StatFlex

It has been 33 years and 10 months since the first edition of StatFlex was released. Recently, users have requested that we officially disclose who developed it, through what process, and how the analysis results are verified as mathematical software. In practice, the person who has consistently overseen its development is myself, Kiyoshi Ichihara, currently affiliated with Yamaguchi University, with the background detailed below. The development of a large-scale software like StatFlex was a project I undertook as a side business while being involved in education and research as a university faculty member. This required a significant amount of funding, making it essential to distribute StatFlex as a paid product. Therefore, I funded the production costs out of my own pocket and requested an acquaintance's company to handle the sales. Although I launched a university-born venture in 2003, I did not explicitly clarify my involvement because of my awareness that it was a side project. Fortunately, having retired and gained a more independent position, I founded the Medical Statistics Laboratory Co., Ltd. in 2018. I have taken this opportunity to clearly identify the developing entity of StatFlex, disclose the development process to date, and also open the verification process of the statistical analysis results to the public. Through this, I hope that everyone using StatFlex will be able to publish their analysis results without hesitation. Another question that might arise regarding the development process of StatFlex is that my area of expertise is medicine—specifically laboratory medicine, clinical epidemiology, and clinical endocrinology—and looking at my background, I lack an academic foundation in statistics and information technology, which are required for developing StatFlex. In fact, my skills in both fields were entirely self-taught by thoroughly reading Western books. Achieving this was largely due to my ability to acquire knowledge in English, a skill I developed during my university days. In the summer of 1970, as a second-year medical student at Yamaguchi University, I traveled around North America for 80 days by bus and hitchhiking, and was overwhelmed by the scale of the US highway networks, skyscrapers, and shopping malls. I deeply felt that for Japan to compete, the first step had to be improving our English proficiency. Therefore, I decided to self-study medicine as much as possible using Western books, and mostly achieved that goal by graduation. After graduating from Yamaguchi University in 1975, I entered the Graduate School of Medicine (Laboratory Medicine) at Osaka University, where I immediately realized that statistics is indispensable for clinical research. Soon after, I encountered two statistics books that made a profound impression on me: Introductory Statistics by T.H. Wonnacott and Nonparametric Statistics for the Behavioral Sciences by S. Siegel. Following this, I read through all major statistics books in English, started receiving requests for statistical processing, and held on-campus seminars. By 1982, I was organizing nationwide intensive statistics courses. On the other hand, the trigger for learning the programming skills needed to develop StatFlex came right after entering graduate school in 1975, when a request was made to our laboratory to test a costly programmable calculator (2,000 steps). That was my first encounter with computers, but I read through several thick English instruction manuals within a few weeks and understood what could be done. I immediately developed my own programs for regression lines and logistic curve regression, which were necessary for laboratory work. Given this background, combined with the fact that I was good at mathematics in high school, I believe that self-study was possible because the mathematical logic of statistics and programming techniques naturally suited my thought process.