Slow Registry

Old-Fashioned Tools for Building High-Quality, Reproducible Research

Best practices that do not rely on AI and never go out of style, from three decades in the field.

Danielle Boyce, MPH, DPA

Start reading The habits at a glance Open the workbook

Slow Registry is a free book for the people who run patient registries: the parent who started a registry because no one else would, the patient advocacy group leader running a research project without a research background, the coordinator who keeps a registry going, and the data scientist who has never worked inside a clinical study. It brings the reading, the case studies, the graphics, the exercises, and the workbook together in one place, so you can read a chapter, practice it on a fictional registry, and then apply it to your own.

What the book teaches

Many patient advocacy groups are building registries now, often with AI tools that can draft a consent form, a survey, or a summary report in minutes. Those tools are useful, and the registries that struggle are usually missing the work no tool can do: a person who knows the participants, a protocol and consent settled before enrollment, a weekly look at who is in the cohort, and conversations with the researchers and regulators who will use the data. Each chapter teaches one of those habits and shows what AI can do for that part of the work and what only a person can do.

Slow cooking and slow stitching are choices people make because some things come out better when they are done by hand, at the pace the work requires. A registry is like that.

Chapters

Part One: Before anyone enrolls

1. Decide who the data are for
2. Start with the analysis plan
3. Look at who is already in the space
4. Decide what to capture before you go looking
5. Put a person between the form and the participant
6. Know who belongs on the data team
7. Settle the protocol, dictionary, and consent before enrollment opens

Part Two: Running the registry

8. Verify the diagnosis before you count it
9. Keep records someone else could read
10. Track every sample and every link
11. Hold the weekly meeting
12. Go and find out why the data are missing
13. Stay in touch on a schedule

Part Three: The community and the field

14. Be part of the community, not a website
15. Keep the researchers close
16. Have the conversations only you can have

Part Four: AI on top of all this

17. Read what the machine wrote
18. Build so that a machine can use it

The closing chapter walks a year in the life of the practice registry, from the protocol to the report a sponsor would read.

Practice materials and tools

The Lantern Registry

A hypothetical registry with synthetic data, built with the problems real registry exports have. Every chapter's exercise uses it.

The interactive lab

Click through the export, flag what needs a query or a correction, and check your answers.

The Slow Registry Workbook

A fill-in template for each chapter, for your own registry. Fill it in online or print it and work by hand.

The AI prompt library

A prompt for each chapter's topic, each paired with the step only a person can do.

Using this book with a group

For a board, a team, or a class, with discussion prompts and the answer key.

The case studies

Every chapter opens with a case study from the author's own work with registries. Details are changed to protect the identities of the individuals and institutions involved, and the changes do not alter the substance of the lessons.

Version 0.9. This is a living book, revised as it grows and as readers contribute case studies, rules, and questions.