Discussion Guide — Chapter 38

Prompt 1 — Open with the number. Insulin is 51 amino acids and is a biological product. Semaglutide's backbone is 31 and it is a drug. Nothing chemical happens at residue 41. Is drawing the line at 40 defensible?

Listen for: students who defend bright-line rules on administrability grounds (predictability, low litigation cost, no case-by-case discretion) versus students who attack the arbitrariness. The sophisticated move — reward it — is separating the effects of drawing a line somewhere from the effects of drawing it at 40. Watch for the assumption that a "scientific" line would be better; press on what a non-arbitrary line would even look like given that the peptide/protein boundary in Chapter 1 is also a convention.

Prompt 2 — The taxonomy under pressure. Someone tells you five compounds are "not FDA approved." Does that sentence carry any information at all?

Listen for: whether students reach for the five cases unprompted. The answer you want is that the sentence carries information only once you know which case applies, and that four of the five cases say nothing about the molecule. Push back if a student says the phrase is "meaningless" — it is not; it is ambiguous across five states, which is a different and more useful diagnosis. If nobody raises case (2), supply it: there is one situation where non-approval genuinely is a negative signal, and refusing to acknowledge it makes the whole taxonomy look like special pleading.

Prompt 3 — The independence pair. BPC-157 is rated ❌ and is prohibited under WADA's S0. An athlete says: "They wouldn't ban it if it didn't work." What is wrong with that?

Listen for: whether students locate the error in the text of the rule rather than in the athlete's character. The strong answer quotes S0's construction — a rule about a property, containing no efficacy clause — and then makes the inversion explicit: S0 captures compounds precisely because nobody knows whether they work. Also listen for the second-order observation that under S0's logic, approval makes a compound less prohibited, which is counterintuitive enough to be worth surfacing.

Prompt 4 — Nesiritide, and whether approval can be wrong. Approved in 2001 on a symptom endpoint; a 7,000-patient trial in 2011 found no benefit on the endpoints that mattered most. Was the 2001 approval a mistake?

Listen for: resistance to the binary. The best answers argue that the decision can have been reasonable on the evidence available and that the episode is a genuine cautionary example, and then address the real question — which error a regulator should prefer to risk, and for which patients. Push students who say "they should have waited for outcome data" to price the delay in a seriously ill population. Push students who defend the approval to say what would have made them uncomfortable.

Prompt 5 — The suppression argument, steelmanned. "There are no trials of these peptides because nobody can make money on them." Is that true? And what does it license?

Listen for: two distinct moves. First, whether students can steelman it properly — the funding argument is largely correct and is a real defect in how medical knowledge is produced. Second, and this is the one that separates the class, whether they can then identify the inference it does not license: that any particular unstudied compound works. The same economics predict that useless compounds go unstudied at the same rate. If a student gets that far, ask them what evidence would distinguish the two, and watch them reinvent Chapter 5.

Prompt 6 — The column discipline. Why does this chapter insist that Field 10 be recorded in a column entirely separate from the evidence rating? Isn't that just bookkeeping?

Listen for: recognition that physical separation is what makes the discipline enforceable — that blending the two in one cell is how the error happens, quietly, without anyone deciding to make it. Then push to the self-audit: ask which direction each student would slip in. Students who trust institutions will soften ratings for approved compounds; students who don't will harden ratings for prohibited ones. Most students can name someone else's bias instantly and their own not at all, which is exactly the finding Chapter 40 asks them to have ready.