2005 Scientific Session Abstracts
Automated Data Collection using Scanning and Optical Character Recognition
J. Mark Tuthill (mtuthil1@hfhs.org), David Chen, Paul Tranchida. Henry Ford Medical Center, Detroit, MI
Content: Commercial hardware and software applications are available to facilitate and improve the collection of data in a variety of settings applicable to the workflow in a pathology department. In this project we discuss our experience with one solution for collecting data from surveys and in cataloging surgical pathology accession forms.
Technology: Hardware: Personal computer running Windows 2000 (Microsoft, Redmond, WA) Panasonic KVS-2065 high-speed scanner (Panasonic Corporation of North America, Secaucus, NJ). Software: TELEform Enterprise Server (Scantron, Irvine, CA), ApplicationXtender (EMC, Pleasanton, CA).
Design: Forms were designed and then made available to the users either printed on paper or online after being exported as an HTML file using a plug-in (E-forms). The user populates the form with data either by writing directly on the sheets or populating fields in the E-form using a computer keyboard. The paper forms are then scanned and the images are stored in a directory. The images are subsequently examined with optical recognition technology to interpret handwriting and/or markings indicating user choice. With E-forms the data is directly obtained from the user input on the web page. Verification can then be performed to confirm the validity of the entered information; this is done with computer-assisted visual inspection and/or cross-correlation with a table containing all possible valid results. Export of data is supported for a wide variety of databases with optional linkage to the scanned images. The value of each field in the database can be configured during design so that certain choices are converted to text or numeric codes that facilitate visual inspection or statistical analyses. Further, using ApplicationXtender, images of scanned forms can be searched, retrieved, and viewed.
Results: We are in the preliminary phase of staging each of our projects into clinical and administrative operations. We see marked timesavings in data entry. Tabulation errors are eliminated as well allowing for better data analysis. Validation of patient identification and other critical parameters via verification helps ensure accurate ownership of the data. The ability to electronically manage images of the clinical requisitions has clear merits especially when added to the benefit of electronic re-verification of patient data. The ability to perform web-based data collection facilitates response rates and the ease of data collection.
Conclusion: The flexibility and portability of "smart form" technology can make it an attractive option for pathology departments seeking to streamline their data collection projects.