NOW FOR AN ORIGINAL PAPER ASSIGNMENT:Persuasive Nature of Case Material

3.2. Research approach and process

We applied content analysis to gain insights from the cases collected. Content analysis is a method for extracting various themes and topics from text, and it can be understood as, “an empirically grounded meth- od, exploratory in process, and predictive or inferential in intent.” Spe- cifically, this study followed inductive content analysis, because the knowledge about big data implementation in health care is fragmented (Raghupathi and Raghupathi, 2014). A three-phase research process for inductive content analysis (i.e., preparation, organizing, and reporting) suggested by Elo and Kyngäs (2008) was performed in order to ensure a better understanding of big data analytics capabilities and benefits in the healthcare context.

The preparation phase starts with selecting the “themes” (informa- tive and persuasive nature of case material), which can be sentences, paragraphs, or a portion of a page (Elo and Kyngäs, 2008). For this study, themes from casematerials were captured by a senior consultant who has over 15 years working experience with a multinational tech- nology and consulting corporation headquartered in the United States, and currently is involved in several big data analytics projects. The senior consultant manually highlighted the textual contents that completely describe how a big data analytics solution and its function- alities create the big-data-enabled IT capabilities and potential benefits while reading through all 26 big data cases for a couple of times. Subse- quently, a total of 136 statements directly related to the IT capabilities and 179 statements related to the potential benefits were obtained and recorded in a Microsoft Excel spreadsheet.

The second phase is to organize the qualitative data emerged from phase one through open coding, creating categories and

abstraction (Elo and Kyngäs, 2008). In the process of open coding, the 136 statements were analyzed by one of the authors, and then grouped into preliminary conceptual themes based on their similar- ities. The purpose is to reduce the number of categories by collapsing those that are similar into broader higher order generic categories (Burnard, 1991; Dey, 1993; Downe-Wamboldt, 1992).