Providing hospitals with a quality benchmark that identifies (sub)sets of patients for whom treatment outcome can be improved
A quality registry typically collects data from patients, calculates a quality indicator in a single hospital (e.g. % of patients getting the guideline-recommended treatment) and compares it to other hospitals, i.e. benchmarking. All current quality registries use a data centralization approach, which has privacy and data control concerns.
Under the Techruption flag, TNO, CZ, MAASTRO, CBZ and the UM-BISS institute worked together to prototype a distributed quality registry using secure multiparty computation (sMPC) as its underlying technology. We found that sMPC works well and is one of the technologies that can be used in the PHT for privacy preserving distributed learning.
sMPC is able to calculate the benchmark in a privacy preserving manner and without hospital seeing other hospitals' performance.
sMPC is an excellent additional technology stack which can be used in the Personal Health Train.
In a new project, funded by the National Health Care Institute (Zorginstituut Nederland), we implement PHT and FAIR principles for benchmarking in colorectal cancer.
TNO / CZ / MAASTRO Clinic / Maastricht University / CBS