ABSTRACT
This is probably the most technical chapter in the book and contains the few formulae that we’ve felt necessary to include. As elsewhere, the most statistical sections are marked with (*), as is the whole chapter because of this. We go through how to produce a risk-adjusted measure for each healthcare unit, which compares the unit with the benchmark. This might be post-operative mortality or unplanned admissions by general practitioner. For the purposes of illustration throughout this chapter, we’ll consider a binary outcome measure, specifically mortality, unless stated otherwise. Outcomes that can be assumed to have a roughly normal distribution, such as blood pressure, weight or patient experience scores, are often turned into binary measures such as “high” versus “normal”. If they aren’t, then things are a bit simpler, though many of the issues that we’ll discuss still apply.
