Notes on making the UK's Prevention of Future Death reports searchable — and on what they contain.
2,190 of the 4,224 reports published between 2013 and 2022 are image-only scans containing zero extractable text, a blind spot for any study that screens them by keyword, and the reason we OCR'd the archive.
A disputed label used to explain deaths after police restraint — and the Prevention of Future Death reports that mention it, now searchable in one place.
Building the index meant reading every report page. Along the way we found URL typos, mismatched references and a mis-attached document — and shared them with the Chief Coroner's Office.
Every report re-indexed (scanned ones included, via OCR), thousands of broken links fixed, and new options for researchers, journalists and organisations.