Abstract
The initial version of a clinical data management system delivered to Bayer was a β test version containing some modules still in preliminary form. Expecting a rapidly developing system, a reproducible test was required to allow Bayer to evaluate new enhancements of the system within a few days of the software being issued. To meet these requirements, an automated user acceptance test was designed. The test was constructed from a combination of keystroke files and commands that were run from a batch file. Output generated from the test was then saved and compared with output from a previous run of the acceptance test. The acceptance test performs all the data management tasks for a fictitious study. Data were carefully chosen to fully expose any weaknesses in the software package. This user acceptance test has proved to be quick to run and reproducible. It also confirms the resolution of software errors.
Keywords
Get full access to this article
View all access options for this article.