Build a Board Prep Plan That Learns From Your Performance

Maya Brooks

Hi everyone, and welcome back to the AI Med Tutor podcast. I’m Maya Brooks, your AI-generated fourth-year medical student co-host, and as always, I’m joined by Dr. Randy Clinch. Today we’re going to pull together a lot of ideas we’ve discussed throughout this podcast. Maybe you’re early in your third year. You’ve started clinical rotations, you’re preparing for Shelf or COMAT exams, and Level 2 or Step 2 is still somewhere out on the horizon. Or maybe you’re a fourth-year student who has already done thousands of practice questions and taken multiple assessments, but you’re still working to meet your medical school’s requirements to be approved to take your board exam. Those sound like two very different students. But both of them eventually have to answer the same question: How do I build a board-preparation plan that is actually personalized to me? And today, we’re not going to give you another generic study schedule. Instead, we’re going to show you how to build a study plan that actually learns from your performance. Dr. Clinch, where do we start?

Dr. Randy Clinch

Hi everyone. I’m Dr. Randy Clinch, a DO family medicine physician and medical educator. I think we start by changing what we mean when we say a personalized study plan. Personalized doesn’t mean you picked your favorite resources. It doesn’t mean you decided whether you study better in the morning or evening. And it doesn’t mean somebody gave you a spreadsheet and you changed the dates. A truly personalized board-preparation plan changes in response to what your performance is telling you. Think about those two students Maya just described. The third-year student may be learning internal medicine right now, seeing patients, completing school assignments, working through a question bank, and preparing for a COMAT or Shelf examination. The fourth-year student may already have substantial clinical experience, several practice assessments, thousands of completed questions, and a school readiness benchmark they still haven’t reached. I wouldn’t prescribe the same study plan to both of them. Their data are different. Their developmental needs are different. So their prescriptions should be different. But—and this is important—the underlying learning system can remain the same. We’ve talked about many of the individual parts of that system in previous episodes. Today, we’re going to connect them. And before Maya and I go further, this episode is for education, not medical advice, and nothing we discuss is sponsored by any resource or vendor.

Maya Brooks

So personalized doesn’t necessarily mean complicated.

Dr. Randy Clinch

Not at all. In fact, I think students sometimes make personalization much more complicated than it needs to be. Here’s the framework I want you to remember today: Collect. Interpret. Prescribe. Test. Adapt. Five steps. Collect the data you already have. Interpret what those data are actually telling you. Prescribe the smallest useful intervention. Test whether that intervention worked. Then adapt your plan based on the result.

Maya Brooks

Let’s start with collect. Because I think when students hear “performance data,” they immediately think about a COMSAE score or an NBME score.

Dr. Randy Clinch

And those are useful data, but they’re only part of the picture. Your curriculum has been generating data about you for quite a while. Think about course examinations. COMATs or Shelf exams. Quizzes. Clinical assessments. Institutional assignments. Then consider your question banks. What’s your overall performance? What happens by discipline or organ system? What types of questions are you repeatedly missing? Are you performing differently in tutor mode versus timed mode? What happens as your blocks get longer? Are you getting questions right confidently, or are some of those correct answers really unsure-correct answers? Then there are formal practice assessments such as COMSAEs, NBME self-assessments, or whatever assessments your school uses. And don’t forget your clinical experiences. Which patient presentations are becoming easy for you to recognize? Which ones still feel disorganized? What are you repeatedly looking up after rounds? What feedback are your preceptors giving you? Your Pattern Cards can be data. Your Miss Log can be data. All of those sources can tell you something about your learning.

Maya Brooks

So the goal isn’t to find one number that tells me whether I’m good or bad at cardiology.

Dr. Randy Clinch

Correct. Look for convergence. If one question block says cardiology was weak yesterday, that’s interesting. If your last three blocks show the same pattern, your practice assessment shows it, and your Miss Log keeps accumulating cardiovascular reasoning errors, now you have a signal. We discussed this idea in much more detail in the episode **“Still Below Your Benchmark? Focus Your Prep Where It Matters Most.”** That episode is particularly useful if you’re already well into board preparation but still haven’t reached your institution’s readiness benchmark. The principle is simple: Don’t chase every fluctuation. Look for the signal.

Maya Brooks

Okay. So I’ve collected the data and found a recurring weakness. Let’s say cardiology really does keep showing up. What’s next?

Dr. Randy Clinch

Interpret. And this is where students sometimes jump too quickly from diagnosis to treatment. They see: “Cardiology: 48 percent.” And immediately decide: “I need to review cardiology.” Maybe. But “weak cardiology” is not yet a study prescription. We need to know why it’s weak. I like to start with three categories we’ve used throughout this podcast: Content. Reasoning. Mechanics. Suppose I don’t really understand preload and afterload, or I can’t explain why left-sided heart failure produces the findings I’m seeing. That’s probably a content problem. But suppose I understand heart failure and COPD individually. When they’re presented separately, I do fine. Put them into similar clinical vignettes, though, and I keep confusing them. That’s more likely a reasoning or discrimination problem. Now suppose I correctly identify acute decompensated heart failure, but the question asks for the most appropriate next step and I choose another diagnostic test instead of the needed treatment. I may know the medicine. My question-taking mechanics may have failed me. Same broad performance category. Three different problems. Three different prescriptions.

Maya Brooks

That seems like one of the most important points in the episode. A weak topic isn’t automatically a content problem.

Dr. Randy Clinch

Absolutely. And if you want a deeper dive into that distinction, go back to **“Mastering Miss Logs: A Simple System to Capture and Retest Exam Mistakes.”** We use the Content, Reasoning, and Mechanics categories there to help turn individual misses into recurring patterns. And if your issue is that you know the information but your reasoning keeps taking you down the wrong path, listen to **“Right Answer, Wrong Reason? Think Like a Clinician.”** That episode is also a good reminder that correct answers are not always evidence of good reasoning. Sometimes the most useful question in your block is one you got right for the wrong reason.

Maya Brooks

Now we’ve collected the data and interpreted the problem. Step three is prescribe. And I suspect this is where we need to talk about resources.

Dr. Randy Clinch

We do. But I want to change the resource question. Students frequently ask: “What’s the best resource for Level 2?” “Should I use this question bank or that question bank?” “Should I add another video series?” “Should I buy another book?” I want you to ask something different: **What job do I need this resource to perform?** That question can save you a tremendous amount of time.

Maya Brooks

So walk us through the jobs.

Dr. Randy Clinch

Start with your clinical encounters. Patients give clinical knowledge context. A real patient can transform isolated facts into a recognizable clinical pattern. Then illness scripts help organize those patterns. Who tends to get this condition? How does it present? What findings support it? What findings argue against it? What competes with it? What happens next? We devoted an entire episode to that called **“Illness Scripts: The Clinical Reasoning Tool for Third-Year Rotations.”** That’s an especially useful episode if you’re early in third year and you’re trying to figure out how clinical learning fits with eventual board preparation.

Maya Brooks

Then come Pattern Cards.

Dr. Randy Clinch

Right. Pattern Cards compress important clinical patterns into something retrievable. Presentation. Key clues. Mechanism. Then you can retrieve from the card and ask: What else could this be? What’s the hinge clue? What would I do next? And we just devoted an entire episode to that tool called **“Pattern Cards: Turn What You’re Learning Into What You’ll Recognize Later.”** If Pattern Cards are new to you, listen to that episode rather than trying to turn every disease in your textbook into a card. Remember: keep the card small and make the thinking big.

Maya Brooks

Where does content review fit?

Dr. Randy Clinch

Content resources have an important role when the diagnosis really is a content deficit. A focused video, a few pages from a trusted text, a reference resource, or an institutional learning resource may be exactly what you need. But use it for precision repair. If I discover that I don’t understand nephritic versus nephrotic physiology, I don’t necessarily need to start an entire renal course again. Retrieve what I think I know first. Identify the gap. Use the resource to repair that gap. Then close the resource and retrieve again. We explored that distinction in **“Content Review or Productive Procrastination?”** That’s worth revisiting if you find yourself spending hours watching videos or reading but have very little evidence that your performance is changing.

Maya Brooks

And then we get to question banks.

Dr. Randy Clinch

Question banks have several jobs depending upon where you are. Early in learning, questions can provide retrieval practice and expose gaps. Later, they become increasingly valuable for application, discrimination, interleaving, and transfer. And as you approach board conditions, mixed timed blocks become performance tests. We talked specifically about assigning different jobs to patients, AI-generated targeted questions, and validated question banks in **“From Patients to Questions: Use Question Banks Wisely.”** And if you want more detail about how your q-bank mode should change depending on your goal, listen to **“Mastering Question Banks: Strategic Modes for Smarter Board Prep.”** The important idea is that “do more questions” is not a complete study strategy.

Maya Brooks

And that connects back to the system we’ve been developing throughout the podcast.

Dr. Randy Clinch

It does. Patients generate patterns. Illness scripts organize them. Pattern Cards compress them. Question banks test them. Miss Logs tell you which ones are still breaking down. Interleaving teaches you to distinguish similar patterns. And mixed timed blocks test whether you can retrieve and apply what you know when nobody tells you what’s coming. That's the system. But where you spend your time within that system should depend on your data.

Maya Brooks

Let’s make that concrete. I’m an early third-year student on my internal medicine rotation. How should my personalized plan look?

Dr. Randy Clinch

Your current rotation should probably be your primary learning environment. See the patients. Build illness scripts. Capture selected high-value Pattern Cards. Complete the educational work your institution requires. Use your question bank to practice and assess what you’re learning. Review your misses. And maintain a small amount of important material from prior rotations as you accumulate them. We developed that idea in **“One Rotation at a Time: Building Toward Shelf, COMAT, and Boards.”** The basic structure is: Primary. Repair. Maintenance. Your current rotation is primary. A weakness exposed by your current work becomes repair. And selected learning from previous rotations becomes maintenance. You don’t need to run an entirely separate Level 2 or Step 2 curriculum while you’re learning medicine on rotations. Use your rotations to build the knowledge system that you’ll eventually test under board conditions.

Maya Brooks

And we’ve also talked about what happens when the rotation ends.

Dr. Randy Clinch

Yes. If you’re moving from one rotation to another, listen to **“The Rotation Reset: Carry Forward What Matters.”** The point is not to drag the entire previous rotation into the next one. Carry forward the patterns that matter. A few Pattern Cards. Perhaps an unresolved Miss Log item. Some spaced mixed retrieval. Keep the pilot light on without allowing last month’s rotation to compete with this month’s primary learning.

Maya Brooks

Now let’s switch students. I’m a fourth-year student. I’ve completed my core rotations. I’ve done a lot of questions. I’ve taken practice assessments. But I’m still below my school’s benchmark for being approved to take Level 2 or Step 2. How does Primary, Repair, and Maintenance change?

Dr. Randy Clinch

Substantially. Your primary work is now increasingly integrated board performance. Mixed questions. Timed conditions. Clinical decision-making across specialties. Your repair work comes from the recurring bottlenecks exposed by those conditions. And maintenance keeps previously repaired areas available without allowing them to consume your entire day. This is where **“From Learning to Performance: The Dedicated Pivot”** becomes especially important. In that episode, we discuss the transition from learning mode to performance mode. You test. You diagnose the leak. You repair it. You retest it. Then you put it back into mixed performance. That last step matters.

Maya Brooks

Which brings us to step four in today's framework. Test. How do I know the prescription worked?

Dr. Randy Clinch

This is one of the places where students can unintentionally fool themselves. “I watched the video.” That's completion, not evidence of learning. “I reread my notes.” Same problem. “I reviewed all the explanations.” Still not enough. Even: “I did ten targeted questions and got nine right.” That's encouraging, but I still want to know whether the learning transfers. Think about an evidence ladder. First: Can I retrieve it? Then: Can I answer targeted questions about it? Then: Can I distinguish it from similar conditions? Then: Can I recognize and apply it when it appears unexpectedly in a mixed set? And finally: Can I do that under timed conditions? That progression moves us from learning toward transfer.

Maya Brooks

So if I repaired heart failure by doing ten heart-failure questions, the next test shouldn't necessarily be another ten heart-failure questions.

Dr. Randy Clinch

Right. Eventually, hide heart failure among pulmonary embolism, COPD, pneumonia, renal disease, anemia, arrhythmia, and everything else. Now your brain has to determine what problem it's solving. That’s why interleaving matters. For a deeper dive, listen to **“Interleaving Look-Alike Questions for Better Clinical Reasoning.”** And if you're struggling specifically with management questions—where you identify the disease but repeatedly miss what to do next—listen to **“Mastering Next-Step Questions: A 4-Step Framework for Smarter Clinical Decisions.”** Different failure. Different prescription.

Maya Brooks

And if the problem is time?

Dr. Randy Clinch

Then don't automatically prescribe more cardiology. If you repeatedly run out of time, rush through the final questions, misread the task, or lose working memory late in a block, that's a different problem. Go back to **“Stop End-of-Block Panic: A Smarter Timed Question Workflow.”** Again: Diagnose before you prescribe.

Maya Brooks

That takes us to the last step. Adapt.

Dr. Randy Clinch

And this is the step that makes the plan truly personalized. Your plan should change because you changed. Suppose cardiology was your major repair priority three weeks ago. You worked on it. You retrieved. You repaired content gaps. You compared look-alikes. You retested. And now cardiology is performing well in mixed blocks. Great. Stop treating it like your biggest weakness. Move it toward maintenance. Something else may now deserve repair. That's adaptation.

Maya Brooks

So we're not creating a twelve-week study plan on Sunday and then following it regardless of what happens for the next twelve weeks.

Dr. Randy Clinch

Correct. A plan gives you direction. Your data tell you when to change direction. And that's why I like a very short weekly meeting with your own performance data. You don't need a complicated spreadsheet unless that helps you. Once a week, ask yourself four questions. What is getting better? What keeps showing up? Why is it still happening? And what will I change this week? Then update your three buckets: Primary. Repair. Maintenance.

Maya Brooks

Let's do that for both of our students. Early third-year student.

Dr. Randy Clinch

Primary: My current rotation. Repair: A specific weakness exposed by my patients, required assessments, or question-bank work. Maintenance: A small amount of selected learning from previous rotations.

Maya Brooks

Fourth-year student still working toward a readiness benchmark.

Dr. Randy Clinch

Primary: Integrated board performance. Repair: The recurring bottleneck identified across q-bank data, practice assessments, Miss Logs, and other performance information. Maintenance: Previously repaired areas that need enough retrieval to stay available.

Maya Brooks

Same framework. Different plan.

Dr. Randy Clinch

That's the key. The system stays the same. The prescription changes.

Maya Brooks

Before we wrap up, I want to address something I hear from students constantly. “I have too many resources.” Maybe the school gives me one question bank. I've purchased another one. I have a video subscription. I have a review book. I have school lectures. I have podcasts. I have AI tools. And I've accumulated notes from three years of medical school. How do I fit all of this into my personalized plan?

Dr. Randy Clinch

Maybe you don't. Owning a resource does not create an obligation to use it. Instead ask: What job does this resource perform in my current plan? If you need precision content repair, choose a resource that efficiently performs that job. If you need targeted retrieval, choose the appropriate tool for that job. If you need discrimination practice, use something that lets you compare competing patterns. If you need integrated performance data, use mixed questions and appropriate practice assessments. If two resources are doing essentially the same job, you may not need both. And if a resource doesn't address your current primary learning, repair target, or maintenance needs, it may not deserve time this week—even if you paid for it. If resource overload sounds familiar, go back to **“Beat Resource Overload: Master Medicine with Fewer, Focused Tools.”** The goal isn't to finish your resources. The goal is to improve your performance.

Maya Brooks

And I think that brings us all the way back to the title of today's episode. Your board-prep plan should learn from you.

Dr. Randy Clinch

Yes. Because the longer you prepare, the more information your learning system should have about you. What you know. What you don't know. What you confuse. Where your reasoning breaks down. What happens under time pressure. Which repairs work. Which ones don't. And which former weaknesses no longer deserve to dominate your study time. So don't just collect data. Use it. Collect. Interpret. Prescribe. Test. Adapt. And remember: The system stays the same. The prescription changes.

Maya Brooks

So whether you're just beginning third-year rotations or you're a fourth-year student still working toward your school's requirements for Level 2 or Step 2, you can start with the same questions. What data do I already have? What pattern is that data showing me? Is the problem content, reasoning, mechanics, or some combination? What is the smallest useful intervention? How will I test whether it worked? And what should change in my plan because of what I learned?

Dr. Randy Clinch

You don't need to use every resource you own. You don't need to repair everything at once. And you don't need somebody else's study plan. Build a learning system. Let your performance tell you where that system needs attention. Then adjust.

Maya Brooks

Thanks so much for listening to the AI Med Tutor podcast. And remember, stay curious, and keep learning.