United States

Background
Pack-year history is an invaluable clinical tool for measuring tobacco-associated disease risk and determining eligibility for lung cancer screening with low-dose computed tomography. Patient-reported measures of smoking history have demonstrated strong test-retest reliability across various survey methods, but the variation between research and clinical settings suggests that non-standardized smoking history interview questions in community practice may not always reveal the true extent of a patient’s tobacco exposure. To date, no study has compared patient responses to differently phrased questions assessing the same measure of smoking history. We hypothesize there is a significant discordance between patient-reported number of years smoked and calculated elapsed time between age at initiation and age at cessation (or present age if current smoker).

Methods
This was a two-month prospective survey-based study conducted at the Western Michigan University Homer Stryker M.D. School of Medicine&#39;s Internal Medicine resident clinic in Kalamazoo, Michigan, including patients with a documented smoking history in their electronic medical records. Through verbal interviews, we collected data on number of years smoked, packs per day, and age of smoking initiation/cessation. Primary outcome variables were total years smoked and calculated pack-years. Secondary outcome variables were gender, age, and age of smoking initiation.

Results
There were 124 respondents (36% male, 63% female, 1% non-binary) with an average age of 55 and average smoking history of 1.2 packs per day. Of them, 46% were current and 54% were former smokers. Using a paired sample one-tail t-test, we compared patient-reported years smoked (M=22.06, SD=12.44) to calculated years based on age of initiation/cessation (M=24.03, SD=13.53), which showed a significant mean difference of 1.97 years (95% CI 1.16- 2.78, t=4.05, p&lt;0.001). This discordance translated to a 21.29% increase in pack-years (from 24.32 to 29.50 pack-years) and a mean difference of 5.21 pack-years (95% CI 1.04-9.38, t=2.07, p=0.02).

Conclusion
Our study underscores the importance of critically evaluating the ways in which we gather a patient’s smoking history. When compared to pack-years calculated using age of smoking initiation/cessation, our sample population tended to underestimate their exposure by over 5 pack-years—a clinically significant difference for someone at high risk who might not otherwise receive screening imaging. Often a pivotal moment in adolescence, participants recalled their age of smoking initiation with greater precision compared to their estimations of total exposure. These results call for urgent reevaluation of the tobacco history interview, its notation in electronic medical records, and the methodology of taking a substance abuse history in general.

AMA Research Challenge 2023

Rethinking the Smoking History Interview: A Prospective Survey-based Study

Background
Pack-year history is an invaluable clinical tool for measuring tobacco-associated disease risk and determining eligibility for lung cancer screening with low-dose computed tomography. Patient-reported measures of smoking history have demonstrated strong test-retest reliability across various survey methods, but the variation between research and clinical settings suggests that non-standardized smoking history interview questions in community practice may not always reveal the true extent of a patient’s tobacco exposure. To date, no study has compared patient responses to differently phrased questions assessing the same measure of smoking history. We hypothesize there is a significant discordance between patient-reported number of years smoked and calculated elapsed time between age at initiation and age at cessation (or present age if current smoker).

Methods
This was a two-month prospective survey-based study conducted at the Western Michigan University Homer Stryker M.D. School of Medicine's Internal Medicine resident clinic in Kalamazoo, Michigan, including patients with a documented smoking history in their electronic medical records. Through verbal interviews, we collected data on number of years smoked, packs per day, and age of smoking initiation/cessation. Primary outcome variables were total years smoked and calculated pack-years. Secondary outcome variables were gender, age, and age of smoking initiation.

Results
There were 124 respondents (36% male, 63% female, 1% non-binary) with an average age of 55 and average smoking history of 1.2 packs per day. Of them, 46% were current and 54% were former smokers. Using a paired sample one-tail t-test, we compared patient-reported years smoked (M=22.06, SD=12.44) to calculated years based on age of initiation/cessation (M=24.03, SD=13.53), which showed a significant mean difference of 1.97 years (95% CI 1.16- 2.78, t=4.05, p<0.001). This discordance translated to a 21.29% increase in pack-years (from 24.32 to 29.50 pack-years) and a mean difference of 5.21 pack-years (95% CI 1.04-9.38, t=2.07, p=0.02).

Conclusion
Our study underscores the importance of critically evaluating the ways in which we gather a patient’s smoking history. When compared to pack-years calculated using age of smoking initiation/cessation, our sample population tended to underestimate their exposure by over 5 pack-years—a clinically significant difference for someone at high risk who might not otherwise receive screening imaging. Often a pivotal moment in adolescence, participants recalled their age of smoking initiation with greater precision compared to their estimations of total exposure. These results call for urgent reevaluation of the tobacco history interview, its notation in electronic medical records, and the methodology of taking a substance abuse history in general.

poster

Welcome to the **largest national, multi-specialty research** event featuring research from nearly 700 medical students, residents, fellows, and international medical graduates.

Use the buttons below or navigation menu to access the event.
 


[![](https://assets.underline.io/uploads/markdown_image/1/image/3e33a8c754cc49610b30950617d813e7.png)](https://underline.io/events/434/sessions?searchGroup=all)
 
[![](https://assets.underline.io/uploads/markdown_image/1/image/551404d429c4e274f9a54cbdd7081918.png)](https://underline.io/events/434/posters?searchGroup=all)
 
[![](https://assets.underline.io/uploads/markdown_image/1/image/677d92f566b880cebd36905fc738fc10.png)](https://underline.io/events/434/lounge)
 
[![](https://assets.underline.io/uploads/markdown_image/1/image/c7392833c7261c3874c5b43fa2f4df19.png)](https://underline.io/events/434/help-desk)
 

#### Event Details

The 2023 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
* **Attend education sessions** with physician experts on key research topics

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. 

### Grand Prize 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 February 4, 2024 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 this link [here](https://www.ama-assn.org/about/events/ama-research-challenge) to register for free.**

Registration Is Required

Join the AMA Research Challenge poster symposium and semifinals to explore top research posters, score posters in the semifinals (an AMA member exclusive opportunity) and attend education sessions with experienced physicians to learn tips to advance your research and career. 

Using deep learning to augment radiologist interpretation of brain MRIs for Alzheimer’s disease

Objective: Femoral endarterectomy is a common vascular procedure. Previous data have shown that it is associated with elevated risk of surgical site infection (SSI) that is estimated to be 15% or higher. Obesity is a known risk factor for SSI. Presence of an overhanging pannus may be associated with increased risk of wound infections, due to decreased local hygiene and increased frictional forces. The aim of the current study was to study the relationship between pannus size and SSI after femoral endarterectomy performed at a rural tertiary medical center. 
Methods: Retrospective chart review was used to evaluate femoral endarterectomy performed between January 1, 2010 and December 31, 2019. Patients were excluded if they did not have a CTA or CT abdomen/pelvis within 6 months of endarterectomy. Cases were identified using CPT code, and basic demographic data and preoperative risk factors were identified. Incidence of postoperative wound infection was recorded. Perioperative CT scans were evaluated using TeraRecon to measure the area of the subcutaneous tissue overlying the common femoral artery on the operative leg, hypothesized to be a surrogate for pannus size. 
Results: The total number of participants was 250, with a 17% incidence of SSI. There was a significant difference in pannus size in those who developed postoperative SSI (30.5 cm2 ± 17.2 cm2) compared to those who did not (24.5 cm2 ± 13.7 cm2) (p =0.022). Pannus area was on average 6cm2 larger in those who developed postoperative SSI. BMI, obesity, hypertension, and diabetes were associated with pannus area but were not strong predictors of SSI. BMI and pannus size were strongly correlated, although BMI was more variable as pannus size increased. A 5cm2 increase in pannus size was associated with an absolute relative risk of SSI of 1.10. In patients with pannus size ≤26cm2, 12% had an SSI; in patients with pannus >26cm2, 24% had an SSI; the relative risk of SSI for pannus >26cm2 was 1.99. 
Conclusions: This study highlights a novel method to measure pannus size, which is a stronger predictor than BMI alone for postoperative SSI after femoral endarterectomy. This can help guide preoperative risk identification and decision making, and assign more aggressive postoperative infection prevention measures to patients at risk of SSI.

Relation of pannus size to postoperative wound infection after femoral endarterectomy

Background
Physician burnout, characterized by emotional exhaustion and a diminished sense of personal accomplishment, affects nearly 25% to 60% of primary care physicians (PCPs). With the ever- increasing burden of non-clinical tasks that compete with direct patient care, burnout rates are only projected to increase. Machine learning (ML), a subset of artificial intelligence where computers learn and make decisions from data, has been previously applied to healthcare to optimize processes and enhance decision-making. We hypothesized that ML-based natural language processing (NLP) could streamline non-clinical tasks by automating patient assessment form (PAF) completion for outpatient PCPs.

Methods
We developed and internally validated a ML-enabled workflow that combined note extraction from the EMR, protected health information de-identification, and NLP and federated learning to automate PAF completion. PAFs were auto-populated with ICD-10 codes matched directly to patient clinic notes. We externally validated our workflow in a real-world setting by applying it to a primary care clinic serving patients who are Medicare Advantage beneficiaries, following their annual wellness visits. PCPs then reviewed and approved completed PAFs with one click, triggering the workflow to auto-generate screening orders for at-risk patients.

Results
Our ML-enabled workflow was trained on data from 6,000 synthetic notes and achieved a ten-fold cross-validation accuracy of 96%. We further externally validated our workflow in a real-world, suburban outpatient clinic. Over five months, PAFs were completed for 179 patients, requiring only six hours of provider review time with a 99.4% approval rate. Compared to the previous year, the ML workflow decreased manual work from 30 hours across three staff to six hours for one provider, yielding overhead savings of $1,680, reimbursement incentives of $8,365, and an additional $8,480 from increased patient engagement. The system also led to a 3.2x increase in preventative and diagnostic order submissions for at-risk patients who have historically failed to complete necessary screenings.

Conclusion
Our study demonstrates the application of AI and ML in improving healthcare providers' operational efficiency and fiscal accountability, as well as enhancing patient care management, particularly for at-risk individuals needing preventative and diagnostic screenings. While future studies could expand on this preliminary data to investigate impacts on metrics like CMS Star and risk adjustment factor (RAF), the current results underline the potential for time savings, reduced healthcare costs, and improved provider satisfaction and rates of burnout. Given these promising findings, further exploration and development of ML-based workflows in healthcare are warranted.

Utilizing a ML-Enabled EMR Provider Workflow to Improve Non-Clinical Tasks

Background
Racial minorities and socioeconomically disadvantaged populations are commonly insured through the government-funded Medicaid program. Medicaid has historically not covered the cost of care for these populations. Additionally, these populations often suffer from nonmedical social determinants of health, further increasing the cost of their care. The goal of this study was to determine the impact of Medicaid funding on the profit margins of academic medical centers (AMCs), both governments funded (GAMCs) and non-profit (NPAMCs).

Methods
We used institutional financial data from the National Academy for State Health Policy and Rice University’s Baker Institute of Public Policy, available between 2011 and 2019. The data set included 4,128 hospitals of which 452 were registered with the Association of American Medical Colleges and reported non-profit or government ownership. Aggregate panel data were analyzed by year and medical center. The primary outcome was net profit margin defined as the net income divided by net patient revenue. The main exposure of interest was Medicaid coverage, defined as the proportion of admissions at the hospital level with Medicaid as the primary insurer. Linear regression models examined the impact of Medicaid coverage on net profit margin when adjusted for facility size. Profit margin change was described using basis points – a financial unit of measure where a 1% change is equivalent to 100 basis points.

Results
We identified 378 AMCs under non-profit and 74 AMCs under government ownership. On average Medicaid insurance coverage represented 16% and 20% (p<0.001) of the payor mix among NPAMCs vs. GAMCs, respectively. Increasing proportion of Medicaid coverage was significantly associated with decreasing net profit margins, regardless of institution type. However, the magnitude of profit margin change was more substantial among government centers. For GAMCs, every 1% or 100 basis point increase in Medicaid proportion was associated with an average profit margin decrease of 0.19% or 19 basis points (p=0.002). For NPAMCs, every 1% or 100 basis point increase in Medicaid proportion was associated with an average profit margin decrease of 0.13% or 13 basis points (p=0.001).


Conclusion
Our findings highlight the negative impact of Medicaid funding on profit margins of academic medical centers. This leads to a vicious cycle in which the hospitals that treat high number of these vulnerable, Medicaid-insured patient cohorts become financially drained and unable to invest in capital expenditures and personnel resources. The resulting hospital-level resource depletion impacts access and quality of care for disadvantaged patient populations, and greatly augments health care disparities.

Medicaid Health Insurance Associated with Decreasing Profit Margins among Academic Medical Centers

Background: Tumor lysis syndrome (TLS) is an oncologic emergency that can occur spontaneously or after chemotherapy administration. Rasburicase, allopurinol, and intravenous fluids are the cornerstone treatments for hyperuricemia. Unfortunately, there is a lack of consistent and evidence-based pattern usage of rasburicase in clinical practice. Inappropriate use can range from 30 to 70%, which leads to institutional financial toxicity, in addition to the unnecessary side effects from rasburicase. Therefore, every attempt should be made to ensure that clinicians follow established TLS guidelines when prescribing rasburicase. 
Methods: Change in the Electronic Medical Record (EMR) system was implemented to add an order set for rasburicase in September 2021. Instead of the previous version of a single-click, ready-to-sign order, the newly incorporated order set had multiple selections that followed a guideline-based algorithm with the aim to ensure the appropriateness of rasburicase use. To investigate if the intervention was successful, all rasburicase orders were reviewed and compared for appropriateness by segregating them into 2 groups (A) before intervention (March-September 2021), and (b) after intervention (October 2021-April 2022). Statistical analysis was done using chi-square, and T-test for categorical and continuous variables respectively.
Results: The study included 67 patients from March 2021-April 2022 [n=31 in pre-intervention, and n=36 in the post-intervention group] (Table 1). Most of the patients had hematological malignancies both in the pre-intervention (n=22/31) and post-intervention group (n=28/36). We found that with the intervention, there was a significant improvement in the percentage of clinicians following the appropriate guidelines while prescribing rasburicase (67.74% versus 88.96%; p = 0.03). With the implementation of this new order set, we were able to reduce inappropriate orders by 21.22%. A review of inappropriate orders showed that isolated hyperuricemia of <12 mg/dL (without TLS) was the most common reason for prescribing rasburicase (n=4/10; 40% in pre-intervention and 3/4; 75% in the post-intervention group). We also noted that 25.81%, and 16.67% of the patients in the pre-, and post-intervention groups were not prescribed allopurinol, respectively (p<0.001). 
Conclusion: A newly implemented and multistep, algorithm-based order set helped us to reduce inappropriate rasburicase use, which significantly reduced unwanted side effects, and also financial toxicity.

Implementation Of A New Electronic Medical Record Algorithm To Streamline The Appropriate Use Of Rasburicase

Background:
Arab Americans are one of the largest ethnic groups in the United States. However, the mental health of Arab Americans and their opinions on mental health remain largely understudied despite high rates of mental illness and lower rates of treatment. Unique cultural conceptualizations of mental illness might generate stigma, which contributes to disparities in treatment. We assessed the attitudes and perspectives of Arab Americans about mental illness to address barriers to treatment.

Methods:
Responses were collected through an anonymous bilingual Qualtrics survey from both Arab and non-Arab participants aged 18-years and older. Participants were assessed on stigma towards mental illness and mental health treatment. conceptualizations of mental illness origins, preferred resources for treatment, and perceived barriers to accessing treatment. Comparisons were performed between Arabs and non-Arabs and among Arabs using different demographic parameters. Data analysis was conducted on SPSS using chi-squared analyses, ANCOVA, and T-tests. 

Results:
Compared to non-Arabs, Arabs have significantly higher average stigma towards mental illness (t(292) = -3.42, p < .001) and were more likely to rate psychiatric medications as addictive and unhelpful (t(292) = -2.93, p = .004; t(292) = -3.09, p = .013). Additionally, among Arabs, being Christian, male, and older increased stigma (t(194) = 2.10, p=.037; t(216) = 4.889, p < .001; t(216) = 2.16, p = .037). Regarding beliefs about mental illness, Arabs were less likely to believe in a biological explanation for mental illness compared to non-Arabs (χ² = 10.629, p = .001). Arabs were also less likely than non-Arabs to choose treatment from a therapist or primary care provider (χ² = 5.24, p = .022; χ² = 5.001; p = .025) and more likely to defer mental health treatment entirely (χ² = 6.410; p = .011). Among Arabs, females were more likely to prefer a therapist than males (χ² = 10.562, p = .001). Finally, Arabs were more likely than non-Arabs to endorse fear of being called “crazy” and provider cultural differences as barriers to seeking mental health treatment (χ² = 6.430, p = .011; χ² = 20.274, p < .001). 

Conclusion: 
Our results demonstrate an increased stigma against mental illness among Arab Americans and hesitancy to seek treatment due to social repercussions from the community and cultural discrepancies. 

Significance:
These are among the first findings describing mental illness stigma and barriers to mental health treatment access in the Arab American population and may inform future community interventions providing culturally competent care to Arab Americans.

Exploring Arab American Mental Health Conceptualization and Stigma in Southeast Michigan

Calendar and Volume Trends in CTA Interpretation: Exam Volume Doesn’t Explain It All

Background 
Peer feedback in medical education is a well-studied means of helping students promote self-reflection, identify knowledge gaps, and foster professionalism and communication skills. While multiple medical schools have documented positive outcomes from peer feedback instruction, it is not clear that the quality of peer feedback was specifically assessed during these interventions. Our study evaluates the impact of a “feedback on feedback” exercise on students’ confidence in delivering and receiving peer feedback. We also evaluate student perceptions of instruction on peer feedback and the “feedback on feedback” exercise.

Methods
We created a peer feedback session for first-year medical students at UC Davis for AY2022-23. To prepare for the session, students watched a pre-recorded video on delivering effective feedback that included two demonstration peer feedback scenarios. In class, students provided their peers with feedback following standardized patient encounters. Feedback providers subsequently received feedback on the quality of their feedback from a faculty or fourth-year medical student facilitator. We assessed students’ perceptions of this session and confidence surrounding peer feedback with pre- and post-session surveys and analyzed the data using a two-tailed t-test. 

Results
Ninety-three percent (104/112) and 92% (98/107) of participating medical students responded to the pre- and post-session surveys, respectively. There was a significant difference in students’ self-reported confidence in their ability to provide their peers with effective feedback in a clinical setting before (M = 3.91, SD = 0.68) and after (M = 4.13, 0.60) the session; t(199) = -2.42, p = 0.02. There was also a significant difference in students’ self-reported confidence in their peers’ ability to provide them with effective feedback in a clinical setting before (M = 3.97, SD = 0.66) and after (M = 4.16, 0.58) the session; t(200) = -2.19, p = 0.03. Additionally, most students (33/41, 80%) agreed or strongly agreed that the “feedback on feedback” they received improved their ability to give effective feedback in the future. Students frequently cited learning about different feedback frameworks as the most useful part of the module (19/64, 30%), followed by positive effects on their professional development (15/64, 23%).

Conclusion
Providing medical students with formal training in delivering effective peer feedback, including providing them with feedback on the quality of their feedback, can increase their confidence in their ability to deliver and receive feedback – a skill that is critical for their role as future physicians and educators. 


"Let's Unpack My Feedback!" - An Approach to Enhance Medical Students' Peer Feedback

Background: 
Following devastating hurricanes, such as Hurricane Ike and Hurricane Harvey, Galveston has struggled to replenish its previously existing affordable housing infrastructure due to socioeconomic and political forces. This has caused many to fall into homelessness or live in unstable housing situations. According to the 2022 Point in Time Count, there are 166 reported individuals experiencing homelessness in Galveston County. Street Medicine Galveston aims to meet unhoused individuals where they are in the community through Street Rounds, care for their physical and social needs, and bridge gaps in collaboration between Galveston organizations serving this population. 

Methods: 
We provide essential living necessities, seasonal supplies (winter gear, bug spray, sunscreen, etc.), wound care, and healthy meals from produce donated by community gardens and Galveston’s Own Farmers Market. Additionally, we have created a community network with several other community organizations that serve the unhoused community which allows us to refer individuals to the appropriate resources based on their unique needs.

Results: 
Through Street Rounds, we have learned that meeting unhoused individuals where they are within the community provides a unique solution to several barriers to care including lack of transportation and limited cellular/ internet access. Additionally, levels of education and awareness of available social resources vary widely within this population. By consolidating a community resource guide, our team bridges these gaps by providing recommendations for what community resources to visit based on their expressed needs. Lastly, within the past year, we have been able to serve 253 individuals thus far, 94 of which are recurrent clients and 174 who are new clients. This is a greater number of individuals than were reported within the Galveston County Point in Time Count. 

Conclusions: 
Through Street Rounds, we were able to collaborate and form trusting relationships with the unhoused community as well as several community partners. In the span of 10 months, we were able to serve a total of 253 people experiencing homelessness. With the help of donations, we were able to successfully provide winter gear for more than 50 people and assisted with referrals to other services in Galveston. In the future, we hope to continue exploring different avenues like collaborating with social services, and medical specialties like Psychiatry to provide further care.

Street Medicine Galveston: How We Created a Community Resource Network

Background: Despite advances in mental healthcare in the U.S., only 46.2% of adults diagnosed with any mental illness and 64.5% of those with severe mental illness received mental health treatment in 2020. Primary care providers (PCPs) are at the front lines of identifying and treating mental health conditions and are essential in bridging the mental health treatment gap. The present study examines the relationship between having a PCP and perceived unmet mental health treatment need in adults living in New York City (NYC).  
Methods: This cross-sectional study uses data from the 2019 and 2020 NYC community health survey (CHS). The two years were compared to account for any effect of the COVID-19 pandemic on access to PCPs. Unadjusted odds ratios were used to identify predictors of having a PCP. A binary logistic regression model assessed the association between having a PCP and perceived unmet mental health treatment need, adjusting for sociodemographic and clinical covariates. Post-stratification survey weights were applied using SPSS complex samples.   
Results: In 2019 and 2020, 25.1% and 29.7% of those with non-specific psychological distress (NSPD), respectively, had perceived unmet mental health treatment need. Having no insurance (2019: OR 0.082, 95% CI 0.063, 0.105, p=0.000; 2020: OR 0.171, 95% CI 0.137, 0.214 p=0.000), having perceived unmet medical (2019: OR 0.471 95% CI 0.358, 0.620 p<0.001; 2020: OR 0.656, 95% CI 0.522, 0.825 p<0.001), and mental health treatment need (2019: OR 0.580 95% CI 0.376, 0.894 p=0.014; 2020: OR 0.613 95% CI 0.435, 0.865, p=0.005) were associated with reduced odds of having a PCP. Adults lacking a PCP had increased odds of perceived unmet mental health treatment need in unadjusted analysis, but after adjusting for confounders, this relationship no longer reached statistical significance. Having NSPD (2019: OR 8.632 95% CI 5.506, 13.532, p=0.000; 2020: OR 11.108, 95% CI 7.404, 16.664) and perceived unmet medical treatment need (2019: OR 5.332 95% CI 3.466, 8.203, p <0.001; 2020: OR 3.5857 95% CI 2.583, 5.758, p<0.001) was associated with increased odds of perceived unmet mental health treatment need.   
Conclusion: More than a quarter of NYC adults needing mental health treatment perceived unmet mental health treatment need. To improve access to quality healthcare both within and outside the primary care setting, there needs to be improved patient health literacy, enhanced healthcare professional training in cultural competence and mental health management, and regular mental health screenings and integration of behavioral health into primary care.

Downloads

Next from AMA Research Challenge 2023

Using deep learning to augment radiologist interpretation of brain MRIs for Alzheimer’s disease