Faculty Bibliography

BACKGROUND: Healthcare systems rose to the challenges of COVID-19 by creating or expanding telehealth programs to ensure that patients could access care from home. Traditionally, though, physicians receive limited formal telemedicine training, which made preparedness for this transition uneven. We designed a survey for General Internal Medicine (GIM) physicians within our diverse health system to describe experiences with providing virtual patient care; with the ultimate goal of identifying actionable recommendations for health system leaders and medical educators.
METHOD(S): Surveys were sent to all faculty outpatient GIM physicians working at NYU Langone Health, NYC Health + Hospitals/Bellevue and Gouverneur, and the VA NY Harbor Health System (n=378) in May & June of 2020. Survey items consisted of Likert and open-ended questions on experience with televisits (13 items) and attitudes toward care (24 items). Specific questions covered barriers to communication over remote modalities.
RESULT(S): 195/378 (52%) responded to the survey. 96% of providers reported having problems establishing a connection from the patient's end while 84% reported difficultly establishing connection on the provider's end. Using interpreter services over the phone was also problematic for providers, with 38% reporting troubles. Regarding teamness, 35% of physicians found it difficult to share information with healthcare team members during virtual visits and 42% found it difficult to work collaboratively with team members, both when compared to in-person visits. When subdivided, 24% of private and 40% of public providers found info sharing more difficult (p<0.04). 31% of private providers and 45% of public found team collaboration more difficult (ns). Physicians also identified challenges in several domains including physical exams (97%), establishing relationships with new patients (74%), taking a good history (48%), and educating patients (35%). In thematic analysis of open-ended comments, themes emerged related to technological challenges, new systems issues, and new patient/provider communication experiences. Positives noted by physicians included easier communication with patients who often struggle with keeping in-person appointments, easier remote monitoring, and a more thorough understanding of patients' home lives.
CONCLUSION(S): Provider experience differences were rooted in the type of technology employed. Safety-net physicians conducted mostly telephonic visits while private outpatient physicians utilized video visits, despite both using the same brand of electronic medical record system. As we consider a new normal and prolonged community transmission of COVID-19, it is essential to establish telemedicine training, tools, and protocols that meet the needs of both patients and physicians across diverse settings. LEARNING OBJECTIVE #1: Describe challenges and barriers to effective communication and clinical skill utilization during televisits LEARNING OBJECTIVE #2: Conceptualize recommendations for educational curricula and health service improvement areas

LEARNING OBJECTIVES 1: Describe a longitudinal curriculum to train internal medicine (IM) residents in point-of-care ultrasound (POCUS). LEARNING OBJECTIVES 2: Recognize the impact of decreased patient contact on residents' retention of POCUS skills. SETTINGAND PARTICIPANTS:Despite thewell-documented benefits of POCUS, internal medicine residents receive little formal training. We implemented a curriculumin the 2019 academic year to train 55 PGY-2 IMresidents in POCUS across four urban teaching hospitals and a method to evaluate its efficacy. As the COVID pandemic hit, we additionally sought to understand the impact of COVID on the efficacy of our curriculum and to ascertain from IM residents their barriers to using POCUS during the COVID pandemic. DESCRIPTION: The curriculum was composed of three workshops, consisting of lectures and hands-on practice covering lung, cardiac, abdominal, and lower extremity vascular views. Following the workshops, we sought to consolidate learners' knowledge with a subsequent year-long skill building phase. The skill-building phase was truncated due to the pandemic.A hands-on assessment was performed prior to the course and not repeated at course conclusion due to social distancing concerns. An online knowledge test was administered before the course, immediately following the course, and at one year. A survey assessing attitudes and barriers to POCUS was administered before the course and at one year. EVALUATION: No resident passed the pre-course hands-on assessment. Prior to the course, the average resident score was 54% on the online knowledge quiz; directly after the workshop series, the average rose to 78%. At one year, the average score on the online knowledge quiz was 74%, a statistically significant decrease (p=0.04). Ninety-one percent of residents reported performing POCUS at least once/month prior to the pandemic. During the pandemic, scanning activity decreased; 67% residents reported they scanned rarely or never. DISCUSSION/ REFLECTION / LESSONS LEARNED: Our course led to significant improvement of knowledge regarding ultrasound technology and image interpretation, however this decayed at one year, likely due to lack of skill reinforcement. Though POCUS was widely used prior to the pandemic, usage dropped at the pandemic's peak, despite its utility as both a diagnostic and therapeutic tool. The most commonly cited reason for lack of use was concern regarding contamination and infectious exposure. While the COVID pandemic disrupted our curriculum, it also highlighted opportunities to incorporate POCUS into clinical practice and reinforced the importance of continued longitudinal practice to retain learned skills

BACKGROUND: New York City became the epicenter of the COVID-19 pandemic in the US, reporting its first case of SARS-CoV-2 on March 1, 2020. Patients with co-morbid conditions such as hypertension and diabetes are disproportionately impacted by COVID-19. Hospital systems have been burdened nationwide, including community and safety-net hospitals who serve medically underserved populations, placing them at risk from a resource-needs standpoint. Our study aim is to describe the clinical presentation and outcomes of hospitalized patients with COVID-19, and to highlight the burden on community hospitals, in order to guide health policy and resource allocation in future crises.
METHOD(S): We conducted a retrospective case series of patients admitted to NYU Langone Hospital - Brooklyn between March 13th and April 4th, 2020. Reverse-transcriptase polymerase chain reaction nasopharyngeal swab confirmed infection with the SARS-CoV-2 virus. Clinical demographics were obtained from the electronic health record (Epic Hyperspace, Madison, WI). The primary outcome was time-to-event, defined as transfer to an intensive care unit, mechanical ventilation or death from time of admission. Statistical analysis was performed using Stata SE 16 (StataCorp, College Station, TX).
RESULT(S): There were 561 patients admitted with a median age of 61 years(IQR 48-74). See Table 1. The median time to composite event was 4.13 days(IQR: 2.23-7.97).
CONCLUSION(S): Our results show that the impact of COVID-19 on a community hospital is similar to what has been reported in the literature for tertiary centers, implying that safety-net hospitals can play an integral role in future impact mitigation. These implications hold true as the pandemic continues to disproportionately affect those with chronic diseases. As cases of COVID-19 near 20 million, our experience positions us as harbingers who can provide insight for resource allocation across the US. LEARNING OBJECTIVE #1: Patient Care: Identify the characteristics in patients with COVID-19 associated with increased risk for hospitalization LEARNING OBJECTIVE #2: Medical Knowledge: Understand the outcomes related to COVID-19 in a diverse population

STATEMENT OF PROBLEMOR QUESTION (ONE SENTENCE): To determine the feasibility and acceptability of using a virtual-only model for initiating and maintaining patients on PrEP (pre-exposure prophylaxis) for HIV prevention. LEARNING OBJECTIVES 1: Participants will be able to identify 3 key considerations in developing a clinical workflow for virtual PrEP initiation. LEARNING OBJECTIVES 2: Participants will be able to discuss 3-5 challenges associated with virtual PrEP initiation, and identify strategies to address these challenges. DESCRIPTION OF PROGRAM/INTERVENTION, INCLUDING ORGANIZATIONAL CONTEXT (E.G. INPATIENT VS. OUTPATIENT, PRACTICE OR COMMUNITY CHARACTERISTICS): The Family Health Centers at NYU Langone (FHC) is a federally qualified health center network with 8 clinical sites in Brooklyn, NY, primarily serving a low-income, immigrant community. Since 2016, FHC has operated a focused outreach program to promote PrEP to high-risk individuals, using targeted strategies to engage those not currently in PrEP care. Our intervention sought to expand on our successful outreach model by using tele health to remove geographic barriers to participation. We developed clinical and patient navigation workflows to enable patients to initiate and continue PrEP through virtual visits. For necessary labs, patients were supported in identifying a lab collection site convenient to their home. Patient navigation staff played a key role in risk reduction education, benefits navigation, and facilitating compliance with labs and virtual care. MEASURES OF SUCCESS (DISCUSS QUALITATIVE AND/OR QUANTITATIVEMETRICSWHICHWILL BEUSEDTOEVALUATE PROGRAM/INTERVENTION): The key measure of success is PrEP uptake and continuation among the virtual visits cohort. Additional evaluation measures include the referral source of patients for virtual PrEP initiation, patient demographics, and HIV risk-these measures will enable us to assess whether we are reaching a more diverse or higher risk population through this program. FINDINGS TO DATE (IT IS NOT SUFFICIENT TO STATE FINDINGS WILL BE DISCUSSED): The pilot project launched in October 2020. In the three months since project launch, 8 patients were served through this program. Six of the patients (75%) had been initially engaged with the FHC through the HIV prevention program, while two were existing FHC patients-one of whom had previously been in standard PrEP care, but struggled to make the in-person visits. Six patients were cisgender men who have sex with men, while two were transgender women. Virtual PrEP provided an opportunity to link patients to other needed healthcare services, including vaccination and STI treatment. KEY LESSONS FOR DISSEMINATION (WHAT CAN OTHERS TAKE AWAY FOR IMPLEMENTATION TO THEIR PRACTICE OR COMMUNITY): The tele health PrEP pilot program enabled us to reach a diverse group of high-risk patients, a majority of whom had not previously been engaged in care within our health system, and we anticipate continued growth this program as we expand our outreach to additional geographic areas. Navigation staff were key in overcoming some of the barriers associated with the virtual model by building relationships with the patients and serving as a reliable source of support for patients encountering logistical barriers. PrEP initiation by tele health must account for additional logistical considerations-most notably, ensuring patient compliance with labs-but it is a feasible approach for engaging high-risk patients in HIV prevention services

Purpose: To carry out a large international validation of how dose prediction quality translates to plan quality in a radiotherapy knowledge-based planning (KBP) process.
Method(s): We collected dose predictions for head-and-neck cancer radiotherapy from 21 different research groups internationally who participated in the OpenKBP Grand Challenge. Each research group used the same training dataset (n=200) and validation dataset (n=40) to develop their methods. These methods predicted dose on a testing dataset (n=100), and those 2100 unique dose predictions were input to a previously published plan optimization method to generate 2100 treatment plans. The predictions and plans were compared to the ground truth dose via: (1)error, the mean absolute voxel-by-voxel difference in dose; and (2) quality, the mean and maximum deviation across 23 dose-volume histogram (DVH) criteria.
Result(s): The range in median prediction error among the top 20 methods was 2.3Gy to 12.0Gy, which was 6.8Gy wider than the range in median plan error of 2.1Gy to 5.0Gy. One method also achieved significantly lower prediction error (P<0.05; one-sided Wilcoxon test) than all the other methods, however, it generated plans with error that was not significantly lower than 28.6% of the other methods. Additionally, predicted dose was consistently lower quality than plan dose. Half (n=1050) of all predictions and plans had an average deviation that was 0.1Gy worse and 0.8Gy better than the ground truth dose, respectively. Similarly, half of all predictions had a maximum deviation that was 3.7Gy worse than the ground truth dose, which was 1.0Gy worse than half of all plans.
Conclusion(s): Many dose prediction methods can achieve low error, however, optimization often improves upon the predictions and eliminates significant differences between prediction methods. Thus, it is critical that we improve the optimization stage in KBP to get better utility out of the existing high-quality dose prediction methods

BACKGROUND: Patients who are immigrants, notably those with undocumented status, face challenges to equitable healthcare access. By understanding immigration status as a social determinant of health (SDOH), physicians can begin to address such disparities. However, few undergraduate medical curricula include formal longitudinal instruction addressing immigration. We conducted a needs assessment of a medical school's curricular content in teaching medical students to address immigration as a SDOH.
METHOD(S): MS1-3 students from a school in Bronx, NY where 35% of the patient population are immigrants, received a 13-question email survey via surveymonkey.com. Students were assessed on three primary areas based on a literature review on sanctuary doctoring and SDOH: 1) Knowledge of immigrants' barriers to care (4-point scale, strongly disagree to strongly agree); 2) Confidence in assessing patient immigration status, taking an immigration history, and advocating for patients at risk of deportation (3-point scale, not confident to very confident); and 3) Frequency of assessing patients' immigration status, identifying immigration status when presenting cases, and referring undocumented patients to social/legal resources (4-point scale, never to always). Outcomes were compared between pre-clinical (MS1-2) and clinical (MS3) students.
RESULT(S): Among 539 students, 159 (29.5%) responded, with 104 preclinical and 55 clinical students. 79.2% strongly agreed that undocumented immigration status limits healthcare access. Few students reported being very confident in asking about immigration status (8.8%), taking an immigration history (12.6%), providing legal information (2.5%) and advocating for patients at risk of deportation (6.3%). Compared to the pre-clinical cohort, clinical students were significantly more confident in taking an immigration history (p=0.04) but not in other skills. Few students endorsed frequently or always asking patients about immigration status (3.2%), identifying immigration status when presenting patients (4.5%), and referring undocumented patients to appropriate resources (8.3%). There were no significant differences in frequencies of use of clinical skills pertaining to care of immigrant patients in the pre- and clinical cohorts.
CONCLUSION(S): Students are aware of barriers that immigrant patients face but lack confidence and experience in identifying and supporting undocumented patients. Our results will inform a revision of the longitudinal curriculum, including didactics and practical activities. LEARNING OBJECTIVE #1: 1. Assess students' skills and confidence in identifying and advocating for undocumented immigrant patients in clinical practice LEARNING OBJECTIVE #2: 2. Assess students' knowledge of immigrants' barriers to care

BACKGROUND: Clinical reasoning (CR) is a core component of medical training, yet residents often receive little feedback on their CR documentation. Here we describe the process of developing a machine learning (ML) algorithm for feedback on CR documentation to increase the frequency and quality of feedback in this domain.
METHOD(S): To create this algorithm, note quality first had to be rated by gold standard human rating. We selected the IDEA Assessment Tool-a note rating instrument across four domains (I=Interpretive summary, D=Differential diagnosis, E=Explanation of reasoning, A=Alternative diagnoses explained) that uses a 3-point Likert scale without descriptive anchors. To develop descriptive anchors we conducted an iterative process reviewing notes from the EHR written by medicine residents and validated the Revised-IDEA Assessment Tool using Messick's framework- content validity, response process, relation to other variables, internal structure, and consequences. Using the Hofstee standard setting method, cutoffs for high quality clinical reasoning for the IDEA and DEA scores were set. We then created a dataset of expertrated notes to create the ML algorithm. First, a natural language processing software was applied to the set of notes that enabled recognition and automatic encoding of clinical information as a diagnosis or disease (D's), a sign or symptom (E or A), or semantic qualifier (e.g. most likely). Input variables to the ML algorithm included counts of D's, E/A's, semantic qualifiers, and proximity of semantic qualifiers to disease/ diagnosis. ML output focused on DEA quality and was binarized to low or high quality CR. Finally, 200 notes were randomly selected for human validation review comparing ML output to human rated DEA score.
RESULT(S): The IDEA and DEA scores ranged from 0-10 and 0-6, respectively. IDEA score of >= 6.5 and a DEA score of >= 3 was deemed high quality. 252 notes were rated to create the dataset and 20% were rated by 3 raters with high intraclass correlation 0.84 (95% CI 0.74-0.90). 120 of these notes comprised the testing set for ML model development. The logistic regression model was the best performing model with an AUC 0.87 and a positive predictive value (PPV) of 0.65. 48 (40%) of the notes were high quality. There was substantial interrater reliability between ML output and human rating on the 200 note validation set with a Cohen's Kappa 0.64.
CONCLUSION(S): We have developed a ML algorithm for feedback on CR documentation that we hypothesize will increase the frequency and quality of feedback in this domain. We have subsequently developed a dashboard that will display the output of the ML model. Next steps will be to provide internal medicine residents' feedback on their CR documentation using this dashboard and assess the impact this has on their documentation quality. LEARNING OBJECTIVE #1: Describe the importance of high quality documentation of clinical reasoning. LEARNING OBJECTIVE #2: Identify machine learning as a novel assessment tool for feedback on clinical reasoning documentation