Power and sample size are determined by numerous factors depending on the question under study. You could have N=10 be statistically powerful and N=1,000,000 be statistically meaningless. It depends ENTIRELY on the subject under study. More is not always better and in many cases completely unnecessary.
https://sphweb.bumc.bu.edu/otlt/MPH-Modules/BS/BS704_Power/B...
Here is a decent guide. You may wish to read it so you can learn when you can dismiss a study based on N. You never mentioned, did you back out the study numbers and determine the N=17 number was outside the range of statistical power? If so, would you mind posting your calculations?