Chapter 44 — Further Reading

A caution specific to this chapter. Almost everything published on the societal effects of metabolic drugs is one of three things: a market projection produced to be persuasive, an opinion essay with no falsifier, or a genuinely careful analysis of a narrow question that gets quoted far beyond what it established. The reading below is weighted toward foundational works that explain the mechanisms, because those age well, and away from current commentary, which does not.

If a source contains a confident number about the future, treat that number as the thing to investigate rather than the thing you learned.

Tier 1 — Start here

These are the works that give you the concepts the chapter runs on. None is about GLP-1 drugs, which is the point: the mechanisms predate the molecules.

Erving Goffman, Stigma: Notes on the Management of Spoiled Identity (1963). The founding text. Short, readable, and the source of the vocabulary everything since has used. Read it for the distinction between discredited and discreditable conditions — which maps directly onto §44.4's question about what a visible body communicates.

Susan Sontag, Illness as Metaphor (1978), and AIDS and Its Metaphors (1989). The epigraph of this chapter comes from the first. Sontag's argument — that the metaphors we attach to a disease do real damage to the people who have it, and that the damage is worst where causation is poorly understood — is the intellectual ancestor of §44.4. The second book is where she works through what happens when a stigmatized condition becomes treatable, which is exactly Case Study 2's problem.

Peter Conrad, The Medicalization of Society (2007). The standard treatment of §44.5. Conrad is neither a cheerleader nor a critic of medicalization; he traces how conditions move into and out of medical jurisdiction and who gains from each move. Read the chapters on the shift of the engine of medicalization from the medical profession toward markets and consumers — that shift is the whole of §44.5's third cost.

Bruce Link and Jo Phelan, "Conceptualizing Stigma," Annual Review of Sociology (2001). If you read one journal article from this list, read this one. It decomposes stigma into components — labeling, stereotyping, separation, status loss, discrimination — and the decomposition is what makes it possible to ask which component an effective treatment would move. Most public argument about stigma fails because it treats stigma as a single quantity.

Tier 2 — Go deeper

Michael Bliss, The Discovery of Insulin (1982). The definitive history behind Case Study 1, including the patent decision and the intentions behind it. Read it and then read anything about insulin pricing a century later; the gap between the two is the most instructive hundred years in peptide medicine.

The attribution-theory literature on stigma, beginning with the experimental work of the 1980s demonstrating that perceived controllability predicts helping behavior, anger, and pity toward people with a range of conditions. This is the empirical foundation for the claim in §44.4 that both stigma mechanisms run through the same variable. Search terms: attributional analysis of reactions to stigmas, controllability attributions health stigma.

The health economics literature on budget impact versus cost-effectiveness. These are different analyses answering different questions, and conflating them produces most of the confused public argument described in §44.3. Any health-economics methods textbook covers the distinction; the useful exercise is to find one budget-impact analysis and one cost-effectiveness analysis of the same therapy and identify why they reach different-sounding conclusions.

Marion Nestle, Food Politics (2002), and Michael Moss, Salt Sugar Fat (2013). Background for §44.2 and for Chapter 12's food-environment position. Both are about how the industry responds to demand and to pressure, which is precisely the mechanism §44.2 asks you to reason about. Note that both books predate this drug class entirely — read them for the response function, not for predictions.

Work on the diffusion of medical innovation and equity. The general finding that new effective technologies are adopted first by better-resourced populations, sometimes widening health gaps before narrowing them, is well established across many technologies. Search terms: inverse equity hypothesis, intervention-generated inequality. This literature is §44.6's foundation and it is older and better developed than most readers expect.

Regulatory and health-technology assessment documents for this drug class in your own country. Dry, public, and far more informative than any commentary about them. These are the documents that actually determine coverage, and reading one teaches you more about §44.3 than a dozen articles about it.

Tier 3 — For the committed

Primary literature on real-world persistence and discontinuation. Conditions 1 in both §44.7 and §44.8 turn entirely on this, and it is one of the few questions in the chapter where informative evidence is arriving now. Search terms: GLP-1 receptor agonist persistence, real-world discontinuation, treatment persistence obesity pharmacotherapy. Read for how "discontinuation" is defined; the definition drives the number more than the population does.

Methods literature on natural experiments and difference-in-differences. If you want to evaluate population-level claims rather than merely be skeptical of them, this is the toolkit. Any modern applied-econometrics text covers it. The payoff is being able to say precisely why a given analysis does or does not identify a causal effect, instead of gesturing at confounding.

Validated weight-stigma instruments and their disagreements. §44.4 argues that instruments diverge and that claims about stigma trends rarely disclose which was used. Reading the instruments themselves — what they ask, what they claim to measure, how explicit and implicit measures relate — is the fastest way to see why the chapter refuses to rate the claim.

Biosimilar market analyses. Chapter 32 gave the mechanism; the empirical literature gives the magnitudes across classes and countries. This is the evidence base behind §44.6's ⚠️, and it is where you would go to see whether the direction is holding.

Historical work on public health interventions displaced by pharmacological ones. §44.5's crowding-out mechanism is asserted as a general pattern; go and check it. It is more contested than the chapter's brisk treatment suggests, and finding out where it has and has not occurred is a genuinely useful exercise in the discipline this book teaches.

If you only do one thing

Read Goffman's Stigma, and then write your dated prediction from the Dossier section.

Goffman takes an evening and gives you the vocabulary for the one section of this chapter most likely to affect how you and the people around you are actually treated. Writing the prediction takes ten minutes and gives you something almost nobody who publishes on this subject has: a claim you can check, on a date you have set, against a falsifier you specified in advance.

One of those two things teaches you what other people have already worked out. The other tells you, in five years, something true about yourself. Do both, but if you do only one, do the second — it is the only item on this page that no book can give you.