Measurement contract

A reporting contract linking claim cohorts, latency, unit economics and defects to investment decisions.

Product decision
Measure completed value and residual work together; never improve a rate by hiding held claims.

Document type

Measurement contract

Lifecycle stage

Operate & learn · 16 / Operate & learn

Status

Sample format · headline results from my healthcare claims work

Product decision

Measure completed value and residual work together; never improve a rate by hiding held claims.

Portfolio sample from my healthcare product experience. Requirements, thresholds and review records are illustrative; they are not employer documents or historical sign-offs.

In this document 7 sections

Reported outcomes

Paper-to-EDI conversion used OCR and AI with custom rules: images and indexes went to OnBase storage, while EDI went to automated claim processing. The conversion saved 15 minutes per claim on about 300,000 paper claims per month. Homegrown EDI shredding captured complete claim data directly into a normalized datamart before adjudication. Capturing complete claim data reduced enhancement turnaround by 50% and improved received-to-paid mapping and CMS reporting. The LifePRO engine replacement, daily reconciliation and paper-to-EDI conversion were separate projects.

1. saved 15 minutes per claim
2. about 300,000 paper claims per month
3. Homegrown EDI shredding captured complete claim data before adjudication
4. reduced enhancement turnaround by 50% improved received-to-paid mapping and CMS reporting

Cohort, window and source

Sample reporting contract: group claims by receipt week, form type and intake channel, with a stated as-of timestamp. Deduplicate by intake identity. Report eligible and all-received cohorts side by side. Keep pending cohorts visible rather than treating them as final outcomes. Use reconciled event history, operator actions, API traces and allocated operating costs.

1. group claims by receipt week, form type and intake channel with a stated as-of timestamp
2. Deduplicate by intake identity
3. Report eligible and all-received cohorts side by side Keep pending cohorts visible rather than treating them as final outcomes
4. Use reconciled event history, operator actions, API traces and allocated operating costs

Input and output measures

Compare equivalent cohorts; isolate changes in form mix and source quality.

Measure	Definition	Product decision
Input completeness	Claims with all required source fields / all unique received claims, by form and channel	Improve upstream intake or narrow eligible scope.
Net automation	Eligible received claims reaching final disposition without manual work / all eligible received claims; show pending share alongside	Expand only when completion rises without hidden rework.
Throughput	Distinct claims reaching final disposition per operating hour; show automation and manual paths separately	Plan capacity using actual end-to-end output.
Processing speed	p50/p95 time from receipt to final disposition for completed claims; pending age reported separately	Prioritize the stage driving tail delay.
Provider API latency	p50/p95 elapsed time from API request acceptance to response, separated by source-query/cache path; report timeout rate	Choose performance work without sacrificing match quality.

Defect, cost and operating measures

Numerators, denominators and observation coverage remain visible in the review.

Measure	Definition	Product decision
Extraction correction rate	Claims requiring an extraction correction / extracted claims reviewed; disclose sampling coverage	Tune capture or field review before automating more forms.
False-match rate	Confirmed incorrect provider matches / reviewed matched claims; disclose review sample and confidence limits	Pause matching changes and investigate identity risk.
EDI rejection rate	Distinct rejected transactions / distinct receiver-evaluated transactions; pending acknowledgement reported separately	Correct mappings or receiver contract, not OCR indiscriminately.
Duplicate defect rate	Confirmed unintended duplicate live submissions / distinct submitted claim identities	Stop rollout and repair retry/recovery behavior.
Cost per completed claim	Allocated extraction + compute + support + manual correction cost / distinct completed claims in period; publish backlog movement	Fund improvements with measurable total-cost benefit.
Reconciliation completeness	Received identities with exactly one known current state / all unique received identities at cutoff	Investigate every missing identity before migration expansion.
Exception age & ownership	p50/p95 age of unresolved items from hold time to cutoff; unassigned items / all open exceptions	Adjust staffing, workflow or source correction.

Mutually exclusive reporting states

Use one current state per intake identity: Received, Needs review, Ready, Submitted awaiting acknowledgement, Accepted awaiting disposition, Final disposition received, or Closed with reason. Rejection and correction events stay in history; they do not create an additional claim in the inventory. Acknowledgement acceptance is not the final adjudication outcome.

1. Received, Needs review, Ready
2. Submitted awaiting acknowledgement
3. Accepted awaiting disposition
4. Final disposition received
5. Closed with reason

Rejection and correction events stay in history; they do not create an additional claim in the inventory.

PBM measures use the right unit

For a pharmacy product, cost to fill uses the agreed fulfillment-cost scope divided by completed fills, paired with turnaround time, reversals and quality signals. I would keep that pharmacy reporting separate from medical cost per claim. Clinical and medication-safety measures require specialist definitions; claims automation volume is not a substitute.

1. cost to fill uses the agreed fulfillment-cost scope divided by completed fills paired with turnaround time, reversals and quality signals
2. keep that pharmacy reporting separate from medical cost per claim
3. Clinical and medication-safety measures require specialist definitions claims automation volume is not a substitute

How the review changes the roadmap

If source completeness falls, prioritize intake evidence. If identity exceptions dominate, prioritize baseline provider resolution. If lookup latency rises without quality change, investigate source and cache behavior. If automation rises but rework increases, hold expansion. A positive weekly volume trend cannot overrule a duplicate or identity defect.

1. If source completeness falls prioritize intake evidence
2. If identity exceptions dominate prioritize baseline provider resolution
3. If lookup latency rises without quality change investigate source and cache behavior
4. If automation rises but rework increases hold expansion

A positive weekly volume trend cannot overrule a duplicate or identity defect.