Survey-ready attestation for a public health laboratory

The integrity page puts the idea in one line: survey-ready is a weekly habit, not a week in April. A public health laboratory is asked to show personnel qualifications, competency for the tests those people perform, validation of the methods, proficiency testing, reagent and quality-control records, and the path a result took before it was reported. The records exist, in most laboratories, somewhere. The failure mode is that they exist in six inboxes, a shared drive, and the memory of one supervisor who is on leave.

Attestation, as Salus uses the word, means a package a director can export and a reviewer can follow. It does not mean Salus attests that the laboratory is compliant. It does not mean Global Hands stands in for the laboratory director at the survey. The platform page is explicit that the product does not guarantee compliance, readiness, or a clean survey. Anyone who hears 'survey-ready' as 'survey-proof' is hearing a claim this site does not make.

The package is decision support for the people who already run the quality system. It is built from the same record as the work: the accession, the result, the verification, the transmission, the SOP version, and the human accept or reject on anything an agent drafted. If the work was never recorded, the export cannot invent it. That is a feature. A pretty binder full of reconstructed documents is how laboratories get hurt.

Personnel and competency come first. The evidence list asks for current CLIA personnel files for high-complexity testing, competency assessments tied to the tests each person actually performs, and an exclusion-screening log with review signatures. The screening log covers the OIG List of Excluded Individuals and Entities and the complementary System for Award Management check. The rule the page cites is 42 CFR 1001.1901: no payment under Medicare, Medicaid, or other federal health care programs for items or services furnished, ordered, or prescribed by an excluded person or entity. Who was screened, which list, which date, who reviewed an alert, and what was done: that is the artifact. A one-time check at hire is not the program, because LEIE updates monthly.

Methods and validation come second. Every live assay, including a laboratory-developed test, needs a validation or verification packet. SOP history needs effective dates, not only the current PDF. Proficiency testing enrollment and scored results need to be findable. The workbench can draft the outline from a catalog card, including the LOINC and SNOMED picks, the specimen types, and the human gates. The outline says, in the draft itself, that human review is required and that the document is not a completed validation. A template that pretends to be the study is the opposite of attestation.

Reporting and integrity come third. The list asks for the ELR companion map, a written statement of what the laboratory does not do, and an audit trail covering accession, result, verify, and transmit. The statement of limits belongs in the package because surveyors and partners ask what the software is allowed to decide. The honest answer on this site is: not a patient-level diagnosis, and not an unsupervised release. Writing that down is part of being survey-ready. Hiding it is not.

If Conrad drafted an accession, a coding check, or a validation outline, the package should show the draft, the person who accepted or rejected it, and the time. The control log on the persona, which records status and scope changes, belongs with that story. A surveyor who asks whether artificial intelligence released the result should be answerable from the record: the agent proposed, the named person verified, the named person transmitted. If the record cannot show that, do not paper over it with a policy sentence that the system does not enforce.

The 2026 CLIA request for information, recorded on the regulatory log, already treats AI in post-analytic work as a quality-system topic. Waiting for a final rule to decide whether drafts are stored is how a laboratory ends up describing a practice it cannot demonstrate. Store the draft. Store the decision. Store the rejection. The kill switch is an event, not a slogan.

None of that converts the software into a certified medical device or into the person of the director. Attestation in this product is evidence the laboratory can show. The laboratory's own certificate, its director, and its survey remain the legal facts. Global Hands does not hold a CLIA number for the customer's laboratory and does not bill federal programs for the laboratory's testing.

The practical cadence the integrity content implies is ordinary. Screen against LEIE on the monthly update and record SAM as the complement. File competency when the test menu changes, not only on a hire anniversary. When an SOP takes effect, keep the prior version and the date. When a validation is signed, keep the raw tables, not only the narrative. When a message is transmitted, keep the map that was live that day, including LOINC, SNOMED, and any partner override a human approved.

Exports exist so that cadence can be handed to someone who was not in the room. The site links a survey evidence checklist, a validation outline template, and an LEIE screening log as downloads. They are aids. They are not completed records, and they are not a score. This article will not invent a percent of laboratories that passed a survey with them, because the site does not publish a survey result.

Use the integrity page before a survey, and use it in ordinary weeks. If a document on the list cannot be produced, that gap is the work. Salus can hold the trail you actually create. It cannot attest to a trail you skipped, and it will not call that gap a success.