Skip to Main Content

Medicine meets computer science

Yale Medicine Magazine, Autumn 2023 (Issue 171) Obesity Special Report
by Mary Ann Littell


LUCILA OHNO-MACHADO, MD, PhD, MBA, is the Waldemar von Zedtwitz Professor of Medicine and Biomedical Informatics and Data Science, deputy dean for biomedical informatics, and chair of the new section of biomedical informatics and data science. Born in Brazil, she came to the United States to earn her PhD and stayed to build a career in academia. She recently joined YSM to create a new department aimed at supporting the use of biomedical informatics and data science through applied research, with the goal of improving human health and eliminating disparities. Yale Medicine Magazine spoke with her about her current initiatives and future plans.

How did you first become interested in computers and medicine?

My first encounter with a computer-like device was in high school, when we used programmable calculators in physics class. I was fascinated. After class, our professor taught us how to program them, and that was super-interesting. I thought, how can I use this in my career?

As a medical student at the University of São Paulo, I learned about the emerging field of medical informatics. The school had just established a medical informatics residency, so I became one of its first residents in this specialty. After that, I obtained an MBA, focused on health care administration and IT. Having exhausted my educational possibilities there, I applied to graduate schools in many countries and was lucky to get a spot at Stanford, where I earned a PhD in medical information science and computer science. At the time, HIV-related research was exploding. So that’s how I got my start—in informatics applied to HIV.

Describe the work Yale recruited you to do.

At the University of California, San Diego, I built a biomedical informatics program from scratch. Creating something that didn’t exist before was exciting—like being at a start-up. Now I get the chance to do it again at Yale! Building a department of biomedical informatics and data science includes three pillars: training, research, and a service component. It is a wonderful opportunity that will take some years to bring to fruition.

Right now, we are recruiting a multidisciplinary faculty in different subspecialty areas, with a goal of 40 to 50 new faculty. Informatics brings together people in basic science, clinical care, community engagement, health disparities, computers, engineering, and other areas. It is a huge collaboration focusing on essentially anything that involves data—which is everything.

Settling into our dedicated space, currently located on the 9th floor of 100 College, and moving to the 10th floor of 101 College when completed, we’re setting up systems and processes, managing a large portfolio of research grants, and expanding the existing training program to educate biomedical informatics professionals. For our service component, we will collaborate with biomedical researchers and clinicians to take on the questions and challenges facing the health care system and biomedical research (basic, translational, and clinical).

How is the intersection of computer science and medicine transforming research and patient care?

In the past, and it’s still true today, the experience of individual expert clinicians has been key to making a diagnosis, prescribing treatment, and understanding what treatments work and don’t work—and in what subtypes of patients.

But we are moving past that point. Data now permits us to gain that expertise from what’s called the “learning health care system.” This is a concept that came about in the last decade or so. Essentially, it means utilizing all the data collected every day in electronic health records and other systems. For example, when patients have an adverse event from a medication, there is no mechanism in place for automatic surveillance or monitoring. But now, patients’ physicians can record such incidents into electronic health records, where they become part of a larger record. By accessing the data from a larger collection of patients being treated, we establish patterns that tell us what we are doing right and what we can do better.

Of course, the question arises: With large amounts of computerized data, how do you use it for research while protecting the privacy of patients? In my informatics research, we’ve devised ways to distribute this data in a privacy-preserving manner by building models that allow accessibility, while keeping the data private within the institution.

What’s the most misunderstood aspect of this field?

I would say it’s the perceived danger or threat of artificial intelligence. Used correctly, AI should be viewed as a helpful tool, rather than a threat. It can be impactful because it processes much more information than one brain can. AI will not only uncover health disparities, but also propose ways to mitigate and possibly eliminate them.

Machine learning, a subset of AI, uses algorithms to automatically recognize patterns from data, and then apply that information to better decision-making. It requires a large amount of data. When machine learning processes millions of records, it can discover disparities and demonstrate when outcomes from a particular subgroup are worse than the larger group. Imagine yourself as a clinician seeing many patients. One patient in a particular subgroup does not respond well to a certain treatment. Because your experience is limited to your own patients, you wouldn’t know that the same thing is happening to the clinician next door. By aggregating the data of a large group of patients, you’ll know that this treatment, in this subgroup, is associated with poor outcomes.

What is the biggest challenge in biomedical informatics?

This field is relatively new, compared to other specialties in medicine. It takes time for a new field to gain status as an established science, attracting young scientists to the field. The fact that this field is interdisciplinary also creates challenges. Combining a mathematical/statistical area with a clinical/biomedical research area makes the training harder, because you must be trained in both disciplines.

What is unique about working within the Yale community?

With deep resources and an excellent faculty, Yale offers solid foundations we can build on. A primary reason why I came here is for the opportunity to create a new department in an outstanding institution. There are few such departments; perhaps only five or six universities have a large, impactful biomedical informatics department. To create one at a university like Yale can change the way other institutions view the field, igniting many new efforts.

Looking toward the future of biomedical informatics and data science, what excites you the most?

I’m excited because the field of biomedical informatics and data science has assumed its place of importance in academia. This field will have a substantial impact on human health and the health care system.

I’m also excited because here at Yale, every department wants to have their own point person in biomedical informatics and data science, and we will work together to ensure that everyone benefits. We’re not just throwing our hats into the ring, saying, “We’ll build a department,” and then checking off that box. We are building THE department, and we want to make it succeed and have a major impact inside and outside the university.

To nominate a subject for Q&A, write to: or Yale Medicine Magazine, 50 Division Street, 2 Science Park, Floor 2, New Haven, CT 06511.

Previous Article
Skin in the game