Written by: Laxmi Keerthi Ravula, BHMS, MPH
I have been working in the Houston Methodist Research Institute’s Patient Engagement Lab for almost two years as a Clinical Research Coordinator. I joined this lab to improve health and prevent disease through applying patient-engaged techniques and access to education to address health disparities. I strongly believe in applying the power of effective communication and research methods to benefit patients and the public. I think we can all contribute to improving the world by sharing our knowledge and promoting diversity, equity, and inclusion. I aim to bridge gaps in our health system.
Attending the RICE University Artificial Intelligence (AI)/Machine Learning (ML) Conference 2024 changed my perspective on AI and how we can use it in healthcare systems. I also learned about using ML algorithms in the medical field, something of which I was previously unaware. I discovered that ChatGPT is being used for medical diagnosis and that doctors will likely utilize it in the future. Terms like Med LEE and Med BERT were introduced to me during the conference, and I learned that they are used in the medical domain for specific medical datasets and structured electronic health records to predict diseases. This Machine Learning Model is meaningful because BERT can turn words into numbers. Through this, I learned that we can train ML models with textual data so that this algorithm can be beneficial for chatbots for medical queries, for example, Health Monitoring and follow-up care with patients. We can also use AI to remind patients to take their medications, and provide them with educational resources. AI is also being used to extract social determinants of health (SDOH) to understand who needs better access to care.
I also learned about Meta AI Llama, a large medical language model used for conversing with patients online, providing information, and assisting in medical diagnosis through apps and doctor chats. Large language models like ChatGPT can produce human-like responses, improving communication between healthcare personnel and patients across language barriers. These models are crucial in the healthcare field, as they can provide basic medical advice without needing direct consultation with healthcare experts, aid in health education, and answer patient inquiries. Furthermore, large language models can help address operational challenges in low- and remote-income countries (LMICs) with a shortage of healthcare providers. They can support time-consuming administrative tasks such as generating discharge reports, summarizing patient stays, and gathering patient histories. Additionally, these models have the potential to bypass infrastructure restrictions.
During the conference, I had the opportunity to attend a special workshop where I learned more about public health and patient engagement equity. To help people on a large scale, we need to have a diverse team with various skills, including doctors, physicians, research teams and data scientists. Everyone has equal importance, and there should be reciprocity, valuing people and unlocking their potential. A speaker asked us to think about the barriers faced by patients, and I suggested that telehealth communication could be a solution to remove these barriers. This would help patients have easy access to providers to talk about their issues, improve access to specialists, reduce transportation costs, decrease exposure to illness, and be accessible to patients who live in remote areas. I also believe there should be a team available 24/7 to answer patient questions, and this policy should be implemented by the government in all hospitals to improve the quality of healthcare. Free disease campaigns in clinical and community settings and training regarding AI and Telehealth apps are significant and should be funded by the government and policies.
At my first workplace conference since completing my Masters, I had the opportunity to present a poster about our research for the first time. My first poster presentation was titled, “Listening to Patients at Higher Risk of Not Completing Kidney Transplant Evaluation: Planning Intervention to Support Underserved Communities.” My favorite moment was when people who visited our poster said, “This poster is full of facts and the groundbreaking reality of our system.” It allowed me to voice patient challenges and improvements that can be made in our hospitals and systems. I enjoyed presenting this and received praise for our efforts in talking to patients and determining why they were dropping out even though they were eligible for kidney transplants, as well as for highlighting financial and knowledge challenges. The best moment was when people appreciated our concept of talking to patients, interviewing them, and bringing out their challenges and solutions to enhance access to kidney transplants.
People were amazed by the way we used a machine-learning model to understand the demographics of people dealing with various barriers and were surprised about the fact that 54% dropped out from kidney transplant even though they were still eligible. People at the conference loved the concept and requested that we continue to use Artificial Intelligence and Machine learning in research and bring more interventions from the patients’ perspective. During these two hours, I gained a deep understanding of the significance of engaging with patients and listening to their experiences, barriers, and challenges so that we can better support them. We can all contribute to making healthcare more accessible and better. The government must invest in and implement policies and campaigns to educate patients about AI applications and usage.
I want to express my gratitude to my director, Dr. Amy Waterman, and my Patient Engagement Lab team for giving me this opportunity to present our research work at the RICE University AI Health Conference 2024. In the future, I aim to incorporate machine learning algorithms into healthcare research to enhance patient outcomes and experiences, and to eliminate barriers, particularly in underserved communities.