Australian Institute of Health and Welfare
This server provides plain-English, structured access to Australian Institute of Health and Welfare (AIHW) statistics via six MCP tools.
search_datasets — fuzzy-search the curated AIHW catalog by keyword.
describe_dataset — explore a dataset's filterable dimensions, measures, units, and source URLs.
get_data — query datasets with filters, measures, period ranges, and output formats (records, series, csv).
latest — quickly retrieve the most recent observation(s) per measure.
top_n — rank rows by a measure and return top or bottom N results.
list_curated — enumerate all curated dataset IDs.
Coverage includes long-term mortality (GRIM), regional mortality (MORT), cancer incidence/mortality, health expenditure, youth justice detention, and public hospitals. It supports time-series filtering, multi-value filters, canonical state/postcode aliases, and includes source URLs and CC-BY attribution in every response.
aihw-mcp
mcp-name: io.ausdata/aihw-mcp
MCP server for Australian Institute of Health and Welfare statistics. Plain-English access to long-term mortality (GRIM), regional mortality (MORT), cancer incidence and mortality (ACIM), national health expenditure, youth justice detention, and the public hospitals register — all from a single uvx command.
Hosted access? For cross-source queries, webhooks, an always-on REST API, and a uniform response envelope across all 9 sources, see ausdata.io — free tier available (500 calls/mo, no card).
"How have diabetes deaths changed since 1980?"
"What's the age-standardised mortality rate in the Sydney - Inner West SA3?"
"Show me breast cancer incidence in women aged 50–54 over time."
"How much did Australia spend on public hospitals in 2022-23?"
"How many young people are in detention in NSW vs VIC?"
"List all Principal referral hospitals in Queensland."
"Top 5 causes of death in 2023."Sister to abs-mcp (Australian Bureau of Statistics), rba-mcp (Reserve Bank of Australia), ato-mcp (Australian Taxation Office), and au-weather-mcp (Australian weather). The five together cover the macro / regulator / tax / health / climate layer of Australian official data.
Install
# Run on demand via uvx (recommended)
uvx --upgrade aihw-mcp
# Or install permanently
pip install aihw-mcpClaude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"aihw": { "command": "uvx", "args": ["--upgrade", "aihw-mcp"] }
}
}Why
--upgrade?uvx aihw-mcp(without the flag) uses whatever wheel is cached and never adopts new PyPI releases on its own.--upgrademakes uvx check PyPI on each launch and pull a newer release if one exists. To verify which version is currently serving you, look at theserver_versionfield on anyDataResponse.
Claude Code / Cursor
claude mcp add aihw --command uvx --args -- --upgrade aihw-mcpRelated MCP server: aemo-mcp
Auto-updating data
Beyond the wheel-level --upgrade, the server has a second auto-update path inside the data layer: when AIHW refreshes GRIM with another year of deaths data or publishes a new MORT release, aihw-mcp resolves the new resource URL via data.gov.au's CKAN API at fetch time and uses the freshest match. Hard-coded YAML URLs are the safe fallback if discovery fails. You do not need to wait for a new wheel release to get new yearly data — just delete ~/.aihw-mcp/cache.db to force a refresh, or wait for the 7-day TTL to expire.
What it exposes
Six tools, all plain-English in, structured out:
Tool | Purpose |
| Fuzzy-search the curated catalog by keyword |
| List a dataset's filterable dimensions and returnable measures |
| Query with |
| Last observation per measure (shortcut) |
| Rank rows by a measure, return top (or bottom) N |
| Enumerate the curated dataset IDs |
Every response is the same shape — dataset_id, dataset_name, query, period, unit, row_count, records, aihw_url, attribution, server_version — across every curated dataset.
Time-series datasets (GRIM, MORT, ACIM, Health Expenditure, Youth Justice) accept start_period and end_period on get_data — e.g. start_period="2000", end_period="2010" narrows GRIM to that decade. latest() returns the most-recent observation per measure, sorted by the dataset's declared period dimension (not by source row order). Error messages include fuzzy "did you mean?" suggestions when you typo a filter or measure name.
Curated datasets (6 in v0.1)
ID | What it is | Period | Coverage |
| National long-term mortality: deaths × cause × year × sex × age band | 1907 → present | ~370k rows, 3 measures |
| Regional mortality: deaths + premature/avoidable deaths × State / SA3 / SA4 / PHN | 2019 → present | ~15k rows, 15 measures |
| Cancer incidence + mortality counts × year × sex × type × 5-year age band | 1968 → present | ~9k rows, 19 age columns |
| Real expenditure by financial year × state × area × source (Government / non-Govt) | 1997-98 → present | ~7k rows, AUD millions |
| Avg nightly youth-detention pop × quarter × state × sex × Indigenous × legal status | 2008 → present | ~42k rows |
| Directory of every public hospital × state × peer group × remoteness × LHN | 2016-17 reference | ~700 hospitals |
Adding a new dataset is a single YAML drop into src/aihw_mcp/data/curated/ — see CONTRIBUTING.md.
Example queries (paste into Claude)
Cross-source compatibility. The
statefilter on HEALTH_EXPENDITURE, PUBLIC_HOSPITALS, YOUTH_JUSTICE_DETENTION, and other location-aware datasets accepts canonical state codes ("NSW"), full names ("New South Wales"), case-insensitive variants ("nsw"), ISO 3166-2 ("AU-NSW"), and 4-digit postcodes ("2000"→ NSW). Powered byaus-identity— the same input format works across abs-mcp, ato-mcp, apra-mcp, aihw-mcp, and asic-mcp.
Public-health research: "For GRIM_DEATHS, give me the deaths and age-standardised rate for 'Diabetes' for Persons every year from 1980 to the latest, so I can chart the trajectory."
Health-tech / regional analysis: "Using MORT_GEOGRAPHY, list the 10 SA3 regions with the highest age-standardised mortality rate for Persons in the most recent year."
Oncology: "From CANCER_INCIDENCE_MORTALITY, give me Breast cancer incidence in Females across the 50–54 age band for every available year, plus the same age band's mortality."
Health-policy: "From HEALTH_EXPENDITURE, what was the real spend on 'Public hospitals' in NSW in 2022-23, broken down by broad source (Government vs Non-government)?"
Criminal-justice tech: "Using YOUTH_JUSTICE_DETENTION, compare the average nightly youth-detention population in NSW vs VIC in 'Jun qtr 2017', for both Indigenous and Total."
Hospital-tech / market intel: "From PUBLIC_HOSPITALS, list every 'Principal referral' hospital with their state and Local Hospital Network. Then count how many there are per state."
Each prompt resolves to one or two get_data / top_n calls. The response includes the source URL so the agent can cite it back.
Architecture
Same shape as the sister packages — client → cache → parsing → shaping → server:
client.pywrapshttpxwith a SQLite-backed disk cache (per-resource TTL).parsing.pyreads CSV (viapandas) and XLSX (viaopenpyxl/pandas). Header rows + sheet names live in the curated YAML so future format quirks are a YAML edit, not a code change.curated.pyloads dataset specs fromdata/curated/*.yaml— each one declares its dimensions, measures, dimension value enums, source/download URLs, format, and parse layout.shaping.pytransforms the parsed DataFrame intoDataResponse(records / series / csv).server.pyis the FastMCP entrypoint — six tools, full input validation with helpful "Try X" hints on error.
Cache lives under ~/.aihw-mcp/cache.db. Most AIHW datasets refresh once a year; the TTLs are tuned for that cadence.
Attribution
Data sourced from the Australian Institute of Health and Welfare (AIHW) via data.gov.au. Licensed under Creative Commons Attribution 3.0 Australia (CC BY 3.0 AU). The MCP server is MIT-licensed; the data carries the upstream CC-BY 3.0 AU licence, which is echoed in every response's attribution field.
Sister MCPs (Australian Public Data portfolio)
Want all 9 sources behind one REST API? The hosted gateway at ausdata.io adds cross-source joins, full history, webhooks, and HMAC-signed responses on top of these MCPs — free tier (500 calls/mo, no card).
abs-mcp — Australian Bureau of Statistics (CPI, unemployment, ERP, building approvals)
rba-mcp — Reserve Bank of Australia (cash rate, lending stats, exchange rates)
ato-mcp — Australian Taxation Office (tax stats, ACNC charities)
apra-mcp — Australian Prudential Regulation Authority (banking, insurance, super)
aihw-mcp — this one. National mortality, regional health, cancer, expenditure, youth justice, hospitals.
asic-mcp — Australian Securities and Investments Commission (company registers)
aemo-mcp — Australian Energy Market Operator (NEM dispatch, spot prices, generation)
au-weather-mcp — Open-Meteo (Bureau of Meteorology aggregator)
wgea-mcp — Workplace Gender Equality Agency
aus-identity — Postcode / state / ABN normalisation helper used by all sisters
The portfolio is designed to compose: an agent can ask for "unemployment + cash rate + median income + mortality + climate" for postcode 2000 and one shot fans out across multiple MCPs.
Roadmap (next iterations)
v0.2: MORT_TABLE_2 (deaths by leading cause × region); intercountry adoptions; AIHW mental-health data tables (XLSX path)
v0.3: hosted version with x402 per-call paywall; programmatic SEO pages
v0.4: listing on MCPay + Apify; paid tier for high-volume agent users
CHANGELOG tracks every release.
Development
git clone https://github.com/Bigred97/aihw-mcp.git
cd aihw-mcp
uv venv
uv pip install -e ".[dev]"
pytest # unit tests, no network
pytest -m live # integration tests against data.gov.auIssues, ideas, and contributions welcome: github.com/Bigred97/aihw-mcp/issues.
Available Tools
6 toolsdescribe_datasetA
Describe a dataset's filterable dimensions, returnable measures, units, and source.
Use this before calling get_data on a new dataset — it tells you the valid filter keys ('sex', 'year', 'state'), the valid filter values ('Females', 'Males', 'Persons'), the measure aliases ('deaths', 'crude_rate_per_100000'), and the canonical source URL.
Returns: DatasetDetail with id, name, description, period_coverage, list of dimensions, list of measures (each with key, source_column, unit, description), and source_url + download_url.
Inputs: dataset_id (required). Returns: describe_dataset shape.
get_dataA
Query a curated AIHW dataset and return observations.
Examples: # Deaths from diabetes, all years and sexes resp = await get_data( "GRIM_DEATHS", filters={"cause_of_death": "Diabetes"}, measures="deaths", )
# Breast cancer incidence in females over time
resp = await get_data(
"CANCER_INCIDENCE_MORTALITY",
filters={"cancer_type": "Breast cancer", "sex": "Female", "type": "Incidence"},
)
# Public hospitals in NSW with peer group "Principal referral"
resp = await get_data(
"PUBLIC_HOSPITALS",
filters={"state": "NSW", "peer_group_name": "Principal referral"},
)Returns: DataResponse with records (or csv), unit, period bounds, row_count, source URL, and CC-BY attribution.
Inputs: dataset_id (required), format, filters, measures, +2 more. Returns: get_data / latest shape — uniform across all curated datasets. `records` carrie…
latestA
Return the most recent observation(s) per measure for a dataset.
For transposed time-series tables this trims to the most-recent period. For wide single-year tables (most AIHW datasets) it returns the same shape as get_data — there is only one period in those tables.
When the curated YAML declares a headline_slice, those filter values
are applied automatically so the no-filter call returns ONE canonical
headline row (e.g. HEALTH_EXPENDITURE → national total, all areas, all
sources). User-supplied filters override the headline_slice per-key —
latest("HEALTH_EXPENDITURE", filters={"state": "NSW"}) keeps the
area/source defaults and returns the NSW latest-year total.
Examples: # Latest year of GRIM data for All causes combined resp = await latest("GRIM_DEATHS", filters={"cause_of_death": "All causes combined"})
Inputs: dataset_id (required), filters, measures. Returns: get_data / latest shape — uniform across all curated datasets. `records` carrie…
list_curatedA
List every curated dataset ID in this version of aihw-mcp.
These are the datasets where get_data accepts plain-English filter keys and returns aliased, well-typed measure columns. Each ID is documented via describe_dataset.
Returns: Sorted list of dataset IDs.
Inputs: no named parameters. Returns: result
search_datasetsA
Fuzzy-search the curated AIHW dataset catalog.
All datasets ship hand-curated in v0.1: long-term mortality (GRIM), regional mortality (MORT), cancer incidence and mortality, health expenditure, youth justice detention, and the public hospitals register.
Examples: # Find a dataset that gives deaths by cause results = await search_datasets("mortality cause of death") # → [{id: 'GRIM_DEATHS', name: 'GRIM — long-term mortality', ...}]
# Discover what's available on cancer
results = await search_datasets("cancer")Returns: List of DatasetSummary (id, name, description, update_frequency, is_curated), ranked by relevance.
Inputs: query (required), limit. Returns: result
top_nA
Return the N rows with the largest (or smallest) value of a measure.
This is the most common agent workflow: "show me the top 10 X by Y". Without this tool, an agent would call get_data, receive the full table, and then sort/slice locally — wasting tokens and turns. top_n does the rank server-side and returns only the requested rows.
Examples: # Top 10 causes of death in 2023 (Persons) top_n("GRIM_DEATHS", "deaths", n=10, filters={"sex": "Persons", "year": "2023"})
# 20 SA3 regions with the highest age-standardised mortality
top_n("MORT_GEOGRAPHY", "age_standardised_rate_per_100000",
filters={"category": "Statistical Area Level 3 (SA3)",
"sex": "Persons", "year": "2023"}, n=20)
# 5 lowest-funded health expenditure areas in NSW
top_n("HEALTH_EXPENDITURE", "real_expenditure_millions",
filters={"state": "NSW", "financial_year": "2022-23"},
n=5, direction="bottom")Returns:
DataResponse with at most n records, sorted by measure value
in the requested direction. Other fields (period, unit, attribution)
match a regular get_data call.
Inputs: measure (required), dataset_id (required), n, filters, +1 more. Returns: get_data / latest shape — uniform across all curated datasets. `records` carrie…
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 6 tools
Each tool has a clearly distinct purpose: describe_dataset explains schema, get_data queries data, latest returns most recent observations, list_curated lists dataset IDs, search_datasets searches catalog, top_n ranks rows. No ambiguity.
All tool names use consistent snake_case pattern (describe_dataset, get_data, latest, list_curated, search_datasets, top_n). No mixing of conventions.
6 tools is well-scoped for a dataset querying server. Each tool covers a necessary function without redundancy.
The tool surface covers full read lifecycle: discovery (list, search, describe), query (get, latest, top_n). No obvious gaps for the intended purpose.
Maintenance
Related MCP Connectors
data.gov.au MCP — Australia open-data CKAN catalogue.
Hosted MCP server exposing US hospital procedure cost data to AI assistants
WHO GHO MCP — World Health Organization Global Health Observatory (free, no auth)
Hosted MCP server for finding authoritative primary data sources and official portals.
Related MCP Servers
- AlicenseAqualityBmaintenanceOne-call Australian workplace gender data plumbing via WGEA — cited responses for workforce, pay-gap and inclusion analysis, not a data broker.2640 PyPIMIT
- AlicenseAqualityFmaintenanceOne-call Australian energy-market plumbing via AEMO — cited, structured responses for market data and analysis, not a data broker.5129 PyPIMIT
- AlicenseAqualityFmaintenanceCited Australian stats via the ausdata.io gateway — stable AU.* series IDs, source_url + retrieved_at on every response. Free tier. Not a data broker; upgrade for Embed / signed / webhooks.2832 npm2MIT
- AlicenseNot gradedqualityDmaintenanceAn MCP server that provides comprehensive medical information by querying authoritative APIs including FDA, WHO, PubMed, RxNorm, and Australia's PBS, enabling drug searches, health statistics, medical literature, drug nomenclature, and PBS subsidy queries.MIT