Instructor Notes — Chapter 8

Teaching notes

What this chapter is actually for

It looks like a drug monograph. It is actually Chapter 5 applied at full scale to the drug students came for, and the three deliverables are:

1. The population lesson — STEP 1 versus STEP 2. Students must never again quote a trial figure without its population.
2. The relative/absolute discipline — SELECT, worked in both framings with an NNT.
3. The chronic-therapy reframe — weight returns, this is not a failure, and the expectation that it should be permanent comes from the moral framing of weight rather than from pharmacology.

If students leave able to recite that semaglutide produces 15% weight loss, the chapter has failed. If they leave asking "in whom, over what period, and which analysis?" reflexively, it has succeeded.

Common misconceptions

"Ozempic and Wegovy are different drugs." Same molecule, different dose and approved indication. Nearly universal, and it matters enormously for insurance (Ch 12).

"15% is what people lose." The population-swap error, and it will be in the room. Figure 8.3 exists for it. Ask which trial the figure came from before letting the discussion proceed.

"20% reduction in heart attacks is huge." Have students convert it. The moment they compute ~1.5 percentage points and an NNT of roughly 67 is the moment Chapter 5 §5.8 becomes real rather than abstract.

"There's a black box warning for thyroid cancer, so it causes cancer." And the opposite — "that's just a rat thing." Both are in the room and both are wrong. §8.7's Hype Check gives the accurate version, and exercise 8.13 makes students produce it under a word limit, which forces precision.

"The weight comes back, so it doesn't work." The chapter's central correction. The comparison to antihypertensives is the tool. Then push further: ask why nobody says statins don't work.

"The side effects are being hidden." Have students open the label's adverse reactions table. Everything is there, with rates, next to a placebo column. The information is not hidden; it is unread.

The hardest point to teach

That an excellent drug can have unresolved mechanism, and that this is normal.

Students expect that a well-evidenced drug is a well-understood one. SELECT establishes a cardiovascular benefit and nobody can say why — six candidate mechanisms, none settled.

Students find this destabilizing, and the destabilization is productive. Connect it to Chapter 2: aspirin was used for seventy years before anyone knew how it worked; insulin for thirty. Clinical efficacy and mechanistic understanding are independent variables, and a field that required mechanistic certainty before use would have almost no drugs.

Then invert it: several compounds in Part III have crisp, satisfying, fully worked-out mechanisms and no evidence of clinical benefit. Understanding and evidence come apart in both directions.

Demonstrations that work

Read the label, live. Project the Wegovy label on DailyMed. Read Section 1 (Indications) aloud — students are surprised by how narrow it is, and by the "as an adjunct to a reduced-calorie diet and increased physical activity" clause, which is the trial protocol written into the label. Then Section 6 (Adverse Reactions) and its placebo column. Ten minutes, and it recalibrates everything.

The conversion, on the board. SELECT: 20% relative. Baseline 8%. Compute 6.4%, absolute 1.6, NNT ≈ 62. (The chapter uses ~1.5 points and ~65–70 from the canon figures; either is fine and the discrepancy is itself worth noting — it is rounding, and students should see how easily a derived figure drifts.)

Two headlines. Have students write the maximally impressive and maximally underwhelming accurate headlines for SELECT, then the honest one, then count words. The realization that the honest version cannot be a headline connects Chapter 6 to a drug they care about.

The dossier entry, side by side. Have students bring their Chapter 8 twelve-field entry and compare it to the model. The Field 5 divergence is the teaching moment.

Timing

Two sessions is better. If you have one 75-minute session:

Cut §8.2, §8.4, §8.6, and §8.8 to assigned reading if you must. Do not cut §8.3, §8.5, or §8.9.

Assessment notes

Discriminating items: 8.8, 8.11, 8.13, 8.23, 8.28.

8.13 (the thyroid warning in under 100 words) is the best single item in the chapter — it requires holding two positions simultaneously under a word limit, which is exactly the skill.

8.23 (is cost-driven discontinuation a treatment failure?) has no clean answer and should be graded on the quality of the distinctions drawn. Watch for answers that frame regain as personal failure; that is the error the chapter is written to prevent, and it should be marked as such.

8.30 (the twelve-field entry) is the major written assignment for Part II. Grade Field 5's "what does not exist" line most carefully — it is where effort and evidence-base are distinguishable.