Several years ago, I was one of few Luddites in my social circle still using a “dumb” phone. Other than calls, and (text only!) messaging, it had no other functionality. Once I made the switch, a whole new magical world unveiled itself. With the tap of a button, I could read my email, listen to a podcast, find a nearby restaurant, order a cab, snap a photo, and so on. What was once considered science fiction seems an indispensable part of life today.
A key task in a biostatistician's life is to designs studies using the appropriate sample size. At the time of writing, if one searches “free sample size calculator” in Apple’s App Store (which has an estimated 2 million+ apps available), you get three results EastLite, E-Valuate and WhatStat.
All three tools offer their own unique features, but there is also common functionality. [Full disclosure: I was the product manager for Cytel’s East Lite app.]
East Lite includes computations for one and two-sample designs, for normal, binomial, and survival endpoints, and superiority or noninferiority trials.
At Cytel, we are known for tacking complex adaptive methods, but very often the starting point for any trial is to build on something basic, such as a conventional two-arm design with a single endpoint. I decided to run the same simple scenario on all three apps: What is the sample size required to compare two proportions (0.2 vs 0.1) with alpha at 5%, to achieve a power of 80%? East Lite gives a total sample size of 398 patients, identical to that obtained from the desktop version of East. The E-valuate and WhatStat apps give similar answers of 404 total patients (202 per arm), and 392 total patients (196 per arm), respectively.
So, if you’re a statistician armed with an iPhone, and you ever need to calculate a simple sample size on the run, these gifts to community are certainly worth checking out. Meanwhile, I’m going to confirm their calculations using pen & paper…
To download East Lite on the Apple App Store click the button below.
About the Author
Charles Liu, PhD, is a Senior Product Manager at Cytel. Before joining Cytel in 2012, he was a postdoctoral fellow at Boston University. His current research interests include: Enrollment & events prediction, Go/No-Go decision making, and Program & Portfolio optimization.