The aim of this study was to determine the accuracy of dysphagia coding using the International Classification of Diseases version 9 (ICD-9) code 787.2. We used the administrative database of a tertiary hospital and sequential videofluorographic swallowing study (VFSS) reports for patients admitted to the same hospital from January to June 2007. The VFSS reports were abstracted and the hospital's database was queried to abstract the coding associated with the admission during which the VFSS was performed. The VFSS and administrative data were merged for data analysis. Dysphagia was coded (using code 787.2) in 36 of 168 cases that had a VFSS. Of these, 34 had dysphagia diagnosed by VFSS (our gold standard) and one had a prior history of dysphagia. Code 787.2 had sensitivity of 22.8, specificity of 89.5, and positive and negative predictive values of 94.4 and 12.9, respectively. Dysphagia was largely undercoded in this database, but when the code was present those individuals were very likely to be dysphagic. Selection of dysphagic cases using the ICD-9 code is appropriate for within-group comparisons. Absence of the code, however, is not a good predictor of the absence of dysphagia.