Canadian Healthy Aging Policy Observatory
A research-infrastructure platform that monitors, quantifies, and evaluates the effect of aging policy across Canadian governments — starting with Nova Scotia and the federal level as a replicable template.
What this is
The Observatory is a research instrument, not a news site. It is part of the Healthy Aging Intelligence Lab (HAIL) — a long-horizon effort to build durable, reusable assets for studying how policy shapes healthy aging: a versioned policy knowledge base, a defensible indicator system, reproducible data pipelines, and a literature knowledge base. The goal is to support rigorous research and, over time, to grow into shared infrastructure that others can build on.
What it does
Policy Library
A structured, timestamped record of aging-related policies across Canadian governments — department, budget, target population, KPIs, lifecycle, and AI-assisted summaries.
Data Hub
Versioned indicators drawn from Statistics Canada, CIHI, and Nova Scotia open data — each observation carries full lineage (source, dataset version, checksum) for reproducibility.
HAPI
The Healthy Aging Policy Index — a transparent, weight-audited composite across six domains (Health, Independence, Social Participation, Financial Security, Care Access, Digital Inclusion).
Policy Analytics
Quasi-experimental analysis — interrupted time series around real policy interventions — that is careful to separate association from causation.
AI Research Assistant
Retrieval over the policy library and literature to accelerate literature reviews, always with citations back to the underlying source.
Data & methodology
- Provenance first. Every indicator value is stored with its source, dataset version, and content checksum, so any number on the site can be traced back to the exact upstream table it came from.
- Transparent index. HAPI’s normalization ranges and domain weights are documented method choices, sensitivity-tested against equal / expert / empirical weighting schemes rather than asserted.
- Honest about causation. Descriptive trends are labelled as association; causal claims are reserved for quasi-experimental designs (interrupted time series, with their assumptions and limits stated). Association is never dressed up as cause.
- Open by default. The methodology, data catalog, and code are public on GitHub for scrutiny and replication.
Glossary
The abbreviations, domain terms, and coined names used across the platform.
The platform
- Observatory
- How we describe this platform: a research instrument that continuously monitors, quantifies, and evaluates aging policy — not a news site or a static report.
- HAILHealthy Aging Intelligence Lab
- The long-horizon research lab this Observatory belongs to, building durable, reusable assets for studying how policy shapes healthy aging.
- HAPIHealthy Aging Policy Index
- Our own composite index (0–100) scoring how well a jurisdiction supports healthy aging, across six domains, with transparent weights and normalization.
- Evidence pack
- The exact set of policies, indicators, and literature the AI assistant is given for a topic. Every claim in a generated draft cites an item from this pack — nothing else.
HAPI domains
- Six domains
- HAPI is built from Health, Independence, Social Participation, Financial Security, Care Access, and Digital Inclusion — each scored from documented indicators and combined by audited weights.
- Indicator
- A single measured signal (e.g. an ADL-limitation rate) with a fixed definition, data source, normalization range, and direction (higher-is-better or lower-is-better).
- Care Access
- The v1-priority domain — access to home care and long-term care — chosen because it maps to high-signal CIHI/IRRS and Nova Scotia data.
Methods & analytics
- Association vs. Causal
- Two labels we apply strictly. Association = a descriptive trend or correlation. Causal = a claim backed by a quasi-experimental design. The site never upgrades one to the other.
- ITSInterrupted Time Series
- A quasi-experimental design that models an outcome's trend before vs. after a policy's effective date, used for careful causal claims.
- DiDDifference-in-Differences
- Compares a jurisdiction that adopted a policy against one that did not, before vs. after — relying on a parallel-trends assumption.
- Synthetic Control
- Builds a weighted 'synthetic' comparison jurisdiction from a donor pool to estimate what would have happened without the policy.
- RAGRetrieval-Augmented Generation
- The assistant retrieves real policies and literature first, then drafts from that grounded material — so output stays traceable rather than invented.
- KPIKey Performance Indicator
- A target metric attached to a policy record (e.g. a stated wait-time goal).
Data & provenance
- Lineage
- The chain that ties any number on the site back to its origin: Observation → DatasetVersion → DataSource. Shown on the Data Hub for reproducibility.
- Observation
- One immutable data point — an indicator's value for a jurisdiction at a point in time — stored with a checksum so it can't silently change.
- DatasetVersion
- A specific retrieval of an upstream dataset (with its retrieval date and content checksum). Re-running with unchanged upstream is a no-op.
- ETLExtract, Transform, Load
- The pipeline stages that pull raw data from a source, clean it, and load validated observations into the database.
Sources & institutions
- StatCanStatistics Canada
- Canada's national statistical agency — a primary source for demography and outcome indicators.
- WDSWeb Data Service
- Statistics Canada's API for programmatically retrieving data tables.
- CIHICanadian Institute for Health Information
- National source for health-system data, including home care and long-term care reporting.
- IRRSIntegrated interRAI Reporting System
- CIHI's consolidated home-care/long-term-care reporting system that replaces the legacy HCRS and CCRS. Care Access data follows IRRS going forward.
- Census
- Statistics Canada's national census — the backbone for population denominators and internet-use (Digital Inclusion) tables.
Geography & domain
- NSNova Scotia
- The first province modelled in depth — chosen as a replicable template that later extends to the rest of Canada.
- Federal
- The Government of Canada level, tracked alongside Nova Scotia so provincial and national policy can be compared.
- LTCLong-Term Care
- Residential care for older adults who can no longer live independently — a core focus of the Care Access domain.
- ADLActivities of Daily Living
- Basic self-care tasks (bathing, dressing, eating). Limitation rates are a standard measure of functional independence.
- PIPEDAPersonal Information Protection and Electronic Documents Act
- Canada's federal privacy law — the baseline the platform's privacy and data-handling design respects.
Get in touch
Questions, collaborations, or data requests are welcome — see the contact page.