The last update of RSG is online, I have added heatmaps which can be useful when having a lot of data to plot. To test it I have simulated data for 100 000 eyes, over 60 000 with target 0 with a mean and standard deviation postop SE conditioned to preop SE and age and with one operator having results different to the others.

This is the SE preop vs postop plot:




That is a thick cloud of points. Let's turn the heatmap on:




We can see now the points density, being green low density and red high density, we can also see that the points fall in fixed SE postop values shown by the red lines, also that preop there are more within -6 and -1 and that there are much less around 0 SE preop, as expected. At the bottom left we see that there are 62 239 eyes represented.

The following graph shows the graphic separated by doctor:




Well, it seems that Dr. E red cloud is less intense and larger, indicating a greater variability in the outcomes (it could also be that there are much less eyes than the others, but here all have the same, that can be easily checked in the histograms tab). It is also a bit lower than the rest so the mean SE is also lower that the others.

You can test it here.




You may have been in a situation where you have to do an eye test with projectors in different rooms and you think that people in one room has better visual acuity than in the other. Or you may bring a chart from another room that was designed for that room distance but you don't have that space.

Well, with this calculator you can check your optotypes or adapt them to other rooms without having to do yourself all the calculations: 


In the ophthalmoinnovations blog I read a post where they wanted to know if the aspheric laser treatments where really aspheric so they put it to the test. The result was yes but to claim this they only give the mean Q values for the planoscan - not aspheric treament - group and the aspheric treatment group. I asked J.R.Villada, the author, if I could have a look at the data to see it better with graphs and he kindly agreed. So I had two set of Q values, one for the planoscan and one for the aspheric treatments with Q values preop and postop. Now first things first, let's look at the distributions:




There are 97 eyes for the aspheric treatment and 75 eyes for the planoscan, the distributions are reasonably normal, although the postops after the aspheric treatment distribution is a bit right skewed. Let's make some boxplots to see both groups:




Well it seems indeed that the aspheric treatment is doing its job making the corneas less oblates. A t-test shows that the preop patients group means are not significantly different, means -0.18 and -0.17 and p value of 0.45, but after the treatment the means are significant, 0.10 and 0.31 and with a p value of 2.3e-7 or 0.00000023, virtually zero.