analyzed residency letters of recommendation for agentic and communal language, but this has not been applied to spoken language. Our objective was to analyze intraoperative spoken language by attending and resident surgeons for the use of agentic and communal language.

Design

We completed a linguistic inquiry and word count (LIWC) analysis on 16 operating room transcripts (total time 615 minutes) between attendings and resident surgeons for categories associated with agentic and communal speech. Wilcoxon signed rank and Mann-Whitney U tests were used to compare attending versus resident and male versus female speech patterns for word count; “I,” clout, and power (agentic categories); and “we,” authentic, social (communal categories).

Setting

Midwestern academic university teaching hospital.

Participants

Sixteen male (9 attendings, 7 residents) and 16 female (7 attendings, 9 residents) surgeons, from 6 surgical specialties, most commonly from General Surgery.

Results

Attending surgeons used more words per minute than residents (40.01 vs 16.92, p < 0.01), were less likely to use “I” (3.18 vs 5.53, p < 0.01), and spoke more language of “clout” (75.82 vs 55.47, p < 0.01). There were no significant differences between attendings and residents in use of analytic speech (23.72 vs 24.67, p = 0.32), “causation” (1.20 vs 1.08, p = 0.72), or “cognitive processing” (10.20 vs 10.54, p = 0.74). Residents used more speech with “emotional tone” (92.91 vs 79.92, p = 0.03), “positive emotion” (4.98 vs 3.86, p = 0.04), more “assent” language (4.89 vs 3.09, p < 0.01), and more “informal” language (9.27 vs 6.77, p < 0.01). There were no gender differences, except for male residents speaking with greater certainty than female residents, although by less than 1% of the total word count.

Conclusions

In the operating room, attending surgeons were more likely to use agentic language compared to resident surgeons based on LIWC analysis. These differences did not depend on gender and likely relate to surgeon experience and confidence, learning versus teaching, and power dynamics.

Key words

communication, surgery, surgical education, gender, language

Placental dysfunction is a leading cause of numerous pregnancy complications, including preeclampsia, preterm birth, fetal growth restrictions, placental abruption, and late spontaneous abortion [...].

IMPORTANCE: Sacral neuromodulation (SNM) requires removal for infectious complications in 3-11%.

OBJECTIVE: The objective of this study was to examine the effect of preoperative antibiotic choice on all-cause SNM device removal rates.

STUDY DESIGN: This was a retrospective cohort analysis, using the Health Facts Database, representing more than 750 hospitals. We included female patients undergoing SNM implantation from 2010 to 2018. Univariate and multivariate logistic regression identified factors associated with removal. Thirty-five comorbidities were evaluated. Those with P < 0.2 on univariate analysis were included in the multivariate analysis. We decided a priori to include prophylactic antibiotic choice in the final model.

RESULTS: Of 1,433 patients, 170 (11.9%) had device removal. Patients were 63.0 ± 14.9 years old, predominantly Caucasian (90.0%), treated in urban hospitals (94.1%), and married (54.2%). A total of 11.8% were obese, and 18.0% smoked. Those in the removal cohort were more likely from the Northeastern United States; 52.3% received first-gen cephalosporins (CPSN), 7.4% second- or third-generation CPSNs, 9.1% vancomycin, 13.4% aminoglycosides, 4.6% clindamycin, and 13.3% fluoroquinolones. Compared to vancomycin, more removals were associated with first-generation CPSNs (odds ratio [OR] = 3.1, 95% confidence interval [1.4, 6.8]); clindamycin (OR = 3.2, [1.2, 8.4]); second/third-generation CPSNs (OR = 3.1, [1.3, 7.6]); and aminoglycosides (OR = 3.1, [1.3, 7.4]). Additionally, patients treated in the Northeast were more likely to undergo removal (OR = 1.9, [1.0, 3.7]).

CONCLUSIONS: Vancomycin as a prophylactic antibiotic was associated with fewer device removals compared to most antibiotics in this retrospective cohort analysis. While prospective trials could confirm this benefit, low removal rates may make this impractical.