Reading the Federal Rulebooks · Hanh D. Brown

There are two federal rulebooks. Almost everybody looking for a rule finds one of them and stops.

One holds what a rule requires today. It is arranged by subject, the way a reference book is arranged. The second holds every change ever made to that rule, in date order, with the agency’s own reasoning attached.

They answer different questions. Most people know one. The book nobody opens is the one carrying the dates.

Gaps like that cost people money. A rule read in the first book looks permanent and settled on the page. Nothing about it hints that it changed in February. Only the second book carries that fact.

How is the Code of Federal Regulations organized?#

By subject, into fifty numbered titles. The Code of Federal Regulations (CFR) gives each title a book of its own, covering one field of federal activity. Chapters name the agency behind the rules. Parts name a programme. Sections carry the requirement.

Four levels. The numbering has held for decades, which is why an address written down twenty years ago still works today.

Arrangement is the useful part here. Nothing sits in date order here. Everything sits with its neighbours instead.

So a part about hospital payment sits beside a part about hospital eligibility. Find one and you have found the room the other lives in. That is why browsing beats searching in this book, which is the opposite of how most people use the web.

A plain description of that structure sits with the National Archives, at the shortest correct explanation I have found. The whole thing fits on a single page.

Where do Medicare rules live in the Code?#

In title 42, under the agency that runs the programme. The exact address sits in Check a Policy Claim, along with the three parts carrying most of the payment rules. This page will not repeat it.

The habit is worth more than the address itself. Learn one title number for the subject you care about. Every future search then starts inside the right book. One number. Learned once.

That single fact does more work than any search tool. A keyword search across fifty titles returns noise, because the same ordinary word appears in farm rules and banking rules and health rules alike. The same search inside one title returns a shelf you can read. Narrow first. Then search.

How do you read a citation like 42 CFR 422.101?#

Left to right, widest first. It is like a street address written backwards, starting with the county and ending at the door of the house you want. The number 42 names the title. The letters name the rulebook. Everything after that narrows, one step at a time, until only one rule is left standing.

Take one apart by hand and you will never have to do it slowly again. Title 42 is the health book, covering public health and the big benefit programmes together. Part 422 is one programme inside it. The number after the dot is the section.

So 422.101 points at a specific requirement. In this case it is the section on requirements relating to basic benefits, inside the Medicare Advantage rules. It is not a chapter heading and not a summary. It is the rule itself, on its own page, in the words that a hospital or a plan is actually held to. Read those words and you have read the law on that point.

Anything further right narrows further still. A letter in brackets points at a paragraph. Two of them point at a sentence. Precision runs rightward. Read to the end.

One warning about the numbers, and it catches careful people. The same part number appears in more than one title. A citation with no title number in front of it is not an address at all, so always carry the title.

That is the whole skill. One minute to learn. It saves the hour I used to lose hunting by keyword and finding near misses instead.

Which parts decide what Medicare pays?#

A small number of them, and knowing which ones turns a long search into a short one. Payment rules cluster into a handful of parts rather than spreading evenly through the book, so a reader who learns the cluster can go straight to it and skip everything else.

Four groups carry most of what I get asked about, and almost every question I see lands in one of them:

• Payment rules. What the programme pays a hospital or a plan, and on what basis.
• Coverage rules. Which items and services are covered at all.
• Enrolment rules. Who may join, when, and under what conditions.
• Appeal rules. What happens after a decision goes against you.

Each group keeps its own parts and its own numbering, which is why a reader who knows the group can find the number in a minute. This page names the shape rather than the numbers. Numbers move between editions. The shape has not moved in years, and a reader who holds the shape can find the numbers in a minute.

Knowing which group your question sits in is most of the work. A question about money is a payment question, and it will sit near other questions about money. A question about whether something is covered at all belongs to a different group entirely, in a different part of the book, written by people answering a different problem.

Mixing the two is the wasted hour I run into most often in this work. You can read a payment rule very carefully and still not learn whether the thing is covered.

How are statute, rule and guidance different?#

By who wrote the paper and by how much weight it carries. A statute is written by Congress. A rule is written by an agency to carry that statute out, and it goes through public comment before it becomes final.

Guidance is the third thing. Manuals, agency web pages, answers to common questions. All of it explains what the agency believes its own rule means. None of it was voted on.

That material is genuinely useful. It is often far clearer than the rule itself. Most of it binds nobody.

Medicare is stricter still. In Azar v. Allina Health Services the Supreme Court held that a payment policy the government had posted without notice and comment had to be vacated.

The statute is narrower than people assume and still wide enough to matter. It reaches a rule, a requirement, or any other statement of policy that establishes or changes a substantive legal standard about coverage, payment, or who is eligible. That has to go through rulemaking, whatever the paper calls itself.

One thing is carved out by name: a national coverage determination. Everything else in that range is caught. The hospitals won.

There is a practical test, short enough to remember. Ask what the document would be worth in a dispute. A statute or a final rule settles the question. A manual page explains it and then defers to the rule sitting behind it.

The one question that cuts through almost everything is what this page would be worth in a dispute. It works even when you cannot follow the substance, because it asks about the standing of the paper rather than the content on it. Ask it early. Ask it of every source somebody hands you.

Keep the three separate on paper and most of the bad policy writing I read becomes visible at once. The usual failure is a manual page quoted as though it were the rule. It arrives in a confident hand, with a real link attached, and everything about it looks finished. Check what wrote it.

How do you find the rule behind a denial letter?#

Start with the letter in your hand and ask it one question. Does it name a rule. Everything follows from that answer, and the two routes are genuinely different amounts of work, so it is worth settling before you begin anything else.

A letter naming a rule has handed you the address, which is more than most letters manage. Read the section it points at. Then check the second rulebook for the date that section last changed. Ten minutes and a phone will do it, and the ten minutes are worth spending before you write a single line of an appeal.

The other route starts empty. A letter naming no rule is not a dead end. It is a request you are entitled to make. Ask in writing which provision the decision rests on, and keep the answer.

Put that request in writing rather than making it by phone. Never by phone. A written request creates a record that survives a change of staff, a change of contractor, and a year of delay. A written answer gives you something to check, and a decision that cannot name its rule is a decision worth appealing.

Then go to the second book. Search the Federal Register for the part number. Every change to it appears in date order, with the agency’s reasoning attached to each one.

It is like a photograph beside a film reel. The first book shows the law standing still. The second shows it moving. A decision made last spring rested on whatever the law said last spring. That is rarely what you are reading today. Check the date.

How do you know whether a rule exists at all?#

Often you cannot, and an empty search proves very little on its own. Nothing found might mean that no rule exists anywhere in the book. It might equally mean the rule is sitting under a word you never thought to try.

So separate the two before concluding anything. Try the subject in the agency’s own vocabulary rather than your own. Then the programme name. Then the part number, if you hold one. Vary the words. Agencies rarely use the phrase an ordinary person would reach for, and the gap between those two vocabularies is where most of my failed searches have lived.

Three attempts is a reasonable floor, and a fourth is worth it if somebody suggests a word you had not considered. Only then is an absence worth mentioning at all.

Keep a note of what you tried while you try it. The list of failed search terms is what turns a shrug into evidence, and it is the difference between telling somebody you looked and being able to show them where.

Even then, mention it as an absence you searched for. Never as a fact about the law. What a checker certifies is the same discipline pointed at a rulebook instead of a footnote.

The reason to be careful is not academic. A family deciding whether to appeal is deciding on what somebody found. There is no rule and I could not find one are different sentences.

Those two lead to opposite choices, and the gap between them is a person’s afternoon of searching. The people I meet who most need the distinction usually have the least time to make it, which is the case for readers who need it most.