Perhaps the most important usability study begins when a medical device has been approved for market and purchased by customers. Real world use problems and usability assessment isn’t readily obtainable unless manufacturers systematically obtain this data through post market surveys, interviews, or questionnaires. Unfortnately, adverse event reports, such as the FDA’s Manufacturer and User Device Experience (MAUDE) database. MAUDE narratives remain vague regarding the relationship between patient harm and user interface problems, because authors are not directed to report problems in the “language” of human factors.
Recently, we began applying systems query language search routines to MAUDE event narratives. These queries are directed at user interaction problems, i.e. perceptual, cognitive, and physical interaction with devices. The search terms were derived from our earlier work with the FDA in developing the UPCARE model which had identified a taxonomy expressing the ability, or inability, of a user to interact with a device.
Through a partnership with rationalhealthcare.us, we have downloaded over 4 million MAUDE narratives (2008 through June, 2013). With this data, downloaded from the MAUDE public database, we can easily isolate a particular device family’s event data over the past six plus years. Then, particular searches can be performed to sort the narratives in their order of relationship to the search terms entered. Subsequently, the user can study the narratives of interest about a particular set of use errors or user interactions.