We took the definitions from the Mandela rules. We gave what we would consider to be a very conservative threshold for defining something as either solitary confinement or torture, so our numbers are probably lower than most people would have put them. We then looked at how many prisoners were in each of these regions and the proportion or the rate per thousand prisoners for this kind of treatment. We did find, as you already said in your question, enormous variability across regions. I think it's important.
Remember that the terms “solitary confinement” and “torture” are in effect technical definitions drawn from the Mandela rules. Here Jane Sprott and I are, criminologists talking about torture rates in different regions in our country. We found it rather disturbing that when we looked to see whether anybody had calculated torture rates in western countries, we weren't able to find them.