9.3 Open-Ended, Multiple-Choice, and Yes/No

9.3 Open-Ended, Multiple-Choice, and Yes/No

Quick Answer: Use open-ended when you need a free-text clinical answer. Use multiple-choice with at least one clinically valid option plus required Other, please specify, and never highlight a favorite. Use yes/no only to clarify something already documented—not to introduce a new diagnosis, and not from a dietitian or wound-nurse diagnosis alone.

Format is how you ask, not why you ask. The 2026 ACDIS/AHIMA brief still requires every format to be nonleading, to carry patient-specific sourced indicators, to omit reimbursement and quality-outcome language, and to preserve provider judgment. A beautifully designed template that fails those tests is still noncompliant. Conversely, a simple open-ended question that cites the right findings can be fully compliant.

Choose the format that matches the documentation gap. Asking the wrong shape of question is one of the highest-yield CCDS traps: yes/no used to birth a new diagnosis, multiple-choice that offers only the billable option, or open-ended stems so vague that the provider cannot tell what is unclear.

Open-ended queries

An open-ended query presents relevant clinical indicators and a nonleading statement that identifies the issue, then leaves space for the provider to document the condition, acuity, etiology, or relationship in their own words. Because you are not handing the provider a short list of diagnoses, this format is useful when several explanations could fit, when you are capturing a verbal query, or when you need the provider to describe clinical significance rather than tick a box.

Open-ended is not a license to hint. "Based on the rising creatinine, oxygen need, and pressor use, what is the diagnosis you are treating as the reason for admission?" is a question. "Please document the septic shock that these indicators obviously show" is a leading speech. Keep the stem to the gap: conflicting notes, missing specificity, or an uncertain term that needs a later-in-the-stay update.

The tradeoff is follow-up. Free-text answers can be incomplete ("yes, it's worse"), off-target, or still unspecified. If the response does not provide the clarity requested, follow organizational escalation rather than editing the provider's words into a code. A second query is appropriate when new information appears or when the first answer needs a further, still-nonleading precision—not when you simply disliked the first answer.

Multiple-choice queries under the 2026 brief

Multiple-choice may be used in any circumstance that needs structured options. The stem remains a nonleading statement plus sourced indicators. Providing options that are supported by those indicators is not, by itself, introducing new information or automatically leading.

The 2026 construction rules candidates must not invert:

• Include at least one clinically valid option—meaning an option actually supported by the indicators in this record.
• Include a required open-ended path such as Other, please specify or Other explanation of clinical findings, please specify. That path is how the provider records an alternative the template did not list.
• There is no mandated minimum number of diagnosis options beyond that clinically valid choice plus the Other path. You do not need four homemade diagnoses to look balanced. You also have no mandated maximum; list as many as the case warrants.
• Options must be clinically relevant. Do not offer a diagnosis contradicted by the sourced data (an opposite electrolyte disorder when the labs show the reverse; a volume-overload label when the indicators describe depletion only).
• Options must be reportable under UHDDS and the ICD-10-CM Official Guidelines when you are asking about inpatient diagnoses or procedures.
• Options may be listed in any order. Do not put the hoped-for MCC first as a wink.
• Do not use arrows, bolding, highlighting, underlining, or similar emphasis to mark a preferred answer. Do not highlight a favorite indicator in the stem either if that emphasis would steer the response.

Optional extra choices—not required in every multiple-choice query—include unknown, unable to determine, not clinically significant, integral to, inherent to, unable to rule out, or similar wording. Know that unable to determine is not the same as unable to rule out and is not an uncertain diagnosis such as possible or probable. If the provider selects unable to determine, decide case by case whether to escalate; do not treat it as a coded possible condition.

Multiple-choice is the optimal format for clinical validation (a documented diagnosis that may lack support). Validation option sets typically still include Other plus choices such as ruled out, no longer valid after study, confirmed with additional support, or an alternative diagnosis. That topic is expanded in the next chapter; the format rule you need here is that validation is not a yes/no fishing trip for a new code.

Trap: Building a multiple-choice list that is really one desired diagnosis, two impossible distractors, and no Other line. That is a leading questionnaire dressed as a menu.

Yes/no queries: a narrow inpatient tool

A yes/no query clarifies an already-documented diagnosis or relationship. In any setting, it may not be used to introduce a new, previously undocumented diagnosis. The query must reference the documented issue and related clinical indicators, and it must not signal a desired answer. Phrase it so that yes or no is a coherent reply—and then add the third door the 2026 brief requires.

Unable to determine is required on POA queries and on yes/no queries generally. POA is an inpatient construct (conditions present at the inpatient order, including those that began in the ED or observation before that order). A POA yes/no that offers only yes or no, with no clinically unable-to-determine choice, is incomplete.

Yes/no is appropriate for:

1. Determining POA status of a condition that is already documented, including the unable-to-determine option.
2. Substantiating a medical diagnosis already present in provider documentation, such as asking the attending to interpret a pathologist's biopsy diagnosis or a radiologist's imaging impression that already lives in the record.
3. Establishing or negating cause-and-effect between documented conditions (manifestation and etiology, a complication relationship, or conflict between two providers' notes).

Yes/no is not appropriate when the only "diagnosis" in the chart comes from a clinician who is not authorized to establish that medical diagnosis within their scope of practice. Documentation of wound etiology by a wound-care nurse, or malnutrition by a registered dietitian (RD), does not convert into a physician diagnosis through a yes/no. In those situations use multiple-choice, with clinically supported options and Other, please specify, so the provider can independently name, reject, or replace the condition. Ancillary notes remain excellent clinical indicators; they are not a back door to yes/no a new coded diagnosis into existence.

Trap: "Does the patient have acute respiratory failure? Yes/No" when no provider has documented that diagnosis. That is a new-diagnosis query wearing a yes/no costume. Use open-ended or multiple-choice instead.

Choosing a format on the exam

Titles visible to providers must not steer the answer (a generic "CDI query — respiratory status" is safer than a title that names the hoped-for failure code). Regardless of format, if the authenticated response is in the permanent record, it can support coding; if later notes conflict, query again. Format never authorizes reimbursement language.