Interprofessional Collaboration: A Qualitative Study of Non-Physician Perspectives on Resident Competency.

BACKGROUND: The Association of American Medical Colleges (AAMC) includes the ability to collaborate in an interprofessional team as a core professional activity that trainees should be able to complete on day 1 of residency (Med Sci Educ. 26:797-800, 2016). The training that medical students require in order to achieve this competency, however, is not well established (Med Sci Educ. 26:457-61, 2016), and few studies have examined non-physician healthcare professionals' perspectives regarding resident physicians' interprofessional skills. OBJECTIVE: This study aims to describe non-physicians' views on barriers to collaboration with physicians, as well as factors that contribute to good collaborative relationships. PARTICIPANTS: Nurses, social workers, case managers, dietitians, rehabilitation therapists, and pharmacists at one academic medical center, largely working in the inpatient setting. APPROACH: A qualitative study design was employed. Data were collected from individual interviews and focus groups comprising non-physician healthcare professionals. KEY RESULTS: Knowledge gaps identified as impeding interprofessional collaboration included inadequate understanding of current roles, potential roles, and processes for non-physician healthcare professionals. Specific physician behaviors that were identified as contributing to good collaborative relationships included mutual support such as backing up other team members and prioritizing multidisciplinary rounds, and communication including keeping team members informed, asking for their input, physicians explaining their rationale, and practicing joint problem-solving with non-physicians. CONCLUSIONS: Discussion of how physician trainees can best learn to collaborate as members of an interprofessional team must include non-physician perspectives. Training designed to provide medical students and residents with a better understanding of non-physician roles and to enhance mutual support and communication skills may be critical in achieving the AAMC's goals of making physicians effective members of interprofessional teams, and thus improving patient-centered care. We hope that medical educators will include these areas identified as important by non-physicians in targeted team training for their learners.

An Efficient Deep Neural Network to Classify Large 3D Images With Small Objects.

3D imaging enables accurate diagnosis by providing spatial information about organ anatomy. However, using 3D images to train AI models is computationally challenging because they consist of 10x or 100x more pixels than their 2D counterparts. To be trained with high-resolution 3D images, convolutional neural networks resort to downsampling them or projecting them to 2D. We propose an effective alternative, a neural network that enables efficient classification of full-resolution 3D medical images. Compared to off-the-shelf convolutional neural networks, our network, 3D Globally-Aware Multiple Instance Classifier (3D-GMIC), uses 77.98%-90.05% less GPU memory and 91.23%-96.02% less computation. While it is trained only with image-level labels, without segmentation labels, it explains its predictions by providing pixel-level saliency maps. On a dataset collected at NYU Langone Health, including 85,526 patients with full-field 2D mammography (FFDM), synthetic 2D mammography, and 3D mammography, 3D-GMIC achieves an AUC of 0.831 (95% CI: 0.769-0.887) in classifying breasts with malignant findings using 3D mammography. This is comparable to the performance of GMIC on FFDM (0.816, 95% CI: 0.737-0.878) and synthetic 2D (0.826, 95% CI: 0.754-0.884), which demonstrates that 3D-GMIC successfully classified large 3D images despite focusing computation on a smaller percentage of its input compared to GMIC. Therefore, 3D-GMIC identifies and utilizes extremely small regions of interest from 3D images consisting of hundreds of millions of pixels, dramatically reducing associated computational challenges. 3D-GMIC generalizes well to BCS-DBT, an external dataset from Duke University Hospital, achieving an AUC of 0.848 (95% CI: 0.798-0.896).

Incorporation of a Social Virtual Reality Platform into the Residency Recruitment Season.

RATIONALE AND OBJECTIVES: The Covid-19 pandemic ushered a sudden need for residency programs to develop innovative socially distant and remote approaches to effectively promote their program. Here we describe our experience using the social virtual reality (VR) platform Mozilla Hubs for the pre-interview social during the 2020-2021 radiology residency virtual recruitment season, provide results of a survey sent to assess applicants' attitudes towards the VR pre-interview social, and outline additional use-cases for the emerging technology. MATERIALS AND METHODS: A VR Meeting Hall dedicated to the pre-interview social was designed in Mozilla Hubs. To assess applicants' impressions of the Mozilla Hubs pre-interview social, applicants were sent an optional web-based survey. Survey respondents were asked to respond to a series of eleven statements using a five-point Likert scale of perceived agreement: Strongly Agree, Agree, Neutral, Disagree, Strongly Disagree. Statements were designed to gauge applicants' attitudes towards the Mozilla Hubs pre-interview social and its usefulness in helping them learn about the residency program, particularly in comparison with pre-interview socials held on conventional video conferencing software (CVCS). RESULTS: Of the 120 residency applicants invited to the Mozilla Hubs pre-interview social, 111 (93%) attended. Of these, 68 (61%) participated in the anonymous survey. Most applicants reported a better overall experience with Mozilla Hubs compared to CVCS (47/68, 69%), with 10% (7/68) reporting a worse overall experience, and 21% (14/68) neutral. Most applicants reported the Mozilla Hubs pre-interview social allowed them to better assess residency culture than did pre-interview socials using CVCS (41/68, 60%). Seventy-two percent of applicants reported that the Mozilla Hubs pre-interview social positively impacted their decision to strongly consider the residency program (49/68). CONCLUSION: Radiology residency applicants overall preferred a pre-interview social hosted on a social VR platform, Mozilla Hubs, compared to those hosted on CVCS. Applicants reported the use of a social VR platform reflected positively on the residency and positively impacted their decision to strongly consider the program.

Artificial intelligence system reduces false-positive findings in the interpretation of breast ultrasound exams.

Though consistently shown to detect mammographically occult cancers, breast ultrasound has been noted to have high false-positive rates. In this work, we present an AI system that achieves radiologist-level accuracy in identifying breast cancer in ultrasound images. Developed on 288,767 exams, consisting of 5,442,907 B-mode and Color Doppler images, the AI achieves an area under the receiver operating characteristic curve (AUROC) of 0.976 on a test set consisting of 44,755 exams. In a retrospective reader study, the AI achieves a higher AUROC than the average of ten board-certified breast radiologists (AUROC: 0.962 AI, 0.924 ± 0.02 radiologists). With the help of the AI, radiologists decrease their false positive rates by 37.3% and reduce requested biopsies by 27.8%, while maintaining the same level of sensitivity. This highlights the potential of AI in improving the accuracy, consistency, and efficiency of breast ultrasound diagnosis.

An Unusual Case of Invasive Kaposi's Sarcoma with Primary Effusion Lymphoma in an HIV Positive Patient: Case Report and Literature Review.

We report a case of AIDS-related Kaposi's sarcoma (KS) with Primary Effusion Lymphoma (PEL) in a 28-year-old, African American male. Kaposi's sarcoma is an AIDS defining disease and typically will disseminate early in the course of the disease affecting the skin, mucous membranes, gastrointestinal tract, lymph nodes, and lungs. This case reports an unusual presentation of the disease along with primary effusion lymphoma. Although the most common organ systems affected by KS are the respiratory and the gastrointestinal systems, the lungs of this patient did not show any evidence of KS. Additionally, the patient demonstrates the rarely seen liver and unique pancreatic involvement by KS along with unusual synchronous bilateral pleural and peritoneal cavity involvement by PEL, adding to the distinct pattern of invasive AIDS-related Kaposi's sarcoma.