Final Exam — Cumulative

Sixty points. Two hours. Closed book unless your instructor states otherwise; the four-line rating format may be reproduced on the board, and simple arithmetic is expected (a calculator is permitted and unnecessary).

Cumulative across Chapters 1–44, weighted to Parts IV–VIII. The method from Part I is assumed throughout and is tested by being used rather than by being recited.

Every invented example on this exam is labeled [constructed teaching example]. Nothing labeled that way is real. The constructed trial in Section D is arithmetic built for this exam and describes no actual study.

Section A — Multiple choice (15 points, 1 each)

A1. Endogenous opioid peptides such as the endorphins and enkephalins are best described as: a) Synthetic analogs of plant alkaloids. b) Peptides produced by the body that act at opioid receptors. c) Neurotransmitters unrelated to the receptors targeted by opioid drugs. d) Hormones secreted exclusively by the adrenal gland.

A2. Intranasal oxytocin has generated an extensive literature and comparatively modest clinical results. Under the rules, the correct response to "but the mechanism is compelling" is: a) Upgrade to ⚠️ on mechanism. b) Downgrade to ❌ because the field is overhyped. c) Hold the rating where the outcome evidence puts it; mechanism licenses a hypothesis, not a rating. d) Decline to rate the claim at all.

A3. Substance P antagonists engaged their molecular target as designed and did not deliver the expected analgesic benefit. The lesson the book draws is: a) The mechanism was wrong. b) The trials were too small. c) Confirmed target engagement does not imply clinical benefit. d) Pain is not mediated by peptides.

A4. A peptide approved for a specific sexual-desire indication in a defined population, and discussed far more broadly than that indication, is a case of: a) Rule 6 — one molecule, many ratings, with an approved claim and unapproved claims held separately. b) Rule 4 — downgrading with distaste. c) A regulatory error. d) A compound with no evidence of any kind.

A5. Antimicrobial peptides are widespread in innate immunity, and translating them into systemic human drugs has been difficult. The most useful thing to record about this in a dossier is: a) The origin story. b) In field 5, what is conspicuously absent — and in field 4, the delivery constraint. c) Field 2, since natural origin supports the claim. d) Nothing; the class is not worth tracking.

A6. B-type natriuretic peptide is used clinically most prominently as: a) A treatment for hypertension. b) A diagnostic biomarker in the assessment of heart failure. c) An appetite suppressant. d) A vaccine adjuvant.

A7. Cosmetic peptides face a delivery problem that is specific to their route. It is that: a) They are destroyed by sunlight. b) The stratum corneum is a substantial barrier to large, hydrophilic molecules. c) They cannot be manufactured at cosmetic scale. d) They are banned in most jurisdictions.

A8. Solid-phase peptide synthesis works by: a) Extracting peptides from animal tissue. b) Assembling the chain stepwise on an insoluble support, then cleaving it off. c) Fermenting engineered bacteria. d) Cutting larger proteins with proteases.

A9. A common strategy for extending a peptide drug's duration of action is: a) Increasing the dose only. b) Attaching a fatty-acid chain that promotes albumin binding, and blocking the site where a degrading enzyme cuts. c) Reducing the number of amino acids. d) Switching from injection to inhalation.

A10. A certificate of analysis accompanying a vial establishes, at most: a) That the vial in your hand contains what the label says. b) Something about a sample that was tested — not necessarily about your vial. c) That the compound is approved. d) That the compound is effective.

A11. Phage display and related library methods are techniques for: a) Manufacturing peptides at scale. b) Discovering peptides that bind a chosen target. c) Measuring purity. d) Extending half-life.

A12. Chapter 38 establishes that "not approved" has five distinct meanings. Which pair are opposite signals wearing the same words? a) Approved elsewhere only / used off-label. b) Never submitted / submitted and rejected. c) Not regulated as a drug / off-label. d) Off-label / approved elsewhere only.

A13. In the Chapter 39 conversation, the most important reason not to respond dismissively to a patient's disclosure is that: a) It is impolite. b) It changes what the patient takes. c) It teaches concealment, and a patient who conceals is in more danger, not less. d) It exposes the clinician to liability.

A14. The book's account of why a compound's ratings drift upward over time, and almost never downward, is that: a) The evidence genuinely improves for most compounds. b) News about a compound you are watching is selected for being newsworthy, and negative results are less newsworthy. c) Regulators progressively approve more indications. d) Readers become better at evaluation and find more support.

A15. A drug is approved in one jurisdiction for one indication in one population. Under the rules, the approval: a) Supports a ✅ for any use of that molecule. b) Is a judgment about a specific claim, and adjusts no rating for any other claim. c) Downgrades all off-label claims to ❌. d) Belongs in field 6 alongside the rating.

Section B — Short answer (21 points, 3 each)

Two to four sentences each.

B1. A peptide is approved for one oncology indication and is widely discussed for a second, untested one. Write the two field-6 rows this produces, each with a claim, a rating, and a date placeholder. State why they do not contradict each other.

B2. Chapter 34 draws a distinction between what analysis can establish about a sample and what a buyer wants to know about a vial. Explain the gap, and name the dossier field where it belongs.

B3. Explain why an origin story — "found in a natural source," "an exact fragment of a human protein" — carries no evidentiary weight. Give one example in each direction from the book.

B4. Chapter 39 asks readers to convert a verdict into a question a clinician could answer. Convert this verdict: "There isn't enough evidence for this." Then say what the conversion tests.

B5. Chapter 12 and Chapter 44 both discuss access and cost. Explain why a supply, price, or access fact belongs in field 10 and must never adjust field 6, and give the failure that occurs if the two are allowed to interact.

B6. Chapter 41 describes a brand name entering ordinary language as a common noun. Name one way this linguistic shift makes evaluation harder, in terms of Rule 1.

B7. You wrote field 12 for a compound in week three. It is now the end of the term and your rating has moved from ⚠️ to ✅. State the one check you owe yourself before accepting that upgrade, and name what you are guarding against.

Section C — The two kinds of ❌ (10 points)

Consider two compounds, both of which could appear at ❌ in a summary table:

  • Compound 1. A peptide that reached a large, well-conducted randomized outcome trial for a cardiovascular indication and did not improve the clinical outcome. It is little used today.
  • Compound 2. A peptide with substantial animal work and enthusiastic informal use for tissue healing, which has never been through an adequate human outcome trial.

C1. (2 points) Tag each ❌ with its kind.

C2. (3 points) Which compound do we know more about, and in what sense? Be precise about what "more" means here.

C3. (3 points) State what would move each rating, and explain why the two answers are different in kind, not merely in degree.

C4. (2 points) A patient asks about each compound. Give the one-sentence honest answer for each — the sentence you would actually say — and make sure the two sentences are not interchangeable.

Section D — Relative to absolute (8 points)

[constructed teaching example] — the trial below is invented for this exam. The arithmetic is real; the study is not. Do not repeat these figures as though they described any actual trial.

Press release headline: "NEW ANALYSIS: TREATMENT CUTS MAJOR EVENTS BY 25 PERCENT"

From the (constructed) results table:

Arm Participants Participants with a major event over 3 years
Placebo 5,000 300
Treatment 5,000 225

D1. (2 points) Calculate the event rate in each arm, as a percentage.

D2. (2 points) Calculate the relative risk reduction. Show that the headline is arithmetically correct.

D3. (2 points) Calculate the absolute risk reduction, in percentage points, and the approximate number needed to treat over three years.

D4. (2 points) The headline is not false. Explain, in two sentences, what it nevertheless leaves a reader believing, and state the standing course rule about reporting relative figures.

Section E — Synthesis (6 points)

Answer in roughly one page.

The book's closing position is that you will forget most of its content and that this is fine, because what should survive is a reflex rather than a set of verdicts.

State that reflex as a set of questions, and then defend or challenge the book's claim that it transfers beyond peptides. Use at least two specific cases from Parts IV–VIII to support your position. A well-argued challenge is worth full credit.

Answer key ### Section A (15 points) | Item | Answer | Note | |---|---|---| | **A1** | **b** | Endogenous peptides acting at the same receptor family targeted by opioid drugs (Chapter 20). | | **A2** | **c** | Rule 3. Note that (b) is the Rule 4 error and should be marked wrong just as firmly. | | **A3** | **c** | Chapter 22, the book's cleanest demonstration that mechanism does not license a rating. | | **A4** | **a** | Chapter 24. The approved claim and the broader informal claims are separate rows with separate evidence. | | **A5** | **b** | Chapter 25. The interesting facts are the missing human outcome data and the delivery constraint, not the natural origin. | | **A6** | **b** | Chapter 28. Diagnostic use, not therapeutic — a distinction students frequently blur. | | **A7** | **b** | Chapter 30. Route is the whole argument for cosmetic peptides; field 4 is where it lives. | | **A8** | **b** | Chapter 32. Stepwise assembly on an insoluble support, then cleavage. | | **A9** | **b** | Chapter 33. Both halves are needed for full credit in a short-answer variant of this item. | | **A10** | **b** | Chapter 34. A certificate describes a sample, not a vial. | | **A11** | **b** | Chapter 35. Discovery, not manufacture. | | **A12** | **b** | Chapter 38. "Never submitted" and "submitted and rejected" are opposite signals in identical words. | | **A13** | **c** | Chapter 39. (b) is wrong and is the reason the error persists — dismissal does *not* change behavior, only disclosure. | | **A14** | **b** | Chapter 6 and Appendix C §C.6. Selection in the information stream, not a change in reasoning quality. | | **A15** | **b** | Rule 1 plus Chapter 38. (d) is a tempting distractor and wrong: regulatory status is field 10, never field 6. | ### Section B (21 points, 3 each) **B1.** Two rows, each with claim + rating + date, e.g.: 
[Approved indication, in the population named on the label]   ✅  [date]
[The widely discussed untested indication]                    ❌  [date]
                                                (kind: evidence ABSENT)
Full credit requires the explicit statement that the rows **do not contradict** because Rule 1 attaches a rating to a claim and Rule 6 permits many per molecule — they are answers to different questions. Deduct 1 if the untested row is left untagged for the kind of ❌. **B2.** Analysis establishes identity and purity **of the material that was tested**. The buyer wants to know about the vial in their hand, which was not the tested material and may differ in content, concentration, degradation, or contamination. The gap is therefore between a **sample** and a **vial**, and it belongs in **field 10** (preparation quality — what could be established, and what cannot), with the consequences appearing in **field 9**. Accept either field for full credit if the reasoning is correct; accept field 1 only if the student frames it as the "is the same name reliably the same molecule?" line. **B3.** Origin describes where a molecule came from, which is a historical fact with no bearing on whether a claim about it holds. Examples in the two directions, both from the book: **exenatide was found in the venom of a lizard and is well supported**, and **several compounds the book rates ❌ are exact fragments of human proteins**. Full credit requires both directions — students who give only the first have shown that a strange origin is compatible with good evidence but not that a reassuring origin is compatible with none, and the second is the one marketing exploits. **B4.** A satisfactory conversion: *"What would you want to see before you'd consider this?"* — or *"Is there anything with better evidence for what I'm actually trying to fix?"* The conversion tests whether the verdict was a conclusion or a feeling: **a verdict that cannot be turned into an answerable question was probably a feeling.** It also changes the conversation from one the clinician must rebut into one they can answer. **B5.** Access, price, supply, and legality are facts about a market and a regulatory environment; a rating is a statement about evidence for a claim. Letting them interact produces the error in both directions: a compound becomes ✅ because it is available and endorsed, or ❌ because it is expensive, restricted, or sold badly. **The failure is that the rating stops tracking the evidence and starts tracking the market**, at which point it can be moved by a price change or a shortage — events that contain no information about whether the drug works. **B6.** A brand name used as a common noun **absorbs distinct claims into one word.** Once the name means a category, a drug, a dose, a look, and an outcome all at once, a sentence using it has no determinate population and no determinate endpoint — so Rule 1's requirement cannot even be applied without first asking the speaker what they meant. Accept any answer that lands on the collapse of specificity; do not require the student to use the phrase "common noun." **B7.** Check the current rating against **what you wrote in field 12 in week three** — specifically, whether the finding you named as the thing that would move it has actually occurred. Not against the latest headline. You are guarding against **rating drift**: the upgrade may have been driven by an information stream selected for good news rather than by the readout you specified. Full credit requires naming the drift mechanism, not merely saying "be careful." ### Section C — The two kinds of ❌ (10 points) **C1 (2 points).** Compound 1: **❌ evidence present and negative.** Compound 2: **❌ evidence absent.** 1 point each. **C2 (3 points).** **We know more about Compound 1.** "More" means: for Compound 1 we have a tested answer — an adequate trial, in a defined population, on a defined endpoint, produced a negative result, which is a *finding*. For Compound 2 we have no answer at all; the ❌ records **ignorance**, not refutation. The apparent paradox — that the compound with the *worse* result is the one better understood — is the point of the distinction. Award 1 point for the correct direction, 2 more for a precise account of what "more" means. **A student who says the two are equivalent, or that Compound 2 is better understood, receives 0 for this part regardless of the prose.** **C3 (3 points).** Compound 1 moves on **replication or reframing**: a trial in a different population or on a different endpoint, or a credible reanalysis showing the original was misdirected — the bar is high because a real answer already exists and must be overturned or bounded. Compound 2 moves on **the first adequate trial**: there is nothing to overturn, only a vacancy to fill. **The difference is in kind: one requires displacing knowledge, the other requires producing it for the first time.** Full credit requires that last distinction stated explicitly. **C4 (2 points).** Two sentences that are not interchangeable, e.g.: - Compound 1: *"This one was properly tested for that use, and it didn't improve the outcome — that's a real answer, not a gap."* - Compound 2: *"Nobody has run the trial that would tell us — so I can't tell you it works, and I also can't tell you it doesn't."* 1 point each. **Deduct both points if the two sentences would work equally well for either compound** — that interchangeability is exactly the failure the item tests. ### Section D — Relative to absolute (8 points) **D1 (2 points).** Placebo: 300 / 5,000 = **6.0%**. Treatment: 225 / 5,000 = **4.5%**. 1 point each. **D2 (2 points).** Relative risk reduction = (6.0 − 4.5) / 6.0 = 1.5 / 6.0 = **0.25 = 25%**. The headline is arithmetically correct. Award 1 point for the calculation and 1 for explicitly confirming the headline is not false. **D3 (2 points).** Absolute risk reduction = 6.0% − 4.5% = **1.5 percentage points**. Number needed to treat = 1 / 0.015 ≈ **67 people treated for three years to prevent one event**. 1 point each. Accept 66–67. **Deduct 1 for reporting the ARR as "1.5%" without the words "percentage points"** if your course has enforced that convention; note it in feedback either way, because the ambiguity between "1.5%" as a difference and as a rate is precisely what the item is about. **D4 (2 points).** The headline leaves a reader believing that a quarter of the people who take the treatment are spared an event, when the correct reading is that roughly 1.5 in 100 are — the same finding, a very different expectation. Both numbers describe one result; they differ in denominator, and the denominator is what the headline omits. The standing rule: **any relative figure must be accompanied by the absolute figure, or by an explicit statement that the absolute figure was not reported.** Award 1 point for the substance and 1 for the rule. Accept students who additionally observe that **neither figure alone tells you whether 1.5 percentage points over three years is worth it** — that judgment requires the endpoint's severity, the harms, the duration, and the person, and the book's position is that this is the patient's judgment to make with the numbers in hand rather than the headline's judgment to make for them. ### Section E — Synthesis (6 points) **The reflex, stated as questions** (2 points). Expect a recognizable version of: *For whom? Measured how? Compared to what? How would we know if it were false?* Accept close variants; require at least three of the four, and require that "how would we know if it were false" (field 12) appears in some form — without it the student has described a summarizing habit rather than a falsifiability reflex. **Two specific cases from Parts IV–VIII** (2 points, 1 each). Any accurate use. Strong choices include Chapter 22 (mechanism confirmed, benefit absent), Chapter 28 (a peptide better known as a diagnostic than as a therapy, and a failed outcome trial), Chapter 30 (route as the whole argument), Chapter 34 (certificate versus vial), Chapter 38 (the five meanings of "not approved"), Chapter 41 (a brand name collapsing distinct claims into one word). Award the point for correct and load-bearing use, not for mention. **Position defended or challenged** (2 points). **A well-argued challenge earns full credit.** Genuinely strong challenges available to students include: that the reflex is expensive in time and most people correctly outsource these judgments to institutions; that it produces reflexive contrarianism as a failure mode (see the participation rubric); that domains without randomized trials — nutrition, exercise, most of policy — cannot supply the evidence the reflex demands, so the honest output is permanent suspension of judgment, which is not usable; or that the reflex assumes access to primary literature that most readers do not have. Reward any of these when supported. **Do not reward agreement that merely restates the book.** **A grading note.** This section is where the course's stated principle is most visible to students, so apply it visibly: *a well-reasoned position the instructor disagrees with outscores an agreeable restatement.* Several students will hedge, expecting the book's position to be the right answer. Mark those at 3–4 and say why in one line.

Related: Chapters 1–44 · Appendix A · Appendix C · Appendix D · Appendix G · Appendix H · Appendix M · Where Students Get Stuck (struggles 3, 5, 8, 10) · Midterm Exam · Practical Evaluation Exam · Claim Evaluation Rubric