Health Product Shortages Canada Viewer

A simpler way to browse Canadian drug shortage and discontinuation reports from Health Product Shortages Canada. The data is refreshed twice a day.

Live site: https://zeptotech.github.io/drugshortages/

This is an independent tool for healthcare professionals. It is not affiliated with Health Canada. Always confirm details with the official report, which is linked from each entry. Use at your own discretion.

Features

• Shortages: tabs for Actual shortage, Anticipated shortage and Resolved / avoided (last 30 days), each with a count. The Actual shortage tab shows how many days each shortage has lasted. When a company never entered an actual start date, the expected start is used and marked "expected".
• Discontinuations: tabs for To be discontinued and Discontinued (last 30 days).
• What's new: reports updated in the past 7 days, grouped by day into new, updated, and resolved/discontinued (like the official homepage). Bars at the top filter by day and by kind of change. Where the daily change log knows exactly what changed (e.g. a new estimated end date), that is shown too.
• Company summary: one row per company with its open reports (actual, anticipated, Tier 3, to be discontinued), overdue count and %, and late count. Sortable; click a company to see its reports. Only open reports are counted, since closed ones are kept for 30 days only.
• Search by brand, ingredient, DIN, company or ATC code. Tab counts update as you type, so you can see where the matches are.
• Filters: Tier 3 only, Overdue only, drug class (ATC group).
• Tier 3 tag on the shortages with the highest potential impact.
• Overdue tag on reports whose expected date passed more than 2 days ago with no actual date entered, as on the official site (see Definitions below).
• Late tag on reports filed more than 5 days after the anticipated or actual start date (the API's late_submission flag).
• New / Updated column for reports created or updated in the last 7 days.
• Detail view: click a row (or Tab to it and press Enter) to see every field, including mitigation, company comments and Health Canada comments, plus a link to the official report.
• Shareable links: the view, tab, search and filters are kept in the URL.
• Sortable columns, copy-to-clipboard DINs, / to jump to search, and a phone layout.
• Theme button: Auto (follows your device), Light or Dark. The choice is remembered.
• Last updated time under the title, in your own time zone. It warns if the update hasn't run in over 36 hours.

How it works

There is no server. A scheduled GitHub Actions job is the backend, and the site is static files served by GitHub Pages.

Midnight + noon Eastern (GitHub Actions)          Anytime (browser)
┌───────────────────────────────────────┐   ┌───────────────────────────┐
│ scripts/fetch_and_diff.py             │   │ index.html + app.js       │
│  1. log in to the API (repo secrets)  │   │  loads data/*.json        │
│  2. pull current + recent reports     │   │  search, filters, sorting │
│  3. compare with the previous data    │   │  all run in the browser   │
│  4. write data/*.json                 │   │  (no API calls)           │
│ commit data/ → deploy GitHub Pages ───┼──►│                           │
└───────────────────────────────────────┘   └───────────────────────────┘

The API password lives only in GitHub Actions secrets. The browser never talks to the API.

What the daily job fetches

Statuses	How	Requests
active_confirmed, anticipated_shortage, to_be_discontinued	Full pull, sorted by id ascending, 100 per page (as the API requires for multi-report extraction)	~25
resolved, avoided_shortage, discontinued, reversed	Updated in the last 30 days only, newest first	~5
Reports that left the current lists since the last run	One detail lookup each, to see what they became	a few

The API allows 1,000 requests per hour, and a run uses about 30–45.

The job keeps only the English fields the site needs, compares them with the previous snapshot, and records what changed. It aborts without touching the data if the login fails, the API errors, or a status count suddenly drops by more than 5% from the previous run's (use --force to override a real drop). When that check stops a run, the log lists the current status of a sample of the missing reports, which shows whether the drop is real or the API left them out.

Data files

File	Contents
data/shortages.json	Current shortages plus those resolved or avoided in the last 30 days
data/discontinuations.json	Upcoming discontinuations plus those discontinued or reversed in the last 30 days
data/meta.json	Last check time and counts per status
data/changes/YYYY-MM-DD.json	New, updated and status-changed reports for that day (last 30 days; older files are deleted)
data/changes/index.json	Days with changes in the last 30 days, newest first

Each report is written on its own line, so the daily data commits show exactly which reports changed.

Definitions

These follow the official site's summary report:

• Late: the report was created more than 5 days after the anticipated or actual start date. This comes from the API.
• Overdue: an anticipated or estimated date in the report passed more than 2 calendar days ago and no corresponding actual date has been entered. The site works this out in the browser each day.
  • Anticipated shortage: the expected or actual start date has passed.
  • Actual shortage: the expected start date has passed and there is no actual start date, or the estimated end date has passed and there is no actual end date.
  • To be discontinued: the expected discontinuation date has passed and there is no actual discontinuation date.

Setup

1. API account: create and verify an account on healthproductshortages.ca.
2. Secrets: in the repo, go to Settings → Secrets and variables → Actions and add HPS_EMAIL and HPS_PASSWORD.
3. Pages: go to Settings → Pages → Source and choose GitHub Actions.
4. Run it: go to Actions → Update data and deploy → Run workflow, or wait for the next scheduled run.

The workflow (.github/workflows/update-data.yml):

• On schedule or "Run workflow": fetches data, commits data/ if it changed, and deploys.
• On a push to main: only redeploys the site.

The job runs twice a day, 12 hours apart (GitHub cron always uses UTC):

• 0 5 * * *: 05:00 UTC, which is midnight EST in winter and 1 AM EDT in summer.
• 0 17 * * *: 17:00 UTC, which is noon EST in winter and 1 PM EDT in summer.

A second run on the same day adds to that day's change log rather than replacing it, so "What's new" still shows one list per day. GitHub can start scheduled runs late when it is busy, sometimes by hours, so the actual time varies.

Local development

Fetch data, using only the Python 3.9+ standard library (nothing to install):

read "HPS_EMAIL?Email: "; read -s "HPS_PASSWORD?Password: "; echo
HPS_EMAIL=$HPS_EMAIL HPS_PASSWORD=$HPS_PASSWORD python3 scripts/fetch_and_diff.py

Preview the site:

python3 -m http.server 8000
# open http://localhost:8000

The bot commits data daily, so run git pull before starting work.

The code is commented for readers new to it. app.js starts with an overview of how the page works.

Project layout

index.html, app.js, styles.css      the website (no build step)
scripts/fetch_and_diff.py           daily fetch, compare and write
.github/workflows/update-data.yml   schedule, commit and deploy
data/                               generated, committed by the bot

Data source

Data comes from the Health Product Shortages Canada public API. Only English fields are used.