23.3 Detecting Leading, Biased and Unsupported Queries
Key Takeaways
Review the source record, question, options, title, formatting and related conversation together.
Yes/no is permitted for appropriate clarification of existing documentation; it should not introduce a previously undocumented diagnosis.
A supported multiple-choice option plus an alternate-explanation opportunity can satisfy the answer-choice rule; a fixed trio of exit options is not universally required.
Unsupported choices or selective evidence remain defects even when an Other field is present.
Audit methods and technology controls are locally designed. A query defect warrants appropriate review and correction, not an automatic fraud or repayment conclusion.
Detecting Leading, Biased and Unsupported Queries
Review the whole communication when assessing a query: the source record, question, answer options, provider-visible title, formatting and any related conversation. The central issue is whether it seeks accurate clarification while preserving the provider's independent judgment. A polished template or an “Other” field cannot cure unsupported options or pressure to produce a preferred answer.
The August 2026 AHIMA/ACDIS brief supplies the professional query framework. It does not make every deviation a federal offense. A compliance finding should identify the actual defect, affected communication and required correction. Any separate legal analysis must address the applicable law and facts.
Identify the mechanism of influence
| Defect | What the reviewer should examine | Example of the problem |
|---|---|---|
| Direction toward a preferred answer | Wording, visual emphasis, preselected responses and conversation | A new diagnosis is already checked before the provider sees the form |
| Unsupported response options | Whether each proposed clinical answer fits the record | Offering hypernatremia solely from a documented sodium of 122 mEq/L |
| Selective evidence | Omitted findings or clinical explanations that materially change the question | Quoting one abnormal result while hiding the provider's documented explanation and relevant follow-up |
| Outcome pressure | References to reimbursement, quality results or denial avoidance | Asking the provider to confirm a diagnosis so the account retains an MCC |
| Improper format or recipient | Whether the format and provider role fit the clarification | A yes/no query introducing a diagnosis found only in an ancillary clinician's assessment |
These concerns may overlap. Avoid treating them as a rigid official taxonomy or inferring motive from a single word. “Length of stay,” for example, may describe actual clinical chronology. The defect is using payment or performance consequences to influence diagnosis, not the mere presence of every term that might also appear in a financial discussion.
Distinguish valid formats from leading use
A yes/no format is not inherently leading. Under the 2026 brief it is appropriate for specified clarification of an already documented diagnosis or relationship, with the required response opportunities. It should not introduce a previously undocumented diagnosis. A query based only on a registered dietitian's malnutrition assessment requires provider-driven multiple-choice clarification when needed; an ancillary cosignature is not reliable diagnostic adoption.
Multiple choice must include at least one supported clinical option and an avenue for an alternative explanation. There is no fixed minimum number of diagnoses. A supported option plus “Other, please specify” can meet the answer-choice requirement. That does not automatically certify the remainder of the query: review its purpose, indicators, source citations and neutrality too.
“Unable to determine,” “ruled out” and “not clinically significant” are useful when suitable to the circumstances; they are not an obligatory trio on every multiple-choice form. POA yes/no queries specifically require yes, no and unable to determine. An inability to determine does not mean an uncertain diagnosis such as probable or unable to rule out.
Apply the rules to concrete communications
New diagnosis from a narrow observation. A query asks, “Do you agree that sodium 131 and confusion represent metabolic encephalopathy due to hyponatremia?” The stem introduces a diagnosis and causal explanation that the provider has not documented. The reviewer should assess the full record before deciding whether a query is justified, then use a suitable format with supported choices and alternate explanation. Do not infer encephalopathy, delirium subtype or causation from a sodium value alone.
Response menu built around a payment target. A respiratory query lists only variants of acute respiratory failure despite a record that supports other explanations of the presentation. “Other” does not excuse including unsupported variants. Review baseline status, provider assessments and relevant observations; revise the question and options around the actual unresolved issue. Do not assume edema automatically supports heart failure or that all chronic heart-failure codes lack CC status.
Blood loss after surgery. A hemoglobin fall with no transfusion does not itself prove either acute blood-loss anemia or harmless hemodilution. Review the operative record, fluid course, provider assessment, monitoring and other care. Avoid demanding the diagnosis from the laboratory change; also avoid a blanket rule that no symptoms or transfusion makes all clarification impermissible. The concern and reporting requirements must be evaluated in context.
A diagnosis is already supported and clear. Do not query merely to replace it with a more favorable code. First apply the Index, Tabular List and reporting rules. A genuine conflict, missing specificity or support concern may justify clarification; the existence of a higher-paying classification does not.
Audit the communication and its use
A useful audit records the sampled population, actual query version, encounter, indicator sources, format, recipient, response and disposition. Include vendor-generated prompts, automated nudges and patient-specific verbal clarification. Compare reviewers' conclusions and calibrate disagreements before presenting a rate as established performance.
Sample size, frequency and any focused pre-delivery review are organizational decisions. A keyword filter may help find communications for review, but it is not required NLP infrastructure and cannot replace judgment. A filter can miss a leading paraphrase or flag a harmless quotation from the record. The 2026 framework permits qualified oversight through review before delivery or ongoing structured audit; it does not mandate physician-advisor approval of every complex query.
Use findings to correct the relevant unit: a flawed shared template may affect many users, while a correct template can still be misused in one encounter. Preserve appropriate versions and disposition history. Distinguish draft, delivered-unanswered, answered and claim-paid situations rather than applying one “void and delete code” instruction to all of them. Section 24.1 develops that remediation workflow.
Separate query defects from claim and legal findings
An improper query does not establish that the underlying condition was absent, that every related payment was excessive or that all elements of a False Claims Act violation are met. It also cannot be ignored simply because a provider signed the answer. Review clinical documentation, code assignment, claim impact and any necessary HIM/provider amendment with the appropriate owners.
For a Medicare Part A/B overpayment, apply the current identification, deadline, suspension and reporting provisions of 42 CFR 401.305. Its framework does not automatically govern every commercial payment. False Claims Act analysis distinguishes the alleged conduct, knowledge and other applicable elements; a query defect or missed general deadline alone is not a complete statutory finding. Escalate credible concerns promptly, preserve evidence, and document the correction and follow-up rather than declare liability from an audit label.
Sources: 2026 AHIMA/ACDIS query brief, FY2027 ICD-10-CM guidelines, 31 USC 3729, and 42 CFR 401.305.
A multiple-choice query contains one clinically supported diagnosis option and an Other field allowing the provider to explain the findings. Which conclusion is justified from that fact alone?
It is necessarily noncompliant because every query needs at least three named diagnoses.
The answer-choice structure can meet the 2026 rule, but the purpose, indicators, wording, recipient and remaining elements still need review.
It is fully compliant regardless of the clinical indicators or a title directing the provider to confirm the diagnosis.
It must add ruled out and unable to determine to every query before any other review can occur.
A provider has not documented encephalopathy. A query asks, Do you agree that sodium 131 and confusion mean metabolic encephalopathy due to hyponatremia? What should the reviewer identify?
Any yes/no query is forbidden, even when it clarifies an already documented condition.
The sodium value establishes the diagnosis and causal relationship, so the query is unnecessary but harmless.
The only defect is the lack of an MCC label next to the diagnosis.
The question introduces and directs a new diagnostic and causal conclusion; review the full record and use a suitable neutral clarification if warranted.
An audit finds financially steering queries, including several linked to paid claims. Which response properly separates the findings?
Stop the defective practice, preserve the communications, review clinical documentation and coding/claim consequences, and apply the relevant correction and legal processes.
Declare every linked diagnosis absent and every payment an overpayment solely because the query was defective.
Keep all payments without review because the providers signed the query responses.
Declare a completed FCA violation from the wording alone and apply one refund deadline to every payer.
Sections you finish are checked off in the contents.