R (www.r-project.org). It's free, has a large community support and you find that it requires you to think about the questions you are asking the data instead of just applying "out-of-the-box" testing. And yes, you need to spend the time understanding the statistic reasoning, the conditions under which a test applies, or even better start moving from inferential statistics to statistical modelling. Enjoy!
The most intuitive program I have ever used for biomedical statistics and plotting of graphs is GraphPad Prism. It is available for both Mac and Windows.
A more powerful and not quite as intuitive program is Origin Pro. It is available only for Windows, but it can be run in a VM on linux or Mac.
Statistica is menu-driven and very comprehensive, but the graphing is not as good as Origin Pro or Prism.
JMP and JMP Pro are very nice for data exploration and visualization. Much more intuitive than SAS. They are available for Mac and Windows.
SAS is a giant and has been a "gold standard", but I think its popularity is waning while the popularity of R is growing.
You could also take a look at Stata (somewhat like R; cross-platform; steep learning curve) and SPSS (menu driven; cross-platform).
Finally, there is R, which has a very steep learning curve, but its capabilities are vast and continually growing and improving. It is also free and cross-platform.
There is no such term "The best statistical package" in biomedical sciences. I believe each statistical package has its own pros and cons. In academics , people prefer to use " R" whereas in pharmaceuticals, "SAS" is in demand. It is always an added advantage to know more than one language . If you have any background in coding then MATLAB , R, SAS all are good but you can also go for GUI interface softwares like SPSS, Rattle R and Rstudio if you are not comfortable with coding. I hope this helps.