A Universal Translator for Health Languages

The potential to transform healthcare management through information technology continues to grow, as more systems, devices, and people than ever before provide and collect health data. Unfortunately, progress is slow because the multitude of systems to integrate do not speak a common language. We propose to create a universal translator between health languages, which will allow multiple systems to co-exist, yet share key information. We intend to do this based on a proven, sound statistical inference methodology that is, in turn, powered by a machine learning approach. Combining both methods yields a systems that is both accurate and flexible, with the potential to significantly the reduce the cost of current methods. We have begun to demonstrate this to be the case in working with a prominent health informatics company on a similar challenge. Through this SBIR, we intend to extend that prototypical system to further improve the translation technology, while integrating the resulting system with important emerging sources like 3M's Healthcare Data Dictionary, all the while navigating towards the open sourcing of the end result.