Faculty Bibliography

IMPORTANCE/UNASSIGNED:Hospital course (HC) summarization represents an increasingly onerous discharge summary component for physicians. Literature supports large language models (LLMs) for HC summarization, but whether physicians can effectively partner with electronic health record-embedded LLMs to draft HCs is unknown. OBJECTIVES/UNASSIGNED:To compare the editing effort required by time-constrained resident physicians to improve LLM- vs physician-generated HCs toward a novel 4Cs (complete, concise, cohesive, and confabulation-free) HC. DESIGN, SETTING, AND PARTICIPANTS/UNASSIGNED:Quality improvement study using a convenience sample of 10 internal medicine resident editors, 8 hospitalist evaluators, and randomly selected general medicine admissions in December 2023 lasting 4 to 8 days at New York University Langone Health. EXPOSURES/UNASSIGNED:Residents and hospitalists reviewed randomly assigned patient medical records for 10 minutes. Residents blinded to author type who edited each HC pair (physician and LLM) for quality in 3 minutes, followed by comparative ratings by attending hospitalists. MAIN OUTCOMES AND MEASURES/UNASSIGNED:Editing effort was quantified by analyzing the edits that occurred on the HC pairs after controlling for length (percentage edited) and the degree to which the original HCs' meaning was altered (semantic change). Hospitalists compared edited HC pairs with A/B testing on the 4Cs (5-point Likert scales converted to 10-point bidirectional scales). RESULTS/UNASSIGNED:Among 100 admissions, compared with physician HCs, residents edited a smaller percentage of LLM HCs (LLM mean [SD], 31.5% [16.6%] vs physicians, 44.8% [20.0%]; P < .001). Additionally, LLM HCs required less semantic change (LLM mean [SD], 2.4% [1.6%] vs physicians, 4.9% [3.5%]; P < .001). Attending physicians deemed LLM HCs to be more complete (mean [SD] difference LLM vs physicians on 10-point bidirectional scale, 3.00 [5.28]; P < .001), similarly concise (mean [SD], -1.02 [6.08]; P = .20), and cohesive (mean [SD], 0.70 [6.14]; P = .60), but with more confabulations (mean [SD], -0.98 [3.53]; P = .002). The composite scores were similar (mean [SD] difference LLM vs physician on 40-point bidirectional scale, 1.70 [14.24]; P = .46). CONCLUSIONS AND RELEVANCE/UNASSIGNED:Electronic health record-embedded LLM HCs required less editing than physician-generated HCs to approach a quality standard, resulting in HCs that were comparably or more complete, concise, and cohesive, but contained more confabulations. Despite the potential influence of artificial time constraints, this study supports the feasibility of a physician-LLM partnership for writing HCs and provides a basis for monitoring LLM HCs in clinical practice.

BACKGROUND:Traditional introductory point-of-care ultrasound (POCUS) courses are resource intensive, typically requiring 2-3 days at a remote site, consisting of lectures and hands-on components. Social distancing requirements resulting from the COVID-19 pandemic led us to create a novel hybrid course curriculum consisting of virtual and in-person components. METHODS:Faculty, chief residents, fellows and advanced practice providers (APPs) in the Department of Medicine were invited to participate in the hybrid curriculum. The course structure included 4 modules of recorded lectures, quizzes, online image interpretation sessions, online case discussions, and hands-on sessions at the bedside of course participant's patients. The components of the course were delivered over approximately 8 months. Those participants who completed a minimum of 3 modules over the year were invited for final assessments. Results from the hybrid curriculum cohort were compared to the year-end data from a prior traditional in-person cohort. RESULTS:Participant knowledge scores were not different between traditional (n = 19) and hybrid (n = 24) groups (81% and 84%, respectively, P = 0.9). There was no change in POCUS skills as measured by the hands-on test from both groups at end-of-course (76% and 76%, respectively, P = 0.93). Confidence ratings were similar across groups from 2.73 traditional to 3.0 hybrid (out of possible 4, P = 0.46). Participants rated the course highly, with an average overall rating of 4.6 out 5. CONCLUSIONS:A hybrid virtual and in-person POCUS course was highly rated and as successful as a traditional course in improving learner knowledge, hands-on skill and confidence at 8 months after course initiation. These results support expanding virtual elements of POCUS educational curricula.

Needs and Objectives: The U.S. surgical population is becoming increasingly medically complex, increasing the risk of post-operative complications. Surgeons have traditionally consulted internists and sub-specialists for medical management during inpatient admissions, but this has failed to decrease complications or healthcare costs. To address this issue hospitalists have taken on the role of co-managing patients admitted to surgical services. However, internal medicine residencies don't adequately prepare trainees for this role. Safe and effective modern medical care requires training in medical co-management (MCM), yet there exists no robust theory-and evidence-based curricula to teach these competencies. We designed a curriculum to fill this need. Setting and Participants: Thirteen hospitalists with 0-7 years' clinical experience in an academic medical center in New York, NY. Hospitalists spend 50% of their clinical time on an MCM service covering general surgery, vascular surgery, neurosurgery, general neurology/epilepsy, and orthopedic surgery. Description: We developed a yearlong curriculum based on the Society for Hospital Medicine's guidelines on creating MCM teams and other published frameworks. We applied evidence-based learning theories including: Adult Learning Theory, Cognitivism, Constructivism, and Ericsson's Theory of Expertise. We developed a conceptual framework incorporating key stakeholders in MCM (medical attending, surgical attending, PCP, mid-levels, and patients/families) and 6 core content topics (Roles & Responsibilities, Communication Strategies, Trust & Respect, Common Complications, Transitions of Care, and Deliberate Practice). The curriculum includes two parts: a week-long intensive orientation with Objective Structured Clinical Examinations (OSCEs) and workshops in the main content areas, and a series of monthly continuing professional development (CPD) sessions to facilitate deliberate practice and improve skills needed to manage niche patients on surgical services. These CPD skills are based upon needs identified by the hospitalists and surgical teams. Evaluation: Effectiveness will be measured at the program and hospitalist level. The program will be assessed by patient quality metrics (eg, LOS and readmission rates) and satisfaction by the surgical teams yearly. We will evaluate the impact on hospitalists using retrospective pre-post surveys to measure changes in clinical knowledge, confidence and engagement. Assessments will be collected using Qualtrics survey software. Discussion/Reflection/Lessons Learned: In the early stages of this curriculum, 79% of participants found the CPD sessions "very or extremely helpful." Participants have shown increasing engagement in the curriculum as evidenced by proposals of future session topics and attendance. Data collection is ongoing

Learning Objective #1: Discuss typical, atypical clinical & laboratory presentations of acute mononucleosis Learning Objective #2: Review indications for treatment beyond supportive care CASE: A 33 year-old female with 1 prior spontaneous abortion presented with facial swelling, arthralgias and fevers to 102F for 2 weeks. The patient reported environmental exposures as a native of Australia such as outbreaks of Ross River Fever, encounters with flying foxes, a layover in Hong Kong, and a child with a febrile illness. She sought outpatient care and initial bloods revealed WBC 2.52K/uL, platelets 105K/uL and AST/ALT 88/70 ALP 117 U/L. She was prescribed antibiotics without effect. A week later, she developed pleuritic chest pain, was found to have a small pericardial effusion on outpatient echocardiogram and referred to the Emergency Department. Triage vitals were normal. Her exam was notable for periorbital edema, cervical lymphadenopathy and pain in bilateral wrists without effusion. Labs showed recovering blood counts (WBC 9.3K/uL Hgb 13.1g/dL Plt 157 K/uL), unremarkable UA, microalbumin/creatinine ratio and CK, but worsening LFTs with AST 505 ALT 578 ALP 788 U/L. She was admitted for further work-up with rheumatology and infectious disease input. While hospitalized, the patient developed new night sweats and a sore throat. Repeat echocar-diogram revealed a trace pericardial effusion and abdominal ultrasound was normal. C3/C4 levels, beta-2 glycoprotein, cardiolipin, Ro/La, lupus anticoagulant, histone, centromere, and mitochondrial antibodies all returned negative, as did respiratory viral panel, HIV, hepatitis serologies, thick and thin smears, Lyme, Anaplasma and Babesia serologies, Ross River fever, Dengue and Chikungunya. ANA titer and dsDNA Ab were < 1:40 and 9 respectively, but EBV viral capsid IgM returned positive consistent with acute mononucleosis. IMPACT: The triad of acute EBV is well known, but atypical presentations provide diagnostic challenges and warrant further evaluation DISCUSSION: The triad of acute Epstein-Barr viral infection involves high fevers, lymphadenopathy, and pharyngitis, all present in our patient. However, she also exhibited less typical disease features. Her periorbital edema, known as "Hoagland's sign," is caused by viral replication obstructing lymphatic drainage of the nasopharynx. Similarly, infected tonsillar B-cells instigate secretion of polyclonal antibodies (including heterophile and non-specific autoantibodies). This process is normally accompanied by leukocytosis with atypical lymphocytes, but our case presented initially with leukopenia. Further, her degree of transaminitis (levels > 10x) normal is usually restricted to the immunocompromised. Finally, EBV may also cause transient myo-or pericarditis as noted in select case reports. The mainstay of treatment is supportive, although steroids and acyclovir are used in cases of laryngeal edema, liver failure, or hemolytic anemia and thrombocytopenia. These agents have not been proven to reduce the length or severity of illness

LEARNING OBJECTIVE #1: Recognize clinical features of chronic eosinophilic pneumonia LEARNING OBJECTIVE #2: Diagnose eosinophilic pneumonia when imaging is atypical CASE: A 42-year-old male with history of hypertension and asthma presented with intermittent cough and progressive dyspnea over eight months. His symptoms were more pronounced in the two months leading to admission, during which time he had outpatient treatment with 5 courses of simultaneous oral antibiotics and steroids. Outpatient CT scan of the chest during this time showed diffuse bilateral ground glass opacities, interpreted as atypical infection and inflammatory changes. His symptoms temporarily improved with therapy; however, he was ultimately admitted due to progression of symptoms. On admission, he reported dyspnea at rest, cough productive of yellow sputum, and 20 lb unintentional weight loss over six months. He denied fevers, recent travel, or smoking. Initial vital signs and exam were normal. Labs were notable forWBC 14,000 with 5% eosinophils (750/muL) and elevated ESR and CRP. A repeat CT of the chest showed airspace consolidations primarily in a central and peribronchovascular distribution, with differential diagnosis including infection, organizing pneumonia, vasculitis, chronic eosinophilic pneumonia (CEP), and neoplasm. Blood cultures, HIV, 1,3 beta-d-glucan, galactomannan, ANA, p-ANCA, and c-ANCA were unremarkable. For tissue diagnosis, patient underwent video-assisted thoracoscopy with wedge resection, complicated by an apical pneumothorax requiring chest tube placement. Pathology showed numerous eosinophils in alveolar airspaces, consistent with CEP. Patient was started on high dose steroids with clinical improvement. He was discharged home with a chest tube and continued steroid treatment. IMPACT: In future practice, diagnoses other than infection should be considered earlier in patients with a history of atopy who fail multiple courses of outpatient antibiotics for presumed pneumonia. Furthermore, CEP can be diagnosed with elevated eosinophil count in broncho-alveolar lavage (BAL) fluid. For a patient with peripheral eosinophilia and symptoms consistent with CEP, BAL is the less invasive and more appropriate first diagnostic test over open lung biopsy. DISCUSSION: This patient's chronic dyspnea and cough, weight loss, lack of improvement with antibiotics, and unrevealing infectious and rheumatologic workup made CEP and cryptogenic organizing penumonia (COP) leading differential diagnoses. The classic radiographic appearance of eosinophilic pneumonia is peripheral upper-lobe ground glass infiltrates, which is seen in approximately two thirds of patients with this disease. About three quarters of patients have peripheral eosinophilia. While imaging of COP may appear similar to CEP, peripheral eosinophilia is not typically present, and open lung biopsy is required tomake the diagnosis. This patient's classic symptoms along with peripheral eosinophilia pointed toward a diagnosis of CEP even in the absence of typical radiographic findings

A 49-year-old woman with cholangiocarcinoma metastatic to the lungs presented with new-onset unrelenting headaches. A lumbar puncture revealed malignant cells consistent with leptomeningeal metastasis from her cholangiocarcinoma. Magnetic resonance imaging (MRI) of the brain revealed leptomeningeal enhancement. An intrathecal (IT) catheter was placed and IT chemotherapy was initiated with methotrexate. Her case is notable for the rarity of cholangiocarcinoma spread to the leptomeninges, the use of IT chemotherapy with cytologic and potentially symptomatic response, and a possible survival benefit in comparison to previously reported cases of leptomeningeal carcinomatosis secondary to cholangiocarcinoma.

Rats and mice are considered resistant and susceptible hosts, respectively, for experimental cryptococcosis. For both species, alveolar macrophages (AM) are central components of the host response to pulmonary Cryptococcus neoformans infection. We explored the role of AM in three strains of mice and three strains of rats during cryptococcal infection by comparing the outcome of infection after macrophage depletion using liposomal clodronate. AM depletion was associated with enhancement and amelioration of disease in rats and mice, respectively, as measured by lung fungal burden. The apparent protective role for AM in rats correlated with enhanced anti-cryptococcal activity as measured by phagocytic activity, oxidative burst, lysozyme secretion, and ability to limit intracellular growth of C. neoformans. Furthermore, rat AM were more resistant to lysis in association with intracellular infection. In summary, differences in AM function in rats and mice suggest an explanation for the species differences in susceptibility to C. neoformans based on the inherent efficacy of a central effector cell of the innate immune system.

The production of melanin pigments is associated with virulence for many microbes. Melanin is believed to contribute to microbial virulence by protecting microbial cells from oxidative attack during infection. However, there is also evidence from various systems that melanins have immunomodulatory properties, which conceivably could contribute to virulence by altering immune responses. To investigate the effect of melanin on the immune response, we compared the murine pulmonary responses to infection with melanized and nonmelanized Cryptococcus neoformans cells. Infection with melanized cells resulted in a greater fungal burden during the early stages of infection and was associated with higher levels of interleukin-4 and MCP-1 and greater numbers of infiltrating leukocytes. Infection with laccase-positive (melanotic) C. neoformans cells also elicited higher MCP-1 levels and more infiltrating leukocytes than did infection with laccase-negative cells. Melanization interfered with phagocytosis in vivo for encapsulated C. neoformans but not for nonencapsulated cells. The results provide strong evidence that cryptococcal melanization can influence the immune response to infection and suggest that immunomodulation is an additional mechanism by which the pigment contributes to virulence.