Vol.:(0123456789) 1 3 International Journal of Data Science and Analytics https://doi.org/10.1007/s41060-019-00176-2 REGULAR PAPER Problems with research methods in medical device big data analytics Kenneth David Strang 1 Received: 18 August 2018 / Accepted: 23 January 2019 © Springer Nature Switzerland AG 2019 Abstract This paper reviews the literature as well as subject matter expert opinions and examines how research methods are being applied in medical device big data analytics. The focus of the study is to identify benefits and illustrate problems when applying certain research methods with healthcare big data. The intended audience is high-level healthcare decision makers, data science researchers, and healthcare big data practitioners. The key results address unintended access to healthcare data, statistical sampling violations with the use of healthcare big data, and the challenges associated with statistical false positives in big data. Solutions for these problems are proposed along with recommendations for further research. Keywords Healthcare big data · Big data research methods · Statistical techniques · Privacy 1 Introduction Big data analytics science is relatively new as compared to other disciplines because the paradigm was only formally recognized by 2011 [1–3]. Numerous manuscripts have been published about applying novel techniques for analyzing large data [4], but studies are generally not generalizable, and they do not address the arising big data analytics chal- lenges. Therefore, many researchers have called for more studies about big data analytics [1, 2, 5–8], and particularly in industries such as health care where a huge volume and velocity of information are generated daily [9, 10]. Approximately 47% of the big data analytics literature has focused on data mining, analytic models, cloud comput- ing, machine learning, social media, electronic data process- ing, algorithms, databases, map reduce, human behavior, as well as healthcare security and privacy [3, 11]. Only roughly 2% of the big data analytics literature has concentrated on exploring scientific research methods [3]. Not surprisingly, some authors have called for more studies about how to apply research techniques for big data analytics because conventional methods may not be practical or accurate [12]. The purpose of this study is to examine how common research methods are being applied in big data analytics. The objective is to highlight the benefits and shortcomings of contemporary big data analytics techniques. In particular, the goal is to identify and illustrate serious problems when applying conventional techniques in big data analytics, as well as to recommend solutions for these problematic issues. 2 Methodology The systems-thinking action research method was applied. The action research method starts by the researcher review- ing the literature either before or after the analysis, so as to validate or improve upon existing theories [13]. This sys- tems-thinking technique differs from the critical analysis method in that the latter attempts to find gaps or inaccura- cies in the literature using only the literature with deductive reasoning, but the former also collects practitioner or pro- cess data and attempts to find a solution to an institutional problem [13]. This method is ideal for examining the com- plicated hidden big data analytics problems in the healthcare discipline which is dominated by subject matter specialists and leading edge technology. Our research design is pragmatic, with a manuscript containing an introduction to the problem(s), methodol- ogy explanation, literature review, subject matter expert discussions, synthesis and assessment of data, recommen- dations to solve problem(s), conclusions and a reference * Kenneth David Strang ceo@multinations.org 1 School of Management and Technology, APPC Research and Walden University, 100 Washington Ave South, Suite 900, Minneapolis, MN 55401, USA