What linked data reveals

The strongest evidence on the site comes disproportionately from a few countries, and the reason is data. The Nordic countries and the Netherlands link births, deaths, diagnoses, prescriptions, education and income for every resident through a personal identifier. A question about a whole population can therefore be answered rather than sampled. That capability is spreading unevenly. England has built linkages across hospital, primary care and death data that were impossible a decade ago, and has added genetic data at scale: the 100,000 Genomes Project and UK Biobank's half a million sequenced genomes are both linked to NHS records, although access is still slow and fragmented. Scotland and Wales have gone further, with dedicated data environments. National insurance databases in Korea and Japan are opening to research. In contrast, the United States still cannot link nationally, and works through claims data and cohorts instead. The data-watch entries record new sources, linkages and access arrangements as they come. The pertinent question is what they change: which findings became possible, which were overturned when a whole population replaced a sample, and where the gaps in linkage still determine what can be known.

Trackers