From Learning to Performance: The Dedicated Pivot

Maya Brooks

Hello everyone, and welcome back to the AI Med Tutor Podcast. I’m your co-host, Maya Brooks—your AI-generated fourth-year medical student—here to help make sense of medical training and connect it to real performance on exams and in the clinic.

Dr. Randy Clinch

And I’m Dr. Randy Clinch, a DO family medicine physician and medical educator. Today we’re talking about the dedicated pivot: when and how to shift from learning mode to performance mode. A lot of students hear “do random timed blocks” and they think that means the strategy is simply volume—more random questions. That’s not a strategy. Random blocks are a tool, and tools only work when you use the data they generate to decide what to fix next. In this episode, we’ll show you how to use random sets correctly, when to use smaller focused sets, what to stop doing as you pivot, and how to maintain two lightweight learning tools—Pattern Cards and a Miss Log—so your performance actually improves. Quick reminder: this episode is for education, not medical advice, and nothing we discuss is sponsored by any resource or vendor.

Maya Brooks

This is so important because students can grind through hundreds of random questions and still feel like they’re not getting better.

Dr. Randy Clinch

Right—and if your score isn’t moving, it usually means your loop is incomplete. You’re testing, but you’re not repairing.

Maya Brooks

Let’s define the two modes. What do you mean by learning mode and performance mode?

Dr. Randy Clinch

Learning mode is when your primary goal is building understanding and pattern recognition. You’re okay going slower because you’re trying to learn mechanisms and reduce confusion between look-alikes. You may use more targeted sets, tutor mode, and deeper review. Performance mode is when your primary goal is executing under exam conditions—timed blocks, mixed questions, pacing, and decision-making under pressure. You still learn in performance mode, but your learning is driven by what breaks under time constraints.

Maya Brooks

So learning mode is “build the machine,” and performance mode is “race the machine.”

Dr. Randy Clinch

That’s a helpful way to frame it. Dedicated requires both. The mistake is living in only one mode.

Maya Brooks

Students always ask, “When should I pivot?” What’s your answer?

Dr. Randy Clinch

You pivot gradually, not overnight. Here are a few signals you’re ready to increase performance mode. One: you’ve built a baseline foundation in your major weak areas—enough that random questions aren’t pure chaos. Two: your misses are increasingly about reasoning choices, hinge clues, and mechanics, not “I’ve never heard of this.” Three: you need pacing and stamina practice because you’re running out of time or getting sloppy late in blocks. When those are true, you increase timed and random blocks—but you keep targeted repair as a smaller, strategic layer.

Maya Brooks

Now let’s address the myth directly: “If I just do a lot of random questions, I’ll improve.”

Dr. Randy Clinch

Random questions can improve you, but only if you use them as a diagnostic. If you do random sets and your only output is a percentage score, you’re leaving most of the value on the table. Random sets are supposed to answer: what patterns are leaking points under realistic conditions? What kind of misses are they—content, reasoning, or mechanics? What is recurring? What is risky? Then you use that data to plan the next day’s focused repair.

Maya Brooks

So random sets aren’t the full plan. They’re the measurement part of the plan.

Dr. Randy Clinch

Exactly right. Random sets generate the signal. Focused sets apply the fix.

Maya Brooks

How should students review random sets in a way that actually leads to improvement?

Dr. Randy Clinch

Here’s a simple review workflow for a random timed block. First, separate your questions into three piles: solid correct, unsure correct, and miss. Second, for every miss and every unsure correct, label the miss type: content, reasoning, or mechanics. Third, identify the repeats: what did you miss more than once in this block or across recent blocks? Fourth, choose one primary leak to repair today—just one. Then build a small targeted set to test that leak after you repair it.

Maya Brooks

That stops students from trying to fix ten things at once.

Dr. Randy Clinch

Right. One leak at a time, with proof.

Maya Brooks

Define content versus reasoning versus mechanics quickly.

Dr. Randy Clinch

Content is “I didn’t know or understand the concept or mechanism.” Reasoning is “I knew facts, but I chose the wrong story or missed the hinge clue.” Mechanics is “I answered the wrong task, missed a stability or urgency cue, or got pulled into an answer-choice trap.” Different miss types need different fixes.

Maya Brooks

Now we need to define Pattern Cards and a Miss Log for listeners who are new to the podcast.

Dr. Randy Clinch

Good call. A Pattern Card is a short way to capture a testable clinical pattern so you can recognize it again. It has three parts: presentation, key clues, and mechanism. You create Pattern Cards from high-value misses or shaky corrects—especially look-alike discriminations. If you want more detail, look back through prior episode titles for “Pattern Cards” and “illness scripts.” A Miss Log is your running list of the patterns you keep missing or keep getting right for shaky reasons, plus what you’re going to do about them—your fix and your retest plan. For more detail, look for prior episode titles that include “Miss Log” or “spaced repetition.”

Maya Brooks

So Pattern Cards help you remember patterns, and the Miss Log helps you decide what to re-test next.

Dr. Randy Clinch

Well said. They keep learning organized during performance mode.

Maya Brooks

How should the ratio of targeted or tutor questions change as students move deeper into dedicated?

Dr. Randy Clinch

Here’s a practical way to think about it. Early dedicated often includes more targeted learning: targeted sets, tutor mode, deeper review—because you’re building foundations. Mid dedicated is mixed: some timed random blocks, but still a strong repair layer. Late dedicated becomes more performance-heavy: more timed random, more mixed blocks, more stamina practice. But even late dedicated, you still need targeted micro-repair because random blocks keep revealing leaks.

Maya Brooks

So the pivot isn’t abandoning targeted work. It’s shrinking targeted work into a strategic layer.

Dr. Randy Clinch

That’s the key idea. Targeted is the tool you use to fix what random blocks reveal.

Maya Brooks

When should a student choose a smaller focused set, and what should it look like?

Dr. Randy Clinch

Use focused sets in three situations. First, you found a content gap in a repeated topic—like acid-base compensation or a pharm mechanism—and you need precision repair plus retest. Second, you found a reasoning gap between look-alikes—like heart failure versus COPD, nephritic versus nephrotic—and you need discrimination reps, ideally interleaved. Third, you found a mechanics leak—like missing stability cues or task identification—and you need a set of next-step questions to practice that micro-skill. A focused set should be small: often 5 to 15 questions. The goal is not endurance; the goal is to apply a fix and prove it worked.

Maya Brooks

So random builds realism and exposes leaks; focused sets fix the leaks efficiently.

Dr. Randy Clinch

Exactly right.

Maya Brooks

Give an example of using data from a random block to pick a focused set for tomorrow.

Dr. Randy Clinch

Let’s say you did a timed random block and your misses cluster around shortness of breath stems. When you triage, you realize it’s mostly reasoning: you’re mixing up heart failure and COPD. Your focused repair tomorrow is an interleaved mini-set: a few heart failure questions alternating with a few COPD questions. You create one Pattern Card for each pattern with presentation, key clues, and mechanism. Then you retest with another small mixed set later in the week to see if you improved under random conditions. If your question bank is running out of specific questions for you to use, check out out episode about how to use OpenEvidence.com to help generate questions.

Maya Brooks

And if the misses were mechanics, like missing instability cues?

Dr. Randy Clinch

Then your focused set is next-step questions where the stem includes urgency or instability. You practice one micro-rule: label the task, scan for stability cues while reading, then choose the safest high-yield move. You don’t need 50 in a row. You need enough reps to make the habit automatic, then you test it again in a random block.

Maya Brooks

What should students stop doing as they pivot into performance mode?

Dr. Randy Clinch

Here are high-value “stop doing” moves. Stop chasing new resources mid-dedicated. Stop turning review into rewriting a textbook. Stop taking random block scores personally—treat them as data. Stop spending five minutes on one question during a timed block. Stop rumination—don’t carry question ten into question thirteen. Your job in performance mode is clean execution with a smart repair layer.

Maya Brooks

That rumination point resonates with students because they feel it happening.

Dr. Randy Clinch

Right. Rumination eats working memory and steals accuracy from the question in front of you.

Maya Brooks

Can you give a simple weekly template that reflects the pivot?

Dr. Randy Clinch

Here’s a simple week. Two timed random blocks early in the week to generate data. After each block, identify one primary leak and build one focused repair set the next day—5 to 15 questions—plus one Pattern Card or one Miss Log entry for the highest-leverage issue. Midweek, do another timed random block to test if repairs are working. End of week, do a mixed block for stamina and integration. Throughout the week, spend a short, consistent amount of time reviewing your Pattern Cards and Miss Log so you don’t forget what you just worked on.

Maya Brooks

So you’re cycling: test realistically, repair precisely, then retest.

Dr. Randy Clinch

Yes—test, repair, retest. That loop is the engine.

Maya Brooks

Quick-start version. If a student wants to pivot starting tomorrow, what do they do?

Dr. Randy Clinch

Tomorrow, do one timed mixed block. Review it by sorting questions into solid correct, unsure correct, and miss. Label each miss as content, reasoning, or mechanics. Identify one repeating leak. Then do a small focused set—5 to 15 questions—that targets that leak, and capture one Pattern Card if it’s a pattern confusion or one Miss Log entry if it’s a recurring weakness you need to revisit. Then retest that skill in your next timed random block. That’s the pivot.

Maya Brooks

OK - Recap time.

Dr. Randy Clinch

Random timed blocks are valuable, but they are not a stand-alone strategy. Their job is to generate data—what breaks under realistic conditions. Your job is to turn that data into a fix: targeted content repair, discrimination training with Pattern Cards, or mechanics micro-rules. Then you prove the fix with a small focused set and a retest in another random block. Keep Pattern Cards and a Miss Log as lightweight tools to capture what matters and guide what you revisit. That is how you pivot from learning mode to performance mode without wasting your time.

Maya Brooks

That’s it for today’s episode of the AI Med Tutor Podcast. If you know someone in dedicated who’s doing tons of random questions but not improving, send them this episode.

Dr. Randy Clinch

And remember: the strategy isn’t “do more random questions.” The strategy is “use random questions to find leaks, then fix them efficiently.”

Maya Brooks

We’ll see you next week everyone! And in the meantime—stay curious and keep learning!