Chapter 36 — Exercises

How to use these. No answers are provided. Items marked † are harder or open-ended and are good candidates for discussion, a study group, or a written response of a page or more.

Several exercises use a [constructed teaching example] — an announcement written for this book, not a real one. They are labeled every time. Nothing in this file is medical advice, and none of it should be read as describing how to obtain or use any compound.

Everything here is dated to the chapter's frame: as of this writing, in 2026.

Part A — The framing problem (§36.1)

a. In one sentence, state why this chapter cannot rate a claim about a compound that has not completed its trials. Your answer should refer to what a rating requires, not to how hard prediction is.

b. Give one example, from your own reading in the last year, of a sentence about the future of medicine that could not be wrong. Rewrite it so that it could be.

c. The chapter claims that a hype-heavy future chapter would "retroactively discredit" the thirty-five chapters before it. Explain the argument. Then state one reason someone might disagree.

d. † Chapter 5 established that a rating attaches to a claim with a population and an endpoint. Construct a claim about a compound in development that is rateable today and one about the same compound that is not. What exactly is the difference between them?

e. Distinguish the two kinds of statement the chapter says it can honestly make. For each, give an example from anywhere in the chapter and say how a reader could check it.

Part B — The six questions (§36.2)

f. Write out the six questions from memory. Then rank them by how much time each takes to answer, and separately by how much each tells you. Where do the two rankings disagree, and why does the chapter say that matters?

g. For each phase — preclinical, 1, 2, 3 — write one sentence describing what a positive result at that stage does and does not license you to believe.

h. Explain phase 2 optimism in your own words without using the word "bias." Then explain why it is not evidence of dishonesty.

i. † The chapter declines to state a single number for the proportion of compounds entering human trials that reach approval. Give two reasons that refusal is defensible, and one reason a reader might find it frustrating.

j. A trial reports a statistically significant reduction in a blood marker. Name three separate things you still do not know about whether the drug helps anyone.

k. Explain the difference between beating placebo and beating standard of care. Then explain why Chapter 28's sacubitril/valsartan result is cited in this chapter rather than some other trial.

l. Rank these sources from most to least informative about a compound's efficacy, and justify the two adjacent pairs you found hardest to order: a conference abstract; a company press release with topline numbers; a peer-reviewed publication of the full trial; a quarterly earnings call slide; a news article summarizing the press release.

m. † "A mechanism story does not move the base rate." Defend this claim against the strongest objection you can construct — namely, that a well-understood mechanism really does make success more likely. Where does the objection have force, and where does it fail?

Part C — Applying the six questions to constructed announcements

Items n through s all refer to [constructed teaching example] text. These are invented for teaching and describe no real compound, company, or trial.

n. [constructed teaching example]

"BIOTECH COMPANY ANNOUNCES POSITIVE TOPLINE RESULTS FOR NOVEL PEPTIDE IN METABOLIC DISEASE. The company today announced that its lead candidate met its primary endpoint in a Phase 2 study, with a statistically significant improvement versus placebo. 'These results validate our mechanism and position us to advance rapidly,' said the chief executive. Full results will be presented at a future scientific meeting."

Run all six questions. For each, state whether the announcement answers it, and what the silence implies. Then write one sentence stating what you actually know about the compound.

o. Using the same announcement, list every phrase doing rhetorical rather than informational work. For each, say what a reader is likely to infer and why the inference is not licensed.

p. † [constructed teaching example]

"First-in-human data presented today showed the candidate was generally well tolerated across all dose cohorts, with pharmacokinetics supporting once-monthly dosing. Investigators noted encouraging trends in exploratory biomarkers."

This announcement is accurate and contains no false statement. Explain, question by question, why it nonetheless supports almost no conclusion about efficacy. Pay particular attention to the words "trends" and "exploratory."

q. [constructed teaching example]

"In a randomized trial of 4,200 adults with established disease, the candidate reduced the primary composite endpoint of cardiovascular death, myocardial infarction, and stroke by a statistically significant margin compared with an active comparator already recommended in guidelines. Results were published simultaneously in a peer-reviewed journal."

Run the six questions. Which are answered, and how strong is this claim relative to item n? Name the two features doing the most work.

r. Compare n, p, and q side by side in a table with the six questions as rows. Then write two sentences: one on what the table makes visible, and one on why press coverage of all three would likely use similar language.

s. † Write your own [constructed teaching example] press release that is fully compliant — one that answers all six questions honestly, in a paragraph. Then say why you think real releases rarely read this way, distinguishing reasons that are dishonest from reasons that are not.

Part D — Oral delivery and duration (§36.3, §36.4)

t. Explain, to someone who has read only Chapter 1, why swallowing a peptide is normally futile, and what SNAC does about it. Do not use the word "bioavailability" until you have defined it.

u. Oral semaglutide's bioavailability is on the order of 1 percent. Trace the consequences of that number through to (i) tablet content, (ii) manufacturing capacity, and (iii) supply. Then say which of those three a reader is most likely to encounter in press coverage, and which is most consequential.

v. State the chapter's conclusion about oral peptides in one sentence, then argue that it is not a defeat for peptide science. Use the captopril story from Chapter 35.

w. † Rank these three attributes of orforglipron by how much each would change access in a low-income country, and defend your ordering: no absorption enhancer; no cold chain; conventional small-molecule manufacturing cost.

x. List the three arguments the chapter gives against assuming longer duration is better. Which of the three is the one engineering cannot solve, and why not?

y. † A company announces a formulation dosed once every three months. Write five questions you would want answered before forming a view, at least two of which concern reversibility.

Part E — Multi-agonists and the muscle question (§36.5, §36.6)

z. State, in one sentence each, what tirzepatide, retatrutide, cagrilintide, and CagriSema are and what stage each is at as of this writing.

aa. Explain why "produces more weight loss than tirzepatide" and "prevents more heart attacks than tirzepatide" are different claims, and why the second takes so much longer to answer.

bb. † Reread the CagriSema framing in §36.5. Write two headlines about the same topline number — one for a financial audience and one for a clinical audience — that are both accurate. Then say what this exercise demonstrates about question 5.

cc. State what is established about lean mass loss during weight loss, in one sentence, and then state the four open questions the chapter identifies.

dd. Why is a DEXA reading a surrogate? Name three functional measurements that are not, and say what each one is actually capturing.

ee. † Section 36.6 calls lean mass on a scan an especially treacherous surrogate, because an agent that adds lean mass to the scan has mechanically improved the very thing being measured. Explain why that makes the surrogate worse than an ordinary one. Then name a case from elsewhere in medicine with the same structure — an intervention acting directly on the measurement used to judge it.

ff. Read the name bimagrumab using Chapter 1 §1.8. What class is it, and what follows for manufacturing and cost? Why does the chapter make a point of raising this in a book about peptides?

Part F — Conjugates, the barrier, supply, and forecasting (§36.7–§36.10)

gg. Explain how a peptide-drug conjugate inverts the book's usual framing of peptide properties. Name one property that is a liability everywhere else and an advantage here.

hh. Write the two separate ratings the chapter assigns in §36.7 and state, in one sentence each, why collapsing them into a single verdict would be misleading.

ii. † "It acts on the brain" and "it crosses the blood-brain barrier" are different claims. Describe the two anatomical routes that make this so, then find a piece of real coverage of GLP-1 drugs and assess whether it conflates them.

jj. Explain why a peptide shortage can persist even when there is drug substance in a warehouse. Name the second bottleneck and say why it surprises people.

kk. Separate these into two lists — "changes when patents expire" and "does not change when patents expire" — and justify the placement of each: price; the size of the treatment effect; who can afford treatment; the strength of the evidence; the gray market's main selling point; the population in which the drug was studied.

ll. † Write your own three lists in the format of §36.10 — would not surprise me, would surprise me, would bet against but cannot rule out — for a field you actually know something about. Date them. Then state, honestly, which list was hardest to fill and why.

mm. † The chapter argues that a forecast differs from a prediction by naming what would prove it wrong. Apply this test to three sources you currently rely on. What did you find, and did anything about your reading habits change?