# Assessment Boundary

**Packet ID:** `HIT-IR-CIGNA-PXDX-001`

## Decision process

Evaluate the post-service determination that a submitted claim was not medically necessary through the PxDx procedure-to-diagnosis review workflow described in the packet sources.

## Institutional unit

Cigna Corporation and Cigna Health and Life Insurance Company, limited to the described PxDx review process.

## Period

From reported 2022 workflow activity through the federal court order entered on 31 March 2025.

The start date is a conservative evidence boundary. It is not a claim about the exact launch date or entire commercial lifespan of PxDx.

## System terminology

Treat PxDx as an algorithmic claims-review workflow. The assessment does not require a conclusion that PxDx was machine-learning artificial intelligence.

## Human authority under examination

The relevant human role is the medical director or physician reviewer represented as reviewing, approving, signing, rejecting, modifying, or escalating a claim disposition.

## Required findings

Assign findings for:

- Counsel
- Judgment
- Command
- Correction
- Repair
- Reform
- Telemetry Integrity

## Evidence boundary

The packet consists of:

1. one investigative report describing the workflow and reported internal data;
2. one follow-up report containing institutional response and regulatory scrutiny;
3. one federal pleading-stage order.

Use the evidentiary status assigned in the source manifest. Do not convert reporting, allegation, institutional denial, or procedural plausibility into an adjudicated merits finding.

## Excluded questions

Do not determine:

- whether Cigna violated law;
- whether every PxDx decision followed one pattern;
- whether the reported workflow caused a particular patient's injury;
- whether algorithmic review is categorically inferior to human review;
- whether PxDx qualifies under a specific statutory definition of artificial intelligence;
- whether the author's existing assessment is correct.

## Sampling rule

This is a retrospective institutional-process reconstruction from the fixed public record. It is not a representative statistical sample of claim files or medical-director decisions.

## Interpretation rule

A public record may provide affirmative evidence of a practice, affirmative evidence of absence, or insufficient evidence. Scorers must distinguish these states explicitly and explain any boundary-dependent judgment.