An Algorithm for Predicting Oral and Pharyngeal Cancers Using the Seer-Medicare Linked Database

• Activity Number: 75
• Type: Contributed
• Date/Time: Sunday, July 31, 2011 : 4:00 PM to 5:50 PM
• Sponsor: Section on Health Policy Statistics
• Author(s): John David Keighley*+ and Jonathan D. Manhnken and Matthew Stuart Mayo
• Companies: University of Kansas Medical Center and University of Kansas Medical Center and University of Kansas Medical Center
• Keywords: secondary ; data ; analysis ; health ; outcomes ; research

Abstract

If a researcher wants to consider the full picture of treatment for a cancer patient they can not go to a record from a cancer registry. When data is transmitted using the North American Association of Central Cancer Registries (NAACCR ) format there is room for one code and one date for each of the six different types of treatment. It is possible to obtain a more complete picture of the treatment of a cancer patient using billing records if they are available. The most complete source available for this purpose is the SEER-Medicare Linked Database. SEER is considered to be the gold standard for cancer registries with a case reporting completeness of 98% of the cases diagnosed in their catchment area. This project developed a new method to identify oral and pharyngeal cancer cases using medicare ICD 9 diagnosis and procedure codes. The algorithm began using methods developed in record-linkage software and was completed in a new form. This method resulted in a sensitivity of over 94% and a specificity of over 96%. This method can be used to identify oral and pharyngeal cancer cases using Medicare records in non-SEER areas which then allows a reasearcher to consider treatment for the