Colorectal cancer (CRC) is the second leading cause of cancer death in the United States. Regular screening for CRC leads to early detection and a substantial reduction in CRC-related morbidity and mortality. Despite this evidence, adherence to CRC screening recommendations remains poor. Given the benefits of screening for CRC, interventions need to be developed that will ultimately increase its utilization. These interventions must be targeted towards individuals who have never been screened or are not current with their screening, and must take into account risk factors [e.g., family history, inflammatory bowel disease, history of previous adenomas]. A significant barrier to applying CRC screening interventions is the ability to accurately and quickly identify patients in need of screening. In our proposed project, we will create and evaluate an innovative electronic data collection system to enhance our understanding of the delivery and utilization of CRC screening. We will use as a test site the primary care clinics affiliated with Vanderbilt University Medical Center, but the approach developed will be general enough to work wherever adequate electronic records are available. This data collection system will build upon our prior work using the Knowledge Map Concept Identifier (KMCI), which was originally developed to extract biomedical concepts represented in the medical school curriculum at Vanderbilt University School of Medicine. We will build a tool using KMCI as its core to detect receipt of CRC screening and to provide its context (e.g., individual risk factors, which CRC test was performed, when, and the results). Utilizing information from the electronic records and billing (claims) records, this system will allow for rapid and accurate identification of which patients have and have not received CRC screening tests, and for identification of those patients that require follow-up after their screening tests. We will determine the test characteristics of KMCI (sensitivity, specificity) for identifying receipt of CRC screening, and compare this to an automated review of billing records. We hypothesize that KMCI will perform better than an automated review of billing records in detecting receipt of CRC screening, and will provide richer, contextual information regarding the screening tests. ? ? ?
| Denny, Joshua C; Choma, Neesha N; Peterson, Josh F et al. (2012) Natural language processing improves identification of colorectal cancer testing in the electronic medical record. Med Decis Making 32:188-97 |
| Denny, Joshua C; Peterson, Josh F; Choma, Neesha N et al. (2010) Extracting timing and status descriptors for colonoscopy testing from electronic medical records. J Am Med Inform Assoc 17:383-8 |
| Denny, Joshua C; Spickard 3rd, Anderson; Johnson, Kevin B et al. (2009) Evaluation of a method to identify and categorize section headers in clinical documents. J Am Med Inform Assoc 16:806-15 |
| Denny, Joshua C; Miller, Randolph A; Johnson, Kevin B et al. (2008) Development and evaluation of a clinical note section header terminology. AMIA Annu Symp Proc :156-60 |
