United States

Background
More than 90% of patient facing health literature is written at a readability level above the recommended sixth grade cutoff, resulting in patient confusion, worse health outcomes, and preventable hospital readmissions. Our aim was to leverage natural language processing (NLP) and artificial intelligence (AI) to improve the readability of hospital discharge instructions, which previous studies show can reduce readmissions by up to 50%.

Methods
We analyzed discharge instructions for adults hospitalized for heart failure, determined by discharge summary designation, at the Hershey Medical Center from 2016 to 2021 (n=18,615). Discharge instructions were collected from medical records. Readability scores were calculated using the SMOG index. Discharge instructions were manually rewritten via a standardized process to improve readability, such as eliminating long sentences and exchanging complex terms with lay terms. Afterward, readability scores were re-calculated. The original and rewritten instructions were then processed via AI-based machine learning.

Results
From 80 discharge instructions, the average SMOG reading level was 10.6 ± 0.2 for the original discharge instructions, 9.4 ± 0.2 for rewritten instructions, and 9.3 ± 0.2 for AI-simplified instructions. In 26% of instances, the algorithm made appropriate improvements to the readability level (improved by 1.3 ± 0.3 grade levels).

Conclusion
Our NLP system learns meaningful phrase-level simplifications from human-simplified discharge instructions.

AMA Research Challenge 2022

Artificial Intelligence Tool to Improve the Readability of Discharge Instructions for Heart Failure

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

ACL 2024

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

poster

Welcome to the largest national, multi-specialty research event **featuring research from 750 medical students, residents and fellows, and international medical graduates.**


#### Event Details
The 2022 American Medical Association Research Challenge includes a poster symposium hall and a semifinals hall with the top scored research. Log in at any time throughout the event to:
* **View posters and presentations** on a variety of topics and specialties
* **Score posters in the semifinals** to help decide the top five finalists (AMA member exclusive opportunity)
* **Connect with participants** and provide feedback to help advance their research

At the conclusion of this event, the AMA will announce the top five finalists who will advance to the AMA Research Challenge finals, present their research to an elite panel of judges and compete to win a $10,000 grand prize sponsored by Laurel Road.
 

#### Attend Live Forums
**Oct. 20, 6 — 7 p.m. Central**
 **How to plan a career in academic medicine** 
 Learn how to plan a career in academic medicine and get your questions answered. [Access Here](https://underline.io/events/363/lounge)
 
 **Speakers**
![](https://assets.underline.io/uploads/markdown_image/1/image/d5adb133dfc58eb0a6d469e9cedcd5d4.jpg)
Ricardo Correa, MD, EdD
![](https://assets.underline.io/uploads/markdown_image/1/image/5f9437ab87861538bf02b9ca7c0d926e.jpg)
Stuart M. Greenstein, MD
![](https://assets.underline.io/uploads/markdown_image/1/image/05b1b637c99286d558c371f8e8b185ce.jpg)
Stephanie Strohbeen, MD
 
 **Oct. 22, 11 a.m. — Noon Central**
 **How to get published**
 Learn tips to get your research published and get your questions answered. [Access Here](https://underline.io/events/363/lounge)
 
 **Speakers**
![](https://assets.underline.io/uploads/markdown_image/1/image/8ce6235e957d31cb5e6f9097cc6f6972.jpg)
Jessica Li, MD/PhD Candidate
![](https://assets.underline.io/uploads/markdown_image/1/image/9950ef2e0237f54cb3d8280334d42491.jpg)
Charles Lopresto, DO

### Sponsor
AMA would like to thank Laurel Road for sponsoring a $10,000 grand prize for the winner of the AMA Research Challenge. Laurel Road is not involved in the judging or scoring process to select the winner. The winner will be announced on Dec. 7 during the AMA Research Challenge finals. Payment of the grand prize will be directly from Laurel Road. See the AMA Research Challenge Official Rules for terms and conditions.

![](https://assets.underline.io/uploads/markdown_image/1/image/611e331d4a0eba0920f240cfc04b4d77.png)

**There is no registration fee for this event! **

Register to explore research in a variety of topics. AMA members get exclusive access to score posters in the semifinals.

**Please use the link below to register for free if you have not already done so:**

https://www.ama-assn.org/about/events/ama-research-challenge

Registration Is Required

Welcome to the largest national, multi-specialty research event featuring research from 750 medical students, residents and fellows, and international medical graduates.


Out-of-hospital sudden cardiac arrest (OHCA) kills over 350,000 Americans annually with a survival rate of only 9%. OHCA outcomes improve with bystander CPR and early use of an AED, however, less than one-fourth of the population feels confident in performing CPR, and only one-third feels confident after CPR education. We conducted a survey study aimed at better understand bystander perceptions of CPR and potential barriers towards performing CPR. 24-question surveys were distributed at community CPR classes in May 2022 prior to participant training. Surveys were given in both English and Nepali to explore differences among the refugee population. In addition to basic demographics, participants were surveyed about their knowledge, perceptions, and barriers towards bystander CPR. 86 participated with ages ranging from 13-82 (mean 44.6) with 55 (64%) identifying as female. 50 (66.6%) completed at least 2 years of college. Nepalese respondents were significantly less likely to call 911 or performing CPR due to law enforcement concerns, immigration status and language barrier. English respondents were more concerned with "doing something wrong." Fears over calling 911 were significantly less common than performing bystander CPR (50.5% vs 87.8%). English respondents placed significantly more importance on a patient's or families wishes when deciding when to initiate CPR, while Nepalese were more likely to consider a person's occupation or role in society. Identifying barriers provides avenues for increased public awareness and more inclusive discussion during CPR training events. 

Barriers to Bystander CPR and AED Intervention

Background: Hospitals across the country have worked to reduce readmission rates, but the efficacy of such interventions is challenging to interpret due to the disjointed nature of programs as well as heterogeneous patient populations and healthcare systems. This project aims to describe the current landscape of data on transitions of care and assess for opportunities to optimize such transitions in a large, multi-hospital health system.

Methods: To understand current best practices, we conducted two interviews of large hospital systems and a PubMed literature review of transitions of care from the inpatient setting to the home, with a focus on medically complex and vulnerable populations. Of the 320 articles reviewed, 93 were selected for analysis. This was followed by 17 internal stakeholder interviews at five Johns Hopkins Medicine affiliated hospitals. The results of the literature review and external interviews were compared to programs at Johns Hopkins affiliated hospitals and synthesized into recommendations as a process map that delineated a risk-based discharge algorithm.

Results: Our literature review revealed six major interventions: 1) medication reconciliation; 2) phone calls; 3) home visits; 4) care navigation; 5) post-discharge clinics; and 6) telemedicine. Notably, single-center studies demonstrated that the post-discharge clinic showed efficacy in reducing 30- and 90-day hospital readmissions and healthcare utilization. However, there was mixed evidence for all other interventions. Stakeholder interviews revealed differing approaches to these interventions. Within the five Johns Hopkins Medicine affiliated hospitals, all offered medication reconciliation, phone calls, home visits, and care navigation. All but one offered telemedicine and only two hospitals had post discharge clinics. Lastly, the structuring and management of each of these services varied across the five hospitals.

Conclusion: Transitions of care programs are heterogeneous, define interventions differently, address distinct patient needs, and lack clear options for scaling up. Even within the Johns Hopkins affiliated hospitals, there were varying programs designed to achieve the same goals. Thus, we need a standardization of programs and research methods to better understand transitions of care and allow for the application of these programs to diverse healthcare systems and contexts.

Best Practices: A Health System’s Experience with Inpatient to Outpatient Transitions of Care

Objective
Food insecurity (FI) is a prevalent health-related social risk that can be exacerbated by pediatric hospitalization. Community-based food supports can mitigate FI; however, little is known about rates of food support knowledge, use, and need among caregivers of hospitalized children. We quantified patterns of food support knowledge, use, and need among parents of hospitalized children and analyzed the relationship between food support use and need and food security. We hypothesized that knowledge and use of food supports would be positively associated with food security.

Methods
We performed a cross-sectional analysis of parents (N = 639) of children admitted to an urban academic hospital from 2020-2022. We described self-reported food support knowledge, use, and need by food security status and examined for bivariate associations. We then analyzed the odds of being food secure by use and need of food supports.

Results
Among marginally food secure/food insecure participants (53%), 27% and 46% reported no knowledge and use of food supports, respectively, while 55% reported needing food support. Food support knowledge was positively associated with food security (P = 0.02). Participants reporting use of food supports were more likely to be food secure than those who reported no use of food supports but indicated a need for them (aOR: 1.99, 95% CI: 1.05-3.76). 

Conclusion
Parents of hospitalized children exhibit a high prevalence of FI and food support need but limited knowledge and use of food supports. A hospital-based strategy that enhances knowledge and use of community-based food supports may help mitigate FI.


Do Parents of Hospitalized Children Experiencing Food Insecurity Know About and Use Community-Based Food Supports?

Background
Roughly 1 in 8 people in South Carolina lacks health insurance. Those without insurance are at a higher risk of skin cancers due to lack of preventative screening. Student-run clinics play an important role in bridging the gap and providing preventative care to those who otherwise may not have access. The CARES Medical Clinic is a student-run free clinic for uninsured patients affiliated with the Medical University of South Carolina. During monthly dermatology clinic staffed by university faculty, patients receive comprehensive dermatologic care, including procedures such as cryotherapy and biopsies. The objective of this study is to characterize the financial value of dermatology clinics for uninsured patients seeking care from a student-run free clinic.
Methods
This is a single-site retrospective chart review of patients who attended Dermatology Night from December 2019 -May 2022. A cost analysis was performed using Medicare reimbursement payments to quantify the value of the care provided.
Results
Patients seen by a dermatologist were identified for chart review. Patient ages ranged from 19 to 73 years, with median age 43. 54.7% identified as female. A procedure was indicated for 27 patients; of those, 20 received the appropriate procedure. Overall costs saved for patients were $7807.82, or $433.77 per night
Conclusion
Access and affordability of dermatologic care are limited and vary across health systems. It is our hope that the results demonstrate the value of dermatologic care for underserved patients in community clinics.

Economic Value of Dermatology Clinic at a Student-Run Free Medical Clinic

Background
We surmised the COVID-19 pandemic and associated lockdown measures may have had a gender-based disparate impact on researchers due to “second-shift” responsibilities. Consequently, women may have been temporarily distracted from publishing first-author research, which proxies for academic productivity.

Methods
We selected high-impact journals (9 Pain Management and 5 Internal Medicine) for review, retrieved the summer issues from 2018 and 2019 (pre-COVID) and 2020 and 2021 (postCOVID), divided the articles into Original and Non-Original Research, separated the Pain journals into Interventional and Non-interventional and created a database listing the first authors of each publication. Two reviewers independently identified first-author gender based on personal and professional websites, social media, and authors’ institutions’ published information. A third reviewer settled inconsistencies.

Results
The total number of publications (Retrieved=1105; Analyzed=1064) in high-impact
interventional pain management journals increased year-over-year between 2018 and 2021. Between 2019 and 2020 in particular, the sum of male first-author publications increased 25.68% and the sum of female first-author publications decreased by 25.00%. This disparity was statistically significant (p=0.044). In 2020, the first year of the COVID pandemic, compared to the other years, a gender disparity occurred in the publication of unblinded non-original research (p=0.046). First-author gender did not impact blinded original research over the same time period. 

Conclusion
If the number of publications correlated with the number of women pain researchers, the differences should have remained relatively stable. However, the divergence between male and female first-authors became significant in summer 2019, prior to COVID. In 2018, perhaps in response to record-high fentanyl overdoses, the Drug Enforcement Agency (DEA) investigated more clinicians than ever before. Medicare revamped its audit process. Federal datasets suggest
both agencies target junior physicians with fewer resources to defend against recoupments. These providers are more likely to be women. The current process for DEA and Medicare investigations may have a deleterious and disparate impact on women in terms of clinical research and academic promotions as well as a negative impact on patient care, research advances in pain management, and the opioid epidemic overall. Our research suggests women face both implicit and overt biases when measuring academic productivity. Further research is needed.

Gender Disparities in High Impact Pain Management Publications Across the COVID Pandemic

Background: The COVID-19 pandemic significantly changed the undergraduate medical education (UME) NRMP application process (NAP). Changes included virtual interviews, the suspension or elimination of away rotations, elimination of COMLEX 2 PE/USMLE STEP 2 CS licensing exams, and increased online rotations. The purpose of this study was to survey medical students from two colleges of osteopathic medicine regarding the impact the pandemic had on their 2021-2022 application process in order to better understand challenges faced by students.

Methods: An IRB-approved 20 question survey was developed and distributed after the 2022 match via email and social media. Likert-scale questions with a range of 1-5 and a comment box were used to assess student experiences related to changes the pandemic had on their match process. Several areas were assessed, e.g., away rotations, virtual interviews, board requirement changes, advising, psychological well-being, etc. 

Results: 83 students from the Liberty University College of Osteopathic Medicine and the Kansas City University College of Osteopathic Medicine participated in the study. 94.81% matched in their first-choice specialty. The number of away rotations in matched specialties ranged from 1 to 9 (M=2.36, SD=0.91). Only 39.74% matched at an audition rotation site. 50.77% of students reported the suspension of the third board exam impacted their application process very highly. Approximately one-third of students rated both stress and anxiety as high impacts. 72.05% of participants sought feedback from students from the 2020-2021 NAP and 48.98% of participants found this feedback to be highly helpful. 

Conclusion: Changes in the 2021-2022 NAP due to COVID-19 positively and negatively impacted students. Student comments indicated that virtual interviews opened up more opportunities to interview and save money. Students also commented that due to the strain of the pandemic, they had less time with their attendings and that this could have negatively impacted their letters of recommendation. Overall, cancellation of the PE and CS was viewed positively. And many students sought out feedback from applicants of the 2020-2021 NAP to be better prepared to face the challenges of the continued pandemic. 


Impact of the COVID-19 Pandemic on Medical Students in the 2021-2022 NRMP Application Process

Abstract Title: Improving ACEs Implementation at Frayser Health Center ​

Authors: Christian Mark Agatep, Mai Nojima, Ashlynn Rae Torres 

Background 

The Adverse Childhood Experiences (ACEs) survey is an important pediatric screening tool that predicts physical and mental health outcomes across an individual’s lifespan1. While the value of ACEs surveys has been recognized within federally qualified health centers (FQHC), little is understood about its implementation in these resource-constrained settings. Through a partnership between a FQHC and medical students enrolled in a Health Systems Science (HSS) Service-Learning course, we analyzed how health systems can impact patient and community health. Our community-engaged aim was to identify common themes about experiences in implementing an ACEs survey at the FQHC and to address these themes to facilitate its successful adoption and sustainability. 

Methods 

We implemented a Plan-Do- Study-Act quality improvement cycle between January 2021 – May 2022 at a FQHC in Southern California. We interviewed and surveyed stakeholders, including FQHC leadership and staff members. Thematic analyses were performed on qualitative data to co-identify successful workflow processes and opportunities for improvement. 



Results 

We interviewed 17 staff members, inquiring about their understanding of ACEs, the importance of survey implementation, and their specific roles in the screening process. Common themes that emerged were that staff a) do not receive consistent ACEs training and b) do not receive training about the rationale for administering ACEs surveys. 



Conclusion 

We found that using a HSS approach helped identify specific workflow steps with patient screening processes, define which staff delivers each step, and understand the interprofessional collaboration for efficient and effective screening. Through this we brought together multiple stakeholders to appreciate the context and implementation of an ACEs survey at a FQHC serving a high needs community. To address the training gaps, we developed a consolidated ACEs training pamphlet in English and Spanish for staff members that provides an overview of ACEs screening and referral resources to complement their clinical knowledge of trauma-informed care. Our next step is to ensure clinic staff receive consistent training so they can screen eligible children at the FQHC in an equitable matter. By doing so, early interventions can be identified to connect patients and their families to appropriate community resources and trauma-informed care.

Improving ACEs Implementation at Frayser Health Center ​

Background: Annual well-child visits (AWCV) are important for evaluating children’s well-being. During AWCV, one’s health, development, and anticipatory guidance are discussed, along with any unmet medical/social needs. Many families have barriers to attending AWCV due to social factors including transportation, housing, and childcare. Addressing socioeconomic barriers may improve access to AWCVs. The CATCH program utilizes tele-outreach to schedule AWCVs for patients lost to follow-up and offers access to community-based resources to address families’ unmet social needs in Pittsburgh, PA. This study assessed the efficacy of the CATCH program in scheduling pediatric AWCV and providing access to community resources.
Methods: This was a retrospective analysis of pediatric patients lost to follow-up (not seen for AWCV for 12 months since last visit), from 1/2021-6/2022. Families were contacted through phone calls. Each call inquired about interest in scheduling AWCV, identified preferred clinics, assessed families’ unmet social needs, and provided access to local resources. Telephone encounters were made in EPIC to notify pediatric offices of scheduled AWCVs. Chart review was performed to determine attendance rates of scheduled appointments. 
Results: A total of 459 families were contacted, of which 266 (58%) were unreachable. Of the 193 families reached, 79 (17%) had already attended AWCV within a year, and 22 (5%) patients had reached adulthood and transitioned care, and were thus excluded from our sample. Among the remaining 92 families, 41 AWCVs (45%) were scheduled. Cited reasons for declining scheduling assistance included residence/pediatrician change, guardianship issues, and personal preference. 19 of 41 scheduled appointments (46%) were attended. Community resources were universally provided, with 32 families connected with over 71 resources. These spanned areas of health care/insurance, COVID-19, food, housing, transportation, legal services, childcare, and general community resource hubs. 
Conclusion: The tele-outreach of the CATCH program increased rates of AWCV scheduling, and connected families with local resources to address unmet social needs. This model offers support for phone call-based outreach to contact families of children lost to follow-up and suggests its role in addressing health care and socioeconomic barriers. CATCH may serve as a model for other tele-outreach programs in underserved communities.

Pediatric Tele-Outreach Program Increases Well Child Visit Scheduling and Addresses Unmet Social Needs 

Background
During the height of the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) pandemic, also known as COVID-19, clinicians warned the public of a significant symptom of COVID-19 called "silent hypoxemia," where patients presented with low oxygen saturation levels and no dyspnea. Pulse oximetry is a critical component of evaluating the severity of arterial deoxygenation. However, studies have shown discrepancies in these readings amongst deep skin tone individuals or individuals with higher levels of active melanocytes. Data from ecent studies have shown that black patients had almost three times the frequency of occult hypoxemia that was not detected by pulse oximetry. It is hypothesized that the importance of pulse oximetry in medical decision-making has some implications that may play a significant role in adverse COVID-19 health outcomes in marginalized communities.

Methods
The design of this study is a systematic review and data extraction of relevant articles that discuss pulse oximetry and racial disparities associated with pulse oximetry technology. Data examining COVID-19 related data that analyzed the rate of infection, hospitalization, and mortality among minorities was also considered.4, 5, 6.

Results
Researchers identified 21,431 citations with relative pulse oximetry accuracy and data related to COVID-19 health outcomes. Compared to their white counterparts, inaccurate pulse oximetry readings in black patients are more than doubled. Some data extractions provided significant evidence that the inaccuracy of these pulse oximetry readings could result in late intubation and or unnecessary intubations, which may have played an essential role in the increased mortality rate for black patients.

Conclusion
Our aim is to show that inequalities in pulse oximetry readings could affect COVID-19 outcomes for African-Americans. Based on the literature, inaccurate pulse oximetry readings could lead to delayed medical treatment of patients in need of further respiratory assistance, thus supporting the hypothesis. This delay is costly as it can increase the frequency of African-Americans needing advanced respiratory intervention such as intubation or ventilators. It is vital that others research topics related to this to prevent additional injuries or deaths.

Racial Disparities among Pulse Oximetry readings and Possible effects on COVID -19 outcomes

LEND is a program designed to address marked disparities in identifying and treating Autism Spectrum Disorder (ASD) in marginalized communities. Studies show that black children are less likely than white children to be diagnosed with ASD. Early and accurate detection of ASD is critical for access to early intervention which optimizes the outcomes. We aimed our project to equip immigrant communities with ASD treatments and available services, to provide a safe platform to explore cultural barriers, and to provide practical solutions to increase access to services.
Methods: It’s a cross-sectional study, a flier distributed through email to parents and colleagues. Three workshops were held, two in English and one in Chinese, which included a zoom presentation and an open discussion at the end of the presentation, pre and post assessment surveys used to assess knowledge. Presentation included basic facts about ASD, assessment, available treatments and resources, and transition from childhood to adulthood.
Results: Seventeen subjects participated in our workshops. In the first English workshop, 7 participated, in the second English workshop, 5 participated, and in the third Chinese workshop, 5 participated, with 100% versus 40% versus 100% respectively response to surveys. The data indicate that the audience became more familiar with diagnosis and treatments based on a 7-question survey.
Conclusions: We observed significant uncertainty in impartial hearings, proper school placements, Medicaid waivers, and OPWDD. Limitations include low attendance and lack of audience participation. Future research should consider event promotions in advance and live presentations in community settings.



Raising Autism Awareness For Disadvantaged Communities as Part of Leadership in Education and Neurodevelopmental and Related Disabilities (LEND) Program

Premium content

Next from AMA Research Challenge 2022

Barriers to Bystander CPR and AED Intervention

Similar lecture

IgnitionInnovators at "Discharge Me!": Chain-of-Thought Instruction Finetuning Large Language Models for Discharge Summaries

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES