Global health data connectors you can add to Claude today
Written in October 2026. These connectors are all about a year old or newer and changing quickly, so check each source's own page before relying on this.
Two of the best known sources in global health, IHME and The DHS Program, now have connectors in Claude's directory, so you can ask about their data in plain language and get published numbers back. They're easy to set up, but they don't measure the same things, and that difference matters more than which one is easier.
Who these groups are
- The DHS Program has supported nationally representative household surveys since 1984, more than 400 of them in over 90 countries, mostly low- and middle-income. It's implemented by ICF.
- IHME, the Institute for Health Metrics and Evaluation, is an independent research organization at the University of Washington School of Medicine, founded in 2007. It runs the Global Burden of Disease study, with more than 20,000 collaborators in 167 countries and territories.
- The World Bank Group works to end poverty and supports development in about 130 countries. Data360 is where it brings its development data together.
What's available
| Demographic and Health Surveys | IHME GBD | World Bank Data360 | |
|---|---|---|---|
| Name in Claude | Demographic and Health Surveys | IHME GBD | Not listed, add as a custom connector |
| Setup | Easy. Search, connect, authorize | Medium. Also needs a free IHME account | Hard. Add it by address (below) |
| What the numbers are | Results from nationally representative household surveys | Modeled estimates that combine many sources | Indicators from across the World Bank Group and partners |
| Coverage | 400+ surveys in 90+ countries, for the years each survey ran | 204 countries and territories, 1990 to 2023, with subnational detail for 17 countries | 200+ economies, about 300 million data points |
| Topics | Fertility, family planning, maternal and child health, gender, HIV, malaria, nutrition | Deaths, DALYs, risk factors, life expectancy, fertility, population | Digital, infrastructure, people, planet and prosperity, including health |
| Also works in | ChatGPT and other MCP assistants | Claude only | Any MCP client |
| Cost | Free | Free for non-commercial use | Free and open source |
In Claude, open Connectors in settings, search the name exactly as it appears above and click Connect. For the World Bank, add a custom connector with this address: https://maimcpext.worldbank.org/ext/data360/mcp. Its code is on GitHub, last updated in September 2026, and its docs note the hosted version may require a token.
The difference that matters
IHME's numbers come from GBD models that combine many sources into comparable estimates for every country and year in the study. DHS numbers come straight from household surveys, for the countries and years a survey was run. So the same question, like under-5 mortality in a given country, can return different values. That's not an error, it's a different method, and it's worth knowing which one you're citing.
Data360 is different again. It goes wide across development topics rather than deep on health, which helps when health sits next to poverty, education or climate in the same question.
Try these
- IHME: "Please generate a Disease Profile for diabetes in Algeria."
- DHS: "Please show a visualization of modern contraceptive use and fertility trends in Malawi."
Both prompts come from the publishers' own pages.
Worth watching
The UN launched the UN System Data Commons in September 2026 at data.un.org, with statistics from about 20 UN agencies and support for MCP. There isn't a documented way to connect to it from Claude yet.
A few habits that help
- Ask which indicator, survey or release a number came from. These connectors can tell you.
- Save the question with the answer. These sources update, and last year's number may not match today's.
- Have a person check the output before it goes in a report.
None of these replace actually getting the data and doing your own analysis in deterministic tools like Python or Excel. They're a fast way to explore and find the right source. The analysis you stand behind still happens in tools that give the same answer every time.