Validation Odyssey: From big data to local intelligence

The main objective of this thesis is advancing the methodology of validation studies of case-finding algorithms that exploit diversity across available data, rather than collecting new data.
The case study that led us to this advancement was the assessment of the capacity of the Italian administrative database to capture cases of chronic disease to get estimates for the compliance with standards of care. Primary care medical records were the main comparative source. Part I of this thesis is focussed on this topic.
In Part II we exploited the results and extended the methodology of Part I to the context of multi-database, multi-national studies.