FDA cures data-entry backlog with digital scanning, crowdsourcing
The Food and Drug Administration tapped a Web-based digital scanning service to repair a monthslong backlog of reports submitted to its drug safety database.
The FDA last June notified the public that “unforeseen issues in its data entry operations” had slowed its ability to record reports submitted to the FDA Adverse Event Reporting System (FAERS), a database of dangerous drug interactions and other accidents.
The FDA receives about 900,000 adverse drug event reports annually, of which 90 percent are electronic and 10 percent are submitted on paper.
In order to accelerate data entry, the FDA announced a deal with Captricity, a digital optical character recognition scanning service, to speed its paper-reports processing.
Captricity used both its OCR scanning technology and manual data entry by workers supplied by Amazon Mechanical Turk to digitize the paper reports. Amazon Mechanical Turk is an electronic service that solicits individuals via the Internet for digital freelance work.
Captricity, which is a product of the Code for America Accelerator program, offers a software-as-a-service scanning application that gives customers a high degree of control and choice over what content to extract from a document.
The content is presented in a machine readable format. An application programming interface is also available for use by developers. The company has said its solution is as accurate as manual data entry for the FDA, but eight times cheaper and 50 times faster.
Using the service, the FDA was able to reduce its “backlog to zero” in a matter of weeks, the company’s CEO Kuang Chen told CivSource.
The company has also received authority to operate within the FDA based on FedRAMP moderate security protocols. Captricity’s user data is also stored on Amazon’s FedRAMP compliant clouds.
Posted by GCN Staff on Jan 08, 2014 at 11:49 AM