Why Biostatistics Matters More Than Ever in the Age of AI
Goran Trajkovski, Ph.D., has spent his career helping people make sense of complex data. With advanced degrees spanning mathematics, computer science, health care management and leadership, he has led research labs, designed data science curricula, authored more than 20 books and hundreds of publications, and advised organizations on artificial intelligence and assessment.
Along the way, Goran’s passion has remained consistent: helping students develop the critical-thinking skills to confidently interpret data and apply it to real-world problems.
That practical approach leads the way when he teaches our Introduction to Biostatistics course. Designed for aspiring health care professionals, researchers and anyone working with clinical data, the course emphasizes reasoning over rote calculation, using real health care datasets and hands-on analysis in R to prepare students for graduate study and professional practice. Goran shares why biostatistics has never been more important in the age of AI, how he helps students overcome their fear of statistics and the real-world skills they'll take with them long after the course ends.
Many of them have quietly decided at some point that they are not quantitative people, usually based on one bad experience with an algebra class 15 years ago. Watching that belief come apart is the best part of the job.
You joined our esteemed instructors in 2025. What drew you to teaching our students?
They are working adults who signed up on purpose, often after working a full day, because a specific gap is standing between them and the next thing they want to do.
The other draw was the department's mission. Biostatistics is a course in:
Clinical Laboratory Scientist Preparatory Program
Certificate Program in Clinical Research Conduct and Management
Post-Baccalaureate Health Professions Program
Post-Baccalaureate Program in Psychology
That means the students are heading into labs, clinics and graduate programs where the statistics are not decorative. Building something that carries that much weight was an appealing problem.
How would you describe your students?
They are capable, motivated and a little wary of math. Many of them have quietly decided at some point that they are not quantitative people, usually based on one bad experience with an algebra class 15 years ago. Watching that belief come apart is the best part of the job.
In three words, how would you describe your teaching style?
Concrete, patient, unhurried.
The audiences that surprise me are those already working in health care without a quantitative title. None of these roles list biostatistics in the job description, and all of them are materially improved by it.
How does your professional experience shape what you teach?
I have spent years designing assessments, and I still write courses on experimental design and causal inference in R, so I know the difference between a student who can execute a procedure and one who understands what the procedure claims. The course is built to produce the second kind.
Also, I have worked on the health care side, and I have seen how often a defensible analysis gets misread in a meeting. We spend real time on interpretation and how you respond when someone asks what the result means.
What types of students would benefit most from taking this course? Are there audiences who may not realize biostatistics could strengthen their careers?
The obvious beneficiaries are people heading to medical, dental, nursing, physician assistant or public health programs, as well as clinical laboratory scientists and anyone planning graduate work in the biosciences. Every one of them will read primary literature for the rest of their career, and biostatistics is the difference between reading a paper and evaluating one.
The audiences that surprise me are those already working in health care without a quantitative title. Nurses moving into quality-improvement roles. Lab managers who are suddenly accountable for turnaround metrics. Clinical research coordinators who assemble the data that someone else analyzes and would like to know whether that analysis is any good. Health policy and health administration people. Regulatory and compliance staff. None of these roles list biostatistics in the job description, and all of them are materially improved by it.
How do you support students during the class?
The course is built in short modules, so nobody has to hold an hour of new material in their head at a time. Every technique arrives attached to a real dataset and a real question, never as an abstraction. I keep the R work scaffolded, so students who have never written a line of code are producing useful output in the first weeks.
I answer questions quickly, and I answer the question underneath the question when I can see one. I am explicit about where the hard parts are because being told in advance that a concept is genuinely difficult is a form of support. It stops people from concluding that the difficulty is a personal failing.
The role has shifted from producing the answer to interrogating it. The person who understands what a model assumes is now the person standing between an organization and an expensive mistake.
Many people hear “biostatistics” and immediately worry they'll need to be a math expert. What would you say to someone who feels intimidated by the subject?
The arithmetic is the part that the computer does. What remains is reasoning, and reasoning is a skill this person has already practiced for years in their clinical or laboratory work.
Biostatistics at this level asks a small number of recurring questions. Compared to what? How much of this could be chance? Who was actually measured and who was left out?
Those are judgment questions, not calculation questions. I have taught confident mathematicians who could compute a p-value flawlessly and could not tell you what it meant. I have taught nurses who arrived convinced they were bad at math and who turned out to have excellent statistical instincts because they had spent a decade noticing when a number did not match the patient in front of them. That instinct is the hard part to teach. If someone has it, the rest is mechanics.
How has the explosion of health care data and artificial intelligence changed the role of biostatistics over the past few years?
The volume of health data has grown faster than anyone's capacity to check it, and the tools now produce polished answers in seconds. An AI system will hand you a coefficient, a p-value and a confident paragraph of interpretation. What it will not tell you is that your sample excluded everyone without a stable address, or that the outcome was measured differently in two of the sites, or that you have tested 40 hypotheses and reported one.
The role has shifted from producing the answer to interrogating it. The person who understands what a model assumes is now the person standing between an organization and an expensive mistake.
What excites me right now is that we have built systems that generate statistical output at enormous scale, and we have not correspondingly built the human capacity to evaluate it. That gap is where the next decade of interesting work sits, and it is a gap that gets closed one educated person at a time.
What practical skills will students walk away with that they can immediately apply in their careers or future studies?
They can take a messy dataset into R, clean it, describe it and visualize it without asking anyone for help.
They can choose an appropriate test for a given question and defend the choice.
They can critically read a clinical paper’s methods and results sections, which is the single most durable skill they will use weekly for the rest of their careers.
They can explain a statistical result to someone who has no statistical training, in plain language, without either overstating it or hedging it into meaninglessness.
As both a researcher and educator, what continues to excite you most about biostatistics today?
It is the discipline that decides what counts as evidence. Every claim in medicine passes through it. That is an unusual amount of leverage for a field that most people find dry.
What excites me right now is that we have built systems that generate statistical output at enormous scale, and we have not correspondingly built the human capacity to evaluate it. That gap is where the next decade of interesting work sits, and it is a gap that gets closed one educated person at a time.
What advice would you give to your students on how to best succeed in our current workforce?
Become the person who asks how the number was made. The people who reliably ask that have more influence than their titles suggest.
Learn to explain technical work to people who are not technical. It is undervalued and it compounds.
Outside of work, where can we find you?
In my studio, making things with my hands. I sew handmade purses and work in leather, jute and mixed media. I also write fiction. I co-authored a health care technothriller called Recommended Action, which let me put a decade of thinking about clinical data into a form where the stakes are literal rather than statistical.
