Patricia Brennan, PhD, RN, FAAN, and director of the National Library of Medicine (NLM), delved into the profound impact of public policies on the NLM and, in turn, the library’s contributions to shaping policies during a virtual presentation.
The talk was held Thursday, Sept. 7, as part of the university’s Quality Enhancement Plan (QEP). Brennan’s presentation underscored the intricate relationship between public policies, the NLM’s operations, and the advancement of science and health care. She highlighted the library’s pivotal role in shaping policies and its dedication to fostering open science, promoting data sharing, and utilizing artificial intelligence (AI) to improve access to knowledge — all while ensuring trustworthiness and transparency.
The full presentation can be viewed on YouTube.
“Public policies play a pivotal role in the broader realm of science, health, and society; science alone is insufficient for bringing about meaningful change,” Brennan said. “It must be integrated into and guided by effective policies.”
Diverse policies affect NLM operations, spanning from legislation to administration policies, both at the national and organizational levels. The policies emerge from various sources, including government executive orders, hospital protocols, NIH policies, scientific journal standards, and international agreements.
“Each policy plays a unique role in shaping how the NLM conducts its research, manages data, and interacts with the scientific and health care communities,” Brennan said.
Brennan explained the NLM is particularly invested in policies pertaining to open access, data sharing, information privacy, data management, and international data sharing.
“When we talk about data sharing across countries, these policies are critically important,” Brennan said. “We work with the World Health Organization and we work with many international bodies to make sure the policies foster the open science engagement that the NIH and the United States government are taking forward. We also work closely with the health data standards industry and the health data interest groups because the policies about health data affect the data that is ultimately available for research. In addition, the library helps to provide the standards that codify clinical language.”
Brennan highlighted the library’s commitment to fostering open science engagement, collaborating with international organizations, and ensuring equitable access to scientific information across borders. She also discussed how the NLM actively engages in policy-related areas such as health data standards, public health, and clinical care, especially during the COVID-19 pandemic.
Brennan described the critical importance of government appropriations and efficient government operations in influencing the NLM’s functioning, expressing concerns about potential budget constraints and underscoring the necessity of financial support to continue the library’s mission effectively.
The presentation also described the NLM’s role as a research institution, spending a significant amount on research initiatives each year. These efforts encompass areas like evolutionary genomics, natural language processing, networks, gene regulation, and chromatin research, which contribute to advancing health care and scientific knowledge.
Brennan discussed how the NLM employs AI in its operations, with a specific focus on improving access to scientific literature, automating indexing processes, and enhancing its clinical data research capabilities. She provided insight into the challenges of ensuring that AI applications align with NLM’s goals of trustworthiness, security, fairness, and privacy.
In the context of AI, Brennan discussed the concept of trustable AI, emphasizing the importance of maintaining trust in both the algorithms and the application of the technology. AI should be considered as a partner to users, not a replacement, as highlighted by the AI Risk Management Framework developed by the National Institute of Standards and Technology (NIST) to enhance trustworthiness in AI systems.
“We have computing schemes now that are more improved and have better mathematics which leads to machine learning,” Brennan said. “We see a resurgence of AI right now that is a very specific quantitative type of approach to understanding information from data. There are new approaches in machine learning that we refer to as deep learning, largely meaning that machine learning algorithms are then processed multiple times to come to a more precise solution and natural language, which is a type of artificial intelligence that draws from taking aspects of words in a paragraph or words in a sentence that may be semantic.”
Brennan also outlined the process of building trust in AI, emphasizing the need for transparency, accountability, and thorough documentation at every stage, from problem recognition to algorithm creation, data acquisition, training, evaluation, retraining, and application. She encouraged policymakers to consider these principles as they navigate the complex landscape of AI in health care and beyond.
The Health Policy Forum Series is part of UTHealth Houston’s efforts to continuously enhance the academic programs and student learning environment at the university. Known as Healthcare Policy for Health Professionals, or HP2, the QEP promotes student engagement, critical thinking, and career preparedness, and reflects UTHealth Houston’s commitment to institutional quality and effectiveness.
For more information about the university’s QEP, including opportunities for student and faculty involvement, contact Angela Gomez, EdD, at Angela.P.Gomez@uth.tmc.edu, or Sandra McKay, MD, at Sandra.McKay@uth.tmc.edu.