Every value-based care organization runs on files. Eligibility rosters from health plans, claims extracts, lab and clinical feeds from practices, pharmacy fills, admission and discharge notifications. Most organizations treat the handling of those files as IT plumbing. That is a costly mistake. The quality of what comes in determines the accuracy of risk scores, the reliability of care gap lists, the success rate of encounter submissions and, ultimately, revenue. Data ingestion is a financial strategy, and it deserves the attention of the CFO as much as the CIO.

Where Bad Data Costs Money

Problems introduced at ingestion rarely announce themselves. They show up weeks or months later as symptoms that look like something else.

  • Understated risk scores. If clinical diagnoses fail to match members because of identifier mismatches, or encounters are dropped because of invalid codes, documented conditions never reach the risk adjustment submission. The organization did the clinical work and does not get paid for it.
  • Wrong gap lists. If a lab feed silently stops sending results, or a vendor changes a file layout, members appear to have open quality gaps they actually closed. Outreach teams call people who do not need calls, while members with real gaps wait.
  • Rejected encounters. Encounter submissions fail when member identifiers, provider identifiers, diagnosis codes or dates do not pass payer and CMS edits. Each rejection delays or loses the risk adjustment value of that visit.
  • Misaligned populations. Stale eligibility files lead to outreach for members who have disenrolled and missed outreach for new members, which wastes staff time and leaves care gaps open.
  • Missed transitions. When ADT feeds lag or arrive malformed, the window for timely post-discharge follow-up closes before anyone knows a patient was admitted.

Why the Stakes Are Rising

Recent policy changes make clean, linked encounter data more important, not less.

In the CY 2027 Rate Announcement issued April 6, CMS finalized excluding diagnoses from unlinked chart review records from Medicare Advantage risk score calculations, with an exception for beneficiaries switching between MA organizations. CMS also finalized excluding diagnoses from audio-only services identified by specific modifiers. The practical effect: diagnoses need to be tied to real encounters that are submitted correctly. A diagnosis that exists only in a retrospective chart review record without a linked encounter will no longer count toward payment.

At the same time, CMS announced in 2025 that it would audit all eligible Medicare Advantage contracts in each payment year through Risk Adjustment Data Validation (RADV) and accelerate completion of older audits. Every diagnosis submitted needs to be supported by documentation and traceable to a clean source. Data errors introduced at ingestion can create both underpayment and audit exposure.

What Validated Ingestion Looks Like

Validated ingestion means files are checked before they are loaded into the systems that drive risk, quality and care management decisions. A practical framework has five layers.

  1. Structure. Does the file match the expected layout? Are the right columns present, in the right formats, with valid date and code formats?
  2. Completeness. What are the fill rates for critical fields: member ID, date of birth, date of service, diagnosis codes, rendering provider NPI? A sudden drop in fill rate is often the first sign that something changed upstream.
  3. Validity. Are ICD-10 and CPT codes valid for the date of service? Are NPIs real and active? Do member identifiers match current eligibility?
  4. Reasonableness. Is the volume in line with prior files? A claims file with half the usual rows, or a lab feed with no A1c results this month, deserves investigation before load.
  5. Reconciliation. Do totals tie back to the source? Can every record be traced to the file and row it came from?

The most important principle is timing. A problem caught at the door costs minutes. The same problem discovered after load, when it has already flowed into gap lists, risk scores and submissions, can take weeks to unwind.

Source by Source: What to Watch

Claims

Watch for missing or truncated diagnosis fields, invalid place of service codes, and duplicate or adjusted claims that inflate utilization if not handled correctly.

Eligibility

Track adds, terms and changes month over month. Large unexplained swings usually indicate a file problem, not a membership change. Confirm identifiers are consistent across payers and products.

Clinical

Monitor result fields and code coverage, especially LOINC-coded lab results and structured vitals, which quality measures increasingly depend on. Free-text results do not close gaps.

Pharmacy

Validate NDCs, fill dates and days supply. Adherence measures and suspected condition logic depend on them.

ADT

Check timeliness above all. An admission notice that arrives three days late is far less useful than one that arrives in hours.

Encounter Submissions: Where It All Shows Up

Encounter submission success is one of the clearest indicators of data health, because it reflects every upstream step. When member, provider, diagnosis and service data are complete and valid, encounters pass edits. When they are not, rejections pile up, staff spend time on rework, and risk adjustment value slips.

One national provider group working with Curitics saw successful encounter submissions rise from 82% to 97%. The broader lesson is that the largest gains come from preventing rejections upstream, not from working them faster after the fact.

A Checklist for Finance and Operations Leaders

  • Do we know the fill rate for critical fields in every inbound file, and do we get alerted when it drops?
  • Are files validated before load, or are problems found downstream?
  • What is our encounter acceptance rate by payer, and what are the top rejection reasons?
  • Are all risk adjustment diagnoses linked to submitted encounters?
  • How quickly do ADT notifications reach care managers?
  • Can we trace any risk score or quality numerator back to its source file and row?

Building It Into the Platform

Curitics approaches ingestion as a front door with checks rather than a pipe. Automated ingestion validates claims, eligibility, clinical, pharmacy and ADT files before load, reporting fill rates and format issues with a row-level preview and an AI summary of what changed. Validated data feeds a unified member record, patient-level CMS-HCC V28 RAF scoring and HEDIS gap tracking, and automated 837P claims delivery sends encounters built from complete documentation.

Clean data in does not just mean fewer headaches. It means risk scores that reflect real patient complexity, gap lists that point outreach in the right direction, and encounters that get paid. That makes it one of the highest-return investments a value-based organization can make.