As many SAFTE-FAST users may know, our software is based on the Sleep, Activity, Fatigue, Task Effectiveness (SAFTE) biomathematical model invented by our Chief Scientist and President, Dr. Steve Hursh through efforts sponsored by the United States Department of Transportation (DOT) and Department of Defense (DOD). SAFTE is a three-process model of human fatigue and circadian variation that predicts changes in cognitive performance as a function of factors like prior sleep and time of day. SAFTE-FAST software solutions are a blend of the aforementioned SAFTE model with the Fatigue Avoidance Scheduling Tool (FAST), also co-developed by Dr. Hursh, and an amalgamation of scheduling tools, tweakable modeling parameters, and customizable features packaged into a user-friendly interface with premier customer service to make fatigue modeling more adaptable and accurate than ever.
However, what many SAFTE-FAST users and fatigue researchers may not know is that the original patent for the SAFTE model was filed while Dr. Hursh was an active-duty officer in the United States Army, meaning that the model has fallen under the ownership of the DOD since its invention in the early 2000s. Interestingly, the patent for the SAFTE model recently expired, meaning that use of the model is no longer protected by law. In other words, anyone can now make, use, or sell products that contain the SAFTE model without infringement. You are probably wondering why we would want to make this information known on the SAFTE-FAST blog of all places. Why would we want our customers to know that the SAFTE model is now free for use by anyone and is no longer under patent? Shouldn’t we be keeping that a secret?
The answer to that last question is absolutely not! In fact, we want MORE scientists, researchers, and academics to use the world’s most widely utilized biomathematical model as a part of their current, future, or post hoc research projects. We are confident in the model’s ability to inform future advancements to the state of sleep and circadian science. When researchers use the SAFTE model to explore new insights into fatigue, the whole field learns where the model retains accuracy and where a parameter needs to be tweaked. And since the model is out of patent, this means that we would be able to make improvements based on the latest research.
To make it easier for researchers to access the SAFTE model, we have created the SAFTEr package for R. For those of you who are not familiar with R, R is a GNU project language and environment for statistical computing, data manipulation, calculation and plotting that is free to use to the public. It is a simple and effective tool used by data scientists, researchers, and academic for data handling, data wrangling, statistical analysis, and creating publication-quality plots and graphics.
One of the benefits of the R language is that it is open source and free to use. Further, highly customizable extensions can be created and shared amongst the user community. These extensions, known as “Packages”, are standardized collections of code that can be shared through centralized repositories and hubs and loaded into a user’s environment. This makes it easy for colleagues and collaborators to share solutions to complex computations, allows others to build off of and incorporate other’s packages within their own package, and allows researchers to test reproducibility of findings and share methodology quickly and efficiently.
The SAFTEr package aims to put the complex equations from the SAFTE model into a researcher-friendly package with the benefit of coming from the team that is not only familiar with using the model in operational research, but is led by the creator of the model himself, Dr. Steve Hursh. SAFTEr will have a collection of functions that can take raw or log-based sleep data and create effectiveness predictions graphically or exported as tables for further analysis. The package will allow users to align and graph the independent variables alongside these effectiveness predictions to evaluate the impact of circadian rhythmicity or sleep history on an outcome of interest. This can be especially advantageous when investigating the effects of fatigue on other time-based observations, such as mood, caffeine intake, meal schedules, or drug dosing.
Some of the features of the SAFTEr package include:
Remapping of interval-based time-based data into minute-by-minute epochs of up to three weeks.
Creation of both tables and graphical representations of effectiveness over time
Customizable graphs with the ability to graph observations alongside SAFTE metrics like effectiveness or sleep reservoir
Identification of missing data
An algorithm to convert psychomotor vigilance task (PVT) scores to effectiveness scores
The SAFTEr package will not have many of the same features of SAFTE-FAST that our users have come to love. Many of SAFTE-FAST’s most popular features, like its ability to estimate light exposure based on date and geographic location, its Auto-Sleep algorithm that replaces the need for objective sleep data, its Workload estimator, or Insights are still proprietary. There is a major difference between being able to replicate the math presented in the original SAFTE patent and the ability to provide the full-service software solution that SAFTE-FAST has evolved into over the past two decades. SAFTEr is limited strictly to modeling cognitive effectiveness based on objective sleep/wake inputs. These inputs can be actigraphy, self-report- or even polysomnography from a laboratory study. We believe that the SAFTEr package is something that can be extremely useful to all scientists and researchers who require access to a validated fatigue model for their research projects but do not have the resources to license SAFTE-FAST or know-how to recreate the model from scratch.
IBR plans on releasing the initial version in SAFTEr this September (2023) on GitHub.