In this validation study, the authors used data from ~17,000 UK Biobank participants to identify participants with one or more dementia codes in primary care, hospital admissions or mortality datasets, and compared the coded data to the full-text medical record. Having validated the accuracy of these datasets, they then developed algorithms that can be applied to identify participants with dementia in UK Biobank and other DPUK cohorts. The authors will continue this promising work by investigating sources of potential bias in the data and the generalisability of these findings to older ages and other geographical areas.