The Data Driver: How Tina Hernandez-Boussard Is Shaping Inclusive Health Care
#Methods
Tina Hernandez-Boussard, PhD, exploring the intersections of technology and medicine in bioinformatics, as reflected through the lens of her research
Tina Hernandez-Boussard grew up in a small rural town, where higher education was uncommon. Despite an environment with limited opportunities, she soon discovered a unique passion: data. “Data was my ticket to a different world,” she recalls. Driven by curiosity and determination, Hernandez-Boussard pursued higher education with a focus on bioinformatics, an interdisciplinary field that combines biology, computer science, and data analysis to understand and analyze biological data.
Fast-forward to today, Hernandez-Boussard, PhD, now serves as the associate dean of research at Stanford University and a professor of medicine. Her journey has been driven by a singular mission: to use data and technology to advance health equity and patient care.
Serving Vulnerable Populations with Data-Driven Insights
One of the most compelling aspects of Hernandez-Boussard’s work is her focus on using AI and data analytics to serve vulnerable patient populations, including those battling opioid addiction, cancer patients experiencing depression, and individuals struggling with mental health issues.
A significant part of her research delves into pain management and the use of opioids. When Hernandez-Boussard and her team started working on pain management and opioids, it was before the opioid epidemic had fully emerged. “Prior to the epidemic, the focus was on ensuring that no one had to deal with pain, leading to a significant promotion of opioid prescriptions,” she says.
As the opioid crisis began to unfold in 2010, it became clear that the system had flaws. “We saw that prescriptions for opioids were really designated by system protocol, not personalized care,” she says. This approach didn’t account for previous opioid addiction, other medications the patient might be taking, or their individual pain management needs. Consequently, the lack of personalized medicine contributed to inadequate and sometimes harmful patient care.
Hernandez-Boussard and her team knew they had to take action. By analyzing large datasets, her team identified trends in opioid prescriptions and patient outcomes, allowing them to develop more precise pain management strategies. “We’ve identified features associated with high-risk patients, such as a history of addiction or concurrent medications,” she says. This information enables personalized pain management plans that minimize the risk of addiction.
Similar data-driven methods are used to address the challenges faced by cancer patients experiencing depression. Hernandez-Boussard and her team have been studying depression after a cancer diagnosis. By applying machine learning, a branch of artificial intelligence (AI) that uses algorithms and statistical models to make data-based predictions, they have identified features associated with depression in these patients. One significant finding was the association between loneliness and depression following a cancer diagnosis. “If we can identify recent losses in a patient’s life, like a divorce or the death of a loved one, we can better predict and manage their risk of depression,” explains Hernandez-Boussard.
Integrating data from a variety of sources, her team crafts comprehensive profiles of patients. This holistic approach allows for targeted interventions that address not only the clinical symptoms of depression but also the social and emotional factors at play. For example, if a patient is flagged for significant loneliness following a divorce, the healthcare team can proactively connect them with support groups, counseling services, or community resources. This not only helps to mitigate their risk of depression but fosters a more supportive and responsive care environment.
“We are at the brink of a digital revolution that is going to be equivalent to, if not bigger than, the Industrial Revolution. AI is here. It’s here to stay. We’re using it. It’s being integrated. Understanding how to embrace that is going to be the future.”
– Tina Hernandez-Boussard, PhD
An artistic portrayal of Tina Hernandez-Boussard, PhD, symbolizing the pursuit of health equity through data science and artificial intelligence, as she works to bridge gaps in healthcare outcomes – Courtesy of DALL-E.
Advancing Mental Health Care
Hernandez-Boussard also recognizes the transformative potential of AI in mental health care, especially in identifying and supporting individuals at risk of severe mental health issues. With mental health concerns escalating at an alarming rate, AI offers innovative solutions for early intervention.
Utilizing natural language processing, which focuses on the interaction between computers and human language, Hernandez-Boussard’s team can analyze clinical notes and patient emails to detect language patterns indicative of depression. This capability allows for the identification of high-risk patients and the provision of timely support.
Moreover, Hernandez-Boussard underscores the potential of AI chatbots to bridge the gap during times when human professionals may not be available. “One of the most critical times for suicide risk is at 4 a.m.,” she observes. “During these hours, professionals aren’t available, and while hotlines might be, access can be challenging.” AI chatbots could provide immediate support and resources during these critical moments, offering a lifeline when it’s most needed.
Addressing Bias and Ensuring Equity
Acknowledging the significant influence of AI on health care, Hernandez-Boussard emphasizes the importance of addressing potential biases in these systems. “AI systems often reflect existing biases from historical data,” she says. This can worsen inequalities, especially for marginalized groups.
She stresses the need for diverse datasets to ensure that models represent the entire population. “When analyzing electronic health record (EHR) data, we include social determinants of health like access to food, transportation, and socioeconomic status. This helps us understand a patient’s broader context and its impact on their health.”
Hernandez-Boussard emphasizes cultural humility in developing models from EHRs. “Patients express their feelings differently based on gender and cultural background. A model trained only on non-Hispanic white women won’t work well for other populations,” she says.
Ultimately, Hernandez-Boussard underscores the importance of developing models trained on diverse data. “Including data from various racial, ethnic, and socioeconomic backgrounds helps avoid perpetuating biases and inequalities,” she states.
Ultimately, Hernandez-Boussard views the current moment in healthcare technology as a pivotal one. “We are at the brink of a digital revolution that is going to be equivalent to, if not bigger than, the Industrial Revolution,” she asserts. “AI is here. It’s here to stay. We’re using it. It’s being integrated. Understanding how to embrace that is going to be the future.”
Save as PDF