Dashboard

The pipeline ships with a single-file HTML record viewer for browsing the output bundle. No server and no install: open the HTML in any modern browser and load the JSON. Try the live demo on an invented ALS cohort.

Loading

1. Open dashboard.html in a current browser such as Chrome, Firefox, Safari or Edge.
2. Click the file picker.
3. Select dashboard_data.json.

The viewer reads the JSON in the browser tab; nothing is uploaded and nothing is loaded from the internet. The bundle holds identified records, so open the viewer on a computer inside the approved environment.

Hiding identifiers

Hide identifiers is ticked by default. Before anything is shown or exported, the viewer replaces each patient's name with a pseudonym (PT-NNNN, in order of patient ID), blanks the MRN, shortens the birth date to the year and drops the health system IDs. Text inside documents and notes is shown as written and can still hold names and dates. Untick the box before loading only on an approved machine, for chart review that needs names.

Two view modes

Browse Patients

The default view lists patients in the sidebar, where they can be sorted, searched and filtered to those with or without a name. Click a patient to see their tabs:

Tab	Contents
Docs	Every document linked to the patient, with buttons to filter by format (C-CDA, HTML, RTF, PDF, text)
Visits	Encounters with date, type, class and status
Problems	Conditions with onset date, code and clinical status
Meds	Medications with date, code system and dose
Procs	Procedures
Labs/Obs	Laboratory results and vital signs with value, unit and code
Allergies	Allergies with severity and clinical status
Vaccines	Immunizations with date
CarePlans	Care plans with title and status
Reports	Diagnostic reports with conclusion
Goals	Goals with status
Sections	C-CDA section narratives with full text

Each tab has an "Export CSV" button that exports just that tab for the current patient. The patient columns hold whatever the viewer shows: with identifiers hidden, the pseudonym and birth year.

Keyword Search

Toggle in the sidebar header. Type a term and the dashboard searches across:

• Every document body (CCDA, RTF, HTML, PDF text).
• Every CCDA section narrative.
• Every clinical record's display name, code, and free-text fields.

Hits show up in the main panel grouped by tab category, each with:

• Tab type (Med, Problem, Document, Section, etc.)
• Patient (pseudonym when identifiers are hidden)
• Date
• A snippet around the matched keyword, with the match highlighted.

The sidebar shows patients ordered by hit count. Click any hit to jump to that patient's record.

The "Export hits" button writes a CSV containing every hit with the patient as shown, tab category, date, code, display name, source file and snippet. Snippets are clinical text and can hold identifiers.

Keeping it inside the approved environment

The viewer is one HTML file and the data are one JSON file, so nothing needs installing. The JSON holds identified records, so both stay on computers inside the approved environment, and the viewer is opened there. The counts-only reports are the files written for review outside that environment, after a person has read them in full.

Customizing

The viewer is plain HTML, CSS and JavaScript in one file. Common tweaks:

• Colors: edit the CSS variables in the :root block of the <style> section.
• Tab tables: the per-patient tabs and their columns are built in renderDetail() in the <script> block.
• Embedding a bundle: see How the embedding works.