Q-Cyber Magazine

Healthcare · 20 August 2026 · 3 min read

Sharing health and research data without widening the record

Secondary use of health data should start from DPDP's purpose and minimisation rules and from ABDM's consent artefact, not from a habit of copying identifiable records into a new store.

Abstract graphic of layered document sheets in Knowlcom blue

Hospitals, researchers and public-health programmes often have a legitimate reason to share information. The failure mode is familiar: a copy of identifiable records is made "for analysis," and the copy outlives both the purpose and the consent.

Purpose first, technique second

"Privacy-preserving" is sometimes used as a synonym for a particular cryptographic protocol. This article will not invent protocol performance, or claim a technique that has not been specified in the sources below. The legally established starting points are simpler, and they already change practice.

Under DPDP, a Data Fiduciary must be able to say why personal data is being processed, collect no more than that purpose requires, and not keep it once the purpose has ended. A research extract that still carries a name, a hospital identifier and a full longitudinal history, when the study needed only an aggregated count, fails that test before any advanced method is discussed. Where a new purpose arises, a new notice and a new lawful basis are required; reuse of an old, bundled consent is exactly what the official emphasis on standalone, purpose-specific notices is designed to discourage.

The national digital health design points in the same direction. Federation means records remain with accountable providers and move when the individual has agreed, through a consent artefact that specifies the scope of the transaction. A Health Information User is not entitled to a private mirror of the national ecosystem. If a research or public-health use cannot be expressed as a scoped, consented request, or as an anonymised output from which the individual cannot reasonably be re-identified, it is not yet a sharing design. The Policy's own definition of anonymisation is strict: an irreversible process, not a reversible mask that the holder can later undo.

Exhibit 1

Choosing a sharing path for health or research data

A specified, current purpose and consent exist
They do not
The receiving party needs identifiable records
Use a scoped consent artefact and a federated exchange; do not create a second lasting store
Do not share; define the purpose and obtain a fresh lawful basis, or stop
The receiving party needs only derived or anonymised outputs
Prefer an anonymised or aggregated extract, with a written test against re-identification
Do not share identifiable source data in the hope that a later method will repair the decision

Source: Qualitative decision framework derived from the DPDP principles as stated by the Press Information Bureau on 14 November 2025, and from the Health Data Management Policy definitions of consent artefacts, Health Information Users and anonymisation. No protocol benchmarks are claimed.

Implications

  1. Write the purpose down before the extract is made, and retire the extract when the purpose ends.
  2. Use consent artefacts for identifiable exchange, in the spirit of the national digital health policy, and do not stand up a parallel warehouse.
  3. Reserve the word "anonymised" for irreversible transformations, as the Policy defines them.
  4. Be ready for 90-day rights and for breach notice whenever identifiable data is still held.

What this means for healthcare

Data sharing can support care and science. It cannot be an informal copy-and-keep habit. Providers and research groups that start with purpose, consent and minimisation will be aligned with both the DPDP Rules and the national digital health design. Those that accumulate identifiable "research" copies will not.

Sources

  1. PIB notification of the DPDP Rules, 2025
  2. National Digital Health Mission: Health Data Management Policy