Leveraging the Cloud for Intelligent Clinical Data Registries (ACM-BCB'19)


Public cloud platforms provide an amazing set of capabilities, but it can be an overwhelming challenge to create a design, implementation, and deployment that properly leverages today’s existing public cloud capabilities while not precluding the use of near-future new services and infrastructure. We tackle this challenge in the context of clinical data registries, and create Cloud-based Patient Outcomes Platform (CPOP), our scalable public cloud application for clinical patient data. Doctors are able to visualize collected medical data in different chart formats and patients are able to check their data and submit medical survey forms. The specific domain of interest in this paper is Chronic Rhinosinusitis (CRS), a largely under-recognized chronic disease in our society. The primary barrier to quality improvement in CRS is the difficulty in collecting data from patients, tracking appropriate follow-up time intervals, and analyzing outcomes results in a prospective and ongoing fashion. We describe key aspects and design experiences of CPOP-CRS in Amazon Web Services.We also provide quantitative evaluation of a key feature of CPOP-CRS, which is the ability of CRS doctors to upload an audio clip of a doctor-patient interaction, and have the cloud render a text-based representation, and show a word error rate of 15.6%. We outline next steps in the development of the CPOP/CPOP-CRS, and provide guidance for other users considering the public cloud for their next parallel and cloud-based Bioinformatics and Biomedicine project.

In 10th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics (ACM-BCB ’19), September 7–10, 2019, Niagara Falls, NY, USA. ACM, New York, NY, USA, 8 pages. https://doi.org/10.1145/3307339.3343464