Enhancing Equity in AI: George Mason’s Initiative for Improved Injury Detection
George Mason University has received funding from the National Institutes of Health (NIH) through the AIM-AHEAD program. This program aims to increase the involvement of underrepresented researchers in artificial intelligence (AI) and machine learning (ML). The goal is to use the data from electronic health records (EHR) to address health disparities.
Janusz Wojtusiak leads a project that builds on the EAS-ID initiative focused on developing AI tools for injury data. This project seeks to measure equity and quality in imaging documentation. The team includes experts like Katherine Scafide, David Lattanzi, Eman Elashkar, Jesse Kirkpatrick, and Amin Nayebi Nodoushan. Their research aims to improve bruise detection, particularly for individuals with darker skin tones, who face challenges in assessing injuries from violence.
Current studies show that skin color impacts the effectiveness of AI in healthcare. Medical devices often provide inaccurate results for people with darker skin, delaying essential medical care and worsening existing health issues.
The research targets bruises, a common injury in intimate partner violence cases. One in three people in the U.S. experiences intimate partner violence, with it affecting racial minorities more severely. Survivors with darker skin often find their bruises hard to see, leading to delays in medical attention.
The George Mason team is using Alternate Light Sources (ALS) to enhance bruise visibility across various skin tones. They aim to ensure that AI tools can detect and analyze injuries fairly. The team will develop technical and ethical metrics to evaluate these tools’ performance. Input from clinicians, forensic nurses, and community members will guide their efforts to align with ethical AI practices.
How does the AIM-AHEAD program aim to promote diversity in AI research related to health equity?
Interview with Dr. Janusz Wojtusiak on AI-Driven Health Equity Initiatives at George Mason University
News Directory 3: Thank you for joining us, Dr. Wojtusiak. Your team at George Mason University recently received significant funding from the NIH through the AIM-AHEAD program. Can you tell us more about this initiative and its significance?
Dr. Janusz Wojtusiak: Thank you for having me. The AIM-AHEAD program represents a crucial step in promoting diversity within the realm of artificial intelligence and machine learning, especially in health-related fields. By increasing the participation of underrepresented researchers, we can bring diverse perspectives to the table, which is essential for creating equitable health solutions. Our project, in particular, focuses on leveraging electronic health records to tackle health disparities that affect marginalized communities.
News Directory 3: Your project builds upon the EAS-ID initiative. Can you explain the key objectives of your work and how they relate to imaging documentation?
Dr. Wojtusiak: Absolutely. The EAS-ID initiative aims to develop AI tools for injury data analysis, specifically regarding the quality and accuracy of imaging documentation. A main focus of our work is to improve bruise detection through enhanced imaging. We want to ensure that such documentation is not only accurate but also equitable, meaning that our tools can assess and address potential biases in how injuries are recorded and interpreted across different demographic groups.
News Directory 3: That sounds fascinating! What specific health disparities are you aiming to address through your research?
Dr. Janusz Wojtusiak: We are particularly concerned with disparities in how injuries are diagnosed and documented among various populations. For instance, studies have shown that certain groups may not receive adequate care or their conditions may be misrepresented due to ineffective imaging standards. By refining our AI tools, we aim to create a more standardized and equitable approach to injury documentation which can lead to better treatment outcomes for all patients, especially those in underserved communities.
News Directory 3: Your team includes several experts in various fields. How does interdisciplinary collaboration enhance your research outcomes?
Dr. Janusz Wojtusiak: Interdisciplinary collaboration is at the heart of our project. Our team is comprised of specialists in areas ranging from data science to nursing and healthcare policy. This diversity allows us to approach the problem from multiple angles, whether it’s through improving technical aspects of AI development or understanding the practical implications of our findings in clinical settings. This holistic approach will ultimately enhance the reliability and effectiveness of the tools we develop.
News Directory 3: There’s growing concern regarding bias in AI algorithms. How are you addressing this issue within your project?
Dr. Janusz Wojtusiak: Addressing bias in AI is indeed a challenge we take very seriously. We are actively working to ensure that our datasets are representative of the populations we aim to serve. By using diverse electronic health records and promoting inclusive research practices, we can mitigate the risk of bias. Additionally, we are implementing robust evaluation frameworks to continuously assess our algorithms for fairness and accuracy.
News Directory 3: Looking ahead, what are the next steps for your project and how do you foresee its impact on health disparities?
Dr. Janusz Wojtusiak: The immediate next steps involve refining our AI tools and beginning validation studies to ensure their effectiveness in real-world scenarios. Ultimately, we hope that our work will not only enhance the accuracy of imaging documentation but also inform healthcare policies that prioritize health equity. By showcasing the value of these advanced tools, we aim to encourage broader adoption across health systems, which we believe will significantly improve outcomes for underserved populations.
News Directory 3: Thank you, Dr. Wojtusiak, for sharing your insights. We look forward to following the progress of your important work at George Mason University!
Dr. Janusz Wojtusiak: Thank you for having me. We appreciate your interest and support in advancing health equity through innovative research.
The team has two main goals: creating metrics to assess equity in AI tools and applying these to enhance bruise detection models. They have gathered a large dataset of bruise images taken under different lighting, which will help improve AI performance. The interdisciplinary team unites specialists in informatics, engineering, healthcare, and ethics.
Aligned with AIM-AHEAD’s goals, this project aims to address health inequalities and enhance AI’s role in healthcare. By focusing on the needs of underserved communities, the research team aims to improve injury assessment accuracy for diverse populations.
This project is managed by George Mason’s College of Public Health in collaboration with the College of Engineering and Computing. More details about the project can be found at bruise.gmu.edu.