Reconcile Covered Persons Before 1095-B Reporting

Identify the provider, reporting year, coverage arrangement and source systems included in the review. Establish the applicable reporting treatment before combining feeds. The 2025 instructions distinguish B-series reporting from other coverage reporting and include rules for multiple coverage arrangements.

Define the expected covered population from reviewed enrollment evidence, not solely the current active roster. A person who ended coverage during the year can still belong in the annual reporting population. Keep excluded arrangements or records in an explained scope list rather than silently dropping them.

Separate the responsible individual, subscriber or administrative group from each covered person. A statement recipient can be outside the covered population. Document the relationship so a recipient record does not create a fictitious enrollment row.

Use provider-scoped internal person keys and a controlled crosswalk to source member identifiers. Retain the original source row and group association. Names, addresses or a subscriber number alone may not distinguish all family members or resolve a re-enrollment episode.

Compare potential matches using authorized source evidence. When a source key changes, record whether the change represents the same person, a new person or an unresolved identity question. Do not merge records simply to make the total match an expected number.

Keep the relationship between a person and their enrollment events. Several source rows may represent legitimate periods or plans, while another repeated row may be an accidental duplicate. The reconciliation must preserve coverage evidence even when it resolves repeated administrative records.

Include a supporting person list for every total. A net difference of one does not reveal the actual issues. Group-level and person-level comparisons show where the differences occur and prevent an omission in one family from being offset by a repeated record in another.

These groups are fictional administrative relationships established for the exercise. They do not assert that every coverage arrangement uses the same statement grouping rules.

Fictional worked resolution: the reviewer confirms that Group A contains one accidental repeated row and Group B contains another. Group C is missing one child whose enrollment is supported by an approved source record. The ten feed rows therefore represent only eight unique covered people before the missing child is restored.

The reviewed result is 10 source rows - 2 accidental duplicates + 1 missing person = 9 covered people. The final group counts are three, two and four, matching the expected population. The reviewer retains the removed-row references and added-person evidence.

The team does not simply delete one row to reduce ten to nine. That would leave the omission unresolved and might remove a legitimate person. The worked example shows why a correct net total can conceal an incorrect population.

Once identities are reconciled, verify that every reviewed person belongs to the intended statement relationship and provider. Check any move between subscriber groups, not just whether the person appears somewhere in the export. A person assigned to the wrong recipient can produce a privacy and reporting problem even when the overall count is correct.

The 2025 Part IV instructions require covered-individual information and person-specific month indicators. Review the associated enrollment periods before approving those indicators. Resolving a duplicate person does not establish that all of their coverage events have been combined correctly.

Use the monthly coverage worksheet for that next step. If more people require continuation pages, preserve their association with the original statement rather than treating a page boundary as a new recipient group.

Save expected and final person lists, group counts, source versions and each disposition. Use concise reasons such as confirmed duplicate source row, restored omitted enrollment or unresolved identity match, with the supporting record reference. These are internal review labels, not official agency statuses.

Keep unresolved records visible and assigned to a source owner. State whether they are excluded from the approved output pending review or handled through another documented process. A completed spreadsheet should not imply that unanswered identity questions have been settled.

Preserve the data dictionary version that defines the keys and relationships. A colleague should be able to reproduce why ten rows became nine people and locate each person's final statement association without guessing which source file was used.