Building a Local Health-Data Map of West Bengal, Privacy First

When you screen people across a district over time, you are not only helping each person in front of you. You are accumulating a picture of the community's health that almost nobody currently has — where anaemia clusters, where blood-sugar risk is rising, which blocks are being missed. That picture has real value for deciding where to send scarce resources. It also carries real risk if handled carelessly. Both things are true, and the second governs how we do the first.

Why local health data is scarce and valuable

Most detailed health data in India is concentrated in cities and large institutions. The granular reality of a rural block in North 24 Parganas — how prevalent anaemia actually is street by street, how it changes season to season, which villages never get screened — is mostly invisible. Resources get allocated on coarse averages because the fine-grained map does not exist.

Screening at community scale starts to fill that in. Done over time, it can show where the burden is heaviest, where a problem is emerging before it becomes a crisis, and where the gaps in coverage are. That lets resources be aimed rather than sprayed. It is genuinely useful, and we do not want to pretend otherwise out of false modesty.

Why privacy is the governing constraint, not an afterthought

Health data is among the most sensitive information a person has. A map of community health is built from individual people's results, and those individuals did not consent to be exposed so that we could have a nice dataset. So privacy is not a feature we add at the end. It is the constraint the whole thing is built inside.

A few principles guide how we handle this. Sensitive raw data stays protected and is not casually moved around. Any data that flows into a research or planning dataset is anonymised and aggregated, so the map shows patterns across a community rather than identifiable individuals. Where individual data is used for research, it is with explicit, informed, revocable consent, explained in the person's own language rather than buried in terms nobody reads. We operate under India's data-protection framework, the DPDP Act, and we do not sell data to third parties or insurers. That last point is a commitment, not a marketing line.

The line we will not cross

The distinction that matters is between using aggregated patterns to serve a community and exposing individuals. A heat map of anaemia prevalence by block helps decide where to run the next camp. It does not, and must not, become a way to identify or expose any particular person. The value of the map comes from the aggregate. The aggregate is also exactly what protects the individual, as long as it is built and kept that way.

Why we are talking about this now

We raise it now, while the screened numbers are still modest, precisely because this is the moment to get the principles right. It is far easier to build privacy in from the start than to retrofit it after a dataset has grown and habits have hardened. If community screening is going to generate a valuable health-data map of this region over the coming years — and it can — then the rules for handling it responsibly need to be settled before the data, not after. We would rather move a little slower and be trustworthy with people's health information than move fast and betray the trust the whole model depends on.