Medical records processing · Production

ABBYY Recognition Server Helps Jeghers Medical Index Create a Searchable Digital Archive of 1,000,000 Files

The problem

Jeghers Medical Index held over a million paper medical journal articles stored across physical filing cabinets and folders, accessible only through slow manual lookups, which limited the archive's usefulness even as it aimed to bring the library online.

Workflow diagram · grounded in source
1
Begin scanning paper articles
Trigger
JMI began scanning its paper articles into TIFF files starting in 1997 to launch the digitization effort.
▾ source quote
“the library began scanning articles into tiffs in 1997”
2
Build search database with Thunderstone
Integration
JMI engaged Thunderstone to build an SQL relational database management system integrating a customized search device.
▾ source quote
“They then engaged with Thunderstone to create an SQL relational database management system integrating a customized search device.”
3
OCR conversion via ABBYY Recognition Server
Ai action
DocuSyst used ABBYY Recognition Server to convert the scanned files into archival PDF/A-1a files.
▾ source quote
“And for DocuSyst, this meant using ABBYY Recognition Server.”
4
Initial batch accuracy verification
Validation
After verifying an initial batch exceeded agreed accuracy benchmarks, the remaining articles were processed in 7 batches over a year.
▾ source quote
“After verifying it exceeded the agreed accuracy benchmarks, the remaining articles, nearly a million of them, were sent over in a series of 7 batches over the course of a year.”
5
Batch quality control sampling
Validation
Every completed batch was subjected to verification based on random sampling of the converted files.
▾ source quote
“Every batch completed was subjected to verification, based on random sampling of the converted files.”
6
Integrate PDFs into digital archive
Integration
The resulting archival PDFs were integrated into JMI's digital archive alongside a new search appliance powered by Thunderstone's Texis database system.
▾ source quote
“the resulting archival PDFs have been integrated into JMI’s digital archive – and their new search appliance powered by Thunderstone’s Texis relational database management system.”
7
Appliance indexes documents from OCR
Ai action
The search appliance uses an algorithm to identify document content based on the OCR results, then crawls and tags all PDFs so everything is properly indexed.
▾ source quote
“uses an algorithm to take the content of a document and identify it based on the OCR results. It crawls all the PDFs, collects all the information, then tags it so that everything is properly indexed in the system.”
8
Multi-facet search output
Output
The archive can be searched by author, title, subject, journal, publication date, words, adjacency, paragraph, page, wildcards, and truncation.
▾ source quote
“The archive can be searched by author, title, subject, journal, publication date, by words in an article, by adjacency (meaning the words can be in the same sentence), paragraph, page, wildcards, truncation and much more.”
Reported outcome

DocuSyst converted the archive to searchable PDF/A files using ABBYY Recognition Server, finishing on time and below budget within agreed quality parameters, and JMI's new search system is far faster and more comprehensive than the old manual process.

Reported metrics
Pages processedover 5,000,000
Project plan length108-page
Digitization duration1 year
Files indexed1,000,000 Files
Show all 9 reported metrics
pages processedover 5,000,000
project plan length108-page
digitization duration1 year
files indexed1,000,000 Files
annual searches20,000 searches every year
physical archive scale165 filing cabinets and 44,000 folders
initial test batch size25,000 files
conversion batches7 batches
total articles in libraryover one million
Reported stack
ABBYY Recognition ServerThunderstoneTexisSQL relational database management system
◆ Does this fit your context?

Compare to your context

Tell us your scale, team, and constraints. We'll show what changes at your size, what fails at your scale, and whether this case is a fit, needs adaptation, or won't scale to you. Free demo, no signup.

Compare to your context →
~30 seconds · free
Source
https://www.abbyy.com/customer-stories/automated-ocr-helps-pioneering-medical-archive-create-searchable-index-of-1000000-files/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

DocuSyst converted the archive to searchable PDF/A files using ABBYY Recognition Server, finishing on time and below budget within agreed quality parameters, and JMI's new search system is far faster and more comprehe…

What tools did this team use?

ABBYY Recognition Server, Thunderstone, Texis, SQL relational database management system.

What results were reported?

Pages processed: over 5,000,000; Project plan length: 108-page; Digitization duration: 1 year; Files indexed: 1,000,000 Files (source-reported, not independently verified).

How is this medical records processing AI workflow structured?

Begin scanning paper articles → Build search database with Thunderstone → OCR conversion via ABBYY Recognition Server → Initial batch accuracy verification → Batch quality control sampling → Integrate PDFs into digital archive → Appliance indexes documents from OCR → Multi-facet search output.

Related medical records processing cases

WHAT TO DO WITH THIS

Now compare it to your context

This case is one data point. Whether its pattern fits you depends on your volumes, your stack, and your exception load — that comparison is the step no case study can do for you.

Browse the Medical records processing cases →