Evaluating the portability of health care data to SQL like Big Data environment.

摘要

Big Data deals with huge-volumes of complex, exponentially growing data sets from multiple, sources. With rapid growth in networking we are now able to generate immense amount of data in almost any field imaginable, including physical, biological and biomedical sciences. While most industries have been far more successful at harnessing the value from large-scale integration and analysis of big data, the health care industry is just getting its feet wet. One impediment for the Health Care industry adoption of Big Data analytics has been the dependence of many of their models on the RDBMS technology. With the diversity and amounts of data in health care industry there is an increasing need to evaluate components in big data frameworks and gauge their adaptability to analytics techniques. However, recent developments in the Hadoop ecosystem environment have led to breakthroughs enabling RDBMS like tools in big data environments. In this paper we evaluate the portability of existing RDBMS solutions employing such SQL like big data tools. Our work focuses on benchmarking multiple SQL like big data technologies over HDFS for Study Data Tabulation Model (SDTM) used in clinical trial databases for improving the efficiency of research in clinical trials. We will examine their potential for improving the efficiency of research in big data clinical trials. Publicly available healthcare data (from National Institute of Drug Abuse (NIDA)) is utilized as a test bed to measure key parameters like usability, adaptability and modularity, robustness and efficiency. Our intention is to demonstrate the portability of the execution of ad-hoc SQL queries on the fly occurring in current clinical trial functionality and evaluate if it can be replicated in a big data SQL like back-end system with relative ease and transparency.