Discussion Guide — Chapter 10

1. "A drug that works on heart, kidney, liver, brain, and addiction. What's your first reaction?"

Open cold. Listen for: both reactions — "that's a red flag" and "that must be fundamental." Get both on the board. Then point out that the room has just produced both compression errors before reading a word of evidence, and that the chapter is about neither.

2. "The trial stopped early because it was working. Is the effect bigger or smaller than reported?"

Listen for: the intuition that early stopping means an especially strong effect. Run the simulation. The moment when someone says "oh — it stopped at the luckiest look" is the one to build on.

3. "Sleep apnea probably works by making people smaller. Does that count?"

Listen for: students separating "does it work" from "how does it work." Then push: when would a mechanical explanation be a criticism? (When the claim as marketed implies something the mechanism doesn't support — a specific airway action rather than generic weight loss.) This distinction transfers directly to Chapter 30.

4. "Millions of patients in the database. Two thousand in the trial. Which do you believe?"

The most productive disagreement in the chapter. Listen for: students defending the database on size grounds, then someone raising selection. Then ask the sharper version: for what question is the database better? (Harms — especially rare, distinctive ones.) Getting to "the same dataset is strong evidence about one thing and weak about another" is the target.

5. "Is rating one drug ten different ways useful, or unusable?"

Genuinely open, and important for clinical students. Listen for: the recognition that a patient does not need ten ratings — they need the one for their situation. Then ask what a clinician gives a patient instead, and whether that summary can be honest without being ten sentences long. This is a direct rehearsal for Chapter 39.

6. "Which of these indications would you bet against?"

Close here. Have them write it down and date it. Listen for: reasoning rather than the pick. A student who says "MASH, because the endpoint is a biopsy surrogate and histology-to-outcome translation has failed before" has done the work regardless of whether the bet pays.