Secondary research use permitted after age + pincode generalization
Consent analysis
- Specified purpose
- The original cardiac-outcomes study only.
- Lawful basis
- Research exemption (§17(2)(b))
- Fit
- exceeds consent
The proposed cross-institution model exceeds the study-specific consent; it may proceed only under the §17(2)(b) research exemption, and only if no decision is taken specific to an individual.
Identifiers
- Direct
- Quasi
- Sensitive
Risk findings — every figure from a tool
| Metric | Value | Source tool |
|---|---|---|
| k (raw QI set) | 1 | compute_k |
| records at risk (k<2) | 7 | compute_k |
| prosecutor max risk | 1.0 | reidentification_risk |
| k after age->band, pincode->district | 6 | compute_k |
Linkage reasoning
An 89-year-old female with a rare restrictive cardiomyopathy (I42.5) in a low-population pincode is population-unique against the electoral roll even though the sample shows k=1.
Recommended transforms
| Column | Action | Rationale |
|---|---|---|
| age | generalize to 5-year bands | collapses rare elderly singletons |
| pincode | truncate to district prefix | removes sub-district uniqueness |
DPDP citations
| Assertion | Provision |
|---|---|
| Proposed use exceeds the consented purpose | §6(1) read with §2(za) |
| Research exemption applies subject to standards | §17(2)(b); Rule 16 + Second Schedule |
Residual risk. Journalist risk not computed for the raw set; assess against Census priors before release.
Your control
DPDP provisions cited here are notified but enforceable from 13 May 2027. Anonymization under DPDP is a definitional question (§2(t)) with no prescribed statutory standard — re-identification risk above is assessed per the stated k-anonymity methodology, not against a legal threshold.
Claude reasons; a deterministic engine owns every number. No figure in this memo was produced by the model.