Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


9253.pdf (892.72 KB)

ETD Abstract Container

Abstract Header

Exploring and Developing Algorithm of Predicting Advanced Cancer Stage of Colorectal Cancer Based on Medical Claim Database

Bian, Boyang
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1396522857

Abstract Details

2014, PhD, University of Cincinnati, Pharmacy: Pharmaceutical Sciences/Biopharmaceutics.
Background: Colorectal cancer (CRC) is a type of cancer which develops from uncontrolled cell growth in the colon or rectum. It is the third most commonly diagnosed cancer in males and the second in females. In epidemiologic research for CRC, advanced cancer stage is an important factor for determining disease development and treatment patterns. However, this variable is not available because medical claims databases is retrospective and only original built for financial analysis only. Algorithms to predict advanced CRC stage were developed based on the existing medical information in claims database. Method: Study cohorts were identified from the Surveillance Epidemiology and End Results (SEER)-Medicare database. Two algorithms were constructed based on covariates obtained from the database for different study periods, including demographic, treatment pattern variables. The training set was used to derive predictive equations by using logistic regression model, then applied to validation set for evaluating the predictive characteristics (sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV)). The developed algorithm were applied to MarketScan® Commercial Claims and Encounters Database and tested the predictive values. Results: The algorithm of predicting advanced CRC stage in 1999 to 2003 achieved sensitivity 50.3% and specificity 95.0%, PPV 66.78% and NPV 90.58% while the equation distinguishing CRC stage IV in 2004 to 2007 achieved sensitivity 56.8%, specificity 95.3%, PPV 71.86% and NPV 91.19%. All algorithms made better predictive values than the single ICD-9 metastatic diagnosis as the predictor. Then the algorithm for 1999 to 2003 was applied to MarketScan database. 9484 patients were predicted as non-advanced CRC group while 1097 patients were assigned to advanced CRC group. Conclusion Claims-based algorithms were developed to predict advanced cancer stage. These algorithms were shown to be successful in the recent study period due to the inclusion of new biologic agents, which were utilized in advanced cancer treatment. This predictive algorithm may be applied in claims database and generate cancer stage information, which can assist with epidemiologic study of patients with CRC.
Jianfei Guo, Ph.D. (Committee Chair)
Jane Pruemer, Pharm.D. (Committee Member)
Christina Kelton, Ph.D. (Committee Member)
Wei Pan, Ph.D. (Committee Member)
Patricia Wigle, Pharm.D. (Committee Member)
138 p.
Classical Studies; Pharmaceuticals
colon rectal cancer

Recommended Citations

Citations

  • Bian, B. (2014). Exploring and Developing Algorithm of Predicting Advanced Cancer Stage of Colorectal Cancer Based on Medical Claim Database [Doctoral dissertation, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1396522857

    APA Style (7th edition)

  • Bian, Boyang. Exploring and Developing Algorithm of Predicting Advanced Cancer Stage of Colorectal Cancer Based on Medical Claim Database. 2014. University of Cincinnati, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1396522857.

    MLA Style (8th edition)

  • Bian, Boyang. "Exploring and Developing Algorithm of Predicting Advanced Cancer Stage of Colorectal Cancer Based on Medical Claim Database." Doctoral dissertation, University of Cincinnati, 2014. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1396522857

    Chicago Manual of Style (17th edition)

ucin1396522857
552
© 2014, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.