Case Study 2 — The Rules That Changed: Query Practice Guidance, the Yes/No Reversal, and the Question Nobody Settled

Real and public, with one clearly labeled composite. The professional practice guidance described below is published by the American Health Information Management Association (AHIMA) and the Association of Clinical Documentation Integrity Specialists (ACDIS), has been revised repeatedly over three decades, and is available from those organizations. This case study cites no edition year, no page, and no verbatim text, and it asserts no enforcement figure or survey statistic. That omission is deliberate and is the case study's point: the specific content of query guidance is exactly the kind of thing a textbook must not freeze. Get the current edition from the source.

Background

Chapter 4 §4.9 gave you the query's rules. §38.3 gave you its anatomy, its option set, and its retention. Both sections were written as though those rules were settled.

They are not settled, they have never been settled, and the most useful thing this case study can do is show you how they moved — because you are going to spend a career under guidance that keeps moving.

Start with a fact that surprises most new coders: the rules governing physician queries are not statute and are not regulation. No federal law defines a compliant query. CMS's manuals address documentation requirements without prescribing query wording. What exists instead is professional practice guidance, published by the field's own membership organizations, revised as practice changes and as enforcement teaches the field something new.

AHIMA — the credentialing and professional body for health information management, introduced in Chapter 6 and covered fully in Chapter 39 — has published practice briefs on provider queries since the 1990s. ACDIS, the professional association for clinical documentation integrity specialists, was founded in the 2000s as CDI became a distinct job rather than a coder's side task, and publishes its own guidance and a code of ethics for the specialty. The two organizations now publish query practice guidance jointly, which is itself a small institutional story: two bodies with overlapping members and different origins concluded that a field with two conflicting standards has no standard.

And here is the tension that makes this a case study rather than a citation. That guidance is voluntary in the sense that no statute enacts it. It is not voluntary in any sense that matters to a person whose query practice is being reviewed. It is the standard auditors reference, the standard expert witnesses testify to, the standard a compliance program adopts by reference, and the standard a credentialing organization's ethics code enforces against an individual. A body of professional guidance has become the operative rule for a regulated activity — which works well, and which means that when the guidance changes, the rule changes, with no rulemaking, no comment period, and no notice to anyone who is not a member.

The issue

Three things in query guidance have moved materially over the decades, and one has never been settled at all. Take them in order, because each teaches something different.

1. Yes/no queries: from prohibited to permitted, narrowly — a genuine reversal

The earlier position was straightforward and defensible: a yes/no query is inherently leading. If the only two answers are the one the coder wants and the one they do not, the question has done the work. Guidance discouraged the format and, in practice, many organizations forbade it outright.

The later position permits it in defined circumstances — and the circumstances are the interesting part. A yes/no query is acceptable where a diagnosis is already documented somewhere in the record and the question is whether it can be confirmed, further specified, or associated with another documented condition. It remains unacceptable for introducing a diagnosis the record has never named.

Why the change is right, and why the reasoning matters more than the conclusion: the earlier rule was written against the format and the later rule is written against the function. The test is not how many answer boxes the query has. The test is whether a "yes" tells you something new about the record or something new about the patient. If the diagnosis is in the consultant's note and the question is whether the attending concurs, a yes/no query asks the attending to do exactly what only the attending can do. If the diagnosis appears nowhere and the query proposes it, the number of options is irrelevant — the query supplied the answer.

§38.3 teaches the later rule. An organization still operating the earlier one is being more restrictive than the standard, which costs it clarifications it could legitimately obtain but harms nobody. An organization that adopted the later rule without adopting the reasoning is the one at risk, because "yes/no queries are allowed now" is a sentence that will produce a leading query within a month.

2. Verbal queries: from unaddressed to permitted with an obligation

Early guidance was written for a world of written queries routed through a health information department. Concurrent review made that world obsolete — §38.2's whole method is a reviewer standing at a nurses' station, and the conversation happens whether the guidance addresses it or not.

Current guidance permits the verbal query and attaches the documentation obligation §38.3 states: who asked, whom, when, what was asked, what was answered. The evolution here was not a reversal. It was guidance catching up to a practice that had already changed — and the interval during which the practice existed and the guidance did not is exactly the interval in which organizations invented their own rules, some of which are still in their policy binders.

3. Query construction: steadily more specific, in one direction

Across revisions, the guidance on how a query is built has become more prescriptive and has moved consistently the same way: present the clinical indicators from this record; offer options that include the ones that add nothing; never state the financial consequence; never name the code. §38.3's seven elements are a distillation of that trajectory. The direction of travel has been constant even when specific rules reversed, and it tells you where the next revision will go.

4. Retention: never settled, and it is the most consequential of the four

Should a query and its response be part of the permanent legal health record, or a separate business record? §38.3 gave you both models and told you to pick one. That is not the book hedging. The field has genuinely not converged, and the reasons on each side are real.

Keep it in the record and the provenance of every documentation change is visible to every later reader — including a reviewer who might otherwise wonder where a diagnosis came from. That is transparency, and transparency is usually the right instinct.

Keep it out and the record contains clinical documentation rather than administrative correspondence, and a query is not mistaken for a clinical opinion by a future clinician reading the chart. That is also a real concern, and it is a patient-care argument rather than a billing one.

Both positions are defensible. What is not defensible is not having decided, and the failure below is what "not having decided" looks like three years later.

What happened: a composite

⚠️ The following is a clearly labeled composite, assembled from documented patterns rather than from any single organization. No named entity is described. The mechanism, the sequence, and the outcome are drawn from failures this book has already taught in other settings.

A hospital system stood up a CDI program. The compliance department wrote a query policy, and it was a good one for the year it was written: it named the practice brief it was based on, prohibited yes/no queries entirely, required written queries only, and specified that queries be retained in the CDI software and not filed in the medical record.

Over the following years, four things happened, none of them decisions.

The guidance was revised. Twice. Nobody in the organization subscribed to the revisions, because the policy had been written by a compliance analyst who had since moved to a different department and the CDI manager had assumed compliance owned it.

Practice drifted ahead of the policy. Concurrent reviewers were having verbal conversations at the nurses' station every day, because that is what concurrent review is. The policy said written queries only. Nobody was documenting the verbal ones, because documenting something the policy said should not be happening felt like creating evidence of a violation. A workaround that mostly works becomes institutional knowledge — Chapter 25's Case Study 2, in a new department.

A vendor's software arrived with a template library. The templates were built to current guidance and included yes/no formats for confirming documented diagnoses. They were compliant. They contradicted the policy. Nobody reconciled the two, and reviewers used the templates because the templates were on the screen.

And the retention rule quietly stopped being followed. The CDI system was replaced; queries in the old system were archived somewhere; the new system's default routed the query response into the patient's chart. Nobody chose that. It was a default.

Then a payer requested records on a set of admissions and asked, in the request, for the queries.

What the organization produced was: some queries in the chart and some not, depending on which system era the admission fell in; verbal clarifications with no query record at all, appearing in the chart as progress notes that changed a diagnosis for no visible reason; and query text from a template library the organization's own written policy prohibited.

Nothing in that inventory is a false claim. Every diagnosis was documented by a physician. Every query — including the yes/no ones — was compliant under current guidance. The organization's problem was that it could not produce a coherent account of its own practice, and Chapter 37 §37.1 named the consequence precisely: a documented control that was never operated is worse than no documented control, because it establishes what the organization understood its own obligation to be. The written policy was now evidence against a practice that was, in substance, fine.

What it shows

First, the guidance in this field is a moving object and somebody has to own the subscription. Chapter 6 §6.7 taught the annual update cycle for the code sets: ICD-10-CM every October 1, CPT every January 1, HCPCS Level II and the NCCI edits quarterly. Professional practice guidance has no calendar at all, which makes it harder rather than easier — a revision arrives when it arrives, and nothing in an organization's normal operations announces it. Name a person. That is the whole control, and it is the same answer Chapter 37 §37.10 gives for everything else.

Second, a policy that is more restrictive than the standard is not automatically safe. The organization's yes/no prohibition harmed no patient and cost some legitimate clarifications, which is a defensible trade. What made it dangerous was the gap between the policy and the practice, not the policy's strictness. A rule nobody follows is worse than a rule that is slightly wrong, because the first one is evidence and the second one is only an error.

Third — and this is the one to carry into any job — the reversal on yes/no queries is a lesson about reasoning rather than about queries. The rule changed because the earlier version tested the wrong thing. It tested a format when the actual danger is a function. Every rule in this book that survives does so because it names a function: modifier 25 asserts a significant, separately identifiable service; a documented negative records that something did not happen; the assertion register inventories assertion-makers rather than technologies. Rules written against formats decay. Rules written against functions do not, and when you read a revised standard, look for which kind it is.

Fourth, the default that nobody chose is the same mechanism this book has watched a dozen times, in one more place. A system migration changed where query responses landed, and the change was made by a configuration rather than by a decision. Chapter 37 §37.10 names a system migration, upgrade, or vendor change as one of the two events that force an off-cycle review of the assertion register, and this is why: migration is when accumulated drift arrives all at once.

And fifth, the honest counterweight. None of this means query guidance is unreliable or that professional bodies should not write it. The alternative — federal rulemaking on the wording of a clinical clarification — would be slower, less expert, and far more brittle, and the field's own guidance has been revised in the right direction consistently for three decades. The mechanism works. It works on the assumption that practitioners read it, and that assumption is the part an individual organization can fail.

The outcome

Query practice guidance continues to be published jointly and continues to be revised. The yes/no question is settled in the direction of narrow permission with a functional test. The verbal query is settled and carries a documentation obligation. The retention question remains an organizational policy decision, and the field shows no sign of converging on one answer, which means §38.3's instruction — pick one, write it down, apply it consistently, and make queries retrievable for at least as long as the audit look-back — is likely to remain the operative advice for a long time.

For the composite organization, the fix was unglamorous and took a quarter: one named owner for the query policy, a subscription to the guidance, a reconciliation of the template library against the written policy, a documented rule for verbal queries that matched what reviewers were actually doing, and a single retention decision applied prospectively with the historical inconsistency documented rather than hidden. Nothing about the coding changed. Nothing about the clinical documentation changed. What changed is that the organization could describe its own practice in one page.

The lesson

Know which of your rules came from a statute, which from a regulation, which from a payer contract, and which from professional guidance. They have different revision cycles, different authorities, and different consequences for being wrong, and most practitioners hold them all in one undifferentiated pile marked "the rules." Chapter 5 taught the first, Chapter 22 the second, Chapter 2 the third. This chapter is the fourth, and it is the only one that changes without anyone publishing a notice.

Subscribe to the source and read the revision when it lands. It is a short document, it is written for you, and reading it is the cheapest professional development available in this field. Chapter 40 calls this the reading habit; here is one of its highest-yield applications.

Never cite a query rule from a textbook, including this one. §38.3 teaches the anatomy and the reasoning, both of which are durable. The specific permissions are not, and a policy citing this book — or any book — as authority for what a query may contain is a policy that will be wrong within a few years and will not know it.

And check your own organization's policy against your own organization's practice this month. Not the guidance against the policy — the policy against the practice. The gap between them is where this case study's failure lived, and it is visible in an afternoon to anyone who asks two reviewers how they actually handle a conversation at a nurses' station.

Discussion questions

1. The yes/no reversal replaced a format test with a function test. Write both tests in one sentence each, then apply them to this query: "The record documents CKD stage 3a on the problem list and diabetes in the assessment. Is the kidney disease diabetic in etiology? Yes / No." Does it pass? Does your answer change depending on whether the linkage appears anywhere in the record?

2. Argue the retention question from both sides in a paragraph each, then pick one and write the policy sentence your organization would adopt. Name what you gave up.

3. This case study's composite failed on four separate mechanisms this book has already taught. Name each one and the chapter it came from, and say which of the four you consider most preventable.

4. Chapter 39 covers the credentialing organizations and their ethics codes. Given that professional guidance functions as the operative standard for query practice, what obligation does that place on an individual credential holder that does not fall on an uncredentialed employee doing the same work? Is that fair?

5. "Rules written against formats decay; rules written against functions do not." Test that claim against three rules from earlier in this book — one you think is written against a function, one against a format, and one you cannot classify. What does the third one tell you?

6. Your CDI software vendor ships a template library. Who in your organization is responsible for reconciling it against your written query policy, what would that reconciliation consist of, and on what schedule would it repeat? Write the assertion register entry for the template library.