P07 Information system for a multi-hospital study using optical scanning and hospital database...

1
76S Abstracts show both key and FAX and "check if applicable" fields. The numeric field errors were comparable (2 % FAX, 1% key). There was a discrepancy in s-numeric error rate (15 % FAX, 1% key). Most FAX errors were c~ substitutions for numeric characters (an "1" for a one or a "0" for a zero). Experience of the DE clerk did not affect these rates. FAX DE was faster than key DE (.42 +- .22 v. 1.09 +- .28 min/form). The difference between FAX DE time and key DE time increased with experience. Our experienced clerk entered the FAX form in .23 +- .06 min and key form in 1.13 + .34 min. The inexperienced clerk had respective times of .62 +- .23 and 1.05 +_.13 mon. Key DE times were comparable between the two clerks but FAX DE times showed experience speeds data entry. These results indicate that FAX data entry is a good alternative to key data entry. FAX DE is faster and as accurate as key DE for check-box fields and numeric fields but, special programming is required to check accuracy of s-numeric fields. P07 INFORMATION SYSTEM FOR A MULTI-HOSPITAL STUDY USING OPTICAL SCANNING AND HOSPITAL DATABASE DOWNLOADS Stephen Anderson*, Mazen Abdellatif**, Steve Schreiner*, Mary McDonell*, Domenic Reda** and Stephan Fihn* *VA Medical Center, Seattle **VA Medical Center, Hines We have developed an information system to monitor quality of care for 10,000 patients at several Veterans Affatrs (VA) hospitals. The system uses new technologies for data collection, data flow and reporting. In a traditional clinical trial, nurses at participating sites record patients' data on study forms. These forms are mailed to a coordinating center where they are keyed into a large computer. A separate time consuming step is required to check for errors after the time of data collection. Our approach has been to develop a PC based system that integrates clinical data from the hospital computer with patient reported data (questionnaires). Key components of the system include automated weekly downloads from the hospital computer system ~nd tracking and printing of questionnaires customized for each patient based on their medical conditions. An optical scanning system reads questionnaire data directly into the database where they are integrated with clinical data. The printing and scanning of questionnaires are centralized for economics of scale; electronic data transfers to and from the sites are done weekly. Each site has a single PC for data transfers, hospital downloads, and generating graphical reports for providers. The coordinating center receives data monthly for monitoring and analysis. The system requires less time for data collection and provides better error checking than traditional key data entry. The improved data integrity has eliminated the need for the traditional full-scale edit system. By using on-site PC's, data can be online within one week of collection and timely graphical reports can be created for clinical intervention.

Transcript of P07 Information system for a multi-hospital study using optical scanning and hospital database...

Page 1: P07 Information system for a multi-hospital study using optical scanning and hospital database downloads

76S Abstracts

show both key and FAX and "check if applicable" fields. The numeric field errors were comparable (2 % FAX, 1% key). There was a discrepancy in s-numeric error rate (15 % FAX, 1% key). Most FAX errors were c~ substitutions for numeric characters (an "1" for a one or a "0" for a zero). Experience of the DE clerk did not affect these rates. FAX DE was faster than key DE (.42 +- .22 v. 1.09 +- .28 min/form). The difference between FAX DE time and key DE time increased with experience. Our experienced clerk entered the FAX form in .23 +- .06 min and key form in 1.13 + .34 min. The inexperienced clerk had respective times of .62 +- .23 and 1.05 +_.13 mon. Key DE times were comparable between the two clerks but FAX DE times showed experience speeds data entry. These results indicate that FAX data entry is a good alternative to key data entry. FAX DE is faster and as accurate as key DE for check-box fields and numeric fields but, special programming is required to check accuracy of s-numeric fields.

P07 I N F O R M A T I O N SYSTEM FOR A M U L T I - H O S P I T A L STUDY

USING O P T I C A L SCANNING AND HOSPITAL DATABASE DOWNLOADS

Stephen Anderson*, Mazen Abdellatif**, Steve Schreiner*, Mary McDonell*, Domenic Reda** and Stephan Fihn*

*VA Medical Center, Seattle **VA Medical Center, Hines

We have developed an information system to monitor quality of care for 10,000 patients at several Veterans Affatrs (VA) hospitals. The system uses new technologies for data collection, data flow and reporting.

In a traditional clinical trial, nurses at participating sites record patients' data on study forms. These forms are mailed to a coordinating center where they are keyed into a large computer. A separate time consuming step is required to check for errors after the time of data collection.

Our approach has been to develop a PC based system that integrates clinical data from the hospital computer with patient reported data (questionnaires). Key components of the system include automated weekly downloads from the hospital computer system ~nd tracking and printing of questionnaires customized for each patient based on their medical conditions. An optical scanning system reads questionnaire data directly into the database where they are integrated with clinical data.

The printing and scanning of questionnaires are centralized for economics of scale; electronic data transfers to and from the sites are done weekly. Each site has a single PC for data transfers, hospital downloads, and generating graphical reports for providers. The coordinating center receives data monthly for monitoring and analysis.

The system requires less time for data collection and provides better error checking than traditional key data entry. The improved data integrity has eliminated the need for the traditional full-scale edit system. By using on-site PC's, data can be online within one week of collection and timely graphical reports can be created for clinical intervention.