Shannon Holloway
Holloway
I purchased a new laptop last year - top of the line processors, more memory than I can use, and lighter than my first cell-phone. It had all of the bells and whistles that exist in the daydreams of a software developer. Sadly, it was relegated to checking e-mail. Why? Because I needed a Windows machine, and Microsoft was still releasing Windows 8. I am quite certain that the people at Microsoft are brilliant and have some incredibly creative ideas; but, when you wrap those ideas up in a package that is, well, let’s go with counter-intuitive, those ideas will never be embraced.
What does this have to do with IMPACT? It's my job to ensure that the methods and tools developed by the program project do not experience a similar fate. The Computational Resources and Dissemination Core, or what we refer to as Core B, is a group of investigators dedicated to fostering the adoption of the state-of-the-art methods developed for the program project.
During the initial grant cycle, our goals were simple: create a website (thank you for visiting!), and develop publicly-available software that implements many of the methods developed. In collaboration with investigators, more than 30 SAS, R, or C implementations are available. Though these are important steps forward, we realized that this model for the Core made some assumptions that simply were not valid.
First and foremost, we assumed that all of you were experienced Statisticians! For many of us (I am not a Statistician), the step from reading about a newly developed method in a journal article to using it appropriately in practice is HUGE. Journal articles are not about teaching; they assume that you know the field, its history, and its vocabulary. If you overcome that obstacle, you are confronted with the expectation that to use/implement the method, you must be intimately familiar with a specific statistical environment. How many of us have the time and energy for that?
Core B is aggressively working to fill this gap by developing interactive, web-based tutorials and demonstrations that teach both the methods and the available software implementations. These tools focus on teaching. As a first step, we are using R's Shiny package to develop tutorials for several treatment regime estimation methods developed by the program and implemented as R package DynTxRegime. In addition to introductory material, users can explore subtleties of these methods by using the software implementations under a range of conditions, without requiring fluency in statistical analysis or a foreign computational environment. For advanced users, these tools provide a simple means of “test driving” the software, exploring its capabilities and limitations.
Second, we assumed that you know about our current research! Yes, we publish. And, yes, we host regular symposia, workshops, and short courses to discuss research relevant to the program project. All of which are available on our website. But even I have to admit that finding information on our site is not always intuitive (we are working on that, too). So, unless you explore our site regularly, peruse the table of contents of the innumerable journals dedicated to statistical methods, or attend the appropriate session at JSM, you may not have heard about some of our recent achievements. For example, we have developed a method using decision lists to construct treatment regimes. And, we released a new SAS macro for fitting joint models for longitudinal and survival data (JMfit). What is the point of developing these tools if no one knows about them? We needed a communication hub that streamlined the information provided on the website, providing a more focused view of our program and its achievements. Thus, the IMPACT blog was started. It is an informal way for our researchers to introduce themselves, their research, and their ideas. Not to mention, we will provide periodic updates on new software releases/updates, publications, and event announcements.
A final initiative worth mentioning is our reproducible research repository. There is a growing movement in scientific studies that believes that all publications must be integrated with the tools used to obtain the published results - tools such as computer codes and data sets. Journals such as Nature have recently made fundamental policy changes to promote reproducibility; however, it has not been widely implemented by others. We believe that reproducibility is a fundamental component of making our methods accessible to the broader community. We are developing a website repository that will make available publications and the codes and data sets (when possible) used to obtain the published results.
We are a component of the program project specifically designed to address user needs. We are here for you. If you have any questions, suggestions, or requests for Core B, please drop us a line!
Comments/Discussion
| Submit a question or comment. |
Each month, one of our program investigators will introduce him/herself and will discuss their research, new research directions, or advancements made toward our goal of improving the clinical trial process. Readers are encouraged to send questions or comments. In addition, we will announce new software releases, publications, and upcoming events.
| Subscribe |
Date: February 15, 2016
Date: January 15, 2016
Date: September 17, 2015
Date: July 16, 2015
Date: June 4, 2015