Discussion Guide — Chapter 19

Six prompts, with what to listen for. Each is designed to run 10–15 minutes. Prompts 2 and 5 are the ones most likely to produce genuine disagreement; prompt 6 is the one to close on.

Prompt 1 — "Each element is individually defensible. So who is responsible?"

Put the laundering diagram on the board and ask the class to assign responsibility for the resulting false claim.

Listen for: students trying to find the guilty party and discovering there isn't one. That discovery is the objective. The strong move is recognizing that the harm is emergent — produced by the structure rather than by any participant — which is precisely why individual good faith is not a defense and why "everyone here seems honest" is weak evidence about a market. Watch for the weaker move of settling on the vendor because they took the money; ask that student whether the claim would disappear if the vendor did.

Prompt 2 — "Vial B is real peptide with an honest certificate. What exactly is wrong with it?"

This is the sharpest question in the chapter and it should produce argument.

Listen for: the distinction between fake and unqualified. Many students will initially say nothing is wrong with vial B, and they are half right — that is the productive starting position. Push toward the six questions: identity and purity may be documented, and content, sterility, endotoxin, and residuals almost certainly are not. The best answers arrive at: nobody ever asked whether this is fit to put in a person, and honesty about what was tested is not the same as having tested the right things. Watch for a student who says "so it's fine if you just get one tested" — that is the moment to introduce the sampling problem and the fact that testing does not tell you what it will do in you.

Prompt 3 — "Why is 'nobody has reported problems' almost useless here, when it would be meaningful for an approved drug?"

Listen for: the distinction between harm not occurring and harm not being collected, and the recognition that these are indistinguishable from the outside. Strong students will connect it to Chapter 10 CS2 and notice that the asymmetry they learned — observational data is strong for harms — has a hidden precondition. The best answers go further and note that the specific failure modes here (endotoxin reactions that look like flu, wrong potency that looks like non-response, immunogenicity that looks like tolerance) would not generate attributable reports even if a system existed. If the room gets there, the cadaver hormone counterfactual lands without you having to argue for it.

Prompt 4 — "The compounded semaglutide dosing errors involved no wrong molecule and no dishonest seller. Where did the risk come from?"

Listen for: the word format. Students are trained to look for a bad actor or a bad substance, and this case has neither. The insight is that an engineered delivery device is part of what was studied, and substituting it silently reintroduces an error the system had designed out. Push for generalization: what else in medicine or engineering carries risk in the format rather than the content? Watch for students who conclude "so compounding is bad" — redirect to the actual conclusion, which is that evidence attaches to a specific product delivered a specific way.

Prompt 5 — "Is the critique that regulation protects profits rather than people a good argument?"

Run this one honestly. It is the prompt where you can lose the room or win it.

Listen for: students who can hold both halves. The critique is often fair — development economics really do shape what gets studied, drug pricing really is hard to defend, the commercial interest in shutting down compounded competition really was asserted. And none of it transfers to a container. No company's profit is protected by endotoxin or by a content assay nobody ran. The strongest contributions will distinguish policy arguments, where the critique has force, from physical facts about a vial, where it has none. If a student is angry about drug pricing, agree with them — Chapter 12 agrees with them — and then ask what that anger establishes about what is in the bottle. Do not win this exchange by dismissing the grievance; the grievance is real, and dismissing it is how you lose the students the chapter exists for.

Prompt 6 — "What would you actually say to a friend who tells you they've been using something from an unregulated source for four months?"

Close on this. Have students draft two or three sentences and read a few aloud.

Listen for: tone before content. The instinct to lecture is nearly universal and nearly always counterproductive — ask the class to predict what happens after the lecture, and someone will say the friend stops mentioning it. That is the whole lesson. Strong answers ask a question rather than deliver a verdict, and land on something specific from §19.7 or §19.8: that nobody has a baseline, or that a new symptom will have no differential, or that their doctor's list is missing an entry. The best answers include the honest admission that they are not in a position to tell their friend to stop, and don't try to.

End the session on the chapter's own closing sentence: the worst outcome is not a bad vial; it is a bad vial in a person whose doctor does not know it exists.