John Snow’s cholera theory faced exceptions. Source details made his case.

Before UDS mapping begins, preserve every original program label, record each transformation, and keep the referral trail tied to its source. CaminoCare can document navigation encounters and referral outcomes, while the final UDS mapping, validation, and submission remain with the FQHC’s reporting team.

In 1854, physician John Snow faced a data problem in London. Cholera deaths clustered around the Broad Street water pump, but the prevailing explanation blamed foul air. Snow needed more than a pattern on a map. He traced deaths to addresses, investigated where residents obtained water, and examined cases that appeared to contradict his theory.

His published account, On the Mode of Communication of Cholera, records important exceptions. Workers at a nearby brewery largely avoided the outbreak, and the brewery had its own water source. A workhouse near the pump also had relatively few deaths and its own well. Those source details strengthened the argument because Snow did not flatten every nearby resident into the same category.

The lesson for an outreach lead preparing data for UDS work is practical: preserve what happened before deciding where it belongs.

Keep the original label beside the mapped label

Program names often grow locally. One team records “community outreach.” Another uses “care coordination.” A third writes “referral follow-up,” even when staff performed similar work. The temptation is to replace those labels immediately with standardized reporting categories.

Keep both.

Your working file should retain the exact source label, the proposed mapped value, and the reason for the change. Add the person or rule responsible for the decision and the date it was made. If a label changes later, preserve the earlier value rather than overwriting it.

A simple mapping record might contain:

• The original program, service, or encounter label exactly as entered.
• The proposed UDS destination field or category.
• The local definition used when the source record was created.
• The mapping rule and workbook version applied.
• Any unresolved question that requires reporting or compliance review.

This protects the team when two familiar phrases turn out to mean different things. It also lets reviewers reverse a mapping without reconstructing the source data from email threads.

Snow’s work held up because the address, water source, and outcome remained distinguishable. The map supported the investigation; it did not erase the observations underneath it.

Preserve the referral trail before counting outcomes

A referral record needs enough history to show what staff actually did. At minimum, retain the originating encounter, the care location offered, the referral date, the navigator involved, and the latest known outcome.

Use outcome labels with narrow meanings. “Referral created” should not become “care received.” “Client called” should not become “appointment completed.” If the team does not know whether the person reached care, record that uncertainty plainly.

CaminoCare supports this operational trail through minimal-PII cases, attached clinic referrals, navigation encounters, and records of whether a client reached care. Teams can also export monthly encounter CSVs for their own workflows. Suggested billing codes are carefully labelled as suggestions, so a qualified reviewer still decides what applies.

That distinction matters. CaminoCare provides navigation documentation. It does not calculate, certify, validate, or submit an FQHC’s UDS report. Your reporting team must apply the current UDS Data Mapping and Validation Workbook, resolve local definitions, run validation checks, and approve the submission.

Build a review queue for labels that do not map cleanly

Do not force uncertain records into the nearest category to make the spreadsheet look complete. Create a short exception queue and resolve it before aggregation.

Flag records when:

• The same source label has been used for different activities.
• Two labels appear different but may describe the same work.
• A referral outcome lacks supporting follow-up.
• A required source field is blank or contains free text.
• A proposed mapping depends on an assumption rather than a documented rule.
• A code suggestion has not been reviewed by the appropriate staff member.

Review the queue with outreach, operations, billing, and UDS reporting staff together. The person who understands how the work happened may interpret the label differently from the person who understands the reporting definition. Both views are needed.

Freeze the approved crosswalk for the reporting cycle. If a rule changes, version it and document which records were remapped. This creates a repeatable process next year and gives reviewers a clear path from the reported value back to the encounter.

Run a small trace before the full mapping

Choose a handful of records from different programs and trace each one forward: source label, encounter, referral, outcome, proposed mapping, and exported row. Then trace each reported value backward to its original record.

A broken link in a five-record sample is cheaper to fix than the same break discovered after a full aggregation. Pay special attention to records that changed owners, crossed programs, or received a corrected referral outcome.

John Snow’s investigation became persuasive because he examined the exceptions instead of hiding them. Apply the same discipline here. Preserve the raw label, keep the referral history, and make every transformation reversible before the first UDS mapping rule touches the data.