Generating and visualizing breast re-excision rate from registry and administrative data.

Abstract

e17624 Background: Monitoring quality metrics in near real time can help healthcare providers and organizations improve care delivery, engage patients, and comply with reporting requirements of accreditation bodies. However, use of registry data alone limits the ability to provide near real time quality metrics, while clinical and administrative data may be incomplete or biased. In this study, we focus on generating breast cancer re-excision rates in an automated fashion and visualizing surgical treatment paths in novel ways. Methods: This study used surgical registry codes and administrative CPT codes for women diagnosed with stage I-III breast cancer between 2000 and 2012 at a comprehensive cancer center. We measured the re-excision rate for this patient population and used the Chi-square test for trends to compare the CPT rates against the registry rates. We also used Sankey diagrams to represent surgical paths. Results: 1582 patients had 3208 surgical events in the tumor registry. Corresponding CPT codes were missing for 28% of the 3208 surgical events, most of which were for the first surgical event in the sequence (see table). Based on registry data, 890 of the 1582 patients had a breast conserving surgery as their first surgical event. Of those, 68% did not require any additional surgery, 20% required a re-excision, and 11% advanced to receive a mastectomy. There was no significant difference in the annual re-excision rates between both data sets while there was a significant reduction in re-excisions over time in both. Finally, we developed intuitive Sankey diagrams to illustrate the progression of the surgical course. Conclusions: We were able to calculate similar re-excision rates from manually curated registry data as well as near real time administrative data. Our analysis showed that despite the missing CPT data, administrative and clinical data could be used to generate quality metrics and visualize data in a more real time fashion to help patients and providers make data informed decisions. Number of missing CPT codes at different positions in the surgical event sequence. Position in Surgical Sequence Total # of CPT Codes # of Missing CPT Codes % of Missing Codes in Position 1 2830 771 27.2% 2 612 108 17.6% 3 135 8 5.9% 4 17 0 0.0% 5 1 0 0.0%