How should a lab score the quality of a specialist handover?

Review the handover against the information available to the model at the moment it wrote the referral.

- Attribute customer statements rather than presenting them as verified facts. - Preserve the customer’s stated reason for requesting assistance. - Separate confirmed account facts from unresolved customer assertions. - Record relevant constraints without adding unsupported personal explanations. - Omit sensitive detail that serves no supplied intake requirement. - Identify missing information needed by the receiving role. - Keep the requested next action clear and actionable.

A summary can fail even when every sentence sounds plausible. In healthcare billing, a customer may say that time away from work has affected payment. Expanding that statement into a diagnosis or a prediction about future earnings introduces claims the interaction did not support.

The rubric should distinguish omission from invention. Leaving out a necessary callback constraint can obstruct the handover, while inventing a reason for that constraint distorts the record. Buyers can request annotations linking each required summary fact to its supporting evidence. This makes disagreements about quality reviewable and helps training target precise information handling rather than summary length.