Demonstrating the Feasibility of Statistical Process Control Charts as a Surveillance and Decision Tool for Breast Screening Utilizing Open Source Solutions

Abstract

Healthcare organizations typically suffer from the Data Rich, Information Poor (DRIP) syndrome, which has a direct impact on the ability of organizations to monitor processes which is the basis of surveillance activities. The objectives of this research were to demonstrate the feasibility of Statistical Process Control Charts (SPC) as a surveillance and decision tool, and to identify a data collection framework that satisfies the key quality factors of completeness, accuracy, comparability, usability and timeliness criteria. SPC charts were developed and customized using Open Source Software (OSS). A survey was developed and executed to identify breast pathologists’ familiarity with synoptic reporting elements and their reporting habits. Results showed that SPC charts successfully identified developing trends in monitored processes and correctly isolated special causes from common causes. Perl was identified as an OSS solution for parsing templated pathology reports to populate the relational database. The R language for statistical computing provided an OSS solution for implementing SPC charts. Pathologists’ responses did not correspond with respect to their ratings of reporting elements.