Timely and accurate documentation of medical information remains a formidable challenge in healthcare. This challenge has persisted despite rapid advances in computer technologies. This is largely due to the significant complexity of medical work. This complexity can be represented as an intricate web of collaboration, communication, and situational awareness. The work presented in this thesis explores one aspect of the challenge of medical documentation: medical orders. The project is a first step toward examining examination of the use of speech recognition technology for order entry. A computer script written in PERL was created and tested using a random sample of handwritten orders from several intensive care units at Oregon Health and Science University Hospital with the goal of categorizing the orders into logical groups.