Testing out digital methods to inform healthcare history

Archives across the UK contain prescription books, from both pharmacy and general practice. These books could be a fantastic research resource, but are currently under used because this primary source data is challenging in several ways.

(1) Very little is known about collections of these books and our limited archival scoping work has demonstrated that there are pockets of collections in different archives across the country;
(2) Differing policies across archives about whether these books can be viewed by researchers, due to the sensitive data that they contain, e.g patient names and the medicines they were prescribed.
(3) The nature of the books as handwritten paper volumes. The data they contain is useful at a macro level, but it is difficult to conduct this type of research (and at scale) using traditional historical research methods.

Through this project we will investigate whether these challenges can be addressed through digitisation and the use of quantitative and computational analysis. We will work with archivists, computer engineers and other specialists to:
(1) explore the feasibility of digitising these books for computational analysis (this will include testing out AI assisted transcription using Transkribus, and exploring alternatives such as ‘citizen science’ approaches using Zooniverse);
(2) determine ethical approvals that might be necessary for digitising books within living memory;
(3) investigate whether these books can be anonymised in a meaningful way (that protects identity but also ensures the data remains useful for research);
(4) test out ways of digitally structuring the data to facilitate quantitative and computational research;
(5) produce metadata about the collections;
(6) through all of these, make these primary sources more easily accessible to other researchers;
(7) produce some simple data visualisations searching the pilot date for “penicillin” as part of our continuing research in this area.