From Patients to Questions: Use Question Banks Wisely

Maya Brooks

Hello everyone, and welcome back to the AI Med Tutor Podcast. I’m Maya Brooks, your AI-generated fourth-year medical student. In our last episode in this 4-part series, Dr. Clinch introduced what he called the Rule of One. Learn one clinical pattern deeply from one patient every day. Today we’re going to tackle another idea that might surprise you. Most medical students spend hundreds of hours in question banks. But are they using those question banks the way they were actually designed to be used?

Dr. Randy Clinch

Hi everyone, I’m Dr. Randy Clinch, a DO Family Medicine physician and medical educator. One of the questions I hear from students all the time sounds something like this: “I got home from clinic, so I opened my question bank and started doing questions.” Then I’ll ask, “What questions?” And they’ll say, “Whatever came next.” There’s nothing wrong with working through mixed questions. In fact, later in your preparation that’s incredibly valuable. But immediately after seeing patients all day… I think there’s a better way. Quick reminder: this episode is for education, not medical advice, and nothing we discuss is sponsored by any resource or vendor.

Maya Brooks

So you’re not saying students should stop using question banks. You’re saying they should think differently about when and why they’re using them.

Dr. Randy Clinch

That’s exactly right. Every educational tool has a purpose. The mistake many students make is expecting one tool to accomplish every goal. Imagine a carpenter trying to build an entire house using only a hammer. A hammer is a fantastic tool. But it’s not the right tool for every job. Question banks are the same way.

Maya Brooks

Then let’s start with the obvious question. What is the primary purpose of a question bank?

Dr. Randy Clinch

I think there are four major purposes. First… Performance assessment. Question banks tell you how you’re doing. Second… Interleaving. They intentionally mix diseases together so you learn to distinguish similar conditions. Third… Exam simulation. Real board exams don’t tell you today’s questions are all about heart failure. Question banks teach you to recognize patterns without knowing what’s coming next. And finally… Longitudinal tracking. Over weeks and months they help identify strengths and weaknesses that might otherwise go unnoticed. Those are incredibly valuable functions. But notice something. None of those necessarily makes a question bank the best place to learn today’s patient.

Maya Brooks

That’s an interesting distinction. So learning and assessment aren’t exactly the same thing.

Dr. Randy Clinch

They’re related. But they’re different. Let’s imagine you spent the afternoon caring for a patient with acute decompensated heart failure. You finally understood why elevated filling pressures produce pulmonary edema. You learned why orthopnea develops. You saw jugular venous distention for the first time. Now you get home. Should you immediately complete forty random questions? Maybe. But probably not.

Maya Brooks

Because only one or two of those questions might even involve heart failure.

Dr. Randy Clinch

Exactly. You’ve just experienced what psychologists call a period of heightened learning. Your brain is already thinking about heart failure. That makes this the perfect time to strengthen that new pattern. Not by hoping your question bank happens to show you another heart failure case… But by deliberately practicing exactly what you learned today.

Maya Brooks

And this is where AI can really help.

Dr. Randy Clinch

I think so. One of the tools I’ve been experimenting with is OpenEvidence. Imagine asking it, “Create five board-style questions comparing acute heart failure with COPD exacerbation.” Or, “Write five questions about causes of hyponatremia in heart failure.” Or, “Quiz me on physical examination findings that distinguish heart failure from pneumonia.” Within seconds you’ve created a highly targeted learning experience built around today’s patient. That’s a fundamentally different goal than opening forty random questions.

Maya Brooks

I really like the phrase “targeted learning.” Because that’s exactly what’s happening.

Dr. Randy Clinch

Right. You’re reinforcing a clinical pattern while it’s still fresh. Educational psychologists call this retrieval practice. You’re forcing your brain to retrieve information shortly after learning it. That retrieval strengthens memory far more effectively than simply rereading notes.

Maya Brooks

But some students might wonder, “If AI can generate questions, why not just stop using question banks?”

Dr. Randy Clinch

That’s an important question. Because the answer is… Absolutely not. AI and question banks are partners. Not competitors. Each does something the other doesn’t.

Maya Brooks

Let’s unpack that.

Dr. Randy Clinch

AI excels at targeted retrieval. You control the topic. You control the difficulty. You control the comparison. It’s almost like asking a faculty member to write five questions specifically about the patient you just saw. Question banks serve an entirely different purpose. They expose you to uncertainty. They force you to distinguish similar diseases. They identify weaknesses you didn’t know you had. And they provide standardized explanations that have been carefully written and reviewed. Those are things AI simply isn’t designed to replace.

Maya Brooks

So maybe the biggest mistake students make is asking one resource to do everything.

Dr. Randy Clinch

I think that’s exactly the mistake. Here’s a sentence I’d love every student to remember. Don’t waste a great question bank on a narrow learning objective. If today’s goal is mastering heart failure… Use targeted retrieval. If today’s goal is measuring how well you distinguish heart failure from every other cause of dyspnea… That’s where your question bank shines.

Maya Brooks

Can you walk us through what that might actually look like after clinic?

Dr. Randy Clinch

Sure. You finish clinic. You review today’s illness script. You look over your Pattern Card. Then you spend fifteen or twenty minutes answering five to ten targeted questions about today’s patient. Only after you’ve reinforced today’s learning would I move into a scheduled mixed question block from your primary question bank. Notice what happened. Today’s patient became today’s teacher. Your question bank became today’s assessment. That’s a very different workflow than simply opening random questions and hoping something relevant appears.

Maya Brooks

I hadn’t thought about it that way before. The patient teaches you. The question bank measures you.

Dr. Randy Clinch

Exactly. That’s the shift. Patients teach. Question banks assess. When students separate those two jobs, learning becomes much more intentional.

Maya Brooks

Before we finish, do you have a challenge for everyone listening?

Dr. Randy Clinch

I do. Tomorrow after clinic, don’t immediately open your question bank. Instead, spend fifteen minutes building on today’s patient. Review your illness script. Review your Pattern Card. Generate five focused questions on that exact topic. Then ask yourself, “Do I understand this patient better than I did this afternoon?” If the answer is yes… You’ve had a productive evening.

Maya Brooks

That’s it for today’s episode of the AI Med Tutor podcast everyone—thanks so much for listening! Next time we’ll put everything together into a practical sixty-to-ninety-minute evening routine that you can start using immediately during your clinical rotations.

Dr. Randy Clinch

Remember… The goal isn’t to answer the greatest number of questions. The goal is to become the kind of clinician who recognizes patterns quickly, reasons carefully, and keeps improving one patient at a time.

Maya Brooks

We’ll see you next time everyone on the AI Med Tutor Podcast. And, in the meantime, stay curious and keep learning!