ideasIdea

Privacy-preserving linkage tokens for every cancer data holder

Give each patient a scrambled code that is the same across hospitals, labs and registries, so records can be joined without anyone seeing names.

Record linkage across holders fails without a shared identifier; where national IDs exist (Nordics) linkage is trivial, elsewhere it is probabilistic and lossy. Privacy-preserving record linkage using salted hashes of identifiers (as used by Datavant in the US and Bloom-filter approaches in Australia and Germany) allows joining without exposing identity. The proposal mandates a common tokenisation scheme, run by a public trusted third party, for all holders of cancer data.

Hypothesis

A common token will raise linkage rates between registries, genomic labs and hospitals from below 70 percent (probabilistic) to above 97 percent, at no measurable privacy cost.

Rationale

Where national identifiers exist, near-complete linkage has powered decades of registry research; tokens replicate the function without a national ID.

Proposed test

Tokenise one registry, one genomic lab and one hospital system; measure linkage rate against a gold-standard manually linked subset.

Maturity

early clinical

Who acts

data

Cost to try

Medium ($1M to $50M)

Years to first evidence

Bottlenecks it attacks

Data silos · Records, scans, genomes and outcomes sit in separate systems that cannot talk. Every patient's experience is lost to the next.

Connected

2top

fronts

AI & Computation

bottlenecks

Data silos

ideasIdea

Privacy-preserving linkage tokens for every cancer data holder

Give each patient a scrambled code that is the same across hospitals, labs and registries, so records can be joined without anyone seeing names.

Hypothesis

A common token will raise linkage rates between registries, genomic labs and hospitals from below 70 percent (probabilistic) to above 97 percent, at no measurable privacy cost.

Rationale

Where national identifiers exist, near-complete linkage has powered decades of registry research; tokens replicate the function without a national ID.

Proposed test

Tokenise one registry, one genomic lab and one hospital system; measure linkage rate against a gold-standard manually linked subset.

Maturity

early clinical

Who acts

data

Cost to try

Medium ($1M to $50M)

Years to first evidence

Bottlenecks it attacks

Data silos · Records, scans, genomes and outcomes sit in separate systems that cannot talk. Every patient's experience is lost to the next.

Connected

2top

fronts

AI & Computation

bottlenecks

Data silos

Privacy-preserving linkage tokens for every cancer data holder

Overview

Connected

fronts

bottlenecks

Privacy-preserving linkage tokens for every cancer data holder

Overview

Connected

fronts

bottlenecks