Preface

This book grew out of a simple observation: the people best equipped to understand blockchain systems are often the least served by the material that explains them. Blockchain pedagogy is dominated by two genres. The first is the developer tutorial, which teaches how to call an interface without explaining why the system is built the way it is. The second is the promotional explainer, which supplies confidence in place of understanding. Neither serves a reader who can fit a survival model, prove a bound, or design a sampling plan, and who simply wants to know what is actually true.

We wrote this book for first-year graduate students in public health, and in particular with the students of the biostatistics and epidemiology programmes at UC San Diego’s school of public health in mind. The reader we imagine has a first course in probability, some comfort with algebraic structure, and working R. What such a reader brings to this subject is, in our view, an unusual advantage, and the book is organised to exploit it. Nakamoto consensus is a statement about the tail probability of a random walk. The ten-minute block time is the set point of a feedback controller. Committee selection in proof-of-stake is probability-proportional- to-size sampling, familiar from complex health surveys. Wealth concentration on a chain is a Lorenz curve, the same instrument we use to measure inequity in vaccine coverage. Almost every load-bearing idea in this field is one the reader already owns in another language, and the book’s purpose is to perform the translation.

Why this matters now, and to you

A first-year student of medicine or public health, browsing this book in a shop, is entitled to a blunt question: why should a health scientist spend ten weeks on a technology born to move cryptocurrency? We shall answer it just as bluntly. The verifiable-data machinery this book takes apart has already arrived in health, at a scale that makes a working knowledge of it a matter of professional competence rather than curiosity, and the pace at which it is arriving is increasing.

Consider what is already in production. The World Health Organization launched, in 2023, a Global Digital Health Certification Network (World Health Organization, 2023), a global public-key trust directory through which a vaccination record, and in time an immunisation history, an international patient summary, or a clinician’s licence, may be verified across borders by cryptographic signature rather than by a telephone call to the issuing authority. In the United States the SMART Health Cards standard (SMART Health Cards / VCI, 2024) put cryptographically signed health records into the hands of hundreds of millions of people during and after the COVID-19 pandemic. Europe’s medicine-verification system has, since 2019, checked the authenticity of prescription drugs across an entire continent, and the United States’ Drug Supply Chain Security Act has, across the mid-2020s, phased in an interoperable package-level tracing requirement for the drug supply, which the pharmaceutical MediLedger network (U.S. Food and Drug Administration, 2023) serves with a production ledger that uses zero-knowledge proofs to verify a package’s legitimacy without exposing commercial data. Estonia has, since the 2010s, protected the integrity of its national health records with a blockchain-style hash-linked timestamping service (Guardtime, 2016), committing only cryptographic digests, never patient data, to the ledger.

Now the honest half, for it is the more instructive. Scarcely any of these successes is the thing the promoters of the last decade promised. The pattern, which we shall meet throughout the book, is stark: the verifiable-data ideas that succeeded in health quietly shed the word ‘blockchain’ and won as cryptographic- signature infrastructure, while the ventures that clung to tokens and ‘data marketplaces’ largely failed. Nebula Genomics, which promised consumers ownership of their genome on a chain, closed its consumer service in 2025, and systematic reviews through the middle of the decade find that the great majority of patient-data blockchain ventures never left the pilot stage. A reader who cannot tell the two apart is at the mercy of whoever is selling.

That distinction is precisely the competence this book exists to build, and it is why the moment is now. The health professional of the coming decade will not, for the most part, be asked to build these systems. She will be asked to judge them: to sit on the committee evaluating a vendor’s ‘immutable, blockchain-secured’ patient ledger, to advise a trial sponsor on whether a distributed consent registry earns its cost, to decide whether a supply-chain attestation proves what it claims. The pressure will only grow, for reasons beyond the deployments already listed. Verifiable credentials are spreading from vaccination into licensure and records, so that a clinician’s own credentials will increasingly be cryptographic objects. And as predictive artificial intelligence enters the clinic, the demand for tamper-evident provenance, an auditable trail from a recommendation back to the data and the model that produced it, is becoming a regulatory expectation. Rules that took effect in the United States in 2025 required certified health-information technology to disclose the source attributes, the training data among them, of predictive clinical algorithms (U.S. Office of the National Coordinator for Health Information Technology, 2024), though we should note in the same breath that the requirement is contested and its future uncertain. The direction of travel is not in doubt even where the details are.

The essential skill, then, is not to program a chain but to evaluate a claim about one, and that skill rests on the mathematics a statistician already commands. This book sets out to build it.

The organizing concept: applications in public health

The overriding concept of the book is the application of blockchain ideas to public-health data problems. That is not a veneer on a general cryptocurrency course; it is the spine. The property that makes a blockchain worth a health scientist’s attention is integrity without a custodian: a record that is tamper-evident, a provenance chain that anyone can verify, and a computation whose result can be audited without trusting the party who ran it. These are precisely the guarantees that clinical- trial data provenance, pharmaceutical and vaccine supply-chain verification, health-record integrity, and programmable consent ask for.

Every chapter is built around this frame. Each opens with a public-health data problem, develops the relevant mechanism and its mathematics as the means of addressing it, and closes by returning to the problem to ask, honestly, whether the mechanism earns its considerable cost. A single running case study threads the entire book: a multi-institution clinical-trial data-integrity ledger, in which several hospitals that do not fully trust one another must agree on an append-only record of enrolment, randomisation, and outcome events. A vaccine cold-chain provenance system recurs as a secondary example. By the end the reader will have seen one realistic application developed against every layer of the stack.

We should be candid about the negative result as well. For most health-data problems a well-governed database with an audit log is simpler, cheaper, and sufficient, and the chapters say so wherever it is true. Knowing when the elaborate machine is unnecessary is itself a statistician’s contribution, and we treat that judgement as part of the subject rather than an aside.

The sceptical stance

The book maintains a critical lens by default, not as a pose but as a discipline. The cryptography at the base of these systems solved a genuinely hard problem, namely open-membership Byzantine agreement without a trusted party. Much of what has been built on top of that achievement is speculation, and some of it is fraud. Both statements are true, and a calibrated mind holds both without collapsing into either evangelism or dismissal.

To keep the analysis honest, we classify every claim we encounter into one of four falsifiability tiers (protocol fact, empirical claim, economic claim, narrative), and we ask the reader to demote each assertion to its true tier before believing it. The reader is also asked to keep a calibrated forecasting record, scored by the Brier score, so that scepticism itself is held to account rather than merely asserted. Epidemiologists who have followed the epidemic-forecasting literature will recognise the discipline.

What this book covers

The ten chapters map to a ten-week course, organised in four parts.

  1. Foundations. The consensus problem (Byzantine agreement in the permissionless setting) and the cryptographic primitives (hash functions, Merkle trees, digital signatures over elliptic curves), studied empirically in R.
  2. Bitcoin and Nakamoto consensus. Proof-of-work and the backbone protocol; the statistics of mining as a renewal and control system; and the security economics of attacks.
  3. Ethereum and programmable state. The account model and the virtual machine; decentralised finance treated as applied mathematics; and maximal extractable value as market microstructure.
  4. Proof-of-stake and on-chain data. Stake-based consensus, finality, and restaking risk; and the statistical analysis of on-chain data, including tokenomics, wealth concentration, and forecast calibration.

What this book does not cover

The book is not a developer manual. We do not teach Solidity programming, wallet operation, smart-contract security auditing, or the practical mechanics of trading; these are covered well elsewhere, and pointers appear in the relevant Further reading sections. Nor is the book investment advice. Where we treat markets and tokenomics we do so analytically and, the reader will find, sceptically.

Prerequisites and companion material

We assume a first graduate course in probability, comfort with groups and finite fields at the level of a good undergraduate algebra course, and R at a working level. The source-rmd/ directory accompanying the book contains the self-study curriculum and a fully runnable R companion from which much of the book’s tested code is drawn. Every computational result in the book is generated at build time from the code shown beside it, so the printed results and the code that produced them cannot drift apart.

Chapter template

Each content chapter follows the same sequence: learning objectives, orientation, the statistician’s contribution, content sections with collapsible check-your-understanding callouts, a worked example, a section on collaborating with an LLM, exercises, and further reading. The pattern repeats deliberately; by the third chapter the reader will know where to find each component. The template puts human judgement and verification at the centre of every chapter, which is where, in our view, they belong.