Design of an internet course for training medical researchers in Bayesian statistical methods

Abstract

Access to statistical information is at an all-time high, and the information age is fuelling this access at an extraordinary pace. This access increases the capacity for medical researchers to use statistics to guide decision making, yet few courses teach methods to do so. Rarely does statistics training include methods for incorporating statistical output into decision making. Mass education and educational reform is needed. Technological advances of the past decade make this goal possible, and allow us to dramatically change how we use, teach, and think about statistics. This paper covers the conceptual development of an Internet continuing-education course designed to teach the basics the Bayesian statistics to medical researchers. Two questions are discussed: Why the Internet, and why the Bayesian paradigm?

Learning from experience

“…On the other hand some statisticians recommend that all of test results, confidence intervals, and power analysis or estimation of effects should be presented. Others favour using a Bayesian approach in inference, for example see Lecoutre (1999) who gives arguments in support of this with many references, and Stangl (2001) who discusses why medical researchers should be trained in Bayesian methods.…”

“…Most statisticians would prefer to be collaborators and joint authors rather than thanked in a foot-note if their contribution is at all substantial, but the survey results reported by Godino et al (2001) suggest that it will be some time before statisticians are readily accepted as joint authors. In all cases statisticians must be given the opportunity to veto any incorrect statistical work or misunderstandings on the part of the researchers, and mistakes can occur as Stangl (2001) comments. Saying that a statistician was involved does not give work statistical credibility.…”

Statistical thinking for novice researchers in the biological sciences

“…The other major issue that arises from this project is that of using Internet or Intranet facilities. Stangl (2001) lists four interrelated reasons that increase the chance of success of the Internet as a teaching tool over other technologies such as video and CD-ROM. One of these is interactivity and she points out its drawbacks of complexity and expense in developing and delivering the materials and of human factors such as time and the feeling of information overload.…”

Training of statisticians and clinical researchers worldwide to collaborate as co-investigators within country clinical epidemiology units: the experience of the international clinical epidemiology network (INCLEN)

“…As part of the continuing training in the future, web based distance learning courses similar to those presented by Lee (2001) and Stangl (2001), but using 'case studies,' are currently being developed by the regional training centres. Researchers will be able to post problems on a web site.…”