A Model-driven Approach to Design Multi-device Forms for Capturing Data in a Medical Study Environment

Abstract

During medical studies a large amount of data has to be captured for later evaluation. Instead of doing this by pen and paper forms recent approaches use electronic data forms that directly digitize the entered data and store it in a central database. The us-age of such electronic forms on mobile devices allows the study staff to carry the forms with them such that they do not have to move to dedicated data capturing sta-tions inside the study facility. The high diversity of available mobile devices (smartphones and tablets) requires developing electronic data forms in such a way that they adapt to the platform specif-ic features. This is especially challenging due to the differences of the devices’ dis-play sizes. Furthermore, the devices offer different possibilities to connect peripheral devices for automating the data capturing process. These differences require a solu-tion for designing multi-device data forms that are usable on the different device types without generating a dedicated form for each device from scratch. This thesis presents a model-driven method that allows the design of multi-device data forms. The approach works on four model layers that are traversed one after an-other. Based on the definition of the data to be captured, an abstract version of the form’s layout is gathered. This is refined by a concrete definition for each device type until the final user interface of the data form is generated. The transitions between the model layers are done by automatic transformation processes. The developed ap-proach is prototypically implemented in the context of a study management software at the German Aerospace Center (DLR). The results of a usability study that tested the developed model-driven approach con-cerning its applicability for the future user group show that the test users were able to complete the given task at an average of about 86%. The test users additionally ex-pressed a high satisfaction when using the implemented prototype. Thus, this model-driven method is assumed to be applicable to design multi-device data forms for use in a medical study environment.