NHSN Nursing Home Data Validation: Scope, Census and Submitted Versions

Choose the reporting scope before choosing the report

The CDC respiratory-pathogens resource hub separates resident and healthcare-personnel materials, forms and import instructions. Its resident protocol and Table of Instructions are labeled March 2026. Monthly reporting-plan selections used in other LTCF reporting contexts are not a universal substitute for these weekly field instructions.

Write the facility, reporting area, field, collection period, instruction version and source extract on the review record. Resolve applicability through qualified reporting personnel. This guide is an original validation method, not a complete NHSN reporting manual or a surveillance case-definition guide.

The resident count has a precise definition

The linked resident Table of Instructions, page 2, defines Question 1 using residents occupying a bed for at least one day—at least 24 hours—during the collection week, counting each person once. The file’s URL retains an older-looking name; use the document linked from the current CDC hub and check its contents.

That field is not a closing-day bed census and not resident-days. Other fields have separate instructions. Do not use this single definition to infer vaccination, positive-test or hospitalization counts.

Work three counts from one fictional population

Assume the seven-day collection window and all day boundaries have been set correctly under the applicable protocol. Every listed day below represents a full 24 hours in this simplified example.

The weekly qualifying population is eight people. The accumulated time is 46 resident-days. The closing-day population is seven. All three arithmetic results can be correct while only one answers Question 1.

A convenience query that returns everyone with any arrival or departure timestamp in the week is not enough to implement the 24-hour criterion. Test actual dated occupancy data and edge cases against current instructions. If a person has more than one stay, verify person-level deduplication rather than counting stay rows. Do not invent rules for interrupted or partial stays from this simplified example.

Check the identities behind an equal total

Suppose the reviewed source population is residents A through H. The prepared extract contains A through G plus J. Both total eight. The reconciliation must explain missing H and unexpected J before anyone assumes the count is supported.

Investigate both lists; the source extract itself may be stale. Keep the original population, discrepancy, reviewer’s rationale and revised extract identity. Automatically adding H and deleting J would substitute an unreviewed edit for an investigation.

The CDC LTCF validation resource distinguishes internal completeness/accuracy review from external validation. Its older examples are methodological resources, not current respiratory field definitions.

A practical source-to-submission ledger

For manual entry, compare the recorded values with the reviewed output. For file import, preserve the file identity and actual returned evidence. Use current module-specific instructions for submission and corrections; the ledger does not prescribe a universal resubmission action.

Location changes and coverage remain separate dependencies

Where the reporting area uses locations, preserve a reviewed crosswalk among facility, local identity, display label and destination mapping. A renamed unit and a newly reorganized unit raise different questions. A label such as North reused by two facilities must not be a cross-facility joining key.

Then test an authorized backup using their own destination access. An EHR export does not grant NHSN access. The backup also needs the reviewed scope, source locations, submission evidence and unfinished discrepancies. A successful login alone does not establish operational coverage.

Preserve the version that was actually submitted

If a later source correction changes eight to seven, keep the earlier value and evidence while qualified staff review the proposed change. After following the applicable procedure, link the resulting record. If no change is warranted, retain the rationale rather than erasing the question.

The useful acceptance result is a reviewer who can reproduce eight, 46 and seven, identify the A–H versus A–G/J mismatch and distinguish approved E3 from submitted E2. Those observations demonstrate data behavior without claiming that every NHSN field or reporting obligation has been validated.