Census

Census compares incoming patient records against your master index over FHIR R4 and decides, record by record, whether you are looking at a new person or a duplicate of someone you already have. It scores demographic overlap, surfaces the close calls, and auto-merges only the matches it is confident about.

Each candidate runs through deterministic checks first, then a probabilistic score on name, date of birth, and contact details so near-twins and data-entry typos do not collapse into one chart by accident. Uncertain pairs go to a review queue with the evidence laid out side by side.

Duplicate rates in patient indexes commonly run into the low double digits before cleanup, and Census steadily pulls that down without merging records it should not. Accuracy depends on how complete your demographics are, and the tool reports its confidence on every decision.