> ## Documentation Index
> Fetch the complete documentation index at: https://docs.encoreos.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Denial Prediction Dashboard

> Monitor denial prediction model performance, payer accuracy breakdown, trend data, and revenue impact over a rolling 90-day window.

The Denial Prediction Dashboard displays AI model performance metrics for claim denial prediction at route `/pm/denial-prediction/dashboard`.

## Overview

The page fetches dashboard data via `useDenialPredictionDashboard` for the current organization, with a hardcoded 90-day lookback window. If no active model exists and no predictions have been made, an empty state prompts that at least 1,000 historical claims are needed. When data is available, the layout shows: an active model badge (version and promotion date), `MetricsKPICards` (summary metrics), `MetricsTrendChart` (full-width trend chart over the period), and a two-column grid with `PayerAccuracyChart` and `RevenueImpactCard`. All chart and card components are from `src/cores/pm/components/denial-prediction/`.

## Who it's for

Requires permission `PM_PERMISSIONS.DENIAL_PREDICTIONS_DASHBOARD` (enforced via `RequirePermission` wrapping the page content).

## Before you start

* You need `pm.denial-predictions.dashboard` permission.
* The dashboard requires an active trained model and sufficient historical data to display metrics.

## Steps

<Steps>
  <Step title="Open the Denial Prediction Dashboard">
    Navigate to `/pm/denial-prediction/dashboard`.
  </Step>

  <Step title="Review model status">
    The badge in the header shows the active model version and when it was promoted. If no active model exists, a "No active model" badge appears.
  </Step>

  <Step title="Review KPI metrics">
    The MetricsKPICards show current prediction performance summary statistics.
  </Step>

  <Step title="Analyze trends">
    The MetricsTrendChart shows how prediction metrics have changed over the 90-day window.
  </Step>

  <Step title="Review payer breakdown and revenue impact">
    The two-column grid shows payer-level prediction accuracy and revenue impact summary.
  </Step>
</Steps>

## Key concepts

| Term                | Meaning in code                                                                |
| ------------------- | ------------------------------------------------------------------------------ |
| `active_model`      | The currently promoted prediction model with `model_version` and `promoted_at` |
| `current_metrics`   | Aggregate prediction performance metrics for the period                        |
| `total_predictions` | Count of predictions made; must be > 0 for dashboard to render                 |
| `trend_data`        | Time-series data fed to `MetricsTrendChart`                                    |
| `payer_breakdown`   | Per-payer accuracy data fed to `PayerAccuracyChart`                            |
| `revenue_impact`    | Revenue impact summary fed to `RevenueImpactCard`                              |

## Related

<Columns cols={2}>
  <Card title="Practice Management" icon="briefcase" href="/pm/overview">
    Practice Management core overview.
  </Card>

  <Card title="Governance & parity" icon="scale-balanced" href="/governance/index">
    Documentation coverage and governance.
  </Card>
</Columns>

<Note>
  This page documents shipped product behavior. It is not medical, legal, or
  billing advice. Verify against your organization's policies and applicable
  regulations before using it for clinical, compliance, or billing decisions.
  Protected health information (PHI) shown in the product is governed by your
  tenant's access controls and is never exposed in this documentation.
</Note>

<Accordion title="Documentation sources">
  * src/routes/pm.tsx
  * src/cores/pm/pages/DenialPredictionDashboardPage.tsx
  * src/cores/pm/hooks/useDenialPredictionDashboard.ts
</Accordion>