Goldbaum, GaryGallivan, Mark2012-09-132013-09-142012-09-132012Gallivan_washington_0250O_10541.pdfhttp://hdl.handle.net/1773/20523Thesis (Master's)--University of Washington, 2012OBJECTIVE: To compare the accuracy of ICD-10 coded and text cause of death data in identifying tuberculosis (TB)-related deaths in Washington State. DESIGN: A cross-sectional study was performed comparing TB-related deaths identified through death certificates using either text fields or ICD-10 codes to TB-related deaths identified in the TB registry. RESULTS: All methods using death certificate data overestimated the number of TB-related deaths and had low to moderate agreement with the tuberculosis registry. Several subjects assigned a TB ICD-10 code had no evidence of TB on the death certificate. CONCLUSION: Death certificates are not highly predictive of TB-related deaths. If death certificates are utilized, then TB-related deaths should be identified using direct cause of death text fields.application/pdfen-USCopyright is held by the individual authors.EpidemiologyEpidemiologyThesis